Você está na página 1de 552


were substantially attenuated, mainly due to inclusion of the

constit utional susceptibility score, but remained significant
in each car.egory for sec and in the highest. category for
melanoma. Sunlamp usage or tanning salon attendance (ever
vs never) was a risk factor for melanoma, and the association
remained significant in the multivariate models. There was no
apparent dose- response relation between the frequency of
usage and melanoma risk with multivariate ORs of 2. 06 (95%
CJ 1.1 5- 3.68) for less than 10 times and 2.05 (95% Cl
1.08- 3.90) for 10 or more times. There were non-significant
associations of sunlamp usage or tanning salon attendance with
increased risks of SCC and BCC. Cumulative sun exposure
while wearing a bathing suit remained a significant risk factor
for the three r.ypes of skin cancers in the multivariate models,
aJld the multivariate ORs were simi lar to those iJl age-adjusted
models. This variable signi fica ntly increased the goodness-of-fit
of the model for three cancer types. Among controls, women
who had higher cumulative sun exposure while wearing a
bathing suit were more likely to use a sunlamp or attend a
tanning salon (P. Chi-square, 0.03). Women in the West and
South regions were significantly more likely to be diagnosed
with SCC or BCC compared with those in Nort heast, but
residence in the West and South was not associated with
melanoma risk in both age-adj usted and multivariate models.
OveralL there was no statistically significant difference in risks
associated with these variables for each type of skin cancer in
multivariate models, according to the heterogeneity test.
Interaction between constitutional susceptibility
score and sun exposure while wearing a
bathing suit
We eval uated interactions between the consti tutional suscep-
tibility score and sun exposure while wearing a bathing suit on
skin cancer risk. We observed a significant interaction for
melanoma risk (P, interaction, 0.03) (Table 3). Compared wi th
women with the lowest consti tutional susceptibility score and
the lowest level of sun exposure while wearing a bathing suit,
those with the highest constitutional susceptibility score and
the highest level of sun exposure while wearing a bathing suit
had a significantly increased risk of melanoma (OR, 8.37; 95%
cr 3.07- 22.84) . Controls with high constitutional susceptibility
were less likely to have prolonged su n exposure while wearing
a bathing suit compared witll those with low consti tutional
susceptibill ty (P, Chi-square, 0.04) . No statisticall y significant
interactions were found for the risks of SCC (P, interaction,
0.52) and BCC (P, interaction, 0.20).
Assessment of recall bias
We had the opportunity to examine recall bias as we had
prospectively and retrospectively obtained questionnaire data
for a subset of variables (Table 4). OveralL the reliability of the
responses on natural hair colour and childhood or adolescence
tendency to ran was high. whereas the reliability of childhood
or adolescence tendency to burn assessment was lower. We
found no notable difference in the reproducibility correlations
for the three variables between cases and controls. As shown in
Table 4, the magnit ude of absol ute shift was similar among
cases and controls, except for tendency to burn among sec and
BCC cases, for which cases retrospectively reported increased
Table 3 Interaction between constituti onal susceptibility score and
sun exposure with a bathing suit on melanoma risk
Sun exposure with a bathing s uit (tertile)
score (tertile)
Low Intermediate High
Cases (%) 5 (20.0) 12 (48.0) 8 (32.0)
Controls (%) 62 (27.9) 70 (3 1.5) 90 (40.5)
OR (95% Cl) 1.00 1.92 0.97
(0.63- 5. 90) (0.30-3. 16)
Cases (%) 11 (21.2) 12 (23.1) 29 (55.8)
Controls (%) 78 (33.3) 86 (36.8) 70 (29.9)
OR (95 % C!) 1.73 1.39 4.13
(0.56- 5.32) (0.46-4.23) (1.47- J J .61)
Cases (%) 21 (20.4) 23 (22.3) 59 (57.3)
Controls (%) 87 (38.3) 72 (31.7) 68 (30.0)
OR (95% CI) 2 65 3.02 8.>7
(0.93-7.60) ( 1.06-8.62) (3.07-22.84)
P. interaction. 0.03.
Unconditional logistic regression adjusted for age, family his10ry of skin
cancer. rhe number of lilerime severe sunburns which blistered (none, 1- 5,
6- tl , > tt). sunlamp use or tanning salon attendance (yes/no), and geo-
graphic region. The percentages may not sum to 100 dut' to rounding.
tendency to burn to a greater extent compared with controls.
As a result, the ORs for tendency to burn based on retrospect ive
data were relatively larger than those on prospective data for
sec and BCC.
We examined the associations of constituti onal risk factors and
sun exposure and their interactions with the risks of the three
types of skin cancer simultaneously in a nested case-comrol
st udy with in the Nurses' Hea lth Study (NHS) cohort. The risks
associated with the constitutional suscepti bility score slightly
changed but remained signi fica nt in mu ltivariate models
controlling for other exposure variables. This suggests that
the constitutional susceptibility is an independent risk factor for
all three types of skin cancer.
Sunburn at any age has been shown to be associated with an
increased risk of melanoma


The lifetime sunburn
variable combines exposure intensity and biological response
to sun exposure. We observed significant associations of the
number of severe sunburns with three types of skin cancer in
the age-adjusted models. The attenuation of the associations in
the multiva riate models ind icated that the skin cancer r isk
attributed to severe sunburns was partially explained by other
variables, particularly the constitutional susceptibility score.
However. even though attenuated, this association remained
significant in the multivariate models for melanoma and sec.
suggesting the number of lifetime severe sunburns may be an
independent risk factor.
The usage of indoor tanning devices was previously
associated with an increased risk of melanoma in several

Even though most of the studies, including ours,
performed retrospective assessment, it was reported recently
Table 4 Assessment of recall bias"
Skin tan after Skin reaction to 2
repeated s un exposure or more hours of
in childhood or s unlight in childhood
adolescence or adolescence
Natural hair colour (tendency to tan) (tendency to burn)
Kappa statistics (11) Kappa statistics (11) Kappa statistics (11)
Melanoma cases 0.84 (178) 0.61 (17 1) 0.45 ( 179)
sec cases 0.82 (244) 0.65 (238) 0.44 (243)
BCC cases 0.83 (245) 0.66 (242) 0.40 (242)
Cont rols 0.81 (682) 0.61 (668) 0.42 (679)
Mean change (SE)c Mean change (SE) Mean change (SE)
Melanoma cases 0.10 (0.03) -0.24 (0.05) 0.11 (0.07)
sec cases 0.03 (0.03) -0.1 8 (0.04) 0.21 (0.06)
BCC cases 0.09 (0.03) -0.19 (0.04) 0.26 (0.06)
Cont rols 0.04 (0.02) -0.23 (0.02) 0.09 (0.03)
Age-adjusted OR (95%Cl)d Age-adjusted OR (95%CJ) Age-adjust ed OR (95%CI)
Melanoma cases
p 2.08 (1.37-3.15) 0.55 (0. 39-0.77) 1.94 (1.38-2.71)
R 2.27 (1.53-3.35) 0.58 (0.41- 0.81) 2. 17 (1.54-3.04)
sec cases
1.79 (1.23-2.60) 0.61 (0.45-0.83) 1.64 ( 1.22-2.21)
R l.70 (1.18-2.44) 0.62 (0.46-0.84) 2.04 (1.51 -2.74)
BCC cases
1.73 (1.19-2.53) 0.68 (0.50- 0.93) ]. 55 ( 1.1 5-209)
R 1.86 ( 1.30-2.66) 0.7'5 (0.5'5-1 .00) 2.26 (1.68-3.06)
Hair colour was scored on a 5-point scale. where 1 was black and 5 was red. Tendenry to tan was scored on a 4-point scale. where 1 was practically none
and 4 was deep tan. Tendency to burn was scored on a 5-point scale. where I was practically none and 5 was painful burn with blisters.
b Kappa statistics of the responses 10 the three questions asked before and after skin cancer diagnosis in cases and in matched controls.
c Mean change was calculated as the retrospective questionnaire score minus the 1982 questionnai re score. Numbers in parentheses, standard error ol the
d Age-adjusted OR was calculated as follows: hair colour: blonde or red vs black. dark brown. or light brown; tendency to tan: average or deep tan vs
practically none or light ta n; skin reaction: burn, painrul bum, or painful burn with blisters vs practically none or some redness only. P stands for ORs based on
the 1982 prospective data; R stands for ORs based on the 2000 retrospcctiw data.
that there was substantial reliability in reporting the usc of
sunlamps after melanoma diagnosis
2 5
A prospective study
showed that tanning device use was a significant risk factor for
melanoma with an OR for use more than once/month during
age 10-39 of 1.55 (95% CI 1.04-2.32).
Only a few studies
have evaluated the relationship between tanning device usc
and non-melanoma skin cancer. One population-based case-
cont rol study reported a significantly positive associat ion; the
relative risks were 2.5 (95% CI 1.7-3.8) for sec and 1.5 (95%
CI 1.1- 2.1) for BCC.
We simultaneously evaluated the
sunlamp use or tanning salon attendance in relation to the
three types of skin cancer. The association was strongest and
significant for melanoma compared with SCC and BCC. Most of
the previous studies only adjusted for pigmentation and
phenotype factors. In this study. after additionally controlling
for cumulative sun exposure while wearing a bathing suit,
lifetime severe sunburns. family history of skin cancer. and
geographic region. at baseli ne, the associations did not change
substantially and remained significant for melanoma risk. These
data suggest that the risks associated with sunlamp use were
not likely to be substantially confounded by sun exposure of
other kinds.
The ratio of UVB to UVA emitted by indoor tanning devices
was greatly reduced around 1980.


We did not differ-
ent iate age category or calendar year of the usage of indoor
tanning devices on the questionnaire. Because the age of our
study population at baseline (1976) ranged from 30 to 55, it is
possible that the majority in this study was of older
UVB-emitting devices. However, UV A has a carcinogenic effect
by causing oxidative DNA damage via reactive oxygen species
generated after the absorption of light energy by cellular

Additional studies are warranted to evaluate
the effect of the more contemporary UVA-emitting devices.
We used cumulati ve sun exposure whiJe wearing a bathing
suit as a measurement of recreational and intermi ttent sun
exposure; it was associated with all three types of skin cancer in
our study with the strongest risk for melanoma. The multi-
variate ORs for cumulative sun exposure while wearing
a bathing suit were not substantially confounded by other
Residence in geographic regions can be viewed as a rough
esti mation of chronic sun exposure. Vve observed that
residence in West and South regions was associated with
increased risks of sec and BCC, but not that of melanoma.
consistent with the descriptive epidemiological evidence of a
stronger North-South gradient in the US for the risks of SCC
and BCC than that of melanoma
We observed a significant interaction on a mu ltiplicative
scale between the constitutional susceptibili ty score and sun
exposure while wearing a bathing suit on melanoma risk.
Women with the highest constitutional susceptibility score
and the highest level of sun exposure while wearing a bathing
suit had the highest risk for melanoma. Among controls, there
was evidence of a 'phenotype-behavior' feedback, i.e. the
controls who were more constitutionall y susceptible to sun
exposure had less sun exposure while wearing a bathing suit
compared with those who were less susceptible. In this study,
constitutional susceptibility was the combination of hair colour,
skin colour, childhood tendency to burn, and mole counts.
These identifiable phenotypic phenomena may make people
aware of their susceptibility, resulting in reduced recreat ional
sun exposure.
Few studies examined melanoma risk factors prospec-
tively3132 For retrospective studies. information on sunlight
exposure and skin cancer risk factors is potentially subject to
recall bias as it was gathered after the onset of disease. We
assessed potent ial recall bias by examining the correlations and
the difference in mean changes between the responses on the
prospective and retrospective questionnaires for the three
questions on constitutional factors and comparing odds ratios
calculated for these variables
The reliability of each measure
was approximately the same magnitude among the cases and
the controls and the odds ratios based on the prospective and
retrospect ive questions were similar, except for childhood
and adolescence tendency to bum, which was slightly over-
reported among sec and BCC cases retrospectively. These data
indicated tha t the retrospective assessment was not li kely to
substantiall y bias the estimate of risk in this study, at least for
these variables. Weinstock et a/.
examined recall bias in 143
melanoma cases with the diagnosis between June 1976 and
June 1984 in a nested case-control study conducted in 1984
and 1986 within the NHS. and the authors observed recall bias
in retrospective assessment of abili ty to tan. but not that of hair
colour. In this study, we collected the retrospective question-
nai res in 2002 among 200 melanoma cases who provided blood
samples in 1989 and 1990 and had the diagnosis between June
1990 and June 2000. We did not observe substantial recall bias
for the three variables among melanoma cases. The different
design of the two studies may help explain the discrepancy of
the results.
In summary. the nested case-control design. high follow-up
rate, and high response rate for the retrospective supplemen-
tary questiOimaire strengthen the validity of this study. The
li mitations of the study include self- reported assessment on
pigmentation phenotypes and exposures, which may lead to
misclassification. There is potential limitation in generalizability
of the results in our cohort of nurses, e.g. outdoor occupa-
tions are underrepresented. We observed sunlamp use or
tanning salon attendance remained a significant risk factor for
melanoma in multivariate models. The cumulative sun
exposure while wearing a bathing suit was an independent
risk factor from constitutional susceptibility and other exposure
variables. We observed a significant interact ion between
constitutional susceptibility and sun exposure while wearing
a bathing suit on melanoma risk, suggesting that the inter-
actions between host factors and sun exposure provide useful
itlformation for skin cancer prevention.
This work is supported by NIH CA87969 and CA 113100, and
the Harvard SPORE in Skin Cancer. We thank Gary Chase and
Nicole Williams for their help in supplementary questionnaire
admhlistration and data entry. We also thank Carolyn Guo for
her progranuning support.
Ameri can Cancer Society. Cancer Facts and Figures. Atlanta: American
cancer society, 2004.
Howe HL Wingo PA, Thun MJ eta/. Annual report to the nation on
the status of cancer (1973 through 1998), featuring cancers with
recent increasi ng trends. J Nat/ Cancer lrrst 2001;93:824-42.
AmtStrong BK, Kricker A, English DR. Sun exposure and skin cancer.
Aust J Dennato/ 1997;38:S 1-6.
English DR, Armstrong BK, Kricker A, Fleming C. Sunlight and
cancer. Cancer Causes Control !997;8:271-83.
Brash DE. Sunlight and the onset of skin cancer. Trends Genet
1997; 13:410-14.
Naldi L, Altieri A, lmberti GL, Gal lus S. Bosetti C. La Vecchia C. Sun
exposure, phenotypic characteristics, and cutaneous malignant
melanoma. A.n analysis according ro diUerent clinico-pathological
variants and anatomic locations (Italy). Cancer Causes Control
2005; 16:893-99.
Naldl L, Altieri A, Imberti GL, Giordano l, Gallus S, La Vecchia C.
Cutaneous malignant melanoma in women. Phenotypic character-
istics, sun exposure. and hormonal factors: a case-control study from
llaly. Ann Epidemic/ 2005; 15:545-50.
MacKie RM. Tnddence, risk fa ctors and prevention of melanoma. Eur
J Cancer 1998;34:$3- 6.
Osterlind A. Fair complexion and risk of malignant melanoma of the
skin based on experience from the Danish Cancer Registry. Recent
Results Cancer Res 1993; 128:91-100.
to Masback A, Westerdahl J, Ingvar C, Olsson H. Jonsson N. Cl inical
and histopathological characteristics in relation to actiological risk
factors in cutaneous melanoma: a population-bast>d study. Melanoma
Res 1999;9: 189-97.
Autier P, Dore JF, Gefeller 0 et a/. Melanoma risk and residence in
sunny areas. EORTC Melanoma Co-operative Group. European
Organization for Research and Treatment of Cancer. Br J Cancer
1997;76: 1521-24.
Dwyer T, Blizzard L, Ashbolt R, Plumb J, Berwick M, Stankovich JM.
Cutaneous melanin density of Caucasians measured by spectro
photometry and risk of malignant melanoma, basal cell carcinoma,
and squamous cell carcinoma of the ski n. Am J Eipidemio/
2002; 155:614-21.
Rosso S, Zanetti R. Pippione M, Sancho-Gamier H. Parallel risk
assessment of melanoma and basal cell carcinoma: skin character-
istics and sun exposure. Melanoma Res 1998;8:573- 83.
Elwood JM, Jopson J. Melanoma and sun exposure: an overview of
published studies. 1111 .J Cancer 1997;73: 198-203.
Armstrong BK. Kricker A. The epidemiology of UV induced skin
cancer. J Pltotochem Photobiol B 2001; 63:8- 18.
tt, Colditz GA. Martin P, Stampfer MJ et a/. Validation of questionnaire
info rmation on risk factors and disease outcomes in a prospective
cohort study of women. Am .l Epidemiol 1986;123:894-900.
NaLional Oceanic and Atmospheric Administration. Climatic Atlas of
the United States, 1983.
Mieninen OS. Stratification by a muhivariate confounder score. Am J
Epidemic/ 1976; 104:609- 20.
Marshall RJ, Ch ish olm EM. Hypothesis testing in the polychotomous
logistic model wi th an application to detecli ng gastroimesti nal
cancer. Stat Med 1985; 4:337-44.
Weinstock MA, Colditz GA, Willett WC, Stampfer MJ, Rosner B,
Speizer FE. Recall (report) bias a nd rel iability in th e retrospective
assessmem of melanoma risk. Am J Epidemic! 1991; 133:240-45.
t Veierod MB, Weiderpass E, Thorn M et a!. A prospective study of
pigmentation, sun exposure, and risk o f cutaneous malignant
melanoma in women. J Nat/ Cancer Tnst 2003;95: 1530- 38.
Holly EA. Aston DA, Cress RD, Ahn DK, Kristiansen .J.J. Cuta neous
melanoma in women. I. Exposure to sunlight, abilit y to ta n, and
other r isk factors related to uhraviolct light. Am J Epidemic!
1995; 141:923- 33.
Young AR. T.anning devices- fast track to skin cancer? Pigment Cell Res
2004; 17:2-9.
Gallagh er RP, Spinell i JJ, Lee TK. Ta nn ing beds. sunlamps, and risk
of cutaneous malignant melanoma. Cancer Epidemic/ Biomarkers Prev
2005; 14:562-66.
Beane Freeman LE. Dennis LK, Lynch CF, Lowe .JB, Clarke WR.
Test-retest of seH-reported exposure 10 a rtificial tanning devices,
self-tanning creams. and sun sensitivity showed consistency. J Clin
Epidemic/ 2005;58:430- 32.
r<aragas MR, Stann ard VA. Mott lA, Slattery MJ, Spencer SK,
Weinstock MA. Use of ta nn ing devices and risk of basal cell and
squamous cell skin cancers. J Nat! Ca11cer lnst 2002; 94:224-26.
Spencer JM, Amonette R. Tanning beds and skin cancer: artificial
sun + old sol = real risk. Clin Dermato/ 1998; 16:487- 50 I.
Lazovich D, Sweeney C, Weinstock MA, Berwick M. Re: a prospective
study of pigmenta tion, sun exposure. and risk of cuta neous
malignant melanoma in women. J Nat/ Cancer lmr 2004; 96:335;
author reply 7-8.
r<i elbassa C. Roza L, Epe B. Wavelength dependence of oxida tive
DNA damage induced by UV a nd visible light. Carcincgemsis
1997; 18:811- 16.
Scotto J. Fears TR. F FJ. Incidence of non-melanoma skin cancer in
t he United States. NIH Pub. No. 83-2433, 1983.
3 1
Schneide r .JS, Moore DH 2nd, Sagebiel RW. Risk factors for
melanoma incidence in prospective follow- up. The importance o f
atypical (dysplastic) nevi. Arch Dermatcl l 994; 130:1002-07.
Ch o E, Rosner BA, Feskanich D. Colditz. GA. Risk factors and
individual probabilities of melanoma for whites. J CTin Oncol
2005; 23:2669-75.
Annotating PDFs for eReturn
version 1.0: August 28, 2007; ful l release
1. Introduction
eProof files are self-contained PDF documents for viewing on-screen and for printing. They contain all appropriate formatting
and fonts to ensure correct rendering on-screen and when printing hardcopy. DJS sends eProofs that can be viewed, annotat-
ed, and printed using the free version of Acrobat Reader 7 (or greater). These eProofs are "enabled" with commenting rights,
therefore they can be modified by using special markup tools in Acrobat Reader that are not normally available unless using
the Standard or Professional version.
The screen images in this document were captured on a PC running Adobe Acrobat Reader version 8.1.0. Though some of the
images may differ in appearance from your platform/version, basic functionality remains similar. At the time of this writing, Ac-
robat Reader v8.1.0 is freely available and can be downloaded from: http://www.adobe.com/products/acrobat/readstep2.html
2. Comment & Markup toolbar functionality
Comment & Markup
- Sticky Note
A B c D E F G H
A. Sticky Note tool; B. Text Edits tool; C. Stamp tool; D. Highlight Text tool; E. Callout tool; F. Text Box tool ; G. Various Obj ect tools; H. Pencil tool
A. Show the Comment & Markup toolbar
The Comment & Markup tool bar doesn't appear by default. Do one of
the following:
Select View > Toolbars > Comment & Markup.
Select Tools > Comment & Markup > Show Comment
& Markup Toolbar.
Click the Review & Comment button in the Task toolbar, and choose
Show Comment & Markup Toolbar.
To add or remove tools for this too/bar, right-click the too/bar and select the tool.
Or, select Tools > Customize Too/bars.
B. Select a commenting or markup tool
Do one of the foll owing:
Select a tool from the Comment & Markup toolbar.
Select Tools > Comment & Markup > [tool].
Note: After an initial comment is made. the tool changes back to the Select tool
so that the comment can be moved, resized, or edited. (The Pencil, Highlight
Text, and Line tools stay selected.)
C. Keep a commenting tool selected
Multiple comments can be added without reselecting the tool.
Select the tool to use {but don't use it yet).
Select View > Tool bars > Properties Bar.
Select Keep Tool Selected.
More Tools [g)
0 Convnent e.: Markup Toolbat"'
0 -jii/Sh:ky Note *
< I
0 Edits
D tl! stei'Jl)Tool "'
0 .ifl HiQhfo;jht Text Tool *
0 Text Tool
0 <{] cross O<A Text Tool*
0 File as Comment
0 (' Record lwdo Conmerl *
0 E\Calout Tool*

O O cloodTool
0 )" Arcow Tool
0 / tine Tool*
0 0 Rectangle Tool "
O O ovai Tool *
0 <;' PolyQon line Tool
0 0 Polygon Tool
0 / IPeroci Tool
0 # Pendl Er&Ser Tool
0 ,.:<Show*
g !Jsend
* Only available when doC\lment ri4"lts <tre enabled.
OK j I Cancel
Choose Tools > Customize Too/bars to remove
unnecessary items from the too/bar
(see Section 7 for suggested too/bar layout)
3. The Properties bar
The Properties bar can be used to format text and select options for individual tools.
To view the Properties bar, do one of the following:
Choose View > Toolbars > Properties Bar.
Pop-Up Text Properties
Right-click the toolbar area; choose Properties Bar.
Select [Ctri-E]

4. Using the comment and markup tools
To insert, delete, or replace text, use the Text Edits tool. Select the Text Edits tool, then select the text with the cursor (or
simply position it) and begin typing. A pop-up note will appear based upon the modification (e.g. , inserted text, replacement text,
etc.). Use the Properties bar to format text in pop-up notes. A pop-up note can be minimized by selecting theGJ button inside it.
Sticky Note t Text Edits -
~ =::: Author; klawson I More .. .
induced angiogenesis.
Tumor-induced <mgiogenesis is the fonnation of new b
vessels from existing vasculature in response to
signals from a tumor. Angiogenesis marks
tion from benign solid tumor gro vas
a more progres:o, ivc and potent ially fatal su
beyond which cancer becomes extremely di
existing therapies become ineffecti ve and
decrease ( I). Angiogenesis is a complex
multiple time scales and intricate interplay
ldawson Opt1c
chemical and biomechanical mechanis including cell -cell
and cell-matrix interact ions, cell tor binding,
uential mor-
Replocement Text 812512007 4 23 14 PM
ldawson Opt1ons
is des igncd to
understanding llllis'f
ngiogenic processes will advance
lopment of new therapi es for treat-
iogcnesis-dependent djseases.
~ ~ ~ ~ ~ j E
A. Attached file; B. Highlighted text; C. Crossed-out (strike-through) text; D. Inserted text; E. Replaced text
5. Inserting symbols or special characters
An ' insert symbol' feature is not available for annotations, and copying/pasting symbols or non-keyboard characters from Mi-
crosoft Word does not always work. Use angle brackets<> to indicate these special characters (e.g., <alpha>, <beta>).
6. Editing near watermarks and hyperlinked text
eProof documents often contain watermarks and/or hyperlinked text. Selecting characters near these items can be difficult us-
ing the mouse alone. To edit an eProof which contains text in these areas, do the following:
Without selecting the watermark or hyperlink, place the cursor near the area for editing.
Use the arrow keys to move the cursor beside the text to be edited.
Hold down the shift key while simultaneously using arrow keys to select the block of text, if necessary.
Insert, replace, or delete text, as needed.
7. Summary of main functions
Insert text- Use Text Edits tool (position cursor and begin typing)
Replace text- Use Text Edits tool (select text and begin typing)
Delete text- Use Text Edits tool (select text and press delete key)
Highlight text- Use Highlight Text tool (select text)
Attach a file- Use the Attach a File with Comment tool (select tool. position
cursor and click mouse, select fi le)
8. Reviewing changes
To review all changes, do the following:
Select the Show button on the Comment & Markup toolbar.
Select Show Comments List.
Comment & Markup
aJ Text Edits
Suggested too/bar layout
Note: Selecting a correction in the list will highlight the corresponding item in the document, and vice versa.
-t;.:::: show
~ Expand All 18 0 Next 0 PreviOUS I ~ Reply 5 ../, Sel Status - l"JJ Checkrnark I .... Show ~ Sott By ' r Search Optrons "
E3 Page I
11 '-'
I!] klawson example_resource_library.jpg
11 ...., [f!]
I! c.... [!]
klawson phases,
this is a test
ll w
Use the Comments Jist to review all changes
9. The eReturn process
A. An email is received that contains a link to the eProof of the article:
http://eproofing. dartmouthjournals. com/pdfproofi ng/jou rnal1234. pdf
B. Click on the link to open the proof with the internet browser. Select "Save As" from the browser's 'File' menu to save a
copy of the PDF to the desktop or other folder.
C. Close the browser and open the saved PDF file with Acrobat.
D. Make corrections using Acrobat's Comment & Markup tools.
E. Save the PDF file, now with annotations, and return according to the instructions provided by the journal manager.
Skin cancer meets vitamin D: The way forward
for dermatology and public health
Martin A. Weinstock, MD, PhD," and Arnold M. Moses, MOb
Providence, Rhode I sland, and Syracuse, New York
itamin D is an essential vitamin that, in recent
years, has become increasingly important to
the dermatology community because of new
evidence of its effects on health and disease and
because of the role of ultraviolet B (UVB) radiation in
its synthesis in the skin as well as in cutaneous
Hence, by promoting protection of
the skin against UVB for cancer prevemion, we may
risk increasing rates or severity of other diseases.
Dermatologists should consider discussing vitamin D
with their patients when UVB exposure issues are
Childhood rickets has been recognized for hun-
<.lreds, if not thousands, of years. Ro.man children in
the first century AD seem to have been afflicted with
rickets; it became more prevalent when people
began to dwell in crowded, smoky, and sunless
cities. In the mid-17th cemury there were multiple
publications in England on the subject of rickets.
1822, children living in \Varsaw were noted to have a
rickets more commonly than children living in the
surrounding countryside and for the first time the
idea was expressed that this disease was caused by
lack of sunshine.
Later in the 19th century, sun
bathing was described as being helpful in preventing
rickets. In 1921 McCollum identified a substance
found in certain fats which could prevent rickets. It
was the fourth "vitamin" identified and thus named
From the Dermatoepidemiology Unit, VA Medical Center, t he
Departments of Dermatology and Community Health, Brown
University, and the Department of Dermat ology, Rhode Island
Hospital, Providence; and the Department of Medicine, Divi
sion of Endocrinology and Metabolism, State University of New
York Upstate Medical University, Syracuse.b
Funding sources: None.
Conflicts of interest: None declared.
Reprint requests: Martin A. Weinstock, MD, PhD, Dermatoepidemiology
Unit, VA Medical Center-1 1 1D, 830 Chalkstone Ave, Providence, Rl
02908. E-mail: maw@brown.edu.
Published online .
J Am Acad Dermatol 2009;:
2009 by t he American Academy of Dermatology, Inc.
doi:l o. 101 6/j.jaad.2009.04.01 6
Abbreviations used:
25(0H)D: calcidiol
1,25(0H)D: 1,25 dihyclroxyvitamin D
vitamin D. Hess and Weinstock
reported in 1924 that
irradiated vegetable oils could also have this etiect.
In the 1930s the chemical structure of vitamin D and
the fact that it was produced in the skin by ultraviolet
irradiation of 7-dehydrocholesterol was discovered.
Soon after, this precursor was added to milk and
some other foods that were then subjected to ultra-
violet irradiation.
Cholecalciferol (vitamin D
), the natural animal
vitamin D, is the form currently added to fortify foods
and is synthesized as a dietary supplement by irra-
diating 7-dehydrocholesterol extracted from lanolin
found in sheep's wool. Where animal products are
not desired, ergocalciferol (vitamin D
) can be made
by irradiat ing the fungal steroid ergosterol.
Ergocalciferol is approved by the Food and Drug
Administration as a pharmaceutical agent in the
United Stares, but this is not so for cholecalciferol.
There is some controversy about the relative biolog-
ical effects of ergocalciferol and cholecalciferol.
Some believe that both forms are equall y active,
whereas others claim that cholecalciferol has a more
favorable profile.

When obtained exogenously from food or sup-
plements, vitamin D is absorbed in the small intes-
tine. Natural food sources of vitamin D include cod,
salmon, eel, and egg yolks, but at least in the United
States, most dietary vitamin D is derived from fOJti-
fied food such as milk, orange juice, and breakfast
Actually, most people in the United States
have tO resort to taking vitamin D supplements
alone, or with calcium or in a multivitamin to obtain
the recommended daily amounts. Uncommonly,
absorption from jejunum and ileum may be impaired
by diseases, such as gluten-sensitive enteropathy,
EDI 5.0 DTO ymjd6240 15 May 2009 5:05pm
2 Weinstock and Moses
and large amounts of vitamin D are required.
Following absorption, vitamin D is transported in
the blood on a binding protein along with vitamin D
derived from the skin, to the liver where it is
QI passively 25-hydroxylated to calcidiol [25(0H)Dl.
This in turn may be stored in adipose tissue or be
transported to the kidneys and other tissues where it
is 1-hydroxylated to the active compound 1,25
dihydroxyvitamin D [1,25{0H)D]; calcitriol). This
second hydroxylation is regulated by a variety of
factors, including parathyroid hormone, phosphate,
and magnesium. Hence levels of calcitriol may not
reflect the level of calcidiol stores. Calcitriol regulates
calcium and phosphorus concentrations, which in
turn leads to proper mineralization of the skeleton to
form strong (and growing) bones and prevents
rickets in children and osteoporosis and osteomala-
cia in adults. The regulation of blood levels of
calcium and phosphorus is through directly or indi-
rectly increasing calcium and phosphorus absorp-
tion from the small intestine, decreasing urine
phosphorus excretion, and decreasing secretion of
parathyroid hormone.
The musculoskeletal consequences of inadequate
vitamin D have been extensively studied, although
considerable controversies remain
In addition to
rickets in children, and osteomalacia and osteoporo-
sis in adults (see above), falls and fractures in the
elderly are also important complications. Reduced
falls and fractures in the elderly were found in
systematic reviews of randomized trials of vitamin
D supplementationi0'
Other trials have failed to
confirm this finding, perhaps due in pa1t to variation
in baseline vitamin D or calcium status, or to selection
of populations studied, or other factors. Hence some
recenl systematic reviews have not found that sup-
plementation is effective for fracture reduction.
Among malignancies, the greatest body of evi-
dence supports a role of vitamin D in preventing
colorectal cancer,
although cancers of the breast
and prostate have also been prominently linked to
vitamin D.
However, the one large randomized trial
that has been published, which involved 36,000
women, found no difference in incidence of colo-
rectal or breasl cancer during a 7-year interven-
tion. 18 19 That trial involved supplememation with
400 IU of vitamin D plus calcium, and the 25(01-I)D
levels achieved were not reported. This trial has been
criticized because the close of vitamin supplement
and compliance may not have been sufficient to
produce a measurable effect that might have been
achieved with higher ingested doses.
M ONTH 2009
Other malignancies have been associated with
vitamin D deficiency in observational studies, but
randomized trial evidence of a beneficial effect in
specific cancers is Jacking.
Ecological studies, such
as those associating malignancies with latitude of
residence, have also been published, but confound-
ing can be a major issue in these analyses, and
conflicting trends have been noted
One notewor-
thy randomized trial with a planned outcome of
fracture compared 1000 TU of vitamin D with 1500
mg of calcium to calcium alone and to placebo and
found a difference in the proportions of each group
that developed any cancer during the trial: 2.9%,
3.8%, and 6.9% in the calcium plus D, calcium alone,
and placebo groups, respectively.
The appropriate
interpretation of this finding remains unclear, partic-
ularly in the context of other studies that. have
suggested a possible role of calcium
2 3
The differ-
ences among groups were not specific with respect
to the type of cancer.
The frequency of multiple other health problems
has been shown to be inversely correlated with
vitamin D levels in some studies. These diseases
include multiple sclerosis,
diabetes mellitus,
among others. Current evidence is
insufficient. to infer a cause-and-effect relationship.
Death is not. the primary outcome in most, if not.
all, vitamin D trials published to elate. However,
mortality information is often available, particularly
since these trials are typically performed with a
substantial number of frail elderly participants. A
recently published systematic review collected 18
published randomized trials of vitamin D supple-
mentation, usually with participants who were
elderly, had low baseline vitamin D levels, or had
a history of musculoskeletal conditions consistent
with inadequate vitamin D.
The intervention was
usually oral ingestion of vitamin D, typically in the
400 to 880 IU range, which some view as likely to
be suboptimal. The baseline mean 25(0H)D levels
in these trials were typically less than 20 ng/ml and
increased to the 20 to 40 ng/mL range after sup-
plementation (when measured). Over 6 months to
7 years follow-up in these varied studies, no one
study found a significant difference between ran-
domized groups in mortality, but meta-analysis
revealed an overall statistically significant 7% de-
creased mortality in the vitamin D supplement
group compared to controls. We do not know the
proportion of this difference in mortality that can be
attributed to musculoskeletal, cardiovascular, neo-
plastic, or other specific causes. Although most of
these studies were of the elderly, the largest si ngle
study among them was in women 50 to 79 years of
EDI 5.0 DTO ymjd6240 15 May 2009 5:05pm
Hence, we can conclude that vitamin D supple-
mentation appears to enhance musculoskeletal health
and reduce mortality in some groups that include a
substantial propo1tion of elderly individuals and
others at high risk of musculoskeletal problems. The
extent to which this can be applied to more general
populations with various baseline 25(0H)D levels
remains to be documented. We have evidence that
vitamin D may have an influence on cancer, cardio-
vascular disease, autoimmune disease, and infection,
but this evidence, although suggestive, is quite varia-
hie in strength and does not allow the conclusion that
any of these associations are proven.
These are
subjects of ongoing investigation.
Since 25(0H)D is the accepted marker of vitamin
D status, it is useful to estimate an optimal level.
Measures such as intestinal calcium absorption and
parathyroid hormone levels suggest that maximal
effects are attained by about 30 to 35 ng/ mL.
Other outcomes, such as lower extremity function
and fractures, may be more directly relevant, and the
estimates based on these tend to be in approximately
the same range.
Risk of cardiovascular events in the
Framingham Offspring cohort study suggested lower
optimal levels, about 20 to 25 ng/mL.
The nation-
ally representative National Health and Nutrition
Examination Survey data suggested that overall
mortality was minimized at levels of about 30 to 40
Given concerns about potential confound-
ing factors, we cannot be cettain about these esti-
mates of optimal levels, but the preponderance of
published evidence would support values of 25 to 40
ng/mL (63 to 100 nmol/L)
; 30 ng/ mL (75 nmol/L) is
a common estimate given current knowledge.
extent to which this estimate should va1y with age,
calcium intake, and other factors remains to be
clarified , and these estimates may be revised as
further evidence is published.
ln efforts to achieve optimal vi tamin D status, the 3
sources of vitamin D must be considered. The first of
these is diet. Most dietary vitamin D intake in typical
diets is due to fortification of foods, such as mi lk and
some cereals and orange juices in the United States,
although this varies substantially hy country. Few
foods are naturally rich sources of this vitamin.
Photosynthesis of vitamin D in the skin requires
ultraviolet radiation in the "B" range (UVB)
and is
generally the largest source of vitamin D. The available
flux of UVB depends on multiple factors, including
latitude, time of day, and season of the year which
determine the angle of the sun from the horizon, and
Weinstock and Moses 3
by cloud cover and shade. Other factors that deter-
mine the ability of the skin to make vitamin D include
clothing worn, skin color, and age (there is a partic-
ularly important decline in the ability of skin ro
synthesize this vitamin with advancing age,
the elderly respond to oral supplementation
) . Other
factors may determine vitamin D levels as well.
series of studies has been published in a variety of
sunny, relatively equatorial locations, including
Santiago (Chil e),'
southern Arizona,
south Flori da,
and Queensland (Australia),
showing that a substantial portion of the population
has low 25(0H)D levels.
Skin cancer prevention efforts did not cause the
problem of low vitamin D levels, and abandoning
sun protection will not solve it. Dermatologists
should be aware that sun protection recommenda-
tions may impede the effectiveness of vitamin D
photosynthesis in the skin
and that it may be
unrealistic to presume that levels on the order of 30
ng/ mL will consistentl y be obtained from incidental
sun exposure. The vast majority of the randomized
trials that have documented the beneficial effects of
increased vitamin D have achieved that increase by
oral supplementation.
The existing Institute of Medicine guidelines rec-
ommend daily intake of 200 lU of vitamin D for
children and adults under the age of 50 years, 400 IU
for those between 50 and 70 years of age, and 600 IU
for those over the age of 70 years
These intake
levels may be increased in light of ongoing research. Q2
Toxicity of excess vitamin D may be manifest as
hypercalcemia or hypercalciuria; alt hough our
knowledge of the effects of high doses of vitamin
Dis limited, a recent review suggested that 10,000 IU
daily is a safe upper limit for adult intake
American Academy of Pediatrics recently increased
its recommended intake from 200 to 400 lU per clay
for all children and adolescents.
Some have sug-
gested that 3000 to 5000 IU daily may be required to
optimize 25(0H)D levels
Although factors including obesity, bariatric sur-
gery, diabetes, skin color, dieta1y intake, dieta1y
supplements, and sun exposure may affect
25(01-I)D levels, history and physical examination
cannot reliably predict these levels.
Clinicians necessarily address low 25(0H)D levels
in individual patients. Individual levels can be
measured, and supplementation recommended.
Common recommendations include 400 or 1000 IU
daily with or without calcium or in a multivitamin;
10,000 IU weekly or eve1y 10 days; or 50,000 IU
EDI 5.0 DTO ymjd6240 15 May 2009 5:05pm
4 Weinstock and Moses
monthly. Subsequent measurement of 25(0H)D
levels can be quite useful since there is great varia-
tion in the increment of 25(0H)D achieved from a
given dose of vitamin D supplement.
For patients with 25(0H)D levels less than
15 ng/ mL, higher doses of supplementation are
needed for a limited time to achieve the recommen-
ded levels within a reasonable period. A common,
effective, and safe method of addressing this problem
is to prescribe oral supplements of 50,000 IU vitamin
D weekl y for 8 weeks and then switch to the lower
doses mentioned above. Patients whose 25(0H)D
Q3 levels are not normalized with this approach may
have a malabsorptive syndrome and should be
referred for evaluation and alternative management.
Some have suggested that use of tanning parlors is
a recommended source of vitamin D,
and the
commercial indoor tanning industty has used this
for marketing purposes. However, patrons of these
facilities generally cannot determine the flux of UVB
delivered by the tanning booths and cannot deter-
mine the amount of vitamin D, if any, being gener-
ated in their skin from this exposure. It is clear,
however, that indoor tanning has been associated
with skin cancers, including squamous cell carci-
noma52 and melanoma,
and is a less reliable and
more expensive source of vitamin D than oral
supplementation for the general population.
For dermatologists and other clinicians whose
focus is on the patient before them, determination of
25(0H)D levels and appropriate supplementation
based on those results may be the preferred option
for managing skin cancer prevention in the context
of potentially low levels of vitamin D, panicularly
among those at high risk for future ultraviolet-
induced skin cancers, such as those whose untanned
skin color is lighter than a light to medium brown.
However, from the public health point of view,
increasing fortification of foods may be a preferable
method for management of vitamin D status that
would reach the entire population, including those
who do not have regular medical care, and fortifi ca-
tion would not require individuals to change routine
We thank Kimberly Marcolivio for her assistance.
1. Lim HW, Gilchrest BA, Cooper KD, et al. Sunlight, tanning
Q4 booths, and vitamin D. J Am Acad Dermatol 2005;52:868-76.
2. O'Riordan JL. Rickets in the 17th century. J Bone Miner Res
2006;21 :1506-1 0.
3. Mozolowski W. Jedrzej Snaidecki (1768-1838) on the cure of
rickets. Nature 1939;143:121.
4. Hess AF, Weinstock M. Antirickitic imparted to inert fluids and
green vegetables by ultra-violet irradiation. J Bioi Chem 1924;
M ONTH 2009
5. Holick MF, Biancuzzo RM, Chen TC, et al. Vitamin D2 is as
effective as vitamin D3 in maintaining circulating concentra-
tions of 25-hydroxyvitamin D. J Clin Endocrinol Metab 2008;93:
6. Trang HM, Cole DE, Rubin LA, Pierratos A, Siu S, Vieth R.
Evidence that vitamin 03 increases serum 25-hydroxyvitamin
D more efficiently t han does vitamin D2. Am J Cli n Nutr 1998;
7. Armas LA, Hollis BW, Heaney RP. Vitamin D2 is much less
effective than vitamin D3 in humans. J Clin Endocrinol Metab
2004;89:5387 -91.
8. Whiting SJ, Calvo MS. Overview of the proceedings from
Experimental Biology 2005 symposium: Opt imizing Vitamin D
Intake for Populations with Special Needs: Barriers to Effective
Food Fortification and Supplementation. J Nut r 2006;136:
9. Wolpowitz D, Gilchrest BA. The vitamin D questions: how
much do you need and how should you get it? J Am Acad
Dermatol 2006;54:301-17.
10. Bischoff-Ferrari HA, Dawson-Hughes B, Willett WC, et al. Effect
of Vitamin D on falls: a meta-analysis. JAMA 2004;291:1999-
11. Bischoff-Ferrari HA, Willett WC, Wong JB, Giovannucci E,
Dietrich T, Dawson-Hughes B. Fracture prevention with
vitamin D supplementation: a meta-analysis of randomized
controlled trials. JAMA 2005;293:2257-64.
12. O'Donnell S, Moher D, Thomas K, Hanley DA, Cranney A.
Systematic review of the benefits and harms of calcitriol and
alfacalcidol for fractures and falls. J Bone Miner Metab 2008;26:
13. Cranney A, Weiler HA, O'Donnell S, Puil L. Summary of
evidence-based review on vitamin D efficacy and safety in
relation to bone health. Am J Clin Nutr 2008;88:513S-9S.
14. Dawson-Hughes B. Serum 25-hydroxyvitamin D and functional
out comes in the elderly. Am J Clin Nutr 2008;88:537S-40S.
15. Bischoff-Ferrari HA, Kiel DP, Dawson-Hughes B, Orav JE, Li R,
Spiegelman D, et al. Dietary calcium and serum 25-hydrox-
yvitamin D status in relation to bone mineral density among
U.S. adults. J Bone Miner Res 2009;24:935-42.
16. Gorham ED, Garland CF, Garland FC, et al. Optimal vit amin 0
status for colorectal cancer prevention: a quantitative meta
analysis. Am J Prev Med 2007;32:210-6.
17. Garland CF, Gorham ED, Mohr SB, et al. Vitamin D and
prevention of breast cancer: pooled analysis. J Steroid
Biochem Mol Bioi 2007;103:708-11.
18. Wactawski-Wende J, Kotchen JM, Anderson GL, et al. Calcium
plus vitamin D supplementation and the risk of colorectal
cancer. N Engl J Med 2006;3S4:684-96.
19. Chlebowski RT, Johnson KC, Kooperberg C, et al. Calcium plus
vitamin D supplementation and t he risk of breast cancer.
J Natl Cancer lnst 2008;100:1581-91.
20. , International Agency for Research on Cancer. Vitami n D and
cancer Vol 5. Lyon: International Agency for Research on
Cancer; 2008.
21. Waltz P, Chodick G. Assessment of ecological regression in the
study of colon, breast , ovary, non-Hodgkin' s lymphoma, or
prostate cancer and residential UV. Eur J Cancer Prev 2008;1 7:
22. Lappe JM, Travers-Gustafson D, Davies KM, Recker RR, Heaney
RP. Vitamin D and calcium supplementation reduces cancer
risk: results of a randomized trial. Am J Clin Nutr 2007;85:
23. Grau MV, Baron JA, Sandier RS, et al. Vitamin D, calcium
supplementat ion, and colorectal adenomas: results of a ran-
domized trial. J Natl Cancer lnst 2003;95:1765-71.
EDI 5-0 DTO ymjd6240 15 May 2009 5:05 pm
24. Munger KL, Levin Ll, Hollis BW, Howard NS, Ascherio A. Serum
25-hydroxyvit amin D levels and risk of multiple sclerosis.
JAMA 2006;296:2832-8.
25. Scragg R, Sowers M, Bell C. Serum 25-hydroxyvitamin D,
diabetes, and ethnicity in the Third National Health and
Nut rition Examinat ion Survey. Diabetes Care 2004;27:2813-8.
26. Cannell JJ, Viet h R, Umhau JC, et al. Epidemic influenza and
vitamin D. Epidemiol Infect 2006;134:1 129-40.
27. Autier P, Gandi ni S. Vitamin D supplementation and tot al
mortality: a met a-analysis of randomized controlled t rials. Arch
Int ern Med 2007;167: 1730-7.
28. Jackson RD, LaCroix AZ, Gass M, et al. Calcium plus vitamin D
supplementation and the risk of fractures. N Engl J Med 2006;
29. Chapuy MC, Preziosi P, Maamer M, et al. Prevalence of vitamin
D insufficiency in an adult normal population. Osteoporos lnt
30. Heaney RP. Functional indices of vitamin D status and
ramificat ions of vitamin D deficiency. Am J Clin Nutr 2004;
80(6 Suppi):1706S-9S.
31 . Bischoff-Ferrari HA, Dietrich T, Orav EJ, et al. Higher 25-
hydroxyvitamin D concentrations are associated wit h better
lower-extremity function in both adive and inact ive persons
aged > or =60 y. Am J Clin Nutr 2004;80:752-8.
32. Wang TJ, Pencina MJ, Booth SL, et al. Vitamin D deficiency
and risk of cardiovascular disease. Circulat ion 2008;117:
33. Melamed ML, Michos ED, Post W, Astor B. 25-Hydroxyvitamin
D levels and the risk of mortality in the general population.
Arch Intern Med 2008;168:1629-37.
34. Bischoff-Ferrari HA, Giovannucci E, Willett WC, Dietrich T,
Dawson-Hughes B. Estimation of optimal serum concentra-
tions of 25-hydroxyvitamin D for multiple health outcomes.
Am J Clin Nutr 2006;84:18-28.
35. Vieth R, Bischoff-Ferrari H, Boucher BJ, et al. The urgent need
t o recommend an intake of vitamin D t hat is effedive. Am
J Clin Nutr 2007;85:649-50.
36. Maclaughlin JA, Anderson RR, Holick MF. Spectral charac-
t er of sunlight modulat es phot osynthesis of previtamin D3
and its photoisomers in human skin. Science 1982;216:
37. Holick MF, Matsuoka LV, Wortsman J. Age, vitamin D, and solar
ult raviolet. Lancet 1989;2:11 04-5.
38. Harris SS, Dawson-Hughes B. Plasma vitami n D and 250HD
responses of young and old men to supplementation with
vitamin D3. J Am Coli Nutr 2002;21 :357-62.
Weinstock and Moses 5
39. Wortsman J, Matsuoka LV, Chen TC, Lu Z, Holick MF. De-
creased bioavailability of vitami n D in obesity. Am J Clin Nut r
40. Binkley N, Novotny R, Krueger D, et al. Low vitamin D status
despite abundant sun exposure. J Clin Endocrinol Metab 2007;
41. Gonzalez G, Alvarado JN, Rojas A, Navarrete C. Velasquez CG,
Arteaga E. High prevalence of vitamin D deficiency in Chilean
healthy postmenopausal women with normal sun exposure:
additional evidence for a worldwide concern. Menopause
42. Jacobs ET, Alberts DS, Foote JA, et al. Vitamin D insufficiency
in southern Arizona. Am J Clin Nutr 2008;87:608-1 3.
43. Levis S, Gomez A, Jimenez C, et al. Vitamin D deficiency and
seasonal variation in an adult Sout h Florida populat ion. J Clin
Endocrinol Met ab 2005;90:1557-62.
44. Kimlin M, Harrison S, Nowak M, Moore M, Brodie A, Lang C.
Does a high UV environment ensure adequate vitamin D
st atus? J Photochem Photobiol B 2007;89:139-47.
45. Mat suoka L Y, Ide L, Wortsman J, Maclaughli n JA, Holick MF.
Sunscreens ~ u p p r s s cutaneous vitamin D3 synthesis. J Clin
Endocrinol Met ab 1987;64:1 165-8.
46. Standing Committ ee on t he Scientific Evaluation of Diet ary
Reference Intakes Food and Nutrition Board. Food and
Nutrition Board. Institute of Medicine. Calcium, phosphorus,
magnesium, vitamin D and fluoride. Dietary Reference Intakes.
Washington (DC): National Academy Press; 1997. p. 250-87.
47. Hathcock JN, Shao A, Vieth R, Heaney R. Risk assessment for
vitamin D. Am J Clin Nutr 2007;85:6-18.
48. Wagner CL, Greer FR. Prevention of rickets and vitamin d
deficiency in infants, children, and adolescents. Pediat rics
2008;122:1 142-52.
49. Aloia JF, Patel M, Dimaano R, et al. Vitamin D intake to attain a
desired serum 25-hydroxyvitamin D concentration. Am J Clin
Nutr 2008;87:1952-8.
SO. Heaney RP, Davies KM, Chen TC, Holick MF, Barger-Lux MJ.
Human serum 25-hydroxycholecalciferol response to ex-
tended oral dosing with cholecalciferol. Am J Clin Nutr 2003;
77:204- 10.
51 . Holick MF. Vitamin D deficiency. N Engl J Med 2007;357:266-81.
52. Karagas MR, Stannard VA, Mott LA, Slattery MJ, Spencer SK,
Weinstock MA. Use of tanning devices and risk of basal cell and
squamous cell ski n cancers. J Nat l Cancer lnst 2002;94:224-6.
53. The association of use of sunbeds with cut aneous malignant
melanoma and other skin cancers: a systematic review. lnt
J Cancer 2007;120:11 16-22. Q5
EDI 5.0 DTO ymjd6240 15 May 2009 5:05 pm
Journal: YMJD
Article no.: 6240
Dear Author,
Author Query Form
During the preparation of your manuscript for typesetting, some questions have arisen. These are listed below. Please check your
typeset proof carefully and mark any corrections in the margin of the proof or compile them as a separate list*. This form should
then be returned with your marked proof/list of corrections to Elsevier. Please use black ink for all comments on proofs and other
Disk use
In some instances we may be unable to process the electronic file of your article and/or artwork. In that case we have, for
efficiency reasons, proceeded by usi ng the hard copy of your manuscript. If this is the case the reasons are indicated below:
r Disk damaged r Incompatible file format r LaTeX file for non-LaTeX joumal
r Virus infected r Discrepancies between electronic file and (peer-reviewed, therefore definitive) hard copy
r Other: ...... .... ........ ........ .................... ........ ........................ ........................................ ........................ ........ ....................... .
We have proceeded as follows:
r ManuSClipt scanned r Manuscript keyed in r Artwork scanned
r Files only partly used (parts processed differently: ..................... . .................... . .. .. .......................................... )
If discrepancies were noted between the literature list and the text references, the following may apply:
r The references listed below were noted in the text but appear to be missing from your literature list. Please complete the list or
remove the references from the text.
r Uncited references: This section comprises references that occur in the reference li st but not in the body of the text. Please
position each reference in the text or delete it. Any reference not dealt with will be retained in this section.
Queries and/or remarks
Query Details required Author's response
(Ql) Note that abbreviation has been changed; thi s one appears
more frequently in the literature. OK?
(Q2) Note insertion of "intake levels"; done to clarify "These";
(Q3) Note rewording; meaning retained?
(Q4) For all references ending in "et al": please supply up to 6
authors before "eta!"; per journal style.
(Q5) Ref 53: Are there no authors for this? Please advise and
supply, if applicable.
ELSEVIER Many thanks for your assistance
*In case artwork needs revision, guidance can be found at http://authors.elsevier.com/artwork.
Page 1 of 1
I LL Docume n t De l i v e r y
NLM -- W1 A126 (Gen); W1 A126 (Ref ); Film S02755; E-Journal w/ILL access
PHONE : 202- 326- 2395
FAX : 202- 326- 2732
E- MAIL : library@ftc . gov
2009-06-17 14:39:07
2009-06-18 12:35:55
REG Copy Journal
American Medical Assn. Chicago.
2009 Feb 23; 169( 4 ): 41 6-7 416-7
Golomb B; Aronoff-Spencer E;Steadman M;Wu W;Yan A
Un ique ID.: 0372 44 0
KYB- 27273266
$1 4. 00
W1 A126 (Gen); W1 A126 (Ref); Film S02755;
E-Journal w/ ILL access
N/ A
E-mail : library@ftc.gov
Mail :
For problems or questions. contact NLM a t h ttp : //wwwcf.nlm.nih.
gov/ill/ill_web_form. cfm or phone 301-496- 5511.
Include LIBID and request number.
NLM Collection Access Section . Bet hesda . MD
women, respectively, but observed a similar near expo-
nential increase in hearing loss with age. In agreement with
NHANES data, hearing loss prevalence increased signifi-
cantly with age and was greater in men than in women.
Apart from hypertension, we confirm all associations be-
tween potential risk factors and any level of hearing loss
as reported by Agrawal et al.
Furthermore, these associa-
tions were marginally stronger (except for smoking) for
more severe levels of hearing loss. Thus, we concur with
Agrawal et al' that focusing on modifiable risks may help
to reduce the prevalence of age-related hearing loss. In con-
clusion, data from both the BMES and NHANES highlight
the burden imposed by untreated and/or underrecog-
nized hearing loss and indicate the need for possible strat-
egies to eliminate preventable hearing loss.
Bamini Gopinath, PI1D
Elena Roc1rtchina, MApplStat
Ji ejin Wang, PhD, MMed, MApplStat
Julie Sdmeider, PhD
Stephen R. Leeder, MD, PhD
Paul Mitchell, MD, PhD
Correspondence: Dr Mitchell, Centre for Vision Re-
search; University of Sydney, Westmead Hospital,
Hawkesbury Road, Westmead, New South Wales, Aus-
tralia 2145 (paul_mitchell@wmi.usyd.edu.au).
Author Contributions: Study concept and design: Gopinath,
Schneider, imd Mitchell. Acquisition of
sis and interpretation of Rochtchina, Leeder,
and Mitchell. Drafting of th'e manuscript: Gopinatl1. Criti-
cal revision of the mamucript for important intellectual con-
tent: Rochtchina, Wang, Schneider, Leeder, and Mitchell.
Statist'ical analysis: Rochtchina. Obtained funding: Mitchell.
Administrative, technical, and material support: Leeder. Study
supervision: Wang, Schneider, and Mitchell.
Financial Disclosure: None reported.
1. Agraw:tl Y, Piau EA. Nlparko JK. Prevalence of hearing loss and differences
by demographic cluractulstics among US adults: dat.t from the National Health
and Nutrition Examlrutl.on Sul'\ey, 1999-20<. Arch lnttm Mtd. 2008;168
2. Attebo K. Mitchell P, Smith W. Visual acuity and the a uses of \isualloss in
Australia: the Blue Mountains Study. Ophthalmology. 1996;103(3):357-
3. Whitworth ) A: World Organization, International Society of H)-per-
tension Writing Group. 2003 World Health Organizat ion (WHO)/
International Society or H)-pCrtcnsion (ISH) statement on management or
hn>ertcnsion.J Hypcruns. 2003;21(ll):l983-1992.
Care Quality and FraU Sublects
e read with interest the editorial titled "Im-
proving Care Quality and Reducing Dispari-
ties,"' and we would like to comment, add-
ing peculiar focus on "unequal treatment" given to very
old, frail subjectS.
At present, elderly persons may be considered mem-
bers of a minority population even if their numbers are
......!. '.
reaching levels much higher than other groups (eg, black,
poor, or disabled subjectS). The crucial pointS are, from
one side, clinical prejudices against old age and, from the
other side, the lack of convincing studi es transferring sci-
entific evidence to the real-world conditions of very old
subjectS, characterized by comorbidity and disability (and
very often also by a reduced cognitive function).
To overcome the gap, it could be important to pro-
duce public reports on clinical performances (starting from
the most easily demonstrable events, eg, the rate of cata-
ract removal or hip prosthesis implantation in subjectS
with dementi a), although the demonstration of out-
comes is far more difficult in elderly subjects. These re-
ports would be extremely useful for clinicians in estab-
lishing control conditions for their work and assuring a
high level of care for the most frail, elderly subjects, thus
reducing dispari ties in their access and quality of care.
Renzo Rozzini, MD
Marco Trabucd1i, MD
Corres pondence: Dr Rozzini, Department of Internal
Medicine and Geriatrics, Poliambulanza Hospital, via Ro-
manino 1, Brescia 25122, Italy (renzo.rozzini@iol.it).
1. Clancy C. lmpro,;ng care quality and reducing disparities: physici:lns' roles
Arch Inttm l\ltd. 2008; 168(11):1135-1136.
A Ray of Sunshine
for the VItamin D-Heart Hypothesis
e suggest that the observation by Giovan-
nucci et ai,l linki ng higher plasma 25-
hydroxyvitamin D concentrations to lower
myocardial infarction risk in men, bears fewer discrep-
ancies with existi ng literature than the authors pre-
sume. The authors note that their observational find-
ings disagr ee wi th the findings fr om the l argest
randomized trial of vitami n D, the Women's Health Ini-
tiative (WHI) study, in which the vitamin D +calcium
group exhibited no benefit-nor trend to benefit-
against myocardial infarction (hazard ratio, 1.04; 95% con-
fidence interval, 0.92-1.18).
We propose that this difference may arise, not pri-
marily from the relatively low vitamin D dose of 400 IU/d,
as the authors suggest, but from the inclusion of cal-
cium with vitamin D supplementation in the WHI: a re-
cent double-blind randomized controlled trial in post-
menopausal women reported an apparent increase in
myocardial infarction incidence in those randomized to
calcium .(as calcium citrate) vs placebo.
Moreover, in a
meta-analysis of randomized trials identifying a benefit
of vitamin D supplementation against all-cause mortal-
ity, only 2 vitamin D trials differed in the direction of effect
from the benefit found overall: both of these trials com-
bined vitamin D with calcium Thus,
the supplemental calcium in the vitamin D arm ofWHI
could be speculated to countermand its cardiac ben-
efits, rendering findings across these several studies wholly
ARCH INTERN MED/VOL 169 (NO. 4), FEB 23, 2009
Material may be protected by copyright law (Title 17, U.S. Code)
. i
"May sunshine brighten your heart" is an Irish epi-
gram5: if a causal link is affirmed in randomized trials,
then it could become a medical prescription.
Beatrice Golomb, MD, PhD
Eli Aronoff-Spencer, MD, PltD
Midtael Steadman, MD
Winnie Wu, MS
Arthur Yan, MD
Correspondence: Dr Golomb, Department of Medicine,
University of California, San Diego, 9500 Gilman Dr, Mail
Code 0995, La jolla, CA 92093-0995 (bgolomb@ucsd
Additional Information: At the time of the drafting of
this letter, the authors were members oflnpatient Ward
Team 3, VA San Diego Healthcare Center/UCSD School
of Medicine (June 2008).
1. Giovannucd E, Uu Y, Hollis BW, Rimm EB. 25-hydroxpi tamin D and risk
o myocardial infarction in men: a prospective study. Arch lnttrn Attd. 2008;
2. llsia j. Heiss G. Reo II, et al. CalciumMtamin D supplementation and car
diovascular events. Circulalion. 2007;115(7):846854.
3. Bolland MJ, Barber PA, Doughty RN, etaL Vascular events In healthy older
women rccching calcium supplementation: randomised conuolled trial. BM].
4. Au tier P, Gandini S. Vitamin D supplementation and total mortality: a meta
anal)sis o randomized controlled trials. Arch InumMtd. 2007;167(16):1730-
5. Irish Culture and Customs. httpJ/www.irishcultureandcustoms.com/blessings
lbless.html. Accessed june 2008.
Hemoglobin Ate Testing,
Boarci-RecertJflcatlon Scores,-
' ancl Treatment Outcomes .
he publication by Holmboe et al' tilled "Asso-
ciation Between Maintenance of Certification
Examination Scores and Quality of Care for
Medicare Beneficiaries" provides a valuable insight. A
few points, however, may be added in reviewing this
The that "evidence-based processes of care"
were used in this study seems to have applied in the time
frame used, 2002 and 2003, when patient care was ex-
amined. Recent findings/
however, suggest an up-
dated perspective"-6 in the value of the practice exam-
ined in the study by Holmboe et al.
Though unknown
at the time of the study, the updated perspective in-
cludes the risk-benefit ratio of treatments that may have
resulted from extensive hemoglobin A
e testing. The im-
pact of effective diabetic control may have caused inad-
vertent harm despite well-intended care and stricter imple-
mentation. Even without this possibility, it supports
compliance characteristics of physicians depending on
standards of care, regardless of possible subsequent
changes in standards.
In addition, in the group examining the performance
of diabetes measures, more than 80% of patients had re-
ceived hemoglobin A
e tests, while other components of
the composite measure were noticeably less common. The
highest among them was lipid testing, and this was simi-
lar to the group measuring performance in cardiovascu-
lar disease, for \Vhich no difference was found \vith scores
in board recertification.
It would have been interesting to know if there were
any differences in performance for diabetic manage-
ment other than hemoglobin testing, such as the eye
examination component, in its possible correlation with
board recertification scores. Familiarity with hemoglo-
bin testing may arguably have facilitated its use more
than eye examinations, but its ease of use, when com-
pared with lipid testing, suggests a less-accepted corre-
lation between eye examination outcomes and the skills
of physicians who treat diabetes.
While Holmboe et alHP
> suggest a "link between cog-
nitive skills and quality of care," it could be argued that
"skills" correlate more accurately in complying \vith stan-
dards, as far as standards can ensure quality. An impor-
tant part of physicians' cognitive skills is their ability to
assess whether the basis underlying a general standard
is sound. This may apply before changes in general stan-
dards are more widely recognized.
Simon G. Kassabian, MD
Correspondence: Dr Kassabian, Department of Medi-
cine,jewish Home Lifecare 120 W 106th St, New
York, NY 10025
1. Holmboe ES, Wang Y, Meehan TP, el at. Association between rnainterumce
or certification examination scores and quality o care for Medicare
Arch lnrern Atcd. 2008;168(13):13961403. .
2. Gerstein HC, Miller 1\t E, B)ington RP; The Action to Control Cardiovascular
Risk in Diabetes Study Group. Effects of intensi,e glucose lowering in type 2
diabetes. N Eng/} Med. 2008;358(24):25452559.
3. Patel A, MacMahon S, Chalmers]; The ADVANCE Collaborative Group.
Intensive blood glucose control and ''ascular outcomes in patients " i th type
2 diaberes. N Eng!] Med. 2008;358(24):2560-2572 .
4. Krumholz Lee Ttl. Redefining quality-implications or recent clinical
trials. N Engl} Attd. 2008;358(24):25372539.
S. Cefalu \VT. Glycemic targets and cardlo\'aSCubr disease. N Englj /lied. 2008;
6. Dluhy RG, McMahon GT. Intensive glycemic control in the ACCORD and
. ADVANCE trials. N Eng/] Attd.
Is Quality of Care Only Instrume ntal?
read with great interest the article by Holmboe et
al' and the editorial by Landon
on the link be-
tween certifying examination performance and qual-
ity of care.
As these 2 articles highlight, there are substantial prac-
tical problems \"'ith collecting adequate data to demon-
strate a reliable link between paper-based cognitive as-
sessments and instrumental performance in clinical
practice. In Australia, the Bettering the Evaluation of Care
and Health (BEACH) study has shown similar links be-
tween the results of the fellowship examination and in-
strumental performance in practice.
However, is it sufficient to judge quality of care solely
in instrumental tenus? I would suggest that quality of
care is equally dependent on an interpersonal perfor-
mance as on doing the "right thing" in terms of disease-
specific care. \Ve should not forget that the essence of
health and health care is the personal health experience
and coping by the patient with his or her ailments-
usually multiple ailments.
The challenge ahead will be
ARCH INTERN MEO/VOL 169 (NO. 4), FEB 23, 2009

Material may be protected by copyright law (Title 17, U.S. Code)
Trends in Sunburns, Sun Protection Practices, and Attitudes Toward Sun
Exposure Protection and Tanning Among US Adolescents, 1998- 2004
Vilma Cokkinides, Martin Weinstock, Karen Glanz, Jessica Albano, Elizabeth Ward
and Michael Thun
Pediatrics 2006;118;853-864
DOl: 10.1542/peds.2005-3 109
The online version of this article, along with updated information and services, is
located on the World Wide Web at:
http://www. pediatrics. org/cgil content/full/118/3/85 3
PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly
publication, it has been published continuously since 1948. PEDIATRICS is owned, published,
and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk
Grove Village, Illinois, 60007. Copyright 2006 by the American Academy of Pediatrics. All
rights reserved. Print ISSN: 003 1-4005. Online ISSN: 1098-4275.
American Academy of Pediatrics
Downloaded from www. pedi atrics.org by on October 21, 2009
Trends in Sunburns, Sun Protection Practices, and
Attitudes Toward Sun Exposure Protection and
Tanning Among US Adolescents, 1998-2004
Vilma Cokkinides, PhD, Martin Weinstock, MD, PhDb.c, Karen Glanz, PhD, MPHd, Jessica Albano, MSPH, Elizabeth Ward, PhD,
Michael Thun, MD, MS
Department of Epidemiology and SuNeillance Research, American Cancer Society, Atlanta, Georgia; bDermatoEpidemiology Unit, VA Medteal Center Providence.
Department of Dermatology, Rhode Island Hospital, Providence. Rhode Island; <Departments of Dermatology and Community Health, Brown University, Providence,
Rhode Island; dRollins School of Public Health, Emory University, Atlanta, Georgia
The author> have indicated they have no t\nanctal relationshlpsielevant w this antcle to dtS<:Iose.
ABSTRACT ---------------------------------------------------------------------
BACKGROUND. Sun exposure in childhood is an important ri sk factor for developing
skin cancer as an adult. Despite extensive efforts to reduce sun exposure among
the young, there are no population-based data on trends in sunburns and sun
protection practices in the young. The aim of this study was to describe nationally
representative trend data on sunburns, sun protection, and attitudes related to sun
exposure among US youth.
METHODS. Cross-sectional telephone surveys of youth aged 11 to 18 years in 1998 (N
= 1196) and in 2004 (N = 1613) were conducred using a 2 -stage sampling process
to draw population- based samples. The surveys asked identical questions about
sun protection, number of sunburns experienced, and attitudes toward sun expo-
sure. Time trends were evaluated using pooled logistic regression analysis.
RESULTS. In 2004, 69% of subjects reported having been sunburned during the
summer, not significantly less than in 1998 (72%). There was a significant de-
crease in the percentage of those aged 11 to 15 years who reported sunburns and
a nonsignificant increase among the 16- to 18-year-olds. The proportion of youth
who reported regular sunscreen use increased significantl y from 31% tO 391< ..
Little change occurred in other recommended sun protection practices.
CONCLUSIONS. A small reduction in sunburn frequency and modest increases in sun
protection practices were observed among youth between 1998 and 2004, despite
widespread sun protection campaigns. Nevertheless, the decrease in sunburns
among younger teens may be cause for optimism regarding future trends. Overall,
there was rather limited progress in improving sun protection practices and
reducing sunburns among US youth between 1998 and 2004.
www.pediatrics org/cgi/doi/10. 1542/
doi:l 0.1542/peds.2005-31 09
Other than the financial suppo11 for data
colleaion, 1he funder did not play a role in
the design and conduct of the study; data
colleaion. managemenl, analysis and
Interpretation of the data; or in the
preparation. review. or approval of the
The American Cancer Sociely staff and
coinvestlgatOJs were involved in I or more
of the following: design and conduct of the
study: d<Jta colleaion management:
analysis and interpretation of the data; and
preparatiOn, I\liew, and approval of the
o, Cokkinides had full access to all the data
in the study and takes responsibility for the
Integrity of the data and the accuracy of
the data analysis.The ideas and opiniom
expressed herein are those of the authors.
skin cancer, skin neoplasms prevention
and control. sun exposure, sun protection,
sunburn, sunscreen agents. adolescent.
epidemiology, health surveys
SPF 1 5+ -sun protection factor ;;::15
Ci-confidence interval
Accepted for publication Mar 27, 2006
Address conespondence to Vilma Cokkinides.
PhD, Depanment Epidemiology and
Surveillance Research, American Glncer
Society, 1599 Clifton Rd, NE. Atlanta, GA
30329-4251. E-mail: vcokkini@Cancei.oig
PEDIATRICS OSSN Numbers: P1ir1t, 0031-4005;
Online, 11J93.4275). Copylight 2006 by the
American Academy of Pediatrics
PEDIATRICS Volume 118, Number 3, September 2006 853
Downloaded from www. pediatrics.org by on October 21 , 2009
ORE THAN l mi!Uon cases of basal cell or squa-
mous cell carcinoma and 62 190 cases of invasive
melanoma are expected in 2006 in the United States.' Of
these, only melanoma is included in cancer statistics
published each year by US tumor registries. Although
basal ce!J and squamous cell skin cancers (keratinocyte
carcinomas) are highly treatable, they account for con-
siderable morbidity and health care expenditures. Fur-
thermore, the incidence of both melanoma
and kera-
tinocyte carcinoma4-6 have been increasing.
Extensive epidemiological and biological evidence
implicates sun exposure as the principal cause of mela-
noma and keratinocyte carcinomas of the skin.

estimated 80% of all skin cancers are caused by UV solar
Despite evidence that excessive sun exposure
in childhood contributes to the risk of skin cancers later
in life,
sunburns, which are markers of intense UV
exposure, remain common, especially among adoles-
cents.10-12 Several recent studies also highlight the ongo-
ing inadequacy of reported sun protection practices
among adolescentsY-
Most governmental and nongovernmental efforts to
prevent skin cancer in the United States have sought to
change the individual behaviors of parents and chil-
dren, ~
1 9
without a concomitant emphasis on sun pro-
tection policies, such as those used effectively in Austra-
lia.20 Furthermore, there have been no systematic efforts
to assess the effectiveness of sun protection interven-
tions in youth nationwide or to measure their changes in
attitudes toward sun protection and tanning, despite the
substantial disease burden caused by excessive sun ex-
posure. This study compared the frequency of sunburns
and attitudes and behaviors regarding sun protection in
2 nationally representative surveys of 11 - to 18-year-
olds conducted in 1998 and 2004.
Overall Design and Study Population
The American Cancer Society conducted 2 population-
based, national cross-sectional surveys of noninstitu-
tionalized children, I I to 18 years of age, Jiving in the
continental United States. The initial Sun Survey l was
conducted from August to November L 998 and sampled
1196 boys and girls in this age range. Six years later, Sun
Survey U was conducted in August to November 2004
and sampled l 613 respondents in the same age range.
Both surveys used similar telephone-based sampling
A 2-stage sampling methodology was
used to obtain a representative sample of US households
with telephones stratified by 7 US regions characterized
by levels of ultraviolet radiation.
At the initial contact,
an adult informant was asked if there were any children
between the ages of 11 and 18 living in the household.
Ln households with I child in this age range, he or she
was selected as the respondent. In households with > 1
854 COKKINIDES et al
eligible child, the I child with the most recent birthday
was selected. The parent or guardian having primary
caregiving responsibilities was selected as the parent re-
Both surveys asked identical questions about sun pro-
tection behaviors, sunburn experiences, and sun protec-
tion attitudes. Trained interviewers administered the
surveys using structured telephone questionnaires. Each
interview took an average of 15 minutes to administer.
Up to 14 attempts were made to reach the eligible par-
ticipant. Initially, the interviewer spoke with the parent
or guardian to complete a parent-specific survey and
gather sociodemographic information and sun protec-
tion practices by the parent. Then the selected adolescent
completed the interview. Quality control logs were
maintained throughout the data collection period, and
data entry error rates were low (<3%). Standard re-
sponse rate formulas were used to calculate the screen-
ing, refusal, and response rates for both surveys.
screening rates were - 90% on both surveys; this indi-
cates that a high percentage of the eligible population
was successfully screened for age and household eligi-
bility. The percentage of households in which a parent
and child declined or the parent declined for both to
participate was 25% in 1998 and 27% in 2004. These
refusal rates seem high because a refusal could occur
before the screening process began or at any time during
the screening or interviewing process. The final response
rates, defined as the percentage of the (possibly) eligible
population who completed the interviews, were 58% in
1998 and 44% in 2004; the lower response rate in 2004
was affected by a larger proportion of telephone num-
bers with undetermined status (- 28%) as compared
with the proportion of telephone numbers with unde-
termined status (ie, ring, no answer) in 1998 (13%).
Participation rates in random-digit dialing surveys have
been reponed to range from 22% to 70%, with a median
of 50%. In recent years, declines in response rates in
random-digit dialing surveys have been reported.
this report, we focused the trend analysis only on the
youth data, whereas analysis of parent data wiU be the
subject of future reports.
Subjects were asked about their age, gender. and race,
and a series of questions related to individuals' pheno-
typic susceptibility to the sun. A validated measure of
sun sensitivity
was used to categorize susceptibility to
sun exposure based on 4 phenotypic characteristics: skin
reaction after 1 hour of exposure to the summer sun
(sensitivity to sunburn), skin reaction after repeated ex-
posure to the summer sun (ease of skin's tanning abil-
ity), the natural color of the skin, and the natural color
of the hair.
Subjects were asked 4 questions to determine the
average number of hours per week spent outdoors be-
Downloaded from www. pediatrics.org by on October 21, 2009
tween I 0:00 AM and 4:00 PM on weekdays during the
recent summer, the average number of hours per week
spent outdoors between I 0:00 AM and 4:00 PM on week-
ends during the recent summer, the average number of
days in the past 12 months spent. at the beach between
10:00 AM and 4:00 PM, and the number of days in the
past l2 months spent at an outdoor pool between
1.0:00 AM and 4:00 PM.
Subjects were asked whether they experienced a sun-
burn (defined as any reddening of the skin lasting 2=12
hours, received from being out in the sun) duri ng the
recent summer and, if so, the number of total sunburns
experienced during the summer. A follow-up question
asked those who experienced sunburns whether they
had taken any measures of sun protection before getting
their most serious sunburn in the sununer; response
options included: wearing protective clothing, hats, or
sunglasses; applying sunscreen; or staying in the shade.
A series of questions were used to ascertain the fre-
quency of pract.idng various recommended sun-safe be-
haviors when going outdoors on a very sunny day dur-
ing the summer for > I hour. The sun protective
practices included staying in the shade or under an
umbrella, wearing sun.glasses, wearing a wide-brimmed
hat, wearing protective clothing, such as a long-sleeved
shirt or long pants, applying sunscreen lotion, and ap-
plying a sunscreen with a sun protection factor 2=1 5 (SPF
15 +) when at the beach or pool. Response options for
these questions used a 5-point Likert scale ranging from
"always" to "never." The behavioral outcomes were di-
chotomized into regular (reported practicing the behav-
ior always or often) versus nonregular (reported prac-
ticing the behavior someti me, rarely, or never).
Subjects were asked a series of items using a 4-point
Likert scale ranging from strongly agree to strongly dis-
agree reflecting att.itudes about sun exposure and tan-
ning appeal. Attitudes toward sun protection included 4
items: "protecting my skin from the sun is an easy way
to stay healthy," "using sunscreen lotion allows me to
enjoy the outdoors with less worry," "spending time in
the sun without any protection can increase my chances
of developing cancer," and "my skin won't wrinkle as
fast if [ spend less time in the sun" (our sun protection
attitude index is a summary measure of these 4 items
and has a reliability coefficient Cronbach a = .62) . Two
other statements reflect attitudes related to outdoor sun
exposure and may be considered as separate potential
barriers against reducing sun exposure: "the sun feels
good on my skin" and "avoiding the sun takes the fun
out of being outdoors." These did not correlate with our
sun protection index and were considered separately.
The tanning preference factor reflects an underlying at-
titudina l preference for tan, and it included 2 items: "T
feel healthy when I have a nice tan" and "I look better
when l have a tan" (reliability coefficient Cronbach a =
Statistical Analysis
For both surveys, sampling weights were calculated that
take into account unequal probabilities of selection be-
cause of sample design, nonresponse, and poststratifica-
tion. Data management was conducted using SAS soft-
ware, version 9.0 (SAS Institute, Cary, NC).
SEs were
calculated with SLJDAAN software, version 9.0 (Re-
search Triangle Institute, Research Triangle Park, NC),
to account for the complex sample design; these were
applied to calculate weighted estimates to make results
representative of the noninstitutionalized population of
US youth aged 11 to l8 years old. Using the weighted
corresponding estimates and SEs in each survey year, we
present the weighted differences (change in the esti-
mates between 2004 and 1998) and 95% confidence
intervals (Cis) for the change in the estimates for the
measures of sun safe practices, sunburns, and attitudes
toward sun exposure protection and tanning. In addition
to categorical and univariate analysis of X
and t tests, we
also tested for the statistical significance of overall time
trends (change in the estimates between 1998 and 2004)
with pooled logistic regression analysis using dichoto-
mous outcomes, and the key predictor was time (2004 =
1 and 1998 = 0); these models also tested for significant
potential interactions between time and selected covari -
ates (age, gender, race, and sun sensitivity).
Table I lists the characteristics of youth aged 11 to I 8
years old across the 2 survey years weighted to the US
population of adolescents. ln both surveys, participants
were mostly white, with a mean age of 14 years, and
equally divided by gender. The phenotypic characteris-
tics relevant to sun sensitivity, and the mean sun sensi -
tivity scores were comparable across survey years. With
respect to patterns of sun exposure, there was a reduc-
tion in mean number of hours per week spent outdoors
between J 0:00 Af',, and 4:00 PM on weekdays (in 1998,
I 3.0 hours vs 12 hours in 2004; P = .03). No change was
noted in the mean number of hours per week spent
outdoors between I 0:00 AM and 4:00 PM on weekends.
Between 1998 and 2004, the number of days during the
past year spent at the beach increased significantly from
6.7 days in 1998 to 10.3 days in 2004 (P = .01). You.t.h
reported spending an average of 18 days at the pool
between 10:00 AM and 4:00 PM in both 1998 and 2004.
Time trend interactions with age, gender, and sun sen-
sitivity did not reveal any significant effect.
Sunburn Trends
In 2004, 69% of youth reported having had J or more
sunburns during the past summer compared with 72% 6
years earlier {Table 2). Among those who had experi -
enced a sunburn, an average of 2.9 sunburns was re-
ported, compared with 3. I in 1998. The time-by-age
interaction effect was significant (P < .01). Among the
PEDIATRICS Volume 118, Number 3, September 2006 855
Downloaded from www. pediatrics.org by on October 21 , 2009
TABLE 1 Characteristics of Study Population: US Adolescents, Aged 11 to 18 Years, Sun Survey 1998 and 2004
Characteristics 1998 2004 Change(2004-1998)
Sample Size Weighted% (95% Cl) Sample Size Weighted % (95% Cl)
Difference (95% Cl)
"' zs
(n = 1192) (n = 1613)
Age V>

11-13y 503 40.3 (37. 1 to 43.5) 536 35.5 (32.9 to 38.3) - 4.8 (- 9.0 to - 06) .06
14-15 y 289 24.6 (21.8 to 27.4) 443 27.0 (24.5 to 29.5) 2.4 ( -1.3 to 6.1)
16-18y 400 35.1 (32.0 to 38 2) 624 37.5 (34.9 to 40.3) 2.4 (-1.6 to6.4)
Mean age (95% Cl mean) 14.3 (14 1 to 14.5) 14.5 (14.3 to 146) 0.2 ( -0.8 to 12)
Boys 600 50.7 (47.5 to 539) 822 51.4 (48.6 to 54.2) 0.7 (- 35 to49) .7
Girls 592 49.3 (46. 1 to 52.5) 791 48.6 (45.8 to 51.4) -0.7 ( -4.9 to 3.5)
White 1016 78.2 (75.3 to 81.1) 1390 77.4 (74.6 to 800) -0.8 ( -4.7 to 31) .S
0 Other 176 218 (18.9 to 24.7) 213 22.6 (19.3 to 25.4) 0.8 (- 3.1 to47)
"' Sun sensitivity index 0.
l ow 316 29 8 (26.8 to 32.8) 427 30.7 (28.1 to 33.4) 0.9(- 3.1 to4.9) .9
Medium 589 46.6 (43.4 to 49 8) 778 46.3 (43.6 to 49.2) -0.3 ( -4.5 to 39)
3 High 287 23.6 (20.9 to 26.3) 404 23.0 (20.8 to 25.2) - 0.6 ( - 4.1 to 2 9)
Mean sun sensitivity index 0.5 (0.50 to 0.52) 0.51 (0.48 to 0.52) 0.0 ( - 0.04 to 0.02)

Phenotypic sensitivity to sun exposure

-o Skin reaction after 1-h exposure to summer sun
Severe sunburn with blisters 127 11.5 (9.3 to 13.7) 148 9.5 (7 9 to 112) -2.0 (-4.7 to07) .1

Severe sunburn with peeling 353 28.8 (25.9 to 31 7) 529 32.2 (29.7 to 34.9) 3.4 ( - 0.4 to 72)
Mild sunburn, sorne tanning 384 310 (280 to 34.0) 484 27.2 (24.9 to 29.6) - 38 (- 7.6 toO OJ
0 T uming darker, no sunburn 156 14.1 (11.7 to 16.3) 207 15.6 ( 13.5 to 17.9) 1.5 ( - 1.6 to 4.6)
No sunburn or tanning 163 14.6(12.2to 17.0) 230 15.5 (13.6 to 17.7) 0.9(-2.2to4.0)
'< Skin reaction after repeated exposure to surnmer sun
Repeated sunburns 286 237 (20.3 to 25.9) 401 25.6 (23 1 to 26.8) 19(- 1.7to5.5) .1 ::I
0 No suntan; freckles 73 6.3 (4.710 7.9) 71 4.0 (3.0 10 4.5) -2.3 (-4.1 to -0.4)
Mild tan 291 25.3 (22.4 to 28 2) 388 262 (23.6 to 27.5) 0.9 (-2.9to4.7)
Moderate tan 325 27.9 (25 0 to 30.8) 425 28.0 (25.4 to 29.3) 0.1 (-3.8 to40) (b
Deep tan 202 16.9 (14.5 to 19.3) 232 162 (14.0to17.3) - 0.7 (- 3.9 to 2.5)

- Natural skin color
Very fair 163 14.3 (11.9 to 16.5) 206 11.8(10.2 to 13.6) -2.5 (-5.3 to 0.4) .1 0
'-0 Fair 633 49.0 (45.910 52.3) 882 51.8 (49.0 to 54.5) 2.8 ( -1.4 to 7.0)
Olive 146 12.0 (9.8 to 14.0) 204 12.6(10.8to 14.5) 0.6 (-2.2 to 34)
Light brown 214 21.5 (18.7 to 24.3) 254 190 (16.8 to 215) - 2.5 (- 6.1to 11)
Dark brown/black 27 32 (1.9 to 4.5) 47 4.8 (3.6 to 6.5) 1.6 ( - 0.3 to 3.5)
Sun-exposure characteristics
Number of hours per week spent outdoors on 1183 13.2 (12.5 to 13.9) 1565 12.0( 11.4 to 12.6) - 1.2 (- 2.0 to -0.3) .03
weekdays, weighted mean
Number of hours per week spent outdoors on 1172 7.3 (6.8 to 7 8) 1571 7.1 (6.7 to 7.4) -0.2 (-0.6 to 0.3) .2
weekends, weighted mean
Number of days in past 12 mo spent at the beach, 1185 6.7 (5.9 to 75) 1549 10.3 (9.2 to 11.4) 3.6(2.1 to5.1} .01
weighted mean
Number of days in pasr 12 mo spent at an outdoor 1174 182 (168to 19.6) 1533 18.3 (17.0to 195) 0. 1 (- 1.6to 18) .7
pool, weighted mean
Between the hours of 10:00 1>.'A and 4:00PM.
"' 0.
~ Q
.. - R
N '-"'
g ~
TABLE 2 Trends in the Prevalence of Sunburns and Mean Number of Sunburns Experienced During the Summer in US Youth, Sun Survey 1998 and 2004
Variable Percentage Who Experienced Sunburns During the Among Those Who Experienced Sunburns
Mean No. of Sunburns Percentage Who Reported Applying Sunscreen as a
Form of Sun Protection Before Occurrence of Most
Serious Sunburn
1998,% 2004,% Change Difference 1998, 2004, Change Difference (Mean) 1998, % 2004,% Change Difference
(95% Cl Difference) Mean Mean (95% Cl Difference) (95% Cl Difference)
Overal l 72.2 68.7 - 3.5 ( -7.5 to 0.5) 3. 10 2.90 -02(-06to02) 39.2 473 8.1" (3.1 to 130)
11-13 75.1 67.3 - 7.8 (- 14.5 to - ll)b 2.84 2.69 - 02(- 06to0.3) 47.2 58.5 1 13 (3.0 to 19.6)
14-15 78.9 69.6 -9.3(- 17.1 to - 1.5) 3.23 3.00 -0.2 (- 1.2 toO.?) 36.5 48.0 11.5 (1.9 to 21.1)
16-18 64.3 70.0 5.7 (- 1.3 to 12.7) 3.36 2.97 - 0.4 ( - 1.3 to 0.5) 30.6 36.6 6.0(- 1.9to 13.8)
Boys 72.9 66.0 - 6.9 ( - 13.8 to 0 1) 2.96 2.90 - 0.1 (- 0.7to0.6) 30.2 38.3 8.P(1.4to 14.8)
Girls 72.0 71.5 - 0.5 (- 5.8to48) 324 2.90 - 0.3 ( - 0.9 to 0.3) 48.2 56.1 7 9 (08 to 15 0)
White 79.4 76.3 -3.1 (-7.1 to0.9) 3.21 3.02 -0.2 ( -0.7 to 0.3) 40.2 45.6 5.4 (0.2 to 10.6)
Nonwhite 46.8 43.0 - 3.8 ( - 14.6 to 7.0) 2.56 2. 10 - 0.5( -2.1 to 1.2) 33.6 56.0 22.4 (6.6 to 38.2)
Sun sensitivity index
Low 55.1 52.2 - 29(- 1Uto5.5) 2.53 2.20 - 0.3 (-1 OtoOA) 29.1 49.0 19.9 (9.6 to 302)
Medium 77.0 73.4 - 3.6( - 9.0to 1.8) 3.07 2.83 - 02 (- l.OtoO.S) 40.1 44.0 39(- 3.0to 10.8)
High 85.5 84.5 -1.0(- 7.8 to 58) 3.67 3.55 - 0. 1 (- 09to0.7) 46.2 52.0 5.8 (- 3.8 to 15.4)
P< .001.
b P < .01 interaction term (age by year).
16- to 18-year-olds, sunburn prevalence increased by
6% (P > .05), whereas at younger ages, sunburn prev-
alence decreased significantly by among II- to
13-year-olds and 9% in I 4- to I 5-year-olds (P < .0 l;
Table 2). ln a multivariate analysis controlling for con-
founders (age, gender, race, and sun sensitivity), we
found a positive association between having had sun-
bums and regular use of sunscreen when going outdoors
in the summer (data not shown). Among those who
sunburned, the most common reported measure of pro-
tection before experiencing the most serious sunburn of
the summer was application of sunscreen, whereas other
measures of sun protection were very infrequent
( <2%). Compared with 1998 levels, the prevalence of
wearing sunscreen before experiencing a serious sun-
bum increased significantly (overalL from 39% to 47%;
P < .00 1; Table 2). During the 6-year period, there were
larger trends that showed increases in the percentage of
youth reporting sunscreen use before getting a serious
sunburn among those aged L L to 1. 5 years, boys, and
those with more sun-resistant skin traits and of non-
white race (Table 2).
Sun Protection Practices Trends
Between 1998 and 2004, there were signifi cant increases
in the overall prevalence of regular (always/often) use of
sunscreen (8% increase, from 31% to 39%) when going
outdoors on a sunny day for > L hour. Statistically sig-
nificant increases in the regular use of sunscreen were
observed in most subgroups examined, but these in-
creases were greatest among older youth (who reponed
the least use of sunscreens) and among girls and those
who with mectium or high sun sensitivity (who had the
highest use of sunscreen). The increase was restricted to
white youth. The percentage of respondents who re-
ported regular use of sunscreen with SPF I 5 +, when at
the beach or at the pool, did not change from J 998 to
2004. Other recommended sun-safe practices, such as
wearing sunglasses, wearing protective clothing (long-
sleeve shirts or long pants), and staying in the shade
were less conunon. The regular use of wide-brimmed
hats in this population was very infrequent (- 5% in
2004). No measurable changes in these practices were
observed between 1994 and 2004, except for an increase
in the prevalence of regular use of long-sleeve shins or
long pants among white youth (P < .05; Table 3). Sim-
ilar trends were observed when comparing regular (al-
ways or often) to always use (Table 3).
Trends in Attitudes
There was a small but statistically significant increase in
the percentage of respondents who reported positive
attitudes (strongly agree or agree) toward the benefits of
sun protection (3.51<, increase; P < .05). This change in
attitude mostly involved white youth. This change was
primaril y influenced by 2 attitudinal/beli efs about the
858 COKKINIDES et al
benefits of sun protection (ie, "protecting my skin from
the sun is an easy way to stay healthy" and "spending
time outdoors without any sunscreen increases my
chances of developing cancer"; Table 4). Attitudes re-
lated to tanni ng preferences continued to be common in
2004: 67.8% of youth strongly agree or agree with the
statement "I look better when I have a tan" and 55% of
youth strongly agree or agree with the statement "I feel
healthy when I have a nice tan," and across youth
characteristics, slight declines in trends for these single
items were observed between 1998 and 2004. As mea-
sured by an overall index of these items (tanning pref-
erence), there was a signifi cantly modest decrease in the
prevalence of positive (strongly agree or agree) attitudes
(from 56.5% to 51.4%; 5.1% decrease; P < .05; Table 4).
In 2004, the proportion of youth who reponed strongly
agreeing or agreeing with the statement "the sun feels
good on my skin" significantly increased (11.6%; P <
.001) compared with 1998 levels, and this pattern of
change was consistent across all of the subgroups (Table
There have been extensive efforts by government agen-
cies, foundations, and other organizations in recent
years to promote sun protection behaviors among chil-
dren and adults in the United States for the purpose of
preventing future melanomas and other skin cancers.
Unfortunately, there has been little systematic effort to
assess the effect of these initiatives on trends in behavior.
Particularly lacking are repeated surveys of population-
based samples with consistent methodology to allow
systematic assessment of trends. Most studies have been
cross-sectional surveys.


Some of these studies
have been regional in scope,


others were based on
nationally representative samples,

and 1 had a large
nonpopulation-based sample.
Despite the varied meth-
odology, these surveys have shown that adolescents
have low levels of sun protection behaviors and high
levels of sun exposure and sunburns. Although findings
of these studies have been i11formative, they do not
address the issue of how much change there has oc-
curred in US youths' sun protection and sun exposure.
Our study collected data from 2 time periods on mul-
tiple behavioral and attitudinal end points. We found
that. trends in US youth sun protection were mixed
during the interval of 1998 to 2004. The prevalence of
youth experiencing sunburns during the summer was
high and generally stable between 1998 and 2004
in 2004). However, there were significant trends
of decreasing prevalence of sunburns among younger
teens. Tt is li kely that this age -specific trend may relate in
part to greater parental influence in younger adolescents
(as well as children), because several studies indicate
that parental vigilance and parent's sun protection be-
haviors are related to greater sun protection practices
Downloaded from www. pediatrics.org by on October 21 , 2009
TABLE 3 Trends in the Prevalence of Sun-Safe Behavior Practices When Going Outdoors on a Sunny Day During the Summer for> 1 Hour in US
Youth, Sun Survey 1998 and 2004
Sun Protection Apply Sunscreen When Going Apply Sunscreen With SPF 15 + Wear Wide-Brimmed Hat
Measures Outdoors in Summer at the Beach/Pool
1998, 2004, Change, % 1998, 2004, Change,% 1998, 2004, Change, %
% % (Change, 95% Cl) % % (Change, 95% Cl) % % (Change, 95% Cl)
Total (often/always) 31.4 39.4 8.0 (4.0 to 12.0) 58.4 56.4 - 2 0 (- 6.5 to 2.5) 3.8 4.9 1.1 ( - 0.7 to 2.9)
T oral {always) 9.7 15.8 6. Jd (3.3 to 88) 32.2 29.0 - 3.2( - 7.40 10) 0.4 2.3 1.9 ( - 0.7 to 3.0)
11-13 41.9 48.1 6.2 ( -0.7 to 13.1) 63.4 64.6 1.2 {-5.9 to 8.3) 5.0 5.0 0.0 (-3.1 to3.1)
14-15 26.5 35.5 901>(1.2to 168) 586 53.0 -56(- 14.5 to3.3) 2.0 4.9 2.9{-0.3to6.1)
16-18 22.9 330 10 1 (4.0 to 16.2) 52.0 51.1 -09(-8.6to68) 3.5 4.8 1.3 ( -1.7 to 4.3)
Boys 25.1 30.0 49{- 0.4to 10.2) 56.0 53.1 - 2.9 ( - 9.6 to 38) 3.8 6.5 2.7 (0.1to 5.3)
Girls 37.9 48.6 10.7" (4.9to 16.5) 60.6 59.8 -0.8 (-6.8 to 5.2) 3.8 3.3 -o.s (-2.9 to 1.9)
White 33.5 430 9.5 (5. 1 to 13.9) 60.3 62. 1 18{-2.9to6.5) 32 4.9 1.7 ( -0.2 to 3.6)
Nonwhite 25.2 25.3 0.1 { -9.5to 9.7) 509 36.8 -14.1 (-26.3 to -1.8) 5.7 53 -0.4 (-5.5 ro4.7)
Sun sensitivity index
Low 21.0 23.6 2.6 { - 4.1 to 9.3) 50.8 42.9 - 79( -1 7.0 to 1.2) 5.1 4.5 - 0.6 ( - 4.4 to 3.2)
Medium 30.7 41.4 10.7 (5.0 to 16.4) 57.1 57.2 0.1 (-6.3 to6.5) 2.9 4.9 2.0 ( -0.3 to4.4)
High 46 1 55.6 9.5 {0.9 to 18.6) 69.0 74.9 5.9 (- 2.4 to 14.2) 3.8 5.8 2.0 ( -1.7 to 5.7)
Wear Sunglasses Wear Clothing Stay in Shade or Under Umbrella
(Long Pants or Long-Sleeve Shirt)
1998, 2004, Change,% 1998, 2004, Change, % 1998, 2004, Change,%
% % (Change, 95% Cl) % % (Change, 95% Cl) % % (Change, 95% Cl)
Total (often/always) 32.2 321 -0.1 (-40to3.8) 21.7 22.8 1.1 (-2.5to4.7) 21.8 21.7 -0.1 {-3.7to3.5)
Total {always) 13.2 16.1 2.9(-0.01 !05.8) 7.3 10.8 3.5 {1.0 to 5.0) 2.8 2.2 -0.6 (-2.1 to 0.9)
11-13 23.5 25.4 1.9{- 4.1 to 7.8) 21.2 21.0 - 0.2 ( - 6.0 to 5.6) 23.9 25.6 1.7 ( - 4.3 to 7.7)
14-15 30.4 27.3 -3.1 (-10.6to4.4) 18.9 24.3 5.4(- 1.8to 12.6) 19.4 20.6 1.2 ( -6.0 to 8.4)
16-18 43.4 41.8 -1.6 ( -8.5 to 5.4) 24.1 23.4 - 0.7 {-6.9to 5.6) 20.8 19.0 -1.8 (-7.6 to4.0)
Boys 26.2 24.4 -1.8 { -6.9 to 3.3) 23.9 21.9 -20(-71 to3.1) 21.5 20.5 - 1.0 (-6.0 to4.0)
Girls 38.2 40.1 1.9{ - 4.1to 7.8) 19.4 23.7 4.3 {- 0.8to9.4) 21.9 230 1.1 ( - 4.2to6.4)
White 32.9 32.9 0.0 {-4.3 to 4.3) 18.8 22.7 3.9b {0.2 to 7.6) 18.6 18.8 0.2 (-3.4 to 3.8)
Nonwhite 30.5 290 - 1.5 {- 11.s to 8.5) 32.8 22.7 -10.1b { - 19.8 to - 0.3) 33.3 31.8 - 1.5 (- 11.8 to 8.8)
Sun sensitivity index
Low 34.4 28.7 -5.7 ( -13.4 to 1.9) 261 24.7 -1.4 (-8.7 to 5 9) 22.7 21.4 -1.3 (-8.5 to 5.9)
Medium 31.5 313 - 02 {- 5.7 to 5.3) 20.5 20.4 - 0.1 (- 51 to4.9) 20.6 20.2 - 0.4{- 5.4 to46)
High 30.4 38.1 7.7 ( - 0.3 to 15.7) 18.5 25.3 6.8( - 0.4!0 14.0) 22.7 25.6 2.9 ( - 4.4 to 10.3)
Prevalence of regular sun-safe behavior practiced as always or often.
P< .001.
bf< .05
and reduced likelthood of sunburn experiences in their The beach setting, in particular, provides a greater op-

Between 1998 and 2004, the prevalence portunity for exposure to intense (intermittent) UV ex-
of regularly practicing measures of sun protection were posure and is associated with extended time of UV radi-
low, whereas opportunities for high (intermittent) UV ation exposure. J(>.37
exposure seem to have increased (ie, increased number We also found increasing trends in the proportion of
of days spent at the beach). rnterestingly, among older youth who reported wearing sunscreen before getting
youth, we found a larger trend increase in the numbers their most serious sunburns and that regular use of
of days spent at the beach than in younger youth (data sunscreen was independently associated with experienc-
not shown). However, we lack the information to deter- ing sunburn in the past summer, after adjusting for
mine whether the trend increase in the numbers of days confounders. These findings, in parallel to our findings
spent at the beach is related to proximity to the beach, of an increasing trend in regular sunscreen use as de-
popularity of the beach as family vacation, or more scribed next. may suggest that sunscreen use is marker
affordable and frequent travel to beach by older youth. of excessive UV exposure and that, whereas intending to
PEDIATRICS Volume 118. Number 3, September 2006 859
Downloaded from www. pediatrics.org by on October 21 , 2009
"' zs
TABLE4 Trends in Attitudes Toward Sun Exposure and Tanning Appeal in US Youth, Sun Survey 1998 and 2004
Proport ion of Youth Who Total Age,y m

"Strongly Agree or "Agree"
11-13 14-15 16--18
1998, 2004, Change 1998, 2004, Change 2004-1998, 1998, 2004, Change 2004- 1998, 1998, 2004, Change 2004-1998,
% % 2004- 1998, % % %(95%( 1) % % %(95%(1) % % %(95%CI)
Attitudes on benefits of sun protection
Protecting my skin from the sun is an easy 87.8 92.8 5.0 (2.2 to 7.8) 87.2 92.4 5.2b (0.5 to 9.8) 89.3 93.6 4.3 (-1.1 to 96) 87.4 92.8 5.4b (0.5 to I 0.2)
0 way to stay healthy
Using sunscreen allows me to enjoy the 84.8 84.7 - 0.1 (-3.4 to 3.2) 90.4 89.2 - 1.2(- 5.7to3.3) 83.4 85.1 1.7(-4.7 to8.1) 78.9 79.9 1.0 (- 5.3 to 73)

::I outdoors with less worry
Spending time outdoors without any sun 91.0 95.1 4.1" (1.6 to 6.6) 90.9 94.8 3.9" (0.0 to 7.8) 90.4 95.5 5. 1"(0.1 to 10.1) 91.4 95.1 3.7 (- 0.4 to 7.8)
"' 0. protection can increase my chances of
developing cancer
My skin won't wrinkle as fast if I spend less 73.0 65.0 - s.o(- 12.4 to - 3.6) 68.8 58.0 - 10.8" (- 18.4 to - 3.2) 77.8 68.0 - 9.8(- 18.0to - 1.6) 74.2 69.3 -4.9( - 11.9 to2.1)
time in the sun

Sun protection index (1) values s2 86.3 89.8 3.5b (0.5 to 6.5) 87.8 90.2 2.4 (- 2.2 to 7.0) 87.0 913 4.3 (-1.1 to9.6) 84.0 88.3 4.3 (- 1.1 to9.7)
Attitudes on outdoor sun exposure

The sun feels good on my skin 71.7 83.3 11 6 (7.5 to 1 5.6) 66.0 80.0 14.0 (7.0 to 20.7) 73.8 86.5 12.7< (5.0 to 20.4) 76.4 84.6 8.2 (1.7to 14.6)
(b Avoiding the sun takes the fun out of 44.8 50.0 5.2" (0.7 to 9.7) 42.9 51.1 8.2b (0.8 to 15.5) 51.4 46.6 - 48(-13.7 !04.1) 42.4 51.3 8.9b (1.6 to 161)
being outdoors

Tanning preferences
I look better when I have a tan 71.2 67.8 - 3.4( - 7.7100.9) 64.1 618 - 2.3 (-9.7 to 5.1) 76.7 69.0 - 77( - 161 100.8) 75.3 72.4 -2.9(-9.6to38)
0 I feel healthy when I have a nice tan 53.4 55.0 1.6(- 3. 1 to63) 48.2 50.7 2.5( -5.1 to 10.1) 57.8 57.0 - 08 ( - 10.2 to 8.6) 56.4 57.0 06(- 7 t to8.6)
Tanning preference index (2) values s2 56.5 51.4 -5.1b (-9.5 to -0.7) 49.5 45.6 - 39(-11.1 to 3.3) 630 52.1 - 10.9b(-19.6to -22) 60.2 56.2 -4.0(- 11.2 to32)
Gender Race
Boys Girls White Nonwhite
1998, 2004, Change 2004-1998, 1998, 2004, Change 2004-1998, 1998, 2004, Change 2004- 1998, 1998, 2004, Change 2004- 1998, .....

% o/o %(95%CI) % % %(95%(1) o/o o/o %(95%(1) % % % (95%CI)
0 Attitudes on benefits of sun protection
'-0 Protecting my skin from the sun is an easy 85.6 913 5.7 (1.3 to 10.0) 89.9 94.4 4.5(0.8 to 8.1) 87.9 95.3 7.4 (4.5 to 102) 86.7 84.3 -2.4 (- 10.5 to 57)
way to stay healthy
Using sunscreen allows me to enjoy the
outdoors with less worry
80 l 82.2 2.1 (-2.9to 7.1) 89.5 87.5 - 2.0(-6.1 to2.1) 853 881 28( - 0.4 to60) 82.1 72.7 -9.4 (-18.9to01)
Spending time outdoors without any sun
protection can increase my chances of
89.6 94.9 5.3" (1.6 to 8.9) 92.4 95.3 2.9 ( -0.4 to 6.2) 92.1 97.0 4.9 (2.5 to 7.2) 86.2 890 2.8(- 4.8 to 10.4)
developing cancer
My skin won't wrinkle as fast 1f J spend less 63.5 59.5 - 4.0 (-10.7to 2.7) 79.2 70.6 - 8.6"(- 14.2 to -2.9) 74.8 68.9 - 5.9( - 10.4to - 1.3) 67.0 51.7 - ISJb (- 26.8 to - 3.7)
time in the sun
Sun protection index (1) values s2 83.0 86.8 3.8 ( -0.7 to 8.3) 89.6 92.8 32 (-0.5 to 7.0) 87.9 93.3 5.4 (2.7 to 8.1) 79.5 77.7 - 1.8(- ll.Oto 73)
Attitudes on outdoor sun exposure
The sun feels good on my skm 726 82.6 1 o.oa (4.2 to 15 7) 70.9 841 13.2 (7.5 to 18.8) 76.5 86.0 9.5 (5.5 to 13.5) 54.7 74.6 19.9 (8.8 to 30.9)

"' 0.




.. - R
N '-"'




TABLE 4 Continued
Proportion of Youth Who
"Strongly Agree or "Agree"
Avoiding the sun takes the fun out of
being outdoors
Tanning preference
I look better when I have a tan
l feel healthy when I have a nice tan
Tanning preference index(2) values s2
Attitudes on benefits of sun protection
Protecting my skin from the sun is an easy
way to stay healthy
Using sunscreen allows me to enjoy the
outdoors with less worry
Spending time outdoors without any sun
protection can increase my chances of
developing cancer
My skin won't wrinkle as fast if I spend less
time in the sun
Sun protection index (1) values S2
Attitudes on outdoor sun exposure
The sun feels good on my skin
Avoiding the sun takes the fun out of
being outdoors
Tanning preference
I look better when I have a tan
I feel healthy when I have a nice tan
Tanning preference index (2) values S2
2004, Change 2004-1998, 1998,
% %(95%( 1) %
51.7 8. 7" (2.4 to 15.Q) 46.7
70.6 -2.7(-8Sto31) 69.2
57.2 -0.7 ( - 7.3 to 5.9) 49.0
532 - 1.8) 52.0
Low Sun Sensitivity
2004, Change 2004-1998, 1998,
o/o %(95%(1) o/o
87.0 3.6 (- 3.0 to 1 0.2) 88.6
75.2 -2.3 (-9.8to52) 86.2
92.5 6.6b (0.8 to 12.3) 93.2
59.7 -5.8(-14.61030) 77.0
82.4 5.0 (- 2.0 to 12.0) 89.1
82.9 107- (2.8to 18.5) 73.9
54.6 7.2 (-1.3 to 15.7) 46.6
69.2 -30(-11.2to52) 74.6
56.4 -1.6 ( - 10.5 to 7.3) 54.5
53.5 -6.1 (- 14.5to2.3) 58.5
Girls White Nonwhite
2004, Change 2004-1998, 1998, 2004, Change 2004- 1998, 1998, 2004, Change 2004-1998,
% %(95%(1) o/o % % (95%CI) % % %(95%(1)
48.3 1.6(-4.7to7.9) 43.4 51.0 7.6b (2.7 to 12.3) 49.3 47.0 -2.3 (- 13.7 to 9.1)
65.0 -4.2(- 10Sto20) 76.0 73.6 -24 ( -6.7 to 19) 530 47.8 -5.2 (-16.9 to6.5)
52.5 3.5 (- 31 to 10.1) 58.0 59.0 1.0(-4.0to6.0) 37.8 40.5 2.7 ( - 8.6 to 14.0)
49.5 -2.5 ( -8.7 to 3.7) 60.7 56.7 - 4.0(-8.6to06) 50.5 33.4 - 64)
Medium Sun Sensitivity High Sun Sensitivity
2004, Change 2004-1998, 1998, 2004, Change 2004- 1998,
o/o % (95% en o/o o/o o/o (95% Cl)
95.0 6.4 (2.7 to 1 O.O) 91.2 97.0 5.8 (1.6 to 1 O.O)b
86.9 0.7 (-36to50) 905 93.4 2 9 (-1.9to 7.7)
95.5 2.3 ( -0.8 to 5.4) 92.5 98.2 57" (1.7to 9.6)
67.7 - 9.3 (-15.1 to - 3.4) 74.7 67.9 - 6.8(-15.7 to 2.1)
92.2 3.1 (- O.Sto6.7) 91.7 95.8 4.1 ( -0.3 to 8.5)
85.3 11.4(5.9!0 16.9) 66.4 81.0 14.6 (5.7 to 23.4)
482 1.6(-4.8to8.0) 37.9 48.0 IO.Ob (I I to 18.9)
71.7 -29(-8.7to29) 62.2 58.2 -40(-136to 5.6)
57.3 2.8 (- 3.8 to 9.4) 44.5 48.0 3.5( - 6.3 to 13 1)
54.6 -3.9 (-10.0 to 2.3) 48.5 41.8 - 6.7 (-15.6 to 2.3)
( 1) indicates percentage prevalence score :s2 for benefits of sun protection attitudes index consisting of four items: ' protecting my skin from the sun is an easy way to stay healthy; using sunscreen allows me to enjoy the outdoors with less worry; 'spending time
outdoors without any sun protection can increase my chances of developing cancer.' and 'my skin won't wrinkle as fast if I spend less time in lhe sun'; measured on a 4point Likert these were 1 = strongly agree, 2 = agree, 3 = disagree, and 4 = strongly disagree
(reliabihtycoefficiem Cronbach a =.62). {2) indicates percentage prevalence score :s2 for tanning appeal attitudes index consisting of two nerns: 'I look better when I have a tan and "I feel healthy when I have a nice tan"; measured on a 4point Likert scale these were
1 = strongly agree. 2 = agree. 3 = disagree. and 4 = strongly disagree (reliability coefficient Cronbach a = .69).
protect themselves from the sun, inadequate application
of sunscreen (ie, not enough appl ied or certain areas of
the body missed) or sunscreens degraded or washed off
from swimming, sweating, or other activities may un-
dennine the sunscreen efficacy for prevention of sun-
burns.38-40 Our findings showed very few improvements
in trends for the regular (routine) practice of sun pro-
tection behaviors and in attitudinal preferences to tan-
ning, whereas some trends in barriers to reduced sun
exposure increased. Specifically, between 1998 and
2004, the only positive increases were for the regular use
of sunscreens, and there was a slight and nonsignificant
reduction (2%) in the prevalence of regular tJse of sun-
screen with SPF 15 + when at the beach or pool (58% vs
56%). Sunscreen use continues to be the main form of
protection from the sun practiced by youth. Our study
found no changes, and the regular use of other sun
protection measures (ie, protective clothing, seeking
shade, and wearing sunglasses) was less prevalent. In
particular, the regular use of wide-brimmed hats in this
population was very infrequent. Research is needed to
understand the reasons or barriers (eg, attitudes, norms,
or contextual settings) related to the low adoption of
such practices.
Modest but significant increases in the prevalence of
youths' attitudes related to benefits of sun protection
occurred between J 998 and 2004; youth in general seem
to hold largely positive attitudes (knowledge-related
benefits) toward sun protection (with the exception that
knowledge levels about the effect of UV exposure and
skin photograph-aging effects were moderate). Educa-
tional skin cancer prevention efforts in schools, outdoor
aquatic settings, and media campaigns may have con-
tributed to these improvements in attitudes/knowledge
during the 6-year petiodY''s
In addition, although
trends showed modest decreases, youths' tanning pref-
erences cont.inued to be prevalent, and barriers related
to sun exposure ("sun feels good on my skjn" and
"avoiding the sun takes the fun out of being outdoors")
increased during this period. Tms may suggest that skin
cancer prevention efforts to change such attitudes may
been (somewhat) less effective in youth or that other
factors, such as social norms toward tanrung, fashion
trends, peer pressure, and the lack of policies to support
sun-safe environments, may have been more influential
so as to preclude changes in attitudes.
Because such
attitudes are determinants of sun exposure and sun pro-
tection behaviors, LO.n-
it is important to develop
better educational messages and strategies targeted at
There were several strengths and limitations in this
study. The surveys used the same instruments, which
include questions for multiple behavior and attitude end
points, and tbe same probability sampling methodology
to allow for generalizing estimates to the entire popula-
tion of US youth aged l1 to 18 years. However, the
862 COKKINIDES et al
telephone survey methodology limited the generalizabil -
ity of results to persons with a telephone ( ~ 9 5 of US
Particularly in the most recent survey, a
substantial proportion of households were not reached
despite multiple attempts, or selected subjects did not
complete the telephone interview, and, thus, a potential
for unmeasured bias may be present. The sampling de-
sign allowed for a representative sample of the US pop-
ulation, but certain low-frequency population subgroups
were not well represented, because cost issues did not
permit oversampling. We note that skin cancer is mgher
among the white-skinned population, a group that is
well represented in these surveys. Data used in this
study were based on self-reports and possibly subject to
recall bias or social desirability bias. However, surveys
were conducted in an anonymous fashion, and ques-
tions gathered information about recent events and ex-
periences occurring either during the past summer (2- 4
months before the date of interview) or within the last
Internationally, Australia has served as a model of a
national (in scope) comprehensive skin cancer preven-
tion program, which has used systematic ongoing pop-
ulation-based tracking of disease and behavioral end

5 1
Early in the 1970s, Australia developed a
national health care policy that made skin cancer pre-
vention a public health ptiority; the Australian public
educational campaigns known as Slip! Slop! Slap! (slip
on a shirt, slop on some sunscreen, and slap on a hat)
and the SunSma.rt program, the most recent national
program, involved multipronged approaches in various
settings (schools, worksites, and health care systems), as
well as environmental supports and policies.
In the
United States, several broad-based national, state, and
private skin cancer primary prevention efforts have been
initiated, and most target primarily chi ldren.

light of our findings of mixed results on sun protection
among US youth, there is a need to further expand
efforts in targeting youth a11d to improve the effective-
ness of these programs. In addition to the assessment of
skin cancer disease trends, there is a need for concomi -
tant surveillance activities to monitor sun protection
practices and sunburn rates in youngsters (and in sus-
ceptible high-risk populations), thereby allowing the
United States to measure success in skin cancer preven-
tion and reduction in disease burden as incorporated in
Objective 3.9 of the Healthy People 2010 goals.
Despite moderate increases in the regular use of sun-
screens among US youth between 1998 and 2004, there
has not been adequate adoption of regular use of other
recommended sun protection practices. Moreover, al-
though the decreasing trends in sunburns among
younger teens may give some cause for optimism re-
garding future trends in skin cancer, there was little
Downloaded from www. pediatrics.org by on October 2 1, 2009
overall change in sunburn trends during this period;
thus, summer sunburns experiences among youth re-
main high. The findings suggest that current efforts in
the United States toward skin cancer prevention have
not been sufficienr to have a major impact on overall sun
protection among youth.
Thus, there is a need to
strengthen the implementation of skin cancer preven-
tion programs and interventions along with environ-
mental structural supports and policies




may help shape the social and physical environment and
promote sustainability or these programs. To strengthen
implementation or skin cancer prevention, research is
needed on effective mechanisms to broadJy diffuse skin
cancer prevention intervention and sun-safety policies
in communities
and in schools
and for effective
ways to increase the role of pediatri cians and health care
providers wi th their patients
R about sun safety practices
to minimize risks associated with excessive UV exposure.
Data collection for the 1998 Sun Survey I was funded by
the American Cancer Society as an intramural research
project. Data collection for the 2004 Sun Survey n was
financially supported by Neutrogena Corporation.
I. American Cancer Sodety. Cancer Facts and Figures. Atlanta, GA:
American Cancer Sodety; 2006
2. Jemal A, Devesa S, llartge P, Tucker M. Recent trends in
cutaneous melanoma incidence among whites in the United
States. J Nat/ Cancer lnst. 2001;93:678-683
3. Howe H, Wingo P, Thun M, et al. Annual report to the nation
on the status of cancer ( 1973 through 1998), featuring cancers
with recent increasing trends. J Nat/ Cancer lnst. 2001;93:
4. Christenson L, Borrow111 an T, Vachon C, et al. Incidence of
basa l cell and squamous cell carcinomas in a population
younger than 40 years. JAM/I. 2005;294:681- 690
5. Harris R, Griffith K, Moon T. Trends in the incidence of non-
melanoma skin cancers in southeastern Arizona, 1985- 1996.
JAm Acad Dmuatol. 2001;45:528-536
6. Karagas M, Grccnl>c.:rg E. Spencer S. Stukd T, Molt L. New
Skin Cancer Study Group. Increase in incidence
rates of basal cell and squamous cell skin cancer in New Hamp-
shire. USA. lnt J Cancer. 1999;81:555-559
7. International Agency for Research on Cancer. Monograph on the
Evaluation of Carcinogmic Risks to Humans; Ultraviolet Radiation,
vol. 55. Lyon, France: Int ernational Agency for Research on
Cancer; t 992
8. B, Kricker A. The epidemiology of UV induced skin
cancer. J Phoroclrmr Plrotobiol B. 2001;63:8-18
9. Whiteman D. Whiteman C, Green A. Childhood sun exposure
as a risk factor for melanoma: a systematic review of epidemi -
ologic studies. Cancer Causes Control. 2001;12:69- 82
ro. Davis K, Cokkinidcs V, Weinstock M, O'Connell M, Wingo P.
Summer sunburn and sun exposure among US youths ages 11
to 18: national prevalence and associated factors. Pediatrics.
2002; 110:27- 35
I I. Mermelstein R, Riesenberg L. Changing knowledge and att i-
tudes about ski n cancer risk factors in adolescents. Health Psy-
cho/. 1992;1 1:37 1-376
12. Robinson J, Rademaker A, Sylvester J. Cook B. Summer sun
knowledge, attitudes. and behaviors of Midwest ad-
olescents. Pm Med. I 997;26:364-372
13. Cokkinides V, Johnston-Davis K. Weinstock M, et al. Sun
exposure and sun-protection behaviors and attitudes among
U.S. youth, II to 18ycarsofage.PrMJMed.2001;33:141-15l
14. Alberg A, Herbst R, Genkinger J. Duszynski K. Knowledge.
allitudes, and behaviors toward skin cancer in Maryland
youths. J Ado/esc Health. 2002;31:372-377
15. Geller A, Colditz G, Oliveria S, et al. Use of sunscreen, sun-
burning rates. and tanning bed usc among more than I 0 000
US children and adolescents. Pediatrics. 2002; t09:1009-l 014
16. Glanz K, Saraiya M. Wechsler H. Guiddines for school pro-
grams to prevent skin cancer. MNIWR Recomm Rep. 2002;51 (RR-
4):1- 18
17. Saraiya M, Glanz I<. Briss P, et al. Preventing skin cancer:
findings of the Task Force on Community Preventive Services
On reducing Exposun: tO Ultravioletl.ight. NL'viWR Recomm Rep.
2003;52(RR- I 5): 1- 12
18. Edl ich R. Wimers K, Cox M, et al. National health strategies to
reduce sun exposure in Australia and the United States. J Long
Tenu Eff Med Imp/allis. 2004; 14:215-224
19. Geller A, Cantor M, Miller D, et al. The Environmental PrOtec-
tion National SunWise School Program: sun protec-
tion education in US (1999-2000). JAm Acad Dermato/.
20. Montague M, Borland R. Sinclair C. Slip! Slop! Slap! and
SunSman, 1980-2000: Skin cancer control and 20 years of
population-based campaigning. Health Educ Behav. 2001;28:
21. Waksberg J. Sampling methods for random digit dialing. JAm
Stat Assoc. 1978:73:40- 46
22. National Weather Service. UV Index Forecast. Available at:
www .cpc. nce1). noaa .gov /prod ucts/st ra tOsphere/u v _index /uv _
current.htrnl. July 30, 2004
23. The American Association for Public Opinion Research. Stan-
dard Definitions: Final Disposition of Case Codes and Outcome Rates
and Otucome Rates for Surveys. Ann Arbor, MI: American Asso-
ciation for Publ ic Opinion Research (AAPOR); 2004
24. Massey J. O'Connor D, J<rokt i K. Response Rates in Random Digit
Dialing (RDD) Telephone Surveys. Alexandxia, VA: American Sta-
tist ical Association; 1997
25. Weinstock MA. Assessment of sun sensitivity by questionnaire:
validity or items and formulation of a prcdicrion rule . .! Clin
Epidemic/. 1992;4 5:5117-552
26. SAS Institute, Inc. SAS Procedure Guide. Cary, NC: SAS Insti tute,
Inc; 1999
27. SUDAAN Usa's Manual. Professional Software for Survey Data
Analysis !computer program]. Version 9.0. Research Triangle
Park, NC: ReS('arch Triangle Institute; 2003
28. Saraiya M. Glanz K. Briss P, et at. Interventions to prevem skin
cancer by reducing exposure to ultraviolet radiation: a system-
atic review. Am J Prrv Med. 2004;27:422-466
29. Banks B. Silvennan R. Schwanz R, Tunnessen W. Attitudes of
teenagers toward sun exposure and sunscreen use. Pediatrics.
I 992;89:40-4 3
30. Reynolds K, Blaum J, Je5ter P, Webs H. Soong S, Didememe R.
Predictors of sun exposure in adolescents in a southeastern
U.S. population. J Ado/esc Health. 1996;19:409-41 5
31. Hall H, Jones S, Saraiya M. Pn:valcncc and correlates of sun-
screen use among US high school students. J Sch Health. 200 L;
32. Hall H, Jurgensen C. McDavid K. Kraft J, Breslow R. Protection
from sun exposure in US white children ages 6 months to I I
years. Public Henltlr Rep. 2001;116:353-361
33. Johnson K. Dnvy L. Boyett T, Weathers L, Roetzhcim R. Sun
PFDIATRICS Volume 118, Number 3, September 2006 863
Downloaded from www.pediatrics.org by on October 2 1, 2009
protection practices for children: knowledge, atti tudes. and
parent behaviors. Arch Pediatr Ado/esc Med. 2001; 155:891- 896
34. O'Riordan D, Geller A, Brooks D, Zhang Z, Miller D. Sunburn
reduction through parental role modeling and sunscreen vigi-
lance. J Pediatr. 2003; 142:67-72
35. Turrisi R, Hillhouse J, Heavin S, Robinson J, Adams M, Berry
J. Exami.nation of the short- term effi cacy o[ a parent-based
intervention to prevent skin cancer. J Behav Med. 2004;27:
393- 412
36. Buller D, Borland R. Skin cancer prevention for children: a
critical review. Heallh Educ Behav. 1999;26:317- 343
37. Thieden E, Philipsen P, Sandby-Moller J, Wuil H. Sunburn
related to UV radiation exposure. age, sex. occupation. and sun
bed usc based on time-stamped personal dosimetry and sun
behavior diaries. Arch Derma to/. 2005; 141:482- 488
38. Diffey B. Grice J The influence of sunscreen t ype on photopro-
tection. Br J Dermato/. 1997; 137:103-105
39. Wul H. Sender 1. Lock-Andersen J . Sunscreens used at the
beach do not protect against. erythema: a new defini tion of SPF
is proposed. Photodermatol Photoimmunol Photomed. 1997; 13:
129- 132
40. Autier P, Boniol M, Severi G. Dore J, European Organization
for Research and Treatment of Cancer Melanoma Co-operative
Group. Quantity of sunscreen used by European students. Br J
Derma/of. 2001;144:288-291
41. Hill D. Dixon H. Promoting sun protection in children: ratio-
nale and challenges. Health Educ Behav. 1999;26:409- 4 17
42. Hill D. Skin cancer prevention. Am J Prev Med. 2004;27:
482- 483
43. Erru11ons K. Colditz G. Preventing excess sun exposure: it is
time for a nat ional poli cy. J Nat/ Cancer lnst. 1999;9 1:
44. Lower T, Girgis A, Sanson-Fisher R. The prevalence and pre-
dictors of solar protection use among adolescents. Prev Med.
45. Livingston P, White v, Hayman J. Dobbinson S. Sun exposure
and sun protection behaviours among Australian adolescents:
t rends over time. Prev Med. 2003;37:577-584
46. Branstrom R. Brandberg Y, Holm L, Sjoberg L. Ullen H. Beliefs,
knowledge and attitudes as predictors of sunbathing habits and
use of sun protection among Swedish adolescents. Bur J Cancer
Prev. 2001;10:337- 345
47. Lowe J. Borland R, Stanton W, Baade P, White V. Balanda K.
864 COKKINIDES et al
Sun-sa fe behaviour among secondary school students in Aus-
tralia. Health Educ Res. 2000;15:271- 281
48. Wichstrom L. Predictors of Norwegian adolescents' sunbathing
and use or sunscreen. Health f>sychol. I 994; I 3:412- 420
49. US Bureau of the Census. Phone/ess in America. Washington,
DC: US Department of Commerce Economics and Statistics
Administration . .Bureau of the Census; I 994. Statistical brief
94-1 6
50. Hill D. White V. Marks R. Borland R. Changes in sun-related
attitudes and behaviours. and reduced sunburn prevalence in a
population at high risk of melanoma. Eur J Cancer Prev. 1993;
51. Staples M, Marks R. Giles G. Trends in the incidence of non-
melanocytic skin cancer (NMSC) t r ea ted in Au st rali a
1985-1995: arc p rimary prevent ion programs starting to have
an effect? lnt J Cancer. 1998;78:144- 148
52. Jorgensen C. Wayman J, Green C, CA. G. Using health com-
munications for primary prevent ion o f skin cancer: CDC's
Choose Your Cover campaign. J Womens Health Gend Based Med.
2000;9:471- 475
53. American Cancer Society. Skin Cancer. Atlanta, GA: American
Canct:r Society. Available at: www.cancer.org. Accessed Octo-
ber 10, 2005
54. Lim H. Cooper l<. The health impact of solar radiation and
prevention strategies: Report of the Environment CounciL
American Academy of Dermatology. JAm Acad Dennatol. I 999;
41 :8 1-99
55. Skin Cancer Foundation. Sun and Skin News. New York. NY:
Skin Cancer Foundation; 2002:19.
56. Healthy People 2010 (HP2010)- Volume I. Focus Area: Cancer. Ob-
jective .3- 9: Sun Exposure and Skin Cancer. Rockville. MD: Office
of Disease Prevention and Health Promotion. U.S. Department
oi Health and Human Services; 2000. Available at: www.
health ypeople. gov I document/ htmlfvolLu11e 1 I 0 3ca ncer.htm # _
Toc490540737. Accessed November 30. 2005
57. l ~ n z K, Steffen A. Elliott T, O'Riordan D. Diffusion or an
effective skin cancer prevention program: design. theoretical
foundations, and first-year implementation. Health Psycho/.
58. Gritz ER. Tripp MK, de Moor CA. Eicher SA. Spedale JH. Skin
cancer prevention counseling and cl inical practices or pediatri-
cians. Pediatr Derma to/. 2003;20: I 6 -24
Downloaded from www. pediatrics.org by on October 21, 2009
Trends in Sunburns, Sun Protection Practices, and Attitudes Toward Sun
Exposure Protection and Tanning Among US Adolescents, 1998-2004
Vilma Cokkinides, Martin Weinstock, Karen Glanz, Jessica Albano, Elizabeth Ward
and Michael Thun
Updated Information
& Services
Refer ences
Subspecialty Collections
Permissions & Licensing
Pediatrics 2006; 118;853-864
including high-resolution figures, can be found at:
http://www.pediatrics.org/cgi/content/full/ 118/3/853
This article cites 47 articles, 13 of which you can access for free
http://www. pediatrics.org/cgi/content/full/ 118/3/853#BIBL
This article has been cited by 7 High Wire-hosted articles:
This article, along with others on similar topics, appears in the
following collection(s):
Allergy & Dermatology
Information about reproducing this article in parts (figures,
tables) or in its entirety can be found online at:
Information about ordering reprints can be found online:
American Academy of Pediatrics
Downloaded from www.pediatrics.org by on October 21, 2009
A Prospective Study of Pigmentation, Sun Exposure, and
Risk of Cutaneous Malignant Melanoma in Women
Marit Bragelien Veierd, Elisabete Weiderpass, Magnus Thorn, Johan Hansson,
Eiliv Lund, Bruce Armstrong, Hans-Olov Adami
Background: Although sun exposure is au established cause
of cutaneous malignant melanoma, possible interactions with
host factors remain incompletely understood. Here we re-
port the first results from a large prospective cohort study of
pigmentation factors and sun exposure in relation to mela-
noma risk. Methods: The Women's Lifestyle and Health Co-
hort Study included 106 379 women from Norway and Swe-
den who were aged 30-50 years in 1991 or 1992 when they
completed an extensive questionnaire on personal character-
istics and exposures. Linkages to national registries ensured
complete follow-up through December 31, 1999. Poisson re-
gression models were used to estimate relative risks (RRs).
All statistical tests were two-sided. Results: During an aver-
age follow-up of 8.1 years, 187 cases of melanoma were di-
agnosed. Risk of melanoma was statistically significantly as-
sociated with increasing body surface area (RR for ;:.1.79 m
versus m
= 1.60, 95% confidence interval [CI] = 1.03
to 2.48; Ptrend = .02), number of large asymmetric nevi on the
legs (RR for ;:.7 nevi versus 0 nevi= 5.29, 95% CI = 2.33 to
12.01; Ptrend<.001), hair color (RR for red versus dark
brown or black= 4.05, 95% CI = 2.11 to 7.76; Ptrend<.001),
sunburns per year at ages 10-19, 20-29, and 30-39 years
cPtrend<.001, Ptrend = .03, and Ptrend = .05, respectively), and
use of a device that emits artificial light (solarium) one or
more times per month (P = .04). Conclusions: Our results
confirm previous findings that hair color, number of nevi on
the legs, and history of sunburn are risk factors for mela-
noma and suggest that use of a solarium is also associated
with melanoma risk. Adolescence and early adulthood ap-
pear to be among the most sensitive age periods for the
effects of sunburn and solarium use on melanoma risk. How-
ever, it may be too early to see the full effect of adult expo-
sures in this cohort. [J Natl Cancer lnst 2003;95:1530-8]
Cutaneous malignant melanoma (hereafter called melanoma)
imposes a considerable public health burden. The incidence of
melanoma va1ies more than 150-fold around the world, with the
highest rates occurring among white or predominantly white
populations in Australia, New Zealand, North America, and
northern Europe ( 1 ). Rates of melanoma in Norway and Sweden
have more than tripled since 1958- 1962, the first years that
Affiliations of authors: M. B. Section of Medical Statistics, Univer-
sity of Oslo, Oslo, Norway; E. Weiderpass, Department of Medical Epidemiol-
ogy and Biostatistics, Karolinska lnstitutet, Stockholm, Sweden, and Interna-
tional Agency for Research on Cancer, Lyon, France; M. Th<im. Department of
Surgery, South Stockholm General Hospital , Stockholm; J. Hansson, Depart-
ment of Oncology and Pathology, Karolinska Institute!; E. Lund, Institute of
Conununity Medicine, University of Troms0, Norway; B. Armstrong,
School of Publ ic Health, University of Sydney, Sydney, Australia; H.-0. Adami,
Department of Medical Epidemiology and Biostatistics, Karol inska Institute!.
and Department of Epidemiology, Harvard University, Boston, MA.
Correspondence ro: Marit B. Veier!ild, PhD, Section of Medical Statistics,
University of Oslo, P.O. Box 1122 Blindern, N-0317 Oslo, Norway (e-mail:
See "Notes" following "References."
DOl : 10.1 093/jnci/djg075
.Journal of the National Cancer l nsrilute, Vol. 95, No. 20, Oxford University
Press 2003, all rights reserved.
1530 ARTICLES Joumal of the National Cancer Institute, Vol. 95, No. 20, October 15, 2003

"' c.
reliable information was available from cancer registries; rates
are now higher there than they are elsewhere in Europe (2) and
are predicted to increase (3 ).
Although sun exposure is the major established risk factor for
melanoma (4,5), geographic differences in melanoma incidence
cannot be attributed solely to differences in the intensity of solar
exposure. Within Europe, for example, the incidence of mela-
noma is higher at northern latitudes, which generally have lower
solar intensities, than at southern latitudes, which generally have
higher solar intensities (2), although in both Norway and Swe-
den, an inverse relationship between melanoma incidence and
latitude has been noted (6,7). Hence, the etlect of UV light on
melanoma risk may be strongly modified by other factors, such
as differences in sun sensitivity and the nature of the exposure to
the sun (8).
A number of studies have examined factors that influence the
association between sun exposure and the risk of melanoma. An
intermittent pattern of sun exposure, which is typically assessed
by measures of sun-intensive activities, such as outdoor recre-
ation or vacations, is associated with increased risk of melanoma
(9). In addition, many studies (4,5,9,10) have reported that sun-
burn, which is an indicator of an intermittent pattern of sun
exposure, is positively associated with the risk of melanoma.
Results of many studies have suggested that childhood is a criti-
cal period for sun exposure (9), and ecologic studies have shown
more consistent associations than case-control studies between
childhood sun exposure and melanoma risk ( 11 ). Host factors
such as eye color, hair color, skin color, the number of nevi, and
skin reaction to chronic and acute sun exposure have also been
associated with the risk of melanoma (4, 12).
Most of what is known about the association between sun
exposure and melanoma risk comes from results of case-control
studies. The Nurses' Health Study is, as far as we know, the only
cohort study to examine the association between sun exposure
and malignant melanoma; however, a case- control design within
the cohort was used in these analyses ( 13,14 ). Case-control
studies are limited by the potential for differential bias in recall
of sun exposure between case patients and control subjects
(15, 16). Prospective cohort studies can overcome such limita-
tions because the exposure information is collected prior to dis-
ease occurrence. Here we report the first results from the Nor-
wegian-Swedish Women's Lifestyle and Health Cohort Study,
which was initiated in 1991. This study is the first prospective
cohort study, to our knowledge, to examine the associations
between pigmentation factors and sun exposure and the risk of
malignant melanoma.
Study Population
For practical reasons, women were enrolled in the Norwegi-
an- Swedish Women's Lifestyle and Health Cohort Study in
both 1991 and 1992.ln Norway, a nationwide random sample of
100000 women who were born between 1943 and 1957 (i.e.,
aged 34-49 years at inclusion) was drawn from the National
Population Register at Statistics Norway (Oslo, Norway). In
Sweden, a random sample of 96 000 women who were born
between 1943 and 1962 (i.e., aged 30-50 years at inclusion) and
were residing in the Uppsala Health Care Region (which com-
prises about one-sixth of the Swedish population) was drawn
from the National Population Register at Statistics Sweden
(Stockholm, Sweden).
All women received a letter inviting them to participate in the
study. The letter also requested that they provide written in-
formed consent and contained a comprehensive questionnaire
that was to be completed and returned in a prepaid envelope.
Identical questions relevant to the analysis presented here were
included in the questionnaires sent to women in the two coun-
tries. The study was approved by the Data Inspection Boards in
both countries and by the regional Ethical Committees, and all
women gave written informed consent to participate.
Host Factors and Exposure Information
In the questionnaires, study participants were asked to cat-
egorize their natural hair color (dark brown/black, brown, blond,
or red) and their eye color (brown, gray/green, or blue) and to
categorize the number of asymmetric nevi larger than 5 mm on
their legs from toes to groin (0, 1, 2- 3,4- 6,7- 12, 13- 24, or ;;.25
nevi). A brochure that was included with the questionnaire pro-
vided color pictures with three examples of asymmetric nevi.
Participants recorded their sun sensitivity according to their
reactions to both acute and chronic exposure to the sun. Regard-
ing acute sun exposure, the questionnaire asked the women to
choose from among four categories to describe how their skin
reacts to heavy sun exposure at the beginning of the summer: the
skin turns brown without first becoming red, the skin turns red,
the skin turns red with pain, or the skin turns red with pain and
blisters. The women were asked to describe how their skin reacts
to long-lasting or chronic sun exposure according to four cat-
egories: the skin turns deep brown, brown, or light brown, or the
skin never turns brown.
Participants were asked to report their histories of sunburn
and sunbathing vacations and on the frequency of their use of a
solarium (i.e., a sun bed or a sunlamp that emits artificial UV
light) when they were aged 10-19, 20-29, 30-39, or 40-49
years. For each age period, the participant was asked to report
the number of times per year she had been burned by the sun so
severely that it resulted in pain or blisters that subsequently
peeled by choosing from among five categories: never, one time
per year at most, two or three times per year, four or five times
per year, or six or more times per year. Participants reported the
average number of weeks per year spent on sunbathing vacations
in southern latitudes (typically southern Europe, e.g., Spain or
Greece) or within Norway or Sweden for each age period by
choosing from among five categories: never, I week per year,
2-3 weeks per y e a r ~ weeks per year, or ;;.7 weeks per year.
Participants reported their average use of a solarium during each
age period by choosing from among six categories: never, rarely,
one time per month, two times per month, three or four times per
month, or more than one time per week. The questionnaires also
contained questions about the participant' s current height and
weight, current and past contraceptive use, reproductive history,
prevalent diseases, and lifestyle.
Follow-up and Endpoints
Start of follow-up was defined as the date of receipt of the
returned questionnaire. Person-years were calculated from the
start of follow-up to the date of diagnosis of primary melanoma,
to the date of emigration or death, or to the end of follow-up
(December 31, 1999), whichever occurred first. Each resident of
Norway and Sweden is assigned a unique national registration
number that includes the person's date of birth; those registra-
tion numbers are entered into the nationwide databases that were
Journal of the National Cancer Institute, Vol. 95, No. 20, October 15, 2003 ARTICLES 1531
"' c.
used in this study. By linkage of cohort data to the national
cancer registries in Norway and Sweden, this national registra-
tion number allowed us to identify cancer cases. Information on
death and emigration was gathered by linkage to Statistics Nor-
way and Statistics Sweden.
A total of 57 584 (57.6%) of the Norwegian women and
49 259 (51.3%) of the Swedish women returned completed ques-
tionnaires; the overall response rate was 54.5%. We excluded
four women because of the lack of vital status information in the
available register files, 18 women who had emigrated or died
before the start of follow-up, 198 women who did not adequately
answer the questions regarding sun exposure or personal char-
acteristics (i.e., sun sensitivity of skin, hair color, eye color, and
number of asymmetric nevi), and 244 women who were diag-
nosed with melanoma prior to the start of follow-up.
Statistical Analysis
Participants' geographic regions of residence were defined
according to four categories: the southern region of Norway, the
middle region of Norway, the northern region of Norway, and
the Uppsala Health Care Region in Sweden. The latitudes of the
population center of mass within each Norwegian county, which
were provided by the Norwegian Mapping Authority, together
with the observed number of melanoma cases in those counties,
formed the basis for our definitions of the three Norwegian
regions. The southem region of Norway includes Vest-Agder,
Aust-Agder, Rogaland, Vestfold, 0stfold, and Telemark coun-
ties, with population centers of mass located at 5824' - 5931'
N; the middle region of Norway includes Oslo, Akershus, Bus-
kerud, Hordaland, Oppland, Hedmark, and Sogn og Fjordane
counties, with population centers of mass located at 5958' -
61 30' N; the northern region of Norway includes Ms:6re og
Romsdal, Ss:6r-Trndelag, Nord-Trs:6ndelag, Nordland, Troms,
and Finnmark counties, with population centers of mass located
at 6244'- 7022' N. The Uppsala Health Care Region in Swe-
den has the population center of mass located at 5986' N.
Body surface area was calculated according to the formula (17)

x height
0 725
x 71.84 and categorized by guartiles.
We combined the upper two categories of the variables concern-
ing acute and chronic exposures to sun because of the small
numbers in each category and analyzed nevus counts in three
categories: 0, 1, 2-6, and ~ 7 (only two categories, 0 and ~ 1
were used when testing interaction effects). In the age period-
specific analyses of sunburns, sunbathing vacations, and so-
larium use, we combined the upper categories of these variables
because of small numbers. For each of the variables (sunburns,
sunbathing vacations, and solarium use), new variables were
constructed to combine the exposure during the three age peri-
ods that were recorded for all women (i .e., 10-19,20-29, 30-39
We used Poisson regression analysis to estimate the associa-
tion between sun exposure or personal characteristics and the
risk of melanoma. The statistical significance of independent
variables and interaction effects was tested by using the likeli-
hood ratio test. We tested for trends across categories of vari-
ables by assigning equally spaced values (e.g., I , 2, 3, or 4) to
the categories and treating the variables as continuous variables
in the Poisson regression analysis. All analyses were adjusted for
attained age (i.e., age at study entry plus the duration of follow-
up), which was categorized by 5-year intervals (for analyses of
women aged 40 years or older, we used only two age categories,
<50 years and 50-60 years), and all multivariable models also
included geographic region of residence. The analyses of per-
sonal characteristics included mutual adjustments for statisti-
cally significant variables. The multivariable models used in the
analyses of sunburn, sunbathing vacations, and use of a solarium
included hair color. In addition, each age-specific model for use
of a solarium included the corresponding numbers of age-
specific sunburns and sunbathing vacations. Results are pre-
sented as relative risks (RRs) with 95% confidence intervals
(Cis). All P values are two-sided, and a 5% level of statistical
significance was used.
The final study sample consisted of 106 379 Norwegian and
Swedish women. Dming an average 8.1 years of follow-up (me-
dian = 8.3 years, range = 0.01- 8.6 years) corresponding to
866 668 person-years of observation, 187 incident cases of mela-
noma were reported to the Cancer Registries in Norway and
Sweden. These incident cases occurred among 183 women for
whom melanoma was their first cancer diagnosis and four
women for whom melanoma was their second cancer diagnosis.
All incident cancer cases were histopathologically confirmed as
invasive melanoma. Characteristics of the study cohort and of
the incident cases of malignant melanoma and their frequencies
are summarized in Table 1. Melanomas on the lower limbs were
observed most frequently, followed by melanomas on the trunk.
Classification of subtypes was less frequently performed in Swe-
den than in Norway. Seventy-one percent of the Norwegian cases
were classified as superficial spreading melanoma (Table 1).
Table 2 summarizes the associations between personal char-
acteristics and the risk of melanoma. Calculated body surface
area was positively associated with the risk of melanoma (Ptrend
= .02), as was hair color (P trend<.OO I). Compared with women
who had dark brown or black hair, women with blond hair had
an approximately twofold higher risk of melanoma, whereas
women with red hair had an approximately foUifold higher tisk.
Eye color was not associated with melanoma risk. We also found
no statistically significant association between tanning of the
skin after heavy or repeated sun exposure and the risk of mela-
noma, although an indication of a trend was seen for skin color
after repeated sun exposure. The number of large asymmetric
nevi on the legs was a strong predictor of melanoma risk: women
with seven or more nevi had an approximately fivefold higher
risk of melanoma than women with no nevi (Ptrend<.OO I ). Mu-
tual adjustment for all statistically signi ficant variables li sted in
Table 2 did not appreciably change any of the multivariable
relative risks presented in the table (data not shown).
Risks of melanoma increased with increasing numbers of
sunburns women reported having during the second, third, and
fourth decade of life (Table 3). The estimated risk of melanoma
was highest for women who reported having sunburns during
adolescence (i.e., the 10-19-year age period), whereas no asso-
ciation between risk and sunburns during the fifth decade of life
(i.e., the 40-49-year age period) was observed. Next, we com-
bined the information about the number of sunburns at ages
10-19, 20-29, and 30-39 years into one new variable. Women
who had one or no sunburns per year during these three periods
were used as the reference category. The other categories were
sunburns two or more times per year during the adult years (i.e.,
20-29 years and/or 30-39 years), sunburns two or more times
1532 ARTICLES Joumal of the National Cancer Institute, Vol. 95, No. 20, October 15, 2003
"' c.
Table 1. Characteristics of participants in the Norwegian-Swedish Women's Lifestyle and Health Cohort Study and of the incident cases
of cutaneous malignant melanoma during follow-up from 1991- 1992 through 1999
Mean age at study entry, y (range)
Person-years of follow-up
Number of incident cases of melanoma
Mean age at diagnosis of melanoma, y (range)
Site of melanoma, No. (%)
Upper limb
Lower limb
Histologic type of melanoma, No. (%)
Superficial spreading melanoma
Nodular melanoma
Lentigo malignant melanoma
Malignant melanoma, not othenvise specified
Body surface area in m
t No. (%) (n = 103 333)
~ 1 6 1
1.62- 1.69
:;. ).79
(n = 57 3 ll )*
41. I (34--49)
45.7 (35.4-54.0)
32 (26)
12 (10)
60 (50)
17 (14)
86 (71)
16 (13)
2 (2)
13 985 (25)
13 929 (25)
14361 (26)
13 896 (25)
(n = 49 068)*
39.6 (30-50)
397 686
45.3 (31.9- 57.5)
19 (29)
10 ( IS)
29 (44)
8 (12)
5 (8)
2 (3)
0 (0)
59 (89)
I I 696 (25)
I I 680 (25)
II 802 (25)
I I 984 (25)
(N = 106 379)*
40.4 (30-50)
45.6 (31.9- 57.5)
51 (27)
22 ( 12)
89 (48)
25 (13)
91 (49)
18 ( 10)
2 (I)
76 (41)
25 681 (25)
25 609 (25)
26 163 (25)
25 880 (25)
Skin color after heavy sun exposure in the beginning o f the summer, No. (%) (n = 105 595)
Brown 14 856 (26)
27 584 (49)
11 342 (20)
I 1532 (24)
23 243 (48)
11421 (23)
26 388 (25)
50 827 (48)
22763 (22)
Red with pain
Red with pain and blisters
Skin color after repeated sun exposure, No. (%) (n = 103 31 2)
Deep brown
Light brown
Never brown
Hair color, No. (%) (n = 103 027)
Dark brown, black
Eye color, No. (%) (n = J 02 710)
Gray, green, or mi x
Total No. of asymmetric nevi >5 mm on legs, No.(%) (n = 100980)
4- 6
2999 (5)
8797 ( 16)
31394 (58)
13 453 (25)
900 (2)
9348 (17)
21 500 (39)
22241 (41)
1495 (3)
6345 ( 12)
2 1 062 (39)
27 170 (50)
47 704 (89)
3438 (6)
1595 (3)
416 (I)
324 (I)
261 8 (5)
7979 (16)
30029 (62)
10129 (2 1)
631 ( I)
20939 (43)
12 185 (25)
1506 (3)
6738 (14)
17 130 (36)
24 265 (50)
38 997 (82)
4842 (10)
2424 (5)
713 (2)
527 (l)
16776 ( 16)
61423 (59)
23 582 (23)
1531 ( I )
23 161(23)
42439 (41)
34426 (33)
3001 (3)
13 083 (13)
38 I 92 (37)
5 I 435 (50)
86 701 (86)
8280 (8)
4019 (4)
I 129 (I)
851 (I)
*Because of missing values. the number of women wil l d iffer in the presentation of personal characteristics below. The total number of women (n) is presented
for each personal characteristic.
t Head/neck and skin unspecitied.
*Calculated according to the followi ng formula {17): weight
x height

x 71.84.
per year dming adolescence (i.e., l0- 19 years), and sunburns
two or more times per year during all three age decades (i.e.,
10-19, 20-29, and 30-39 years). We observed increased risk of
melanoma for the upper two categories of this new variable and
a statistically significant positive trend (Table 3). Collapsing the
upper three categories into one gave a multi variable relative risk
of 1.70 (95% CI = 1.23 to 2.34; P = .002) for sunburns two or
more times per year for at least one of the three age decades as
compared with a maximum of one sunburn per year in all three
decades. No statistically significant interaction was found be-
tween this dichotomous sunbum variable and the number of nevi
on the legs (P = .84).
We found suggestive evidence for an association between
increasing risk of melanoma and increasing number of weeks
women spent on sunbathing vacations at ages 30- 39 years
(Table 4). Although most of the point estimates and all of the
trends pertaining to this association were not statistically signifi-
cant, we consistently observed a risk increase of approximately
60%- 70% for the hi ghest compared with the lowest exposure
category for women who took sunbathing vacations between the
ages of I 0 and 39 years. Increased risk, albeit not statisticall y
significant and with no appreciable trend, was also observed
when inforn1ation on sunbathing vacations from these three de-
cades of life was combined into a new variable in a way analo-
Journal of the National Cancer Institute, VoL 95, No. 20, October 15, 2003 ARTICLES 1533
"' c.
Table 2. Relative risks (RRs) and 95% confidence intervals (Cis) of cutaneous malignant melanoma according to personal characteristics*
Body surface area, m
:j: (n = I 03 333)

1.62- 1.69
;;> 1.79
Skin color after heavy sun exposure at the beginning of sunuuer (n = 1 OS 595)
Red with pain/red with pain and blisters
Skin color after repeated sun exposure (11 = 103 312)
Deep brown
Light brown/never brown
Hair color (n = 103 027)
Dark brown, black
Eye color (n = I 02 710)
Gray, green, or mix
Total No. of asymmetric nevi >5 mm on legs (n = 100 980)
*Poisson regression analysis. All statistical tests were two-sided.
t Multivariable models included attained age and region of residence.
No. of cases
Age-adjusted RR (95% Cl) Multivariable RRt (95% Cl)
1.00 (referent) 1.00 (referent)
1.38 (0.87 to 2. 17) 1.37 (0.87 to 2.17)
1.74 (1.13 to 2.68) 1.73 ( 1.12 to 2.66)
1.60 (1.03 to 2.49) 1.60 ( 1.03 to 2.48)
Puend = .02 P.,,.od = .02
1.00 (referent) 1.00 (referent)
1.45 (0.99 to 2.12) 1.45 (0.99 to 2.13)
1.34 (0.88 to 2.06) 1.36 (0.89 to 2.08)
plr<Od = .21 P.,,.od = .19
1.00 (referent) 1.00 (referent)
1.39 (0.87 to 2.22) 1.40 (0.87 to 2.23)
1.62 (0.97 to 2.69) 1.60 (0.96 to 2.67)
ptrcnd = .07 Pcrcnd = .07
1.00 (referent) 1.00 (referent)
1.18 (0.74 to 1.88) 1.16 (0.73 to 1.84)
2.10 (1.35 to 3.26) 1.96 ( 1.25 to 3.07)
4.13 (2.16to7.9l) 4.05 (2.11 to 7.76)
P.,..,, d<.OOl ptrend<.OOI
1.00 (referent) 1.00 (referent)
Ll8 (0.70 to 1.99) 115 (0.68 to 1.94)
1.36 (0.82 to 2.25) 1.33 (0.80 to 2.20)
P,,..," = .17 P,.en<l = .19
1.00 (referent) 1.00 (referent)
2.15 ( 1.41 to 3.28) 2.29 ( 1.50 to 3.49)
2.14 (1.27 to 3.60) 2.30 ( 1.36 to 3.87)
4.92 (2.17 to 1l.l5) 5.29 (2.33 to 12.01)

:j:Calcu1ated acc-ording to the following formula ( 17): weight
x height

x 7 1.84.
gous to that described above for sunburns (Table 4). Collapsing
the upper three categories of this new vatiable gave a multi vari -
able relative risk of 1.51 (95% CI = 0.95 to 2.40; P = .07) for
sunbathing vacations one or more weeks per year in at least one
of the three age decades as compared with never going on sun-
bathing vacations in any of the three decades. No statistically
significant interaction was found between this dichotomous vari-
able for sunbathing vacations and the number of nevi on the legs
(P = .58).
We had limited power to examine the association between the
use of a solarium during adolescence and melanoma risk be-
cause only 2% of the women in the study reported having such
exposure. However, we found that compared with women who
never used a solarium at ages 20-29 years, women who reported
using a solarium once or more per month during that age period
had a relative risk of melanoma of 2.58 (95% CI = 1.48 to 4.50;
Ptrend = .006) (Table 5). Use of a solarium at ages 30- 39 years
and 40-49 years also appeared to be associated with a risk,
although not a statistically significantly increased risk, of mela-
noma (Table 5). In a multivariable analysis of the combined
variable for solarium use duri ng the 10- 39-year age period,
women who used a solatium one or more times per month jn at
least one of the three decades between ages 10 and 39 had a
statistically significantly higher risk of melanoma than women
who had never or rarely used a solarium during those three
decades (RR = 1.55, 95% CI = 1.04 to 2.32; P = .04) (Table 5).
The multivariable models in Tables 3-5 include hair color as
a measure of sun sensitivity. Additional adjustment for skin
color after repeated sun exposure gave similar results and did not
affect the conclusions (data not shown).
Results of our prospective analysis suggest that hair color, the
number of large asymmetric nevi on the legs, and body surface
area are important personal characteristics that contribute to the
risk of melanoma. The number of sunburns was also an impor-
tant predictor of melanoma risk, and the strongest effects were
associated with the number of sunburns women experienced
during adolescence; there was similar, albeit weaker, evidence
for an associat.ion between the number of sunbathing vacations
taken in Norway, Sweden, or more southern latitudes and mela-
noma risk. Using a solarium one or more times per month,
particularly during the 20-29-year age period, adjusted for num-
bers of sunburns and sunbathing vacations, was statistically sig-
nificantly associated with melanoma risk.
The incidence of melanoma observed in our study was higher
among the Norwegian women than among the Swedish women.
The crude incidence rates of melanoma, which we calculated
from the data presented in Table 1, were 25.8 cases per 100000
person-years of follow-up for the Norwegian women and 16.6
cases per 100000 person-years of follow-up for the Swedish
1534 ARTICLES Joumal of the National Cancer Institute, VoL 95, No. 20, October 15, 2003

"' c.
Table 3. Relative risks (RRs) and 95% confidence intervals (Cis) of cutaneous malignant melanoma according to annual number of
sunburns during different age periods*
Age period and number of sunburns No.(%) No. of cases Age-adjusted RR (95% Cl) Multivariable RRt (95% Cl)
I 0-19 years (n = 95 472)
0 2 1 747 (23) 22 1.00 (referent) 1.00 (referent)
52452 (55) 94 1.80 (1. 13 to2.86) 1.64 ( 1.03 to 2.62)
21273 (22) 55 2.70 ( 1.65 to 4.44) 2.42 ( 1.46 to 4.02)
P,rend<.001 P.,
,w<.00 I
20-29 years (n = 97 442)
0 20346 (21) 28 1.00 (referent) 1.00 (referent)
58 458 (60) 102 1.29 (0.85 to 1.96) 1.24 (0.81 to 1. 88)
18638 (19) 43 1.76 (1.09 to 2.84) 1.69 (1.04 to 2.76)
p<rcnd = .02 p<rcnd = .03
30- 39 years (n = 94 850)
0 30 588 (32) 48 1.00 (referent) 1.00 (referent)
54 199 (57) 99 1.15 (0.82 to 1.63) 1.15 (0.81 to 1.62)
10 063 (11) 27 1.71 (1.07 to 2.74) 1.71 (J.06to 2.76)
p<rend = .04 P,rend = .05
40-49 years:t: (n = 45 269)
0 20031 (44) 43 1.00 (referent) 1.00 (referent)
22 260 (49) 44 0.92 (0.61 to 1.41) 0.92 (0.61 to 1.41)
2978 (7) 6 0.94 (0.40 to 2.21) 0 96 (0.41 10 2.27)
p<rcnd = .74 P,rend = .77
Combined, 10-39 years (n = 90633)
1/ycar, I 0- 39 years 64807 (72) 99 1.00 (referent) 1.00 (referent)
20- 29 years and/or 30-39 years 5873 (6) 13 1.4 7 (0.82 to 2.62) 1.54 (0.86 to 2.75)
10-19 years 7357 (8) 20 1.82 (1.13 to 2.95) 1.66 ( 1.02 to 2.70)
10- 39 years 12 595 (14) 34 1.83 (1.24 to 2.70) 1.79 ( 1.20 to 2.68)
Ptrerld<.OOl P, rend = 002
*Poisson regression analysis. All statistical tests were two-sided.
tMultivariable models included attained age, region of residence, and hair color.
*Included only women who were aged 40 years or older when answering the questionnaire.
women. These incidence rates are in accordance with crude in-
cidence rates reported for Norwegian and Swedish women for
1993 through 1997 (23.3 cases per 100000 person-years for
Norwegian women and 17.3 cases per 100000 person-years for
Swedish women) (3). Age-adjusted incidence rates of melanoma
have been consistently hi gher among Norwegian than among
Swedish women since the 1960s.
We observed a strong association between hair color and
melanoma risk but not between eye color and melanoma risk.
These results are consistent with results of a pooled analysis of
data derived from published case-control studies, in which the
reported relative risks were 2.38 (95% CI = 1.90 to 2.97) for
individuals who have red hair compared with those who have
black or dark brown hair and 1.55 (95% CI = 1.35 to 1.78) for
individuals who have blue eyes compared with those who have
brown eyes (12). However, the association we observed between
cutaneous sensiti vi ty to the sun (i. e., burning or tanning) and
melanoma risk was much weaker than that reported in a retro-
spective Australian study (I 8). Our findings, that hair color but
not eye color was statistically significantly associated with mela-
noma risk, agree with those of two Danish case-control studies
(19,20): in addition, the association between melanoma risk and
cutaneous sun sensitivity reported in those two studies was also
much weaker than that for hair color. A Swedish case-control
study (21) also found that hair and eye color and skin type were
statistically signi ficantly associated with melanoma risk, al-
though the associations were considerably weaker for eye color
and skin type than for hair color, whereas an early Norwegian
case-control study (22) that used hospital-based control subjects
found that tolerance to sun exposure, but not hair or eye color,
was associated with melanoma risk. We speculate that hair color
may be the best measure (combining accuracy of measurement
and predictive capacity) of sun sensitivity in homogeneous fair-
skinned populations, such as those of Scandinavia. By contrast,
reported sun sensitivity may be a less reliable measure of sun
sensitivity in these populations because it depends on an indi-
vidual's experience with repeated and quite heavy sun exposure,
which many Scandinavian subjects may not have.
In agreement with the results of several case-control studies
(23- 25), the results of our cohort study show that the number of
asymmetric nevi larger than 5 mm on the legs was the strongest
host risk factor for melanoma. The participants self-reported
such nevi on their legs, guided by color pictures of dysplastic
nevi in a brochure that was encl osed with the questionnaire. The
method we used for this self-reporting has been shown to have
limited accuracy for the diagnosis of one or more dysplastic
nevi, with an estimated sensitivity of 29% and a specificity of
85% (26). Hence, the relative risk of 5.3 for melanoma in the
presence of seven or more large nevi on the legs that we ob-
served in our study may underestimate the excess risk. Increased
surveillance and more frequent excision of suspected lesions
might, on the other hand, spuriously inflate the risk of melanoma
among subjects with asymmetric nevi. However, such an effect
seems unlikely because all incident cases were histopathologi-
cally confirmed invasive malignant melanomas.
Our results confirm the positi ve association between past his-
tory of sunburn and melanoma reported previously by the ma-
jority of case-control studies (9,10). Our effect estimates were
higher for sunburns that occurred during adolescence than for
those that occurred later in life; however, it may be too early to
Journal of the National Cancer Institute, Vol. 95, No. 20, October 15, 2003 ARTICLES 1535

"' c.
Table 4. Relative risks (RRs) and 95% confidence intervals (Cis) of cutaneous malignant melanoma according to the average number of
weeks per year spent on sunbathing vacations to southern latitudes or within Norway or Sweden during different age periods*
Age period and annual weeks on sunbathing vacation No.(%) No. of cases Age-adjusted RR (95% Cl) Multivariable RRt (95% C1)
10-19 years (n = 93418)
0 45 298 (48) 77 1.00 (referent) 1.00 (referent)
I week/year 19 921 (21) 35 1.10 (0.74 to 1.65) 1.21 (0.80 to 1.83)
2- 3 weeks/year 20086 (22) 32 1.02 (0.67 to 1.54) 1.09 (0.71 to 1. 65)
;;;.4 weeks/year 8113 (9) 20 1.56 (0.95 to 2.55) 1.67 (1.01 to 2.74)
ptrend = .22 P,,.,," = .12
20-29 years (n = 96 029)
0 26460 (28) 41 1.00 (referent) I .00 (referent)
I week/year 28 723 (30) 55 1.28 (0.85 to 1.92) 1.36 (0.90 to 2.05)
2-3 weeks/year 32 997 (34) 53 I .08 (0.7 1 to I .62) I . 13 (0.74 to I .70)
;;;.4 weeks/year 7849 (8) 19 1.67 (0.96 to 2.88) 1.79 ( Ul3 to 3. I 1)
P,,..,J = .26 P., .. ," = .18
30-39 years (n = 93 845)
0 24 293 (26) 33 1.00 (referent) 1.00 (referent)
I week/year 28 858 (3 I) 56 1 .42 (0.93 to 2. 19) 1 .49 (0.97 to 2.30)
2- 3 weeks/year 33 144 (35) 65 I .43 (0.94 to 2.18) 1.45 (0.95 to 2.21)
;;;.4 weeks/year 7550 (8) 16 1.56 (0.86 to 2.84) 1.63 (0.89 to 2.97)
pttend = .10 P,,.,," = .08
40-49 years:t: (n = 45 21 1)
0 13806(31) 23 I .00 (referent) 1.00 (referent)
l week/year 12801 (28) 36 1.70 (1.01 to 2.87) 1.87 ( 1.11 to 3.18)
;;;.2-3 weeks/year 18 604 (41) 31 1.01 (0.59 to 1.73) 1.06 (0.61 to 1.81)
ptrend = .87 pto-end = .98
Combined, 10-39 years (n = 88 450)
0, 10-39 years 15799(18) 2 1 1.00 (referent) 1.00 (referent)
;;;. I week/year, 20-29 and/or 30- 39 years 27851 (31) 53 1.42 (0.86 to 2.36) 1.45 (0.87 to 2.40)
;;;. I week/year, I 0-19 years 1751 (2) 3 1.37 (0.4 I tO 4.59) 1.46 (0.43 to 4.92)
;;;. I week/year, I 0-39 years 43 049 (49) 79 1.44 (0.89 to 2.34) 1. 56 (0.95 to 2.56)
P,,..," = .27 P trend = .13
*Poisson regression analysis. All statistical tests were two-sided.
t Multivariable models included attained age, region of residence, and hair color.
*Included only women who were aged 40 years or older when answering the questionnaire.
Table 5. Relative risks (RRs) and 95% confidence intervals (Cls) of cutaneous malignant melanoma according to solarium usc during different age periods*
Age period and solarium use No.(%) No. of cases Age-adjusted RR (95% CJ) Multivariable RRt (95% CJ)
10-19years(n = 85 847)
Never 84 182 (98) 152 1.00 (referent) 1.00 (referent)
Rarely or ;;;. I time/month 1665 (2) 4 1. 65 (0.61 to 4.47) 1. 52 (0.56 to 4. 1 2)
p = .36 p = .44
20-29 years (n = 89 142)
Never 71 133 (80) 123 1.00 (referent) 1.00 (referent)
Rarely 11618 (13) 19 1.16 (0.70 to 1.92) Ill (0.67 to 1.85)
;;;. I time/month 6391 (7) 18 2.32 ( I .35 to 3.99) 2.58 ( 1 .48 to 4 50)
P,,..,"" = .009 P,,...
= .006
30-39 years (n = 87 890)
Never 44338 (50) 78 1.00 (referent) 1.00 (referent)
Rarely 28 383 (32) 51 I .03 (0.72 to 1.48) 0 93 (0.64 to 1 .34)
;;;. I time/month 15169(17) 36 1.40 (0.93 to 2.10) 1.42 (0.93 to 2.16)
P<rend = .15 Ptr<nJ = .19
40-49 yearst (n = 41 409)
Never 17 345 (42) 27 1.00 (referent) 100 (referent)
Rarely 145 14 (35) 33 1.46 (0.88 to 2.43) 1 .39 (0.82 to 2.33)
;;;. I time/month 9550 (23) 22 1.48 (0.84 to 2.60) 1.67 (0.93 to 2.99)
P., . .,, = .14 pmmd = .08
Combined, 10-39 years (n = 79 616)
Never/rarely. 10-39 years 65239 (82) 111 1.00 (referent) 100 (referent)
;;;. I time/month I 0-19, 20-29, or 30-39 years 14 377 ( 18) 34 1.45 (0.98 to 2.14) 1.55 ( 1.04 to 2.32)
p = .07 p = .04
*Poisson regression analysis. All statistical tests were two-sided.
t Multivariable models included attained age, region of residence, hair color, and the corresponding number of age-specific sunburns and weeks on annual summer
only women who were aged 40 years or older when answering the questionnaire.
1536 ARTICLES Joumal of the National Cancer Institute, Vol. 95, No. 20, October 15, 2003

"' c.
see the full effect of sunburns later in tife in our cohort of
women. Systematic reviews of case-control studies ( 10,11) have
not found evidence of an overall stronger effect of sunburns in
early life than in later life. Furthe1more, the rep011ed dose-
response gradients of melanoma risk with frequency of sunburn
were comparable during childhood and adulthood in a recent
large multicenter case-control study from Europe (27). How-
ever, it is possible that the case- control studies have underesti-
mated the effects of sunburn during childhood and adolescence
because of high recall error from the very long recall period for
most subjects. All of our study subjects were younger than 50
years when they answered the questionnaire, giving a shorter
recall period than in many case-control studies (9) that include
subjects up to 70 years old or older.
Sun exposure during sunbathing vacations is usually intense
and intermittent, and results of previous case-control studies
( 10) suggest that there is a positive association between the
incidence of melanoma and hi gh levels of intermittent sun ex-
posure. We recorded the number of sunbathing vacations in
Norway and Sweden (at latitudes higher than 58 N, where UV
levels are low, even in summer) and those in southern latitudes
in the same variable, which may explain the lack of a strong
association between sunbathing vacations and melanoma in our
study. Previous Scandinavian studies show inconsistencies in
their results on sunbathing and melanoma risk. One Swedish
study (21) and a Danish study (28) found associations between
vacations spent in sunny places and melanoma risk, whereas
another Swedish study (29) did not.
Our results provide stronger evidence than those of other
studies that solarium use is associated with an increased risk of
melanoma; we found that overall, regular (i.e., one or more times
per month) solarium use at any age was associated with a sta-
tistically significant 55% increase in risk of melanoma after
adjustment for sun sensi tivity and measures of sun exposure.
Although other studies (30-34) have reported positive associa-
tions between melanoma risk and exposure to artificial UV light,
these associations often apply to specific subgroups of the study
population (e.g. the youngest subjects with melanoma), or they
have not been adjusted for possible confounding with sun ex-
posure. A recent review (35) concluded that there was insuffi-
cient evidence to determine whether or not tanning lamps cause
melanoma. The more consistent and overall statistically signifi-
cant association between melanoma risk and solarium use ob-
served in our study, which may be due to the relative youth of
our cohort, adds substantially to the existing evidence that arti-
ficial UV light for recreational tanning increases fisk of melanoma.
Our study has several important strengths. First, because aU
physicians, hospital departments, and histopathologic laborato-
ries in Norway and Sweden are obliged to report malignant
diseases to the cancer registries, and the cancer registries match
regularly against the death registers at Statistics Norway and
Statistics Sweden, respectively, we had a complete follow-up
and histopathologic confirmation of all incident cases of mela-
noma. Second, our study had a prospective design, such that
detailed infonnation on host factors and sun exposure was col-
lected prior to melanoma diagnosis. Error in measurement of
these factors is inevitable in epidemiologi c studies of skin cancer
(26,36,37) but can be assumed to be non-differential in the pres-
ent study. By contrast, measurement error in case-control stud-
ies may be intluenced by a diagnosis of skin cancer and therefore
may differ in degree between cases and controls (15,16).
Among the limitations of our study were the comparatively
small number of cases, the limited detail about the exposure
measurements, and the relatively short follow-up petiod for solar
and artificial UV light exposure during midlife. In addition, we
did not adjust for the multiple comparisons made in this study.
Instead, we chose to eval uate the individual associations on their
own merits and with respect to results from prior studies. Fi-
nally, because our cohort included only women, our results may
not be generalizable to both sexes. In Norway and Sweden,
incidence rates of melanoma tend to be slightly higher among
women than among men (3). However, previous case- control
studies (9) have not focused on whether there are sex differences
in the associations between pigmentation characteristics or sun
exposure and the risk of melanoma.
The results of our cohort study suggest that public health
recommendations for melanoma prevention should include a
combination of information on inherent predisposition and the
effects of exposure to UV radiation. Hair color and large asym-
metric nevi on the legs were the most important host factors
associated with ri sk, and our results for sunbum, sunbathing
vacations, and use of a solarium support current recommenda-
tions for the avoidance of UV exposure, especially intermittent
exposure, either from natural or from artificial sources. Al-
though our smdy cohort is still too young to fully assess whether
UV exposure during adolescence is more critical than UV ex-
posure during adulthood for melanoma risk, there is great po-
tential to explore this question and the important issue of inter-
actions between risk factors in future follow-up studies of these
Norwegian and Swedish women.
(1) Parkin MD, Pisani P, Ferlay J. Estimates of the worldwide incidence of 25
major cancers in 1990 Tnt J Cancer 1999;80:827-41.
(2) Ferlay J, Bray F, Pisani P, Parkin DM. GLOBOCAN 2000: cancer inci-
dence, mortality and prevalence worldwide. Version 1.0. lARC Cancer-
Base No. 5. Lyon (France): IARC Press; 2001.
(3) B, Fekjaer H, Hakulinen, T, Tryggvad6ttir L, Storm HH, Talback
M, et al. Prediction of cancer incidence in the Nordic countries up to the
year 2020. Eur J Cancer Prev 2002; I I Suppl I :S 1-96.
(4) Armstrong BK, English DR. Cutaneous malignant melanoma. ln: Schot-
tenfeld D. Fraumeni IF Jr, editors. Cancer epidemiology and prevention.
2nd ed. New York (NY): Oxford University Press; 1996. p. 1282-312.
(5) Green A, Trichopoulos D. Skin cancer. In: Adami HO, Hunter D, Tricho-
poulos D. Textbook of cancer epidemiology. New York (NY): Oxford
University Press; 2002. p. 281- 300.
(6) Eklund G, Malec E. Sunlight and incidence of cutaneous malignant mela-
noma. Effect of latitude and domicile in Sweden. Scaod J Plast Recoostr
(7) Magnus K. Incidence of malignant melanoma of the skin in Norway,
1955- 1970. Variations in time a11d space and solar radiation. Cancer 1973;
(8) Armstrong BK. Melanoma of the skin. Br Med Bull 1984;40:346-50.
(9) International Agency for Research on Cancer (IARC). !ARC monographs
on the evaluation of carcinogenic risks in humans. Vol. 55. Solar and
ultraviolet radiation. Lyon (France): IARC; I 992.
{10) Elwood JM, Jopson 1. Melanoma and sun exposure: an overview of pub-
lished srudies. Tnt J Cancer 1997;73: 198-203.
(II) Whiteman DC, Whiteman CA, Green A C. Childhood sun exposure as a
risk factor for melanoma: a systematic review of epidemiologic studies.
Cancer Causes Control 200 I; 12:69- 82.
{12) Bliss JM. Ford D, Swerdlow AJ, Armstrong BK, Cristofolini M, Elwood
JM, et al. Risk of cutaneous melanoma associated with pigmentation char-
acteristics and freckling: systematic overview of I 0 case-control studies. Int
J Cancer 1995;62:367- 76.
Journal of the National Cancer Institute, Vol. 95, No. 20, October 15, 2003 ARTICLES 1537

"' c.
(13) Weinstock MA, Colditz GA, Willett WC, Stampfer MJ, Bronstein BR,
Mihm MC, et al. Moles and site-specific risk of nonfamilial cutaneous
malignant melanoma in women. J Natl Cancer Inst 1989:81:948-52.
(14) Weinstock MA, Colditz GA, Willett WC, Stampfer MJ, Bronstein BR,
Mihm MC, et al. Nonfamilial cutaneous melanoma incidence in women
associated with sun exposure before 20 years of age. Pediatrics 1989:84:
199- 204.
( I 5) Weinstock MA, Colditz GA, Willett WC, Stampfer MJ, Rosner B, Speizer
FE. Recall (report) bias and reliability in the retrospecti ve assessment of
melanoma risk. Am J Epidemiol 1991;133:240-5.
( 16) Cockburn M, Hamilton A, Mack T. Recall bias in self-reported melanoma
risk factors. Am J Epidemiol 2001;153: 1021-6.
(17) Geigy scientific tables. Vol. I. Units of measurement, body fluids. com-
position of the body, nutrition. 8th ed. Basel (Switzerl and): Ciba-Geigy;
1981. p. 227.
(18) Holman CD, Armstrong BK. Pigmentary traits, ethnic origin, benign nevi,
and family history as risk factors for cutaneous malignant melanoma. J Nat!
Cancer Inst 1984;72:257-66.
(19) 0sterli nd A, Tucker MA. Hou-Jensen K, Stone BJ, Engholm G. Jensen
OM. The Danish case-control study of cutaneous malignant melanoma. I.
Importance of host factors. lnt J Cancer 1988;42:200-6.
(20) Lock-Andersen J. Drzewiecki KT. WulfHC. Eye and hair colour. skin type
and constituti ve skin pigmentation as risk factors for basal cell carcinoma
and cutaneous malignant melanoma. A Danish case-control study. Acta
Derm Venereol 1999;79:74-80.
(21) Beitner H, Norell SE, Ringborg U, Wennersten G, Mattson B. Malignant
melanoma: aetiological importance of individual pigmentation and sun ex-
posure. Br J Dermatol I 990; I 22:43-51.
(22) Klepp 0, Magnus K. Some environmental and bodily characteristics of
melanoma patients. A case-control study. lnt J Cancer 1979;23:482- 6.
(23) Augustsson A, Stiemer U, Rosdahl I, Suurkula M. Common and dysplastic
naevi as risk factors for cutaneous malignant melanoma i.n a Swedish
population. Acta Derm Venereol 1991;71:518-24.
(24) Tucker MA, Halpern A, Holly EA, Hartge P, Elder DE, Sagebiel RW, et al.
Clinically recognized dysplastic nevi. A central risk factor for cutaneous
melanoma. JAMA 1997;277:1439-44.
(25) Bataille V, Grulich A, Sasieni P. Swerdlow A. Newton Bishop J. McCarthy
W, et al. The association between naevi and melanoma in populations with
different levels of sun exposure: a joint case-control smdy of melanoma in
the UK and Australia. Br J Cancer 1998;77:505- 10.
(26) Titus-Ernstoff L, Thorn M, Tosteson TD, Brahme EM, Yuen J, Baron JA,
et al. The accuracy of skin self-examination for atypical nevi. Epidemiol-
ogy 1996;7:619-23.
(27) Pfahlberg A, Kolmcl KF, Gcfeller 0 ; FEBlM Study Group. Timing of
excessive ultraviolet radiation and melanoma: epidemiology docs not sup-
port the existence of a critical period of high susceptibility to solar ultra-
violet radiation-induced melanoma. Br J Dermatol 2001 ;144:471-5.
(28) 0sterl ind A, Tucker MA, Stone BJ, Jensen OM. The Danish case-control
study of cutaneous malignant melanoma. II. Importance of UV -light expo-
sure. Lnt J Cancer 1988;42:319- 24.
(29) Westerdahl J. Olsson H, Ingvar C. At what age do sunburn episodes play
a crucial role for the development of malignant melanoma. Eur J Cancer
1994;30A: 1647-54.
(30) Swerdlow AJ, Engl ish JS, MacKie RM. O' Doherty CJ, Hunter JA, Clark J,
et al. Fluorescent lights, ultraviolet lamps, and risk of cutaneous melanoma.
BMJ 1988;297:647-50.
(3 I) Walter SD, Marrett LD, From L, Hertzrnan C. Shannon HS. Roy P. The
association of cutaneous mal ignant melanoma with the use of snnbeds and
sunlamps. Am J Epidemiol 1990;131:232-43.
(32) Autier P, Dore JF, Lejeune F, Koehne! KF, Geffeler 0, Hille P, et al.
Cutaneous malignant melanoma and exposure to sunlamps or sunbeds: an
EORTC multicenter case-control study in Belgium, France and Germany.
lnt J Cancer 1994;58:809-13.
(33) Westerdahl J, Olsson H, Masback A, lngvar C, Jonsson N, Brandt L, et al.
Use of sunbeds and sunlamps and malignant melanoma in southern Swe-
den. Am J Epidemiol 1994;140:69.1-9.
(34) Westerdahl J, Tngvar C. ~ s b a c k A. Jonsson N, Olsson H. Risk of cuta-
neous malignant melanoma in relation to use of sunbeds: further evidence
for UV-A carcinogenicity. Br J Cancer 2000:82:1593- 9.
(35) Swerdlow AJ, Weinstock MA.. Do tarming lamps cause melanoma? An
epidemiologic assessment JAm Acad Dermatol 1998;38:89- 98.
(36) Westerdahl J, Anderson H. Olsson H, l ngvar C. Reproducibility of a self-
administered questionnaire for assessment of melanoma risk. lnt J Epidc
miol 1996:25:245- 51.
(37) English DR. Armstrong BK. Kricker A. Reproducibility of reported mea-
surements of sun exposure in a case-control study. Cancer Epidemiol Bio-
markers Prev 1998;7:857-63.
In Norway, the sur vey was s upported by Public Health Service grant CA-
52449 (to Professor E. Lund) from the National Cancer Institute. National In-
stitutes of Health. U.S. Department of Health and Human Services; granLS 90050
and E 00065 from the Norwegian Cancer Society; and grants from the Aakre
Foundation. ln Sweden, the survey was supported by the Swedish Counci l for
Planning and Coordination of Research, the Swedish Cancer Society. Organon,
Pharmacia. Medical Products Agency. and Scheri ng-Piough.
Manuscript received February 20, 2003; revised May 20, 2003; accepted
August 21,2003.
1538 ARTICLES Joumal of the National Cancer Institute, Vol. 95, No. 20, October 15, 2003
"' c.
nt' the Amenun C.:.llg nf Co.rdinlng:r
It) hr thr A:nctjc:tn C..U9.<u af Caul.inltlgy Fc:w.dation
t'uhll.<hcd hy Eln.-vior bl<!.
Vitamin D Deficiency
An Important, Common
and Easily
Treatable Cardiovascular Risk Factor?
Vol 52, Na. 21, 200l!
ISSN 07351097/081$34.00
doi:lO. 1016(j,j3cc.1008.09.050
John H. Lee, MDt James H. O'Keefe, MD, David Bell, MD,t Donald D. Hensrud, l\IID, MPH,:f:
Michael F. Holick, MD, PHD
Kansas City. Missottri, Birmingham, Alabama; Rochester, l'V!irmcsota.: and Boston, Massachuutts
Vitamin 0 deficiency Is s highly prevalent condition, prosont in approximately 30% to 50% of lha ganerol popula-
tion. A growing body of dotn suggests !hill low 0 lovols may
health. Vitamin D deficiency activates the system and can predispose to hyperten
slon and left vontrloulilr hypertrophy. Additionally. vitamin D deficiency cousos on inercilSO In pornthyrold hor-
mone, which increases insulin raslstoneo ond is ossociotod with diilbotos, hypertension, ond in-
cardiovascular risk. Epidemiologic studies hOve iJSSOclated low 0 levels with coronary
rnctors and adverse cardiovascular outcomes. Vitamin 0 svpplementlltlon Is simple, safe, and IneXPensive.
Lurgorundomized controlled trials are needed to firmly establish the relev<Jnce of vltnmln 0 status tl;) oardlovas
c:ular health. In tha moilnwhllo, monitoring serum 25hyaroxyvltamin o lovols Md correction of vtttlmin o dofJ-
clency Is Indica tad tor optlmlz;.,tlon of musculoskeletal and general (J Am Col! 2008;52:
1949-56) 2008 by tho Am orlean College of Cardiology Found!ltlon
Traditionally, vltamin D has been associated
bone healrh, and it is well tmdcrstood that vitamin D
deficiency lends to rldcct.c; in children nnd osteomalacia and
osteoporosis in adults (l). Howeve.r, it i.s now known
adequate vitamin D status is important for optimal function
of many organs and tissues throughout rhe body, including
the cardiovascular (CV) system (2). Vimmin D receptors
(VDRs) arc present on :1 L'trge variety of cell types, including
myocyret;, ca.rdioroyocytcs, pnncrc<ttic vuscular
e'ndothelial cells, neurons, immune cells, and osteoblast:; (1).
Vitamin D deficiency or insufficiency prevalent in prac-
tically every segment of the U.S. populacioflt including
children and young adults (f). This worldwide pandemic
. remains generally Wliccogn i.7.e.d and untreated.
Evolving dttt:.\ indicate that vltamin D deficiency is
playing an important role in the: genesis of coronary risk
and CV disease:. Vitamin D deficiency seems to
predispose: to hypertension, cliahetes and the metabolic
syndrome, left ventricular hypertrophy, congestive heart
fiillure, and chronic vascular inflammation (1,2). Epidemi-
Vro"l the MJd lkut ln;titutr uml uf Mi..auri, K.1m1 City,
Mi>IJ,lllri: Emlocrinofnc;y, AJp\r.lmr.; Clinic, Roch-
<tcr,IIMinn"r.ot..; 2nd !iB<,."tt>n University Mcdic:\1 Center, ilonon, Mnmchuscm.
Dr. O'Keefe is nn unpnid 5cic:ntitlc COtHultant tu C""ilii)Tb. Fulldl derived from tllis
comp:111y Ate \l!cd for tnarkcring :md potlr.nr CRrdinvuculr Con.,.ltunt<,
tbe I;IOUp vrJCUCC ><ith which Or. O' Kectc lt Dr.. Holi,;lc i p
fur thr Nut:ional Dairy C<luncU me UV 1-'oundntiQn.
Mltnwt:ript receit ... d Jup. 4, 7.006; rc:viaed 111111\llctipr recdl'td 6, 20011,
=.prr,.d Augur 13, '-OOS.
ologic :;cudi!!s have also recently linked vitamin D deficiency
with incrensed risk of mu.jor adverse CV cvco.ts (3). A study
of male health professionals showed a 2-fold risk of myo-
cardial infi\rction (Ml) in t>ubjects who were: vitnolin D
deficient compared with those in the range (4).
Similarly, a rc:cent prospectiw cohort study measured the
vitanlin D levels in 3,258 German adults who were under-
going elective cardiac cathe[en7.ation. During a mean
follow-up of7.7 years, indiv.idullls in the lowest quartile for
baseline serum 25-hydroxyv.ita.rnin D [25(0H)D) had a
risk-adjusted 2-fold increased risk of death, especially CV
death, cornpar.ed with those in the highest of
vit11.mi.n D (5).
This review focuses on the relatlonship between 2 wide-
spread problems: v.itanlin D deficiency and CV disease. The
i.ssues addressed will include:: 1) the: role of vitamin D
deficiency in the genesis of coronary risk factors and adverse
CV 2) how repletion of v.immi n D stores may
improve CV health and prognosis; and 3) practical :tnd
specific for restoring i\nd maintaining a
healthy vi.tumin D in pnticnts, because no
guidelines have been published on this topic yet.
VItamin D Baslc:s
Vitamin. D comes in 2 forms: vita.min D
nnd D
(cholecnlciferol). Vitamin Dz, found in
is the propuct ofuhraviolct B (UVB) (290 ro 315
mm) irradiation of and cnn. be consumed as
Vltlmln D Doficiency and CV Risk
:u1d Acronyms
c::v .. catdrava. culJtr
Ml = III)GC8 rdllll Jnlatetlan
IIV8 Ultnlllfalat 9
11tm 1.4tamln D >c:captor
a supplement or in fortified
foods (1). Vitamin D
, a prod-
uct of UVB irradiation of 7-
is synthe-
in the human epidermis or
consumed in the form of oily
fish, fortified foods, or a :;upplc-
mc:nt. Excessive sunlight
sure c:mnot vitamin D tox-
icity because UVB
ccccss virom.i.n to biologically
ioen isomers (l); ho\vcver, QCccs-
sive oml \':it:unin D intake can cnuse toxicity at very high
doses (6).
Vitamin 0 is converted in the liver to 2S(OH)D, which
is t:he major circulating metabolite of vit:un.in D. Serum
25(0H)D concentrations, which reflect both vitamin D
inrnke nnd endogenous producti.on, should be measured to
clinic:illy :tsscss vitamin D status (1). In the kidney,
25(0H)D is by lcr-hydroxylase to its active form,
1,25-dihydroxyvit:unin D [1,25(0H)
D], which n
vit-.11 role in maintnining boac and health by regu-
lating calcium metabolism. Although 1,25(0HhD is the
active form of vitamin D, it& serum level docs not cortelate
with over:ill vitamin D stntus and thus is generally not
clinically lL<ieful. (1).
Vitamin D in the form of 1,2S(OH)p is a hormone,
because it ic; produced primarily in 1 orgnn (the kidney) and
chen circuhttc.c; throughout the body, where it c.xerts wide-
ranging effects. The VDR is present in mosr: til;sues,
including endodtelium, vasc:nl:lr smooth mut;de, :tnd myo
C.'U'dium (2). In nddition, both vasc:ulM smooth muscle and
endothc.li:U cells may have the ability to convert 25(0H)D
to 1,25(0HhD (7). Circulatory 1,2S(OH)
D crosses the
cell and cytoplasm tU1d reaches the nucleus,
where it to the VDR. The VDR-bound 1,25(0H)
in turn binds to the retinoic acid x-rcceptor and serves u.s a
nuclear tronscription fuctor, altering gene function and
inducing protein synthesis (1). Directly or indirectly,
D regulates over 200 genes, including those
involved in renin production in the kidney, in!\ulin produc
cion in the: p:mcn:as, release of cytokincs .&om lymphocytes,
production of cathclicidin in and growth nnd
proliferation of both vascular smooth muscle cell!1 and
cardiomyocytes (1).
Definition and Prevalence of Vrtamln D Deficiency
Although a consensus regarding the optimal level of serum
25(0H)D has not yet been established, most e.\.-pe.tts define
vitamin D deficiency a 25(0H)D level of <20 nglml (50
nmoVl) and -+itarnin D in:;ufficiency as 21 to 29 ng/ml.
(Table 1). For all studied end points to d11.te, the optimal
concentncion of .lS(OH)D is at lenst 30 ng/ml (8).
A rapidly evolving knowledge base indicates that vitamin
D deficiency is mucl1 more prcvnlcnt thlln previously rec:-
JACC Vol. 52, No. 24, 2008
December 9,
ognizcd and is prese.tu in up tc> SO% of young adult!l (9) and
a.pparemly healthy childrer'l (1), 'lhe Third Nation:U. Health
and Nutrition Examination Survey (NHANES In) re-
ported the prevalence of vitamin D deficiency in the U.S; to
be between 25% and 57% of adults (10).
The prevalence of vitamin D deficiency increases in
proportion to distance from the equ.'\tor because of increased
atmospheric filtering of UVB radiation caused by the
oblique :mgles of the lt-un's rays at higher latitudes. Addition-
ethnic groups with slcin require proportionally mon:
!>1Jn o..-po!>urc to equiV'..tl.cnt :J..!llOWlt<; of vitun.in D
compared with people with lighter skin (11).
Modern humnn cultures produce less vitamin D cutane:-
ously, in part because of increasingly indoor lifestyles and
cffons to minimize sun c::.:>..-po:;u.rc by using and
other sua avoidance strategies. Sunscreen with u sun pro-
tection fac:ror of 15 blocks 99% of the
cut:IJ\eous vimmin D production (12). AdcUtiontill.y, obesity
is associated with vifll.mjn () deficiency (13), probably
because of a dec:rcascd bioav.Wability of vitamin D that is
sequestered in the fat of indivi duals \vith C."<CeSS adipose
(14). After equivalent to UVB radiation or a
do!'e of vitamin D:
obese individuals showed 50%
lower blood levels of vitamins D
und D
compared with
nonobesc individuals, probably because of scque<;tcring of
25(0H)D in adipose tissue (14). Older age also reduces the
capacity for U\fB-induced cutaneous syntheSis of vitamin
D. After equal doses of runl.ight C);posure, a 70-year-old
person produces 75% less vitamin D
than a 20-yeru:-old
person (15). Other risk factors for vit:u:n.in D deficiency arc
listed in Table 2.
VItamin D Deficiency and cv Disease
.Epidemiological studies report that the rates of coronary
heart diabetes, nnd

like vitamin D
deficiency, increase in proportion to increasing dist:mcc
from the equator (16). Defi.cienr or insuf!icitnr serum
25(0H)D lcvcls have been .documented in patients with
myocardial infarction (17)_. stroke (18), heart fiillure (2),
diabetic CV disease {19), and peripheral arterial disease
(20). Recendy, the rcL'\tionship between CV risk and
25(0H)D levels was explored among the 15,088 subject!:
from the NHANES Til national cohort registry. In this
study, 25(0H)D levels were: inversely asso-
ciated with hypertension, diabetes mdlirus, hypcrtriglycer-
idemia, and obesity (21). Other cross-sectional studies have
:ZS.Hydrol(Y'<twnln 0 (nVml)

Vltllmln D S13hn
S1>\'l>n: ddicflongy

Toxic ley
: . We!
. bas
. yea
. nor.
. to!
. . 11lCI
; wa:;
.: Wit)
t, 2003

) re-
.S. to
:s .in
. the
: esity
at is

. ora
Jg of
; the
' -old
: -old
r are
. nary
: (2),
' .
.. ; t ; . ... : . .. ; ... : ; . . t ' : ;.. .. .
Factors . .for Vltamln o.:nencroncy .: : .... , .. ; : .. .... ,
. ' f- '!: .... .,..\; .. ... .. .... . . ... ,.. . ' '
)motllutlon"lh:rd at homebound
? dist:once lrotn :<:qulllat
. >: C<>vcMJP cl otnlniC nnll/o;,,un&cre&n
... Aft ooltu\lon
""'!l+"" l ...
':. Mn!absol'l'tiD<I
Ronal dl5<:;>$0
.' llvOf df5Cl!SO
Me<lleatlons.: ,tllueocoi'\Jeold,, enllrejectlon, om! tillman
vfrus moctlcgtlon'

:corrfirmed the links bcrween vitamio D deficiency and b<1th
hypertension and diabetes (22,23). Additionally, vitrunin D
deficiency predisposes to insulin rt!Sistance, pancreatic beta
cell dyofunction (24), and the: metabolic :l)'lldrome (24,25) .
One reported that n daily intake of 800 ill of vitamin
D compared with a dailyintake of <400 JU of vitamin D
reduced the rilik of type 2 diabetes by cue-third (26). A
study of 10,366 Finnish children who were given 2,000 IU
of vitamin per day thr.oughout che first year of life
ocpccicncc:d a 78% reduced risk of type 1 diabetes over the
ensuing 31 years of follow-up (27). Subsequently, this
finding has been confirmed by a meta-analysis performed on
5 stuclies by t\ group in England (2S).
A co.rcd.ation between vitamin D deficiency and
sequent rn:1jor adverse CV events was found among the
1,739 Framingham Offspring Study participants- who
were free of CV di!iease at baseline: (3). In this prospective
ob!ien'lttional study, 25(0H)D levels were mca.-:urcd at
baseline and subjects were followed up for u rnenn of 5.4
years. The rate of a composite CV end point (fatal or
nonfntal MI, ischemia, stroke:, or heart f.lllure) was 53%
to 80% higher in people with low vitamin D levels. The
CV risk associated with vi. mmin D deficiency
w:1s m:1gnilied in the cohort of Frarningh1ttn offspring
.with hypertension (Figs. 1 and 2).
Vitamin D deficiency predisposes co up-regulation of the
rcnln-gngiotensin-aldosterone l>Y,item .il!ld hrpertrophy of
both the left ventricle and vascular smooth muscle cells {2)
(Fig. 3). In vitamin D- deficient anirnnls there is an in-
creased incidence of hypertc:n. 1on, left vc:ntricul\\r hypertro-
phy, and athetosc.lerosis (29). Human studies indic:\te that
D inhibits renin synthesis,, which may
blood pressure (30). Krause et al. (31} showed that incre:\sed
c:."Posurc to UVB radi.ation in a tanning bed 3 times per
week for 3 months led to a 18Ql!;b increase in 25(0H)D
levels and a 6-mm Hg reduction in both systolic and
diastolic A small, randomized, placebo-controlled
t;tudy of patients with type 2 diabetes and low baseline
25(0H)D levels showed thac a single dose of 100,000 IU of
vitamin D
reduced systolic blood pressure: by a mean of 1.4
Lee et al. .1951
Vlt11mln D Deficiency CV Rlsk
nun Hg and signifiumtly improved endothelial function as
measured hy forearm blood flow (32). J n the NHANES ill
study, the mean blood was about 3 mm Hg
lower in those in the individuals in the hlghest quincile of
serum 25(0H)D levels compnred with those in the lowest
quintilc (22) .
Hyperparathyroidism Increases CV
Chronic vitamin D deficiency causes secondary hyperpara-
thyroidism, which in turn may mediate many of the detri-
mental CV dfects of inadequate vitamin D lcvcb. The
A o.s r------------
B o.a 1""'"""'----------
0 0.25 1---------------
0.2 1----------------



26-(0HJ D !ovals< 'IS nglmL j ..
0.05 1-....;_....;.._ ___

2 3 4 5 6 7
. D .. ':\
, .. . ': :: , . ,. . .. '\ t , . :;:,,. . . : . ,:
Vit:lmin D dllficlancy CV rlsk. CUM!& show 1110 orobabll
lty of m!l]or OCllhlrse CV ownts In partleloant& wlttl 2.!1(0H)O teii81S :::.1.5 nc:/mt
(lll'llon ttnoo) eno 2!S(OHJD l evets <15 ng,tmt (red nnea). Tile tncmesed cv
rtsk was more sppStent In pattanl!. wnn (A) In plstients wllhlllit (8) nype,..
tllnBIOil. Reprlnlt:d, With rtom Wang el. (3), 25(0H)O - 2.S.hy-
Ciroy;ttaonln o: cv =- CVD :o c:.ofdiov=,.cutor dlzc::osc:.
"" <D 1.2
lee etsl,
V'lbmin D und CV Risk
25 (OH) D Level
< 10 nglmL
. 1or: bvo.br. Vitimln:o
. . : . ,; . . . I . . . : . I: . . . .. ... : .. ..
odjustod hiiUII'II ratloa tor m'1)or udvors!t
CV Oom from wan, e1 al. (:l). Mlbrovliltloll!l as In Fi!lurc 1.
threshold for elevation of parathyroid hormone (PTH) is a
25(0H)D level of <30 ng/ml. Furcher decreases in
25(0I-I)D levels will rc.crult in proportionally higher PTH
levels to m:llnt:Un strum and total body calcium (Fig. 4).
Vimrnin D ddiciency reduces intestinal cnlcium nbsorption
by more than 5()1.){, (1). The attend'lT'It decrease in serum
calcium triggers PTI-I releasr;:, which quickly corrects
the calcium levd by mobili1.ati.on of cnlcium from bone,
renal tubular calcium reabsorption, t\ncl
rennl production of 1,25(0H)J).
JACC Vol. 52, No. 24, 2008
Oecemt>et 9, 2000:1949-56
The effects of prim:uy hypc:rparathyroidism on CV out-
come& were shown in a study th:tt reported approximately
40% lower relative rlsks of MJ
stroke, and death in patientS
who had surgic:U parathyroidectomy compared with obser-
vation (33). 'I11i:; UoJ.: between inCJ:e:tsc:d PTH and CV
disease was further corroborated by a. srudy of with
renal failure and necondnry hyperp:u-achyroidism by
decreltSed conversion of25(0H)D to 1,25(0HhDJ. In this
study, patients with a PTH levd 2::250 pg/mJ had a 2-fold
risk of CV compared with those with PTH
<250 pg/ml (34). Additionally, a recent obscrvntionnl study
found that c:lcv:ttcd PTH level-; in elderly individuals was
associated with a doubling of mortality during follow-up
compared with those with norm:U PTH levels (35),
An increased PTH level is with increases in
both blood pressure (36) and myocardial contractility, which
eventually lew to hypertrophy, apoptosis, and fibrosis of
both the left ventridc: nml vascular mcdinl smooth muscle
(2). Vit:\min D deficiency and/or increased P11-i also
prt:dib-posc to calcification of hel\rt v:tlvell, mitral annulus,
:md myocardium, especirllly in paticnt'l; wirl1 moderate or
severe chronic kidney (37).
Chronic kidm:y dilicasc is with markedly in-
creased CV risk (38), which may in part be mediated by
inlldequatc vitamin D lcvck Vitamin D deficiency is asso-
ci:ltcd wid1 increased mortality mtes in d1e Retting of chronic
kidney (39), and rcplcting vimmin D in such
parients improves outcomes. Recent observational srudies of
patients wich chronic lcidney dise-.1sc and
ism fo11nd that the oral admioisrration of 1,25(0H):J):l
- --------- -- - ----
:: _ .. .. ' :., j
r------t Vlremin.o t----..,
. ' . , .....
' ., .. \ .
- . \ .: , : , , ; , . I; , : ' ,; , ..1 , .'' ;: : ,
0 !' .... . . . .. . ' .. . :: . .
Po:;:;lblc m<!Cil&nl9nl' or tCVI nsl< rrcnt viu.mln 0 oonc1ency, PTH = pnruthyroid hormone: RM!:i - renin -'tl!!iotons/n;lldostr.ronc sy51om.
I, 200S
:d by
1 this


is of
te or
'jt in-
d by
es of
Jt<CC Vel. 52. No. 24. 2008
[)ecembcr 9, 2008:1949-56



05 6-10 11-15 ll-2.5 26-30 >30
25-Hydro1tyv1Uimln D (ng/ml)
I!P. -o; < ::" :: :' :
. or a!l'd : : : . ! ..
The ln-..:rso: relotlonsnlp !l9twel!fl 5erurn o levels and serum
p=thylolo hcumone 1"--eH. Reprinto<l, wnn permission. ftOrn Z\ttr:w.,., ct ol. (2).
(also known liS activated vitamin D or cnldtriol) was
associated with signilicandy itnprovcd survival ( 40). A
placebo-controUed study of: 30 pre-dialysis renal fuilure
patients with !;econdacy hyperparathyroidism showed that
treatment with 1,25(0HhD
improved left vontrkul.nr
diastolic function (41). Vitunin D :malogs used in patients
on dialysis b1we been shown to improve survival at all
tested thus far { 42). ..
Low 25(0H).D level:; and incre:tSed PTH lcvdll increase
both inilammation and the risk of adverse CV events
(36). Vit:unin D deficiency increases systemic iofinrolmt-
tion, :lli documl!nted by elevated levels of C-rcactivc protein
and intcrlc:ulcin-10 (2). Furthermore, adnJinlstracion of
1,25(0HhD to vitamin D-deficic:nt individuals down-
regulated inflammatory markers (C-reaccive prorein, nnd so
on) and conferred an anti.prolifecative effect (43). E:rtr:u-cnnl
of 1,25(0HhD occurii through cytokinc stimula-
tion (7) and is locally important in the: paracrine regulation
of cellula.r growth, d.i1fcrentinrion, :md function {44). This
may explain why vitamin D deficiency has been assochtted
with type 1 di:lbcrc.s, cartcer, and multiple sclerosi.o; (45).
Out come Studies
A recent meta-analysis of 18 randomized controlled trials
comprising 57,000 individuals showed that a vitamin D
intake >500 ill/day improved all-cnusr: mort:1lity, in part by
decreasing CV dctlths (46). Moreover, the dati. rcg:u-cling
both efficacy and CV s:tfcty for vit.'U'Tl)n D nppellC to be
to thnt fur calcium supplements. Indeed, calcium
supplemeottltion ha.<; recently been implicated ns possibly
increasing d1e risk of adverse CV c:vents, particularly in
LM atl. 1953
Vlbmin D Dt!ficie"cy und C:V Risk
patients with chronic kidney clisense (47,48). Calcium sup-
plements acutely increa.c;e &rum calcium levels, which might
accelerate c.alcification (49). In cont:rMt, serum
vitamin D lc:vds arc inversely nssociated with coronRry
artery calcific.ation (50).
Ostl!oporosis and ad1crosclerotic CV disease share many
common risk factol'l>, and \1 pathologic link between these 2
highly prevalent age-related cliHeascs been suggested
{51). A latgc! number of midcllc-ngcd to elderly individuals
at risk for both CV disc:!Sc nnd osteoporosis may benefit
from therapies that nrc likely to improve both conclitions,
such as an anti-inflarnmarory diet, daily c:xc.rcise (especiallr
weight-bearing fu.rrm), avoidance of both tobacco :md heavy
alcohol inmke, and possibly vittmin D supplementation.
Vltantin D Myopathy
o.re genc:rnlly the fir.it manifestation of vitnmin D
deficiency. Severe: vir.mun D deficiency with correl>ponding
elevations of PTH were reported in 88% of 'women who
presented \vith muscle pams :md weakness (52). Another
study investigated 150 padenn; \vlth nonspecific musculo-
skeletal pains and reported that 25(0H)D levels were
insufficient in of individuals and severely deficient in
28% (53). A metn-malysis of 5 randomi7..ed clinical trials
reported that vitamin D supplementation reduced the risk
of .fitlls, mosr likely from improved muscle function and
strength (54). Mynlgia, the most common complaint
ported by patients on statin therapy, may be at in put
caused by undedying vitamin D dc.ficicncy. Anecdotally, we
have observed that repletion of 25(0H).I) levels predictably
improves or resolves statin-reh\ted myalgias.
Supplementing Vitamin D
Traditionally, up to 95% of the body's vitamin D
ment comes from the synthesis in the c:piderml!; on sun
exposure, with the remainder ingelited from dictaty sources
(Ta]Jlc: 3) (55). The U.S. government's current rccommen.-
dation for ornl '<itamin Dis 200 IU daily for individuals age
<50 years, 400 ru d.-illy for .between age 50 and
70 yeart>, [Uld 600 ru for those older than age 70
Studies indicate that the average U.S. adult consumes o.bout
230 ill vir:unin D pt!r. ch1y (56). However, ic has been
Cod llnr oil, 1
illl mon, 3 oz
Farmed ""lmon, 3 oz
MQCICG!el, cooked, 3 o.
Tuna nSII, eennod In o!l, 3 w
San!lneJ (Willi on. 1 cu
Mllk. non!:Jt. lat. ana Ylllol&,
\Ol1;:tmfn o.fonlnad .L eul)
Oolol lrom Holick (1) tho tlollonnlln&dMe of H .. lth
IU per &crvf111i!

l 00-250
l9S4 Lee et al.
Vitamin D DeficienCy and CV Risk

:!.! ...
Durutlnn ofVJ<mln D rupplcmaut.tdon (tnonlh.o)

,.: .. : .. . ::, .. .: .. !
-: .i: :;. ;d:, !
ttect or do9& ena dursuon or D supplomoounlon on tM mean
serum 25-llyarol<)"'il;ornio D [25(0HJDJ ooncontrauon acllleved. from Vieth
et st. (62),
estimated that 1,000 to 2,000 IU is oecc:ss:uy to satisfy the
needs for people (8). Mwy experts in the ficld
suggest the recommended daily intnkc: of vitn.min D be
incrc:ascd to at least 800 ro .2,000 IU of vitamin D daily,
doses that are difficult ro achic:ve without rupplemenmtion,
pa.rcicularly in higher latitudes and in of c:xtrc:mc:
winter climate. A dose of vitamm D
up to 2,000 IU daily
has been deemed by the: U.S. Food and Drug Adm..inistra-
tion's nutritional. guidelines to be genernl.l.y recognized as
sue. A rect:nt review concluded that the safe upper limit for
vitamin D consumption is 10,000 IU per dn.y (57); doses
JACC Vol. 51, No. 24, 2008
9, 2008:1949-56
above this increase of renal calculi formati.on, ' c:specially
in with absorptive hypc:rcalc:uria nnd end-stage
renal dise:I.Se patients on dialysis (SS).
A study of 340 children ages 10 to 17 years found that
increasing cl1e intake of oral vitamin D 10-fold, from the:
cw-rcntly recommended dose of 200 to 2,000 IU daily, \-r.\S
required to reach a 25(0H)D lc:vel of 30 ng/m.l (the lower
end of the optimal range) (59). Tho investigators concluded
that doses equivalent to 2,000 IU of vitamin D
daily were
not only sue for adolc::;cents, bur also necessary for nchiev-
ing the: desirable vitamin D lcvc:ls.
The most poten.t 5ourceA Dare sunllght (about
3,000 IU vitamin D
per 5 to 10 min of mid-day, midyc:u.r
exposure of arms legs for n light-skinned C<lucashm) or
prescription oral supplements of 50,000 IU C!lpsule of either
vitumin D
or D
every 2 weeks (1). Among foods, oily fish
have: the highest content of vitamin D
, which mnges from
100 to 1,000 IU per 3.5 (1,60), wherens other sources
such as milk or orange juice forri.6.ed wich vitamin D con min
up to 100 IU per
As a general rule, every 100 IU vitamin D ingested daily
incrca . .c;es the 25(0H)D level by about 1 ng/ml (61,62) (Fig.
5). Over-the-counter dietary supplements of vitamin D
typically comain 400 ro 5,000 ID pc.r capsule. Oral
supplementation with either vitamin D
or D
initially will
increase v.itrunin D lcvds cqu:illy well (63), although the
incrc:uscs in serum 25(0H)D lc:vcls to persist
after a bolus dose ofvitamin D
than (64).
Treatment of vitamin D-defictent individuals should be
initiated "vith 50,000 IU of vitamin D
or D
weekly for a
period of 8 to 12 weeks. Once the initial repletion phase is
complete, maintenance thcrnpy can be continued in 1 of 3
1) 50,000 IU vitamin D
or D
every 2 weeks; 2)
25(0H)D In 31o 6 months
:Z5(0H)O "' 2&1Jydf'OX)N1tamln o.
J'7nr'7 C'tar. nTift
.:, .. . 12.
,., : : 15,
. ..
aid year
:ian) or
f either
ily fish
'!S from
d daily
. ') o-
. 'lg.
1in D
:. Oral
lly will
uld be
1 for a
1ase is
1 of3
ks; 2}
... ... .
;, .. : ...
-: ,,.; :.
to 2,000 IU vitamin D:! daily; and 3) $Untight
for 5 to 10 min for Caucasians (longer times
........ for people with increased skin pigmentation) be-
the hours of 10 AM to 3 l'M (spring, summer, nnd f.ill)
(F;g. 6).
authors thank Lori J. vVJbon for hc:r help in pt'epara-
of this article.
-----------.. .... .. . .... .. .
requests nnd correspondence: Dr. J:unes H. O'Keefe,
Woi'Il!l!l Road, Suite 2000, City, Mi,;souri 64111.
. Holick MF. Vitnmm D deficiency. N EnglJ Mc.d
Zirn:rmnnn A. Yi=in D prevention with specinl refer-
. ence m ordlov:uu:uhr Ptog Mol Blol2006;92:39-49.
W1111g TJ, Pencion .!VJJ, Booth SL, et ul. Viblmin D ri'k
of ordiov:u<:Ulu Cl.rcuhtion 2008;117!503-11.
4. Gio .... .tnnucci E, Liu Y, BW, Rimm B. 25-hydtoxyvitAmin D
:md risk of myocnn:lial ioEw:tion in men: prospective kch
': Jnrctn Mc:d 2008;1GS:ll74-SO.
::: '5 . .Dohnig H, Pili. S, Sch:un3g1 H, et 11l. Independent :wac.i3clon of low
sctum 25-hydt(l).'}"'lt;ltl'lin d 11cd 1,25-cf.hydtuxyvitRmin D hwcb with
ill-omse 1\od c.1ttdioYUSGUlnr mart'ility. Arch lntem Med 2008;16S:
1340-9. .
6. Koutkin P, Cbcn TC, HolicklV1F. Vl=in D intoocication as6ociJlted
with an over-the-counter ropplcmcnr. N Eng! J Med 2001;345:66-7.
7. Zchndc( D, :BJ;1.od R, Chima RS, et Ill. o.f J.,2S-
dihydroxyvitllmin D(3) by humnn cndothcli:Ll cell! is rc:gubrc:d by
infl:lmmntory cytokincs: n novelnul'OeriOI! cd1
ndbe:.ion. J Am Soc Ncphrol2002;13:621-9.
8. Bischolf-Fcrr:u-1 H:A, P., Willen: WC, Dic:ttich T,
B. R,tim;ltion of optimnl vcrum c.cmcentrllliOtl$ of
D for multiple health outcomes, Am J Clio Nutr
9. Tnogpricha V, Pcuce EN, Chen TC, HoUck 1\IIF. Vitamin D
inrufiicicncy :unoog freeUving young :tdults. Am J M.:d
10. [..,okcr 1\.C, B, Calvo MS, Guntt:t F.:W,
NR. Serum 25-hydro:cyvit:uni.n D of wd 3dults in
two RUbpopul:nioM from NHANES m. Bone 2002;30:
11. Clcmem TL, Ad:uns JS, Heodc.rnon SL, Hollcl: MF. lncrcued skin
pigment the ap:tclty of to vimmin D3. Lancet

12. lVJ!Ir>uok.:l J:.Y, ldt L, Wort:imao J, M:teLnughlin JA, Holick
. &uppral cutaneous vit:Unln D:l synthcsu.J Clin Endoai-
nol Meub 1987;64:1165-8.
13. Wortsm:mJ, LY, Chen TC, Z, Holick MF.
of vitumin D in obesity. Am J Clio Nutr 2000;72:
14, Blwn M, Dolniko\l'llki C, Scyouu1 E, tt 1\.1. Vitnmin D(3) in
Endoc.rinr: 20011;33:90-4.
15. Holick lv!F, Su>lligbt and \it.unin D for bonc-bc:.'llth nnd prrn:acion of
uutoimmune dm::ises, CIUlCcr.l, cwioV':l!CUW cfueuc. Am I Clio
Nutr 2004;80:16785-885.
16. Rostnnd SG, Ultmviolct light may contribute: to wd racial
blood pressure diFcrcnc:cn. Hyp<:rtension '997:30:150-6.
17. S=gg R, R, Ho.ldaWity IM, Lim T, Beuglehole R.
. infi1tctioa is invcmdy wid1 25-hydroX}".oit:unin DJ
11 study. lntJ Epidemio11990;19:559-63.
lB. Poole KE, Lo1eridgc N, B:uker PJ, c:t :U. Reduced viumin Din acute:
stroke. Stroke 2006;37!243-S .
19. Cigolln.i M, hgu11i Mll, Miami V, Gnliotto M, L<lmhunH S, 'fnrWter
C. Serum D3 cooceacr:ttions nnd prenlem:C o(
dhelt:;e nmong type 2 diabetic Clll1:
2006;29:722- 4,
1o:rno:r oac nT<.
Lno et 1955
Vltllmin D Defit:iency and CV Risk
:w. Mcl;&rottl ML, MvntMr P, ED. et nl. Serum 25-
hydroxyviromin D lo:vch and the pri!Vn!ence of peripher.U :mcrbl
from NHANES 2001 to 2004. Anc:dooclcr Thromh
Vase 'Hiol2008;28:1179-85.
21. M:urin$ D, Wolf M, P-111 D. ct Pti:V:Ilcnce of risk
f.lc:rora a.nd the serum levels of 25-hydroxyviuml.n D In the: United
Stutes; dnt:1 &om the Thittl Nutionul He1Uth und Nutrition E.-uuninn-
tion Survey. Arcl\ Intern Mcd 2007;167:1159-65. ,
22. Scragg R., Sow= M, Bc:ll C. S=un 25-hydroxyvitamin D, etbnicity,
blood in che TIW:d Health nnd N11trition
Survey. Am J HypcrtCll5 2007;20:713-9.
13. S=gg R, S.lwcr' M, Bell C. SctUtn 25-hydroxyviwrun D,
cthnicity in the 'Third Nutionul He1tlth 11nd Nutrition &'Ounina-
clon Survey. C:u-e 2004;27:2813-8.
24. Ri:tchy R. V:llldcwu!lc B, Moennun E, ct uL 1,25-Dihyd.roX)"-it&min
D3 humnn e:ywl:inc-induc:cd '-popto-
si.!i via downregul.1tion of the F:IS receptor. Apoprosi.5 2006;11:151-9.
25, Chiu KC, ChuA, Go VT.., MF. D
with inbulin resill'tltncc: llnd bent cell Am J Clin Nutr
26. Pittns AG, Dnwson-Hughe& B, Li T, et al. Vit1min D ruid alcium
ln cc:l:ttinn ro cypc: 2 dio.b<:tc.1 In womc:n. DbhctcR C:trc
27. Hypponeu E, Lnar-<1 E, Reun:tneo. A, ]iltvdin MR, Vlttanen SM.
Int1ke of 0 and ri!lk of type l n bi.rth-(ohort study.
Lmcc:c 2001;'!5S:1500-3.
28. Zipitis cs, Akobeng AI<, Vinunin D 1Uppl=cntation m early
childhood and dsk of type 1 &alx:te&: :1 li)llltc:m:ldc ce\'lew :1nd
mctu-unul)'!'u. Arch ))j, Child 2008;93:517.-7
29. Slmpsaa RU, Hc:nhc:y SH, Nlbbcllnk KA. Clur.u:tcri.:t:ltion ofheut
1111d bloJotl in che vito'lmin .D knocko11c J
Steroid Biochcm Mol Bioi 2007;103:521-4.
30. Ll YC. Vi1:1min D regui:J.tion of che "}'Stem. J Cell
iochcm 2003;88;327-31.
3).. R, Buhri.ng M, Hopfenmullet W, Holick MF, Sb1\tm!1 AM.
Ultr:Iviolct B nnd blood prCRrurc. Lltncet 1998;352:709-10.
32. Sugden JA, Davies jl, With:un MD, N!orris AD, Struthc:n AD.
Vilttmin D c:ndothelilll function in p1ttients '"oitb type: 2
di:tbc:tcJ mclliDJ.q nnd low tif.:lmln D lcvciR. DubctcR Mc:d 2008;25:
33. P, Moaddldc L Cohort study on &co; of p:u:lthyrold
on Jnul.riple outcomtl> in prim:>.')' hypcrpolt;othyroidi.1m, BMJ
34. Soub:uei LP, Chi= TC, ED, c:r :U. Incider.c:e
fitetuno of' curonucy benrt discwsc in elderly p11tic:nts on chronic:
hemodi3lysi,;. T.nt Uml Ncphro1 2006;38:795-SOO.
35. Bjork:mtln MP, Sol'\'!!. AJ, Tuvis RS. sc:rum p:uuthyroid
hormone: prcdic;tl; impaired )'\lrvi"u in 1\ g.:ncnu t1gcd
popubtion. Eur J Endocrinoi2008;1SS:7-19-S3.
36. Ogmd CG, J::cgehrumo MD, Kisturp C, Nichcn SL, Yestcrgitard H.
lncrc:w:d. N-tctminn.l pro-B-typc: peptide
mnrkcn; of inflrurunntion rcbtcd to utherosclcrosili in paticntl; with
prim:l!y hypc.rp:u:tthyroldiam. Clin Endocrinol (Ox) 2005;63:493-S.
37. AnclcMoo P, RydbctJ; E;Willcnheimcr R. hyperpmthyroid-
ism and hClLit rc:view. Eur Hem J 2004;25:1776-87.
38. S11rnnk MJ, Ltm:y AS, Schoohvcnh AC, t!t 111. Kidney diseuse tu 11 risk
flctar for dc:vclopmc:nt of disclll!e: rn:atc:mc:nt from the
Amcric11n H= Associntion Councils on Kidney in Cud.iovuscubr
High Blood PrCM\Il'c Clinic::U C:ll'dio1ogy,
Epidcmiolo;;y Mnd Prevention. Citcullltion 2003;108:2154-69.
39. Wolf M, Shalt A, Gucicnc:-t 0, et :U. Vit:1mln D Jcvdr. wd
mortAlity t\rtiOilg ineitl(:l'lt JllltientJ. Kidney lnt 2007;72:
40. CP, Ahm3d7.3dcli S, JE, K:shnta:-Z:.dch K.
Association of a<:tivutcd vitumin D tn:utment and mortnlity in chronic
k:dncy 1\reh Intern Med 2008:168:397-403.
41. Singh NP, V, 0, Nair M. Elfcct of
of Reeond:uy on c::udiav:lS<:uhr he
modymunics in prc:diuly.;is CKD pnticnts: 11 preliminllt')'
Hemodiallnt 2007;11:417-23.
42. Le': GH, Bennet D, Regidor DL, Knlnnttr-ZRdch K. Imp3ct of
lcidno:J' boac disezc wd ito lllll.'1llgcment on rurvivnl of pKticnts on'
dlnl)'!lis. J Rc:1t Nutr 2007;17!3\l-44.
t lO. d 'TVJ.O.L
1956 lee et 111.
V"rbntin D Dofldoncy lltld CV IUsk
43. S::hldtholfSS, Zittmmum A, C. ll=ha!p HK, Stdlle P,
Komer R. D supplcmc:nutlon improves cytokinc ht
with congestive hem f.tilUK: a double-blitd, t:Uldomiud,
pbccbocontrollcd triaL Am J Clin Nutt .. 1:754-9.
44. T, Morris RC Jr., J\'C.I HE. 1,25-dihydroxy;itnm.in D3
of v:ISculu &mooth muscle L-dls. J Clin I nvut
45. P. Vit:unin D pbyzlology. Frog Mol Biul 2006;?2!4-R.
46. Auticr P, G11ndini !:i. Vitumin D :md tot:U morn.liry:
of r:mdonuua controlled tri.Js. Arch l ntc:rn Mcd
47, Jone' C, W.M.cnbccz T. C:udiov:ucu!M risks of cnlcium
in women. BMJ 2005;336:226-7.
4S. Bolland .Mj
.l:larbc:r P A, Doughty RN, ct V eve no; in healthy
older women nndomi.scd con-
rtollerl cnal. BMJ 200S;336:262-6.
49, &mu HG, Broun J, Kr.u"'r. R, .t Tw11 yc".lt COlllpariaon of
o.nd urboonte c:.fi'eas on c:udiovo.sculnr cu.lciicl!tiun
and hone dcmity. Nephrol Dial ' l'nm"Plunt :!005;20:1 65:1-lll.
SO. Wnn:ul'l K,F., Ahrobt 1\fi., MAlone LL, ct :al.. Active sc:rum. vitamin D
levels uc: in\crsely correlated with coruni\C)'
19?7;?6:j 755-60.
51. JR, IWnaiacr CM, IUUr.vatu DL. c:r Ill. bone mineral
density is com:llllt<i with subdink:d in olrlcr, but not
)'DWlt,OU, Meactn AmcriCin women fl)gl: rhc Sn Anronio F:i:nily
Sturl). lm ::!007;81:430-41.
52. ClctUp H, Mikkd5c:n K, Poul!c:n L, et aL Commonly rccommc:nded
dilly lnt'lke of viwnin V is not sufficient if $llnligbt c.'q'OMC i..;
limited. J .lnt m1 Med 2001T,247:260B.
53. PlomikoffGA, Qltiglcy ]1\lt. Prc:v21ena: of bypovh:uninos.ls Din
with pcni5lc:nt, nonspecific muKUlosla:kllll pnin. Muyo Clin
Ptve 2003;78:1163-70.
JACC Vol. 52, tlo. 24, 2008
54. Bischoff-Fcrt:Lrl HA, :B, WJ..IJ.ert we. Ct :11. Elfecr
of vit'.tmin D 011 fo.IL\: :1. meta. :uulyals. JAMA 2004;291:199?-
SS. Diebt) Supplement Fact Sbtct; Vittunin D. Office of Dieooy Sup-
NIH CUnic::ll Center, Io.-titutes of Hetlth. 2007.
AvnilAble http:/ldlctaty-ilnpplemcnts.lnfo.nlh.gov/f:!.CuhcetJ/
vit:lmiod.asp. July 30, 2()0!l.
SF.. Moore C, Murphy MM, DR, HulickMF. Viumin D in
the United Sutc&, J Am Diet Anoc 2004;104:980-3.
57. Huthc:oek JN, A. Vieth R, Heaney R. for
vittrmin D. Am] Clln Nutr 2007;85:6-18.
58. Dnudun M. Jungc:rA P. Drug-induced renal cnlculli epidemiology,
prcvtntion Knd tn01>1agcmr:m. Drugs 2004:64:245-75.
59. M:uloufj, Nubuai M, Vieth R, cr al. Shorc-rc:rm :md long rerm
of wctlcly highdofic viumin DJ in school children.
J Clln Endoc::rinol Melilb
60. Chcl TC, Chlmch F, Lu Z, ct Foctot3 tl.mr inllucnc:c the cuti\DCOU>
and dlcr.uy su11rt:o.:!l c>f' (). Atcb Biochcm Biophy:i
2007;460:213- 7. .
61. He11t1c:y RP, DavicA KM, Chen TC, Holick MF, .MJ.
Human serum response co extended onl
doai.ng with Am J Clln Nutc 2003;77:204-10.
62. Vieth R. Vitamln D nrpplcmcnt:.tion, 25-hydtoxyviwnin D cunCI:TI-
tnuions, :u1d WCt)'. Am J Clin Nurr 1'J!l9;69:842-5G.
63. Holick Ml?, BiRncu;:7.o RM, Chen TC, ct :U. Vitamin 02 in ;u
cJFtcrivc :u1 viwnin 03 in cireul:lting conccn=cions of
25-hydroxyviu.rnit 1). J Clin Endocriool 2008;93:677-Sl.
64. Arm:u LA, Hollis 'OW, RP. Vltamln 02 j ,; much
than vir:unin D3 in huml\ns. J Clin 'Endocrinol Met\b
Key Wo(d.: viumln D 25- hydro.-t)'vinmin 'D c::Uclwu coron:tty
plcvcntiOr\ " hypcr.:cnsion
. : Ql=

l emma/ ofGerontolog\: MEDICALSC/E/IlC.4)
Cite journal as: J Grromol A Bioi Sci Med Sd
2009. Vol. 64A. No. :\, 5:\9-56 7
The Awhor 2009. Publishetl by O.iford Uni\ersiry Pre.o;s ou belwlj of The Cit'rtllllOiogical Society ofAmerit.'a.
All riglus rtsen-ed. For pamissions. p/e((St e-mail: joumals.permissions@oxjordjoumals.org.
Admm.:t> A<.ce,\S publlt;ation ()II February 16, 2009
doi: I 0.1 093/gerono/glp006
Calcium Plus Vitamin D Supplementation and Mortality in
Postmenopausal Women: The Women's Health Initiative
Calcium-Vitamin D Randomized Controlled Trial
Andrea Z. LaCroix,
Jane Kotchen,2 Gamet Anderson,
Robe1t Brzyski,
Jane A. Cauley,
Steven R. Cummings,5 Margery Gass,6 Karen C. Johnson,
Marcia Ko,8 Joseph Larson,
JoAnn E. Manson,9
Marcia L. Stefanick, IO and Jean Wactawski-Wende
Division of Public Health Sciences, WHl Clinical Coordinating Center, Fred Hutchinson Cancer Research Center, Seattle, Washington.
Division of Epidemiology, Department of Population Health, Health Policy Institute, Medical College of Wisconsin, Milwaukee.
3Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology and Infertility, University of Texas,
Health Science Center, San Antonio.
Department of Epidemiology, University of Pittsburgh, Pennsylvania.
Department of Medicine, Epidemiology and Biostatistics, University of California at San Francisco.
Department of Medicine, University of Cincinnati, Ohio.
Department of Medicine, Health Science Center, University of Tennessee, Memphis.
8Department of Internal Medicine, Women's Health, Mayo Clinic Scottsdale, Arizona.
Department of Epidemiology, Division of Preventive Medicine, Brigham and Women's Hospital, Harvard Medical School,
Boston, Massachusetts.
Stanford Prevention Research Center, School of Medicine, Stanford University, California.
Department of Social and Preventive Medicine, University at Buffalo, New York.
Background. Calcium and vitamin D (CaD) supplementation trials including the Women's Health Initiative (WHI)
trial of CaD have shown nonsignificant reductions in total monality. Tlus report examines intervention effects Oil total and
cause-specific mortality by age and adherence.
Methotls. The WHJ CaD trial w a ~ a randomized, double-blind, placebo-controlled trial that enrolled 36,282 post-
menopausal women aged 51- 82 years from 40 U.S. clinical centers. Women were assigned to 1,000 mg of elemental
calcium carbonate and 400 IU of vitamin D3 daily or placebo with average follow-up of7.0 years.
Results. The hazard ratio (HR} for total mortalily was 0.91 (95% confidence interval [CI], 0.83-1.01} with 744
deaths in women randomized to CaD versus 807 deaths in the placebo group. HRs were in the direction of reduced risk
but nonsignificant for stroke and cancer mortality, but ncar unity for coronary heart disease and other causes of death.
HRs for total mortality were 0.89 in the 29,942 women younger than 70 years (95% CI, 0.79- 1.01) and 0.95 in the
6,340 women aged 70 and older (95% CI, 0.80-1 . 12; p value for age i.nteraclion= .10). No statistically significant in-
teractions were observed for any baseline characteri stics. Treatment effects did not vary significantly by season.
Couclusions. In the WHI CaD trial. supplementation did not have a statistically significant effect on mortality rates but the
findings support the possibility that these supplements may reduce lllOltality rates in postmenopausal women. These data can
neither support nor refute recommendation.s for higher dose vitamin D supplemenlation to reduce cancer or total mortality.
Key Words: Calcium-Vitamin D-Mortality-Cause-specific mortality-Women' s Health Initiative.
ESPITE predictions that some types of micronutri-
ent supplementation might prevent cancer and car-
diovascular disease (CVD) and reduce mortality, the
results of large intervention trials of chemoprevention
have been largely disappointing to date. Several random-
ized controlled trials have shown nonsignificant reduc-
tions in total mortality among women assigned to vitamin
, calcium carbonate alone, or the combination of cal-
cium plus vitamin D (1-4). Among the 36,282 women
who participated in the Women's Health Initiative (WHI)
calcium plus vitamin D (CaD) supplementation trial, the
intention-to-treat hazard ratio (HR) for total mortality
was 0.91 (95% confidence interval [CI], 0.83-1.0 1) (1). A
recent meta-analysis, in which WHI women contributed
nearly two thirds of the data, reported a significant reduc-
tion in mortality among nine trials (4) based only on pre-
viously published data.
We report here the first thorough evaluation of the possi-
ble effects of CaD supplementation on total and cause-
specific mortality in the WHI CaD trial. This repott also
assesses the impact of adherence to CaD supplementation
on treatment effect estimates and determines whether HRs
for total mortality differed according to levels of selected
baseline characteristics, including age.
Design Overview
The WHI CaD trial is a randomized, double-blind, place-
bo-controlled trial designed to determine whether supple-
mentation with calcium (1,000 mg/d) plus vitamin D (400
IU/d) would prevent hip fracture (the primary outcome), other
fractures, or colorectal cancer (designated secondary out-
come). These findings have been previously published (1,5).
Setting and Participants
Between 1993 and 1998, postmenopausal women aged
~ 7 9 years were enrolled in the WHI randomized clinical
trials designed to assess the 1isks and benefits of hormone
therapy (HT) and dietary modification (DM) (6). One year
later, participants in either the HT or the DM trial were in-
vited to enroll in the CaD trial. Exclusion criteria for the CaD
trial included a predicted survival of less than 3 years; a his-
tory of renal calculi or hypercalcemia; or current use of oral
corticosteroids or calcitriol, or daily use of 600 IU or more of
supplemental vitamin D (7). Personal use of calcium supple-
ments up to 1 ,000 mg/d or vitamin D up to 600 IU/d wa<;
permitted. The protocol and consent fonns were approved by
the institutional review board at each pa1ticipating institution.
All women provided written informed consent.
Randomization and Intervention
Details regarding execution of the trial have been de-
scribed previously (I ,6). Briefly, using a permuted-block
algorithm stratified by clinical center and age, 18,176 women
were randomly assigned to receive one tablet of 500 mg of
elemental calcium as calcium carbonate combined with 200
JU of vitamin D3 (GiaxoSmithKline. Pittsburgh, PA) twice
daily, totaling I ,000 mg of elemental calcium and 400 IU of
vitamin D. A total of 18,106 women were randomized to
receive an identical-appearing placebo tablet twice daily.
Blinding of the study was achieved by bottle labeling.
Outcomes and Follow-Up
Participants were contacted semiannually thereafter to
ascertain self-reported medical history updates. When they
could not be contacted, information on vital status was sought
from previously identified proxy informants, National Death
Index searches, and obituary notices. Regardless of adher-
ence, participants were followed up until they died, were lost
to follow-up, requested no further contact, or until the study
ended. Ninety-seven percent of participants were followed to
study completion. Causes of death were determined based on
available medical records, autopsy reports, and the death cer-
tificate in a blinded fashion by local and central physician
adjudicators. For 352 deaths (22.7%), the death certificate
was the only available source of information for determining
cause of death. Ninety-six percent of reported deaths were
centrally adjudicated, 2.5% were locally adjudicated, and
1.5% could not be adjudicated.
Study pills were discontinued if women reported kidney
stones, hypercalcemia, dialysis, or the use of calcitriol or
daily supplements of more than 1,000 IU of vitamin D dur-
ing semiannual contacts. During the study, adherence was
assessed by weighing returned pill bottles.
An independent data and safety monitoring board reviewed
the trial data semiannually. Closeout visits occmTed as
planned between October 1, 2004, and March 31,2005, with
final outcomes assessed before the treatment assignment was
revealed. WHI investigators and National Institutes of Health
sponsors all contributed to the study design and execution.
Nested Case-Control Study of Vitamin D Levels
Baseline serum measurements of 25-hydroxyvitamin D
were available for 323 women who died and I ,962 living
controls in the CaD trial from previous nested case-control
studies (I ,5). Levels of 25-hydroxyvitamin D were mea-
sured using the DiaSorin Liaison chemiluminescent immu-
noassay system at Diasorin headquarters (Stillwater, MN)
as previously described (I ,5).
Statistical Analysis
Primary analyses used time-to-event methods, according
to the intention-to-treat principle. Mortality rates were com-
pared in the two groups with the use of HRs (with 95 per-
cent Cis) and stratified logrank tests from Cox proportional
hazards models (8), stratified according to baseline age and
treatment assignment in the HT and DM uials.
In addition, HRs were calculated by categories of causes
of death (cardiovascular, cancer, other, and for CVD, coro-
nary heart disease (CHD), and cerebrovascular cause of
death separately). Because mortality rates tise markedly
with age, treatment effects were examined separately for
women younger and older than 70 years, the cut point pre-
specified in the protocol for defining older women. Analy-
ses according to adherence to study medication were also
conducted, with participants censored 6 months after first
detection of becoming less than 80% adherent. Inverse
probability weighting methods were used to estimate full-
adherence HRs adjusting for covariates found to be associ-
ated with nonadherence (1,9).
Potential differential effects across categories of important
risk factors for mortality were also tested individually with
the use of a likelihood ratio test for interaction between the
risk factor and the treatment assignment after including both
as main effects. Twenty-three subgroup comparisons were
tested; thus, about one test would be expected to be statisti-
cally significant at the .05 level by chance alone. All reported
p values are two sided and were not adjusted for multiplicity.
Odds ratios for CaD supplementation were estimated using
logistic regression across tertiles of serum 25-hydroxyvitarnin
D. The model stratified by randomization assignment in

AU ages !:::! Ages ;::70

HR =0.95
(95% Cl. 0.83-1.0 I) (95% Cl, 0.80 1.1 2)


Time (years) 0 Time (years)
0 2 4 5 6 7 8 0 2 3 4 5 6 7 8
CaD: CaD:
Event;;. 33 58 89 96 118 131 103 19 33 E\'COL."i II 23 32 37 44 so 3S 29 16
Nat risk 1817618072 17975 17839 17676 17331 14768 9213 4+14 Na1 3173 3155 3129 3087 3038 2940 2327 1262 570
Placebo: Placebo:
Evcms 48 74 98 114 126 114 74 25 18 36 41 31 39 46 42 24 10
N 18106 17993 17872 17720 17541 17174 14617 9105 4385 Na1risk 3167 3145 3103 3056 3012 2921 2317 1262 565

(95%CI, 0.79-1.01)
N Ql ri&k
N ul ris.k
22 35
15003 14917
30 38
14939 I4R4R
Figure I. Cumulative mortality Kaplan-Meier curves by age group.
3 4 5 6 7 8
59 74 81 68 50 17
14752 14638 1244 1 7951 3874
83 87 81 72 50 15
141\1\4 14253 12100 7R4l 3R20
Note: CaD = calcium and vitamin D; Cl = confidence interval; HR =hazard ratio.
another WHI trial and age group, and adjusted for age, ethnic-
ity, and latitude of clinical center. An interaction p value was
calculated for treatment by continuous serum level.
Baseline characteristics were well balanced between the
treatment groups (Table 1). Twenty-nine percent of study
participants were taking at least 500 mg/d of supplemental
calcium on their own. Low calcium intakes ( <800 mg/d)
were noted in approximately one third of study participants.
The mean (SD) duration of follow-up was 7.0 (1.4) years.
At time of closeout, 1,551 deaths had occurred, 744 in the
CaD intervention group and 807 in the placebo group. The
final HR for total mortality was 0.91 (95% Cl, 0.83- 1.01)
(Table 2). CaD HRs were in the direction of reduced risk but
nonsignificant for stroke and cancer mortality, whereas HRs
were close to unity for CHD and other causes of death.
Intention-to-treat HRs by age at baseline ( <70 vs ?:.70
years) suggested lower HRs among younger women for to-
tal, stroke, and other causes of death. The HRs for total mor-
tality were 0.89 in women younger than 70 (95% CI,
0.79- 1.01) compared with 0.95 in women of ages 70 and
Table I. Characteristics of WHl CaD Participants at WHI Screening by Randomization Arm
CaD (N= 18,176) Placebo (N = 18, I 06)
N Mean (SD) or % N Mean (SD) or % p Value
Age at screening, y
Mean (SD) 18,176 62.4 (7.0) 18,106 62.4 (6.9) .97
50-54 2.594 14.3 2.563 14.2
55-59 4,134 22.7 4, 131 22.8
~ 9 8.275 45.5 8,245 45.5
70-79 3,173 17.5 3, 167 17.5
Race/ethnicity .45
White 15.047 82.8 15.106 83.4
Black 1.682 9.3 1,635 9.0
Hispanic 789 4.3 7 18 4.0
American Indian/Native American 77 0.4 72 0.4
Asian/Pacific Islander 369 2.0 353 1.9
Unknown 212 1.2 222 1.2
Education at screening .98
High school or less 4,286 23.6 4,289 23.7
Some school after high school 5,30 1 292 5,248 29.0
College graduate 1,862 10.2 1,856 10.3
Graduate school 4,693 25.8 4,687 25.9
Hormone use at annual visit I .23
Never used 5,814 32.0 5.690 3!.4
Past user 3,004 16.5 2,932 16.2
Current user 9,358 51.5 9,484 52.4
Dai ly supplemental calcium use
Mean (SD) 18,176 322.6 (479.2) 18,106 326.2 (482.4) .48
None. mg 8.009 44.1 7,871 43.5
<500 4.975 27.4 4,922 27.2
>500 5.192 286 5,3 13 29.3
Dai ly iota! calci um (supplements+ diet). mg
Mean (SD) 17,821 1,148.4 (654.2) 17,753 1.154.3 (658.0) .40
<800 6.104 33.6 6.003 33.2
800 to <l ,200 4,715 25.9 4.655 25.7
2: 1.200 7.002 38.5 7.095 39.2
Daily vitami n D (supplements+ diet), IU
Mean (SD) 17.821 365.3 (265.5) 17.753 367.9 (266.0) .36
<200 6,827 37.6 6,671 36.8
200 to <400 3.379 18.6 3.423 18.9
400 to<600 4.188 23.0 4.295 23.7
<:600 3,427 18.9 3,364 18.6
History of hypertension .88
Never hypertensive 6,944 38.2 6,925 38.2
Untreated hypertensive 6.686 36.8 6,692 37.0
Treated hypertensive 4,546 25.0 4.489 24.8
Systolic blood pressure 18.176 127.4 (l7 .I) 18.106 127.5 (17.2) .48
Diastolic blood pressure 18,175 75.9 (9.1) 18,101 75.9 (9.1)) 56
Smoking .50
Never smoked 9,428 52.1 9,325 51.3
Past smoker 7,133 39.4 7,255 39.9
Current smoker 1,356 75 1,405 7.7
Alcohol use 88
Nondrinker 1,863 10.2 1,891 10.4
Past drinker 3,192 17.6 3,209 17.7
<I drink per mo 2.529 13.9 2,520 13.9
<I drink per wk 3.863 21.3 3.758 20.8
I to <7 drinks per wk 4.683 25.8 4,706 260
<:7 drinks per wk 1.910 10.5 .1.900 10.5
Physical activity
Mean (SD) METs/wk 16.546 10.7 ( 12.7) 16.448 10.6 (12.4) .60
0-3.00 5,517 30.4 5,478 30.3
>3.00 to <11.75 5.463 30.1 5.477 30.2
:0:11. 75 5.566 30.6 5,493 30.3
Table I. (Continued)
CaD (N= 18,176) Placebo (N= 18, 106)
N Mean (SD) or% N Mean (SD) or % pYalue
Body mass index. kg!Jn2
Meao (SD)
No. of CVD risk factors
History ofCVD
History of treated diabetes
OM trial arm
Not randomized
HTtrial arm
Not randomized
No. of chronic conditions at screening
Depressive symptoms score
Physical function score
1, 173
29.1 (5.9) 18.011
26.1 4,833
35.6 6.4&3
37.8 6,695
30.5 5.579
67.9 12. 195
1.6 332
6.5 1.221
4.9 &75
43.1 7,738
26.2 4.878
30.7 5.490
22.1 3,957
22.2 4.078
55.7 10.071
1.2 (0.9) 18.106
2.30 (2.55) 18,057
81.5 (20.1) 18,061
29.0 (5.9)
1.2 (0.9)
2.33 (2.59)
81.4 (20.3)
Note: CaD = calcium and vitamin D; CVD = cardiovascular disease; DM = dietary modification; HT = hormone therapy; MET = metabolic equivalent;
WHI = Women's Health Initiative.
older (95% CI, 0.80-1.12; p value for age interaction= .1 0).
HRs for cancer death were similar across age categories.
At trial closure, 76% of women enrolled were still taking
study medications and 59% were taking at least 80% of
study pills. When participants were censored 6 months after
first detection of adherence less than 80%, HRs were 0.87
for total mortality (95% Cl, 0.73-1.04), 0.85 for cancer
mortality (95% CI, 0.66-1.09), and 0.57 (95% Cl, 0.30-
L09) for stroke mortality. HRs suggested greater benefit for
adherent women younger than 70 years, as compared with
older adherent women (data not shown). Tests for age inter-
action were statistically significant for total mortality
(p = .02) but not for the cause-specific categories. Estimated
full-adherence HRs based on the inverse probability weight-
ing method were similar to those for adherent women and
all Cis included 1.0.
Intention-to-treat HRs did not vary significantly for any
of the 23 baseline risk factors examined for mortality. In
addition to the subgroups shown in Table 3, no significant
differences were observed by subgroups of education, al-
cohol intake, history of diabetes, depressive symptoms,
physical function score, and disability in activities of
daily living, or randomization in the dietary modification
trial. In models adjusting for season as a time-dependent
variable, risk of mortality was increased during winter
months, with a HR of 1.14 (95% CI, 0.99-1.32). Treat-
ment effects did not vary significantly by season at time
of death (p value for interaction= .20), although HRs were
observed to be reduced iJJ winter (HR =0.81; 95% CI,
0.66-0.98), summer (HR=0.87; 95% Cl, 0.71-1.06), and
autumn (HR =0.93; 95% CI, 0.77-1.13), but not in spring
(HR= 1.10; 95% Cl, 0.89-1.36).
In the nested case-control study, CaD intervention effects
did not vary significantly by baseline levels of serum 25-
hydroxyvitamin D (p= .66), although the odds ratio was in
the direction of benefit for women in the lowest tertile
(Table 4). Compared with women in the highest tertile of
serum 25-hydroxyvitamin D, there was a significantly in-
creased risk for death for women in the middle and low ter-
tiles. The odds ratio for serum 25-hydroxyvitamin D level
analyzed as a continuous variable was 0.80 (95% CI, 0.67-0.95)
for a difference of 29.9 nmol/L (the interquartile range).
In this large randomized trial of women aged 50-79, CaD
supplementation for an average of 7 years was associated
with a nonsignificant reduction in the risk of death (9% with
95% Cl, 17% reduction to a 1% increase in risk). The evi-
dence was most consistent for a decrease in cancer mortality.
Among the 23 interactions tested, none was statistically
significant. Age, as the strongest predictor of mortality in
the general population, was of particular interest. Among
the women younger than 70 years, CaD supplementation
appeared to reduce risks of total, CVD, and cancer death.ln
older women, only cancer mortality appeared to be reduced.
The interaction between CaD supplementation and age was
not statistically significant for total mortality overall but
was so among adherent women (p = .02).
Because the majority of statistical tests repotted herein
are not significant at the .05 level, a viable explanation for
Table 2. Effect of Calcium With Vitamin D Supplementation on Total, CVD, Cancer, and Other Mortality, According to Randomly Assigned
CaD Placebo
Cases (am1ualized %) Cases (annualized%) HR(95% Cl)
Follow-up time, y (mean SDl 7.01.4 7.01.4
Total monality 744 (0.58) 807 (0.63) 0.91 (0.83- 1.01 )
CVDdeath 226 (0.18) 244 (0.1 9) 0.92 (0.77- 1.10)
CHD death 130 (0.1()) 128 (0.1 0) 1.01 (0.79-1.29)
Cerebrovascular death 54 (0.04) 60 (0.05) 0.89 (0.62-1.29)
Cancer death 344 (0.27) 382 (0.30) 0.89 (0.77-1.03)
Other/unknown death 174 (0.14) 181 (0.14) 0.95 (0.77-1. 17)
Participants younger than 70 years
Follow-up time, y (mean SD) 7.1 1.4 7.1 1.4
Total mortality 466 (0.44) 517 (0.49) 0.89 (0.79-t.OJ)
CVDdcath 115 (0.11) 135 (0.13) 0.85 (0.66-1.08)
CHDdeath 70 (0.07) 70 (0.07) 0.99 (0.71- 1.38)
death 21 (0.02) 33 (0.03) 0.62 (0.36-1 08)
Cancer death 245 (0.23) 268 (0.25) 0.9 1 (0.76-1.08)
Other/unknown death 106 (0.10) I 14 (0. 11) 0.92 (0.70-1.20)
Participants 70 or older
Follow-up time, y (mean SO) 6.7 1.4 6.7 1.5
'lbtal mortality 278 ( 1.30) 290 (137) 0.95 (0.80-1.12)
CVDdeath I l l (0.52) 109 (0.51) 1.01 (0.78-1.32)
CHD death 60 (0.28) 58 (0.27) 1.02 (0.71- 1.47)
Cerebrovascular death 33 (0.15) 27 (0.13) 1.20 (0.72- 2.01)
Cancer death 99 (0.46) 114 (0.54) 0.86 (0.65-1.12)
Other/unknown death 68 (0.32) 67 (0.32) 1.01 (0.72-1.42)
Note: CaD= calcium and vitami n D: CHD =coronary heart disease; Cl =confidence interval: CVD =cardiovascul ar disease; HR = hazard ratio.
these results is chance alone. Among the younger women,
there was a difference favoring intervention for reduced
mortality within the first year, which is difficult to attribute
to intervention. A trial designed to confirm this association
would require more than 90,000 women aged 50 and older
followed for 5 or more years to detect a statistically signifi-
cant reduction in mortality rates of 9%-10%.
Nonetheless, previous randomized controlled trials, most
of which enrolled only older women, have also shown non-
significant reductions in total mortality. A recent meta-
analysis of nine trials testing vitamin D supplementation
reported an overall HR of 0.93 (95% CI, 0.87-0.99); how-
ever, this point estimate largely reflects the contribution of
WHI women who contributed a majority of the person-time.
Six of the remaining eight smaller uials showed nonsignifi-
cant reductions in total mortality ( 4 ). Cause-specific mortal-
ity was examined in only one previous trial and showed
nonsignificant reductions in cardiovascular, total cancer,
and colorectal cancer mortality with very wide Cis (3).
If the results did not occur by chance, multiple pathways
could be involved including any elements of the supple-
ments tested: calcium, vitamin D
, or carbonate, or all. Cal-
cium supplementation has been shown in small experimental
studies to lower blood pressure ( 1 0) and reduce cholesterol
levels (11-13). Observational epidemiological studies have
shown an association of calcium intake with reduced stroke
incidence and mortality ( 14-16) and inconsistent associa-
tions with reduced risk of coronary disease (16, 17). Lower
bone density has been associated with increased mortality
risks (18,19). In the WH1 CaD trial, the intervention group
had statistically significant but small reductions in low-
density lipoprotein, weight, and waist circumference, whereas
blood pressure change was slightly increased compared
with placebo (20). Although the intervention improved hip-
bone density by about 1 %, it produced no significant differ-
ences in CHD or stroke incidence.
Vitamin D insufficiency has now been linked to a broad
spectrum of human diseases from cancer to cardiovascular
to autoimmune conditions (21). A recent ecological study
showed consistently elevated death rates from cancer sites
among U.S. women with less solar ultraviolet-B exposure
(22). Vitamin D receptors are present in many cell types and
1,25-dihydroxyvitami.n D bas been shown to favor cell dif-
ferentiation over proliferation and to inhibit potential for me-
tastasis and invasiveness (23). Serum vitamin D levels were
inversely associated with colon cancer risk in the Nurses'
Health Study (24), in the WHI CaD trial (6), and with colon
cancer mortality in National Health and Nutrition Examina-
tion Survey (25). Supplementation with 1,100 IU of vitamin
reduced the risk of cancer in a recent small trial of healthy
postmenopausal women (26). Two recent prospective studies
link levels of serum 25-hydroxyvitarnin D below 44.4 nmol/L
(27) or 25.2 nmol/L (28) to increased risk of total mortality.
The emerging optimism over vitamin D supplementation
for the prevention of cancer (29) is occurring in the absence
of confirmation from large prevention trials that can deter-
mine the benefits and any currently unknown risks. Proponents
argue that the necessary levels of vitamin D supplementation
Table3. HRs Re lating CaD to Ri sk of Total Mortality St ratified by Selected Baseline Characteristics
CaD (N= 18.176) Placebo(N= 18,106)
No. of Dea1hs % No. of Deaths % HR (CI)' p Value for Interaction
Ethnicity .30
White 607 0.57 679 0.64 0.89 (0.80-0.99)
Black 79 0.68 89 0.78 0.91 (0.67-1.23)
Hispanic 23 0.42 11 0.22 2.28 ( 1.07-4.87)
American Indian 0.93 4 0.79 0.84 (0.16-4.48)
Asian/Pacific Islander 18 0.73 12 0.51 1.60 (0.75- 3.43)
Otber/Unk.oown 12 0.83 12 0.81 0.90 (0.45- 1.80)
HT use at annual visit I . 14
Never used 280 069 275 0.69 100 (0.84- 1.18)
Past user 118 0.56 155 0.75 0.74 (0.58-0.94)
Currem user 346 0.52 377 0.57 0.93 (0.80-1.07)
Calcium supplementation. mg .64
None 366 0.64 369 0.66 0.97 (0.84-1.12)
<500 199 0.57 226 0.66 0.86 (0.71-1.04)
::::soo 179 0.50 212 0.58 0.8710.71-1.06)
Baseline total calcium (supplements) .20
<800 282 0.65 313 0.74 0.88 (0.75- 1.04)
800 to <1,200 186 0.56 204 0.63 0.88 (0.72- 1.08)
2: 1,200 259 0.53 269 0.54 0.98 (0.83- 1.17)
Baseline total vitamin D (supplemems) .59
<200 275 0.56 303 0.64 0.88 (0.74-!.03)
200 to <400 144 0.60 150 0.62 0.98 (0.78-1.23)
400 to <600 170 0.58 189 0.63 0.92 (0.75-1.14)
::-600 138 0.58 144 0.62 0.93 (0.73-1.17)
Latitude of clit1ical center .70
Southern: <35 north 232 0.60 255 0.67 0.8910.74-1.06)
Middle: 35-40 north 209 060 238 0.69 0.86 (0.71- 1.03)
N())1hern: >40 north 303 0.55 314 0.57 0.96 (0.82- 1.13)
Blood prc.<sure .39
Treated 237 0.77 278 0.92 0.84 (0.70- 1.00)
;::: 140/90 305 0.64 307 0.65 0.98 (0.84-1.15)
Normal 202 0.40 222 0.45 0.90 (0.75-!.09)
Smoking .20
Never smoked 326 0.49 322 0.48 1.0 I (0.86-1.17)
Past smoker 300 0.59 355 0.7 1 0.83 (0.71-0.97)
CmTent smoker 108 1.11 120 1.28 0.87 (0.67-1.13)
Pbysical activity IMETs/wk) .06
53.00 237 063 278 0.75 0.85 (0.71- 1.01)
>3.00 to <11.75 2 18 058 249 0.66 088 (0.73- 1.06)
2: 11.75 188 0.49 179 0.47 1.03 (0.84- 1.26)
No. of CVD risk factors .59
None 130 0.32 153 0.38 0.85 (0.67-1.08)
1-2 596 0.69 616 0.73 0.95 (0.85-1.07)
;:::3 18 0.89 38 1.70 0.49 (0.27-0.87)
Body mass index. kgtm2 .49
<25 197 0.58 216 0.63 0.93 (0.76-1.12)
25 to <30 232 0.51 261 0.57 0.89 (0.74- 1.06)
::::30 3 15 066 328 0.70 0.93 (0.80- 1.09)
History of CVD .90
No 645 0.54 694 0.58 0.92 (0.82- 1.02)
Yes 99 1.25 113 1.38 0.91 (0.70- 1.20)
Number of chronic conditions* .67
None 95 0.32 107 0.36 0.86 (0.65-1.14)
300 0.52 298 0.52 1.00 (0.86-1.18)
2 244 0.72 283 0.84 0.86 (0.72-1.02)
3 or more 105 1.49 119 1.73 0.87 (0.66-1.13)
Self-reported health status . 13
Excellent 84 0.39 79 0.36 1.14 (0.83- 1.55)
Very good 296 0.52 305 0.54 0.93 (0.79- 1.09)
Good 273 067 304 0.75 0 89 (0.76- 1.05)
Fair 79 0.99 106 1.28 0 79 (0.59- 1.06)
Poor 12 231 13 3.07 051 (0. 18- 1.43)
Notes: ~ t D =calcium and vitamin D: CJ = coniidenc.: interval; CYD =cardiovascular disease; HR = hazard ratio; HT = h(nmone therapy; MET= metabolic equivalent.
HRs and 95% Cis are derived from Cox proportional hazards models suatified by age and randomization in the WH! dietary modification or HT trial, or both.
t 1> values are for the interaction between the subgroup variable and treatment assignment.
t Number of chronic conditions includes self-report of coronary heart disease. congestive heart failure. treated diabetes. hypertension (treated or untreated). hip
fracture. emphysema. cancer, or artluitis. HT at annual visit J includes randomized treatment assignment in the WHI HT trial aod personal use of HT.
Table 4. ORs* for Total Mortality According to Tertile of Serum 25-Hydroxyvitamin D Level and CaD Study Group, as Determined in a Nested
Case-Control Study
Main Effect ORI (95% Cl) CaD. Cases/Controls Placebo. Cases/Controls OR(95%CD p Value for Interaction
Tertile of serum 25-hydroxy 65
vitamin D, nmoi/L
!::52.5 1.00 53/404 50/425 1.04 (0.69-1.59)
35.4-52.4 1.42 (1.06-1.91) 57/301 59/296 0.96 (0.64-1.45)
<35.4 1.46 ( 1.07- 1.99) 47/270 571266 0.79 (0.51 - 1.23)
Notes: CaD= calcium and vitamin 0; CI = c.onfidence interval; OR =odds ratios.
* ORs were derived from logi stic regressioo 01odels stratified by age group and randomization assignment io the hormone therapy or dietary ruodificarioo trial, or
both, and further adjusted for age. ethniciry (white, black, Hispanic, other), latitude of clinical center, and season of blood draw.
t Serum 25-hydroxyvitamin D main effects models are also adjusted for CaD treatment assignment.
to exert anticancer effects might be 1,500-2,000 IU/d based
on selective findings from observational studies (30,31).
The optimum dose of vitamin D for reducing cancer or mor-
tality is unknown and, if it exists, can only be deteiTnined
from future large randomized trials.
The trends toward reductions in risks of both cardiovascular
and cancer mortality in this trial, if real, suggest mechanisms
that may have broad physiological effects. Vitamin D supple-
mentation has been shown experimentally to increase levels of
the anti-inflammatory cytokine interleukin-10, while prevent-
ing increases in tumor necrosis factor alpha (32). Calcium and
vitamin D supplementation both lower parathyroid h01mone
levels (PTH), and PTH levels have been associated with an
increased risk of mortality in a vitamin D-<leficient cohort of
the very old age (33). Chronic metabolic acidosis increases
with age and declining kidney function even when plasma
measures of acid-base balance are normal. Administration of
bicarbonate has been shown to reverse acidosis (34,35). Ben-
eficial effects of CaD supplementation on cytokine profiles,
PTH, or acidosis might be expected to produce greater bene-
fits in older women, whereas the opposite trend was observed
in this trial. However, it is also possible that mo11ality could be
delayed by improvements in these parameters and eventually
benefits diminish with disease progression in later life.
Strengths of this trial include its large size, diversity, long
duration, excellent retention, and reasonably good adher-
ence for an average of 7 years of follow-up. This study is
limited by Jack of statistical power for detecting interven-
tion effects on mortality, as it was not designed for this pur-
pose. We did not collect postintervention serwn specimens
for exploring potential intermediate effects of the interven-
tion that might have influenced mortality. We cannot distin-
guish between effects of calcium, vitamin D, or carbonate
because the intervention combined these ingredients. We do
not know whether higher supplement doses of vitamin D
would have produced different results.
In conclusion, the WHI CaD tiial supports the hypothesis
that CaD supplementation provides a modest reduction in rates
of total and categories of cause-specific mortality, but the re-
sults are too imprecise to be definitive. Moreover, these data
can neither support nor refute recommendations for higher
dose vitamin D supplementation to reduce cancer or total
mortality. This hypothesis merits further consideration by test-
ing altemative underlying mechanisms and possibly by testing
higher dose supplements in even larger randomized trials.
Supplementary material can be found at: http://biomed.
gerontology j ournals. org/.
The autJJOrS thank the WHJ investigators, staff. and study participants for
their outstanding dedication and commitment. A list of key investigators in-
volved in this research follows: A full listing of WHI investigators can be
found at the following Web site: bttp://www.whi.org. A.Z.L. affinns that ev-
etyone who significantly contributed to this work has been listed in the Ac-
The WHI investigaton; and National Institutes of Health
sponsors all contributed to the design and execution of the study. This work
was supported by the National Heart, Lung, and Blood Institute of the U.S.
Department of Health and Human Services. The active study drug and pla-
cebo were suppl ied by GlaxoSmithKii ne Consumer Healthcare (Pittsburgh).
Clinical Trials Registration: ClinicaiTrials.gov identifier: NCf00000611
Address correspondence to Andrea z. LaCroix, PhD, WHJ Clinical Co-
ordinating Center, Fred Hutchinson Cancer Research Center, J 100 Fairview
Avenue North. M3-A410, P.O. Box 19024. Seattle, WA 9RI09-1024.
Email : alacroix@wbi.org
L Jackson RD. LaCroix AZ, Gass .M, eta! for tbe Women's Health Ini-
tiative Investigators. Calcium plus vitamin D supplementation and
the risk of fractures. N Eng/ J Med. 2006;354:669-683.
2. Prince RL, Devine A. Dhaliwal SS, Dick lM. Effects of calcimn
supplementation on clinical fracture and bone structure: results of a
5-year, double-blind, placebo-controlled ttial in elderly women. Arch
Intern Med. 2006;166:869-875.
3. Trivedi DP, Doll R. Khaw KT. Effect of four monthly oral vitamin D3
{cholecalciferol) supplementation on fractures and mortality in men
and women living in the community: randomized double blind con-
trolled trial. BMJ. 2003;326:1 - 6.
4. Autier P. Gandini S. Vitamin D supplementation and total mottalit.y:
a meta-analysis of randomized controlled trials. Arch Intern Med.
2007;167: 1730-1737.
5. Wactawski-Wende J. Kotchen JM. Anderson GL et al. For the Women's
Health Initiative Investigators. Calcium plus vitamin D supplementation
and the risk of colorectal cancer. N Eng/ J Med. 2006;354:684-696.
6. The Women's Health Initiative Study Group. Design of the Women's
Health Initiative clinical trial and observational study. Control Clin
Trials. 1998;19:61-109.
7. Jackson RD, LaCroix AZ, Cauley JA, et al. The Women' s
Health Initiative Calcium-Vitamin D Trial: overview and baseline
characteristics of participants. Ann Epidemio/, (suppl9) 2003; 13:
8. Cox DR. Regression analysis and life tables. J R Stat Soc. 1972;
34: 187-220.
9. Robins JM, Finkelstein DM. Correcting for noncompliance and depen-
dent censoring in an AIDS clinkal trial with inverse probability of cen-
soring weighted (IPCW) log-rank tests. Biometrics. 2000; 56:779-781.
10. Appel LJ, Moore TJ, Obarzanek E, eta!. A clinical trial of the effects
of dietary pmterns on blood pressure. DASH Collaborative Research
Group. N Eng/ J Med. 1997;336: 11 17- 1124.
11. Groot PHE, Grose WFA, Dijkbuis-Stoffelsma R, Fernandes J,
Ambagtshcer JJ. The effect of oral calcium carbonate administration
on serum lipoproteins of children with famil ial hypercholesterolae-
mia (type II-A). Eur J Pediatr. 1980:135:81-84.
12. Denke MA, Fox MM, Schulte MC. Short-term dietary calcium forti-
fication increases fecal saturated fat content and reduces serum lipids
in men. J Nutr. 1993;123: 1047-1053.
13. Bell L, Halstenson CE, Halstcnson CJ, Macrcs M, Keane WF.
Cholesterol-lowering effects of calcium carbonate in patients with
mild to moderate hypercholesterolemia. Arch Imem Med. 1992; 152:
14. Abboll RD. Curb JD, Rodrigues BL. Sharp DS, Burch lie! CM, Yano K.
Effect of dietary calcium and milk consumption on risk of thromboem-
bolic stroke in older middle-aged men. Stroke. 1996;27:813-818.
15. lso H, Stampfer MJ, Manson JE, et al. Prospective study of calcium,
potassium and magnesium intake and risk of stroke in women. S1roke.
1999;30: 1772-1779.
16. Umesawa M, lso H, Date C, et al. For the JACC Study Group.
Dietary intake of calcium in relation to mortality from cardiovascular
disease: the JACC Study. Stroke. 2006;37:20-26.
17. Bostick RM, Kushi LH, Wu Y, Meyer KA, Sellers TA, Folsom AR.
Relation of calcium. Vitamin D. and dairy food intake to ischemic
heart disease mortality among postmenopausal women. Am J Epide-
miol. 1999;149: 151-161.
18. Yonder Recke P, Hansen MA, Hassager C. The association between
low bone mass at the menopause and cardiovascular mortality. Am .T
Med. 1999:106:273- 278.
19. Browner WS, Seeley DG. Yogt TM, Cummings SR. Non-trauma
mo1tality in elderly women with low bone mineral density. Study of
Osteoporotic Fractures Research Group. Lancet. 1991 ;338:355-358.
20. Hsia J, Heiss G. Ren H, et al. For the Women's Health Initiative In-
vestigators. Calcium/vitamin D supplementation and cardiovascular
events. Circulation. 2007;1 15:846- 854.
21. Grant WB. Epidemiology of disease risks in relation to vitamin D
insufficiency. Prog Biophys Mol Bioi. 2006;92:65-79.
22. Boscoe FP, Schymura MJ. Solar ultraviolet-B exposure and cancer
incidence and mortality in the United States. 1993-2002 BMC
Cancer. 2006;6:264-272.
23. GiovarulUcci E. The epidemiology of vitamin D and cancer incidence
and mortality: a review (Uni ted States). Cancer Causes Control.
24. Feskanich D, Ma J, Fuchs CS, et al. Plasma vitamin D metabolites
and risk of colorcctal cancer in women. Cancer Epidemiol Biomark-
ers Prev. 2004: 13:1502-1508.
25. Freedman DM. Looker AC, Chang SC, Graubard BI. Prospective
study of serum Vitamin D and cancer mortality in the United States.
J Nat/ Cancer fnst. 2007;99:1594-1602.
26. Lappe JM. Travers-Gustafson D, Davies KM, Recker RR. Heaney
RP. Vitamin D and calcium supplementat ion reduces cancer risk:
results of a randomized trial. Am J Clin Nuir. 2007;85:
1586- 1591.
27. Melamed ML, Michos ED, Post W, Astor B. 25-Hydroxyvitamin D
levels and the risk ofmoltality in the general population. Arch lntem
Med. 2008;168: 1629- 1637.
28. Dobnig H. Pilz S, Schamagl H, et at. Independent association of low
serum 25-IJydroxyvitamin D and 1,25-dihydroxyvitamin D levels
with all-cause and cardiovascular mortality. Arch Intern Med. 2008;
168:1340- 1349
29. Vieth R, Bischoff-Ferrari H, Boucher BJ, et a!. The urgent need to
recommend an intake of vitamin D that is effective [Editorial]. Am J
Clin Nutr. 2007;85:649-650.
30. Bischoff-Ferrari HA. The 25-hydroxyvitamin D threshold for better
health. J Steroid Biochem Mol Bioi. 2007;103:614-619.
31 . Giovannucci E. Liu Y, Rimm EB, et al. Prospective study of predic-
tors of vitamin D status and cancer incidence and mortality in men.
J Nat/ Cancerlnst. 2006;98:451-459.
32. Schleithoff SS. Zittermann A. Tenderich G. Berthold HK. Stehle P.
Koerfcr R. Vitamin D supplementation improves cytokine profiles in
patients with congestive heart fai lure: a double-blind, randomized,
placebo-controlled trial. Am J Clin Ntar. 2006:83:754-759.
33. Sambrook PN, Chen JS, March LM, et al. Serum parathyroid
hormone is associated with increased mortality independent of
25-hydroxy vitamin D status, bone mass. and renal function in the
frai l and very old: a cohort study. J Clin Endocrirwl Metab. 2004;
34. Frassetto LA. Morris RC, Sebastian A. Effect of age on blood acid-
base composition in adult humans: mle of age-related renal functional
decline. Am J Physiol. 1996;271 (6 Pt 2):FII I4-Fl 122.
35. Frassetto L. Morris RC. Jr. Sebastian A. Potassium bicarbonate
reduces urinary nitrogen excretion in postmenopausal women. J Clin
Endocrinol Mewb. 1997;82:254-259.
Received July 6, 2008
Accepted December 23, 2008
Decision Editor: Luigi Ferrucci, MD, PhD
l ocation:
Friday. April 18, 2008
302,393 (1)
Newspaper (D)
Circul ation (DMA):
Type (Frequency):
Page: 611
Keyword: dermatology
Researcher Received Industry Funds
Was Backed by Group
Tied to Indoor Tanning
A Boston University researcher who
authored an article in the New England
Journal of Medicine last year recom-
mending the moderate use of tanning
beds as a way to treat or avoid vita-
min-D deficiency has received research
funding from an organization funded
and controlled by the tanning-bed in-
The link to researcher Michael Hol-
ick's work and the tanning industry
wasn't made clear in the article. A
note at the end of the article dis-
closed that Dr. Holick's research was
funded, in part, by the UV Founda-
tion. No information about the foun-
dation was provided.
The nonprofit foundation, accord-
ing to its Web site, is funded by the In-
door Tanning Association as well as
the makers of tanning-bed equipment.
The board of directors is composed en-
tirely of tanning-bed -industry offi-
cials. Boston University was the top re-
cipient of grants from the foundation
from 2004 through 2006, the most re-
cent three years oft he group's Internal
Revenue Service filings. In total, the
university received $162,014 during
that period.
On its Web site, the UV Foundation
said it "has made a commitment of
$150,000 over three years to Boston
University, to continue the efforts of
Dr. Michael Holick, a Vitamin D expert."
The site said the foundation "is dedi-
cated to exploring the positive effects
of UV light and to increasing public
awareness about those benefits."
Dr. Holick said in an interview that
the money was an unrestricted grant
that he used for vitamin-D research.
A report on the funding arrange-
ment was to appear in Friday's edition
of the Cancer Letter, a Washington-
based trade publication.
A spokeswoman for the New Eng-
landJownalsaid "wewereawareofthe
source of funding and disclosed it. We
looked at the Web site and saw that the
inf01mation was available and felt the
inf01mation was publicly available."
Please contl'let lhe poblisher lOt reprints.
Dr. Holick said he complied with all dis-
closure requests from the journal.
Questions over disclosures of poten-
tial conflicts have been growing at med-
ical journals. Two weeks ago, the New
England Journal corrected the record
on a lung-cancer-screening study pub-
lished in 2006 to show some of the au-
thors were receiving previously undis-
closed royalties from General Electric
Co., a big scanner make1; and also get-
ting research funding from tobacco
company Vector Group, the parent of
Liggett Group LLC.
The current NEJM policy for review
articles requires that authors "not
have major research support" from rele-
vant companies. Dr. Holick said the
funds were about 2% of his total re-
search budget. The NEJM spokes-
woman said that amount was within
the jow'Jlal's guidelines.
Dr. Holick's article didn't repmt new
research but made recommendations
based on a review of published studies.
Tanning beds were cited as a "recom-
mended" source of vitamin D when
used in moderation.
That concerned some cancer spe-
cialists, who note that ultraviolet rays
used in indoor tanning have been
linked to an increased risk of cancer in
some studies.
"I was surprised that the New Eng-
land Journal, a very prestigious jow-
nal, would run the article this way,"
said Ma1tin Weinstock, a Brown Univer-
sity professor. He said he
was surprised t e journal would run a
piece by an industry-funded author.
-> The Situation: A Boston University
researcher authored an article last year
recommending moderate use of
tanning beds for vitamin-D deficiency.
-:. Industry Funding: The author has
received research funding from an
organization controlled by the
tanning-bed industry.
-:. Disclosure Issues: Questions over
disclosures of potential conflicts have
been growing at medical journals.
Account: 7470NX (3159)
Page 1 of 1
Vitamin D in the prevention and treatment of coronary heart
Armin Zittermann and Reiner Koerfer
Department of Thoracic and Cardiovascular Surgery,
Heart and Diabetes Center North-Rhine Westfali a,
Ruhr University Bochum, Bad Oeynhausen, Germany
Correspondence to Armin Zittermann, PhD, Associate
Professor, Department of Thoracic and Cardiovascular
Surgery, Heart and Diabetes Center North-Rhine
Westfalia, Ruhr University Bochum, Georgslra6e 11,
32545 Bad Oeynhausen, Germany
Tel: +49 573 1 97 1912; fax: +49 5731 97 2020;
e-mail: azit1ermann@hdz-nrw.de
Current Opini on In Clinical Nut rition and
Met abol ic Care 2008, 11 :752-757
Purpose of review
The pathogenesis of coronary heart disease is of multifactorial origin. Probably, not all
risk factors are satisfactorily understood. This article outlines beneficial vitamin D effects
on cardiac function and the vasculature. In addition, human data associating serum
vitamin D metabolite levels or oral vitamin D dosages or both with coronary heart disease
outcome parameters are reviewed.
Recent findings
There is accumulating evidence that the vitamin D hormone calcitriol exerts important
physiological effects in cardiomyocytes, vascular smooth muscle cells, and the vascular
endotheli um. Low levels of the calcitriol precursor 25-hydoxyvitamin D are associated
with myocardial infarction, congestive heart fail ure, and calcific aortic stenosis. Deficient
calcitriol concentrations probably contribute to the massive vascular calcification seen
in chronic kidney disease. In patients with end-stage renal disease and end-stage heart
failure, very low-circulating calcitriol levels or nonuse of active vitamin D or both are
independently associated with high mortality rates.
Despite these exciting data, it is still too early to recommend exact dosages for the
prevention or therapy of coronary heart disease. Prospective, randomized controlled
trials with different amounts of vitamin D and probably with its active form calcitriol are
needed to determine whether vitamin D can prevent coronary heart disease events and
calcitriol, coronary heart disease, mortality, survival, vascular calcification, vitamin D
Curr Opin Clin Nutr Metab Care 11:752- 757
2008 Wolters Kluwer Health I Lippincott Williams & Wilkins
1363-1 950
Coronary heart disease (CHD), also called coronary artery
disease or ischemic heart disease, is the result of the
accumul ation of atheromatous plaques within the walls of
the arteries that supply the myocardium with oxygen and
nutri ents. After decades of progression, some atheroma-
tous plaques may rupture and may thus severel y restri ct
the flow of oxygen carrying blood to the myocardium. As a
consequence, a heart arrack can occur. l\llyocardial infarc-
tions are the most common cause of sudden death in
western countries. In the USA and Germany, CHD is
responsible for one in fi ve deaths. The WHO estimated
that in 2002, 12.6% of deaths worldwide were from CHD
[1]. CHD is also one of the major causes of congestive
heart failure, another prevalent disease in western
countries that is characterized by fai lure of the heart to
supply adequate blood flow and therefore oxygen deliv-
ery co peripheral ti ssues and organs or co do so only from
elevated filling pressures. The prevalence of congestive
heart failure is 2.0% in patients aged 40- 59 years and
rises to 12.0% in patients who are 80 years and older [2].
This review summarizes recent experimental and epide-
mi ological evidence for a role of vitamin D in the pre-
vention of CHO. In additi on, available data concerning
the effect of vitamin D on cardiovascular and total
mortality are presented.
Vitamin D and the heart
The active form of vitamin D, 1,25-dihydroxyvitamin D
or calcitriol, is the end product of two hydroxylation steps
of vitamin 0: a hepatic 25-hydroxylation and a sub-
sequent renal !a-hydroxylation. Calcitriol exerts geno-
mic and nongenomic effects through a cytosolic vitamin
D receptor (VDR) and a membrane bound receptor.
VDRs have been found in almost all human tissues
and cells, among them cardiomyocyces, endothelial
cells, and vascular smooth muscle cells. Several tissues
also possess an enzymatically active 25-hydroxyvitamin
o-la-hydroxylase system, among them vascular smooth
muscle cells [3]. However, cardiomyocytes do not possess
1363 1950 2008 Wolters Kluwer Heal th ! Lippi ncott W illiams & W ilkins 001:10.1 097/MCO.Ob013e328312c33f
Copyright Lippincott Willi ams & Wilkins. Unauthorized reproduction of this article is prohibited.
such a 1-a-hydroxylase activity [4). Therefore, the car-
diac muscle strongly depends on the circulating blood
concentration of calcitriol. It is noteworthy that renal
calcitriol synthesis and thus circulating calcitriol concen-
trations become substrate dependent, that is, dependent
on the circulating 25-hydroxyvitamin D [25(0H)D] con-
centration, in case of vitamin D deficiency/insufficiency.
lt is therefore an important finding that in postmenopau-
sal women, calcitriol concentrations were reduced by
approximately one-third in patients with deficient
25(0H)D levels ( < 25 nmol/1) compared with adequate
levels (>80 nmol/1) [5]. Moreover, in severely obese
patients (BMI > 40 kg/m
), 25(0H)D levels were
20 nmol/1 lower compared with patients with a BMI less
than 25 kg/m
[6]. These two subgroups also differed
significantly in serum calcitriol levels (by 20 pmol/l). Data
are in line with the fact that both, some postmenopausal
women and obese patients, have an enhanced risk for
Experimental vitamin D effects on cardiomyocytes
Calcitriol affects the growth, proliferation, and morpho-
logy of murine cardiomyocytes (Fig. 1). In detail, treat-
ment with calcitriol increased the expression of the
cardiac muscle protein myotrophin. Calcitriol treatment
also decreased expression of atrial natriuretic peptide, a
biochemical risk marker that is inversely related to car-
diac function. In addition, calcitriol treatment increased
expression and nuclear localization of the VDR in these
cells [7]. Murine cardiomyocytes isolated from VDR
knockout mice show accelerated rates of contraction
and relaxation as compared wi th wild type mice, and
calcitriol directly affected contractility in the wild type
but not the knockout cardiomyocyte [8]. Thus, calcitriol
is an important hormone involved in modulating and
maintaining heart cell structure and function. The afore-
mentioned study group [9] has also demonstrated that
the heart of VOR knockout mice is hypertrophied
because of cellular hypertrophy of heart myofibrils.
Microarray analysis revealed tissue inhibitors of metallo-
proteinases were significantly underexpressed in VDR
knockout mice compared with wi ld type mice [10].
Matrix metalloproteinases (MMPs) are connective tissue
enzymes that are involved in remodeling of the vascular
wall and myocardium. They function to destabilize athero-
matous plaques to cause rupture and thrombosis within
the lumen. Extracellular matrix remodeling mediated by
MMPs contributes to progressive left ventricular remo-
deling, dilation, and heart failure. Experimental data
suggest that calcitriol may prevent the development of
cardiac hypertrophy. In spontaneously hypertensive
heart fai lure rats, calcitriol treatment resulted in lower
heart weight, myocardial collagen levels, left ventricular
diameter, and cardiac output compared with untreated
rats [111.
Vitamin D in the prevention of CHD Zittermann and Koerfer 753
Vitamin D and cardiac events
During the last decade, it became clear that deficient
serum concentrations of vitamin D metabolites are preva-
lent not only in specific patient groups but also in the
general population in western countries and throughout
the world [121. The by far most important reason for this
phenomenon is an inadequate skin exposure ro solar
ultraviolet B radiation, as ultraviolet B-induced skin
synthesis is the major source of vitamin D for humans.
Ecological studies have reported hi gher rates of CHD
with increasing distance from the equator, a phenomenon
that can be attributed to the higher prevalence of vitamin
D deficiency in regions with less exposure to sunlight.
In a nonrandomized prospective study in 1739 Framing-
ham Offspring Study participants, individuals with low
25(0H)O levels ( <37.5 nmol/1) had a multivariable-
adjusted hazard ratio of 1.62 for incident cardiovascular
disease such as myocardial infarction, coronary insuffi-
ciency, and heart failure compared with those with
25(0H)D levels of at least 37.5 nmol/1 [13.]. In the
Women's Health Initiative study, however, myocardial
infarction, ischemic attack, hospitalization rate for heart
failure, and cardiovascular death could not be prevented
by supplementation with 1000 mg calcium and 10 fl-g
vitamin D daily compared with the placebo group [14].
But several li mitations such as the low daily vitamin D
dose and the lack of measurements of serum 2S(0H)D
levels or of calcitriol or of both of the vitamin D arm of
this srudy makes data interpretation difficult [15]. In a
nested case- control study among male participants of the
Health Professionals Follow-up Study [1 6.], men with
low 25(0H)D levels ~ 3 7 . 5 nmol/1) had a relative risk of
2.09 [95% confidence interval (Cl) 1.24- 3.54] of myo-
cardial infarction compared with those considered to
be sufficient (:;75 nmol/1), after adjustment for various
lifestyle and other risk factors. Very recently, data on all-
cause and cardiovascular mortality in association with
vitamin 0 status have been published from a prospective
cohort study of 3258 consecutive male and female
patients scheduled for coronary angiography [171.
During a median follow-up period of 7.7 years, 737
patients (22.6%) died, including 463 deaths from cardio-
vascular causes. Multivariate-adjusted hazard ratios for
patients in the lower two ZS(OH)D quartiles (median,
19.0 and 33.3 nmol/1) were higher for all-cause mortality
(hazard ratio 2.08; 95% CI 1.60-2.70; and hazard ratio
1.53, 95% CI 1.17 -2.01, respectively) and for cardiovas-
cular mortality (hazard ratio 2.22; 95% CI 1.57 - 3.13; and
hazard ratio 1.82, 95% CI 1.29- 2.58, respectively) com-
pared with patients in the highest 25(01-I)O quartile
(median, 71.0 nmol/1). Similar results were obtained for
patients in the lowest calcitriol quartile. A trend toward
lower mortality was already observed in a British vitamin
D intervention trial [18]. T he vitamin D group recei ved
one capsule containing 2500 fl-g vitamin 0
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
754 Functional foods
Figure 1 Biochemical and clinical effects of low calcitriol availability on cardiac and vascular function
Low serum levels of Chronic kidney
25-hydroxyvitamin D disease
t t
Low cellular availabi lity of calcitriol

Enhanced synthesis andjor Reduced synthesis of Disturbed cardiomyo- Reduced synthesis
enhanced deleterious effects MGP, osteopontin, cyte contraction of inhibitors of MMPs
of MMPs, TNF- ex, IL-1, IL-6, IL-10, and type IV Enhanced synthesis of and cardiac muscle
contracting factors, and AGEs collagen natriuretic peptides protein

Vascular calcifi cation Cardiac hypertrophy
Coronary heart disease
Myocardial infarction Heart fail ure
AGEs, advanced glycation end products; IL-1, interleukin-1; IL-2, interleukin-2; MGP, matrix Gla protein; MMP, matrix metalloproteinases; TNF, tumor
necrosis factor.
4 months over 5 years (equivalent w 211-Lg vitamin D per
day). In subgroup analysis, serum 25(01-I)D concen-
trations were 21 nmol/1 higher in the active vi tamin 0
group compared with the placebo group. The rare ratios
for cardiovascular diseases, cancers, and coloreccal can-
cers incidences were 0.90 (0. 77 -1.06), 1.11 (0.86- 1.42),
and 1.02 (0.60-1.74), respectivel y. For mortality, these
ratios were 0.84 (0.65-1.20), 0.86 (0.61 - 1.20), and 0.62
(0.24- 1.60), respectively. Hence, though statistical sig-
nificance was not achieved, incidence rate ratios were
always close to 1.0, whereas mortality rate ratios were
always lower, suggesting that the protective effect of
vitamin D on cardiovascular and cancer mortality is more
effective than irs effect on disease incidence. In line with
this suggestion, a metaanalysis of controlled clinical trials
came co the conclusion that vitamin 0 supplementation
reduced total mortality in middle aged to elderly adul ts
by 7% during a trial size-adj usted mean of 5. 7 years [19 ].
In the vitamin 0 -suppl emented patients and controls,
mean serum 25(0H)O concentrations increased by
40 nmol/1 and decreased by 6.5 nmol/1, respectively.
Although the authors could not evaluate cause-specific
mortality, the relatively immediate effect of a large
enough magni tude co influence coca! mortality would
suggest a preventive effect on Cl-IO morcalicy.
Vitamin D and congestive heart failure
The aforementioned experimencal data and several
clinical results indicate that adequate vitamin 0 metab-
olite levels may contribute to the prevention of conges-
tive heart failure (CHF). In a recent study in 43 men and
17 women \Vith left ventricular ejection fraction of 40%
or less, longer 6 min walk distance was correlated with
higher 25(0H)D levels [20.]. The 6 min walk distance is
a frequently used cest in heart fai lure patients to assess
functi onal cardiac outcome. Other independent vari ables
of the 6 min wal k distance were age, sex, and hi ghly
sensi ti ve C-reactive protein levels. A case-controlled
study [21] showed that lifestyle factors associated with
adequate vi tamin D status such as membership in a sports
club, frequent summer holidays, and li ving in rural areas
were less frequently reported for the period of childhood,
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
adolescence, and young adulthood by CHF patients
compared with healthy controls. Data indicate that
vitamin D insufficiency/deficiency is a causal factor of
CHF and not just the result of disease-related alterations.
Very low serum calci triol levels ( <37.5 pmol/1) have fre-
quently been found in end-stage heart failure patients
[22""]. In this study, patients in the highest calcitriol
tertile had a hazard ratio for an event (death or cardiac
transplantation) of only 0.506 (95% CI 0.334- 0. 767)
compared with patients in the lowest calcitriol tertile,
after adjustment for potential confounders. Data
support the assumption that low vitamin D metabolite
levels are a risk factor for survival and that vitamin D
seems to be a very important protective factor for
cellular health.
Vitamin D and the vasculature
Classical risk factors in the pathogenesis of CHD are
smoking, dyslipoproteinemia, hypertension, disturbed
glucose metabolism, and proinfiammatory processes.
Until recently, vascular calcification was considered to
be a passive process that occurred as a nonspecific
response to vascular damage without clinical significance.
However, there is no-..v accumulating evidence that vas-
cular calcificat ion is an acti ve process [23""]. Almost all
angiographically atherosclerotic lesions are calcified [24].
Vascular calcification can cause thrombosis, arterial rup-
ture, and myocardial infarction. In general, the develop-
ment of tissue calcification requires a preexisting injury as
an inducer, whereas further progression requires the
presence of other promoter factors such as hyperpho-
sphatemia and hypercalcemia or a deficiency in calcifica-
tion repressor factors or both [251.
Experimental vitamin D effects on the vasculature
Beside the heart, the vasculature is an important target
tissue for vitamin D. Available data concerning protective
effects of vitamin 0 on the vasculature have recently
been summarized [23""]. These effects include calcitriol-
mediated downregulation of MMPs and proinflammatory
cytokines such as IL-l, IL-6, and TNF-a (Fig. 1). The
effects also include upregulation of the anti-inflammatory
cytokine IL-10 and of inhibitors of vascular calcification
such as matrix Gla protein (MGP), osteopontin, and
type IV collagen in vascular smooth muscle cells and
osteoblast-like cells. In addition, calcitriol blunts the
deleterious impact of advanced glycation end products
(AGEs) on endothelial cells [26"). AGEs are believed
to induce vascular dysfunct ion in diabetic and uremic
patients and may explain the link between hyper-
glycemia and the development of vascular complications.
In experimental studies with rings of spontaneously
hypertensive rats aorta, calcitriol treatment reduced the
Vitamin D in the prevention of CHD Zittermann and Koerfer 755
acetylcholine-induced and ATP-induced endothelium-
dependent contractions [27"]. Moreover, the acetyl-
choline-induced release of prostacyclin was reduced by
the acute administration of calcitriol. Calcitriol also
reduced the increase in cytosolic free calcium concen-
tration caused by acetylcholine. These data demonstrate
that calcirriol modulates vascular cone by reducing
calcium influx into the endothelium and hence decreas-
ing the production of endothelium-derived contracting
Vitamin D deficiency and vascular calcification
in chronic kidney disease
Obviously, chronic kidney disease (CKD) can be con-
sidered as a human model for the effects of calcitriol
deficiency on vascular calcification and Cl-IO mortality.
In CKD, both calcitriol deficiency and vascular calcifica-
tion are extremely common. DistUrbances in mineral
homeostasis such as hyperphosphatemia and hyperpar-
athyroidism seem to be secondary to calcitri ol deficiency
and are obviously key factors for medial calcification
[23""]. End-stage renal disease (stage 5 CKD) results
in glomerular filtration rates below 15 ml/min/1.73 m
As a consequence, circulating calcitriol levels fall below
37.5 pmol/1 (reference range 40- 150 pmol/1). In end-stage
renal disease, the age-standardized risk for cardiovascular
events and mortality is 37 times and 10-20 times,
respectively, higher than in the general population [23""].
Children on dialysis with calcitriollevels helow 40 pmol/1
or ahove 150 pmol/1 had significantly higher cardiac cal-
cification score than controls and patients wi th calcitriol
levels in the reference range [28""]. A mouse model of
CKD-stimulated atherosclerotic cardiovascular mineral-
ization demonstrated that treatment with calcitriol and
one of its analogs, paricalcitol, was protective against
aortic calcification at dosages sufficient to correct second-
ary hyperparathyroidism. However, higher dosages
stimulated aortic calcification [29"]. Data support the
assumption that both inadequately low and high vitamin
D concentrations are associated with deleterious effects
on the vasculature [30"]. In non-CKD patients, lower
serum 25(0H)D levels and higher serum parathyroid
hormone (PTH) levels were independentl y associated
with calcific aortic stenosis [31"]. In patients with type 2
diabetes and low vitamin D status [25(0H)D levels
<50 nmol/1], vitamin D treatment with a single oral
vitamin 0
bolus of 100000 JU (1500J.Lg vitamin 0
improved endothelial function, measured by flow medi-
ated vasodilatation, in a double-blind, parallel group,
randomized trial [.32"].
Treatment with active vitamin D and total mortality
in chronic kidney disease
Wolf eta/. [33."] reported that in CKD, serum 25(0H)D
levels of at least 25 nmol/1 or serum calcitriol levels of at
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
756 Functional foods
least 37.5 pmol/1 were associated with reduced risk of
early mortality compared with patients who were
untreated and had 25(0H)D and calcitriol levels below
these respective values. Some studies indicate that treat-
ment of CKD with active vitamin D (0.25-0.75 !J..g/d)
reduces all -cause mortality by 16- 24% and cardiovascular
mortality by 70% (30"]. Data also indicate that the differ-
ences in survival between patients treated with different
forms of active vitamin D (calcitriol, doxercalciferol,
and paricalcitol) are small [34] . In a recent study
[35"], patients receivi ng treatment with calcit riol , 0.25-
0.5 !J..g/day, for a median duration of 2.1 years, had a
significantly lower incidence rate ratio [adjusted relative
risk (RR) 0.35, 95% CI 0.23- 0.54] for predialysis
mortality chan untreated patients. Although calcitriol
appears co be associated with greater survival, results
have to be confirmed by the use of prospect ive, rando-
mized tri als.
Several experimental studies demonstrate that the vita-
min D hormone calcitriol plays a pivotal role for normal
physiology of cardiomyocytes and the vasculature. Low
vitamin D metabolite levels are associated with vari ous
CHD outcome parameters such as myocardial infarction,
congestive heart fai lure, and cardiovascular mortality.
There are also exciting data that vitamin D may reduce
total mortality, especially in patients with deficient calci-
triol levels. However, it is still too earl y to recommend
exact dosages for the prevention or therapy of CHD, as
prospective, randomized controlled trials with adequate
amounts of vitamin D are Jacking.
References and recommended reading
Papers of particular interest, published within the annual period of review, have
been highlighted as:
of special interest
of outstanding interest
Additional references related to this topic can also be found in the Current
World Literature section in this issue (p. 8 18).
World Health Organization. The World Health Report 2004 - Changing
History, 2004, 1204. ISBN 924 156265X.2.
2 Parikh Nl, Vasan RS. The epidemiology of heart failure. In: Colucci WS, editor.
Atlas of heart failure, 5th ed. Philadelphia, Pennsylvania: Springer; 2008. pp.
1-1 4.
3 Somjen D, Weisman Y, Kohen F, et at. 25-hydroxyvi tamin 031 alpha-hydro
xylase is expressed in human vascular smooth muscle cells and is upregulated
by parathyroid hormone and estrogenic compounds. Circulation 2005;
4 Hewison M, Zehnder D, Chakraverty R, Adams JS. Vitamin D and barrier
function: a novel role for extra-renal 1 alpha-hydroxylase. Mol Cell Endocrinol
2004; 215:31- 38.
5 Rejnmark L, Vestergaard P, Heickendorff L, et at. Plasma 1,25(0H)2D levels
decrease in postmenopausal women with hypovitaminosis D. Eur J Endocrinol
2008; 158:571-576.
The above is a crosssectional study in 315 healthy postmenopausal women
selected from the local background population. This study demonstrated
that circulating calcitriol levels depend on the availability of its substrate
25-hydroxyvitamin D in case of pl asma 25-hydroxyvitamin D levels below
6 Konradsen S, Ag H, Lindberg F, et at. Serum 1,25-dihydroxy vitamin D is
inversely associated with body mass index. Eur J Nutr 2008; 47:87- 91.
This cross-sectional study investigated the dependency of serum 25-hydroxyvi-
tamin D and calcitriol levels on BMI in 21 87 patients. Obesity was associated with
lower 25-hydroxyvitamin D and calcitriol levels.
7 Nibbelink KA, Tishkoff DX, Hershey SO, eta/. 1,25(0H)2vitamin 03 actions
on cell proliferation, size, gene expression, and receptor localization, in the
HL-1 cardiac myocyte. J Steroid Biochem Mol Bioi 2007; 103:533- 537.
8 Tishkoff OX, Nibbelink Ka, Holmberg KH, eta/. Functional vi tamin D receptor
(VDR) in the ! tubules of cardiac myocytes: VDR knockout cardiomyocyte
contractility. Endocrinology 2008; 149:558 - 564.
This study investigated the subcellular localization of the functional VDR in adult rat
9 Simpson RU, Hershey SH, Nibbelink KA, eta/. Characterization of heart size
and blood pressure in the vitamin D receptor knockout mouse. J Steroid
Biochem Mol Bioi 2007; 103:521 - 524.
This study demonstrated that ablation of the VDR signaling system results in
cellul ar hypertrophy. This model supports the assumption that the vitamin D
hormone calci triol may contribute to the prevention of congestive heart failure.
1 0 Rahman A, Hershey S, Ahmed S, et a/. Heart e<tracellular matrix gene
expression profile in the vitamin D receptor knockout mice. J Steroid Biochem
Mol Biol 2007; 103:416-419.
11 Mancuso P, Rahman A, Hershey SO, el a/. 1,25DihydroxyvitaminD3 treat
ment reduces cardiac hypertrophy and left ventricular diameter in sponta
neously hypertensive heart failure-prone (cp/ +} rats independent of changes
in serum leptin. J Cardiovasc Pharmacol 2008; 51 :559-564.
Another experimental study demonstrating the pivotal role of calcitriol for normal
cardiac function.
12 Holick MF. Vitamin D deficiency. N Engl J Med 2007; 357:266 -
A comprehensive overview about the reasons and consequences of vitamin D
13 Wang TJ, Pencina MJ, Booth SL, et at. Vitamin D deficiency and risk of
cardiovascular disease. Circulation 2008; 117:503-511.
In 17 39 Framingham Offspring Study participants, the multivariableadjusted
hazard ratio lor cardiovascular disease was 1 .62 (95% Cl 1.11 - 2.36} for
individuals with 25-hydroxyvitamin 0 levels of less than 37.5 nmol/1 compared
with those with 25hydroxyvitamin D levels of at least 3 7.5 nmol/1. This effect was
evident in patients with hypertension (hazard ratio 2.13, 95% Cl 1.30-3.48} but
not in those without hypertension (hazard ratio 1.04, 95% Cl 0.55- 1.96). Mean
follow-up was 5.4 years.
14 Hsia J, Heiss G, Ren H, et a/. Calcium/vitamin D supplementation and
cardiovascular events. Circulation 2007; 115:846- 854.
15 Zittermann A, Schleitholl SS, Koerfer R. Letter by Zitterman et at. regarding
article, 'Calcium/vitamin D supplementation and cardiovascular events'. Cir
culation 2007; 116:e85.
A letter to the editor that addresses the problems of appropriate study design and
data interpretation in vitamin D research.
16 Giovannucci E, Liu Y, Hollis BW, Rimm EB, et at. A prospective study of
25hydroxyvitamin D in relation to risk of myocardial infarction in men. Arch
Intern Med 2008; 168:1 174- 1180.
In a nested case- control study among male participants of the Health Proles
sionals Follow-up Study, the multivariableadjusted relative risk l or myocardial
infarction was 2.42 (95% Cl 1.53-3.84) lor individuals with 25-hydroxyvitamin D
levels of less than 37.5 nmol/1 compared with those with 25hydroxyvitamin D
levels of at least 75 nmol/ 1. Follow up was 1 0 years.
17 Dobnig H, Pilz S, Scharnagl H, el a/. Independent association of low serum
25-hydroxyvitamin D and 1,25-dihydroxyvi tamin D levels wi th all-cause and
carduiovascular mortality. Arch Intern Med 2008; 168:1340- 1349.
This large prospective cohort study adds further evidence l or an association of low
vitamin D status with total and cardiovascular mortality.
18 Trivedi DP, Doll R, Khaw KT, et at. Effect of four monthly oral vitamin 03
(chol ecalciferol) supplementation on fractures and mortality in men and
women living in the community: randomised double blind controlled trial.
BMJ 2003; 326:469-472.
19 Autier P, Gandini S. Vitamin D supplementation and total mortality: a meta
analysis of randomized controlled trials. Arch Intern Med 2007; 167: 1730-
A metaanalysis of 1 8 randomized controlled trials demonstrating that vi tamin
supplementation reduced the relati ve risk for total mortality to 0.93 (95% Cl
0.87 - 0.99) in middle-aged and early patients. Mean dose of daily vitamin D was
13.21'-g (ranging from 7.5 to
20 Boxer RS, Dauser OA, Walsh SJ, el at. The association between vitamin D and
inflammation with the 6min walk and frailty in patients with heart failure. Am
Geriatr Soc 2008; 56:454- 461 .
This cross-sectional study was performed in 60 patients with heart failure. Aerobic
capacity was reduced in patients with low 25-hydroxyvitamin D levels.
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
21 Zittermann A, Fischer J, Schleithoff SS, et al. Patients with congestive heart
failure and healthy controls differ in vitamin D-associated lifestyle factors. tnt J
Vitam Nutr Res 2007; 77:280- 288.
A case- control study that investigated lifestyle factors of congestive heart failure
patients before onset of the disease. Results of 1 50 patients were compared with
150 apparently heal thy controls.
22 Zittermann A, Schleithoff SS, Gotting C, el a/. Poor outcome in end-stage
heart failure patients with low circulating calcitriol levels. Eur J Heart Fail 2008;
A very large investigation on 389 congestive heart failure patients i ndicating that
low calcitriollevels are a risk factor for poor outcome (death or cardiac transplanta-
tion). This study supports the assumption that adequate circulating calcitriol levels
play a pivotal role for survival.
23 Zittermann A, Koerfer R. Protective and toxic effects of vitamin D on vascular
calcification: clinical implications. J Mol Aspects Med 2008. [Epub ahead of
A comprehensive review that summarizes available data of vitamin D effects on the
vasculature and on vascular calcification.
24 Margolis JR, Chen JT, Kong Y, et a/. The diagnostic and prognostic sig-
nificance of coronary artery calcification. A report of 800 cases. Radiology
1980; 137:609- 616.
25 Grases F, Sanchis P, Perello J, et a/. Effect of crystallization i nhibitors on
vascular calcifications induced by vitamin D - A pilot study in Sprague-Dawley
rats. Circ J 2007; 71: 1152-1 156.
26 Tal mor Y, Golan E, Benchetrit S, eta/. Calcitriol blunts the deleterious impact
of advanced glycation end products (AGEs) on endothel ial cells. Am J Physiol
Renal Physiol 2008; 294:F1059-F1064.
In this experimental study, calcitriol administration blunted several metabolic
pathways in endothelial cells that are associated with the adverse effects of AGEs.
27 Wong MS, Delansorne R, ManRY, er al. Vi tamin D deri vatives acutely reduces
endothelium-dependent contractions in the aorta of the spontaneously
hypertensive rat. Am J Physiol Heart Circ Physiol 2008; 295:H289-H296.
This study adds further evidence for an important role of vitamin D in the
28 Shroff R, Egerton M, Bride! M, eta/. A Bimodal Association of Vi tamin D Levels
and Vascular Disease in Children on Dialysis. J Am Soc Nephrol 2008;
19:1239- 1246.
An important investigation in children on dialysis supporti ng the assumption of a
bimodal association of calcitriol levels with vascular calcification.
Vitamin D in the prevention of CHD Zittermann and Koerfer 757
29 Mathew S, Lund RJ, Chaudhary LR, el a/. Vitamin D receptor activators can
protect against vascular calcification. J Am Soc Nephrol 2008; 19:1509-
Another important study, performed in a mouse model of CKD, supporti ng the
assumption of a bimodal association of circulating calcitri ol levels with vascular
30 Zittermann A, Schleithoff SS, Koerfer R. Vi tamin D and vascular calcification.
Curr Opinion Lipidol 2007; 18:41 - 46.
A comprehensive review of experimental and clinical data about the association of
vitamin D with vascular calcification.
31 Linhartova K, Veselka J, Sterbakova G, eta/. Parathyroid hormone and vitamin
D levels are independentl y associated with calcific aortic stenosis. Circ J
2008; 72:245 - 250.
This cl ini cal study, performed on 122 consecutive patients, showed that
low 2 5-hydroxyvi tamin D level s were an i ndependent predictor of aortic
32 Sugden JA, Davies Jl, Witham MD, et a/. Vitamin D improves endothelial
function in patients with type 2 diabetes mellitus and low vitamin D levels.
Diabet Med 2008; 25:320-325.
This is one of the first prospective, randomized, placebo-controlled, double-blind
studies demonstrating that vitamin D supplementation can improve endothelial
function i n humans.
33 Wolf M, Shah A, Gutierrez 0 , et a/. Vitamin D level s and early mortality
among i ncident hemodialysis patients. Kidney l nt 2007; 72: 1004-
This is a large cross-sectional analysis on 825 incident US hemodialysis patients.
The investigation demonstrated an association of baseline vitamin D levels with
90-day mortality.
34 Tentori F, Hunt WC, Stidley CA, et a/. Mortality risk among hemodialysis
patients receiving different vitamin D analogs. Kidney tnt 2006; 70:1858-
35 Kovesdy CP, Ahmadzadeh S, Anderson JE, et al. Association of acti vated
vitamin D treatment and mortality in chronic kidney disease. Arch Intern Med
2008; 168:397-403.
This investigation was performed on 520 male US veterans with CKD stages 3 - 5
not yet receiving dial ysis. The study supports the hypothesis that calcitriol may
improve survival.
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Working Group Reports
Volume 1
This report represents the views and expert opinions of an IARC Working
Group that met in Lyon, France
27 - 29 June 2005
IARC Library Cataloguing in Publication Data
IARC Working Group on Risk of Skin Cancer and Exposure to Artificial Ultraviolet Light (2005 : Lyon,
Exposure to artificial UV radiation and skin cancer I views and expert opinions of an IARC Working
Group that met in Lyon, France 27- 29 June 2005.
(IARC Working Group Reports ; 1)
1. Skin Neoplasms - epidemiology 2. Skin Neoplasms- etiology 3. Ultraviolet Rays
4. Risk Assessment I. Title II. Series
ISBN 92 832 2441 8 (NLM Classification: W1)
List of participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
List of abbreviations ... . ...... . .. . . . ....... . ... .. . . ....... . ......................... vii
Preamble ...... . ... . . . . . . . . .. . .. .... . . .. . ..... . .......... . ....... . ........... . .... . ix
Executive summary ...... . .. . .......... . ............... . ... . .............. . ..... . ... xi
Physical characteristics and sources of exposure to artificial UV radiation .. ................ 1
Physical characteristics of UV radiation . .......... . .......... .. ........................ 1
Units and measurements of UV radiation ..... . .... . ............... . . ................... 1
Measurement of ambient solar UV radiation . . . ........... . ........ . .. . ...... . . . . .... . .. 1
Standard erythemal dose (SED) and minimal erythemal dose (MED) ........... .. ........ . .. 2
UV index . ... . .. . .... . .... . ................ . ... . ...... . ...... ... . . .......... . . .. 2
Limit values .... . .... . .... . ...................................... . ........ . ...... 2
Sources of natural and artificial UV radiation . . ................................. . ............ 2
Solar radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Artificial UV radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Comparison of UV spectrum from sunlight and indoor tanning appliances .... . .............. . 5
European and international positions regarding artificial sources of UV radiation ... .. ........ 5
Standard for appliances designed specifically for tanning purposes .. . . . ... . . . ............... 5
National and international scientific policies .............. . . ...... . ...... . . . ............. 6
Regulations . . ... . . . .... . ........................ . ........... .. ........ . ......... 6
Biological effects of exposure to UV radiation relevant to carcinogenesis . . ...... . . .. . ...... 7
Biological lesions induced by UVA and UVB radiation ....................... . .. . ......... 7
DNA damage ......... .. .................... . ........... ... .. . ....... . .. . ........ 7
Cell damage ..... .. ... . .... . ....................... . .. . ........ . ................. 7
UVA, UVB and human skin ................. . ............. . .......... . ... . .......... 8
Differential effect of UVA and UVB on skin cancers ...... . ... . ....... . ............. . ..... 8
Experimental systems ........... . .............................. . .. . , .......... . ... 8
Relevance of experimental data to human skin cancers .. . .... . . . ... . . . ........... . ....... 8
Changes in immune response ...... . ...... . .......... . .... . ........... . .. . ........ . . . 9
Experimental systems ......... . ........... . ........... . ............ .. ... . ......... 9
Studies in humans ............. .. ... . . .. . . .. . .... . ...... . ... .. . . .......... . . . ..... 9
Effects of natural and artificial UV radiation on human skin .... . .................. .. ...... 9
Variety of skin types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Sunburn .. . ..... . ............ . . .. ..... . .............. . . . .......... . ... . ......... 10
Tan acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 0
Prevalence of exposure to artificial UV radiation for tanning purposes ....... . .. . .. . .. . ..... 11
Prevalence of exposure by region/country ... . .. . .................... . . . . . . . .... . . . . . ... 11
Time trends . ... . ..... . .... . . . .. . ........... . .... . . . ....... . .... .. . .. . . . ........... 11
Personal characteristics of adult users ........................ . .. . .. . .... . ...... . . . ... 13
Sex ..... . ........... . .... . .. . ........ . ........... . . . ........... . . . .... . . . .. . . . . 13
Age ... . .... . . . ........ . ... . .... . .... . ............ . ........ . ........... . ........ 14
Exposure to Artificial UV Radiation and Skin Cancer
Skin type .................... . ....... . ............. . . . . .... ......... . ... . ....... 14
Other factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Personal characteristics of adolescent and children users . ...... ... ........... ... ........ 15
Studies of compliance to regulations and recommendations . .. . .... . .... .... ......... ... . 15
Compliance of operators . . . . .... . . .. . .... . ..... . .. . . . . . .. .... . .............. .. . .... 15
Compliance of customers ................ .. .................. . . . .. . ............ . ... 18
Epidemiological data on exposure to artificial UV radiation for cosmetic purposes and
skin cancers ..... . ... . ........ .... ................... ... ... . .. .. . . . .... . . .... . . .. .. 20
Methodology for literature search ............ . ...... ... . .. . .. . ............ .. . .. . ... .. . 20
Melanoma ............................. ........ ......................... .... ....... 21
Description of studies .................. . ......... . ... . . .. . .. . ... . ........ . . .... ... 21
Quantitative approach: meta-analysis ............................................. . . .. 25
Discussion .... . ......... . ................................. . ..................... 33
Basal cell and squamous cell carcinomas .................. . ........................... 38
Description of studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Meta-analysis ............. .. ......... ............... . . ... . . ... . ............. . ........ 40
Quality of studies ................ ..... .' .................... . ...................... 40
Other sources of exposure to artificial UV radiation ............ .... ...................... 41
Medical use ..... . .. . ... .................. . ... .. . ........ . . .. . .... . ....... . .. .. .. 41
Lighting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Effects of artificial UV radiation not relevant to skin carcinogenesis ........... ..... . ..... .. 44
Cutaneous diseases ............. . .. .. .. . . ..... .. . .. . ................... .. .......... 44
Skin ageing .......... ...... ... . ................... . .. . .. .. .. . ......... . . .. . . .... 44
Other skin diseases caused or exacerbated by exposure to UV radiation ..................... 44
Drug-induced photosensitivity .. . .... . .. . ...................... . ..................... 45
Effects on the eyes .. . . ....... . . .............. . ......... .. ..... . .................. . . 45
Cataract ...... . .......... . . . ......... . ..... . .. . .. . .. . ............... . ... .. ...... 45
Intraocular melanoma . . .... . ... ....... . .. . . . . .............. . . .... . .. .... . . . .. . .... 46
UV exposure and vitamin D ...... . .. . . .... .. . .. ... . ............................. . .. . . 46
Vitamin D formation by photosynthesis ........... . .. . . .. . . .. . . ... . . ..... ........... . . . 46
Dietary sources of vitamin D ........ . .... . .. . .. . ...... . . .... . . . . .... . .. . .... . . . .. . .. 46
Vitamin D and exposure to artifical UV radiation for tanning purposes . .............. . .... ... . 48
Vitamin D and xeroderma pigmentosum patients ..... .... . .. . . . ....... .. ....... ... .. .. .. 48
Summary and Conclusion
Summary ............... . ....... . ........ . ..... . . .. . ........ ...... ................ 49
Conclusion . ...... . .......... . . .. . ....................... .. ................... . .... 50
References ..... . ..... ... .. .... .... . ......... ................ . ..... . . . . ........... . 51
Appendix: European and international positions regarding artificial sources of UV radiation ... 61
Establishment of a standard for appliances designed specifically tor tanning purposes ....... 61
National and international scientific policies ....... . ........ . . . .... ................. . .. . 62
Regulations .... .... . .. . . ............... .. ...... . ....... . .... .. ....... . .. . ... . . . .. . 63
Dr Philippe Autier
150 cours Albert Thomas
69008 Lyon
Dr Mathieu Boniol
150 cours Albert Thomas
69008 Lyon
Dr Peter Boyle
150 cours Albert Thomas
69008 Lyon
Mr J. Daniel (Technical Editor)
150 cours Albert Thomas
69008 Lyon
Dr Jean-Francais Dore
Centre Leon Berard
28 rue Laennec
69008 Lyon
Dr Sara Gandini
Division of Biostatistics and Epidemiology
European Institute of Oncology
Professor Adele Green (Chair)
Queensland Institute of Medical Research
PO Royal Brisbane Hospital
Brisbane 4029, Queensland
Professor Julia Newton-Bishop
Cancer Research UK Genetic Epidemiology Div.
St James's University Hospital
Beckett Street
Leeds LS9 7TF
United Kingdom
Professor Martin A. Weinstock
Dermatoepidemiology Unit
Department of Dermatology
Brown University Medical School
VA Medical Center- 111 D
Providence, Rl 02908
Dr Johan Westerdahl [unable to attend]
Department of Surgery
Lund University Hospital
22185 Lund
Dr M. Beatrice Secretan (Coordinator)
150 cours Albert Thomas
69008 Lyon
Dr Stephen D. Walter
Visiting Scientist at IARC until mid-July 2005
Clinical Epidemiology and Biostatistics
McMaster University
1200 Main Street West
Hamilton, Ont. L8N 3Z5
American Conference of Governmental Industrial Hygienists
Basal cell carcinoma
95% confidence interval
Commission lnternationale de I'Eclairage
Degrees of freedom
Graft versus host disease
General practitioner (family doctor)
International Agency for Research on Cancer
International Commission of Non-Ionising Radiation Protection
Immediate pigment darkening
International Organization for Standardization
Minimal erythemal dose
National Radiation Protection Board
National Toxicology Program
Odds ratio
Psoralen photochemotherapy
Relative risk
Squamous cell carcinoma
Standard erythemal dose
United Nations Environment Programme
World Health Organization
The concern that there may be an association between exposure to artificial UV radiation and skin
cancer was reactivated in 2003-4 when the 1Oth Report on Carcinogens published by the National
Toxicology Program in the USA classified UVA radiation as a "Known Carcinogen to Humans".
In October 2004, the French Ministry of Health contacted the Director of the International Agency
for Research on Cancer (IARC) , Dr Peter Boyle, raising a particular concern about the
continuous increase in incidence of melanomas in France and in the world. Since the last IARC
Monograph on ultraviolet (UV) radiation in 1992, a large number of epidemiological and
experimental studies have been conducted on the risks associated with exposure to UV radiation. The
Ministry therefore requested IARC to investigate the possibil ity of reevaluating the carcinogenic risk
associated with this radiation, particularly concerning artificial UV sources and the use of indoor
tanning facilities.
A Working Group and a Secretariat were gathered by Dr Peter Boyle to this end. The Secretariat
met in January to prepare for the meeting of the Working Group in June 2005. The Working Group
met on 27-29 June 2005 to compile the present document.
We have assessed the available evidence relating to possible detrimental health effects of expo-
sure to artificial ultraviolet radiation through use of indoor tanning facilities, in particular whether their
use increases the risk for skin cancer. Epidemiologic studies to date give no consistent evidence that
use of indoor tanning facilities in general is associated with the development of melanoma or skin can-
cer. However, there was a prominent and consistent increase in risk for melanoma in people who first
used indoor tanning facilities in their twenties or teen years.
Limited data suggest that the risk of squamous cell carcinoma is similarly increased after first use
as a teenager. Artificial tanning confers little if any protection against solar damage to the skin, nor
does use of indoor tanning facilities grant protection against vitamin D deficiency. Data also suggest
detrimental effects from use of indoor tanning facilities on the skin's immune response and possibly
on the eyes (ocular melanoma) .
Knowledge of levels of UV exposure during indoor tanning is very imprecise. Moreover, early
studies published had low power to detect long-term associations with artificial UV exposure that
become evident only following a prolonged lag period. Although the available findings are therefore
not conclusive, the strength of the existing evidence suggests that policymakers should consider
enacting measures, such as prohibiting minors and discouraging young adults from using indoor
tanning facilities, to protect the general population from possible additional risk for melanoma and
squamous cell carcinoma.
Physical characteristics and sources of exposure to artificial UV radiation
For most individuals, the main source of
exposure to ultraviolet (UV) radiation is the sun.
Nevertheless, some individuals are exposed to
high doses of UV through artificial sources.
Sunbeds and sunlamps used for tanning purposes
are the main source of deliberate exposure to
artificial UV radiation.
Physical characteristics of UV radiation
UV radiation belongs to the non-ionizing part of
the electromagnetic spectrum and ranges
between 100 nm and 400 nm; 100 nm has been
chosen arbitrarily as the boundary between non-
ionizing and ionizing radiation. UV radiation is
conventionally categorized into 3 regions: UVA
(>315-400 nm), UVB (>28Q-315 nm) and UVC
(> 100-280 nm) (Figure 1 ).
These categories have been confirmed by
the Commission lnternationale de I'Eclairage
(CIE, 1987), although there is variation in usage.
In the medical and biological fields, for example,
320 nm is used as the limit between UVA and
UVB. More recently, it was proposed to
distinguish between UVA-1 (>340-400 nm) and
UVA-2 (320-340 nm).
Units and measurements of UV radiation
Measurement of ambient solar UV radiation
Measurement of ambient solar UV radiation has
been performed worldwide for many years.
However, UV radiation detectors for research or
individual use have been developed only recently.
There are two principal types of instruments:
steady spectroradiometers, which screen the
entirety of the UV spectrum (1 00-400 nm) within
a few minutes, and broad-spectrum dosimeters,
which can measure solar irradiance within a few
seconds. Individual dosimeters, which can easily
be placed at strategic places on individuals, are
of the second type.
Broad-spectrum instruments often include a
weighting factor representative of a given
biological spectrum (e.g. skin erythema). In
current practice, the margin of error for the
measurement is relatively high, around 30%.
The biologically relevant UV radiation dose at
a given wavelength corresponds to the measured
UV radiation multiplied by a weighting factor
specific to the biological endpoint considered
(e.g. erythema, pigmentation, carcinogenesis,
etc.) at that wavelength. For the overall dose (Eeff
Figure 1. Ultraviolet (UV) region of the electromagnetic spectrum
---- Vacuum UV
________ u_v_c _______ _.
uvs u ~ ~
Extreme UV
10 100
Adapted from IARC (1992)
FarUV Middle UV
Wavelength (mn)
315 400
Near UV
300 380
Exposure to Artificial UV Radiation and Skin Cancer
Table 1. Specifications of relative erythemal
Wavelength (>..; nm)
A.< 298
298 <A< 328
328 <A::;; 400
Relative erythemal
(SA.) (weighting factor)
1 0
1 0
From McKinlay & Diffey (1987); International Electrotech-
nical Commission (1989)
expressed in watts per square meter (W.m-2)),
the weighted components are added for all the
wavelengths included in the interval considered.
The specifications of the relative erythemal
effectiveness are defined by the parameters
described in Table 1 .
Standard erythemal dose (SED) and minimal
erythemal dose (MED)
The standard erythemal dose (SED) is a
measure of UV radiation equivalent to an efficient
erythemal exposure of 100 joules per square
meter (J.m-
The clinically observed minimal erythemal
dose (MED) is defined as the minimal amount of
energy required to produce a qualifying
erythemal response, usually after 24h. The
erythemal responses that qualify can be either
just-perceptible reddening or uniform redness
with clearly demarcated borders, depending on
the criterion adopted by the observer.
Since 1997, the Erythemal Efficacy Spectrum
of human skin has become an International
Organization for Standardization/International
Commission on Illumination (ISO/CIE) standard
that allows, by integration with the emission
spectrum of any UV source, calculation of the
erythemal output of this source.
The UV index is a tool designed for communication
with the general public. It is the result of a common
effort between the World Health Organization
(WHO), the United Nations Environment
Programme (UNEP), the World Meteorological
Organization and the International Commission
on Non-Ionising Radiation Protection (ICNIRP),
and is standardized by ISO/CIE. The UV index
expresses the erythemal power of the sun: UV
index = 40 x Eett W.m-
(Table 2).
Limit values
The American Conference of Governmental
Industrial Hygienists (ACGIH) and ICNIRP have
determined the maximal daily dose that a worker
exposed to UV would be able to receive without
acute or long-term effects on the eyes. This dose
has been established at 30 J.m
(eft), which cor-
responds to a little less than 1/3 of SED. The
value takes into account an average DNA repair
capacity in the cells.
There are currently no recommendations for
safe doses for human skin.
Sources of natural and artificial UV
Solar radiation
The sun is the main source of exposure to UV for
most individuals. Sunlight consists of visible light
(40Q-700 nm), infrared radiation (>700 nm) and
UV radiation. The quality (spectrum) and quantity
(intensity) of sunlight are modified during its pas-
sage through the atmosphere. The stratosphere
stops almost all UV radiation <290 nm (UVC) as
well as a large proportion of UVB (70-90%).
Therefore, at ground level, UV radiation
represents about 5% of solar energy, and the
radiation spectrum is between 290 and 400 nm.
An individual's level of exposure to UV varies
with latitude, altitude, time of year, time of day,
clouding of the sky and other atmospheric com-
ponents such as air pollution.
Artificial UV radiation
Artificial sources of UV radiation emit a spectrum
of wavelengths specific to each source. Sources
of artificial UV radiation include various lamps
used in medicine, industry, business and
research, and for domestic and cosmetic purposes.
Physical characteristics and sources of exposure to artificial UV radiation
(Pjrek et a/., 2004), sleep disorders and the
behavioural/activity disorders in dementia
(Skjerve et a/., 2004). The light boxes used for
such treatment can emit light levels up to approxi-
mately 10,000 lux (Pjrek et a/., 2004; Skjerve et
a/., 2004), an intensity 5 to 10 times lower than
that of bright sunlight. The emission spectrum is
variable, and some lamps may contain a small
but non-negligible proportion of UVA and UVB
(Reme et a/., 1996), which however is largely
inferior to that of indoor tanning appliances. It is
noteworthy that the UV component of the light
emitted is not involved in the therapy.
(c) Occupational exposures: Artificial sources of
UV are used in many different ways in the
working environment: some examples include
welding, industrial photoprocesses (e.g. polymer-
ization), sterilization and disinfection (sewage
effluents, drinking water, swimming pools,
operating theatres and research laboratories), pho-
totherapy, UV photography, UV lasers, quality insur-
ance in the food industry, and discotheques. For
some occupations, the UV source is well
contained within an enclosure and, under normal
circumstances, presents no risk of exposure. In
other applications, workers are exposed to some
radiations, usually by reflection or scattering from
adjacent surfaces. Of relevance, indoor tanning
facilities may comprise 20 or more UVA tanning
appliances, thus potentially exposing operators to
high levels (>20W/m
) of UVA radiation (Diffey,
Comparison of UV spectrum from sunlight
and from tanning appliances
During a sunny day on the Mediterranean coast,
the solar UV spectrum at noon contains 4-5% of
UVB and 95-96% of UVA.
When UV output is calculated in terms of
biological activity, as estimated by the erythema-
effective irradiance, the emission of many tanning
appliances is equivalent to or exceeds the emis-
sion of the midday sun in the Mediterranean
(Wester eta/. , 1999; Gerber eta/., 2002). The UV
intensity of powerful tanning units may be 10 to
15 times higher than that of the midday sun
(Gerber et al., 2002) , leading to UVA doses per
unit of time received by the skin during a typical
tanning session well above those experienced dur-
ing daily life or even sunbathing. As a result, the
annual UVA doses received by frequent indoor
tanners may be 1.2 to 4.7 times those received
from the sun, in addition to those received from the
sun (Miller eta/., 1998). This widespread repeated
exposure to high doses of UVA constitutes a new
phenomenon for human beings.
In the 1990s, regulations in some countries
(e.g. Sweden, France) limited to 1.5% the maxi-
mum proportion of UVB in the UV output of
tanning appliances. However, in practice, the UV
output and spectral characteristics of tanning
appliances vary considerably. Surveys in the
United Kingdom on tanning appliances operated
in public or commercial facilities revealed sub-
stantial differences in UV output, mainly for UVB,
for which up to 60-fold differences in output have
been observed (Wright et a/., 1996; McGinley et
a/., 1998). The proportion of UVB in total UV out-
put varied from 0.5 to 4%, and thus emission
spectra similar to that of the sun in the UVB range
were sometimes attained (Gerber eta/., 2002).
These differences are due to tanning appliance
design (e.g. type of fluorescent tubes used as
sources, materials composing filters, distance
from canopy to the skin), tanning appliance
power and tube ageing. Tanning appliances in
commercial facilities may have a greater output in
the UVB range than those used in private prem-
ises (Wright et a/., 1997). With tube ageing, the
output of fluorescent lamps decreases, and the
proportion of UVB decreases more rapidly than
that of UVA.
European and international positions
regarding artificial sources of UV radiation
Full details are given in the Appendix and are
summarized below.
Standard for appliances designed specifically
for tanning purposes
Appliances designed specifically for tanning pur-
poses are defined according to an international
standard prepared by the International
Electrotechnical Commission (IEC 60 335-2-27).
Exposure to Artificial UV Radiation and Skin Cancer
This standard was first established in 1985 and
further modified in 1990, in 1995 and in 2002. A
first amendment was added in 2004 and a
second amendment is currently being voted on
internationally. This standard regulates all
appliances sold worldwide, except for the USA
who are regulated by the Food and Drug
Administration (FDA).
Appliances emitting UV radiation must
belong to one of four types of such appliances,
determined by their wavelength spectrum and
irradiance efficiency (see Appendix for detail).
National and international scientific policies
Several national and international authorities
(ICNIRP, WHO, EUROSKIN, the National
Radiological Protection Board [United Kingdom]
and the National Toxicology Program [USA]) have
adopted explicit positions regarding the use of
UV-emitting appliances for tanning purposes.
These positions are almost invariably accompa-
nied by recommendations targeting the safety of
the customers.
Regulations and recommendations by health
authorities exist in a dozen countries, predomi-
nantly in Western and Northern Europe and the
USA. Details of the regulations for each country
are given in the Appendix.
Biological effects of exposure to UV radiation relevant to carcinogenesis
A large body of literature documents the effects
of UV radiation on different living organisms,
including humans, animals and bacteria.
Experimental as well as epidemiological data
strongly indicate that the spectrum of UV
radiation reaching the Earth's surface is involved
in the development of melanoma (IARC, 1992).
The biological effects of exposure to UV
radiation were described in detail in an IARC
Monograph on UV radiation (IARC, 1992), and
the molecular effects in recent review articles
(Griffiths eta/., 1998; Pfeifer et at., 2005). In this
section, we summarize the aspects most relevant
to the understanding of the biological issues
associated with exposure to artificial sources of
UV radiation.
Biological lesions induced by UVA and UVB
DNA damage
(a) Experimental systems: UVB is a complete
carcinogen that is absorbed by DNA and can
directly damage DNA. DNA damage induced
by UVB irradiation typically includes the
formation of cyclobutane pyrimidine dimers
(CPD) and 6-4 photoproducts (6-4P). If repair
mechanisms fail to restore genomic integrity,
mutations are likely to occur and persist through
subsequent cell divisions. These mutations are
C ~ T and CC ~ TT transversions, commonly
referred to as "UVB fingerprint" or "UVB
signature" mutations. UVB can also induce the
formation of singlet oxygen species (Oi), an
oxidative compound that is highly reactive and
can cause DNA damage indirectly (Griffiths eta/.,
UVA is not readily absorbed by DNA and thus
has no direct impact on DNA. Instead, UVA
induces DNA damage indirectly through the
absorption of UVA photons by other cellular
structures (chromophores), with formation of
reactive oxygen species (such as singlet oxygen
and hydrogen peroxide [H
]) that can transfer
the UVA energy to DNA via mutagenic oxidative
intermediates such as 8-hydroxydeoxyguanosine
(8-0HdG). DNA damage by UVA radiation typi-
cally consists ofT transversions, called "UVA
fingerprint" or "UVA signature" lesions (Dobretsky
et at., 1995).
One study in hamster fibroblasts showed that
UVB produces numerous immediate mutations,
whereas UVA produces fewer immediate muta-
tions and more delayed mutations than UVB
(Dahle & Kvam, 2003).
(b) Effects on humans: The mutagenic properties
of UVA in humans have been confirmed in several
studies (Robert et at. , 1996; see Pfeifer et a/.,
2005; Halliday, 2005 for reviews). The possibility
that indirect DNA damage induced by UVA could
play a major role in melanoma occurrence is
underlined by reports of multiple cutaneous
melanomas developing in patients genetically
highly susceptible to oxidative agents (Pavel et
at., 2003).
Experiments in human volunteers conducted
during the last decade have shown that commer-
cial tanning lamps produce the types of DNA
damage associated with photocarcinogenesis in
human cells. Volunteers whose skin was exposed
to UVA lamps used in tanning appliances show
DNA damage, p53 mutations induced by oxida-
tive damage, and alterations of the p53 protein
similar to those observed after sun exposure or
after UV exposure of experimental animals
(Woollens et a/., 1997; Whitmore et a/., 2001;
Persson et at., 2002).
Studies in humans show that a pre-vacation
artificially-induced tan offers little or no protection
against sun-induced DNA damage (Hemminki et
a/., 1999; Bykov eta/., 2001; Ruegemer eta/., 2002).
Cell damage
UVA and UVB radiation can cause cell damage
through different mechanisms: both UVA and
UVB lead to differential expression of p53 and
Exposure to Artificial UV Radiation and Skin Cancer
bcl-2 proteins, which may play an important role
in regulating UV-induced apoptosis (Wang eta/. ,
1998). DNA repair and apoptosis protect the
cell's integrity against UV-induced damage. One
study conducted in cells from medaka fish sug-
gested that different apoptotic pathways exist
depending on the wavelength, i.e. for long- (UVA)
and for short- (UVB or UVC) wavelength radia-
tions (Nishigaki et al., 1999). Irradiation of
melanocytes with UVA or UVB leads to alter-
ations of different intracellular proteins, suggesting
that UVA and UVB may induce initiation of
melanoma via separate intracellular pathways
(Zhang & Rosdahl, 2003).
UVA, UVB and human skin
In humans UVA penetrates deeper into the skin
than does UVB. Because UVA represents the
majority of the UV spectrum of tanning appli-
ances and of solar radiation reaching the Earth's
surface, far more UVA than UVB reaches the
basal layers of the epidermis, where skin
keratinocytic stem cells and melanocytes are
located. DNA analysis of human squamous cell
carcinoma (SCC) and solar keratosis showed
that UVA fingerprint mutations are mostly detect-
ed in the basal germinative layer of these lesions,
whereas UVB fingerprint mutations are found
predominantly more superficially in these lesions
(Agar eta/., 2004).
Differential effects of UVA and UVB on skin
Experimental systems
Several studies showed that UVA could induce
squamous cell cancers in nude mice, but the abil-
ity of UVA alone (without exogenous photosensi-
tizers such as those used in PUVA therapy -
see Page 41) to induce squamous cell skin can-
cers was about 5000 to 1 0000 times lower than
that of UVB alone {IARC, 1992; de Laat et al.,
1997; Griffiths et al., 1998). Both in-vitro experi-
ments and epidemiological studies have demon-
strated that long-lasting, chronic exposure to
UVB is the main cause of sec of the skin (see
IARC, 1992; Brash et al., 1996 for reviews).
Accordingly, before 1990, only UVB, and not
UVA, was considered to be carcinogenic.
In the 1990s, studies in newborn rodents and
on human foreskin grafted on immunosup-
pressed nude mice have provided compell ing
evidence that high UVB doses were required in
the genesis of melanoma or of melanocytic
tumours considered to be precursor lesions of
melanoma (Mintz & Silvers, 1993; Atillasoy eta/.,
1998; Robinson eta!., 1998; Sauter eta/., 1998;
Robinson et al. , 2000a; Noonan eta/., 2001; van
Schanke et al. , 2005). At the same time, several
in-vivo studies showed that UVA can induce
melanoma in backcross hybrids of freshwater
fishes of the genus Xiphophorus (platyfish and
swordtail; Setlow et a/. , 1993) and melanocytic
tumours in the South American opossum
Monodelphis domestica (Ley, 1997, 2001 ).
However, UVA was less efficient than UVB for the
induction of melanocytic tumours in Monodelphis
domestica (Ley 2001 ), and experiments with UVA
on newborn rodents and on human foreskin could
not reproduce the results obtained with UVB
(Robinson eta/., 2000b; Berking et al. , 2002; de
Fabo et al., 2004; van Schanke eta/. , 2005).
Other studies showed that radiation emitted
by lamps used in tanning appliances (mainly
UVA) could significantly increase the carcino-
genic effect of broad-spectrum UV radiation
(Bech-Thomsen et a/., 1991 , 1992), indicating
the possibil ity of a complex interplay between
UVA and UVB radiation in human skin.
Relevance of experimental data to human
skin cancers
To date, evidence obtained from experimental
studies on the involvement of high UVB doses in
the causation of sec is consistent with observa-
tions in humans. In contrast, experimental studies
provide conflicting results on an implication of
UVB and UVA in the induction of melanoma in
humans. The same uncertainties hold true for
basal cell carcinoma (BCC), a type of tumour that
shares many of the epidemiological characteris-
tics of melanoma.
The relevance of animal models for elucidating
the biological mechanisms involved in the
development of melanoma and BCC remains
Biological effects of exposure to UV radiation relevant to carcinogenesis
questionable, as even engineered mice with
multiple deficiencies in key genes involved in cell
cycle regulation and growth factor synthesis do
not represent a model equivalent to the human
skin. In addition, experiments on animals cannot
reproduce the complex relationship existing in
individuals between highly variable natural sus-
ceptibilities to UV radiation, different sun exposure
behaviours, and exposure to various sources of
UV radiation. In the case of indoor tanning, such
relationships may be critical, as users are more
inclined than the average population to engage in
outdoor tanning activities (Autier et al., 1991 ), and
indoor tanning sessions often precede or follow
active sun exposure or outdoor tanning.
Changes in immune response
Several reports (IARC, 1992, 2001; Ullrich, 2005)
have extensively reviewed the studies on the
effects of UV on the immune system and of the
underlying mechanisms. This section only refers
to studies relevant to UVA and use of indoor
tanning facilities.
Experimental systems
Both UVA and UVB radiation can affect the
immune response that may be involved in the
promotion of melanoma (Kripke, 1974; Singh et
al. , 1995), but the two types of radiation seem to
act differently. UVB can induce immune suppres-
sion at both local and systemic levels whereas
UVA does not induce systemic immune suppres-
sion. However, studies have shown that a number
of local responses induced by UVB radiation on
the skin could be suppressed by a UVB filter, but
the melanoma growth stimulation effect could not
be suppressed (Donawho eta/., 1994; Wolf et al.,
1994). This result suggests that UVA may influ-
ence local immune responses different from
those influenced by UVB.
Studies in humans
Observations in human volunteers have
demonstrated that UV exposure suppresses the
induction of immunity (Cooper eta/., 1992; Tie et
al., 1995; Kelly et a/., 1998). Few studies have
specifically investigated the effects of exposure to
tanning appliances on the systemic and local
immune systems. UV lamps similar to those used
in tanning appliances are used without concomi-
tant use of photosensitizer for treating skin
conditions such as dermatitis and sun allergies,
illustrating the effect of that radiation spectrum on
the skin immune system.
Studies in volunteers have shown that expo-
sure to tanning appliances induces reductions in
blood lymphocyte counts, changes in proportion
of lymphocyte subpopulations, immune response
to known carcinogens applied to the skin, and
changes in the skin immune system (Hersey et
al., 1983, 1988; Rivers et al. , 1989; Clingen et al.,
2001 ). These studies also indicated that UVA and
UVB would affect the immune system via inter-
acting and overlapping mechanisms, depending
on the amount of UVA and UVB emitted (Ciingen
et a/., 2001 ), which would then lead to the
suppression of known immune reactions
(Nghiem et al., 2001, 2002). Hence, these stud-
ies indicate that UVA can suppress established
immune reactions at the skin level, but it remains
to be established how these effects relate to the
induction of neoplastic processes.
Effects of natural and artificial UV radiation
on human skin
Variety of skin types
There is a considerable range of susceptibility of
the human skin to the carcinogenic effects of UV
radiation, and in humans, there is an estimated
1 000-fold variability in DNA repair capacity after
UV exposure (Hemminki et a/., 2001 ).
Susceptibil ity to sun-induced skin damage is
closely related to pigmentary traits, and subjects
having the following characteristics are at
increased risk for developing a skin cancer
(melanoma, SCC and BCC):
Red hair, followed by blond hair, followed by
light brown hair.
Skin phototype (Fitzpatrick, 1988): subjects
who always burn and never tan when going
Exposure to Artificial UV Radiation and Skin Cancer
unprotected in the sun (skin phototype I) have
a much higher risk for skin cancer than sub-
jects who never burn and always develop a
deep tan (skin phototype IV). Intermediate
risk categories are subjects who always burn
then develop a light tan (skin phototype II),
and subjects who sometimes burn and always
develop a tan (skin phototype Ill). Subjects of
skin phototypes V and VI belong to popula-
tions with natural brown or black skin, and are
resistant to sunlight.
Freckles (ephelides) on the face, arms or
shoulders. The skin cancer risk increases with
increasing sensitivity to freckling.
Skin colour: pale colour, followed by
increasing depth of pigmentation.
Eye colour: blue, followed by grey/green eyes,
then by brown eyes.
Subjects with red hair, many freckles and
who never tan are at particularly high risk for skin
Sunburn is the occurrence of painful erythemal
reaction after exposure to UV radiation. Sunburn
during childhood or during adulthood is a risk fac-
tor for melanoma, and the risk increases with
increasing number of sunburns (IARC, 1992).
Skin erythema or sunburns are reported by
18- 55% of users of indoor tanning facilities in
Europe and North America (reviewed in Autier,
2004). Although UVB is more potent than UVA for
triggering sunburn, high fluxes of UVA are capa-
ble of inducing skin erythemal reactions after 1 0
to 20 minutes in subjects susceptible to sunlight
and having moderate tanning ability (Fitzpatrick
skin phototype II}.
Tan acquisition
The production of melanin (tanning) accounts for
part of the protection against UV radiation, but
there is mounting scientific evidence that faculta-
tive tan is triggered by UV-induced DNA damage
in the skin (Pedeux eta/., 1998; Gilchrest & Eller
1999 for a review}. Facultative tanning is now
considered a better indicator of inducible DNA
repair capacity than of efficient photoprotective
skin reaction. Inducible DNA repair capacity
rather than pigmentation itself could result in the
lower incidence of skin cancer observed in
darker-skinned individuals (Young et a/. , 1998;
Agar & Young, 2005; Bohm eta/., 2005}.
In subjects who tan easily, exposure to
tanning appliances will first lead to the oxidation
of melanin already present in superficial
keratinocytic layers of the skin (i.e. immediate
pigment darkening [IPD]). IPD is essentially trig-
gered by UVA (Young, 2004). It develops rapidly
after exposure during an indoor tanning session,
and fades away after a few hours. A more
permanent tan is acquired with accumulation of
exposure, depending on tanning ability and on
the amount of UVB present in the UV spectrum of
the lamps. The permanent tan conferred by
"UVA-tanning" has a uniform and less deep
brown appearance than the tan acquired in the
IPD has no photoprotective effect against
UV-induced erythema (Black et a/., 1985}. A
UVA-induced permanent tan provides practically
no photoprotection either (Gange et a/. , 1985;
Rivers et a/., 1989), and UVA-induced moderate
skin thickening would afford even less photopro-
tection than tanning (Seehan eta/., 1998).
Prevalence of exposure to artificial UV radiation for tanning purposes
The indoor tanning industry developed in Europe
and the USA in the early 1980s, a time when UVA
radiation was thought to be harmless, with the
introduction of tanning applances emitting UVA at
levels similar to or even exceeding those from nat-
ural sunlight. In the USA, indoor tanning is now a
more than $5 billion industry that employs
160,000 persons (Indoor Tanning Association,
2004), and in the United Kingdom the turnover in
the indoor tanning industry exceeds an estimated
100 million per annum (source: www.ray-
watch.co.uk; accessed on 15/06/2005).
Prevalence of exposure by region/country
Indoor tanning is a widespread practice in most
developed countries, particularly in Northern
Europe and the USA, and is gaining popularity
even in sunny countries like Australia.
Few surveys have estimated specifically the
prevalence of indoor tanning among adult popu-
lations. In 1996, a telephone survey was carried
out among white adults (18 to 60 years old) from
the two most densely populated regions
(Montreal and Quebec) of the Province of
Quebec, Canada (Rhainds et a/., 1999). Of the
1 003 respondents, 20% reported having used a
tanning appliance in a commercial tanning facility
at least once during the last 5 years before the
survey. The prevalence of use during the last 12
months before the study was 11 %.
Recently, a brief report describing prevalence
of indoor tanning in Minnesota, USA, derived
from a telephone interview (45% response rate)
concerning quality of life, employment and health
of 802 randomly selected adults, showed that in
2002, 38% of adults had ever used indoor
tanning facilities (Lazovich eta/., 2005).
The prevalence of use of indoor tanning facil-
ities can be estimated from the proportion of
exposed controls in population-based case-con-
trol studies on risk factors for melanoma and
basal and squamous cell skin cancers (Table 3).
The prevalence varies greatly with country,
gender and age. Prevalence of ever having used
indoor tanning facilities ranges from 5% in
Northern Italy to 87% in Swedish women, and is
currently very high in Northern European coun-
tries, particularly in Sweden and the Netherlands.
Prevalence of exposure to tanning appliances
may still be low in some European countries or
populations. In a survey conducted among
33,021 adults older than 30 years attending
health check-up centres in France, only 2% of
subjects reported use of indoor tanning facilities
(Stoebner-Delbarre eta/., 2001 ).
Time trends
The prevalence of indoor tanning is currently
increasing in many countries, and current avail-
able estimates may therefore be rapidly outdated.
In studies conducted approximately 20 years
ago, the practice of indoor tanning was generally
low: 7% in Germany, 18% in Denmark.
Prevalence of exposure to tanning appliances by
the controls included in case-control studies is
higher in the most recent studies than in studies
conducted before 1990 (Table 3).
A survey in Minnesota (Lazovich eta/., 2005)
indicated that prevalence of use has increased
over the last decades. Few men and women had
used a tanning appliance before 1980. Women
were almost twice as likely as men to report
tanning indoors during the 1980s (19% versus
1 0%), but in the following decade, the proportion
of men using indoor tanning facilities approached
that of women (15% versus 17% in the 1990s).
The fact that the prevalence of indoor tanning
has increased during the 1990s can be demon-
strated by comparing prevalence of use as
reported in studies conducted by the same inves-
tigators in the same countries at intervals of
several years.
A case-control study conducted in 1991 in
five centres in Belgium, France and Germany
Table 3. Prevalence of use of indoor tanning facilities by popul ation controls from epidemiological studies
Reference Location Inclusive years Disease Type of No. of Source of c ontrols Age range Prevalence of ever use (il
of recruitment Study controls (years)
Number %
Holman el a/. Western Australia 1980 1981 M Case-control 511 Population, electoral roll, NR NR NR
(1986) matched on age, sex ~
Osterlind et a/. East Denmark Oct. 1982 Mar. 1985 M Case-control 926 Population, National 20-79 168 18
(1988) Population Register
Zanetti eta/. (1988) Torino, Italy May 1984 Oct. 1986 M Population, from the National Case-control 416 NR 21 5 d:
Health Service ::1
Walter et a/. (1990 Southern Ontario, Oct. 1984 Sep. 1986 M Case-control 608 Population, Property tax 20-69 109 18
and 19g9) Canada assessment rolls
Autier el a/. (1994) Germany, France, Jan. 1991 onwards M Case-control 447 Population, door to door ~ 2 120 27
Westerdahl et a/. Sweden July 1966 June 1990 M Case-control 640 Population, National 15-75 159 25 ~
(1994) Population Registry
Holly et al. (1995) San Francisco, Jan. 1961 Dec. 1966 M Case-control 452 Population, random digit 25-59 NR NR
USA telephone dialling
Bajdik el a/. (1996) Alberta, Canada 1983 1984 BCC I SCC Case-control 406 Population, health insurance 25-79 33 8. 1
plan subscriber list
Chen et a/. (1998) Connecticut, USA Jan. 1987 May 1989 M Case-control 512 Population, telephone NR 95 19
random digit dialling
Westerdahl el al. South Health Care Jan. 1995 June 1997 M Case-control 913 Population, National NR 372 4 1
(2000) region, Sweden Populati on Registry
Karagas et a/. New Hampshire, July 1993 June 1995 BCC I SCC Case-control 539 Population, Dept. of 25-74 75 14
(2002) USA Transportation, medicare
Veier0d eta/. Norway and 1991 1992 M Cohort 79616 Population, prospective 10-39 14 377
(2003) Sweden cohort
Bataille eta/. (2004) North East Aug. 1989 July 1993 M Case-control 416 Hospital and general 16-75 110 26
Thames, UK practice, excluding skin
Bataille el a/. {2005) Belgium, France, Dec. 1998 July 2001 M Case-control 622 Sweden, population-based; 18- 50 354 57
Netherlands, France & Belgium. door to
Sweden & UK door; UK & Netherlands, GP
NR. not reported; GP, general practitioner
' BCC, basal cell carcinoma; M, melanoma; SCC, squamous cell carcinoma
<: 1 time/month
Prevalence of exposure to artificial UV radiation for tanning purposes
(Autier eta!., 1994) showed that 19% of controls
had ever exposed themselves to a sunlamp or a
sunbed, this proportion being higher in Germany
(25%) than in Belgium (20%) or in France (6%) .
Of the recorded exposures, 84% had started
after 1979. In a more recent case-control study
conducted by the same investigators between
1998 and 2000 in Belgium, France, Sweden, the
Netherlands and the United Kingdom among
persons younger than 50 years (mean age of
controls, 37 years), 57% of controls had ever
exposed themselves to artificial UV tanning, with
the highest prevalence of use being found in
Sweden (87%) (Bataille eta!., 2005).
According to two studies conducted within the
same population in the south of Sweden in
1988-1990 and in 1995- 1997, the prevalence of
exposure doubled in 7 years. In 1988-1990, 46%
of individuals younger than 30 years had ever
exposed themselves to sun lamps or solaria (56%
of women and 12% of men, these figures being
higher in the group aged 15-24 years) while this
proportion was only 24% among individuals older
than 30 years (31% of women and 16% of
men)(Westerdahl et a!., 1994). After 1995, the
prevalence of solarium use in the population aged
16-80 years was 41 %, but 70% of women and
50% of men aged 18-50 years reported having
ever used a solarium (Westerdahl eta!., 2000).
Personal characteristics of adult users
Use of indoor tanning facilities is more prevalent
among women, particularly among younger age
groups and in Northern countries.
A survey of tanning appliances in commercial
use in Scotland was conducted in 1997 to measure
the spectal irradiance of the different models and
compare this irradiance with UV doses received
during sunbathing (McGinley eta!., 1998). As part
of the study, a questionnaire was distributed to
sunbed users, seeking information about their age,
sex, skin type, frequency of use, attitudes and rea-
sons for use. A total of 205 questionnaires were
collected. The majority of users were women (170
versus 35 men).
A significantly higher proportion of women
and young people (18-34 years old) was found
among tanning bed users in the Montreal-
Quebec survey (Rhainds et a!. , 1999). In the
Minnesota survey (Lazovich eta!., 2005), indoor
tanning was also more prevalent among women
than among men: 45% versus 30%. Among
users, the median number of times used was 1 0
for men and 20 for women (range, 1-600), and
21% of women reported frequent use (defined as
more than 30 times) .
In Europe, a recent case-control study found
use of indoor tanning facilities to be more preva-
lent among women (61 %) than among men
(43%) (Bataille et al., 2005). Another recent
survey explored exposure to tanning appliances
and sun exposure behaviour in a cohort of adult
volunteers. In 2001, a self-administered question-
naire was specifically developed and addressed
to 12 741 adult volunteers in France enrolled in
the SU.VI.MAX cohort (a cohort recruited in 1994
and followed for 8 years, which included men
aged 45-60 years and women aged 35-60
years). Over 60% of the questionnaires were
returned, of which 97% were useable. Among the
7 359 individuals who answered the question-
naire, 1 179 (16%) - 953 women (22%) and 226
men (8%) - reported having ever experienced
indoor tanning. Men and women reported similar
prevalences for regular use (6% and 7%, respec-
tively) and for a duration of at least five years
(1 0% for men and women). Among women, 44%
of users belonged to the youngest age group at
recruitment (35-44 years), versus 33% in non-
users (in rnen, data were not available for this age
group); 48% of female users lived in the North or
in lle-de-France, versus 39% of non-users (45%
and 36% for men, respectively) (Ezzedine eta!. ,
2005) (Table 4).
Bataille et a!. (2005) recently observed that
indoor tanning is becoming more frequent in men
and in younger age groups, with important varia-
tions by country: exposure of men is highest in
Sweden (78%) and Netherlands (60%), while
39% of men in the United Kingdom and 13% in
France reported ever having used indoor tanning
Exposure to Artificial UV Radiation and Skin Cancer
Table 4. Lifetime use of indoor tanning facilities and sun exposure behaviour among 7 359
healthy adults (SU.VI.MAX cohort)
Women Men
Use of indoor tanning facilities Users Non-users Users Non-users
N = 953 (22%) N = 226 (8%)
Regular use 7%
5 years 10%
Residence North of France or lle-de-France 48%
Sunbathing between 11 a.m. and 4 p.m. 56%
Regular sunscreen use during sunbathing 39%
Progressive sun exposure 54%
Nudism 13%
Sunburns in adulthood 93%
Important or extreme tan seeking behaviour 37%
From Ezzedine et at. (2005)
Younger age (<35 years) is significantly associated
with higher likelihood of using indoor tanning
facilities among both men and women.
In an early case-control study conducted in
several countries in Europe (Autier eta/., 1994),
indoor tanning was more prevalent in younger
age groups (31 % among controls < 40 years). In
a more recent case-control study in Europe
(Bataille eta/., 2005), exposure before the age of
15 years was reported in 3% of all controls, but
reached 20% in Sweden. The mean age at first
exposure was 20 years in Sweden, 23 years in
the United Kingdom and 27 years in France.
In the survey conducted in Scotland (McGinley
et a/. , 1998), 73% of users were under 35 years
old, with 32% of users being under 25 years old.
In the Minnesota survey (Lazovich et a/.,
2005), 13% of men and 22% of women reported
first tanning indoors as adolescents.
Skin type
Few studies have analysed specifically the use of
indoor tanning facilities as a function of skin type.
Since most studies have been conducted primari-
ly in relation to skin cancer risk factors, use by skin
type cannot be derived from the reported results.
In the survey conducted in Scotland
(McGinley et a/., 1998), 38% of users described
45% 36%
37% 53% 38%
24% 17% 7%
43% 53% 38%
6% 19% 8%
88% 93% 89%
20% 26% 11 %
their skin phototype as type I or II , and 38% also
indicated that they had experienced an adverse
reaction when using indoor tanning facilities; 31%
of users had more than 1 0 courses of over 5 ses-
sions in a year, and for 16% this amounted to
over 1 00 sessions per year.
In several case--control studies, use of indoor
tanning facilities was more frequent among
controls with a poor ability to tan: for example, 27%
and 31% among controls with blond or red hair,
respectively, in a European study (Autier eta/., 1994).
In the SU.VI.MAX cohort, individuals wi th a
pale complexion were more likely to use indoor
tanning facilities (Ezzedine eta/. , 2005). This was
not the case among controls from a recent
case-control study conducted in Europe, where
approximately one third of controls using indoor
tanning facilities were of phototype I or II (Bataille
et a/., 2005) (Table 5). However, it must be
stressed that in this study, phototype was declared
by participants and it is likely that few of them
perceived themselves as sun-sensitive, as exem-
plified by the very low proportion of persons with
self-reported phototype I in the Swedish popula-
Other factors
Higher education levels or income are significantly
associated with a higher li kelihood of using
indoor tanning facilities among men.
Prevalence of exposure to artificial UV radiation for tanning purposes
Table 5. Prevalence of indoor tanning according
to skin type among controls in a European
case-control study (Bataille et a/., 2005)
Country Phototype (%)
Belgium 13.3 23.3 43.3 20.0
France 6.4 38.7 25.8 29.0
Sweden 1.2 24.7 64.2 9.8
The Netherlands 6.0 17.9 53.0 23.1
United Kingdom 12.7 32.1 39.0 6.9
Data courtesy of V. Bataille.
The most common reasons given for use of
indoor tanning faci lities is to develop a "base tan"
before a holiday and to feel more relaxed
(McGinley eta/. , 1998).
In the SU.VI.MAX survey, the most frequently
report ed motivations tor using artificial tanning
were aesthetic (35%) and skin preparation before
sun exposure (34%) (Ezzedine et a/. , 2005) . In
this cohort, there was a clear link between use of
indoor tanning faci lities and sun-seeking
behaviour (Table 4).
Personal characteristics of adolescent and
children users
Since 1989, a total of 16 studies (18 reports)
have examined indoor tanning among children
and adolescents aged 8-19 years. These studies
are summarised in Table 6 (see Lazovich &
Forster, 2005 for review). Studies were conducted
in Europe (Norway, Sweden and the United
Ki ngdom) , in various locations throughout the
USA (including two nationally representative
samples) and in Australia. Adolescents were
identified through paediatric clinics, schools, as
offspring of adult cohort study participants, or
through random selection of defined populations.
Sample size ranged from 96 to over 15,000. Use
of indoor tanning facilities was defined either as
ever use, or use in the past 6 or 12 months. Given
the differences in the study populations and in
the definition of indoor tanning between studies,
it is not surprising that prevalence estimates vary
greatly. However, all these studies show frequent
use by adolescents and chi ldren, sometimes at a
very young age. According to the most recent
studies, 30% of adolescents in Sweden and 24%
of adolescents in the USA aged 13-19 years
reported ever use of indoor tanning faci lities, and
8% and 12% respectively were frequent users
(1 0 times per year or more). In a recent survey in
the United Kingdom, while 7% of children aged
8- 11 years reported exposure to a sunbed in the
last 6 months, as many as 48% expressed a
desire to use a sunbed (Hamlet & Kennedy,
2004) .
The earl iest studies in Sweden and in the
USA tended to find indoor tanning to be more
prevalent among adolescents with fair skin types
who are more prone to sunburn (Mermelstein &
Riesenberg, 1992; Boldeman et a/., 1996;
Robinson eta/. , 1997). More recent studies in the
USA found either the opposite (Cokkinides eta/.,
2002; Geller eta/., 2002; Demko eta/., 2003) or
no association (Lazovich eta/., 2004).
Studies of compliance to regulations and
Few studies have assessed the compliance of
indoor tanning facility operators or consumers
with recommendations and regulations. In this
section, studies are first summarised and then
data are presented according to each regulation.
Compliance of operators
(a) Study descriptions - overall compliance rates:
In 1991, Oliphant et a/. (1994) surveyed over
1 000 high school students aged 13 to 19 years in
suburban Minnesota (USA) via a self-adminis-
tered questionnaire regarding use of indoor tan-
ning facilities and knowledge about risks of
indoor tanning. The survey assessed compliance
of staff with regulations and recommendations as
reported by the users.
In 1998, Culley et a/. (2001) quantified the
level of compliance by indoor tanning facility
operators with selected federal and state regula-
tions and recommendations. A person posing as
a potential customer visited 54 tanning facilities in
Exposure to Artificial UV Radiation and Skin Cancer
Table 6. Studies of adolescent use of Indoor tanning facilities
Reference Year of Location Population N Age range Prevalence (%)
survey source (years)
assessed In relation to
Boys Girls All
use of indoor tanning
Banks eta/. 1989 Vienna, Adolescents 96 16-19 16 33 23 Gender, age, frequency
(1992) VA, USA seen at nine
Mermelstein 1990 Chicago, 10 schools 1 703 9th and 7 19 NR Gender, age, skin type
& Riesen IL, USA participating
berg (1992) in skin
cancer inter
Oliphant et 1991 St. Paul, One high 1 008 13-19 15 51 34
Gender, age, frequency,
a/. (1994) MN, USA school knowledge of risks,
practice, symptoms
Wichstrom 1992 Norway 56 randomly 15 169 17.3
752 NR
Gender, age, frequency
(1994) selected (mean)
high schools
Boldeman et 1993 Stockholm, 60 randomly 1 252 14-19
32 68 57
Gender, age, knowledge
a/. (1996, Sweden selected of risks, smoking,
1997) classes frequency, skin type,
symptoms, sunbathing,
skin disease, perceived
attractiveness, attitudes
Robinson et 1994 Chicago, Population- 658 11-19
16 8
Gender, age, skin type,
a/. (1997) IL, USA based ran- socio-economic status
dom sample
Brandberg et 1996 Sweden Population 2 615 13, 15, 17
4 16 10
Gender, age, satisfaction
a/. (1998) based ran- with self
dom sample
Boldeman et 1999 Stockholm, Population- 4 060 13-19
19 40 30
Gender, age, frequency,
a/. (2001) Sweden based ran symptoms
dom sample
Lucci eta/. 1999 Dallas & Junior and 210 14-19
NR NR 183 None
(2001) Houston, senior high
Texas, students
Cokkinides 1998
USA Population- 1 192 11-18 5 16 102 Gender, age, race,
eta/. (2002)
based ran- parent education and
dom sample income, residence, sun
sensitivity, skin type,
sunbathing, sun
protection, health-
provider advice,
attitudes, parent tans
Prevalence of exposure to artificial UV radiation for tanning purposes
Table 6 c o n t d ~
Reference Year Location Population N Age range
Prevalence (%)
of source (years) assessed in relation to
Boys Girls All
use of indoor tanning
Geller eta/. 1999 USA Prospective 10 079 12-18 2 14 102 Gender, age, skin type,
(2002) cohort of social factors, sun pro-
off-springs tection, attitudes
of Nurses
Knight et a/. 1999 Bloomington, College 489 ~ 17 38 70 62 Gender, age, frequency,
(2002) Indiana, students 402 17- 22
NR NR 522
skin type, geographical
USA attending
region, reason for using
student tanning bed, believes
health about tanning,
knowledge of risks
Demko eta/. 1996 USA 132 6 903
13-19 11 37 24 Gender, age, frequency,
(2003) schools in sun sensitivity, geogra-
80 commu- phical region, school
nities location, student income,
maternal education,
sunbathing, substance
use, diet, obesity, body
image, physical activity,
body piercing, psycho-
social factors
Hamlet & NR Wish ow 23 primary 1 405 8-11 NR NR 7 Age, frequency, atti-
Kennedy Local schools tudes, exposure at home
(2004) Health or on commercial
Care, UK premises.
Lazovich et 2000 Minneapolis/ Random 1 273
14-17 12 42 30
Gender, age, smoking,
a/. (2004); St. Paul, MN sample of
satisfaction with looks,
Stryker et a/. and Boston, households
depression, sun protec-
(2004) MA. USA likely to
tion, skin cancer risks,
parent and teen
knowledge of risks,
parent and teen atti
tudes, social factors,
parent tans, parent
education, parent con-
cern, parental influence
Paul eta/. 2000 New South Population- 1 509 ~ 15 5 14 10 Gender, age, attitudes,
(2004) Wales,
15-17 4
use of sunscreen
random 18-29 11
sample 30-39 19
Adapted from Lazovich & Forster (2004)
NR, not reported.
Prevalence of ever use, unless otherwise noted.
Past 12 months
Past 6 months
Exposure to Artificial UV Radiation and Skin Cancer
the San Diego, USA metropol itan area.
Compliance with 13 regulations/recommenda-
tions was assessed by either direct query or
observation of the presence/absence of signs and
warning labels. No facility was in compliance with
all 13 selected regulations. The mean number of
regulations complied with was 8.33.
In another study conducted in the San Diego
area, in 2000, Kwon et al. (2002) assessed the
compliance of 60 tanning facilities with recom-
mended exposure schedules by means of a tele-
phone enquiry made by a supposedly prospective
One study, conducted in Australia in 2005,
explored compliance with international recom-
mendations on solarium use in an unregulated
setting: simulated customers visited 176 solaria
in two face-to-face visits for each establishment
and one telephone contact. Few (16%) establish-
ments were compliant with more than 1 0 of the
13 recommendations. Compliance was particu-
larly poor for those recommendations with the
greatest potential for minimising harm: i.e. to dis-
courage or exclude persons at high risk from UV
exposure (Paul et al. , 2005).
(b) Duration/frequency of exposure: In the survey
assessing compliance of staff as reported by the
users (Oliphant eta/., 1994), 26% said they were
never told to limit their time per session.
In a later study from the USA (Culley et al. ,
2001 ), compliance was found to be relatively high
for maximum duration allowed to tan (98%) but was
relatively low for presence of and compliance with
an appropriate shut-off switch (57%). Frequency
allowed to tan had the lowest compliance at 6%;
one facility even allowed two consecutive tanning
In the most recent study from the USA (Kwon
et al. , 2002), only 4 out of 58 tanning salons (7%)
recommended less than 3 sessions in the first
week, and therefore were compliant with the reg-
ulations. All responded with a duration of expo-
sure of less than 30 minutes, but all reported
offering unlimited tanning packages, and less
than 30% limited the exposure to once a day.
(c) Wearing of goggles: In the high school student
survey cited above (Oliphant et al., 1994), less
than half of the customers interviewed (42%) had
always been told to wear goggles, and 28% had
never been.
In a more recent study from the USA (Culley
et al., 2001 ), compliance was found to be high for
provision and sanitation of protective eyewear
(1 00%) and for requirement to use it (89%).
(d) Age restriction: Very few studies have looked
at compliance with age restriction. One study
observed a low compliance (43%) with the
requirement for parental permission for adoles-
cent users aged 14-18 years (Culley et a/.,
2001 ). Low levels of compliance with recommen-
dations relating to age restriction were also found
in a more recent study (Paul eta!., 2005).
(e) Warning of health risks: In the survey assess-
ing compliance of staff as reported by users
(Oliphant et a/., 1994), 50% reported that they
had never received a warning about the health
risks of indoor tanning, and less than half (48%)
had ever noticed a warning sign at the facility. In
another study in the USA (Culley et a/:, 2001 ),
compliance was found to be relatively high for
presence of labels on warning of UV danger and
of exposure (85%) and legibility, accessibility and
correctness of these labels (74%); lower compli -
ance (15-20%) was observed for warning signs
in the tanning area.
(f) Other regulations: In the Australian study (Paul
et al. , 2005), 1% of operators refused access to a
pretending customer with skin phototype I, and
1 0% recommended against solarium use. In the
same study, low levels of compliance were also
found for using a sunbed while taking medica-
tions, for provision of consent forms and for
discussing safety procedures.
Compliance of customers
(a) Study descriptions: The 1991 high school stu-
dent survey in the USA (Oliphant et al. 1994) has
been described above.
McGinley eta/. (1998) conducted a survey of
the output of tanning appliances in use in 1997 in
Scotland. At the same time, questionnaires were
distributed by the indoor tanning facilities to
Prevalence of exposure to artificial UV radiation for tanning purposes
users, seeking information on patterns of expo-
sure and reasons for using sunbeds.
In 1996, a telephone survey was carried out
among adults from the two most densely popu-
lated regions of Quebec, Canada, as described
above (Rhainds et al., i 999). The final sample
included i 003 white persons i 8-60 years old.
Interviewers used a standardised questionnaire
to document exposure habits to artificial UV
radiation sources.
One study was conducted in North Carolina
(USA) to assess adherence of indoor tanning
clients to FDA-recommended exposure limits.
A community-based survey was administered
during routine state inspections of 50 indoor
tanning facilities. At each facility, users' records
were randomly selected (n = 483) for a survey of
exposure (Hornung et al., 2003).
To gain anecdotal evidence that primary
school children were using sunbeds in
Lanarkshire (United Kingdom), school nurses
conducted a short questionnaire in 23 primary
schools in 2003. Children 8-11 years old took part
in the classroom surveys. Positive responses
were counted by a show of hands by the children
(Hamlet & Kennedy, 2004). [This small study was
based on a "hands up" survey, which may have
biased answers through copying of friends'
{b) Duration/frequency of use: In the high school
student survey, 1 i% of users reported tanning
indoors for more than 30 minutes. Those who
reported longer usual tanning sessions were
more likely to tan frequently (Oliphant et a/.,
A user survey demonstrated that 31% of 205
responders had more than 1 0 courses of over
five tanning sessions in a yeaF and, for 16% of
them, this amounted to over i 00 sessions per
year (McGinley eta/., 1998).
In the study by Hornung et al. (2003), out of
483 users, 95% were exceeding the recommended
exposure times. Also, 33% of users started their
first tanning session at or above exposure times
recommended for users in the maintenance
phase of tanning (>4.0 MED). The average dura-
tion of exposure on the first visit was 14.3 min-
utes (range, 3-30 minutes). Compilation of 15
common exposure schedules listed a suggested
range of 2- to 15-minute sessions (average, 5.76
minutes) for the first week of tanning, with
gradual increases over a 4-week or longer period
to a range of 8- to 30-minute maintenance ses-
sions (average, 20.5 minutes). The average peri-
od of tanning for each user was 6.3 weeks. Users
spent approximately 43 minutes per week (range,
5-135 minutes) during an average of 2.4 ses-
sions per week (0.25-7 sessions) (Hornung et
at., 2003).
(c) Wearing of goggles: In the i 991 study of high
school students (Oliphant et al., 1994), 59%
reported always wearing goggles and 17%
reported never wearing them. Those who reported
longer usual tanning sessions were less likely to
use goggles.
In the Scottish survey (McGinley eta/., 1998)
35% of users stated that they never or hardly ever
wore protective goggles.
In the Canadian study (Rhainds eta/., 1999),
70% of 203 tanning bed users wore protective
goggles during tanning sessions.
(d) Age restriction: In the US high school survey,
almost 20% of those aged 14 years or younger
reported using indoor tanning facilities, and half
of the users had had their first session before age
15 years (Oliphant et al. , 1994).
Among 1405 adolescents under 16 years
surveyed in the United Kingdom (Hamlet &
Kennedy, 2004), 7% had used a sunbed in the
last 6 months, of whom sixteen (17%) agreed
that they used a sunbed regularly, i.e. twice a
month or more. Of these 96 adolescent recent
users, 61 (64%) reported using a sunbed in
someone's house, and 23 (24%) had used a
sunbed in a shop or salon.
Epidemiological data on exposure to artificial UV radiation for cosmetic
purposes and skin cancers
As no valid animal model of human melanoma or
other skin cancers exists, evidence of an associ-
ation between indoor tanning facility exposure
and skin cancer must be sought predominantly
from epidemiological studies. Few studies have
addressed this topic specifically, but most skin
cancer studies have included one or more items
about use of indoor tanning facilities. We system-
atically analysed the summary statistics compiled
from the relevant studies in a meta-analysis. The
results have also been discussed qualitatively, to
allow for the large differences in study popula-
tions and study quality.
Since melanoma and other skin cancers
differ somewhat in their aetiology, studies of
melanoma were analysed separately from those
of basal and squamous cell cancers.
Epidemiological evidence from studies investigating
other sources of exposure to artificial UV
radiation has also been presented.
Methodology for literature search
The literature to April 2005 was searched using
the following databases: Pubmed, lSI Web of
Science (Science Citation Index Expanded),
Embase, Pascal, Cochrane library, Lilacs and
Medcarib. The following keywords and their cor-
responding French translations were used for
search in the PASCAL database: "skin cancer",
"squamous cell carcinoma", "SCC", "basal cell
carcinoma", "BCC", "melanoma" for diseases. To
define exposure, the following keywords were
used: "sunbed", "sunlamp" , "artificial UV",
"artificial light", "solaria", "solarium", "indoor tan-
ning", "tanning bed", "tanning parlour", "tanning
salon" and "tanning booth".
We searched for keywords in the title and in
the abstract, when available. We also performed
a manual search of references cited in the selected
articles, and in selected reviews or books on
melanoma and skin cancer. All participants of the
working group and some IARC staff were asked
to report any additional published or submitted
study. No language restriction was applied.
Primary inclusion criteria were developed for
the selection of relevant articles, which were:
case-control, cohort or cross-sectional studies
published as an original article. Ecological
studies, case reports, reviews and editorials were
not considered eligible.
For the meta-analysis, we selected the
articles fulfilling both of the following two criteria:
1 . The article contained sufficient information to
estimate the relative risk and 95% confidence
intervals (odds ratios [OR], relative risks or
crude data and corresponding standard errors,
variance, confidence intervals or P-values of
the significance of the estimates); and
2. The article reported an independent study (in
order to avoid giving additional weight to some
The selected articles were reviewed and data
abstracted by means of a standardized data-
collection protocol. When another article on the
same study was published simultaneously,
additional relevant or missing information was
retrieved from the companion paper. For each
study the following information was retrieved:
General information: year of publication,
recruitment years, study design, study loca-
tion and latitude of the region;
Exposure information: definition of type of
exposure, age at first exposure, duration of
exposure, year of exposure, place of exposure;
Case-control information: inclusion or exclusion
of specific histological types of melanoma,
number and source of cases and controls,
matching design, blinding of interviewers;
Statistical information: statistical methods
used, adjustment for confounding variables
(demographic factors such as age and sex,
Epidemiological data on exposure to artificial UV radiation for cosmetic purposes and skin cancers
baseline host characteristics such as hair, eye
and skin colour, inherent tendency to burn or
tan easily, naevi, sunburns or sun exposure)
and type of effect estimates (odds ratio,
relative risk, standardized incidence ratio)
with corresponding measures of precision,
according to specific exposure category.
The minimal common information about exposure
to indoor tanning devices for all studies was "ever
exposed". For those studies where the
definition of exposure "ever versus never exposed
to indoor tanning facilities" was not present, we
used the information closest to this category.
Since it has been suggested that age at
exposure may influence the relative risk for skin
cancer associated with UV exposure (Whiteman
eta/. , 2001 ), we extracted relative risks associated
with use of indoor tanning facilities before the age
of 35 years where available. Studies used
different age categories for classifying age at first
exposure, so odds ratios for the "young expo-
sure" category were pooled without correction.
We identified 23 studies of use of indoor tanning
facilities and melanoma (Kiepp & Magnus, 1979;
Adam eta/., 1981 ; Gallagher eta/., 1986; Holman
et a/., 1986; Holly et a/., 1987; Swerdlow et a/.,
1988; Osterl ind eta/. , 1988; Zanetti eta/., 1988;
MacKie eta/. , 1989; Beitner eta/., 1990; Walter et
a/., 1990 (and 1 999); Dunn-Lane et a/., 1993;
Garbe eta/., 1993; Westerdahl eta/., 1994; Autier
eta/., 1 994; Holly eta/., 1 995; Chen eta/., 1998;
Westerdahl eta/., 2000; Naldi eta/., 2000; Kaskel
et a/., 2001; Veiemd et a/., 2003; Bataille et a/.,
2004; Bataille et a/., 2005). All studies were
case-control studies, except for one cohort study
(Veier0d eta/., 2003). No cross-sectional studies
were identified. A case-control study was
considered population-based when cases were
derived from a population-based cancer registry
and controls selected from the general population.
Description of studies
(a) Cohort study- Veiered eta/. {2003): The only
published prospective cohort study was conducted
in Norway and Sweden, where 106 379 women
aged 30-50 years at inclusion were recruited
between 1991 and 1992. This population was
selected from the National Population Register
and followed for an average of 8.1 years. Among
these, 187 cases of invasive melanoma were
diagnosed during follow-up. The analysis was
stratified by age at the time of exposure to
sunbeds. Thirty-four cases occurred among the
14 377 women who were exposed at least once
a month during one of three age periods (1 0-19,
20-29 or 30-39 years). The corresponding risk
for melanoma for the entire cohort was 1.55 (con-
fidence interval (CI), 1.04-2.32) when adjusting
for age, region, hair colour, age-specific sunburns
and annual number of weeks of summer vaca-
tions. For the age group 20-29 years, the risk for
melanoma associated with solarium use more
than once a month compared with rarely or never
was 2.58 (CI, 1.48-4.50).
(b) Population-based case-control studies -
Adam eta/. (1981): A case-control study was
conducted in Oxford and the south-western
region of the United Kingdom between 1971 and
1976, recruiting 111 incident cases and 342 con-
trols to study the association between the oral
contraceptive and melanoma in women. Cases
were selected from two cancer registries and
when identified, were contacted through their
General Practitioner (GP) ; controls were selected
from the G P practice lists and matched to cases
for age, marital status and GP practice. Nine
cases and 1 0 controls had ever used sunlamps.
The crude odds ratio calculated [by the Working
Group] was 2.93 (CI, 1.16-7.40). [No estimate was
reported for the exposure to sunlamps. The working
group noted that 169 cases and 507 controls were
selected from the registry, but only 111 cases and
342 controls completed questionnaires.]
Holman eta/. (1986): A case-control study was
conducted in Western Australia between 1 980
and 1981 to evaluate constitutional traits, sunlight
exposure, hormones, diet and other possible risk
factors for cutaneous melanoma. This study
recruited 511 incident cases and 511 controls,
selected from the electoral roll and matched to
cases for age and sex. Past use of sunlamps was

Exposure to Artificial UV Radiation and Skin Cancer
recorded, but only 9% of subjects had used them.
The crude odds ratio for "ever use" compared to
"never use" of sunlamps was 1.1 (CI, 0.6-1.8).
Osterlind eta/. (1988}: A case-control study con-
ducted in East Denmark between October 1982
and March 1985 recruited 474 incident cases and
926 controls aged 2Q-79 years selected from the
National Population Register to study risk factors
for melanoma. Sixty-six cases and 168 controls
had ever used sunbeds, and 50% of controls had
used sunbeds less than 1 0 times. The crude odds
ratio for ever versus never use [calculated by the
Working Group] was 0.73 (CI, 0.53-1.01 ), and no
trend was observed with number of sessions.
Regarding exposure to sunlamps, 45% of cases
and 42% of controls had used sunlamps, with
40% of both cases and controls having used sun-
lamps less than 1 0 times. [No estimate was
reported for the use of a sunlamp.]
Zanetti et a/. (1988): A case-control study inves-
tigating melanoma risk factors was conducted in
Torino, Italy between May 1984 and October
1986. The authors identified 208 incident cases in
the "Registro Tumori Piemonte" registry and
selected 416 controls from National Health
Service files. Of these, 15 cases and 21 controls
had used UVA lamps for tanning purposes. The
risk for melanoma from this exposure was 0.9
(CI , 0.4-2.0) after adjustment for age, hair colour,
skin reaction, sunburn in childhood and educa-
tion level. The use of sunlamp for tanning was
very rare in Italy during the study period, and the
authors warned about the consequent lack of
power of the study.
Walter et a/. (1990): A case-control study,
designed specifically to investigate the
melanoma risk associated with artificial UV expo-
sure, was conducted in southern Ontario,
Canada between October 1984 and September
1986. Recruitment included 583 incident cases
identified from pathology reports and 608 con-
trols selected from property tax assessment rolls.
Controls were matched to cases for sex, age and
place of residence; 152 cases and 109 controls
had ever been exposed to sunlamps or sunbeds.
The risk for melanoma, adjusted for skin reaction
to initial summer exposure, was 1.54 (CI,
1.16-2.05). The relative risk in the youngest age
group (2Q-34 years) was 1.51 (CI, 0.82-2.77).
When duration of exposure to tanning appliances
was analysed by category (never; <12 months;
12 months), a significant trend was observed
both for men (p < 0.01) and for women (p = 0.04).
[This study was initially published in 1990 (Walter
et a/., 1990). Further calculations with new
adjustments were published in 1999 (Walter et
a/. , 1999).]
Westerdahl et a/. (1994}: A case-control study
was conducted in Sweden between July 1988
and June 1990. The authors recruited 400 inci-
dent cases selected from the regional tumour
registry, and 640 controls selected from the
National Population Registry, aged 15 to 75
years. Controls were matched to cases for age,
sex and place of residence. Of these, 111 cases
and 159 controls had ever used sunbeds or sun-
lamps. The relative risk, adjusted for sunburns,
hair colour, naevi number and sunbathing habits
during summer, was 1.3 (CI, 0.9- 1.8). Among
individuals aged $ 30 years, the relative risk was
2.7 (CI, 0.7-9.8). When exposure exceeded 10
sessions per year, the risk for melanoma was
significantly increased over that of never-users
(OR, 1.8; Cl, 1.0-3.2).
Holly et a/. (1995): A case-control study on
melanoma risk factors was conducted in San
Francisco, USA between January 1981 and
December 1986. The study was restricted to
women aged 25-59 years. The authors recruited
452 incident cases ascertained through the
SEER Registry for the San Francisco Bay area
and 452 controls ascertained using telephone
random digit dialling. Controls were frequency-
matched to cases for age in 5-year categories.
Exposure to sunlamps was investigated. No
association was observed for ever using a sun-
lamp (crude OR, 0.94; Cl , 0.74- 1.2). [The
Working Group noted that use of sunlamps by
63% of cases and 62% of controls, as presented
in the text, would result in an odds ratio of 1.05
(CI, 0.79-1.38). Despite this inconsistency, it was
decided to use the estimate given in the table.)
Epidemiological data on exposure to artificial UV radiation for cosmetic purposes and skin cancers
Chen et a/. (1998): A case-control study was
conducted in Connecticut, USA between January
1987 and May 1989. Using the population-based
Rapid Case Ascertainment System, 624 incident
cases were identified and 512 controls ascer-
tained using telephone random digit dialling. Of
these, 141 cases and 95 controls had ever used
a sunlamp or sunbed. The risk for melanoma
associated with sunlamp or sunbed exposure
was 1.13 (CI, 0.82-1.54) after adjustment for
age, sex, cutaneous phenotype index and recre-
ational sun exposure index. In a stratified analy-
sis, the relative risk associated with first exposure
before age 25 years was 1.35 (CI, 0.88-2.08). No
trend was observed in relation to duration of
exposure to sunlamps-or sunbeds.
Westerdahl et a/. (2000): A case-control study
was conducted in the South Health Care region
of Sweden between January 1995 and June
1997. The authors recruited 571 incident cases
identified in the regional tumour registry, and 913
controls matched for age and sex ascertained
from the National Population Registry. Of these,
250 cases and 372 controls had ever used
sunbeds. The risk for melanoma associated with
sunbed exposure was 1.2 (CI, 0.9- 1.6) after
adj ustment for age, sex, history of sunburn, hair
colour, skin type and number of raised naevi. No
change in the estimate was observed after
adjustment for sunbathing habits. In a stratified
analysis, there was a significant increase in risk
when exposure took place before the age of 35
years (OR, 2.3; Cl, 1.2-4.2). No trend relating to
total duration of exposure was observed.
(c) Hospital- or clinic-based case-control studies
Klepp & Magnus (1979): A hospital-based
case-control study was conducted in Oslo,
Norway between January 1974 and May 1975.
The authors enrolled 89 cases and 227 controls
aged 20 years or more to evaluate possible etio-
logical factors for melanoma. Cases were incident
cutaneous melanomas from the Norwegian
Radium Hospital; controls were other cancer
patients in the same hospital. The self-adminis-
tered questionnaire included a question about
use of artificial UV lamps. No estimates were
derived from the results because exposure to UV
lamp was very rare, and there was no difference
between cases and controls.
Gallagher eta/. (1986): A case-control study was
conducted in western Canada between April
1979 and March 1981. To study risk factors for
melanoma, including host factors, sun exposure,
and the use of oral contraceptive for women, 595
incidence cases from dermatology practice and
595 controls from provincial medical plans were
recruited. Controls were matched to cases for
age and sex. The recruitment was limited to indi-
viduals 20-79 years old. No estimate of the risk
was presented. The study showed no association
between sunlamp use and subsequent risk for
melanoma (X
=6.1; 5 df; p=NS), including after
stratifying by sex or by anatomical site exposed to
the sunlamp.
Holly eta/. (1987): A hospital-based case-control
study was conducted in San Francisco (USA)
between April1984 and October 1987. To assess
melanocytic naevi (dysplatic and non-dysplastic
naevi) as risk factor for melanoma, 121 incident
cases were recruited from a melanoma clinic at
the University of California, San Francisco, and
139 controls were recruited among patients in
another clinic at the same university. No estimate
of the risk for melanoma associated with sunbed
use was presented. The patients with cutaneous
melanoma were similar to those in the control
group with respect to their use of tanning salons.
Swerdlow et a/. (1988): A hospital-based case-
control study was conducted in Scotland (United
Kingdom) between 1979 and 1984 to evaluate
the role of fluorescent light and UV lamps on
cutaneous melanoma risk. The authors recruited
180 incident cases from dermatology and plastic
surgery units and 197 hospital inpatients and out-
patients as controls excluding those with malig-
nant disease. Analysis for exposure to tanning
appliances was restricted to 120 controls without
dermatological disease. Only 38 cases and 1 0
controls had ever used UV lamps or sunbeds
(crude OR, 2.94; Cl , 1.40-6.17). Data by age at
first use (before and after age of 30 years) and by
total number of hours of exposure (1-19 hours;
20 hours within the 5 years before presentation)
Exposure to Artificial UV Radiation and Skin Cancer
were also presented. A significant linear trend for
duration of use was observed (p<0.01 ).
Adjustment for hair colour, eye colour, skin type
or sun exposure did not substantially change the
estimates, while a small decrease was observed
when adjusting for number of naevi.
MacKie eta/. (1989): A hospital-based case-control
study of melanoma was conducted in Scotland,
United Kingdom in 1987. The authors identified
280 incident cases (99 men and 181 women)
through the Scottish Cancer Registry; 280 con-
trols (99 men and 181 women) were recruited at
a hospital, excluding patients with dermatological
illness. Controls were matched to cases for age
and sex. In the questionnaire, one item investi-
gated exposure to artificial UV radiation and use
of sunbeds; 33 cases and 8 controls had been
exposed to such sources. The odds ratio was
stratified by sex and adjusted for total number of
naevi, atypical naevi, freckling tendency, history
of severe sunburns, tropical residence for more
than 5 years and skin type. The adjusted odds
ratios were 1.3 (CI, Q.2-7.9) for men and 1.2 (CI,
0.5-3.0) for women. Only 26 cases and 6 controls
had used "modern sunbeds" once or twice weekly
for at least 12 weeks. [Due to stratification by sex,
two estimates from this study were used in the
Beitner et al. (1990): A case-control study was
conducted in Stockholm, Sweden between
February 1978 and December 1983. The authors
recruited 523 incident cases from the
Department of Oncology at Karolinska Hospital
and 505 controls selected from population reg-
istries. Controls were matched to cases for age
and sex. No estimate of the risk was presented.
No increase in the risk for developing cutaneous
malignant melanoma was associated with fre-
quent exposures to solaria.
Dunn-Lane et a/. (1993): A hospital-based
case-control study was conducted in Dublin,
Ireland between 1985 and 1986. The authors
recruited 100 incident cases from seven Dublin
hospitals and 100 controls, admitted for limb
injuries in the accident and emergency and
orthopaedic departments, were recruited.
Controls were matched to cases for age (within 5
years), sex and health broad area of residence.
Seventeen cases and 15 controls had ever used
sunbeds. The crude odds ratio [calculated by the
Working Group] was 1.16 (CI, 0.54-2.47). [No
estimates were reported by the authors.]
Garbe eta/. {1993): A hospital-based case-con-
trol study evaluating risk factors for melanoma
was conducted in Germany between 1984 and
1987. The authors studied 856 cases selected
from the Central Malignant Melanoma Registry of
the German Dermatology Society and 705
controls selected from outpatients presenting at
dermatology clinics. Of these, 66 cases and 50
controls had ever used sunbeds. The relative risk
for melanoma, adjusted for number of naevi, hair
colour, skin type, age and study centre, was 1.5
(CI, 0.9-2.4). [The Working Group noted that the
Central Malignant Melanoma Registry is a volun-
tary registry.]
Autier et a/. (1994): A case-control study of
melanoma was conducted in Europe (Germany,
France, Belgium) from January 1991 onwards.
The authors recruited 420 incident cases from
dermatology practices and cancer centres; 447
controls were selected from neighbourhood by
door-knock. Of these, 11 0 cases and 120 con-
trols had ever been exposed to sunlamps or
sunbeds. While there was no crude association
with melanoma (OR, 0.97; Cl, 0.71-1.32) , in a
stratified analysis total exposure to sunlamp or
sun bed for tanning purposes for more than 1 0
hours and before 1980 showed an increased risk
(OR, 2.12; Cl, 0.84-5.37) after adjustment for
age, sex, hair colour and number of holiday
weeks per year. The risk for melanoma associated
with sunlamp or sunbed use was significantly
increased if exposures for more than 10 hours
were accompanied by a burn to the skin (OR,
7.35; Cl, 1.67-32.3).
Naldi eta/. (2000}: A hospital-based case-control
study of melanoma was conducted in Italy
between June 1992 and February 1995. The
authors recruited 542 incident cases from oncol -
ogy and dermatology centres, and 528 controls
admitted to the hospital for a non-dermatologic or
Epidemiological data on exposure to artificial UV radiation for cosmetic purposes and skin cancers
non-neoplastic illness. Of these, 30 cases and 36
controls were ever exposed to sunbeds or sun-
lamps. The risk for melanoma, adjusted for age,
sex, marital status, education, eye and skin
colour, number of naevi , freckles density, sun-
burns and number of sunny vacations, was 0.78
(CI, 0.45-1.37).
Kaskel eta/. (2001): A hospital-based case-con-
trol study of melanoma was conducted in Munich,
Germany between June 1996 and April 1997.
The authors recruited 271 prevalent cases (diag-
nosed from 5 years to 6 months before inclusion)
from the Tumour Centre in Munich, and 271 con-
trols from hospital departments of general surgery
and ophthalmology. Controls were matched to
cases for age (in 5-year categories), sex and
place of residence. Among the 56 factors
explored, one item investigated exposure to UV
radiation or UV beds more than 5 times per year
compared with 5 times per year or less. In the
analysis of discordant pairs, the crude risk for
artificial UV exposure was 1.0 (CI, 0.6-1.8).
Bataille eta/. (2004): A hospital-based case-con-
trol study of melanoma was conducted in the
North East Thames region (United Kingdom)
between August 1989 and July 1993. The authors
recruited 413 cases and 416 controls aged 16 to
75 years old. Incident cases of histologically con-
firmed melanomas were recruited from hospitals
and general practices. Controls were also recruited
through hospitals and general practices, excluding
patients attending for a skin disease. One hun-
dred cases and 11 0 controls had ever been
exposed to sunbeds. The risk for melanoma
associated with sunbed use was 1.19 (CI,
0.84-1.68), after adjusting for age and sex.
Further adjustment for skin type and other sun
exposure measures did not affect the results. In a
stratified analysis, if sunbed exposure took place
before the age of 45 years, the relative risk was
1.2 (CI, 0.76-1.90). No trend toward increased
risk was observed with increasing lifetime dura-
tion of exposure.
Bataille et a/. {2005) : A case-control study
designed specifically to investigate melanoma
risk associated with sunbed exposure was con-
ducted in Belgium, France, the Netherlands,
Sweden and the United Kingdom between
December 1998 and July 2001. The authors
recruited 597 incident cases from dermatology or
oncology clinics or identified through pathology
laboratories. The method of recruitment of 622
controls differed according to each centre: popu-
lation register in Sweden, neighbourhood
controls in Belgium and France, and general
practices in the Netherlands and the United
Kingdom. Of these, 315 cases and 354 controls
had ever used sunbeds. The risk for melanoma
associated with sunbed use was 0.9 (CI,
0. 71-1 .14) when adjusting for age, sex and skin
type. If exposure to tanning appliances occurred
before age 15 years, the relative risk was 1.82
(CI, 0.92-3.62) . No trends in risk for melanoma
were observed with increasing lifetime exposure
or with increasing time since first exposure. No
association was observed when stratifying by
type of sunbed. [A companion paper warned
about potential biases that could have occurred
in this study: selection bias of controls and mis-
classification of cases who tended to underreport
their exposure (de Vries eta/. , 2005)].
Of these 23 studies, 4 studies were excluded-
in accordance with the selection criteria-
because they did not include estimates of the
relative risk for cutaneous melanoma associated
with exposure to tanning appliances (Kiepp &
Magnus, 1979; Gallagher eta/., 1986; Holly eta/.,
1987; Beitner eta/., 1990).
Another study (Walter eta/., 1990) which pre-
sented an evaluation of "ever" versus "never"
exposed to artificial UV radiation was excluded
because it involved the same population as a later
publication (Walter eta/., 1999); moreover, it pre-
sented crude rather than adjusted relative risks.
However, the estimate for "first exposure before
age 35 years" from the early publication (Walter et
a/. , 1990) was included in the relevant section.
Quantitative approach: meta-analysis
(a) Evaluation of exposure: Four types of expo-
sure to indoor tanning appliances were evaluated:
"ever" versus "never";
"first exposure before age 35 years" versus "never".
Exposure to Artificial UV Radiation and Skin Cancer
In addition, another concept was considered in
order to make a comparison between recent and
distant exposures:
"exposure distant in time" versus "never";
"exposure recent in time" versus "never".
A dose-response model was not considered
for this meta-analysis because of the hetero-
geneity among the categories of duration and
frequency of exposure used by different authors.
(b) Study characteristics: Table 7 provides an
overview of all the studies retrieved, including the
19 studies reporting estimates that could be
included in the meta-analysis (for a total of 7 355
cases). The first (publ ished in 1981) and the last
(published in 2005) studies included were pub-
lished more than 20 years apart. Three
case-control studies presented a time lag
between first recruitment year and publication of
1 0 years or more.
Fifteen studies were carried out in European
countries, four of which were in Scandinavian
countries; two were conducted in the United
States, one in Canada and one in Australia. The
mean latitude of the study centres was 50
(range 25!L59
) ; eight studies were conducted in
countries with average latitude below 50.
(c) Types of estimate presented: Since melanoma
is a rare disease, we ignored the distinction
between the various estimates of relative risk (i.e.
odds ratio, rate ratio, risk ratio), and all measure-
ments were interpreted as odds ratios.
Except for the studies by Kaskel eta/. (2001)
and by Veierod eta/. (2003) , all studies presented
estimates for "ever" versus "never" exposed to
artificial UV radiation (Table 8). Thirteen of 19
studies presented positive estimates for "ever"
versus "never" exposed to sunbed/sunlamps, but
only four were statistically significant. For seven of
these studies it was possible to obtain only crude
relative risks, one adjusted for age and sex only.
The cohort study (Veiemd eta/., 2003) pre-
sented an estimate for the widest age interval
included (1 Q-39 years) , only for the comparison
":?: 1 time per month" versus "never/rarely" . One
study (Kaskel eta/., 2001) presented an estimate
only for the comparison ">5 times per year"
versus "::;: 5 times per year".
Five studies (Swerdlow eta/., 1988; Walter et
a/., 1990; Chen et a/., 1998; Westerdhal et a/.,
2000; Bataille eta/. , 2005) also presented an esti-
mate for first exposure at age s 35 years (Table
9). Veierod eta/. (2003) presented relative risks
for ":?: 1 time per month" versus "never" in the age
period 2Q-29 years; Westerdhal et a/. (1994)
presented estimates of "ever" versus "never" for
individuals younger than 30 years. All relative
risks were adjusted for confounders related to
sun exposure or sun sensitivity, except in the
study by Walter eta/. (1990). All these estimates
were considered for the evaluation of "first expo-
sure before age 35 years" versus "never".
Five studies investigated time since exposure
(Table 1 O) and reported estimates that allowed
comparisons between recent and distant expo-
sure: number of years of exposure before pres-
entation (Swerdlow et a!., 1988; Bataille et a!.,
2005), number of years since last exposure
(Walter et a/., 1990) and age at first exposure
(Autier eta!. , 1994; Chen eta/., 1998).
(d) Selection of data and methods of analysis:
Every measure of association adjusted for the
maximum number of confounding variables and
corresponding confidence interval were trans-
formed into log RR, and the corresponding vari-
ance was calculated using the formula proposed
by Greenland (1987). Where no estimates were
given, crude estimates were calculated from tab-
ular data, using Asymptotic Mantei-Haenszel
estimates to evaluate the 95% Cl of the log odds
Most estimates included all subjects, combining
sexes. One study presented results separately for
women and men with no combined data; both
estimates were included (MacKie eta/., 1989).
The homogeneity of the effects across studies
was assessed using the large sample test based
on the Chi-square statistic (Chi). Since the Chi-
square test has limited power, we considered
statistically significant heterogeneity at the
P=0.1 0 level of association. A further measure of
heterogeneity, H (the square-root of Chi-square
divided by its degrees of freedom), has been con-
sidered in order to make comparisons between
heterogeneities of pooled estimates summarizing
Epidemiological data on exposure to artificial UV radiation for cosmetic purposes and skin cancers
Table 7. Characteristics of studies considered for the meta-analysis on melanoma
Reference Country First Year Number Histological di agnosis Participation
of controls {%)
Cases Controls
Cohort study
Veier0d eta/. (2003) Norway, Sweden 1992 187 106 379
HC invasive M 54.5
Populatlonbased studies
'Adam eta/. (1981) UK 1971 169 207 HCM 68
Gallagher eta/. (1986) Western Canada 1979 595 595 M excluding LMM and ALM 48
'Holman eta/. (1986) Australia 1982 511 511 HC pre-invasive/ invasive M 69
'Osterlind eta/. (1988) Denmark 1985 474 926 HCM excluding LMM 81.7
'Zanetti eta/. (1988) Italy 1984 208 416 M in situ and all other 68.2
Beitner eta/. (1990) Sweden 1978 523 505 HCM (SSM, NM. LMM, 96.2
unclassif. MM)
Walter eta/. (1990) Canada 1984 583 608 HCM in situ and Hutchinson's 81
freckle, LMM
'Westerdahl el a/. (1994) Sweden 1990 400 640 Invasive M 77.4
'Holly el a/. ( 1995) USA 1986 452 930 HCM 77
'Chen eta/. (1998) USA 1969 624 512 HC fi rst primary invasive M 70
'Walter eta/. (1999) Canada 1966 583 608 HCM in situ and Hutchinson's 81
freckle, LMM
'Westerdahl at a/. (2000) Sweden 1997 571 913 HC first primary invasive M 68
Other studies
Klepp & Magnus (1979) Norway 1974 78 131 M NR
Holly el a/. ( 1987) USA 1984 121 139 NMorSSM NR
Swerdlow eta/. ( 1988) UK 1988 160 120 PrimaryM NR
'MacKie eta/. (1989) UK 1987 280 180 Invasive M NR
' Dunn-Lane et a/. (1993) UK 1986 100 100 M excluding LMM and ALM NR
' Garbe eta/. (1993) Germany 1987 280 280 M NR
'Autier eta/. (1994) Belgium, France & 1991 420 447 HCM 78
' Naldi eta/. (2000) Italy 1993 542 538 M NR
' Kaskel eta/. (2001) Germany 1996 271 271 HCM NR
' Bataille eta/. (2004) UK 1993 413 416 M including in situ and LMM NR
' Bataille eta/. (2005) UK 1998 597 622 HC first primary invasive M NR
excluding LMM
included in the meta-analysis; cohort size: response rate.
ALM, acral lentiginous melanoma; HC, histologically confirmed: LMM, lentigo maligna melanoma; M, melanoma; MM. malignant
melanoma; NM, nodular melanoma; NR, not reported; SSM, superficial spreading melanoma.
Exposure to Artifici al UV Radiation and Skin Cancer
Table 8. Estimates included in the evaluation of an association of ever use of indoor tanning
facilities and risk for melanoma
Reference Exposure comparison Adjustment
Adam eta/. (1981) Ever use of sunlamps vs never Crude
Holman et a/. ( 1986) Ever use of sunlamps vs never 1.1 (0.6-1. 8) Cnude
Osterlind et a/. (1988) Ever use of sun beds vs never 0.73 (0.53-1.01) Cnude
Swerdlow eta/. (1988) Ever use of UV lamps/ sunbeds vs 2.94 (1.41-6.17) Cnude
Zanetti et a/. (1988) Use of UVA lamp for tanning 0.9 (0.4-2.0) Age, hair colour, skin reaction,
purpose: yes/no sunburn in childhood, education
MacKie et al. (1989) Ultraviolet use: some vs none 1.3 (0.2- 7.9) Naevi, freckles, sunburns, tropical
(men) residence, phototype
MacKie et al. (1989) Ultraviolet use: some vs none 1.2 (0.5--3.0) Naevi, freckles, sunburns, tropical
(women) residence, phototype
Dunn-Lane eta/. Ever use of sun beds vs never 1.16 (0.54- 2.47) Cnude
Garbe et a/. (1993) Use of sunbeds: yes/no 1.5 (0.9-2.4) Age, naevi, hair colour, phototype,
study cent re
Autier et al. (1994) Ever exposed to 0.97 (0.71-1.32) Crude
sunl amps/sunbeds vs never
Westerdahl eta/. (1994) Ever exposed to 1.3(0.9- 1.8) Sunburns, hair colour, naevi,
sunbedslsunlamps vs never sunbathing
Holly et al. (1995) Ever use of sunlamps vs never 0.94 (0.74-1.2) Crude
Chen et al. (1998) Ever use of sunlamps vs never 1.13 (0.82- 1.54) Sex, age, phenotype, recreational
sun exposure
Walter et al. (1999) Ever use of sunbeds/sunlamps vs 1.54 (1.16-2.05) Sex, age, skin reaction to initial
never summer sun exposure
Naldi et al. (2000) Ever use of sunbeds/sunlamps vs 0.78 (0.45-1 .37) Sex, age, skin, hair, eye, naevi,
never freckles, sunburn, number of sunny
Westerdahl et al. (2000) Ever use of sunbeds vs never 1.2 (0.9-1.6) Sunburns, hair colour, skin type,
raised naevi
Kaskel et al. (2001) Artificial UV radi ation/UV beds: 1.00 (0.6-1 .8) Crude
>5/year vs :s5/year
Veier0d et al. (2003) Solarium use : 2:1/month vs 1.55 (1 .04-2.32) Age. region of residence. hair
(women) never/rarely colour, sunburns, summer vacations
Bataille et al. (2004) Ever use of sun beds vs never 1.19 (0.84- 1.68) Sex, age
Bataille eta/. (2005) Ever use of sunbeds or sunlamps 0.90 (0.71-1.14) Sex, age, skin phototype
vs never
Epidemiological data on exposure to artificial UV radiation for cosmetic purposes and skin cancers
Table 9. Estimates included in the evaluation of an association of first use of indoor tanning
facility in youth and risk for melanoma
Reference Definition
Swerdlow eta/. (1988) Age at first exposure <30 years vs
Relative risk
(95% Cl)
3.8 (0.9--16.5)
Naevi, skin type, hair and
eye colour, sun exposure
Walter eta/. (1990) Age at first use <30 years vs never 1.67 (1.17-2.39) Age
Westerdahl eta/. (1994) Ever use of sunbed at age younger 2.7 (0.7- 9.8) Sunburns, hair colour,
naevi , sunbathing than 30 years
Chen eta/. (1998) Age at first use of sunlamp < 25
years vs never
1 .35 (0.88-2.08) Sex, age, phenotype index,
recreational sun exposure
Westerdahl et a/. (2000) Age at first exposure s; 35 years vs 1 .6 {0.9--2.9) Sunburns, hair colour,
skin type, naevi
Veiemd et al. (2003)
Bataille et at. (2005)
Exposure at age 2D-29:
~ 1 time/month vs never
Ever sunbed use before age
15 years vs never
2.58 (1.48-4.50)
1 .82 {0.92- 3.62)
Age, region of residence,
sunburns, summer
Age, sex, skin type
Table 10. Estimates included in the evaluation of an association of distant and recent exposure
and risk for melanoma
Swerdlow et a/.
Walter eta/. {1990)
Autier eta/. (1994)
Chen eta/. (1998)
Bataille et a/. (2005)
Less than 5 years before
presentation vs never
More than 5 years before
presentation vs never
Less than 5 years since last use
vs never
More than 5 years since last use
vs never
First use in 1980 or later ~ 1 0 hr
of exposure for tanning purposes)
First use before 1980 ~ 1 0 h r of
exposure for tanning purposes)
First use after 1970
First use before 1970
< 6 years between first sunbed use
and interviews
;:: 15 years between first sunbed use
and interviews
Relative risk
(95% Cl)
1.9 (0.6-5.6)
9.1 (2.0-40.6)
Men, 1.52 (0.56-4.25)
Women, 1.24 (0.67-2.31)
Men, 2.00 (1 .21-3.34)
Women, 1.53 {0.96-2.46)
0.99 (0.49--2.00)
2.12 (0.84-5.37)
1.15 {0.64-2.07)
1.33 (0.84-2.12)
0.91 (0.58-1.42)
0.97 (0.70-1.34)
Age, sex,
Age, sex, hair
colour, holiday
weeks spent in
sunny resorts
Sex, age,
phenotype index,
recreational sun
Sex, age, skin
Exposure to Artificial UV Radiation and Skin Cancer
different numbers of studies. Greater values of H
indicate larger heterogeneity (Higgins &
Thompson, 2002).
The summary relative risk was estimated by
pooling the study-specific estimates by random
effects models even when heterogeneity was
found to be not significant and H was very low, in
order to be conservative and to enable generali-
zation of the results. For mixed effects models,
SAS was used (SAS Institute Inc. SAS Windows
version 8.02, 1999, Cary, NC) with PROC MIXED
(van Houwelingen et a/., 2002). These models
allowed taking into account between-study vari-
ability and non-independence of estimates origi-
nating from the same study.
Subgroup analyses and meta-regressions
were carried out to investigate inter-study hetero-
geneity (Colditz eta/., 1995). Heterogeneity was
investigated by looking at all factors concerning
the type of study, analysis, exposure and features
of the population that could influence the esti-
mates. Studies conducted in different populations
living at substantially different latitudes were not
included in the heterogeneity analysis that evalu-
ated latitude.
A sensitivity analysis was conducted to eval-
uate the stability of the pooled estimates and the
influence of individual studies. To verify whether
publication bias might affect the validity of the
estimates, funnel plots were plotted using Copas
and Shi's method (Copas & Shi , 2001) and the
funnel plot regression of Ln(RR) on the sample
size, weighted by the inverse of the pooled vari-
ance (Macaskill eta/., 2001 ).
Table 11. Meta-analysis of all studies included
(e) Pooled estimates: Results of the meta-analy-
sis of all studies included are shown in Table 11
and Figure 2. Between-study heterogeneity was
found significant for being "ever" versus "never"
exposed to artificial UV (Chi=35.40, degrees of
freedom (d.f.) =19, P=0.013). The pooled
estimate indicated a borderline-significant posi-
tive association between "ever" versus "never"
use of sunlamps/sunbed and melanoma (RR,
1.15; Cl, 1.00-1 .31 ).
When "first exposure before age 35 years"
was analysed, a significant 75% increase in risk
was detected (Table 11; Figure 3) and the Chi-
square testing heterogeneity was non-significant
(Chi= 4.95, d.f. = 6, P = 0.55) and H (= 0.91) was
smaller than the value obtained for "ever" versus
"never" (H = 1.37).
The number of studies presenting an
assessment of time since exposure was low
(n = 5); however all studies presented greater
estimates for exposures more distant in time com-
pared to more recent exposures. Heterogeneity was
greater for "distant exposure" (H = 1.65 and Chi
=13.63, d.f. = 5, P = 0.018) than for "recent expo-
sure" (H = 0.67 and Chi = 2.52, d.f. = 5, P = 0.81 ).
It is interesting to note that exposures more
distant in time led to an increased risk compared
with recent exposures, consistently with the higher
risk for "first exposure before age 35 years"
versus "never'' compared to "ever" versus "never".
In order to decrease the influence of biases,
estimates were calculated including only the
cohort and population-based case-control studies
(Table 12). The pooled relative risks were very
similar apart from wider confidence intervals.
Exposure Number of Summary relative risk (95% Cl) Heterogeneity
studies P-value x
Ever use of indoor tanning facility 19 1.15 (1.0G-1.31) 0.013 1.37
First exposure in youth 7 1. 75 (1.35-2.26) 0.55 0.91
Exposure distant in ti me 5 1.49 (0.93-2.38) 0.018 1.65
Exposure recent in ti me 5 1.10 (0.76-1.60) 0.81 0.67
The degrees of freedom for the Chi-square are given by the number of databases included minus one, not by the
number of studies.
Epidemiological data on exposure to artificial UV radiation for cosmetic purposes and skin cancers
Figure 2. Relative risk for cutaneous melanoma associated with ever use of indoor tan-
ning equipment: estimates of 19 studies and summary estimate
Holman eta/.(1986)
o.terlind eta/. ( 1988)
Swerdlow et a/.(1988)
Zanerti., al. (1988)
Mackie eta/.(1989, Men)
Mackie eta/(1989, Women)
Dunn-Lane c/a/.(1993)
Garbe et a/.(1993)
Au tier et a/. ( 1994)
Westerdnhlet al. (1994)
Holly et al (1995)
Wal!eret al. (1999)
Nldi eta/. (2000)
Wes!erdah.l eta/. (2000)
Kaslccl et a/. (2001)
Veierod et a/.(2003)
Bataille I a/. (2004)
Dnta.ille eta/. (2005)
Summary relative risk
-II 1--

1.15 (1.00, 1.31)
1.0 1.5 2.0 2.5 3.0 5.0 7.0
Relative risk
Figure 3. Relative risk for cutaneous melanoma associated with first use of indoor tanning
equipment at age <35 years: estimates of 7 studies and summary estimate
Swerdlow eta/. (1988)
Westerdnbl et al. (1994)
Chen eta/. (1998)
Walter eta/. ( 1999)
Westerdahl eta/. (2000)
Veierod et al. (2003)
Bataille el al. (2005)
Summary relative risk


1. 75 (1.35-2.26)
l.O 1.5 2.0 2.5 3.0 5.0 7.0
Relative risk
Exposure to Artificial UV Radiation and Skin Cancer
Table 12. Meta-analysis of the cohort and population-based case-control studies included
Exposure Number of Summary relative risk Heterogeneity
studies (95% Cl)
P-value X
Ever use of indoor tanning facility 10 1.17 (0.96-1.42) 0.011 1.540
Age at first exposure in youth 5 1.71 {1.25-2.33) 0.435 0.973
Exposure distant in time 2 1.58 (0.25-9.98) ' 0.502 0.830
Exposure recent in time 2 1.24 (0.52- 2.94) 0.762 0.521
'The confidence interval is very wide because this analysis includes only 2 studies, one of which has two estimates.
(f) Heterogeneity analysis: For the comparison of
"ever" versus "never", which included the largest
number of studies, several factors that could
influence the variability among estimates were
investigated. This analysis revealed that studies
with a longer time lag between the first year of
recruitment and 10 years) presented
higher estimates (Table 13). (The cohort study
was excluded from this analysis because of the
nature of the study design.)
Studies carried out in countries at higher
latitudes presented higher relative estimates than
did studies carried out at lower latitudes (Table 13
and Figure 4).
Adjustment for confounders related to sun
exposure and sun sensitivity led to a higher
pooled estimate compared with studies considering
only crude relative risks or relative risks adjusted
only for age and sex (Table 13). In the analysis
restricted to the eight studies that adjusted for
Table 13. Heterogeneity analysis
confounders related to sun exposure and sun
sensitivity, the pooled relative risk remained simi-
lar to the summary estimate for all 19 studies but
the confidence interval widened (RR, 1.19; Cl,
0.33--4.30). The difference between adjusted and
crude pooled relative risks may not be due to the
adjustment in itself but to the fact that well-con-
ducted studies usually adjust for sun exposure
and sun sensitivity, which could be an indicator of
the quality of the analysis.
(g) Sensitivity analysis: A series of analyses were
performed to test the stability and sensitivity of
the analysis (Table 14). Inclusion criteria were
tested by including the estimates reported by
Walter and colleagues in 1990 instead of those
reported in 1999. Also, the studies that did not
report any relative risk (Kiepp & Magnus, 1979;
Gallagher eta/., 1986; Holly eta!. , 1987; Beitner
et a!., 1990) were included by imputing the
Number of Pooled relative risk Heterogeneity
Parameter analysed
studies (95% Cl)
P-value X
Number of years between recruitment and 3 1.38 (0.25- 7.46) 0.16
pub I ication 1 0
Number of years between recruitment and 15 1.06 (0.50-2.27) 0.14
publication <1 0
Estimate adjusted for phototype/sun 10 1.19 (0.45-3.12) 0.17
Crude estimate or estimate adjusted for age 9 1.03 {0.31-3.40) 0.018
and sex only
Latitude of study centre <50 8 1.08 (0.31 - 3. 78) 0.73
Latitude of study center >50 o 11 1.20 (0.41-3.46) 0.003
Epidemiological data on exposure to artificial UV radiation for cosmetic purposes and skin cancers
Figure 4. Correlation between latitude of study centre and relative risk for melanoma
associated with use of indoor tanning facilities

-0.5 +------.,.--- - -----..,r-------.------1
20 30 40 50 60
Latitude (in degrees) for the region of each study
m1ss1ng estimates from data available in the
reports. Where no data at all were presented but
an indication of non-significant effect was given,
a relative risk of 1 and a standard error equal to
the mean standard error of the other studies was
considered. The pooled relative risks did not
change considerably (Table 14).
In order to verify the stability of the results, a
new analysis was carried out taking out the
estimate from the cohort study (Veiemd et a/.,
2003) . The pooled relative risk showed a wider
confidence interval.
The definitions used to evaluate the risk for
"first exposure before age 35 years" differed for
two studies: one study presented an estimate of
"ever" versus "never" for individuals aged 30
years (Westerdahl et a/. , 1994); the other study
(Veiemd et a/., 2003) presented two estimates:
"ever" versus "never" at age 1 0-19 years and ";;::: 1
time/month" versus "never" at age 20-29 years.
For the latter study, the estimate including a larger
number of individuals (age group 20-29 years) was
used for the main analysis of "first exposure before
age 35 years" (only 4 cases were in the exposed
group for the estimate at age 10-19 years). When
both studies were excluded, the pooled estimate did
not change considerably (Table 14).
For the evaluation of recent and distant expo-
sures, Autier et al. ( 1994) reported estimates by
several substrata; for the main analysis we selected
the adjusted relative risk evaluating exposure for
tanning purposes and for a duration of 10 hr or
more. Crude relative risks obtained by merging all
categories were: for "distant exposure" , 1.22
(CI, 0.79-1 .88) and for "recent exposure", 0.82
(CI, 0.56-1 .19). Thus the pooled relative risk for
"distant exposure" remained greater than that for
"recent exposure" (data not shown).
Analysis by Funnel plot regression gave no indi-
cation of publication bias ("ever used sunbed/sun-
lamps", P = 0.80; "first exposure before age 35
years", P = 0.1 0). In addition, analysis by the Copas
and Shi method of trends in the funnel plots
(Figures 5 and 6) gave an indication of non-signifi-
cant asymmetry ("ever used sunbed/sunlamps",
P = 0.37; "first exposure before age 35 years", P =
To establish a causal link between exposure to tan-
ning appliances and melanoma occurrence, studies
should show whether there are dose-effect rela-
tionships and whether exposures distant in time are
Exposure to Artificial UV Radiation and Skin Cancer
Table 14. Sensitivity analysis
Inclusion criteria Number of
Ever use of indoor
tanning facility
Including study by
Walter eta/. (1990)
Including all studies 23
First exposure in
Excluding the
cohort study by
Veier0d eta/. (2003)
Excluding the
cohort study by
Veier0d et a/. (2003)
Including only
those studies with
a specific definition
of first exposure
(studies by Veiered
et a/. 2003 and
Westerdahl eta/.,
2000 excluded)
more strongly associated with melanoma than are
recent exposures. The latter point is important, as
there is most probably a latency period between
exposure and melanoma, thus the carcinogenic
effect of more recent exposures would not yet be
detectable. Also, since the fashion of using indoor
tanning facilities has been increasing steadily, a
lack of distinction between distant and recent expo-
sures may mask an actual increase in risk.
Experimental and epidemiological studies
provide evidence that susceptibility to UV radia-
tion is greater at younger ages (mainly in child-
hood and adolescence) than at older ages (see
page 8; Autier & Dare 1998; Whiteman et at.,
2001). Hence, data analysis should identify
whether exposure to tanning appliances starting at
younger ages was more strongly associated with
melanoma than exposure starting at older ages.
The UV emission spectrum of UV lamps in
indoor tanning appliances has changed over
time: before 1980, many UV lamps produced
large amounts of UVC and UVB, whereas most
UV tanning appliances used after 1985 mainly
emitted in the UVA range (see page 3).
Summary relative
P-value X
risk (95% Cl)
1.15 (1.00-1.32) 0.007
1.14 (1.00-1.30)
1.11 (0.97-1.26) 0.019
1.64 (1.22-2.20)
1.65 (1.17-2.32) 0.709
(a) Case-control studies: Case-control studies of
melanoma providing results on use of indoor tan-
ning facilities have been of variable study design,
and many of them only included one question on
exposure to tanning appliances. Some positive or
negative associations between exposure to tan-
ning appliances and risk for melanoma may have
been due to statistical fluctuations (i.e. alpha or
beta errors) or to design effects.
In some studies, melanoma patients (i.e.
cases) were derived from a small number of der-
matologic clinics, and subjects without melanoma
(i.e. controls) were derived from hospital wards or
outpatient clinics. This way of selecting cases and
controls is prone to many biases: for instance,
control subjects could suffer from a disease asso-
ciated with higher or lower propensity to engage
in indoor or outdoor tanning.
Users of indoor tanning facilities have been
shown to have a greater-than-average propensity
to engage in intentional sun exposure (Autier et
a/., 1991 ), and may have characteristics of
inherited sun sensitivity different from the rest of
the population (see page 9). Hence, a possible
association between exposure to tanning
Epidemiological data on exposure to artificial UV radiation for cosmetic purposes and skin cancers
Figure 5. Investigation by Funnel plot representation of a possible publication bias in the
studies of risk for melanoma associated with use of indoor tanning facilitites included in the

.. !

., 'i:







-0.5 0 0.5 1 1.5
Logarithm of RR
Figure 6. Investigation by Funnel plot representation of a possible publication bias in the
studies of risk for melanoma associated with first use of indoor tanning facilities in youth

0 0.4


0.8 +------......-------,--------,
0 0.5 1 1.5
Logarithm of RR
Exposure to Artificial UV Radiation and Skin Cancer
appliances and risk for melanoma could in fact be
due to greater sun exposure than average, or to
greater use of indoor tanning facilities by subjects
naturally more prone to melanoma. To reduce the
effect of these confounding factors on risk
estimates, it was necessary to adopt statistical
methods (e.g. a multivariate logistic regression
model) allowing the calculation of estimated risks
adjusted for both sun exposure history and host
In order to examine the consistency of the
data on exposure to tanning appliances and risk
for melanoma provided by case-control studies,
we selected those studies among the 19 studies
included in the meta-analysis (see Tables 7 and
8) that had a section specifically exploring
exposure to tanning appliances and results
adjusted for (intermittent) sun exposure and
sun sensitivity (Autier eta/., 1994; Westerdahl et
a/., 1994; Chen et a/., 1998; Westerdahl et a/.,
Table 15 presents adjusted relative risks for
melanoma associated with exposure to tanning
appliances, showing some statistically significant
dose-effect relationship for two studies (Autier et
a/. , 1994; Westerdahl et a/., 1994), a borderline
statistically significant dose-effect relationship in
one study (Chen eta/., 1998), and one study with
a non-significant dose-effect relationship
(Westerdahl eta/., 2000).
Two of the four studies (Autier et a/. , 1994;
Chen et al. , 1998) showed that the highest risk
for melanoma was associated with exposure to
tanning appliances more distant in time (Table
1 0). Three studies (Westerdahl eta/., 1994; Chen
eta/., 1998; Westerdahl et al., 2000) showed that
melanoma risk was highest when exposure to
tanning appliances started at younger ages, i.e.
before approximately 35 years old (Table 9).
However, most associations with exposure dis-
tant in time and with younger age at start did not
reach statistical significance because of the low
number of subjects in the relevant categories of
exposure. Statistical significance first emerged
when all data were combined in a meta-analysis,
resulting in a greater number of subjects in
relevant categories of exposure and thus higher
statistical power (see page 30).
(b) Prospective study: The Norwegian-Swedish
study (Veiemd eta/., 2003) is the only published
prospective cohort study of environmental risk
factors for melanoma. Women in Norway and
Sweden (N=106 379) were followed for an average
of 8.1 years from 1991 until 1999. The study
showed consistent associations between host
characteristics of inherited sun susceptibility,
sunburn history, sun exposure, exposure to tan-
ning appliances and cutaneous melanoma.
During follow-up, 187 cases of melanoma were
diagnosed. After adjustment for intermittent sun
exposure and host characteristics, the adjusted
relative risk for melanoma was 1.55 (CI ,
1 .04-2.32) among the 18% of women aged
1 0-39 years who reported having used sun beds
at least once a month when they were 10-19,
20-29 or 30-39 years old. Twelve sunbed
sessions per year correspond to the typical tan-
ning programme proposed by many commercial
tanning facilities. Thus the 55% increase in
melanoma risk was related to 40 hours or more
of exposure to tanning appliances, assuming an
average of 20 minutes per session. In that
respect, the levels of exposure to tanning
appliances reported in this prospective study
were more comparable with levels reported in
surveys carried out in European countries than
those reported in case-control studies.
In the Scandinavian countries, use of indoor
tanning facilities has been popular since the late
1970s, and the prevalence of use of indoor tanning
facilities in those countries is the highest in the
world. In the Norwegian-Swedish prospective
study, the highest risk for melanoma was found in
women who used indoor tanning faci lities at least
once per month when they were 20 to 29 years old
(RR, 2.58; Cl, 1.48-4.50), and the lowest risks
were found for exposure to tanning appliances at
least once a month during the third (RR, 1.42; Cl,
0.93-2.16) or the fourth decade of life (RR, 1.67;
Cl, 0.93-2.99). These results support the hypo-
thesis by which a latency period is needed before
the impact of exposure to tanning appliances on
melanoma incidence becomes apparent. It also
underlines the greater vulnerability of younger sub-
jects to harmful effects of sunbeds.
Epidemiological data on exposure to artificial UV radiation for cosmetic purposes and skin cancers
Table 15. Duration of exposure to indoor tanning facilities and risk for melanoma in selected
case-control studies
Reference Duration of Cases Controls Estimated 95% Cl
Place & years of study exposure risk
Numbers of cases/control
Autier et a!. ( 1 994) Never used 310 327 1.00 Ref.
Belgium, France, Germany, Exposure starts < 10 hours 36 45 0.75 0.46-1.25
1991- 92 1980 10 hours 19 18 0.99 0.49-2.00
Exposure starts < 10 hours 16 15 1.00 0.47- 2.13
< 1980 10 hours 18 7 2.12 0.84-2.12
Westerdahl eta/. (1994) Never used 282 479 1.0 Ref.
Sweden, 198890 1- 3 sessions/year 44 67 1.1 0.7-1.9
400/640 4-1 0 sessions/year 30 55 1.1 0.7-1.9
> 1 0 sessions/year 41 33 1.8 1.0-3.2
Chen et a/. ( 1998) Never used 483 417 1.00 Ref.
Connecticut, USA, 1987-89 < 10 sunlamp uses 76 50 1.25 0.84-1.84
624/512 1 0 sunlamp uses 63 40 1.15 0.60-2.20
Westerdahl eta!. (2000) Never used 319 538 1.0 Ref.
Sweden, 1995-97 1-125 uses 22 32 2.8 1.0-7.8
571/913 126-250 uses 34 31 3.1 1.3-7.1
> 250 uses 31 37 1.5 0.7-3.2
Duration of exposure, relative risk, and 95% confidences as in published reports. All estimated risks are adjusted for
age, sex, natural sun sensitivity and recreational sun exposure.
The 21 cases and 35 controls who were exposed to sunlamp or sunbed for non-tanning purposes are not reported in
this Table.
(c) Methodological aspects of case-control and
prospective cohort studies: Case-control studies
are prone to two biases inherent in the design.
First, since data are collected retrospectively
(when cases already know they have a
melanoma}, the associations found could be the
result of recall bias, as melanoma patients might
have been more likely to remember past expo-
sures to artificial UV sources (Walter et al. ,
1990). Second, the selection of controls may
have included subjects more (or less) inclined to
have had more frequent exposure to tanning
appliances than average (selection bias).
Among the four case-control studies selected
in Section (a) of this section, three studies (Autier
et al. , 1994; Westerdahl et al., 1994, 2000) used
measures to control for recall bias. Autier et al.
(1994) focused on recall bias in the training of the
interviewers: neither interviewers nor subjects
were informed of the study's objective.
Westerdahl et al. (1994) used a questionnaire
with many variables and stated that at the time of
the interview (1988 to 1990), the population was
unaware of the relationship between exposure to
artificial UV radiation for tanning purposes and
malignant melanoma. Westerdahl et al. (2000)
used identical procedures of data collection for
cases and controls, and collected information
from melanoma patients shortly after diagnosis.
Selection bias of controls was not likely to
have occurred in any of the four selected
case-control studies: three studies (Westerdahl
eta!., 1994, 2000; Chen eta/., 1998) were based
on population-based melanoma registries and
sampling of control subjects. The study by Autier
et a/. (1994) selected cases from multiple
sources (hospital, clinics and melanoma regis-
tries), and controls were chosen in the neigh-
bourhood of cases according to rigorous contact
procedures (Grimes & Schulz, 2005).
The prospective cohort study assessed
exposure to tanning appliances retrospectively
Exposure to Artificial UV Radiation and Skin Cancer
but before diagnosis of melanoma. Thus, this
study was less prone to interview and selection
biases at the inception of the cohort.
Taken together, the four case-control studies
selected and the prospective study offer the con-
clusion that the increased melanoma risk was
associated with exposure to tanning appliances
(mainly when exposure started before the age of
approximately 35 years) and the observed positive
associations are not entirely due to recall or selec-
tion biases.
(d) Type of artificial UV light: Only one study
(Chen et al., 1998) collected information con-
cerning the type of appliance used by showing
subjects pictures of various types of indoor
tanning appliances (e.g. desktop models, floor
models, beds, walk-in booths). The study found a
non-significant elevated risk for melanoma asso-
ciated with the use of desktop sunlamps and
heavyweight floor-model sunbeds and a statisti-
cally significant tripled risk associated with use of
more than two types of sunlamps, compared with
no use of sunlamps.
Before 1980, exposure to artificial UV radia-
tion was more likely to take place at home with
appliances that emitted large amounts of UVB
radiation, whereas exposure in the 1980s
increasingly occurred in commercial salons using
appliances that emitted mainly UVA. The
prospective study provided evidence that the
increased melanoma risk associated with expo-
sure to tanning appliances was not due to the
type of UV lamps used before 1983 (Veiemd et
a/., 2004).
Basal cell and squamous cell carcinomas
Description of studies
Nine case-control studies have addressed the
possible association of artificial UV exposure with
either BCC or SCC of the skin. All studies reported
a risk estimate, except one (Boyd et al., 2002),
which was therefore excluded. A further three
studies that did not distinguish between these
two major types of skin cancer (O'Loughlin et al.,
1985; Herity eta!., 1989; Hogan eta!., 1991) were
also excluded from review because BCCs and
sees have different aetiologies, thus leaving five
studies under consideration (Table 16).
Aubry & MacGibbon (1985): The earliest
case-control study that addressed the possible
Table 16. Characteristics of case-control studies included in the meta-analysis on non-
melanoma skin cancers
Reference Country Number of
Aubry & McGibbon Canada sec: 92
Bajdik eta/. (1996) Canada BCC: 226
SCC: 180
Corona eta/. (2001) Italy BCC: 166
Karagas et af. (2002) USA BCC: 601
Walther et af. (2004) Germany BCC: 213
BCC, basal cell carcinoma; SCC, squamous cell carcinoma
Number of
Population, health
Dermatology Population, Dept.
department of Transportation,
Hospital Hospital
Epidemiological data on exposure to artificial UV radiation for cosmetic purposes and skin cancers
association of artificial UV exposure and
squamous cell carcinoma was conducted in
Montreal, Canada. Its overall aim was to assess
risk factors for sec of the skin with a particular
focus on potential carcinogenic occupational expo-
sures. Eligible cases were histologically diagnosed
with primary invasive cutaneous sec in 1977-78
in 12 hospitals in the Montreal region; 2 controls
per case with no known history of skin cancer,
matched for sex, age and hospital of case diagno-
sis, were selected from those diagnosed in the
same period with specified dermatologic
conditions. Data on standard risk factors for skin
cancer were collected including skin type, occu-
pational and nonoccupational sun exposure as
well as ever-use of long- and round-tube sun-
lamps. The final study population, aged 65 years
on average, comprised 30% of all eligible
patients. There were 92 SCC cases, 4 of whom
reported any exposure to a long-tube sunlamp,
and 17 4 dermatological controls, one of whom
was so exposed, giving an odds ratio of 13.4 after
adjusting for age, sex, eye and hair colour, eth-
nicity, and nonoccupational sun exposure (p <
0.008). (Round-tube sunlamp results were not
reported.) [This study was conducted almost 30
years ago among elderly people; the Working
Group noted major drawbacks, including a hospital-
based study population, controls with skin conditions
and a very low response rate. The risk estimates
were based on a single exposed control, and no
details of artificial UV exposure were obtained.]
Bajdik et a/. (1996): Another study carried out in
Canada that aimed to assess phenotypic, solar
and non-solar risk factors for BCC and SCC of
the skin in men in the province of Alberta also
asked about exposure to non-solar UV light.
Cases were men with a first BCC or SCC histo-
logically diagnosed in 1983-84 and ascertained
through the Alberta Cancer Registry. Controls
were matched for age within 2 years from the
Alberta health insurance plan subscriber list.
Through personal interviews, information about
non-solar UV exposure such as exposure to weld-
ing torches, UV lights and sunlamps was
obtained, as well as standard risk factors. Results
were based on 226 BCC cases (72% of those
ascertained) , 180 sec cases (80%), and 406
eligible controls (71 %). Ever-use of a sunlamp
was reported by 8% of controls (33 of 404) and
9% of BCC cases (23 of 226), giving an odds
ratio of 1.2 (CI , 0.7-2.2); ever-use was reported
by 10% of sec cases (18 of 180), with odds ratio
of 1.4 (CI, 0.7-2.7). Risk estimates were adjusted
for age, skin and hair colour, ethnicity and lifetime
occupational sun exposure. [While this study was
population-based, it was conducted 20 years
ago, was restricted to men of unreported but likely
older ages, and no details of artificial UV expo-
sure were available.]
Corona eta/. (2001): A more recently conducted
hospital-based case-control study of causes of
BCC in Italy assessed non-solar factors as well
as phenotypic and solar factors. Cases of histo-
logically-confirmed BCCs diagnosed in
1995-1997 were ascertained on random days of
the week through a hospital for skin diseases in
Rome. Controls diagnosed with minor skin disor-
ders (e.g. warts, naevi) were drawn from the
same hospital but excluded if they had a history
of skin cancer or UV therapy. Questionnaire data
collected face-to-face included artificial UV expo-
sure as well as standard risk factors regarding
phenotype and patterns of sun exposure. Ever-
use of a sunbed or sunlamp was reported by 20%
of controls (31 of 158) and 11% of BCC cases
(17 of 166). After adjustment for age, sex, family
history of skin cancer, outdoor work and beach
exposure in youth, the relative risk estimate for
BCC was 0.6 (CI, 0.3-1 .2). [This study, carried out
1 0 years ago, had major shortcomings through its
design, namely a convenience sampling frame of
adult dermatologic patients. No details of expo-
sure to tanning appliances were obtained.]
Karagas eta/. (2002): A case-control study con-
ducted in the USA among New Hampshire resi-
dents assessed risk for BCC and SCC in relation
to exposure to artificial UV tanning appliances,
among other factors. Cases of skin cancer diag-
nosed in 1993-1995 were ascertained through a
network of dermatologists and pathology labora-
tories. Controls were a frequency-matched
sample of residents drawn from the Department
of Transportation listing (< 65 yrs) or Medicare
program list (> 65 yrs). Sunlamp/tanning bed use
Exposure to Artificial UV Radiation and Skin Cancer
and age at first and last use as well as standard
skin-cancer risk factor data were obtained
through personal interviews. The study popula-
tion comprised 603 sec cases and 293 sec
cases (78% of those eligible) and 540 (60%) eli-
gible controls. Fourteen percent of controls (75 of
539), 21% (127 of 601) of BCC cases and 22%
(63 of 229) of sec cases reported any exposure
to tanning appliances. After adjustment for age,
sex and sun sensitivity, risk estimates associated
with ever-use of a sunlamp in relation to BCC
were 1.5 (CI, 1.1-2.1) and to SCC, 2.5 (1.7-3.8),
and were similar in men and women. There was
a non-significant trend toward increased risk with
younger age at first use for SCC. Risks were
increased for both BCC (OR, 1.6; Cl, 1.1- 2.3)
and SCC (OR, 2.9; Cl, 1.8-4.7) for first use more
than 20 years previous to enrolment (before
1975). [The strengths of this study conducted 10
years ago were its population-based design and
its availability of some quantitative data regarding
sunlamp use. It lacked power to explore the asso-
ciations with age at first use versus years since
first exposure, and no data were available about
frequency of use.]
Walther et a/. (2004): The most recently pub-
lished study of the association of artificial UV
radiation and BCC was conducted in Germany,
based on 213 patients with sec diagnosed in the
previous 5 years and 411 controls from the same
dermatology department as the cases or the gen-
eral surgery department of the same hospitals.
During an interview patients were asked about
number of times a year they used indoor tanning
faci lities. On crude analysis there was no associa-
tion between recent history of BCC and use of
indoor tanning facilities more than 5 times a year
(OR, 0.7; Cl , 0.3-1 .5).
The meta-analysis was based on the five studies
reporting type-specific risk estimates (Table 17).
Chi-squared test and random effect models were
used to assess heterogeneity, as described on
page 26. Pooled relative risks suggested a signifi-
cant effect of exposure to indoor tanning facilities
for SCC, but not for BCC (Table 18).
The effect estimate seen for BCC was not
much influenced by the estimate reported by
Corona et a/. (2001 ), which indicated a protective
effect of artificial UV radiation for BCC (the weight
of this study was the lowest [w = 8.0]). As above,
this study was not specifically designed to investi-
gate exposure to artificial UV radiation, thus radia-
tion exposure data were not detailed. Excluding
this publication from the analysis changed the
pooled relative risk for BCC, although not substan-
tially (pooled RR, 1.39; Cl , 0.14-13.51).
Regarding SCC as an outcome, the study by
Aubry & MacGibbon (1985) reported findings for
only one type of sunlamp (long-tube type) and
was hospital-based. The weight of this study was
the lowest of this group (w=0.74); nevertheless,
the pooled relative risk for sec excluding this
study was neither stronger not more significant
(pooled RR, 2.16; Cl , 0.24- 19.53).
Funnel plot regression gave indication of no
publication bias (P=0.77 and 0.26 for BCC and
sec, respectively) but results based on so few
estimates are not reliable.
The study by Karagas et a/. (2002) gave the
most detai led results and the trends were
consistent with the results reported for
melanoma. The weight of this study was the high-
est (w = 23.8 for SCC and w = 36.8 for BCC) and
therefore its results were the most influential.
Quality of studies
Only one case-control study (Karagas et a/.,
2002) had a section designed specifically to
explore sunlamp/sunbed use in more detai l than
never/ever use. Results were adjusted for sun
sensitivity but not for sun exposure since adjust-
ment for sun exposure did not change the risk
estimates. Study participants who reported using
sunlamps or sunbeds were more likely to be
women, to be aged under 50 years, to have
sun-sensitive skin, more painful sunburns and a
history of frequent sunbathing (> 4 times per year)
than non-users. Based on age at first use, the rela-
tive risks for BCC and SCC were found to
increase by 10% (OR, 1.1; Cl , 0.9-1 .5) and 20%
(OR, 1.2; Cl, 0.9-1.6) respectively, for each decade
younger the person was at first use of an indoor
tanning facility. The effects of age at first use could
Epidemiological data on exposure to artificial UV radiation for cosmetic purposes and skin cancers
Table 17. Estimates included in the evaluation of an association of ever use of indoor
tanning facilities and risk for non-melanoma skin cancers
Reference Exposure Diagnosis
Aubry & McGibbon Long-tube sunlamp SCC
(1985) use
Bajdik et a/. (1996) Ever use of sun- BCC
lamps sec
Corona et a/.
(2001) Sun bed or sun- BCC
lamp use
Karagas et a/. Any tanning device BCC
(2002) use sec
Walther et a/. Exposure <! 5 BCC
(2004) times/year to artifi-
cial UV radiation/-
UV sunbeds
BCC, basal cell carcinoma; SCC, squamous cell carcinoma
not be distinguished from years since first use
because of the relatively small number of cases in
the study, and there were no semi-quantitative
measurements of artificial UV exposure (e.g. num-
ber of sessions per month, duration of use).
Other sources of exposure to artificial UV
Medical Use
Light treatment has been used for a large number
of medical conditions (see page 4), most
particularly for psoriasis.
(a) PUVA therapy in psoriasis patients: Most
long-term studies looking at risk for skin cancer
resulting from exposure to UV treatment collected
data from a significant number of psoriasis
patients treated with PUVA (see page 4 (b)).
Relative risk
(95% Cl)
13.4 (1.4-130.5)
1.2 (0.7-2.2)
1.4 (0.7-2.7)
0.6 (0.3-1.2)
1.5 (1.1-2.1)
2.5 (1.7-3.8)
0.7 (0.3-1.5)
Age, sex, eye and
hair colour, ethnicity,
sun exposure
Age, ethnic origin,
skin and hair
colour, occupational
sun exposure
Age, sex, pigmen-
tary traits, family
history skin cancer,
outdoor work, num-
ber of weeks spent
at beach before age
20 years
Age, sex, sun
There is clear evidence that PUVA increases the
risk for sec with a relatively short latency period,
although it is difficult to distinguish the contribu-
tion of PUVA from other factors, given that treated
patients have usually received multiple
carcinogenic treatments. For example, SCC in
psoriatic patients treated with PUVA commonly
have UV signature mutations rather than PUVA
signature mutations (Kreimer-Erlacher et a/.,
2003), suggesting that PUVA may act as a
promoter rather than an initiator.
Two large cohorts of psoriasis patients have
been followed up since the 1970s: one of 4799
patients in Sweden (Lindelof et a/., 1999) and
another of 1380 patients in the USA (Stern,
2001 ). In the Swedish cohort the relative risk for
sec was 5.6 in men (CI , 4.4-7.1) and 3.6 in
women (CI, 2.1-5.8). In the cohort in the USA,
one fourth of patients who received more than
2000 J/cm
developed an SCC (Stern & Laird,
Exposure to Artificial UV Radiation and Skin Cancer
Table 18. Meta-analysis of studies of exposure to artificial UV radiation and risk for non-
melanoma skin cancers
Diagnosis Number of
Summary relative risk
(95% Cl)
P-value X
2.25 (1.08-4.70)
1.03 (0.56-1.90)
BCC, basal cell carcinoma; sec, squamous cell carcinoma
1994). The same authors subsequently carried
out a meta-analysis of their own data and all pub-
lished studies with more than 150 patients (Stern
& Lunder, 1998), and found that patients exposed
to high doses of PUVA (more than 200 treat-
ments or more than 2000 J/cm
) had a 14-fold
higher risk for sec than those with <1 00 treat-
ments or <1 000 J/cm
exposure. The risk is
further increased when the patients have also
received methotrexate at some time (Stern &
Laird, 1994) and is greater still with the use of
cyclosporine (Marcil & Stern, 2001 ). There is no
evidence to date that bath PUVA increases the
risk for SCC (Hannuksela-Svahn eta!., 1999) but
the data available relate to only 944 patients who
received relatively low total PUVA doses.
The risk for melanoma after PUVA treatment
is more controversial. In the cohort in the USA,
discussed above, an increased risk for melanoma
has been reported (Stern, 2001 ). Of the 822 par-
ticipants with long-term follow-up, 44% had at
least 200 PUVA treatments and therefore are
called high exposure patients. Sixteen of the
1380 patients developed an invasive melanoma
and 6 developed a melanoma in situ. The authors
reported a 1 0-fold increase in the incidence of
invasive melanoma compared with population
rates in the 27 months prior to publication of the
article. Within the cohort, the risk for melanoma
was greater in those with fair skin (Fitzpatrick skin
type) and those who received high doses of
PUVA (incidence rate ratio, 2.6; Cl, 1.0-.6) for
more than 200 treatments compared with less
than 200. The risk also appeared to have a long
latency in that an elevation in risk appeared only
after 15 years. There did not appear to be any
increased risk in patients who were also treated
with ionizing radiation or methotrexate.
The Swedish cohort (Lindelof et al., 1999)
reported no increased risk for melanoma. This
study was much larger than the study in the USA
and the patients were tracked using the Swedish
Cancer Registry, thereby allowing "complete"
follow-up. Of the 2343 men in the cohort, 8
developed a melanoma compared with the 7.3
expected, and of the 2456 women, 7 developed a
melanoma compared with the 6.3 expected. The
length of follow-up was impressive in this cohort,
as the average length was 16 years and 1 038
patients had been followed for more than 19 years.
Given the considerable size and the duration
of follow-up of the Swedish cohort, the findings
from this cohort are the more persuasive of the
two studies. The difference in findings, however,
remains unexplained. In the Swedish cohort a
proportion of patients had had bath PUVA, which
tends to be associated with lower UVA doses.
There were differences in the treatment protocols
as well (Honigsmann, 2001 ), in that in Europe
schedules are individualized after light testing,
more commonly resulting in reduced time to
clearing and lower doses per treatment course.
These differences may explain the discrepant
risk estimates, but it cannot be excluded that the
data from the study in the USA are subject to
bias, not least because follow-up was substan-
tially incomplete.
Overall, there is a postive association
between PUVA and risk for SCC and there
appears to be a dose-response effect. The risk
was greater for fair-skinned people. The risk for
melanoma is much less clear, even in
fair-skinned populations. The positive dose-
response relationship in the study in the USA
supports the interpretation that the association is
causal. It seems likely, however, that the risk is
associated with high doses of PUVA, is relatively
small and is observed after a long latency.
Epidemiological data on exposure to artificial UV radiation for cosmetic purposes and skin cancers
The data from PUVA studies are important in
that they include large numbers of people who
were studied prospectively. They cannot however
be extrapolated to exposure to tanning appli-
ances because of the presence of psoralen.
Furthermore, the total UV dose received by
psoriasis patients is considerably less than that
received by long-term users of indoor tanning
(b) Broadband and narrow-band UVB in psoriasis
patients: The evidence relating to long-term risk
for skin cancer after UVB therapy is scanty. In the
PUVA cohort study from the United States, there
was no discernible additional effect of exposure
to UVB (Stern & Laird, 1994}. In a study of psori-
atics treated with coal tar and UVB in the 1950s
followed up for 25 years, there was no demon-
strable increased risk for skin cancer, though the
numbers treated were relatively small (n = 280)
(Pittelkow et a/., 1981 ). In a small study of 195
German psoriatics treated with broadband
(n = 69) or narrow-band UVB (n =126) from 1994
to 2000 only one skin cancer had occurred by
2004. This was an in-situ melanoma which devel-
oped in the same year that narrow-band UVB
therapy was begun (Weischer et a!., 2004}.
Though these data are reassuring they cannot
exclude a small increased risk nor a large
increased risk in patients treated with high doses.
(c) UV treatment of other skin diseases: The
immunomodulatory effects of UV radiation are
utilized in the treatment of a variety of skin dis-
eases other than psoriasis. Many of the patients
treated are at increased risk for skin cancer even
without PUVA because of the nature of their
dermatosis (e.g. viti ligo). Others are at further
increased risk because of immunosuppression
which may both characterize the skin disease
and its treatment, such as graft versus host
disease (GVHD) (Furlong eta/., 2002) or cuta-
neous T-cell lymphoma.
A series of 1 03 patients with steroid-resistant
GVHD treated with PUVA received a mean dose
of 41 J/cm
between 1994 and 2000. Only one
sec has developed in this cohort to date
(Furlong eta!., 2002).
PUVA is also very useful, although not cura-
tive, in the treatment of cutaneous T cell lym-
phoma (CTCL) when it is commonly used as part
of multi-modality treatment programmes with
other drugs contributing to risk such as cytotoxi-
cs (McGinnis et at., 2003). Narrow-band UVB has
been reported to be as effective as PUVA in the
treatment of early CTCL in one retrospective
study (Diederen et at., 2003). There is no doubt
that in this patient population there was an
increased risk for sec but it is difficult to appor-
tion risk to PUVA. The risk for melanoma was
reported in a very small series of patients and
therefore cannot be assessed (McGi nnis et a/.,
(a) Fluorescent tubes: Household lights emit
significant amounts of UV radiation (Sayre eta/.,
2004) and several case-control studies have
addressed risk for melanoma associated with
such exposure. The earl iest study suggested an
elevated risk associated with exposure to fluores-
cent lights at work (Beral et at., 1982) but alt sub-
sequent studies failed to identify such a risk
(Rigel et at., 1983; Osterlind et a/., 1988; Walter
et at., 1992; Holly eta/., 1995).
(b) Full spectrum lamps: No data were available
to the Working Group regarding exposure to full-
spectrum lamps intended for domestic and public
use and risk for skin cancer.

Exposure to Artificial UV Light and Skin Cancer
Effects of artificial UV radiation not relevant to skin carcinogenesis
Cutaneous diseases
Skin ageing
Skin ageing is a phenomenon comprising intrin-
sic processes that are largely genetically deter-
mined and extrinsic ageing (or photo-ageing) that
is largely related to sun exposure (Jenkins,
2002) . Both UVB (Del Bino eta/., 2004) and UVA
(Marrot eta/. , 2004) are held to be mediators of
these effects. Intrinsic ageing is characterized by
thinning of the skin accompanied by reduction in
collagen levels. These changes are thought to
result at least in part from diminished cellular pro-
liferative activity in the skin (inevitable cellular
senescence) and from increased expression of
enzymes that degrade the collagen, such as the
metallo-proteinases (Jenkins, 2002). The process
is undoubtedly complex, and one of the drives to
cellular senescence may be chronic exposure to
oxidative stress.
The changes resulting from sun exposure
that are perceived as ageing are loss of elastici-
ty, pigmentary change and deep wrinkling
(Leyden, 1990). Most of these changes result
from damage to the dermis, which is visible his-
tologically as elastotic material. This material is
comprised of degenerate elastic fibres and newly
synthesized dysfunctional elastotic material.
Similarly there appears to be a reduced amount
of collagen I in the dermis and increased
amounts of degenerate collagen. The metallo-
proteinases mediate this degradation, at least in
part, and their activity appears to be increased
both by age and by sun exposure. Increased age
is associated with a diminished ability to repair
damage induced by exposure to UV radiation
(Takahashi et at., 2005).
Comparatively few epidemiological studies
have addressed photo-ageing of the skin, not
least because of the difficulties of measuring it
accurately. Some authors have suggested that
ultrasound measurement is of value (Gniadecka
& Jemec, 1998); others have used silicone
(Green, 1991; Fritschi et a/., 1995) to create
moulds to allow an estimate of the topography of
the skin, a method which appears to be better
evaluated. A large study performed in
Queensland, Australia demonstrated premature
ageing of the skin in a population excessively
exposed to the sun (Green, 1991 ). This was more
marked in men who reported outdoor work or
leisure, and especially those with fair skin. The
presence of photo-ageing was correlated with
skin cancer. The relationship between non-
melanoma skin cancer and solar keratoses is
held to be clear and straightforward (Green et at.,
1999). Experimental studies on UV exposure and
photo-ageing have been reviewed (IARC, 1992).
Very few studies have investigated the rela-
tionship of artificial UV exposure to ageing in
humans. Lentigos similar to PUVA freckles have
been reported to be induced by exposure to tan-
ning appliances (Roth et at., 1989; Kadunce et
at. , 1990), which is of concern given the evidence
that the risk for skin cancer is increased in PUVA
patients. A number of case reports have
described an extreme form of cutaneous ageing
which resulted from very frequent exposure to
tanning appliances in fair-skinned people (Poh-
Fitzpatrick & Ellis, 1989). There have been no
informative epidemiological studies of the role of
indoor tanning facilities in the induction of photo-
There is some evidence that cigarette smoke
exacerbates photo-ageing of the skin (IARC,
2004; Placzek eta/., 2004).
Other skin diseases caused or exacerbated
by exposure to UV radiation
A wide variety of dermatoses are exacerbated by
sun exposure, such as atopic eczema or psoria-
sis if sunburn occurs. Some skin diseases are
directly provoked by sun exposure, the most
common of which is polymorphic light eruption,
which is common in women. It has been reported
in around 20% of healthy women (Millard et at.,
Effects of artificial UV radiation not relevant to skin carcinogenesis
2000). Variants occur, such as a blistering erup-
tion seen on the ears in childhood or actinic
prurigo, where itchy papules and nodules devel-
op after sun exposure. Such photodermatoses
are a nuisance but otherwise relatively trivial
problem. Exposure to tanning appliances may
precipitate such dermatoses (O'Toole & Barnes,
1995). Medical use of artificial UV radiation may
be used to control polymorphic light eruption if
used carefully as a means of desensitization.
Photosensitivity is usual in patients with
lupus even in the absence of a history of a sun-
evoked eruption (Sanders eta/., 2003), and light
testing reveals that the majority of patients react
to both UVA and UVB (Sanders eta/., 2003). The
cutaneous manifestations of lupus are also com-
monly precipitated by exposure to the sun. Photo-
testing with artificial UV radiation sources has
been reported to provoke cutaneous lupus
(Marguery et a/., 2005), and therefore it seems
likely that it may also be provoked by other
sources of artificial radiation such as tanning
More significant photodermatoses occur
more rarely, such as chronic actinic dermatosis,
in which persistent sun-induced eczema occurs.
It is a rare condition, usually seen in elderly men.
It may develop from an allergic dermatitis for
example to pollen or fragrances.
Much more significant are the porphyrias in
which sun exposure may trigger photosensitivity.
The varieties that induce photosensitivity are
variegate porphyria (Mustajoki, 1980) (most
common in South Africans of Dutch descent),
erythropoietic porphyria (Goerz, 1979) and
porphyria cutanea tarda (PCT) . PCT is the most
common and in 80% of cases occurs because of
exposure to estrogens or alcohol. Use of indoor
tanning facilities or other artificial UV sources -
even fluorescent lights - by patients with latent
porphyria is potentially very serious as a result of
the possible induction of sunburn.
The overall dose of UVB and UVA incurred
by most people using bright light therapy is likely
to be considerably less than that received by pso-
riasis patients treated with PUVA. Exposure is
also likely to be limited to the face. It seems likely
therefore that the theoretical risk will relate to
non-melanoma skin cancer rather than
melanoma, but there are no relevant data from
epidemiological studies at present to inform. It
would seem very reasonable however to con-
clude that lamps emitting low levels of UVA would
be preferred to those emitting higher levels.
Case reports also suggest that use of indoor
tanning facilities is associated with development
of drug-induced photodermatoses and exacerba-
tion of lupus erythematosus (Spencer &
Amonette, 1995).
Drug-induced photosensitivity
A variety of commonly used drugs increase cuta-
neous sensitivity to the sun and to artificial UV
sources, and are predicted therefore to increase
the risk for skin cancer. Most drugs have a pho-
totoxic effect rather than a photo-allergenic one
(Moore, 2002). Oral photosensitisers include
tetracyclines, amioderone, diuretics, non-steroidal
anti-inflammatory drugs (NSAIDs) and chlorpro-
mazine (Moore, 2002). Diuretics, antibiotics and
NSAIDs are very widely used drugs, and their
phototoxic effects therefore have the potential to
affect a significant proportron of the population.
Topical agents include plant-derived photosensi-
tisers (psoralens) such as bergamot, widely used
in perfumed products. Use of perfumes has the
potential to increase the photo-damaging effects
of indoor tanning appliances.
Effects on the eyes
Conditions linked to sub-chronic and chronic
exposure to solar uv include pterygium and sec
affecting the cornea; cataract, affecting the ocular
lens; and acute macular degeneration affecting
the retina (Tomany et a/., 2004). Of these,
cataracts of the nuclear and cortical types are the
most widespread and serious UV-related eye
conditions. There is an inverse association
between latitude of residence and cataract sur-
gery in Medicare program data from the USA
(Javitt & Taylor, 1994), and epidemiological studies
conducted in Australia, China, and the USA (see
Taylor, 1994, for review) support a role of UV
exposure in cataract development. Risk for
Exposure to Artificial UV Radiation and Skin Cancer
cortical cataract (opacity of the outer lens) is
related to increasing cumulative UVB exposure,
while risk for nuclear cataract (opacity of the cen-
tral lens) has been shown to be significantly
increased with increasing UV exposure in young
adulthood, consistent with the successive laying
down with age of outer lens fibres on the cortical
layer exposed in earlier life (Neale eta/., 2003).
With regard to artificial UVB, there is sufficient
experimental evidence that exposure causes
cortical lens opacity in the eyes of laboratory ani-
mals (IPCS, 1994).
Intraocular melanoma
Early-life exposure to sunlight may be important
in the development of intraocular melanoma
(Tucker eta/., 1985; Seddon eta/., 1990), more
specifically of choroidal melanoma (Moy, 2001 ).
This is consistent with the observation that after
childhood most UV radiation is screened by the
lens (Zigman, 1983; Lerman, 1984).
Welding equipment and tanning appliances
are sources of intense UV radiation. Five out of
eight epidemiologic studies found a significantly
increased risk for ocular melanoma with welding
exposure, with relative risks ranging from 1.9 to
10.9 (Tucker et a/., 1985; Holly et a/., 1990;
Seddon et a/., 1990; Siemiatycki, 1991 ; Ajani et
a/. , 1992; Holly eta/., 1996; Guenel eta/., 2001 );
in contrast , one study conducted in nine
European countries found an increased risk only
in one country (Lutz eta/., 2005) .
Four case-control studies have examined
the risk for intraocular melanoma in relation to
exposure to UV radiation from sunlamps (Table
19). Whi le the earliest study only found a non-sig-
nificant trend in risk according to frequency of
sunlamp use (Tucker eta/., 1985), the three more
recent studies consistently found an increased
relative risk, ranging from 1.7 to 3.6 (Holly eta/.,
1990; Seddon eta/. , 1990; Vajdic eta/., 2004) .
UV exposure and vitamin D
Vitamin D is an essential nutrient, generally
quantified by measuring circulating levels of 25-
hydroxyvitamin D. There are three major sources
of vitamin D: photosynthesis in the skin, ingestion
in the diet and oral supplementation. Worldwide,
photosynthesis from sunlight is the most common
source of vitamin D.
Vitamin D formation by photosynthesis
Previtamin D3 is produced from 7-dehydroxycho-
lesterol (provitamin D
) by the direct photolytic
action of UVB. The precursor, 7-dehydroxycho-
lesterol, is abundant in human skin although lev-
els decrease with age (Holick et a/., 1989). On
exposure of the skin to sunlight, 7-dehydroxycho-
lesterol in epidermal and dermal cells absorbs
UVB radiation to form previtamin D
. Previtamin
is thermodynamically unstable and is rapidly
transformed by rearrangement of its double
bonds to form vitamin D
{here called vitamin D)
before entering the circulation. Vitamin Dis a pro-
hormone that is converted by 25-hydroxylation in
the liver to the intermediate metabolite 25-
hydroxyvitamin D, which is the main ci rculating
and storage form. With physiological demands for
calcium and phosphorus, 25-hydroxyvitamin D
undergoes 1 a-hydroxylation in the kidney to form
the active hormone, 1 ,25-dihyroxyvitamin D.
Blood levels of 25-hydroxyvitamin D reflect the
availability of vitamin D (Osborne & Hutchinson,
Main target organs for 1 ,25-dihydroxyvitamin
D include the intestine, kidney and bone, but
nuclear receptors have been found in over 30 dif-
ferent tissues, reflecting its many other actions
besides parathyroid activity and serum calcium
homeostasis, analogous to those of classical
steroid hormones. 1 ,25-dihydroxyvitamin D is
also an antiproliferative, prodifferentiation and
proapoptotic agent (Osborne & Hutchinson,
Dietary sources of vitamin 0
There are only a few foods (cod liver oil , oily fish
such as salmon, mackerel and sardines) that are
naturally rich in vitamin D, so in many countries
where oily fish are not widely consumed, food
fortification or vitamin supplements may be needed.
In a global review of vitamin D intake, wide varia-
tions were found in food fortification
practices and contributions from supplement use
Table 19. Case-control studies of exposure to artificial UV radiation and risk for intraocular melanoma
Reference Location, Cases and Controls Age Adjusted Relative Characteristics Comments
Period of (years) Risk Assessed
Tucker era/. Philadelphia, PA, 444 intraocular melanoma NR 1.4 (0.9-2.2); Eye colour, complexion Telephone interview (45 min.); estimates
(1985) USA, January (1 hospital) with cataract: and hair colour, adjusted for history of cataract; sunlight
197 4- June 1979 424 hospital controls with 2.3 (0.9- 5.9) corrective lenses, sun exposure is an important risk factor for
detached retina, matched without cataract: exposure during leisure intraocular melanoma (persons born in the
on age, sex and period of 1.2 (0.7-2.0) time, sun protection, South: RR 2.7 (1.3-5.9). Trend in RR
diagnosis years lived in the South according to frequency of sunlamp use
rising to twofold for frequent use (P=0.10)
Holl y eta/. (1990) 11 Western 407 cases (Ocular 20--74 Exposure to UV or Eye and hair colour, Telephone interview; adjustment on eye
States. USA. Oncology Unit, UCSF); blacklight: naevi, freckles, tendency colour, coffee, effect of 0.5 h exposure to
January 1978-- 870 controls (random digit 3.59 (1.57-8.70) to sunburn, other eye midday summer sun, other UV exposure.
February 1987 dialling in the same Welding burn, conditions, tobacco,
geographic area as the sunburn to eye, coffee. tea. alcohol
patients), snow blindness:
2 age/sex matched 7.17 (2.5-20.57) m
controls per patient
Seddon et a/. New England, 197 cases 17--88 Use of sunlamps: Ancestry from Northern Only choroid and ciliary body melanomas.
(1990) USA 385 matched population 3.4 (1.1-10.3); or Southern latitudes, 2 independent comparisons. Occasional or

controls (random digit with random digit latitude of residence, skin frequent versus never use. ![
dialling) dialled controls colour, naevi
2.3 (1.2--4.3) with
337 cases; sibling controls
all USA 800 sibling controls
Vajdic eta/. (2004) Australia, 290 cases (246 with 19--79 Use of sunlamps: Eye colour, host Population-based prospective recruitment
1996-1998 melanoma of choroid or 1. 7 (1.0--2.8) characteristics. lifetime of cases (all ophthalmologists and cancer
ciliary body); Welding: residence, work calendar, registries). Telephone interview. Risk

893 population controls 1.2 (0.8--1 .7) sun exposure, sun increases with increasing duration of use 0
(electoral rolls) protective wear and is greater for exposures begun before
the age of 21 years and after 1980.
Sunlamp use or welding not associated
with iris or conjunctival melanoma.
NR, not reported.

Exposure to Artificial UV Radiation and Skin Cancer
(Calvo et a/., 2005). In Canada and the USA,
where fortification is mandatory for staple foods
such as milk and margarine and optional for other
classes of food, vitamin D intake was generally
2-3 f.!g higher than in either Australia, Ireland,
Scotland or the United Kingdom where fortifica-
tion of staples like margarine and breakfast cere-
als is optional, or European and other countries
where food fortification is restricted. Mean daily
vitamin D intakes were reported to be highest in
young adult Caucasian men and women in North
America (8.1 and 7.3 f.!g/d) due to milk fortifica-
tion and in Japanese women (7.1 f.!g/d) due to
high fish consumption. Norwegian men and
women, who also have high fish consumption,
had higher levels (6.8 and 5.9 mg/d) than their
British counterparts (4.2 and 3.7 f.,lg/d).
Contributions by dietary supplements to mean
daily vitamin D intakes ranged from 49% in
Norwegian women to 12% in British men, and on
average contributions from supplements
increased with age and were more common in
women (Calvo eta/., 2005) .
Vitamin D and exposure to artificial UV radia-
tion for tanning purposes
Available data are inadequate to assess the
effect of exposure to UV in indoor tanning facili-
ties on vitamin D status. Our current understanding
of the photosynthesis of vitamin D in the skin
would suggest that this type of artificial UV expo-
sure would be effective in induction of vitamin D
photosynthesis only to the extent that it contains
UVB, as opposed to UVA radiation. Practically
speaking, the usefulness of these facilities for
correcting vitamin D insufficiency is limited by the
inability of consumers to ascertain the UVB flux
to which they are being exposed in a tanning
session, the expense and inconvenience of these
sessions compared with oral vitamin D supple-
mentation, and the other health consequences of
using these facilities, as outlined in the other
chapters of this document.
The nature of the network of photo- and ther-
moreactions that are involved in vitamin D syn-
thesis in vitro is well established. As measured in
vitro the greatest sensitivity of the conversion of
7-dehydroxycholesterol to previtamin D lies in the
UVB part of the solar spectrum, 280-320 nm. It is
therefore similar to the erythema action spec-
trum, except for a sharper downturn in provitamin
D absorptivity in the UVA spectral region from a
maximum at 282 nm (Galkin & Terenetskaya,
Seasonal and latitude variations in UVB
intensity markedly affect vitamin D synthesis,
with lowest relative production occurring at high lat-
itudes during winter months. Also, ageing
decreases the capacity of skin to produce vitamin
D (Holick et al., 1989). Finally, compared with fair-
skinned people, those with darkly pigmented skin
are less efficient at producing vitamin D and
require 10-50 times the level of sun exposure to
produce the same amount (Clemens et al.,
1982). In addition to season, latitude, ageing, and
skin pigmentation, vitamin D photosynthesis may
be influenced by factors that affect the intensity of
skin exposure to UVB.
For light-skinned adults, a few minutes per
day with the face and hands in bright sunshine is
sufficient to cover daily needs in vitamin D.
Intense UVB exposure may generate little vitamin
D beyond that achieved by more modest expo-
sure because previtamin D and vitamin D can
readily convert to other photoproducts that have
little or no vitamin D action.
Vitamin D and xeroderma pigmentosum patients
A further indication that necessary amounts of
vitamin D may be provided through dietary
sources comes from a study of patients suffering
from xeroderma pigmentosum, a rare disease
associated with a deficiency in UV-induced DNA
lesion repair (Setlow eta/., 1969), and character-
ized by extreme sensitivity to sunlight (Kraemer
et al., 1994). To prevent the development of skin
cancers at an early age, these patients wear
protective clothing and use sunscreens when
outdoors. A six-year follow-up study of eight
children with xeroderma pigmentosum showed a
normal vitamin D intake and that normal vitamin
D levels can be maintained in ambulatory
patients despite rigorous sun protection (Sollitto
et al., 1997).
Summary and conclusion
UV radiation wavelengths range between 100 nm
and 400 nm and are broadly categorised into UVA
(>315-400 nm), UVB (>28Q-315 nm) and UVC
(1 OQ-280 nm). The portion of solar UV radiation
that reaches the earth's surface is composed pre-
dominantly of UVA and less than 5% UVB.
The sun is the main source of UV for most
individuals. Sources of artificial UV radiation are
used during indoor tanning, for medical appl ica-
tions and in some occupations. Indoor tanning
facilities in general deliver higher relative intensi-
ties and higher proportions of UVA compared
with solar UV radiation, but there are wide varia-
Several national and international organisa-
tions have presented recommendations regard-
ing the use of indoor tanning facilities, but few
countries regulate access and use.
The few studies that have addressed the bio-
logical changes in the skin induced by indoor
tanning have shown that they are similar to those
induced by sunlight.
Many studies have substantiated the carcino-
genic effects of UV radiation. Experimental stud-
ies in humans have shown that in the basal lay-
ers of the epidermis, where melanocytes are
located, UVA induces more DNA damage than
does UVB.
Both UVA and UVB radiation can affect the
immune system: while UVB induces immunosup-
pression at both the local and systemic levels,
UVA does not induce systemic immune suppres-
sion. Exposure to tanning appliances has also
been shown to induce changes in the skin
immune system, including reduced skin test
responses, changes in lymphocyte populations,
and depression of NK cell activity.
Prevalence of indoor tanning varies greatly
between countries; it is widespread in Northern
Europe and North America, particularly among
women and young people. Indoor tanning has
increased considerably since the early 1980s.
Several studies show common use by adoles-
Exposure to Artificial UV Light and Skin Cancer
cents, and sometimes by children. The most fre-
quent motivations for indoor tanning are the
acquisition of a so-called "safe" tan and skin
preparation before sun exposure. Limited evi-
dence suggests that compliance with recommen-
dations and regulations by indoor tanning facility
operators and customers is poor.
Twenty-three published studies (22
case-control, one cohort) in light-skinned popu-
lations investigated the association between
indoor tanning and risk for melanoma, and 7
case-control studies for keratinocytic skin can-
cers. Characterisation of exposure was highly
variable across reports.
The summary relative risk for ever versus
never use of indoor tanning facilities from the 19
informative studies was 1.15 (1.0Q-1.31 ). When
the analysis was restricted to the nine population-
based case-control studies and the cohort study,
the summary relative risk was 1.17 (0.96-1 .42).
There was no consistent evidence for a
dose-response relationship between indoor tan-
ning exposure and risk for melanoma.
All studies that examined age at first expo-
sure found an increased risk for melanoma when
exposure started before approximately 30 years
of age, with a summary relative risk estimate of
1.75 (1.35-2.26).
Studies on exposure to indoor tanning appli-
ances and squamous cell carcinoma found some
evidence for an increased risk for squamous cell
carcinoma, especially when age at first use was
below 20 years. Studies on basal cell carcinoma
did not support an association with use of indoor
tanning faeilities.
Investigation of the association between
indoor tanning and skin cancers poses challeng-
ing problems, as the fashion of indoor tanning is
still very recent. Associations after long latency
periods, such as may be expected for melanoma
and basal cell carcinoma, may not yet be
Artificial UV sources are used to treat a
variety of skin conditions, predominantly psoriasis.
Exposure to Artificial UV Radiation and Skin Cancer
Broadband UVB has been used for many years
and, more recently, narrow-band UVB, but there
are few data on which to base estimates of risk
for skin cancer. PUVA therapy increases the risk
for squamous cell carcinoma. Data concerning
the risk for melanoma as a result of PUVA thera-
PY are conflicting, but to date it seems likely that
any increased risk for melanoma is small and that
the latency is in excess of 20 years.
Case reports suggest that use of indoor tan-
ning facilities is associated with the development
of drug-induced photodermatoses and exacerba-
tion of lupus erythematosus.
UV exposure is related to eye damage,
including cataracts, corneal squamous cell carci-
noma and ocular melanoma. Several epidemio-
logical studies have shown an association
between artifiCial UV exposure and ocular
melanoma, especially if exposure occurred in
adolescence or young adulthood.
Sources of vitamin D include photosynthesis
in the skin in response to exposure to UVB radi-
ation, oral intake from consumption of food and
dietary supplements. In cases of insufficiency,
supplementation through oral intake is recom-
mended. Indoor tanning may produce vitamin D
photosynthesis in the skin depending on the
amount of UVB radiation, if any, in their emission
spectrum, although the emission spectrum is
generally unknown to consumers and operators.
The use of indoor tanning facilities is widespread
in Europe and North America, and this impels con-
sideration of the risk for adverse health conse-
quences, particularly melanoma. Consideration is
hampered by the relative recency of widespread
use and the limitations of available studies.
Our systematic review of published studies,
conducted mainly in North America and Europe,
of the association of indoor tanning facility use
with melanoma revealed an association of early
age at first use (less than approximately 30
years) with melanoma risk. These studies consis-
tently indicated a moderate strength of associa-
tion, with a summary relative risk of 1.75
(1 .35-2.26). The association with ever use of
these facilities, or use more than 15 to 20 years
prior to diagnosis of melanoma, was weak, and
evidence regarding a dose-response relation-
ship was scanty. The evidence is limited by varia-
tion in characterization of exposure, potential
confounding by sun exposure or other variables,
and the low power to detect associations that
become evident only following a prolonged lag
period after exposure.
The association between indoor tanning facility
use and melanoma risk is consistent with the
knowledge that melanoma is caused by exposure
to solar radiation. Exposure to sunlight in child-
hood has been established as an important
contributing factor for melanoma risk in adults.
Although the contexts of exposure to sun and of
indoor tanning differ, both deliver UV radiation,
and the health effects would therefore be expected
to be similar. The limited evidence for an associ-
ation between indoor tanning and squamous cell
carcinoma is consistent with the known associa-
tion of sun exposure with that cancer. In light of
the known effects of UV radiation on the skin, the
biological plausibility of a causal association
between use of indoor tanning facilities and risk
for melanoma and squamous cell carcinoma is
On balance, the evidence pertaining to the
strength, consistency, dose-response and tem-
poral sequence of the association of the use of
indoor tanning equipment with melanoma risk,
and of the coherence and biologic plausibility of
the association, leads us to conclude that there is
convincing evidence to support a causal relation-
ship, particularly with exposure before the age of
35 years. This evidence is strongly suggestive
and further studies could clarify our understand-
ing of this association and allow more definitive
We are cognizant of the importance of this
issue for the health of light-skinned populations.
The strength of the existing evidence suggests
that policymakers should consider enacting
measures, such as prohibiting minors and dis-
couraging young adults from using indoor tanning
facilities, to protect the general population from
possible additional risk for melanoma.
Adam SA, Sheaves JK, Wright NH, et al. (1981) . A
case-control study of the possible association
between oral contraceptives and malignant
melanoma. Br J Cancer, 44:45-50.
AFSSE (2005) [Ultraviolet radiation - Current
knowledge relating to exposure and health
risks], Report, Maisons-Aifort, France (in
Agar N, Young AR (2005) . Melanogenesis: a photo-
protective response to DNA damage? Mutat
Res, 571 :121-132.
Agar NS, Halliday GM, Barnetson RStC, et al.
(2004). The basal layer in human squamous
tumors harbors more UVA than UVB fingerprint
mutations: A role for UVA in human skin car-
cinogenesis. Proc Natl Acad Sciences USA,
Ajani UA, Seddon JM, Hsieh CC, et al. (1992).
Occupation and risk of uveal melanoma: an
exploratory study. Cancer, 70:2891-2900.
Atillasoy ES, Seykora JT, Soballe PW, et al. (1998).
UVB induces atypical melanocytic lesions and
melanoma in human skin. Am J Pathol,
152:1179-1 186.
Aubry F, MacGibbon B (1985). Risk factors of squa-
mous cell carcinoma of the skin. Cancer,
55:907- 911.
Autier P (2004). Perspectives in melanoma preven-
tion: the case of sunbeds. Eur J Cancer,
Autier P, Dore JF (1998) . 1nfluence of sun exposures
during childhood and during adult hood on
melanoma risk. lnt J Cancer, 77:533-537.
Autier P, Joarlette M, Lejeune F, et al. (1991).
Cutaneous malignant melanoma and exposure
to sunlamps and sunbeds: a descriptive study in
Belgium. Melanoma Res, 1:69-74.
Autier P, Dore JF, Lejeune F, et al. (1 994), for the
EORTC Melanoma Cooperative Group.
Cutaneous malignant melanoma and exposure
to sunlamps or sunbeds: an EORTC multicenter
case-control study in Belgium, France and
Germany. lnt J Cancer, 58:809-813.
Bajdik CD, Gallagher RP, Astrakianakis G, et al.
( 1996) . Non-solar ultraviolet radiation and the
risk of basal and squamous cell skin cancer. Br
J Cancer, 73:1612-1614.
Exposure to Artificial UV Light and Skin Cancer
Banks BA, Silverman RA, Schwartz RH, et al.
(1992). Attitudes of teenagers toward sun expo-
sure and sunscreen use. Pediatrics, 89:40--42.
Bataille V, Winnett A, Sasieni P, et al. (2004).
Exposure to the sun and sun beds and the risk of
cutaneous melanoma in the UK: a case-control
study. Eur J Cancer, 40:29-435.
Bataille V, Boniol M, DeVries E, et al. A multicentre
epidemiological study on sunbed use and cuta-
neous melanoma in Europe. Eur J Cancer 2005,
41 : 2141-2149.
Bech-Thomsen N, Wulf HC, Poulsen T, et al. (1991) .
Photocarcinogenesis in hairless mice induced
by ultraviolet A tanning devices with or without
subsequent solar-simulated ultraviolet irradia-
tion. Photodermatol Photoimmunol Photomed,
Bech-Thomsen N, Poulsen T, Christensen FG, et al.
(1992). UVA tanning devices interact with solar-
simulated UV radiation in skin tumor develop-
ment in hairless mice. Arch Dermatol Res,
Beitner H, Norell SE, Ringborg U, et al. (1990).
Malignant melanoma: aetiological importance of
individual pigmentation and sun exposure. Br J
Dermatol, 122:43-51.
Beral V, Evans S, Shaw H, Milton G (1982).
Malignant melanoma and exposure to fluores-
cent lighting at work. Lancet, 2:29Q-293.
Berking C, Takemoto R, Binder RL, et al. (2002).
Photocarcinogenesis in human adult skin grafts.
Carcinogenesis, 23: 181-187.
Black G, Matzinger E, Gange RW (1985). Lack of
photoprotection against UVB-induced erythema
by immediate pigmentation induced by 382 nm
radiation. J Invest Dermatol, 85:448-449.
Bohm M, Wolff I, Scholzen TE, et al. (2005). Alpha-
Melanocyte-stimulating hormone protects from
ultraviolet radiation-induced apoptosis and DNA
damage. J Bioi Chem, 280:5795-5802.
Boldeman C, Beitner H, Jansson B, Nilsson B, Ullen
H (1996) . Sunbed use in elation to phenotype,
erythema, sunscreen use and skin diseases. A
questionnaire survey among Swedish adoles-
cents. BrJ Dermatol, 135:712-716.
Boldeman C, Jansson B, Nilsson B, Ullen H (1997).
Sunbed use in Swedish urban adolescents relat-
ed to behavioral characteristics. Prev Med,
Exposure to Artificial UV Radiation and Skin Cancer
Boldeman C, Branstrom R, Dal H, et al. (2001).
Tanning habits and sunburn in a Swedish popu-
lation age 13-50 years. Eur J Cancer,
Boyd AS, Shyr Y, King LE (2002). Basal cell carci-
noma in young women: An evaluation of the
association of tanning bed use and smoking. J
Am Acad Dermatol, 46:706-709.
Brandberg Y, Ullen H, Sjoberg L, Holm LE (1998) .
Sunbathing and sunbed use related to self
image in a randomized sample of Swedish ado-
lescents. Eur J Cancer Prev, 7:321-329.
Brash DE, Ziegler A, Jonason AS, et al. (1996).
Sunlight and sunburn in human skin cancer:
p53, apoptosis, and tumor promotion. J Jnvestig
Derma to/ Symp Proc, 1:136-142.
Bykov VJ, Marcusson JA, Hemminki K (2001 ).
Protective effects of tanning on cutaneous
melanoma. Dermatology, 202:22-26.
Calvo MS, Whiting SJ, Baron CN (2005). Vitamin D
intake: A global perspective of current status. J
Nutr, 135:31 Q-316.
Chen YT, Dubrow R, Zheng T, et al. (1998). Sunlamp
use and the risk of cutaneous malignant
melanoma: a population-based case-control
study in Connecticut, USA. tnt J Epidemiol,
Clemens TL, Adams JS, Henderson SL, et al.
(1982) . Increased skin pigment reduces the
capacity of skin to synthesise vitamin D
Lancet, 319:74-76.
Clingan PH, Berneburg M, Petit-Frere C, et al.
(2001). Contrasting effects of an ultraviolet B
and an ultraviolet A tanning lamp on interleukin-
6, tumour necrosis factor-alpha and intercellular
adhesion molecule-1 expression. Br J Oermatol,

Cokkinides VE, Weinstock MA, O'Connell MC, Thun
MJ (2002). Use of indoor tanning sunlamps by
US youths, ages 1118 years, and by their par-
ents or guardian caregivers: prevalence and cor-
relates. Pediatrics, 1 09:1124-1130.
Colditz GA, Burdick E, Mosteller F (1995).
Heterogeneity in meta-analysis of data from epi-
demiologic studies: a commentary. Am J
Epidemiol, 142:371-382.
Commission Internationals de I'Eclairage
(International Commission on Illumination)
(1987). Vocabulaire International de I'Eclairage
[International Lighting Vocabulary] (CI E
Publication No. 17.4), 4th ed., Geneva, Bureau
Central de Ia Commission Electrotechnique
lnternationale, p. 3.
Cooper KD, Oberhelman L, Hamilton TA, et al.
( 1992). UV exposure reduces immunization
rates and promotes tolerance to epicutaneous
antigens in humans: relationship to dose, CD1 a
DR+ epidermal macrophage induction, and
Langerhans cell depletion. Proc Nat/ Acad Sci
USA, 89:8497-8501.
Copas JB, Shi JQ (2001). A sensitivity analysis for
publication bias in systematic reviews. Stat
Methods Med Res, 10:251-265.
Corona R, Dogliotti E, D'Errico M, et al. (2001 ). Risk
factors for basal cell carcinoma in a
Mediterranean population. Arch Dermatol,
Culley CA, Mayer JA, Eckhardt L, et al. (2001 ).
Compliance with federal and state legislation by
indoor tanning facilities in San Diego. JAm Acad
Dahle J, Kvam E (2003). Induction of delayed muta
tions and chromosomal instability in fibroblasts
after UVA, UVB, and X-radiation. Cancer Res,
De Fabo EC, Noonan FP, Fears T, Merlino G (2004).
Ultraviolet B but not ultraviolet A radiation initi
ates melanoma. Cancer Res,
Del Sino S, Vioux C, Rossio-Pasquier P, et al.
(2004). Ultraviolet B induces hyperproliferation
and modification of epidermal differentiation in
normal human skin grafted on to nude mice. Br
J Dermatol,
Dellavalle RP, Parker ER, Cersonsky N, et al.
(2003). Youth access laws: in the dark at the tan-
ning parlor? Arch Dermato/, 139:443-448.
Demko CA, Borawski EA, Dehanne SM, et al.
(2003). Use of indoor tanning facilities by white
adolescents in the United States. Arch Pediatr
Ado/esc Med, 157:854-860.
DeVries E, Boniol M, Severi G, et al. Public aware-
ness about risk factors could pose problems for
case-control studies: the example of sun bed use
and cutaneous melanoma. Eur J Cancer 2005,
in press.
De Winter S, Pavel S (2000). [Tanning beds: effect
on skin cancer risk unclear]. Ned Tijdschr
Geneeskd, 144:467-470 (in Dutch).
Diederen PV, van Weelden H, Sanders CJ, et al.
(2003). Narrowband UVB and psoralen-UVA in
the treatment of early-stage mycosis fungoides:
A retrospective study. J Am Acad Dermatol,
Diffey BL (1990). Human exposure to ultraviolet
radiation. Semin Dermatol, 9:2-10.
Diffey BL, Langley FC (1986). Evaluation of
Ultraviolet Radiation Hazards in Hospitals
(Report No. 49), London, Institute of Physical
Sciences in Medicine.
Diffey BL, Farr PM, Ferguson J, et al. (1990). Tanning
with ultraviolet A sun beds. Br Med J, 301: 773-774.
Donawho CK, Wolf P, Kripke ML (1994). Enhanced
development of murine melanoma in UV-irradiat-
ed skin : UV dose response, waveband depend-
ence, and relation to inflammation. Melanoma
Res, 4:93-100.
Drobetsky EA, Turcotte J, Chateauneuf A (1995). A
role for ultraviolet A in solar mutagenesis. Proc
Nat/ Acad Sci USA, 92:2350-2354.
Dunn-Lane J, Herity B, Moriarty MJ, Conroy R
(1993). A case control study of malignant
melanoma. lr Med J, 86:57-59 (Erratum in lr
Med J (1993), 86:135).
Ezzedine K, Guinot C, Mauger E, et al. (2005) . Use
of tanning devices: attitudes, beliefs and behav-
iour in 1,179 French adults. EuroSkin workshop,
Lyon, 19-20 May 2005 (meeting abstract).
Fitzpatrick TB (1988). The validity and practicality of
sun-reactive skin types I through VI. Arch
Dermatol, 124:869-871.
Francis SO, Burkhardt DL, Dellavalle RP (2005).
2005: A banner year for new US youth access
tanning restrictions. Arch Dermatol, 141 :524-525.
Fritschi L, Battistutta D, Strutton GM, Green AA
(1995). A non-invasive measure of photoageing.
lnt J Epidemiol, 24: 150-154.
Furlong T, Leisenring W, Storb R, et al. (2002).
Psoralen and ultraviolet A irradiation (PUVA) as
therapy for steroid-resistant cutaneous acute
graft-versus-host disease. Bioi Blood Marrow
Transplant, 8:206-212.
Galkin ON, Terenetskaya IP (1999). Vitamin D bio-
dosimeter: Basic characteristics and potential
applications. J Photochem Photobiol B:Biology,
Gallagher RP, Elwood JM, Hill GB (1986). Risk fac-
tors for cutaneous malignant melanoma: the
Western Canada Melanoma Study. Recent
Results Cancer Res, 1 02:38-55.
Gange RW, Blackett AD, Matzinger EA, et al. (1985).
Comparative protection efficiency of UVA- and
UVB-induced tans against erythema and forma-
tion of endonuclease-sensitive sites in DNA by
UVB in human skin. J Invest Oermatol,
Garbe C, Weiss J, Kruger S, et al. (1993). The
German melanoma registry and environmental
risk factors implied. Recent Results Cancer Res,
Geller AC, Colditz GA, Oliveria S, et al. (2002). Use
of sunscreen, sunburning rates, and tanning bed
use among more than 10,000 US children and
adolescents. Pediatrics, 109:1009-1 014.
Gerber B, Mathys P, Moser M, et al. (2002).
Ultraviolet emission spectra of sunbeds.
Photochem Photobiol, 76:664-668.
Gilchrest BA, Eller MS (1999). DNA photodamage
stimulates melanogenesis and other photopro-
tective responses. J Invest Oermatol Symp Proc,
Gniadecka M, Jemec GB (1998). Quantitative eval-
uation of chronological ageing and photoageing
in vivo: studies on skin echogenicity and thick-
ness. BrJ Oermatol, 139:815-821.
Goerz G (1979). [Erythropoietic protoporphyria].
Fortschr Med, 97:1625-1630 (in German).
Green AC (1991). Premature ageing of the skin in a
Queensland population. Med J Aust,
155:473-474, 477-478.
Green A, Whiteman D, Frost C, Battistutta D (1999).
Sun exposure, skin cancers and related skin
conditions. J Epidemio/, 9:S7-13.
Greenland S (1987) . Quantitative methods in the
review of epidemiologic literature. Epi Reviews,
Greinert R, McKinlay A, Breitbart EW (2001 ). The
European Society of skin cancer prevention -
EUROSKIN: towards the promotion and harmo-
nization of skin cancer prevention in Europe.
Recommendations. Eur J Cancer Prev,
Griffiths HR, Mistry P, Herbert KE, et al. (1998).
Molecular and cellular effects of ultraviolet light-
induced genotoxicity. Grit Rev Clin Lab Sci,
Grimes DA, Schulz KF (2005). Compared to what?
Finding controls for case-control studies. Lancet,
Guenel P, Laforest L, Cyr D, et al. (2001 ).
Occupational risk factors, ultraviolet radiation,
and ocular melanoma : a case-control study in
France. Cancer Causes Control, 12:451-459.
Halliday G (2005). Inflammation, gene mutation and
photoimmunosuppression in response to UVR-
induced oxidative damage contributes to photo-
carcinogenesis. Mut Res, 571:107-120.
Halliday GM, Agar NS, Barnetson AS, et al. (2005).
UV-A fingerprint mutations in human skin can-
cer. Photochem Photobiol, 81:3-8.
Hamlet N, Kennedy K (2004). Reconnaissance
study of sunbed use by primary school children
in Lanarkshire. J Public Health, 26:31-33.

Exposure to Artificial UV Radiation and Skin Cancer
Hannuksela-Svahn A, Sigurgeirsson B, Pukkala E,
et al. (1999). Trioxsalen bath PUVA did not
increase the risk of squamous cell skin carcino-
ma and cutaneous malignant melanoma in a joint
analysis of 944 Swedish and Finnish patients
with psoriasis. Br J Dermatol, 141:497-501.
Hefferren JJ, Cooley RO, Hall JB, et al. (1971 ). Use
of ultraviolet illumination in oral diagnosis. JAm
dent Assoc, 82:1353-1360.
Hemminki K, Bykov VJ, Marcussen JA (1999). Re:
Sunscreen use and duration of sun exposure: A
double-blind, randomised trial. J Nat/ Cancer
lnst, 91:2046-2047.
Hemminki K, Xu G, Le Curieux F (2001). Ultraviolet
radiation-induced photoproducts in human skin
DNA as biomarkers of damage and its repair.
IARC Scientific Publications, 154:69-79.
Henseler T, Wolff K, Honigsmann H, et at. (1981).
Oral 8-methoxypsoralen photochemotherapy of
psoriasis. The European PUVA study: A cooper-
ative study among 18 European centres. Lancet,
Herity B, O'Loughlin C, Moriarty MJ, Conroy R
(1989). Risk factors for non-melanoma skin can-
cer. J Irish Med, 82:151-152.
Hersey P, Bradley M, Hasic E, et al. (1983).
Immunological effects of solarium exposure.
Lancet, 1 :545-548.
Hersey P, MacDonald M, Henderson C, et at. (1988) .
Suppression of natural killer cell activity in
humans by radiation from solarium lamps
depleted of UVB. J Invest Dermatol,
Higgins JP, Thompson SG (2002). Quantifying het-
erogeneity in a meta-analysis. Stat Med,
Hogan DJ, ToT, Wilson ER, et al. (1991 ). A study of
acne treatments as risk factors for skin cancer of
the head and neck. Br J Dermatol, 125:343-348.
Holick MF, Matsuoka LY, Wortsman J (1989). Age,
vitamin D, and solar ultraviolet. Lancet,
Holman CD, Armstrong BK, Heenan PJ, et al.
(1986). The causes of malignant melanoma:
results from the West Australian Lions
Melanoma Research Project. Recent Results
Cancer Res, 102:18-37.
Holly EA, Kelly JW, Shpall SN, et al. (1987). Number
of melanocytic nevi as a major risk factor for
malignant melanoma. J Am Acad Dermatol,
Holly EA, Aston DA, Char DH, et at. (1990). Uveal
melanoma in relation to ultraviolet light exposure
and host factors. Cancer Res, 50:5773-5777.
Holly EA, Aston DA, Cress RD, et al. (1995).
Cutaneous melanoma in women. I. Exposure to
sunlight, ability to tan, and other risk factors
related to ultraviolet light. Am J Epidemiol,
141 :923-933.
Holly EA, Aston DA, Ahn DK, et al. (1996).
Intraocular melanoma linked to occupations and
chemical exposures. Epidemiology, 7:55-61.
Honigsmann H (2001). Phototherapy for psoriasis.
Clin exp Dermatol, 26:343-350.
Hornung RL, Magee KH, Lee WJ, et al. (2003).
Tanning facility use: are we exceeding Food and
Drug Administration limits. JAm Acad Dermatol,
IARC (1992). /ARC Monographs on the Evaluation
of Carcinogenic Risks to Humans, Vol. 55, Solar
and ultraviolet radiation, Lyon, International
Agency for Research on Cancer.
IARC (2001). !ARC Handbooks of Cancer
Prevention, Vol. 5, Sunscreens, Lyon, IARC
IARC (2004). /ARC Monographs on the Evaluation
of Carcinogenic Risks to Humans, Vol. 83,
Tobacco smoke and involuntary smoking, Lyon,
International Agency for Research on Cancer.
ICNIRP (2002). General approach to protection
against non-ionizing radiation. Health Physics,
ICNIRP (2004). Guidelines on limits of exposure to
ultraviolet radiation of wavelengths between 180
nm and 400 nm (incoherent optical radiation).
Health Physics, 87:171-186.
Indoor Tanning Association (2004). Statistics and
facts, Sep 30, 2004. Available at: http://www.thei-
ta.com/client_files/news/estimates.pdf (accessed
26 July 2005).
International Electrotechnical Commission (1987).
Safety of Household and Similar Electrical
Appliances, Part 2, Particular Requirements tor
Ultraviolet and Infrared Radiation Skin
Treatment Appliances tor Household Use, IEC
60335-2-27, Geneva, International Electro-
technical Commission.
International Electrotechnical Commission (1989).
Amendment, Geneva, International Electro-
technical Commission.
International Electrotechnical Commission (1995).
Household and similar electrical appliances -
Safety - Part 2-27: Particular requirements for
appliances for skin exposure to ultraviolet and
infrared radiation, IEC 60335-2-27, 3rd edition,
Geneva, International Electrotechnical Com-
International Electrotechnical Commission (2003).
Household and similar electrical appliances -
Safety - Part 2-27: Particular requirements for
appliances for skin exposure to ultraviolet and
infrared radiation, IEC 60335-2-27, 4th edition,
Geneva, International Electrotechnical
IPCS (1994). Ultraviolet radiation, Environmental
Health Criteria 160, Geneva, International
Programme on Chemical Safety, World Health
Organisation, United Nations Environment
Javitt JC, Taylor HR (1994). Cataract and latitude.
Doc Ophtha!mol, 88:307-325.
Jenkins G (2002). Molecular mechanisms of skin
ageing. Mech Ageing Dev, 123:801-810.
Kadunce DP, Piepkorn MW, Zone JJ (1990).
Persistent melanocytic lesions associated with
cosmetic tanning bed use: 'Sunbed lentigines'. J
Am Acad Dermato!, 23:1029-1031.
Karagas MR, Stannard VA, Mott LA, et al. (2002).
Use of tanning devices and risk of basal cell and
squamous cell skin cancer. J Nat! Cancer lnst,
Kaskel P, Sander S, Kron M, et al. (2001 ). Outdoor
activities in childhood: a protective factor for
cutaneous melanoma? Results of a case-control
study in 271 matched pairs. Br J Dermatol,
Kelly DA, Walker SL, McGregor JM, Young AR
(1998). A single exposure of solar simulated
radiation suppresses contact hypersensitivity
responses both locally and systemically in
humans: quantitative studies with high-frequency
ultrasound. J Photochem Photobiol 8,
Klepp 0, Magnus K (1979). Some environmental
and bodily characteristics of melanoma patients.
A case-control study. lnt J Cancer, 23:482-486.
Knight JM, Kirincich AN, Farmer ER, Hood AF
(2002). Awareness of the risks of tanning lamps
does not influence behavior among college stu-
dents. Arch Dermatol, 138:1311-1315.
Kraemer KH, Lee MM, Andrews AD, Lambert WC
(1994). The role of sunlight and DNA repair in
melanoma and non-melanoma skin cancer: The
xeroderma pigmentosa paradigm. Arch
Dermatol, 130:1018-1021.
Kreimer-Erlacher H, Seidl H, Back B, et al. (2003).
High frequency of ultraviolet mutations at the
INK4a-ARF locus in squamous cell carcinomas
from psoralen-plus-ultraviolet-A-treated psoria-
sis patients. J invest Dermatol, 120:676-682.
Kripke ML (1974) . Antigenicity of murine skin tumors
induced by UV light. J Nat! Cancer lnst,
Kwon HT, Mayer JA, Walker KK, et al. (2002).
Promotion of frequent tanning sessions by
indoor tanning facilities: two studies. JAm Acad
Dermato!, 46:700-705.
de Laat A, van der Leun JC, de Gruijl FR (1997).
Carcinogenesis induced by UVA (365-nm) radia-
tion: the dose-time dependence of tumor forma-
tion in hairless mice. Carcinogenesis,
Lazovich D, Forster J (2005). Indoor tanning by ado-
lescents: prevalence, practices and policies. Eur
J Cancer, 41 :20-37.
Lazovich D, Forster JL, Sorensen G, et al. (2004).
Characteristics associated with adolescents
intention to initiate or continue indoor tanning
use. Arch Pediatr Ado/esc Med, 158:918-924.
Lazovich D, Sweeney C, Forster J (2005).
Prevalence of indoor tanning use in Minnesota,
2002. Arch Dermatol, 141:523-524.
Lerman S (1984). Biophysical aspects of corneal
and lenticular transparency. Curr Eye Res,
Ley RD (1997). Ultraviolet radiation A-induced pre-
cursors of cutaneous melanoma in Monodelphis
domestica. Cancer Res, 57:3682-3684.
Ley RD (2001 ). Dose response for ultraviolet radia-
tion A-induced focal melanocytic hyperplasia
and nonmelanoma skin tumors in Monodelphis
domestica. Photochem Photobiol, 73:20-23.
Leyden JJ (1990). Clinical features of ageing skin.
Br J Dermato!, 122 (Suppl. 35):1-3.
Lim HW, Cyr WH, DeFabo E, et al. , for the American
Academy of Dermatology Association, American
Society for Photobiology, US Food and Drug
Administration (2004). Scientific and regulatory
issues related to indoor tanning. J Am Acad
Dermatol, 51:781-784.
Lindelof B, Sigurgeirsson B, Tegner E, et al. (1999) .
PUVA and cancer risk: The Swedish follow-up
study. Br J Derma to/, 141 : 1 08-112.
Lucci A, Citro HW, Wilson L (2001). Assessment of
knowledge of melanoma risk factors, prevention,
and detection principles in Texas teenagers.
J Surg Res, 97:179-183.
Lutz JM, Cree I, Sabroe S, et al. (2005).
Occupational risks for uveal melanoma results
from a case-control study in nine European
countries. Cancer Causes Control, 16:437-447.
Macaskill P, Walter SO, lrwig L (2001 ). A comparison
of methods to detect publication bias in meta-
analysis. Stat Med, 20:641-654.


Exposure to Artificial UV Radiation and Skin Cancer
MacKie RM, Freudenberger T, Aitchison TC (1989).
Personal risk-factor chart for cutaneous
melanoma. Lancet 2:487-90.
Marcil I, Stern RS (2001 ). Squamous-cell cancer of
the skin in patients given PUVA and cyclosporin:
Nested cohort crossover study. Lancet,
358:1 042-1 045.
Marguery MC, Lamant L, Bayle P, et al. (2005).
Lupus-like phototriggering in a young woman
with benign summer li ght eruption.
Photodermatol photoimmunol Photomed,
Marrot L, Belaidi JP, Lejeune F, et al. (2004).
Photostability of sunscreen products influences
the efficiency of protection with regard to UV-
induced genotoxic or photoageing-related end-
points. Br J Dermatol, 151:1234-1244 .
McGinley J, Martin CJ, MacKie RM (1998). Sunbeds
in current use in Scotland: a survey of their out-
put and patterns of use. Br J Dermatol,
McGinnis KS, Shapiro M, Vittorio CC, et al. (2003).
Psoralen plus long-wave UV-A (PUVA) and
bexarotene therapy: An effective and synergistic
combined adjunct to therapy for patients with
advanced cutaneous T-cell lymphoma. Arch
Dermatol, 139:771- 775.
McKinlay AF, Diffey BL (1987). A reference action
spectrum for ultraviolet induced erythema in
human skin. CIE (Commission lnternationale de
I'Eclairage) J, 6:17-22.
Mermelstein RJ, Riesenberg LA (1992). Changing
knowledge and attitudes about skin cancer risk
factors in adolescents. Health Psycho/,
11 :371-376.
Millard TP, Bataille V, Snieder H, et al. (2000). The
heritability of polymorphic light eruption. J invest
Dermatol, 115:67-70.
Miller SA, Hamilton SL, Wester UG, Cyr WH (1998).
An analysis of UVA emissions from sunlamps
and the potential importance for melanoma.
Photochem Photobiol, 68:63-70.
Mintz B, Silvers WK (1993). Transgenic mouse
model of malignant skin melanoma. Proc Nat/
Acad Sci USA, 90:8817-8821.
Moore DE (2002). Drug-induced cutaneous photo-
sensitivity: Incidence, mechanism, prevention
and management. Drug Sat, 25:345-372.
Moy CS (2001). Evidence for the role of sunlight
exposure in the etiology of choroidal melanoma.
Arch Ophthalmol, 119:430-431.
Mustajoki P (1980). Variegate porphyria. Twelve
years' experience in Finland. Q J Med,
Naldi L, Gallus S, lmberti GL, et al. , for the Italian
Group for Epidemiological Research in
Dermatology (2000). Sunlamps and sunbeds
and the risk of cutaneous melanoma. Eur J
Cancer Prev, 9:133-134.
National Radiation Protection Board (NRPB)
(2002). Statement by the advisory group on non-
ionizing radiation, use of sunbed and cosmetic
tanning. In : Health Effects from Ultraviolet
Radiation, 13:279-282.
National Toxicology Program (2002). Report on
Carcinogens, 10th Edition, Substances Profiles,
National Toxicology Program, Research Triangle
Park, NC.
Neale RE, Purdie JL, Hirst LW, Green AC (2003).
Sun exposure as a risk factor for nuclear
cataract. Epidemiology, 14:707-712.
Nghiem OX, Kazimi N, Clydesdale G, et al. (2001).
Ultraviolet A radiation suppresses an estab-
lished immune response: implications for sun-
screen design. J lnvestDermatol, 117:1193-99.
Nghiem OX, Kazimi N, Mitchell DL, et al. (2002).
Mechanisms underlying the suppression of
established immune responses by ultraviolet
radiation. J Invest Dermatol, 119:600-608.
Nishigaki R, Mitani H, Tsuchida N, Shima A (1999).
Effect of cyclobutane pyrimidine dimers on
apoptosis induced by different wavelengths of
UV. Photochem Photobio/, 70:228-35.
Noonan FP, Recio JA, Takayama H, et al. (2001 ).
Neonatal sunburn and melanoma in mice.
Nature, 413:271-272.
Oliphant JA, Forster JL, McBride CM (1994). The
use of commercial tanning facilities by suburban
Minnesota adolescents. Am J Public Health,
O'Loughlin C, Moriarty MJ, Herity B, et al. (1985) . A
Re-appraisal of risk factors for skin carcinoma in
Ireland. lr J med Science, 154:61-65.
Osborne JE, Hutchinson PE (2002). Vitamin D and
systemic cancer: is this relevant to malignant
melanoma? BrJ Dermatol, 147:197-213.
Osterlind A, Tucker MA, Stone BJ, et al. (1988) . The
Danish case-control study of cutaneous malig-
nant melanoma. II. Importance of UV-Iight expo-
sure. tnt J Cancer, 42:319-24.
O'Toole E, Barnes L (1995). Sun-bed-induced
vesiculobullous polymorphic light eruption. Br J
Dermatol, 132:171 .
Paul CL, Girgis A, Tzelepis F, Walsh R (2004).
Solaria use by minors in Australia: is there a
cause for concern? Aust N Z J Public Health,
Paul CL, Stacey F, Girgis A, et al. (2005). Solaria
compliance in an unregulated environment: The
Australian experience. Eur J Cancer, 41:
Pavel S, Smit NP, Van Der Meulen H, et al. (2003).
Homozygous germline mutation of
CDKN2A/p16 and glucose-6-phosphate dehy-
drogenase deficiency in a multiple melanoma
case. Melanoma Res, 13:171-178.
Pedeux A, Al-lrani N, Marteau C, et al. (1998) .
Thymidine dinucleotide induce S phase cell
cycle arrest in addition to increased melanogen-
esis in human melanocytes. J Invest Dermatol,
Persson AE, Edstrom DW, Backvall H, et al. (2002).
The mutagenic effect of ultraviolet-A 1 on human
skin demonstrated by sequencing the p53 gene
in single keratinocytes. Photodermatol
Photoimmunol Photomed, 18:287-293.
Pfeifer GP, You YH, Besaratinia A {2005). Mutations
induced by ultraviolet light. Mutat Res,
Pittelkow MR. Perry HO, Muller SA, et al. (1981).
Skin cancer in patients with psoriasis treated
with coal tar. A 25-year follow-up study. Arch
Dermatol, 117:465-468.
Pjrek E, Winkler D, Stastny J, et al. (2004). Bright
light therapy in seasonal affective disorder-
Does it suffice? Eur Neuropsychopharmacol,
14:347-351 .
Placzek M, Kerkmann U, Bell S, et al. (2004).
Tobacco smoke is phototoxic. Br J Dermatol,
150:991- 993.
Poh-Fitzpatrick MB, Ellis DL (1989). Porphyrialike
bullous dermatosis after chronic intense tanning
bed and/or sunlight exposure. Arch Dermatol,
Reme CE, Rol P, Grothmann K, et al. {1996). Bright
light therapy in focus: lamp emission spectra and
ocular safety. Techno/ Health Care, 4:403-413.
Rhainds M, De Guire L, Claveau J (1999). A popu-
lation-based survey on the use of artificial tan-
ning devices in the province of Quebec, Canada.
JAm Acad Dermatol, 40:572- 576.
Rigel DS, Friedman RJ, Levenstein MJ, et al. {1983).
Relationship of fluorescent lights to malignant
melanoma: another view. J Dermatol Surg
Oncol, 9:836-838.
Rivers JK, Norris PG, Murphy GM, et al. {1989).
UVA sunbeds: tanning, photoprotection, acute
adverse effects and immunological changes. Br
J Dermatol, 120:767- 777.
Robert C. Muel B, Benoit A, et al. (1996). Cell sur-
vival and shuttle vector mutagenesis induced by
Exposure to Artificial UV Light and Skin Cancer
UVA and UVB radiation in a human cell line.
J Invest Dermatol, 106:721-728.
Robinson ES, Hubbard GB, Colon G, et al. (1998).
Low-dose ultraviolet exposure early in develop-
ment can lead to widespread melanoma in the
opossum model. tnt J Exp Patho/, 79:235-244.
Robinson ES, Hill RH Jr, Kripke ML, et al. (2000a).
The Monodelphis melanoma model: initial report
on large ultraviolet A exposures of suckling
young. Photochem Photobiol, 71:743-746.
Robinson ES, Hubbard GB, Dooley TP (2000b).
Metastatic melanoma in an adult opossum
(Monodelphis domestica) after short-term inter-
mittent UVB exposure. Arch Dermatol Res,
292:469-471 .
Robinson JK, Rademaker AW, Sylvester JA, Cook B
(1997). Summer sun exposure: knowledge, atti-
tudes and behaviours of midwest adolescents.
Prev Med, 26:364-372.
Roenigk HH Jr (1979). Photochemotherapy for
mycosis fungoides: Long-term followup study.
Cancer Treat Rep, 63:669-673.
Roth DE, Hodge SJ, Callen JP (1989). Possible
ultraviolet A-induced lentigines: A side effect of
chronic tanning salon usage. J Am Acad
Dermatol, 20:950-954.
Ruegemer J, Schuetz B, Hermann K, et al. (2002).
UV-induced skin changes due to regular use of
commercial sunbeds. Photodermatol
Photoimmunol Photomed, 18:223-227.
Sanders CJ, Van Weelden H, Kazzaz GA, et al.
(2003). Photosensitivity in patients with lupus
erythematosus: A clinical and photobiological
study of 1 00 patients using a prolonged pho-
totest protocol. Br J Derma to/, 149:131-137.
Sauter EA. Klein-Szanto AJ, Atillasoy E, et al.
{1998). Ultraviolet B-induced squamous epithe-
lial and melanocytic cell changes in a xenograft
model of cancer development in human skin.
Mol Carcinog, 23:168-174.
Sayre AM, Dowdy JC, Poh-Fitzpatrick M (2004).
Dermatological risk of indoor ultraviolet expo-
sure from contemporary lighting sources.
Photochem Photobiol, 80:47- 51.
Seddon JM, Gragoudas ES, Glynn RJ, et al. (1990).
Host factors, UV radiation, and risk of uveal
melanoma. A case-control study. Arch
Ophtalmol, 1 08:127 4-1280.
Sheehan JM, Patten CS, Young AR (1998). Tanning
in human skin types II and Ill offers modest pho-
toprotection against erythema. Photochem
Photobiol, 68:588-592.
Setlow AB, Regan JD, German J, et al. (1969).
Evidence that xeroderma pigmentosum cells do

Exposure to Artificial UV Radiation and Skin Cancer
not periorm the first step in the repair of ultravi
olet damage to their DNA. Proc Nat/ Acad Sci
USA, 64:1035-1041 .
Setlow RB, Grist E, Thompson K, et al. (1993).
Wavelengths effective in induction of malignant
melanoma. Proc Nat/ Acad Sci USA, 90:
Siemiatycki J (1991 ). Risk factors for cancer in the
work place. Vol. 1, Boca Raton, CRC Press.
Singh RK, Gutman M, Reich R, et al. (1995).
Ultraviolet 8 irradiation promotes tumorigenic
and metastatic properties in primary cutaneous
melanoma via induction of interleukin-8. Cancer
Res, 55:369-374.
Skjerve A, Holsten F, Aarsland D, et al. (2004).
Improvement in behavioral symptoms and
advance of activity acrophase after short-term
bright light treatment in severe dementia.
Psychiatry clin Neurosci, 58:343-347.
Sellitto RB, Kraemer KH, DiGiovanna JJ (1997).
Normal vitamin D levels can be maintained
despite rigorous photoprotection: six years'
experience with xeroderma pigmentosum. JAm
Acad Dermatol, 37:942-947.
Spencer JM, Amonette RA (1995). Indoor tanning:
risks, benefits, and future trends. J Am Acad
Dermatol, 33:288-298.
Stern AS (2001 ). The risk of melanoma in associa-
tion with long-term exposure to PUVA. J Am
Acad Dermatol, 44:755-761.
Stern RS, Laird N (1994). The carcinogenic risk of
treatments for severe psoriasis.
Photochemotherapy Follow-up Study. Cancer,
Stern AS, Lunder EJ (1998) . Risk of squamous cell
carcinoma and methoxsalen (psoralen) and UV-
A radiation (PUVA). A meta-analysis. Arch
Dermatol, 134:1582-1585.
Stoebner-Delbarre A, Thezenas S, Kuntz C, et al.
(2001 ). Connaissances, attitudes et comporte-
ments des adultes vis-a-vis de !'exposition
solaire en France. In: Aubin F. Humbert P (eds)
Rayonnement ultraviolet et peau. Paris, John
Libbey Eurotext, pp 135-140 (in French).
Stryker JE, Lazovich D, Forster JL, et al. (2004).
Maternal/female caregiver influences on adoles-
cent indoor tanning. J Ado/esc Health,
Swerdlow AJ, English JS, MacKie RM, et al. (1988).
Fluorescent lights, ultraviolet lamps, and risk of
cutaneous melanoma. Br med J, 297:647-650.
Takahashi Y, Moriwaki S, Sugiyama Y, Endo Y, et al.
(2005). Decreased gene expression responsible
for post-ultraviolet DNA repair synthesis in
aging: A possible mechanism of age-related
reduction in DNA repair capacity. J Invest
Dermatol, 124:435-442.
Taylor HR (1994). Ocular effects of UV-8 exposure.
Doc Ophthalmol, 88:285-93.
Tie C, Golomb C, Taylor JR, Streilein JW (1995).
Suppressive and enhancing effects of ultraviolet
B radiation on expression of contact hypersensi-
tivity in man. J Invest Dermatol, 104:18-22.
Tomany SC, Cruickshanks KJ, Klein R, et al. (2004).
Sunlight and the 1 0-year incidence of age-relat-
ed maculopathy: the Beaver Dam Eye Study.
Arch Ophtha/mol, 122:750-757. Erratum in:
Arch Ophtha/mol (2005), 123:362.
Tucker MA, Shields JA, Hartge P, et al. (1985).
Sunlight exposure as risk factor for intraocular
malignant melanoma. N Eng/ J Med,
Ullrich SE (2005). Mechanisms underlying UV-
induced immune suppression. Mut Res,
US FDA (1985) Code Fed. Regul., 21 CFR Part
1040 (Docket No. 82N-0188). Sunlamp
Products; Periormance Standard - Final Rule.
Federal Register, Vol. 50, No. 173, US
Department of Health and Human Services.
Food and Drug Administration.
US FDA (2004) Code Fed. Regul., 21 CFR 1040.20.
Sunlamps and sunlamp products; Periormance
standards. Revised as of April 1, 2004, US
Department of Health and Human Services,
Food and Drug Administration.
Vajdic CM, Kricker A, Giblin M, et al. (2004). Artificial
ultraviolet radiation and ocular melanoma in
Australia. /nt J Cancer, 112:896-900.
van Houwelingen HC, Arends LA, Stijnen T (2002).
Advanced methods in meta-analysis: multivari-
ate approach and meta-regression. Stat Med,
van Schanke A, Jongsma MJ, Bisschop R, et al.
(2005). Single UVB overexposure stimulates
melanocyte proliferation in murine skin, in con-
trast to fractionated or UVA-1 exposure. J Invest
Dermato/, 124:241-247.
Veiemd MB, Weiderpass E, Thorn M, et al. (2003).
A prospective study of pigmentation, sun expo-
sure, and risk of cutaneous malignant
melanoma in women. J Nat/ Cancer lnst,
Veierod MB, Weiderpasss E. Lund E, et al. (2004).
Re: prospective study of pigmentation, sun
exposure, and risk of cutaneous malignant
melanoma in women. J Nat/ Cancer lnt,
Walter SD, Marrett LD, From L, et al. (1990) . The
association of cutaneous malignant melanoma
with the use of sunbeds and sunlamps. Am J
Epidemiol, 131:232-243.
Walter SD, Marrett LD, Shannon HS, et al. (1992).
The association of cutaneous malignant
melanoma and fluorescent light exposure. Am J
Epidemio/, 135:749--762.
Walter SD, King WD, Marrett LD (1999). Association
of cutaneous malignant melanoma with intermit-
tent exposure to ultraviolet radiation: Results of
a case-control study in Ontario, Canada. lnt J
Epidemiol, 28:418-427.
Walther U, Kron M, SanderS, et al. (2004). Risk and
protective factors for sporadic basal cell carcino-
ma: results of a two-centre case-control study in
southern Germany. Clinical actinic elastosis may
be a protective factor. Br J Dermatol,
151 :170-178.
Wang Y, Rosenstein B, Goldwyn S, et al. (1998) .
Differential regulation of P53 and Bcl-2 expres-
sion by ultraviolet A and B. J Invest Dermatol,
111 :380-384.
Weischer M, Blum A, Eberhard F, et al. (2004). No
evidence for increased skin cancer risk in psori-
asis patients treated with broadband or narrow-
band UVB phototherapy: A first retrospective
study. Acta Derm Venereal, 84:370-374.
Wester U, Boldemann C, Jansson B, Ullen H (1999).
Population UV-dose and skin area -Do sunbeds
rival the sun? Health Physics, 77:436-440.
Westerdahl J, Olsson H, Masback A, et al. (1994).
Use of sunbeds or sunlamps and malignant
melanoma in southern Sweden. Am J Epidemiol,
Westerdahl J, lngvar C, Masback A, et al. (2000).
Risk of cutaneous malignant melanoma in rela-
tion to use of sunbeds: Further evidence for UV-
A carcinogenicity. Br J Cancer, 82:1593-1599.
Whiteman DC, Whiteman CA, Green A (2001).
Childhood sun exposure as a risk factor for
melanoma: A systematic review of epidemiologic
studies. Cancer Causes Control, 12:69-82.
Whitmore SE, Morison WL, Patten CS, et al. (2001).
Tanning salon exposure and molecular alter-
ations. JAm Acad Dermatol, 44:775-780.
WHO (2003). Artificial tanning sunbeds - risks and
guidance, Geneva, World Health Organisation.
Wichstrom L (1994). Predictors of Norwegian ado-
lescents' sunbathing and use of sunscreen.
Health Psycho/, 13:412-420.
Wolf P, Donawho CK, Kripke ML (1994). Effect of
sunscreens on UV radiation-induced enhance-
ment of melanoma growth in mice. J Nat/ Cancer
lnst, 86:99--105.
Woollens A, Clingen PH, Price ML, et al. (1997).
Induction of mutagenic damage in human fibrob-
lasts after exposure to artificial tanning lamps. Br
J Dermatol, 137:687-692.
Wright AL, Hart GC, Kernahan E, et al. (1996).
Survey of the variation in ultraviolet outputs from
ultraviolet A sunbeds in Bradford. Photodermatol
Photoimmunol Photomed, 12:12-16.
Wright A, Hart G, Kernahan L (1997). Dangers of
sun beds are greater in the commercial sector. Br
Med J, 314:1280-1281.
Young AR, Patten CS, Nikaido 0 , et al. (1998).
Human melanocytes and keratinocytes exposed
to UVB or UVA in vivo show comparable levels
of thymine dimers. J Invest Dermatol,
Young AR (2004). Tanning devices-Fast track to
skin cancer? Pigment Cell Res, 17:2-9.
Zanetti R, Rosso F, Faggiano F, et al. (1988). A
case-control study on cutaneous malignant
melanoma in the province of Torino, Italy. Rev
Epidemiol Sante Publique, 36:309-317 [in
Zhang H, Rosdahl I (2003). Ultraviolet A and B dif-
ferently induce intracellular protein expression in
human skin melanocytes - A speculation of
separate pathways in initiation of melanoma.
Carcinogenesis, 24:1929--1934.
Zigman S (1983). Effects of near ultraviolet radiation
on the lens and retina. Doc Ophtatmol,
Appendix: European and international positions regarding artificial
sources of UV radiation
Establishment of a standard for appliances
designed specifically for tanning purposes
Appliances designed specifically for tanning pur-
poses are defined in the international standard
prepared by the International Electrotechnical
Commission (IEC 60 335-2-27). This standard
was first established in 1985 (IEC, 1987) and
slightly modified in 1989 (2nd edition; IEC, 1989),
in 1995 (3rd edition; IEC, 1995) and in 2002 (4th
edition; IEC, 2003). The USA follow the recom-
mendations of the Food and Drug Administration
(US FDA, 1985), which were recently updated
(Lim et al., 2004; US FDA, 2004).
According to the 3rd edition of the IEC stan-
dard, appliances emitting UV radiation must
belong to one of the types described below:
Type UV-1 appliance: appliance with a UV
source such that the biological effect is trig-
gered by radiation wavelengths >320nm, and
which is characterized by relatively high irra-
diance efficiency in the range 32D-400nm.
Type UV-2 appliance: appliance with a UV
source such that the biological effect is trig-
gered by radiation wavelengths above or
below 320nm, and which is characterized by
relatively high irradiance efficiency in the
range 320-400nm.
Type UV-3 appliance: appliance with a UV
source such that the biological effect is
triggered by radiation wavelengths above or
below 320nm and which is characterized by
restricted irradiance over the entire range of
UV radiation.
Type UV-4 appliance: appliance with a UV
source such that the biological effect is
triggered primarily by radiation wavelengths
Table I shows the physical characteristics of the
According to the standard, "the appliances
must not be toxic or represent similar hazard. The
appliances emitting UV radiation must not emit
radiation in dangerous amounts and their irradi-
ance efficiency must be within the values
specified in Table [1]". In addition, the standard
states that the verification of conformity must be
performed by 1) determining the ageing of the
appliance before measurement and 2) respecting
a distance of 0.3m.
The guidelines recommend that the exposure
time for the first session on untanned skin should
correspond to an effective dose not exceeding
1 00 J/m
; this is approximately equivalent to 1
MED for subjects with sun-reactive skin type I.
The annual exposure should not exceed an effec-
tive dose of 25 kJ/m
(IEC, 1989).
Although these guidelines form the basis of
several national standards on the use of tanning
appliances, it should be noted that variations
exist; for example, in the Netherlands, Norway
Table I. Type of UV appliances according to their irradiance efficiency
Type of UV
lrradlance efficiency In W/m
250 nm < A. < 320 nm 320 nm < A. < 400 nm
< 0.0005
0.0005 to 0.15
< 0.15
2: 0.15
A., radiation wavelength
2: 0.15
2: 0.15
< 0.15
< 0.15
Exposure to Artificial UV Radiation and Skin Cancer
and Sweden, certain UV appliances are not per-
mitted. Regulations concerning the use of
tanning appliances are in force in only a few
countries, but many others have published advice
on their use, including information on adverse
effects, as well as guidelines on manufacturing
In 2004, amendment 1 (2004-2007) to the
4th edition of the standard (2002-2009) added
type-5 appliances to the standard: appliance with
a UV source such that the biological effect is trig-
gered by radiation wavelengths above or below
320nm and which is characterized by a relatively
high irradiance efficiency over the entire range of
UV radiation.
The second amendment is currently being
voted internationally. This amendment would sup-
press the current classes and distinguish only
two classes: appliances for sale to the general
public (formerly type-3 appliances) and appli-
ances for professionals making UV available to
the public.
National and international scientific policies
Several international authorities have adopted a
defined position regarding specifically the use of
UV-emitting appliances for tanning purposes.
These positions are almost invariably accompa-
nied by recommendations targeted at the safety
of customers.
The International Commission on Non-Ionizing
Radiation Protection (ICNIRP) is an independent
group of experts convened to evaluate available
data on the effects of non-ionizing radiation to
humans. ICNIRP proposes exposure limits to UV
radiation for the general population and in occu-
pational settings for the eye and for the skin, for
an 8-hour exposure period (ICNIRP, 2004).
In its statement relating to UV-emitting
appliances for tanning purposes, ICNIRP, after
considering the effects of UV radiation on the
skin and the different types of existing
appliances, concluded that use of UV-emitting
appliances for tanning and other non-medical
purposes should be discouraged. High-risk
individuals must be particularly warned against
the use of tanning appliances. These include:
individuals with skin phototypes I and II;
children and adolescents under the age of 18
individuals with a large number of naevi;
individuals with a tendency to have freckles;
individuals who had frequent sunburns during
individuals with pre-malignant and malignant
skin tumours;
individuals with actinic skin ageing;
individuals who have applied cosmetics on
their skin; and
individuals who are taking medication must
seek advice from their doctor to determine
whether their medication renders them more
sensitive to UV.
If, in spite of the above-mentioned recommenda-
tions, individuals decide to use tanning
appliances, a number of measures must be
implemented to minimize the risk. These
measures apply specifically to skin phototype I
and II, children, individuals with increased sensi-
tivity due to the use of medication or cosmetics,
or individuals with a skiR cancer-related pathology.
World Health Organization (WHO)
In 2003, WHO published a document entitled:
"Artificial tanning sunbeds: risks and guidances"
in the framework of the INTERSUN program
(WHO, 2003). This document is based on recom-
mendations cited by other organisations such as
ICNIRP, EUROSKIN and the National
Radiological Protection Board (NRPB), among
others. Specifically, WHO recommends that cos-
tumers carefully read the recommendations and
sign the consent form before each tanning ses-
sion, so as to make them fully aware of their
EUROSKIN dedicated an international meeting
to the problems arising from the use of tanning
appliances. The outcome of the conference was
published in the European Journal of Cancer
Prevention (Greinert eta/. , 2001 ). The document
presents general statements about individuals
who should avoid such practices, and makes
specific recommendations on the information to
be given to customers and on how to use UV-
emitting appliances.
National Radiological Protection Board
In 2002, a group of public health scientists in the
United Kingdom published a report through the
National Radiological Protection Board on the
health effects of UV radiation (NRPB, 2002). The
document advises against the use of UV-emitting
appliances for tanning purposes and recom-
mends that the potential risks for detrimental
health effects be clearly outlined to the users and
to the general population at large.
National Toxicology Program (NTP)
In 2002, the National Toxicology Program in the
USA published the 1Oth Report on Carcinogens
(NTP, 2002) in which UVA, UVB and UVC
radiations were included in the list of "known car-
cinogens to humans". One chapter of the docu-
ment is dedicated to solar radiation and exposure
to UV-emitting appliances. In fact, the American
Medical Association had asked in 1994 for a
complete ban of exposure to UV for non-medical
purposes. In the USA, 27 states have a regulation
regarding availability of indoor tanning facilities to
the general population.
Regulations and recommendations of health
authorities from those countries where they are
available are listed in Table II. The following find-
ings can be highlighted:
For consumer safety reasons, Scandinavian
countries authorise only type-3 appliances with
an emission limit of 0.15 w.m
for both UVA and
UVB, i.e. a total UV intensity of 0.3 W.m
. Some
countries specify that these appl iances may also
be sold to individuals.
The organisations responsible for radiopro-
tection and the health authorities of five
Scandinavian countries (Denmark, Finland,
Iceland, Norway and Sweden) released a joint
public health advice recommending that more
stringent safety procedures be adopted regarding
the use of UV tanning appliances. This advice is
in line with the position of international (WHO,
ICNIRP), European (EUROSKIN) and national
The countries that produce the majority of
UV tubes and UV tanning appliances (Germany,
Italy and the Netherlands) have no legislation
limiting the manufacturers.
The USA has its own restrictive regulation
imposed by the FDA.
The European Commission has raised a
concem about the lack of upper limits for the
dose rate of type-1 and type-2 appliances (Type-
4 appliances are not concerned).
The legislation in Spain requires that each
individual sign a book, a registry and a consent
form, and be given a tanning booklet specifying
the details of the sessions in accordance with the
characteristics of the UV appliance. This initiative
corresponds to the recommendations of WHO
(WHO, 2003).
The maximal cumulative annual dose of UV
radiation currently established at 15 kJ.m-2 is cal-
culated from the standard IEC 60 335-2-27. This
cumulative dose by far exceeds the dose
received from ambient natural UV. Finland
wishes to set a maximal annual cumulative dose
of 5 kJ.m2. In fact, the maximal annual cumula-
tive dose should be adjusted to the phototype, i.e.
9 kJ.m-2 (NMSC) for phototype II, 15 kJ.m-2
(NMSC) for phototype Ill and 21 kJ.m-2 (NMSC)
for phototype IV.
It is noteworthy that few countries regulate
indoor tanning, and when they do, regulations are
mostly silent on use of these appliances by ado-
lescents. According to a recent review by
Dellavalle et al. (2003), only France has adopted
the age of 18 years as the legal minimum age for
indoor tanning. In the USA, only 6 states have in
place minimum age limits for tanning patrons:
California, Illinois, New Hampshire and North
Carolina restrict access to individuals younger
than 14 years old, while Texas and Wisconsin
restrict access to adolescents younger than 13 and
16 years old, respectively (Francis eta/., 2005).
In France, technical controls are performed
periodically in all registered tanning
Exposure to Artificial UV Radiation and Skin Cancer
Table II. Regulations and recommendations from health authorities in those countries where
information is available
Presnorme 6norm S 1132 : "Safety rules during the use of solaria emitting UV radiations'
(1 January 2002)
"Royal decision on requirements for exploiting solaria" (2000)
Territory, Province and State Committee on radioprotection : "Guidelines to owners, operators and
users of tanning salons". Enforcement of regulations enacted to implement the" law relating to
appliances emitting radiations - RED Act : Regulations of sun lamps , (2002-2003)
Bundesamt fUr Strahlenschutz, Munich : ' Certification of solaria" (proposal)
Currently, UV exposure in solaria must respect the German standard DIN 5050-2 (June 1998)
Decree on the limitation of public exposure to non-ionising radiation (294/2002) chapter 4,
"Ultraviolet radiation"; SS 9.1 "Safety of solaria" (1989); SONT 9.1 "Safety and control conditions
of solaria" (project) (2003)
Decree no 97-617 (30 May 1997) and regulations of 10 September 1997, 09 December 1997 and
16 September 2002 enacted to implement the law
lstituto Superiore di Sanita "Devices must conform to the technical norm IEC 60335-2-27"
(1995-2003); detailed information available at http://www.iss.iVsitp/sole/abbrarVIeggi.html
No formal regulation.
Several reports from the Dutch Health Council (1987, 1994)
Regulations of 8 April 1983 for solaria/alpine sun. Delegation of authority regulation fixing by royal
decree the usage of UV radiations for cosmetic purposes (01 July 1983)
19574 Royal decree 100212002: "Regulation of sale and use of tanning appliances emitting
UV radiations" (27 September 2002)
Regulatory code concerning sunbeds (SSI FS 1998:2) "The regulation of tanni ng appliances
used by the public fulfils the criteria of the standard EN 60335-2-27." (28 September 1998 and
03 November1 998)
FDA Sunlamp Performance Standard 21CFR1040.20 combined wi th a guide (1986) on timers
and frequency of exposure (updated 01 April 2004)
establishments since 1999. The proportion of
establishments 90mpliant with the technical
requirements increased from 51% in 1999 to
72% in 2003. Periodic visits in 2002 and 2003 to
additional establishments showed a compliance
of 85% and 81%, respectively. Non-compliance
was mainly for minor infractions (AFSSE, 2005).
tanning facilities in 2002 and 2003. The following
comments were noted: the facility operators are
well informed of the regulations in place; the
compulsory declaration was generally satisfactory;
in most places, at least one person had the
required qualifications, but the need for
continuing education was not perceived clearly;
information to the customers were usually
available to the users, but information on risks
was often lacking (AFSSE, 2005).
In addition, the Direction Generale de Ia
Concurrence, de Ia Consommation et de Ia
Repression des Fraudes (DGCCRF) proceeded
to a series of controls and enquiries about indoor
Solar and Ultraviolet Radiation
This publication represents the views and expert opinions
of an IARC Working Group on the
Evaluation of Carcinogenic Risks to Humans,
which met in Lyon,
11-18 February 1992
In 1969, the International Agency for Research on Cancer {IARC) initiated a programme
on the evaluation of the carcinogenic risk of chemicals to humans involving the production of
critically evaluated monographs on individual chemicals. In 1980 and 1986, the programme
was expanded to include the evaluation of the carcinogenic risk associated with exposures to
complex mixtures and other agents.
The objective of the programme is to elaborate and publish in the form of monographs
critical reviews of data on carcinogenicity for agents to which humans are known to be
exposed, and on specific exposure situations, to evaluate these data in terms of human risk
with the help of international working groups of experts in chemical carcinogenesis and
related fields; and to indicate where additional research efforts are needed.
This project is supported by PHS Grant No. 2-UO 1 CA33193- 10 awarded by the US
National Cancer Institute, Department of Health and Human Services. Additional support
has been provided since 1986 by the Commission of the European Communities.
International Agency for Research on Cancer 1992
ISBN 92 832 1255 X
ISSN 0250-9555
All rights reserved. Application for rights of reproduction or translation, in part or in t ~
should be made to the International Agency for Research on Cancer.
Distributed for the International Agency for Research on Cancer
by the Secretariat of the World Health Organization
NO'fE TO THE READER . ... . . .. ..... . .... .. . ... .... .. .... . . ... .. ...... . 11
UST OF PARTICIPANTS . . .. .. ..... .. . . . ... . . . . . . . . ... .. . ... .. . . .. . . .. .. . 13 I
Background ... . ... . . . ..... . ...... . ... . .... . ... . .. ... . ...... . .... . . . . 19
Objective and Scope .. .... .. ... ... ..... . ....... .... .. . . . . ..... . . ..... . 19
Selection of Topics for Monographs ... .. .. ... . . .... .... . .. . .... . .... .. .. . 20
Data for Monographs . . ... ..... .. . ..... . . . . ... . . . ... . . . . . . ... . ... . .. . . 21
The Working Group . . .. .... . . .. . . . . . . ... . . .... .. .. .. . ... . ... .. .. .... . 21
Working Procedures . ..... .. .. . .. . .. . ... . . .. . . . . . .. ... . ... ...... . . .. . . 21
Exposure Data ... . .. ..... . .. . . . ... . . . ... .... . . .. .. .. . ........ . . .. .. . . 22
Evidence for Carcinogenicity in Humans ... ... . . . . .... . .. . .. . . . . . ... . . . . . . 23
Studies of Cancer in Experimental Animals . .. .. . .. .. .. .. .. .......... ... . . 27
Other Relevant Data .. ... . .. .... . . ....... ... ... ... .. . .... .. . .. ... .. . . 29
Summary of Data Reported . .. . .. ... .. .. .... ... . . .. .. .. .. . ... ... . . .. .. . 30
Evaluation .... . . .. .... .. . ... . . ... . . . ..... . .. . .. .. .. .. . . ..... . . .. .. . . 32
References . . .... ....... . .... .. .. .. .......... . . ... .. . .. . . . . ... . .. . .. . 36
GENERAl. REMARKS .. ... .. .. .. .. .. . . ... .. . . . .. .. .. ........ . .. . . .. .. . . 39
1. Exposure data ... ..... . .. . . ..... . . ...... . .. ... . .. .... .. . .... . .. . .. ... . 43
1.1 Nomenclature . .. ... . . . . . . .. . . . ...... . . . .. ... .. .. ... .. ... . . . . . .. . 43
1.1.1 Optical radiation ... . .. . .. ... ... . ....... . . . ........ . . .. .. ... . 43
1.1.2 Quantities and units ... . . .. . .. .. . ........ . . .... ...... . .. ... . . 45
1.1.3 Units of biologically effective ultraviolet radiation ...... . .. .. .... . 46
1.2 Methods for measuring ultraviolet radiation ..... .. ...... ... . . . .. . .. .. . 47
1.2.1 Spectroradiometry .. . ..... .. .. . ...... . . .. . . ...... .. .. .. .. . . . 47
1.2.2 Wavelength-independent (thermal) detectors . ..... .. ... . ... ... . . . 48
1.2.3 Wavelength-dependent detectors . ... . ... .. ... .. .... ... . ...... . . 48
1.3 Sources and exposures . . .... . ... . .... .... . .... . . ... . ..... .... .. ... . 49
1.3.1 Solar ultraviolet radiation ....... . . . . . . ... . ........ . ..... .. .. . 50
(a) Measurements of terrestrial solar radiation ... . . . . .. .. . . . .. . . 54
(b) Personal exposures ... . ... . . ... .. . .. ... .. . ... . .. . .. .. .. . . 57
1.3.2 Exposure to artificial sources of ultraviolet radiation . . . . . .. . . . . .. . 58
(a) Sources ... .. ... ... . . . ... .. .. . . .. .. ...... . ..... . .. . .. .. . 58
(i) Incandescent sources .... . ... .... . . .. .. .. . .. ... . ... .. . 58
. (ii) Gas discharge lamps . . . .. ... ... . .. ... .. ... . ... .. .. . . . 59
(iii) Arc lamps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
(iv) Fluorescent lamps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
(v) Metal halide lamps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
(vi) Electrodeless lamps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
(b) Human exposure ...... .. .... . ...................... .. .. . 60
(i) Cosmetic use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
(ii) Medical and dental applications . . . . . . . . . . . . . . . . . . . . . . . . 63
(iii) Occupational exposures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
(iv) General lighting ..................... .. . . . . . . . . . . . . . 70
(c) Regulations and guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
(i) Cosmetic use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
(ii) Occupational exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
2. Studies of cancer in humans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
2.1 Solar radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
2.1.1 Nonmelanocytic skin cancer ... . .. . . . . .......... . . . . . ... ....... 73
(a) Case reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
(i) Studies of xeroderma pigmentosum patients . . . . . . . . . . . . . . 73
(ii) Studies of transplant recipients . . . . . . . . . . . . . . . . . . . . . . . . 73
(b) Descriptive studies . .. .... . ......... .. .. ........ ... .. ..... 74
(i) Host factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
(ii) Anatomical distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
(iii) Geographical variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
(iv) Migration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
(v) Occupation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
(c) Cross-sectional studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
(d) Case-control studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
(e) Cohort studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
(j) Collation of results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
2.1.2 Cancer of the lip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
(a) Descriptive studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
(i) Geographical variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
(ii) Occupation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
(b) Case-control studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
2.1.3 Malignant melanoma of the skin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
(a) Case reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
(b) Descriptive studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
(i) Sex distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
(ii) Age distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
(iii) Anatomical distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
(iv) Ethnic origin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
(v) Geographical variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
(vi) Migration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
(vii) Socioeconomic status and occupation . . . . . . . . . . . . . . . . . 99
(c) Case-control studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
(i) Australia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
(ii) Europe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
(iii) North America . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
(d) Collation of results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
(i) Total sun exposure: potential exposure by place of residence 113
(ii) Biological response to total sun exposure . . . . . . . . . . . . . . . 113
(iii) Total sun exposure assessed by questionnaire . . . . . . . . . . . . 115
(iv) Short periods of residence implying high potential exposure 115
(v) Occupational exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
(vi) Intermittent exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
(vii) Sunburn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
2.1.4 Malignant melanoma of the eye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
(a) Case reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
(b) Descriptive studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
(i) Ethnic origin . . . . . . .. . . . . .. .. . . . . . . . .. . . . .. . .. . .. . . 125
(ii) Place of birth and residence . . . . . . . . . . . . . . . . . . . . . . . . . . 125
(iii) Occupation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
(iv) History of skin cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
(c) Case-control studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
2.1.5 Other cancers . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . 130
2.2 Artificial sources of ultraviolet radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
2.2.1 Nonmelanocytic skin cancer. . .. .......... . . . . . . . . . . . . . . . . . . . . 130
2.2.2 Malignant melanoma of the skin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
2.2.3 Malignant melanoma of the eye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
2.3 Premalignant conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
2.3.1 Basal-ceH naevus syndrome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
2.3.2 Dysplastic naevus syndrome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
2.4 Molecular genetics of human skin cancers . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
2.4.1 ras Gene mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
2.4.2 p53 Gene mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
3. Studies of cancer in animals. . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
3.1 Experimental conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
3.1.1 Species studied . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
3.1.2 Wavelength ranges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
3.1.3 Measured doses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
3.1.4 Protocols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
3.2 Broad-spectrum radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
3.2.1 Sunlight ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
3.2.2 Solar-simulated radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
3.2.3 Sources emitting UVC, UVB and UVA radiation . . . . . . . . . . . . . . . . 142
3.3 Sources emitting mainly UVB radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
3.3.1 Mouse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
3.3.2 Rat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
3.3.3 Hamster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
3.3.4 Guinea-pig. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
3.3.5 Fish . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
3.3 .6 Opossum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
3.4 Sources emitting mainly UVC radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
3.4.1 Mouse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
3.4.2 Rat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
3.5 Sources emitting mainly UVA radiation . . . . . ... ..... . ... . .......... . 148
3.6 Interaction of wavelengths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
3.6.1 Interaction of exposures given on the same day . . . . . . . . . . . . . . . . . . 150
3.6.2 Long-term interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
3.7 Additional experimental observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
3.7.1 Tumour types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
3.7.2 Dose and effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
3.7.3 Dose delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
3. 7.4 Action spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
3.7.5 Pigmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
3.8 Administration with known chemical carcinogens . . . . . . . . . . . . . . . . . . . . . 155
3.8.1 Administration with polycyclic aromatic hydrocarbons . . . . . . . . . . . . 156
(a) 3,4-Benzo[a]pyrene ........ ... . . ..... .. ... . .... ... ...... 156
(b) 7,12-Dimethylbenz[a]anthracene ... . ..................... . 156
3.8.2 Administration with other agents with promoting activity . . . . . . . . . . 157
(a) Croton oil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
(b) 12-0-Tetradecanoylphorbol 13-acetate . . . . . . . . . . . . . . . . . . . . . 158
(c) Benzoyl perm<ide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
(d) Methyl ethyl ketone peroxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
3.9 Interaction with immunosuppressive agents . . . . . . . . . . . . . . . . . . . . . . . . . . 160
3.10 Molecular genetics of animal skin tumours induced by ultraviolet radiation 161
4. Other relevant data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
4.1 Transmission and absorption in biological tissues . . . . . . . . . . . . . . . . . . . . . . 163
4.1.1 Epidermis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
(a) Humans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
(b) Experimental systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
(c) Epidermal chromophores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
(d) Enhancement of epidermal penetration of ultraviolet radiation . 166
4.1.2 Eye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
(a) Humans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
(b) Experimental systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
4.2 Adverse effects (other than cancer) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
4.2.1 Epidermis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
(a) Humans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
(i) Erythema and pigmentation (sunburn and suntanning) . . . . 167
(ii) Pigmented naevi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
(iii) Ultrastructural changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
(iv) Keratosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
(v) Photosensitivity disorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
(b) Experimental systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
(c) Comparison of humans and animals . . . . . . . . . . . . . . . . . . . . . . . 174
4.2.2 Immune response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
(a) Humans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
(i) Contact hypersensitivity (allergy) . . . . . . . . . . . . . . . . . . . . . . 175
(ii) Lymphocytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
(iii) Infectious diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
(iv) Photosensitive diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
(b) Experimental systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
(i) Contact hypersensitivity ..... ........... . . .'. . . . . . . . . . 177
(ii) Delayed hypersensitivity to injected antigens . . . . . . . . . . . . 179
(iii) Immunology of ultraviolet-induced skin cancer . . . . . . . . . . 180
(iv) Transplantation immunity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
(v) Infectious diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
(vi) Human lymphocytes in vitro ......... . . .. . .............. 182
(c) Comparison of humans and animals . . . . . . . . . . . . . . . . . . . . . . . 182
4.2.3 Eye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
(a) Humans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
(i) Anterior eye (cornea, conjunctiva) . . . . . . . . . . . . . . . . . . . . . 183
(ii) Lens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
(iii) Posterior eye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
(b) Experimental systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
(i) Anterior eye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
(ii) Lens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
(iii) Posterior eye .. ... ...... . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
(c) Comparison of humans and animals . . . . . . . . . . . . . . . . . . . . . . . 184
4.3 Photoproduct formation ........... . ................ .. .... ; . . . . . . . 185
4.3.1 DNA photoproducts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
(a) Cyclobutane-type pyrimidine dimers . . . . . . . . . . . . . . . . . . . . . . . 185
(b) Pyrimidine- pyrimidone (6-4) photoproducts . . . . . . . . . . . . . . . . . 186
(c) Thymine glycols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
(d) Cytosine damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
(e) Purine damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
(/) DNA strand breaks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
(g) DNA-protein cross-links. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
4.3.2 Other chromophores and targets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
(a) Chromophores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
(b) Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
4.4 Human excision repair disorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
4.4.1 Xeroderma pigmentosum ................................. ... 191
4.4.2 Trichothiodystrophy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
4.4.3 Cockayne 's syndrome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
4.4.4 Role of immunosuppression. . ......... . . . . . . . . . . . . . . . . . . . . . . . 193
4.5 Genetic and related effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
4.5.1 Humans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
(a) Epidermis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
(i) Broad-spectrum ultraviolet radiation, including solar
simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
(ii) UVA radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
(iii) UVB radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
(iv) UVC radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
(b) Lymphocytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
(i) Broad-spectrum ultraviolet radiation . . . . . . . . . . . . . . . . . . . 198
(ii) UVA radiation ..................... . .............. . 199
(iii) UVB radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
4.5.2 Experimental systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
(a) DNA damage ... ... ............... .. ... .. . .... ... .. .. . . 199
(b) Mutagenicity . ........ . .... . .. . ... . ......... . . .. ..... .. 200
(c) Chromosomal effects .. ....................... . .......... 202
(d) Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
(e) Effects of cellular and viral gene expression . . . . . . . . . . . . . . . . . 204
5. Summary of data reported and evaluation ..................... -. . . . . . . . . . . 217
5.1 Exposuredata . . . . . . .... .. ... . . ...... ............. ... ........... 217
5.2 Human carcinogenicity data ....... . .... . ........ .. .... .... . .. ... . . 218
5.2.1 Solar radiation ......................... . . . ..... . ... .. . . . .. 218
(a) Nonmelanocytic skin cancer . ......... . . .. .... .... ........ 218
(b) Cancer of the lip ........................ . ... ........... 219
(c) Malignant melanoma of the skin ...... .. .. ....... . ... ..... 219
(d) Melanoma of the eye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2ZO
(e) Other cancers .............. . ........ . .. .............. .. 220
5.2.2 Artificial sources of ultraviolet radiation . . . . . . . . . . . . . . . . . . . . . . . 220
5.2.3 Molecular genetics of human skin cancers . . . . . . . . . . . . . . . . . . . . . . 221
5.3 Carcinogenicity in experimental animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
5.4 Other relevant data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
5.4.1 Transmission and absorption ........ . .. . ..... .. ... ....... .. .. 222
5.4.2 Effects on the skin ................ . ..... . ..... ... .. ... . . ... 222
5.4.3 Effects on the immune response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
5.4.4 DNA photoproducts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
5.4.5 Genetic and related effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
5.5 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
6. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
SUMMARY OF FINAL EVALUATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
GLOSSARY OF TERMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
Appendix 1. Topical sunscreens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
1. General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
2. Protective effects . .... . ... . . ..... . .. .. .. . . ..... . . . ..... .. .. . .. : . . . . . . . 286
2.1 Against DNA damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
2.2 Against acute and chronic actinic damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
2.3 Against immunological alterations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
2.4 Against tumour formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
3. Adverse effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
3.1 Acute toxicity . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
3.2 Chronic toxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
3.3 Reduced vitamin D synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
4. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
CUMULATIVE INDEX TO THE MONOGRAPHS SERIES . . . . . . . . . . . . . . . . . . . 291 ~ - - - - - - ' __________________________ _
The term 'carcinogenic risk' in the /ARC Monographs series is taken to mean the proba-
bility that exposure to an agent will lead to cancer in humans.
Inclusion of an agent in the Monographs does not imply that it is a carcinogen, only that
the published data have been examined. Equally, the fact that an agent has not yet been
evaluated in a monograph does not mean that it is not carcinogenic.
The evaluations of carcinogenic risk are made by international working groups of in-
dependent scientists and are qualitative in nature. No recommendation is given for regu-
lation or legislation.
Anyone who is aware of published data that may alter the evaluation of the carcinogenic
risk of an agent to humans is encouraged to make this information available to the Unit of
Carcinogen Identification and Evaluation, International Agency for Research on Cancer,
150 cours Albert Thomas, 69372 Lyon Cedex 08, France, in order that the agent may be
considered for re-evaluation by a future Working Group.
Although every effort is made to prepare the monographs as accurately as possible,
mistakes may occur. Readers are requested to communicate any errors to the Unit of
Carcinogen Identification and Evaluation, so that corrections can be reported in future
Lyon, 11-18 February 1992
C. Arlett, MRC Cell Mutation Unit, University of Brighton, Falmer, Brighton BNl 9RR,
United Kingdom
B. Bridges, MRC Cell Mutation Unit, University of Brighton, Falmer, Brighton BNl 9RR,
United Kingdom (Chairman)
A Br!ISgger, Department of Genetics, Institute for Cancer Research, Montebello, 0310 Oslo
3, Norway
B.L. Diffey, Regional Medical Physics Department, Dryburn Hospital, Durham DHl 5TW,
United Kingdom
J .M. Elwood, Hugh Adam Cancer Epidemiology Unit, University of Otago Medical School,
PO Box 913, Dunedin, New Zealand
E.A. Emmett, Worksafe Australia, National Occupational Health and Safety Commission,
92 Parramatta Road, Camperdown, NSW 2050, Australia
D. English, NHMRC Research Unit, The Queen Elizabeth II Medical Centre, University of
Western Australia, Nedlands, WA 6009, Australia
P.D. Forbes, Temple University, Biohazards Control Office, Environmental Health and
Safety Offices, 3307 North Broad Street, Philadelphia, PA 19140, USA
R.P. Gallagher, British Columbia Cancer Agency, 600 West lOth Avenue, Vancouver, BC
V52 4E6, Canada
J.W. Grisham, Department of Pathology, University of North Carolina, Brinkhous-Bullitt
Building CB# 7525, Chapel Hill, NC 27599, USA
Unable to attend: J. Marshall, Department of Ophthalmology, Block 8, UNDS, StThomas's Hospital,
London SEl 7EH, United Kingdom
K. Kraemer, National Cancer Institute, Division of Cancer Etiology, Laboratory of
Molecular Carcinogenesis, Building 37, Room 3D-06, Bethesda, MD 20892, USA
J.C. van der Leun, Institute of Dermatology, University Hospital Utrecht, Heidelberglaan
100, 3584 CX Utrecht, The Netherlands
T. Mack, University of Southern California, School of Medicine, Department of Preventive
Medicine, Parkview Medical Building, 1420 San Pablo Street, Los Angeles, CA
90033-9987, USA (Vice-Chairman)
W.L. Morison, Johns Hopkins University, Baltimore, MD 21205, USA
S. Olin, ILSI Risk Science Institute, 1126 Sixteenth Street NW, Washington DC 20036, USA
A 0sterlind, Danish Cancer Registry, Rosenvaengets Hovedvej 35, Box 839, 2100
Copenhagen, Denmark
D.H. Sliney, Laser Branch, US Army Environmental Hygiene Agency, Aberdeen Proving
Ground, MD 21010-5422, USA
F. Stenbiick, Department of Pathology, University of Oulu, Kajaanintie 52 D, 90220 Oulu 22,
R.M. 1}rrrell, Swiss Institute for Experimental Cancer Research, 1066 Epalinges-sur-
Lausanne, Switzerland
AR. Young, Photobiology Department, StJohn's Institute of Dermatology, StThomas's
Hospital, London SE1 7EH, United Kingdom
Representative of the US Food and Drug Administration
J.Z. Beer, US Food and Drug Administration, Center for Devices and Radiological Health,
12709 Twinbrook Parkway, Rockville, MD 20852, USA
R.C. Burton, Department of Surgery, John Hunter Hospital, Locked Bag #1, Newcastle Mail
Centre, Newcastle, NSW 2310, Australia
C.J. Portier, National Institute of Environmental Health Sciences, PO Box 12233, Res.earch
Triangle Park, NC 27709, USA
B. Armstrong, Deputy Director
H. Bartsch, Unit of Environmental and Host Factors
P. Boffetta, Unit of Analytical Epidemiology
J .R.P. Cabral, Unit of Mechanisms of Carcinogenesis
E. Cardis, Director's Office
M. Friesen, Unit of Environmental and Host Factors
M.-J. Ghess, Unit of Carcinogen Identification and Evaluation
J. Hall, Unit of Mechanisms of Carcinogenesis
E. Heseltine, Lajarthe, St Leon-sur-Vezere, France
A Kricker, Unit of Descriptive Epidemiology
V. Krutovskikh, Unit of Multistage Carcinogenesis
D. McGregor, Unit of Carcinogen Identification and Evaluation
D. Mietton, Unit of Carcinogen Identification and Evaluation
H. M0ller, Unit of Carcinogen Identification and Evaluation
R. Montesano, Unit of Mechanisms of Carcinogenesis
T. Nakazawa, Unit of Multistage Carcinogenesis
I. O'Neill, Unit of Environmental and Host Factors
M. Parkin, Unit of Descriptive Epidemiology
C. Partensky, Unit of Carcinogen Identification and Evaluation
I. Peterschmitt, Unit of Carcinogen Identification and Evaluation, Geneva, Switzerland
D. Shuker, Unit of Environmental and Host Factors
L. Tomatis, Director
H. Vainio, Unit of Carcinogen Identification and Evaluation
1. Wilbourn, Unit of Carcinogen Identification and Evaluation
H. Yamasaki, Unit of Multistage Carcinogenesis
Secretarial assistance
1. Cazeaux
M. Lezere
S. Reynaud
In 1969, the International Agency for Research on Cancer (IARC) initiated a pro-
gramme to evaluate the carcinogenic risk of chemicals to humans and to produce mono-
graphs on individual chemicals. The Monographs programme has since been expanded to
include consideration of exposures to complex mixtures of chemicals (which occur, for
example, in some occupations and as a result of human habits) and of exposures to other
agents, such as radiation and viruses. With Supplement 6 (IARC, 1987a), the title of the series
was modified from /ARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals
to Humans to /ARC Monographs on the Evaluation of Carcinogenic Risks to Humans, in order
to reflect the widened scope of the programme.
The criteria established in 1971 to evaluate carcinogenic risk to humans were adopted by
the working groups whose deliberations resulted in the first 16 volumes of the /ARC
Monographs series. Those criteria were subsequently updated by further ad-hoc working
groups (IARC, 1977, 1978, 1979, 1982, 1983. 1987b, 1988, 199la; Vainio eta/., 1992).
The objective of the programme is to prepare, with the help of international working
groups of experts, and to publish in the form of monographs, critical reviews and evaluations
of evidence on the carcinogenicity of a wide range of human exposures. The Monographs may
also indicate where additional research efforts are needed.
The Monographs represent the first step in carcinogenic risk assessment, which involves
examination of all relevant information in order to assess the strength of the available
evidence that certain exposures could alter the incidence of cancer in humans. The second
step is quantitative risk estimation. Detailed, quantitative evaluations of epidemiological
data may be made in the Monographs, but without extrapolation beyond the range of the data
This project is supported by PHS Grant No.2 UOI CA33193-10 awarded by the US National Cancerlnstitute,
Department of Health and Human Services. Since 1986, the programme has also been supported by the Com-
mission of the European Communities.
available. Quantitative extrapolation from experimental data to the human situation is not
The term ' carcinogen' is used in these monographs to denote an exposure that is capable
of increasing the incidence of malignant neoplasms; the induction of benign neoplasms may
in some circumstances (seep. 28) contribute to the judgement that the exposure is carcino-
genic. The terms 'neoplasm' and 'tumour' are used interchangeably.
Some epidemiological and experimental studies indicate that different agents may act at
different stages in the carcinogenic process, and several different mechanisms may be
involved. The aim of the Monographs has been, from their inception, to evaluate evidence of
carcinogenicity at any stage in the carcinogenesis process, independently of the underlying
mechanisms. Information on mechanisms may, however, be used in making the overall eval-
uation (IARC, 1991a; Vainio eta/., 1992; see also pp. 33-34).
The Monographs may assist national and international authorities in making risk
assessments and in formulating decisions concerning any necessary preventive measures.
The evaluations of IARC working groups are scientific, qualitative judgements about the
evidence for or against carcinogenicity provided by the available data. These evaluations
represent only one part of the body of information on which regulatory measures may be
based. Other components of regulatory decisions may vary from one situation to another and
from country to country, responding to different socioeconomic and national priorities.
Therefore, no recommendation is given with regard to regulation or legislation, which are the
responsibility of individual governments and/or other international organizations.
The /ARC Monographs are recognized as an authoritative source of information on the
carcinogenicity of a wide range of human exposures. A users' survey, made in 1988, indicated
that the Monographs are consulted by various agencies in 57 countries. Each volume is
generally printed in 4000 copies for distribution to governments, regulatory bodies and
interested scientists. The Monographs are also available via the Distribution and Sales
Service of the World Health Organization.
Topics are selected on the basis of two main criteria: (a) there is evidence of human
exposure, and (b) there is some evidence or suspicion of carcinogenicity. The term 'agent' is
used to include individual chemical compounds, groups of related chemical compounds,
physical agents (such as radiation) and biological factors (such as viruses). Exposures to
mixtures of agents may occur in occupational exposures and as a result of personal and
cultural habits (like smoking and dietary practices). Chemical analogues and compounds
with biological or physical characteristics similar to those of suspected carcinogens may also
be considered, even in the absence of data on a possible carcinogenic effect in humans or
experimental animals.
The scientific literature is surveyed for published data relevant to an assessment of
carcinogenicity. The IARC surveys of chemicals being tested for carcinogenicity (IARC,
1973-1990) and directories of on-going research in cancer epidemiology (IARC, 1976-
1991) often indicate those exposures that may be scheduled for future meetings. Ad-hoc
working groups convened by IARC in 1984, 1989 and 1991 gave recommendations as to
which agents should be evaluated in the /ARC Monographs series (IARC, 1984, 1989, 1991b ).
As significant new data on subjects on which monographs have already been prepared
become available, re-evaluations are made at subsequent meetings, and revised monographs
are published.
The Monographs do not necessarily cite all the literature concerning the subject of an
evaluation. Only those data considered by the Working Group to be relevant to making the
evaluation are included.
With regard to biological and epidemiological data, only reports that have been
published or accepted for publication in the openly available scientific literature are
reviewed by the working groups. In certain instances, government agency reports that have
undergone peer review and are widely available are considered. Exceptions may be made on
an ad-hoc basis to include unpublished reports that are in their final form and publicly
available, if their inclusion is considered pertinent to making a final evaluation (see pp. 32
et seq.). In the sections on chemical and physical properties, on analysis, on production and
use and on occurrence, unpublished sources of information may be used.
Reviews and evaluations are formulated by a working group of experts. The tasks of the
group are: (i) to ascertain that all appropriate data have been collected; (ii) to select the data
relevant for the evaluation on the basis of scientific merit; (iii) to prepare accurate summaries
of the data to enable the reader to follow the reasoning of the Working Group; (iv) to
evaluate the results of experimental and epidemiological studies on cancer; (v) to evaluate
data relevant to the understanding of mechanism of action; and (vi) to make an overall
evaluation of the carcinogenicity of the exposure to humans.
Working Group participants who contributed to the considerations and evaluations
within a particular volume are listed, with their addresses, at the beginning of each publi-
cation. Each participant who is a member of a working group serves as an individual scientist
and not as a representative of any organization, government or industry. In addition,
nominees of national and international agencies and industrial associations may be invited as
Approximately one year in advance of a meeting of a working group, the topics of the
monographs are announced and participants are selected by IARC staff in consultation with
other experts. Subsequently, relevant biological and epidemiological data are collected by
IARC from recognized sources of information on carcinogenesis, including data storage and
retrieval systems such as BIOSIS, Chemical Abstracts, CANCERLIT, MEDLINE and
TOXUNE-including EMIC and ETIC for data on genetic and related effects and
teratogenicity, respectively.
For chemicals and some compJex mixtures, the major collection of data and the
preparation of first drafts of the sections on chemical and physical properties, on analysis, on
production and use and on occurrence are carried out under a separate contract funded by
the US National Cancer Institute. Representatives from industrial associations may assist in
the preparation of sections on production and use. Information on production and trade is
obtained from governmental and trade publications and, in some cases, by direct contact
with industries. Separate production data on some agents may not be available because their
publication could disclose confidential information. Information on uses may be obtained
from published sources but is often complemented by direct contact with manufacturers.
Efforts are made to supplement this information with data from other national and
international sources.
Six months before the meeting, the material obtained is sent to meeting participants, or
is used by IARC staff, to prepare sections for the first drafts of monographs. The first drafts
are compiled by IARC staff and sent, prior to the meeting, to all participants of the Working
Group for review.
The Working Group meets in Lyon for seven to eight days to discuss and finalize the texts
of the monographs and to formulate the evaluations. After the meeting, the master copy of
each monograph is verified by consulting the original literature, edited and prepared for
publication. The aim is to publish monographs within nine months of the Working Group
The available studies are summarized by the Working Group, with particular regard to
the qualitative aspects discussed below. In general, numerical findings are indicated as they
appear in the original report; units are converted when necessary for easier comparison. The
Working Group may conduct additional analyses of the published data.and use them in their
assessment of the evidence; the results of such supplementary analyses are given in square
brackets. When an important aspect of a study, directly impinging on its interpretation,
should be brought to the attention of the reader, a comment is given in square brackets.
Sections that indicate the extent of past and present human exposure, the sources of
exposure, the people most likely to be exposed and the factors that contribute to the exposure
are included at the beginning of each monograph.
Most monographs on individual chemicals, groups of chemicals or complex mixtures
include sections on chemical and physical data, on analysis, on production and use and on
occurrence. In monographs on, for example, physical agents, biological factors, occupational
exposures and cultural habits, other sections may be included, such as: historical pers-
pectives, description of an industry or habit, chemistry of the complex mixture or taxonomy.
For chemical exposures, the Chemical Abstracts Services Registry Number, the latest
Chemical Abstracts Primary Name and the IUPAC Systematic Name are recorded; other
synonyms are given, but the list is not necessarily comprehensive. For biological agents,
taxonomy and structure are described, and the degree of variability is given, when applicable.
Information on chemical and physical properties and, in particular, data relevant to
identification, occurrence and biological activity are included. For biological agents, mode of
replication, life cycle, target cells, persistence and latency, host response and description of
nonmalignant disease caused by them are given. A description of technical products of
chemicals includes trades names, relevant specifications and available information on
.._ ______________ - - - - - - - - ------
composition and impurities. Some of the trade names given may be those of mixtures in
which the agent being evaluated is only one of the ingredients.
The purpose of the section on analysis is to give the reader an overview of current
methods, with emphasis on those widely used for regulatory purposes. Methods for moni-
toring human exposure are also given, when available. No critical evaluation or recommen-
dation of any of the methods is meant or implied. The !ARC publishes a series of volumes,
Environmental Carcinogens: Methods of Analysis and Exposure Measurement (IARC, 1978-
91), that describe validated methods for analysing a wide variety of chemicals and mixtures.
For biological agents, methods of detection and exposure assessment are described, inclu-
ding their sensitivity, specificity and reproducibility.
The dates of first synthesis and of first commercial production of a chemical or mixture
are provided; for agents which do not occur naturally, this information may allow a reaso-
nable estimate to be made of the date before which no human exposure to the agent could
have occurred. The dates of first reported occurrence of an exposure are also provided. In
addition, methods of synthesis used in past and present commercial production and different
methods of production which may give rise to different impurities are described.
Data on production, international trade and uses are obtained for representative
regions, which usually include Europe, Japan and the USA It should not, however, be
inferred that those areas or nations are necessarily the sole or major sources or users of the
agent. Some identified uses may not be current or major applications, and the coverage is not
necessarily comprehensive. In the case of drugs, mention of their therapeutic uses does not
necessarily represent current practice nor does it imply judgement as to their therapeutic
Information on the occurrence of an agent or mixture in the environment is obtained
from data derived from the monitoring and surveillance of levels in occupational envi-
ronments, air, water, soil, foods and animal and human tissues. When available, data on the
generation, persistence and bioaccumulation of the agent are also included. In the case of
mixtures, industries, occupations or processes, information is given about all agents present.
For processes, industries and occupations, a historical description is also given, noting
variations in chemical composition, physical properties and levels of occupational exposure
with time. For biological agents, the epidemiology of infection is described.
Statements concerning regulations and guidelines (e.g., pesticide registrations, maximal
levels pennitted in foods, occupational exposure limits) are included for some countries as
indications of potential exposures, but they may not reflect the most recent situation, since
such limits are continuously reviewed and modified. The absence of information on regula-
tory status for a country should not be taken to imply that that country does not have
regulations with regard to the exposure. For biological agents, legislation and control,
including vaccines and therapy, are described.
(a) Types of studies considered
Three types of epidemiological studies of cancer contribute to the assessment of carcino-
genicity in humans-cohort studies, case-control studies and correlation studies. Rarely,
results from randomized trials may be available. Case reports of cancer in humans may also
be reviewed.
Cohort and case-control studies relate individual exposures under study to the occur-
rence of cancer in individuals and provide an estimate of relative risk (ratio of incidence in
those exposed to incidence in those not exposed) as the main measure of association.
In correlation studies, the units of investigation are usually whole populations (e.g., in
particular geographical areas or at particular times), and cancer frequency is related to a
summary measure of the exposure of the population to the agent, mixture or exposure
circumstance under study. Because individual exposure is not documented, however, a causal
relationship is less easy to infer from correlation studies than from cohort and case-control
studies. Case reports generally arise from a suspicion, based on clinical experience, that the
concurrence of two events- that is, a particular exposure and occurrence of a cancer-has
happened rather more frequently than would be expected by chance. Case reports usually
lack complete ascertainment of cases in any population, definition or enumeration of the
population at risk and estimation of the expected number of cases in the absence of exposure.
The uncertainties surrounding interpretation of case reports and correlation studies make
them inadequate, except in rare instances, to form the sole basis for inferring a causal
relationship. When taken together with case- control and cohort studies, however, relevant
case reports or correlation studies may add materially to the judgement that a causal
relationship is present.
Epidemiological studies of benign neoplasms, presumed preneoplastic lesions and other
end-points thought to be relevant to cancer are also reviewed by working groups. They may,
in some instances, strengthen inferences drawn from studies of cancer itself.
(b) Quality of studies considered
The Monographs are not intended to summarize all published studies. Those that are
judged to be inadequate or irrelevant to the evaluation are generally omitted. They may be
mentioned briefly, particularly when the information is considered to be a useful supplement
to that in other reports or when they provide the only data available. Their inclusion does not
imply acceptance of the adequacy of the study design or of the analysis and interpretation of
the results, and limitations are clearly outlined in square brackets at the end of the study
It is necessary to take into account the possible roles of bias, confounding and chance in
the interpretation of epidemiological studies. By 'bias' is meant the operation of factors -in
study design or execution that lead erroneously to a stronger or weaker association than in
fact exists between disease and an agent, mixture or exposure circumstance. By 'confounding'
is meant a situation in which the relationship with disease is made to appear stronger or to
appear weaker than it truly is as a result of an association between the apparent causal factor
and another factor that is associated with either an increase or decrease in the incidence of
the disease. In evaluating the extent to which these factors have been minimized in an
individual study, working groups consider a number of aspects of design and analysis as
described in the report of the study. Most of these considerations apply equally to
case-control, cohort and correlation studies. Lack of clarity of any of these aspects in the
L---------------- - ---
reporting of a study can decrease its credibility and the weight given to it in the final
evaluation of the exposure.
Firstly, the study population, disease (or diseases) and exposure should have been well
defined by the authors. Cases of disease in the study population should have been identified
in a way that was independent of the exposure of interest, and exposure should have been
assessed in a way that was not related to disease status.
Secondly, the authors should have taken account in the study design and analysis of other
variables that can influence the risk of disease and may have been related to the exposure of
interest. Potential confounding by such variables should have been dealt with either in the
design of the study, such as by matching, or in the analysis, by statistical adjustment. In cohort
studies, comparisons with local rates of disease may be more appropriate than those with
national rates. Internal comparisons of disease frequency among individuals at different
levels of exposure should also have been made in the study.
Thirdly, the authors should have reported the basic data on which the condusions are
founded, even if sophisticated statistical analyses were employed. At the very least, they
should have given the numbers of exposed and unexposed cases and controls in a case-
control study and the numbers of cases observed and expected in a cohort study. Further
tabulations by time since exposure began and other temporal factors are also important. In a
cohort study, data on all cancer sites and all causes of death should have been given, to reveal
the possibility of reporting bias. In a case-control study, the effects of investigated factors
other than the exposure of interest should have been reported.
Finally, the statistical methods used to obtain estimates of relative risk, absolute rates of
cancer, confidence intervals and significance tests, and to adjust for confounding should have
been clearly stated by the authors. The methods used should preferably have been the
generally accepted techniques that have been refined since the mid-1970s. These methods
have been reviewed for case-control studies (Breslow & Day, 1980) and for cohort studies
(Breslow & Day, 1987).
(c) Inferences about mechanism of action
Detailed analyses of both relative and absolute risks in relation to temporal variables,
such as age at first exposure, time since first exposure, duration of exposure, cumulative
exposure and time since exposure ceased, are reviewed and summarized when available. The
analysis of temporal relationships can be useful in f9rmulating models of carcinogenesis. In
particular, such analyses may suggest whether a carcinogen acts early or late in the process of
carcinogenesis, although at best they allow only indirect inferences about the mechanism of
aCtion. Special attention is given to measurements of biological markers of carcinogen
exposure or action, such as DNA or protein adducts, as well as markers of early steps in the
carcinogenic process, such as proto-oncogene mutation, when these are incorporated into
epidemiological studies focused on cancer incidence or mortality. Such measurements may
allow inferences to be made about putative mechanisms of action (IARC, 199la; Vainio et al.,
(d) Criteria for causality
After the quality of individual epidemiological studies of cancer has been summarized
and assessed, a judgement is made concerning the strength of evidence that the agent,
mixture or exposure circumstance in question is carcinogenic for humans. In making their
judgement, the Working Group considers several criteria for causality. A strong association
(i.e., a large relative risk) is more likely to indicate causality than a weak association,
although it is recognized that relative risks of small magnitude do not imply lack of causality
and may be important if the disease is common. Associations that are replicated in several
studies of the same design or using different epidemiological approaches or under different
circumstances of exposure are more likely to represent a causal relationship than isolated
observations from single studies. If there are inconsistent results among investigations,
possible reasons are sought (such as differences in amount of exposure), and results of studies
judged to be of high quality are given more weight than those from studies judged to be
methodologically less sound. When suspicion of carcinogenicity arises largely from a single
study, these data are not combined with those from later studies in any subsequent
reassessment of the strength of the evidence.
If the risk of the disease in question increases with the amount of exposure, this is
considered to be a strong indication of causality, although absence of a graded response is
not necessarily evidence against a causal relationship. Demonstration of a decline in risk
after cessation of or reduction in exposure in individuals or in whole populations also
supports a causal interpretation of the findings.
Although a carcinogen may act upon more than one target, the specificity of an asso-
ciation (i.e., an increased occurrence of cancer at one anatomical site or of one morpho-
logical type) adds plausibility to a causal relationship, particularly when excess cancer
occurrence is limited to one morphological type within the same organ.
Although rarely available, results from randomized trials showing different rates among
exposed and unexposed individuals provide particularly strong evidence for causality.
When several epidemiological studies show little or no indication of an association
between an exposure and cancer, the judgement may be made that, in the aggregate, they
show evidence of lack of carcinogenicity. Such a judgement requires first of an that the
studies giving rise to it meet, to a sufficient degree, the standards of design and analysis
described above. Specifically, the possibility that bias, confounding or misclassification of
exposure or outcome could explain the observed results should be considered and excluded
with reasonable certainty. In addi.tion, all studies that are judged to be methodologically
sound should be consistent with a relative risk ofunityfor any observed level of exposure and,
when considered together, should provide a pooled estimate of relative risk which is "at or
near unity and has a narrow confidence interval , due to sufficient population size. Moreover,
no individual study nor the pooled results of all the studies should show any consistent
tendency for relative risk of cancer to increase with increasing level of exposure. It is
important to note that evidence of lack of carcinogenicity obtained in this way from several
epidemiological studies cap. apply only to the type(s) of cancer studied and to dose levels and
intervals between first exposure and observation of disease that are the same as or less than
those observed in all the studies. Experience with human cancer indicates that, in some cases,
the period from first exposure to the development of clinical cancer is seldom less than 20
years; latent periods substantially shorter than 30 years cannot provide evidence for lack of
For several agents (e.g., aflatoxins, 4-aminobiphenyl, bis(chloromethyl)ether, diethyl-
stilboestrol, melphalan, 8-methoxypsoralen (methoxsalen) plus ultraviolet radiation, mus-
tard gas and vinyl chloride), evidence of carcinogenicity in experimental animals preceded
evidence obtained from epidemiological studies or case reports. Information compiled from
the first 41 volumes of the IARC Monographs (Wilbourn et al. , 1986) shows that, of the 44
agents and mixtures for which there is sufficient or limited evidence of carcinogenicity to
humans (seep. 32), all37 that have been tested adequately produce cancer in at least one
animal species. Although this association cannot establish that all agents and mixtures that
cause cancer in experimental animals also cause cancer in humans, nevertheless, in the
absence or adequate data on humans, it is biologically plausible and prudent to regard
agents and mixtures for which there is sufficient evidence (see p. 33) of carcinogenicity in
experimental animals as if they presented a carcinogenic risk to humans. The possibility that
a given agent may cause cancer through a species-specific mechanism which does not operate
in humans (see p. 34) should also be taken into consideration.
The nature and extent of impurities or contaminants present in the chemical or mixture
being evaluated are given when available. Animal strain, sex, numbers per group, age at start
of treatment and survival are reported.
Other types of studies summarized include: experiments in which the agent or mixture
was administered in conjunction with known carcinogens or factors that modify carcinogenic
effects; studies in which the end-point was not cancer but a defined precancerous lesion; and
experiments on the carcinogenicity of known metabolites and derivatives.
For experimental studies of mixtures, consideration is given to the possibility of changes
in the physicochemical properties of the test substance during coJlection, storage, extraction,
concentration and delivery. Chemical and toxicological interactions of the components of
mixtures may result in nonlinear dose-response relationships.
An assessment is made as to the relevance to human exposure of samples tested in
experimental systems, which may involve consideration of: (i) physical and chemical charac-
teristics, (ii) constituent substances that indicate the presence of a class of substances, (iii) the
results of tests for genetic and related effects, including genetic activity profiles, DNA adduct
profiles, proto-oncogene mutation and expression and suppressor gene inactivation. The
relevance of results obtained with viral strains analogous to that being evaluated in the
monograph must also be considered.
(a) Qualitative aspects
An assessment of carcinogenicity involves several considerations of qualitative im-
portance, including (i) the experimental conditions under which the test was performed,
including route and schedule of exposure, species, strain, sex, age, duration of follow-up;
(ii) the consistency of the results, for example, across species and target organ(s); (iii) the
spectrum of neoplastic response, from preneoplastic lesions and benign tumours to mali-
gnant neoplasms; and (iv) the possible role of modifying factors.
As mentioned earlier (p. 21}, the Monographs are not intended to summarize all
published studies. Those studies in experimental animals that are inadequate (e.g., too short
a duration, too few animals, poor survival; see below) or are judged irrelevant to the
evaluation are generally omitted. Guidelines for conducting adequate long-term carcino-
genicity experiments have been outlined (e.g., Montesano et a/. , 1986).
Considerations of importance to the Working Group in the interpretation and eva-
luation of a particular study include: (i) how clearly the agent was defined and, in the case of
mixtures, how adequately the sample characterization was reported; (ii) whether the dose
was adequately monitored, particularly in inhalation experiments; (iii) whether the doses and
duration of treatment were appropriate and whether the survival of treated animals was
similar to that of controls; (iv) whether there were adequate numbers of animals per group;
(v) whether animals of both sexes were used; (vi) whether animals were allocated randomly to
groups; (vii) whether the duration of observation was adequate; and (viii) whether the data
were adequately reported. If available, recent data on the incidence of specific tumours in
historical controls, as well as in concurrent controls, should be taken into account in the
evaluation of tumour response.
When benign tumours occur together with and originate from the same cell type in an
organ or tissue as malignant tumours in a particular study and appear to represent a stage in
the progression to malignancy, it may be valid to combine them in assessing tumour inci-
dence (Huff eta/., 1989). The occurrence of lesions presumed to be preneoplastic may in
certain instances aid in assessing the biological plausibility of any neoplastic response
observed. If an agent or mixture induces only benign neoplasms that appear to be end-points
that do not readily undergo transition to malignancy, it should nevertheless be suspected of
being a carcinogen and it requires further investigation.
(b) Quantitative aspects
The probability that tumours will occur may depend on the species, sex, strain and age of
the animal, the dose of the carcinogen and the route and length of exposure. Evidence of an
increased incidence of neoplasms with increased level of exposure strengthens the inference
of a causal association between the exposure and the development of neoplasms.
The form of the dose-response relationship can vary widely, depending on the particular
agent under study and the target organ. Since many chemicals require metabolic activation
before being converted into their reactive intermediates, both metabolic and pharmaco-
kinetic aspects are important in determining the dose-response pattern. Saturation of steps
such as absorption, activation, inactivation and elimination may produce nonlinearity in the
dose-response relationship, as could saturation of processes such as DNA repair (Hoe] et al.,
1983; Gart et al . 1986).
(c) Statistical analysis of long-term experiments in animals
Factors considered by the Working Group include the adequacy of the information given
for each treatment group: (i) the number of animals studied and the number examined
histologically, (ii) the number of animals with a given tumour type and (iii) length of survival.
The statistical methods used should be clearly stated and should be the generally accepted
techniques refined for this purpose (Peto eta/. , 1980; Gart eta/., 1986). When there is no
difference in survival between control and treatment groups, the Working Group usually
compares the proportions of animals developing each tumour type in each of the groups.
Otherwise, consideration is given as to whether or not appropriate adjustments have been
- - ---------------- - - - - - - - -
made for differences in survival. These adjustments can include: comparisons of the
proportions of tumour-bearing animals among the effective number of animals (alive at the
time the first tumour is discovered), in the case where most differences in survival occur
before tumours appear; life-table methods, when tumours are visible or when they may be
considered 'fatal' because mortality rapidly follows tumour development; .and the Mantel-
Haenszel test or logistic regression, when occult tumours do not affect the animals' risk of
dying but are 'incidental' findings at autopsy.
In practice, classifying tumours as fatal or incidental may be difficult. Several survival-
adjusted methods have been developed that do not require this distinction (Gart et al., 1986),
although they have not been fully evaluated.
(a) Absorption, distribution, metabolism and excretion
Concise information is given on absorption, distribution (including placental transfer)
and excretion in both humans and experimental animals. Kinetic factors that may affect the
dose-response relationship, such as saturation of uptake, protein binding, metabolic activa-
tion, detoxification and DNA repair processes, are mentioned. Studies that indicate the
metabolic fate of the agent in humans and in experimental animals are summarized briefly,
and comparisons of data from humans and animals are made when possible. Comparative
information on the relationship between exposure and the dose that reaches the target site
may be of particular importance for extrapolation between species.
(b) Toxic effects
Data are given on acute and chronic toxic effects (other than cancer), such as organ
toxicity, increased cell proliferation, immunotoxicity and endocrine effects. The presence
and toxicological significance of cellular receptors is described.
(c) Reproductive and developmental effects
Effects on reproduction, teratogenicity, fetotoxicity and embryotoxicity are also sum-
marized briefly.
(d) Genetic and related effects
Thsts of genetic and related effects are described in view of the relevance of gene
mutation and chromosomal damage to carcinogenesis (Vainio eta/., 1992).
The adequacy of the reporting of sample characterization is considered and, where
necessary, commented upon; with regard to complex mixtures, such comments are similar to
those described for animal carcinogenicity tests on p. 28. The available data are interpreted
critically by phylogenetic group according to the end-points detected, which may incJude
DNA damage, gene mutation, sister chroJTlatid exchange, micronucleus formation, chromo-
somal aberrations, aneuploidy and celJ 1 msformation. The concentrations employed are
given, and mention is made of whether use of an exogenous metabolic system affected the
test result. These data are given as listings of test systems, data and references; bar graphs
(activity -profiles) and corresponding summary tables with detailed information on the
preparation of the profiles (Waters et a/., 1987) are given in appendices.
Positive results in tests using prokaryotes, lower eukaryotes, plants, insects and cultured
mammalian cells suggest that genetic and related effects could occur in mammals. Results
from such tests may also give information about the types of genetic effect produced and
about the involvement of metabolic activation. Some end-points described are clearly
genetic in nature (e.g., gene mutations and chromosomal aberrations), while others are to a
greater or lesser degree associated with genetic effects (e.g., unscheduled DNA synthesis).
In-vitro tests for tumour-promoting activity and for cell transformation may be sensitive to
changes that are not necessarily the result of genetic alterations but that may have specific
relevance to the process of carcinogenesis. A critical appraisal of these tests has been
published (Montesano et al., 1986).
Genetic or other activity manifest in experimental mammals and humans is regarded as
being of greater relevance than that in other organisms. The demonstration that an agent or
mixture can induce gene and chromosomal mutations in whole mammals indicates that it
may have carcinogenic activity, although this activity may not be detectably expressed in any
or all species. Relative potency in tests for mutagenicity and related effects is not a reliable
indicator of carcinogenic potency. Negative results in tests for mutagenicity in selected
tissues from animals treated in vivo provide less weight, partly because they do not exclude
the possibility of an effect in tissues other than those examined. Moreover, negative results in
short-term tests with genetic end-points cannot be considered to provide evidence to rule out
carcinogenicity of agents or mixtures that act through other mechanisms (e.g., receptor
mediated effects, cellular toxicity with regenerative proliferation, peroxisome proliferation)
(Vainio et al., 1992). Factors that may lead to misleading results in short-term tests have been
discussed in detail elsewhere (Montesano eta!., 1986).
When available, data relevant to mechanisms of carcinogenesis that do not involve
structural changes at the level of the gene are also described.
The adequacy of epidemiological studies of reproductive outcome and genetic and
related effects in humans is evaluated by the same criteria as are applied to epidemiological
studies of cancer.
(e) Structure-activity considerations
This section describes structure-activity relationships that may be relevant to an evalua-
tion of the carcinogenicity of an agent.
In this section, the relevant epidemiological and experimental data are summarized.
Only reports, other than in abstract form, that meet the criteria outlined on p. 21 are
considered for evaluating carcinogenicity. Inadequate studies are generally not summarized:
such studies are usually identified by a square-bracketed comment in the preceding text.
(a) Exposures
Human exposure is summarized on the basis of elements such as production, use,
occurrence in the environment and determinations in human tissues and body fluids.
Quantitative data are given when available.
(b) Carcinogenicity in humans
Results of epidemiological studies that are considered to be pertinent to an assessment
of human carcinogenicity are summarized. When relevant, case reports and correlation
studies are also summarized.
(c) Carcinogenicity in experimental animals
Data relevant to an evaluation of carcinogenicity in animals are summarized. For each
animal species and route of administration, it is stated whether an increased incidence of
neoplasms or preneoplastic lesions was observed, and the tumour sites are indicated. If the
agent or mixture produced tumours after prenatal exposure or in singledose experiments,
this is also indicated. Negative findings are also summarized. Dose-response and other
quantitative data may be given when available.
(d) Other data relevant to an evaluation of carcinogenicity and its mechanisms
Data on biological effects in humans are of particular relevance are summarized.
These may include toxicological, kinetic and metabolic considerations and evidence of DNA
binding, persistence of DNA lesions or genetic damage in exposed .humans. Toxicological
information, such as that on cytotoxicity and regeneration, receptor binding and hormonal
and immunological effects, and data on kinetics and metabolism in experimental animals are
given when considered relevant to the possible mechanism of the carcinogenic action of the
agent. The results of tests for genetic and related effects are summarized for whole mammals,
cultured mammalian cells and nonmammalian systems.
When available, comparisons of such data for humans and for animals, and particularly
animals that have developed cancer, are described.
Structure-activity relationships are mentioned when relevant.
For the agent, mixture or exposure circumstance being evaluated, the available data on
end-points or other phenomena relevant to mechanisms of carcinogenesis from studies in
humans, experimental animals and tissue and cell test systems are summarized within one or
more of the following descriptive dimensions:
(i) Evidence of genotoxicity (i.e., structural changes at the level of the gene): for
example, structure-activity considerations, adduct formation, mutagenicity (effect on speci-
fic genes), chromosomal mutation/aneuploidy
(ii) Evidence of effects on the expression of relevant genes (i.e., functional changes at
the intracellular level): for example, alterations to the structure or quantity of the product of
a proto-oncogene or tumour suppressor gene, alterations to metabolic activation/inacti-
vation/DNA repair
(iii) Evidence of relevant effects on cell behaviour (i.e . morphological or behavioural
changes at the cellular or tissue level): for example, induction of mitogenesis, compensatory
cell proliferation, preneoplasia and hyperplasia, survival of premalignant or malignant cells
(immortalization, immunosuppression), effects on metastatic potential
(iv) Evidence from dose and time relationships of carcinogenic effects and interactions
between agents: for example, early/late stage, as inferred from epidemiological studies;
initiation/promotion/progression/malignant as defined in animal carcinogeni-
city experiments; toxicokinetics
These dimensions are not mutually exclusive, and an agent may fall within more than one
of them. Thus, for example, the action of an agent on the expression of relevant genes could
be summarized under both the first and second dimension, even if it were known with
reasonable certainty that those effects resulted from genotoxicity.
Evaluations of the strength of the evidence for carcinogenicity arising from human and
experimental animal data are made, using standard terms.
It is recognized that the criteria for these evaluations, described below, cannot
encompass all of the factors that may be relevant to an evaluation of carcinogenicity. In
considering all of the relevant data, the Working Group may assign the agent, mixture or
exposure circumstance to a higher or lower category than a strict interpretation of these
criteria would indicate.
(a) Degrees of evidence for carcinogenicity in humans and in experimental animals and
supporting evidence
These categories refer only to the strength of the evidence that an exposure is carcino-
genic and not to the extent of its carcinogenic activity (potency) nor to the mechanisms
involved. A classification may change as new information becomes available.
An evaluation of degree of evidence, whether for a single agent or a mixture, is limited to
the materials tested, as defined physically, chemically or biologically. When the agents
evaluated are considered by the Working Group to be sufficiently closely related, they may
be grouped together for the purpose of a single evaluation of degree of evidence.
(i) Carcinogenicity in humans
The applicability of an evaluation of the carcinogenicity of a mixture, process, occu-
pation or industry on the basis of evidence from epidemiological studies depends on the
variability over time and place of the mixtures, processes, occupations and industries. The
Working Group seeks to identify the specific exposure, process or activity which is considered
most likely to be responsible for any excess risk. The evaluation is focused as narrowly as the
available data on exposure and other aspects permit.
The evidence relevant to carcinogenicity from studies in humans is classified into one of
the following categories:
Sufficient evidence of carcinogenicity: The Working Group considers that a causal
relationship has been established between exposure to the agent, mixture or exposure
circumstance and human canGer. That is, a positive relationship has been observed between
the exposure and cancer in studies in which chance, bias and confounding could be ruled out
with reasonable confidence.
Limited evidence of carcinogenicity: A positive association has been observed between
exposure to the agent, mixture or exposure circumstance and cancer for which a causal
interpretation is considered by the Working Group to be credible, but chance, bias or
confounding could not be ruled out with reasonable confidence.
----------- ------- - - - -- --
Inadequate evidence of carcinogenicity: The available studies are of insufficient quality,
consistency or statistical power to permit a conclusion regarding the presence or absence of a
causal association, or no data on cancer in humans are available.
Evidence suggesting lack of carcinogenicity: There are several adequate studies covering
the full range of levels of exposure that human beings are known to encounter, which are
mutually consistent in not showing a positive association between exposure to the agent,
mixture or exposure circumstance and any studied cancer at any observed level of exposure.
A conclusion of 'evidence suggesting lack of carcinogenicity' is inevitably limited to the
cancer sites, conditions and levels of exposure and length of observation covered by the
available studies. In addition, the possibility of a very small risk at the levels of exposure
studied can never be excluded.
In some instances, the above categories may be used to classify the degree of evidence
related to carcinogenicity in specific organs or tissues.
(ii) Carcinogenicity in experimental animals
The evidence relevant to carcinogenicity in experimental animals is classified into one of
the following categories:
Sufficient evidence of carcinogenicity: The Working Group considers that a causal
relationship has been established between the agent or mixture and an increased incidence of
malignant neoplasms or of an appropriate combination of benign and malignant neoplasms
in (a) two or more species of animals or (b) in two or more independent studies in one species
carried out at different times or in different laboratories or under different protocols.
Exceptionally, a single study in one species might be considered to provide sufficient
evidence of carcinogenicity when malignant neoplasms occur to an unusual degree with
regard to incidence, site, type of tumour or age at onset.
Limited evidence of carcinogenicity: The data suggest a carcinogenic effect but are limited
for making a definitive evaluation because, e.g., (a) the evidence of carcinogenicity is
restricted to a single experiment; or (b) there are unresolved questions regarding the
adequacy of the design, conduct or interpretation of the study; or (c) the agent or mixture
increases the incidence only of benign neoplasms or lesions of uncertain neoplastic potential,
or of certain neoplasms which may occur spontaneously in high incidences in certain strains.
Inadequate evidence of carcinogenicity: The studies cannot be interpreted as showing
either the presence or absence of a carcinogenic effect because of major qualitative or quan-
titative limitations, or no data on cancer in experimental animals are available.
Evidence suggesting lack of carcinogenicity: Adequate studies involving at least two
species are available which show that, within the limits of the tests used, the agent or mixture
is not carcinogenic. A conclusion of evidence suggesting lack of carcinogenicity is inevitably
limited to the species, tumour sites and levels of exposure studied.
(b) Other data relevant to an evaluation of carcinogenicity
Other evidence judged to be relevant to an evaluation of carcinogenicity and of
sufficient importance to affect the overall evaluation is then described. This may include data
on preneoplastic lesions, tumour pathology, genetic and related effects, structure-activity
relationships, metabolism and and physicochemical parameters.
Data relevant to mechanisms of the carcinogenic action are also evaluated. The strength
of the evidence that any carcinogenic effect observed is due to a particular mechanism is
assessed, using terms such as weak, moderate or strong. Then, the Working Group assesses if
that particular mechanism is likely to be operative in humans. The strongest indications that
a particular mechanism operates in humans come from data on humans or biological
specimens obtained from exposed humans. The data may be considered to be especially
relevant if they show that the agent in question has caused changes in exposed humans that
are on the causal pathway to carcinogenesis. Such data may, however, never become
available. because it is at least conceivable that certain compounds may be kept from human
use solely on the basis of evidence of their toxicity and/or carcinogenicity in experimental
For complex exposures, including occupational and industrial exposures, chemical com-
position and the potential contribution of carcinogens known to be present are considered by
the Working Group in its overall evaluation of human carcinogenicity. The Working Group
also determines the extent to which the materials tested in experimental systems are related
to those to which humans are exposed.
(c) Overall evaluation
Finally, the body of evidence is considered as a whole, in order to reach an overall eval-
uation of the carcinogenicity to humans of an agent, mixture or circumstance of exposure.
An evaluation may be made for a group of chemical compounds that have been eval-
uated by the Working Group. In addition, when supporting data indicate that other, related
compounds for which there is no direct evidence of capacity to induce cancer in humans or in
animals may also be carcinogenic, a statement describing the rationale for this conclusion is
added to the evaluation narrative; an additional evaluation may be made for this broader
group of compounds if the strength of the evidence warrants it.
The agent, mixture or exposure circumstance is described according to the wording of
one of the following categories, and the designated group is given. The categorization of an
agent, mixture or exposure circumstance is a matter of scientific judgement, reflecting the
strength of the evidence derived from studies in humans and in experimental animals and
from other relevant data.
Group 1- The agent (mixture) is carcinogenic to humans.
The exposure circumstance entails exposures that are carcinogenic to humans.
This category is used when there is sufficient evidence of carcinogenicity in humans.
Exceptionally, an agent (mixture) may be placed in this category when evidence in humans is
less than sufficient but there is sufficient evidence of carcinogenicity in experimental animals
and strong evidence in exposed humans that the agent (mixture) acts through a relevant
mechanism of carcinogenicity.
Group 2
This category includes agents, mixtures and exposure circumstances for which, at one
extreme, the degree of evidence of carcinogenicity in humans is almost sufficient, as well as
those for which, at the other extreme, there are no human data but for which there is evidence
of carcinogenicity in experimental animals. Agents, mixtures and exposure circumstances are
assigned to either group 2A (probably carcinogenic to humans) or group 2B (possibly
carcinogenic to humans) on the basis of epidemiological and experimental evidence of
carcinogenicity and other relevant data.
Group 2A-The agent (mixture) is probably carcinogenic to humans.
The exposure circumstance entails exposures that are probably carcinogenic to humans.
This category is used when there is limited evidence of carcinogenicity in humans and
sufficient evidence of carcinogenicity in experimental animals. In some cases, an agent
(mixture) may be classified in this category when there is inadequate evidence of carcino-
genicity in humans and sufficient evidence of carcinogenicity in experimental animals and
strong evidence that the carcinogenesis is mediated by a mechanism that also operates in
humans. Exceptionally, an agent, mixture or exposure circumstance may be classified in this
category solely on the basis of limited evidence of carcinogenicity in humans.
Group 2B-The agent (mixture) is possibly carcinogenic to humans.
The exposure circumstance entails exposures that are possibly carcinogenic to humans.
This category is used for agents, mixtures and exposure circumstances for which there is
limited evidence of carcinogenicity in humans and less than sufficient evidence of carcino-
genicity in experimental animals. It may also be used when there is inadequate evidence of
carcinogenicity in humans but there is sufficient evidence of carcinogenicity in experimental
animals. In some instances, an agent, mixture or exposure circumstance for which there is
inadequate evidence of carcinogenicity in humans but limited evidence of carcinogenicity in
experimental animals together with supporting evidence from other relevant data may be
placed in this group.
Group 3-The agent (mixture or exposure circumstance) is not classifiable as to its carcino-
genicity to humans.
This category is used most commonly for agents, mixtures and exposure circumstances
for which the evidence of carcinogenicity is inadequate in humans and inadequate or limited
in experimental animals.
Exceptionally, agents (mixtures) for which the evidence of carcinogenicity is inadequate
in humans but sufficient in experimental animals may be placed in this category when there is
strong evidence that the mechanism of carcinogenicity in experimental animals does not
operate in humans.
Agents, mixtures and exposure circumstances that do not fall into any other group are
also placed in this category.
Group 4-The agent (mixture) is probably not carcinogenic to humans.
This category is used for agents or mixtures for which there is evidence suggesting lack of
carcinogenicity in humans and in experimental animals. In some instances, agents or mixtures
for which there is inadequate evidence of carcinogenicity in humans but evidence suggesting
lack of carcinogenicity in experimental animals, consistently and strongly supported by a
broad range of other relevant data, may be classified in this group.
Breslow, N.E. & Day, N.E. (1980) Statistical Methods in Cancer Research, Vol. 1. The Analysis of Case-
control Studies (!ARC Scientific Publications No. 32), Lyon, IARC
Breslow, N.E. & Day, N.E. (1987) Statistical Methods in Cancer Research, Vol. 2, The Design and
Analysis of Cohort Studies (IARC Scientific Publications No. 82), Lyon, !ARC
Gart, J.J., Krewski, D., Lee, P.N., Tarone, R.E. & Wahrendorf, J. (1986) Statistical Methods in Cancer
Research, Vol. 3, The Design and Analysis of Long-term Animal Expenments (IARC Scientific
Publications No. 79), Lyon, !ARC
Hoe!, D.G., Kaplan, N.L. & Anderson, M.W. (1983) Implication of nonlinear kinetics on risk
estimation in carcinogenesis. Science, 219, 1032-1037
Huff, J.E., Eustis, S.L. & Haseman, J.K. (1989) Occurrence and relevance of chemically induced
benign neoplasms in long-term carcinogenicity studies. Cancer Metastasis Rev., 8, 1-21
!ARC (1973-1990) Information Bulletin on the Survey of Chemicals Being Tested for Carcinogenicity/-
Directory of Agents Being Tested for Carcinogenicity, Numbers 1- 14, Lyon
Number 1 (1973) 52 pages
Number 2 (1973) 77 pages
Number 3 (1974) 67 pages
Number 4 (1974) 97 pages
Number 5 (1975) 88 pages
Number 6 (1976} 360 pages
Number 7 (1978) 460 pages
Number 8 (1979) 604 pages
Number 9 (1981) 294 pages
Number 10 (1983) 326 pages
Number 11 (1984) 370 pages
Number 12 (1986) 385 pages
Number 13 (1988} 404 pages
Number 14 (1990) 369 pages
!ARC (1976-1991)
Directory of On-going Research in Cancer Epidemiology 1976. Edited by C.S. Muir & G. Wagner,
Directory of On-going Research in Cancer Epidemiology 1977 (IARC Scientific Publications
No. 17). Edited by C.S. Muir & G. Wagner, Lyon
Directory of On-going Research in Cancer Epidemiology 1978 (IARC Scientific Publications
No. 26). Edited by C.S. Muir & G. Wagner, Lyon
Directory of On-going Research in Cancer Epidemiology 1979 (IARC Scientific Publications
No. 28). Edited by C.S. Muir & G. Wagner, Lyon
Directory of On-going Research in Cancer Epidemiology 1980 (IARC Scientific Publications
No. 35). Edited by C.S. Muir & G. Wagner, Lyon
Directory of On-going Research in Cancer Epidemiology 1981 (!ARC Scientific Publications
No. 38). Edited by C.S. Muir & G. Wagner, Lyon
Directory of On-going Research in Cancer Epidemiology 1982 (!ARC Scientific Publications
No. 46). Edited by C.S. Muir & G. Wagner, Lyon
Directory of On-going Research in Cancer Epidemiology 1983 (IARC Scientific Publications
No. 50). Edited by C.S. Muir & G. Wagner, Lyon
Directory of On-going Research in Cancer Epidemiology 1984 (IARC Scientific Publications
No. 62). Edited by C.S. Muir & G. Wagner, Lyon
Directory of On-going Research in Cancer Epidemiology 1985 (IARC Scientific Publications
No. 69). Edited by C.S. Muir & G. Wagner, Lyon
Directory of On-going Research in Cancer Epidemiology 1986 (IARC Scientific Publications
No. 80). Edited by C.S. Muir & G. Wagner, Lyon
Directory of On-going Research in Cancer Epidemiology 1987 (IARC Scientific Publications
No. 86). Edited by D.M. Parkin & J. Wahrendorf, Lyon
Directory of On-going Research in Cancer Epidemiology 1988 (IARC Scientific Publications
No. 93). Edited by M. Coleman & J. Wahrendorf, Lyon
Directory of On-going Research in Cancer Epidemiology 1989/90 (IARC Scientific Publications
No. 101). Edited by M. Coleman & J. Wahrendorf, Lyon
Directory of On-going Research in Cancer Epidemiology 1991 (IARC Scientific Publications
No. 110). Edited by M. Coleman & J. Wahrendorf, Lyon
IARC (1977) !ARC Monographs Programme on the Evaluation of the Carcinogenic Risk of Chemicals to
Humans. Preamble (IARC intern. tech. Rep. No. 77/002), Lyon
IARC (1978) Chemicals with Sufficient Evidence of Carcinogenicity in Experimental Animals-IARC
Monographs Volumes 1-17 (IARC intern. tech. Rep. No. 78/003), Lyon
IARC (1978-1991) Environmental Carcinogens. Methods of Analysis and Exposure Measurement:
Vol. 1. Analysis ofVolatile Nitrosamines in Food (IARC Scientific Publications No. 18). Edited by
R. Preussmann, M. Castegnaro, E.A. Walker & A.E. Wasserman (1978)
Vol. 2. Methods for the Measurement of Vinyl Chloride in Poly(vinyl chloride), Air; Water and
Foodstuffs (IARC Scientific Publications No. 22). Edited by D.C.M. Squirrell & W. Thain
Vol. 3. Analysis of Polycyclic Aromatic Hydrocarbons in Environmental Samples (IARC Scientific
Publications No. 29). Edited by M. Castegnaro, P. Bogovski, H. Kunte & E.A. Walker(1979)
Vol. 4. Some Aromatic Amines and Azo Dyes in the General and Industrial Environment (!ARC
Scientific Publications No. 40). Edited by L. Fishbein, M. Castegnaro, I.K. O'Neill & H.
Bartsch (1981)
Vol. 5. Some Mycoto.xins (IARC Scientific Publications No. 44). Edited by L. Stoloff, M.
Castegnaro, P. Scott, I.K. O'Neill & H. Bartsch (1983)
Vol. 6. N-Nitroso Compounds (IARC Scientific Publications No. 45). Edited by R. Preussmann,
I.K. O'Neill, G. Eisenbrand, B. Spiegelhalder & H. Bartsch (1983)
Vol. 7. Some Volatile Halogenated Hydrocarbons (IARC Scientific Publications No. 68). Edited by
L. Fishbein & I.K. O'Neill (1985)
Vol. 8. Some Metals: As, Be, Cd, Cr. Ni, Pb, Se, Zn (IARC Scientific Publications No. 71). Edited
by I.K. O'Neill, P. Schuller & L. Fishbein (1986)
Vol. 9. Passive Smoking (IARC Scientific Publications No. 81). Edited by I.K. O'Neill, K.D.
Brunnemann, B. Dodet & D. Hoffmann (1987)
Vol. 10. Benzene and Alkylated Benzenes (IARC Scientific Publications No. 85). Edited by L.
Fishbein & I.K. O'Neill (1988)
Vol. 11. Polychlorinated Dioxins and Dibenzofurans (IARC Scientific Publications No. 108) ..
Edited by C. Rappe, H.R. Buser, B. Dodet & I.K. O'Neill (1991)
IARC (1979) Criteria to Select Chemicals for IARC Monographs (IARC intern. tech. Rep. No. 79/003),
IARC (1982) /ARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans,
Supplement 4, Chemicals, Industrial Processes and Industries Associated with Cancer in Humans
(/ARC Monographs, Volumes 1 to 29), Lyon
IARC (1983) Approaches to Classifying Chemical Carcinogens According to Mechanism of Action
(!ARC intern. tech. Rep. No. 83/001), Lyon
IARC (1984) Chemicals and Exposures to Complex Mixtures Recommended for Evaluation in IARC
Monographs and Chemicals and Complex Mixtures Recommended for Long-term Carcinogenicity
Testing (IARC intern. tech. Rep. No. 84/002), Lyon
IARC (1987a) !ARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Supplement 6,
Genetic and Related Effects: An Updating of Selected !ARC Monographs from Volumes 1 to 42,
IARC (1987b) /ARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Supplement 7,
Overall Evaluations of Carcinogenicity: An Updating ofiARC Monographs Volumes 1 to 42, Lyon
!ARC (1988) Report of an /ARC Working Group to Review the Approaches and Processes Used to
Evaluate the Carcinogenicity of Mixtures and Groups of Chemicals (IARC intern. tech. Rep.
No. 88/002), Lyon
!ARC (1989) Chemicals, Groups of Chemicals, Mixtures and Exposure Circumstances to be Evaluated in
Future /ARC Monographs, Report of an ad hoc Working Group (!ARC intern. tech. Rep. No.
89/004), Lyon
IARC (1991a) A Consensus Report of an IARC Monographs Working Group on the Use of Mechanims
of Carcinogenesis in Risk Identification (!ARC intern. tech. Rep. No. 91/002), Lyon
!ARC (199lb) Report of an Ad-hoc !ARC Monographs Advisory Group on Viruses and Other Biological
Agents Such as Parasites (IARC intern. tech. Rep. No. 911001), Lyon
Montesano, R., Bartsch, H., Yainio, H., Wilbourn, J. & Yamasaki, H., eds (1986) Long-term and
Slwrt-term Assays for Carcinogenesis-A Critical Appraisal (IARC Scientific Publications No. 83),
Lyon, !ARC .
Peto, R., Pike, M.C., Day, N.E., Gray, R.G., Lee, P.N., Parish, S., Peto, J., Richards, S. & Wahrendorf,J.
(1980) Guidelines for simple, sensitive significance tests for carcinogenic effects in long-term
animal experiments. In: /ARC Monographs on the Evaluation of the Carcinogenic Risk of Chem-
icals to Humans, Supplement 2, Long-term and Short-term Screening Assays for Carcinogens: A
Critical Appraisal, Lyon, pp. 311-426
Vainio, H., Magee, P., McGregor, D. & McMichael, A., eds (1992) Mechanisms of Carcinogenesis in
Risk Identification (IARC Scientific Publications No. 116), Lyon, IARC
Waters, M.D., Stack, H.F., Brady, A.L., Lohman, P.H.M., Haroun, L. & Yainio, H. (1987) Apendix 1.
Activity profiles for genetic and related tests. In: !ARC Monographs on the Evaluation of
Carcinogenic Risks to Humans, Suppl. 6, Genetic and Related Effects: An Updating of Selected
!ARC Monographs from Volumes 1 to. 42, Lyon, !ARC, pp. 687-696
Wilbourn, J., Haroun, L., Heseltine, E., Kaldor, J., Partensky, C. & Yainio, H. (1986) Response of
experimental animals to human carcinogens: an analysis based upon the IARC Monographs
Programme. Carcinogenesis, 7, 1853- 1863
---------- ----- ---- -- - - -- - -
This fifty-fifth volume of /ARC Monographs contains evaluations of carcinogenic risks
associated with human exposure to solar and ultraviolet (UV) radiation from medical and
cosmetic devices, general illumination and industrial sources. Ultraviolet radiation (UVR)
was considered previously (IARC, 1986) in a volume in which furocoumarins were evaluated.
Since some of these compounds are used clinically in conjunction with ultraviolet A (UVA)
radiation, information on the carcinogenic effects of UVR alone was provided in an
appendix; however, no evaluation was made at that time.
Solar radiation is largely optical radiation (UV. visible and infrared), although both
shorter wavelength (ionizing) and longer wavelength (microwaves and radiofrequency)
radiation is present. UVR lies in the interval 100-400 nm and is further subdivided into UVA
(315-400 nm), UVB (280-315 nm) and UVC (100-280 nm). The UV component of
terrestrial radiation from the sun comprises about 95% UVA and 5% UVB; UVC is removed
from extraterrestrial radiation by stratopheric ozone. Before the beginning of this century,
the sun was essentially the only source of UVR; with the advent of artificial sources, the
opportunity for additional exposure, not only to UVA and UVB but also to UVC, has
increased. It should be stressed that the distinction of UVR into UVA, UVB and UVC ranges
has no biological basis, and the potential of UVR for causing damage to biomolecules, cells,
tissues and organisms varies enormously over the spectral region from 250 to 400 nm.
UVA raqiation is one of the components of solar emissions and of emissions from
medical lamps and lamps used for cosmetic purposes. UVB radiation is present in solar
emissions, from lamps used in medicine and for cosmetic purposes and in certain lamps used
for general illumination, such as unshielded fluorescent and tungsten-halogen lamps. It
causes sunburn relatively easily and is immunosuppressive; it can cause ocular cataracts. The
possibility that the UVB component of solar radiation will increase as a result of depletion of
the ozone layer is a matter of concern. This question was not addressed in the present
Human exposure to UVC radiation is uncommon and is related to the use of germicidal
and tungsten-halogen lamps, phototherapy and welding arcs. Thus, very little is known about
the effects ofUVC on humans, although a great deal of information is available on the effects
of radiation in this range on biomolecules, cells and viruses.
In the USA. skin melanoma has been second only to lung cancer in its rate of increase in
incidence over the last 40 years: the incidence has been increasing by about 5% per year. The
major sites have been male trunk and female leg. Mortality from melanoma may now be
falling in younger generations (at least in the USA) due, possibly, to changes in sun exposure
(Scotto et al., 1991). There is also evidence that the incidence of nonmelanocytic skin cancer
is increasing in some white-skinned populations (Gallagher et al., 1990). Constitutional risk
factors, e.g., skin type, hair and eye colour and specific subtypes of exposure (for example,
occupational and recreational), have been assessed in individual studies or sections of the
monographs but have not been included in the evaluations.
UYR is ubiquitous and cannot be totally avoided. An appendix to this volume presents a
discussion on the use of topical sunscreens, taking into consideration both potentially
beneficial, protective effects and possible adverse reactions. The biological effects of
combinations of psoralens and UYR were not considered since these were the subjects of
separate monographs (IARC, 1980, 1986, 1987) in the !ARC Monographs series.
Gallagher, R.P., Ma, B., McLean, DJ., Yang, C.P., Ho, V., Carruthers, J.A. & Warshawski, L.M. (1990)
Trends in basal cell carcinoma, squamous cell carcinoma, and melanoma of the skin from 1973
through 1987. J. Am. Acad. Dermatol., 23, 413-421
IARC (1980) !ARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans,
Vol. 24, Some Pharmaceutical D r u ~ Lyon, pp. 101-124
!ARC (1986) /ARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans,
Vol. 40, Some Naturally Occurring Synthetic Food Components, Furocoumarins and Ultraviolet
Radiation, Lyon, pp. 317-371
IARC (1987) /ARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Suppl. 7, Overall
Evaluations of Carcinogenicity: An Updating of IARC Monographs Volumes 1 to 42, Lyon,
pp. 242-245
Scotto, J ., Pitcher, H. & Lee, J.A.H. (1991) Indications of future decreasing trends in skin-melanoma
mortality among whites in the United States. Int. 1. Cancer, 49, 490-497
1.1 Nomenclature
1.1.1 Optical radiation
1. Exposure Data
Optical radiation is radiant energy within a broad region of the electromagnetic
spectrum that includes ultraviolet (UV), visible (light) and infrared radiation. Ultraviolet
radiation (UVR) is characterized by wavelengths between 10 and 400 nm- bordered on the
one side by x rays and on the other by visible light (Fig. 1 ). Solar .radiation is largely optical
radiation, although ionizing radiation (i.e. , cosmic rays, gamma rays and x rays, which have
wavelengths less than approximately 10 nm) and radio-frequency radiation (i.e., wavelengths
greater than 1 mm: microwaves and longer radio waves) are also present in the spectrum.
The optical radiation spectrum is generally considered to fall between 10 nm and 1 mm,
and several different conventions have been developed to describe different bands within
this spectrum. It is important to recognize that no single convention is uniquely 'correct' but
that each may be useful for a particular branch of science and technology. For example, in
optics, it is convenient to separate the spectrum into different bands on the basis of the
transmission and absorption properties of optical materials (e.g., glass and quartz). In one
optical convention, shown in Figure 1, UVR is divided into vacuum UV, extending from 10 to
180 nm; middle UV, from 180 nm to 300 nm; and near UV, from 300 nm to 380 or 400 nm.
Meteorological scientists typically define optical spectral regions on the basis of atmospheric
windows. Some spectral designations are based on uses, e.g., 'germicidal' and 'black-light'
For the purposes of this monograph, the photobiological designations of the Com-
~ s s o n Internationale de l' Eclairage (CIE, International Commission on Illumination) are
the most relevant and are used throughout to define the approximate spectral regions in
which certain biological absorption properties and biological interaction mechanisms may
dominate (Commission Internationale de l'Eclairage, 1987). The CIE bands are: UVC
(100-280 nm), UVB (280-315 nm) and UVA (315-400 nm). Visible light is the region
between 400 nm and 780 nm.
It is important to recognize that these spectral band designations are merely short-hand
notations and cannot be considered to designate fine dividing lines below which an effect is
present and above which it does not occur. The reader should also be alerted to the fact that
the CIE nomenclature is not always followed rigorously and that some authors introduce
slight variations; for example, distinguishing b.etween UVB and UVA at 320 rather than
315 nm (frequently used in the USA) and defining UVC as200-280 nm(Moseley, 1988). The
German Industrial Standard (DIN 5031) defines UVA as radiation between 315 and 380 nm
(Mutzhas, 1986).
Figure 1. Electromagnetic spectrum with enlargement of ultraviolet (UV) region
Violet Red
Cosmic rays, gamma rays, x rays
Infrared Radio waves
X rays -.J
100 180
Vacuum UV
Extreme UV FarUV
280 315
Middle uv Near UV
I I ,
10 100 180
Wavelength (nanometres)
Adapted from WHO (1979), Morison (1983a), Sylvania (undated)
300 380
From the viewpoint of photochemistry and photobiology, interactions of optical
radiation with matter are considered to occur when one photon interacts with one molecule
to produce a photochemically altered molecule or two dissociated molecules (Phillips, 1983;
Smith, 1989). In any photochemical interaction, the energy of the individual photon is
important, since this must be sufficient to alter a molecular bond. The photon energy is
generally expressed in terms of electron volts (eV). A wavelength of 10 nm corresponds to a
photon energy of 124 eV, and 400 nm to an energy of 3.1 eV (WHO, 1979). The number of
altered molecules produced relative to the number of absorbed photons is referred to as the
'quantum yield' (Phillips, 1983). The efficacy of photochemical interaction per incident
quantum and the photobiological effects per unit radiant exposure typically vary widely with
wavelength. A quantitative plot of such spectral variation, usualJy normalized to unity at the
most effective wavelength, is referred to as an 'action spectrum' (Jagger, 1985).
- - --- - - - - -
1.1.2 Quantities and units
1\vo systems of quantities and units are used to describe the characteristics of light and
light sources: the radiometric and the photometric systems. Radiometry can be applied to all
optical sources and to all exposures to optical radiation (including solar radiation and UVR).
Photometry can be used only to describe visible light sources, and photometric quantities are
used in illumination engineering. The basic photometric unit is the lumen, which is defined in
terms ofthe spectral response of the human eye (specifically, the spectral response ofthe CIE
'standard observer'), i.e., the action spectrum of vision, which is initially a photochemical
process. It is important to recognize that radiometric quantities and units are absolute, while
photometric quantities and units are related to standardized human perception; the
relationship between the two sets of units varies significantly with the spectrum of radiation.
The effects of optical radiation (including light), other than vision, must therefore be
measured and quantified in terms of radiometric units and spectral characteristics rather
than photometric units. This is particularly important in relation to the photobiological
effects of UVR. Most lamps used for illumination are rated by manufacturers only in
photometric terms (e.g., lumen output) and not in terms of UVR emission (Phillips, 1983).
The most important radiometric quantities and units commonly used to describe optical
radiation are given in Table 1. Certain terms are used primarily to describe source charac-
teristics, e.g., radiance, radiant intensity; whereas other terms are generally used to describe
exposure (irradiance, radiant exposure). The term 'spectral' placed before any of the quan-
tities implies restriction to a unit wavelength band, e.g., spectral irradiance (watts per square
metre per nanometre) (Moseley, 1988). For a more detailed discussion of these parameters,
see various standard textbooks on radiometry, such as Boyd (1983).
The quantities of radiometry are expressed in terms of absolute energy (Jagger, 1985).
Radiant intensity is the power emitted per unit solid angle of a source. Radiance is the radiant
intensity per unit area of source. Thus, a fluorescent lamp does not have very high radiance in
comparison to the filament of a flashlight bulb, even though it has a high radiant power
output. The radiometric term expressed in units of watts per square metre (dose rate) is
irradiance, which is also the power striking a unit area of surface.
The energy of UVR falling on a unit surface area of an object was defined in 1954 by the
First International Congress of Photobiology as the 'dose'; it has also been referred to as
'exposure dose'. The equivalent radiometric quantity is radiant exposure, expressed in joules
per square centimetre or per square metre. Radiant exposure has been referred to as 'energy
fluence' in some texts; however, fluence is a radiometric quantity, with the same units as
radiant exposure, but referring to energy arriving at a plane of unit area from all directions,
including backscatter. Thus, fluence is quite correctly of value in describing an exposure dose
at a depth inside tissue; it has, however, seldom been calculated in photobiological studies of
the effects of UVR, in which the radiant exposure incident upon the skin is normally
measured. Radiant exposure is the amount of energy crossing a unit area of space normal to
the direction of propagation of a beam of UVR. If the radiant energy arrives from many
directions, as from the sky, then the fluence at one point is the sum of all the component
fluences entering a unit sphere of space. The energy fluence rate is the power that crosses a
unit area normal to the direction of propagation, or the energy per unit area per unit time
Table 1. Some basic terminology used to quantify optical radiation
Term International Definition SI unit Synonyms and comments
Wavelength >.. nm Nanometre = tQ-
m (also called
millimicron, mJJ.)
Radiant energy
Oe l: (Pe X dt) J Joule; 1 joule = 1 watt x second;
total energy contained in a radiation
field or total energy delivered to a
given receiver by such a radiation
Radiant flux Pe dQe/dt w Watt; rate of delivery of radiant
energy ('radiant power'); also
expressed as q,
Ee dPe/dA
Wfm2 Radiant flux arriving over a given
area ('fluence rate', ' dose rate',
'intensity', 'radiant incidence'). In
photobiology, has also been
expressed in W/cm2, mW/cm2 and
Radiant intensity Ie dPeld.n W/sr Watt/steradian; radiant flux emitted
by source into a given solid angle
(solid angle expressed in steradians)
Radiance dPeldA X d.n Wfm2 x sr Watt/m 2 X steradian; radiant flux
per unit solid angle per unit area
emitted by an extended source
Radiant exposure He Ee X t
Jt m2 Radiant energy delivered to a given
area (' fluence', 'exposure dose' ,
'dose'); t = time in seconds. Has
also been expressed as J/cm2,
m.J/cm2 and J.l]/cm2
Adapted from WHO (1979), Boyd (1983), Jagger (1985), Hoffman (1987) and Weast (1989)
/s or W/m
) . The terms dose (J/m
) and dose rate (W/m
) pertain to the energy and
power, respectively, striking a unit surface area of an irradiated object (Jagger, 1985).
In terms of visible light perceived by humans, the photometric analogue of the radiance
of a source is luminance (brightness), and irradiance is illuminance (measured in 'lux' or
lumen per square metre). In photometry, the lumen is the unit of luminous power (Jagger,
1.1.3 Units of biologically effective ultraviolet radiation
In addition to general radiometric quantities, specialized quantities of effective irra-
diance relative to a specified photochemical action spectrum are used in photochemistry and
photobiology. Effective radiant exposures to produce erythema (Jagger, 1985) or photo-
keratitis are examples. Effective irradiance or radiant exposure is not limited to photo-
biology, and a similar approach has been used to quantify the photocuring of inks, in photo-
polymerization (Phillips, 1983) and in assessing the hazards of UVR. In order to weight a
source spectrally, the general formula involves an action spectrum and a spectral radiometric
quantity. The effective irradiance of a given photobiological process is defined as:
I EA X SA X fl.l
expressed in W/m
, where Ex. is the spectral irradiance (W/m
x nm) at wavelength A. (nm)
and Ax. is the wavelength interval 0\.
) used in the summation (in nm). Sx. is a measure of
the effectiveness of radiation of wavelength A. (nm), relative to some reference wavelength, in
producing a particular biological end-point. As it is a ratio, Sx. has no units {American
Conference of Governmental Industrial Hygienists, 1991).
Effective irradiance is equivalent to a hypothetical irradiance of monochromatic radia-
tion with a wavelength at which Sx. is equal to unity. The time integral of effective irradiance is
the effective radiant exposure (also called the ' effective dose').
A unit of effective dose commonly used in cutaneous photobiology is the 'minimal
erythema dose' (MED). One MED has been defined as the lowest radiant exposure to UVR
that is sufficient to produce erythema with sharp margins 24 h after exposure (Morison,
1983a). Another end-point often used in cutaneous photobiology is a just-perceptible
reddening of exposed skin; the dose of UVR necessary to produce this 'minimal perceptible
erythema' is sometimes also referred to as an MED. In unacclimatized, white-skinned
populations, there is an approximately four-fold range in the MED of exposure to UVB
radiation (Diffey & Fdrr, 1989). When the term MED is used as a unit of exposure dose,
however, a representative value is chosen for sun-sensitive individuals. If, in the above
expression for effective irradiance, Sx. is chosen as the reference action spectrum for
erythema (McKinlay & Diffey, 1987) and a value of200 11m
at wavelengths for which Sx. is
equal to unity is assumed for the MED, the dose (expressed in MED) received after an
exposure period oft seconds is
t X I: E.t X Sl X fl.J200.
Notwithstanding the difficulties of interpreting accurately the magnitude of such an
imprecise unit as the MED, it has the advantage over radiometric units of being related to the
biological consequences of the exposure.
1.2 Methods for measuring ultraviolet radiation
UVR can be measured by chemical or physical detectors, often in conjunction with a
monochromator or band-pass filter for wavelength selection. Physical detectors include
radiometric devices, which depend for their response on the heating effect of the radiation,
and photoelectric devices, in which incident photons are detected by a quantum effect such as
the production of electrons. Chemical detectors include photographic emulsions,
actinometric solutions and UV-sensitive plastic films.
1.2.1 Spectroradiometry
The fundamental way of characterizing a source of UVR is on the basis of its spectral
power distribution in a graph (or table) which indicates the radiated power as a function of
wavelength. The data are obtained by a technique known as spectroradiometry. Spectral
measurements are often not required as ends in themselves but are used to calculate
biologically weighted radiometric quantities. A spectroradiometer comprises three essential
components (Gibson & Diffey, 1989):
(i) input optics, such as an integrating sphere or Teflon diffuser, which collects the
incident radiation and conducts it to
(ii) the entrance slit of a monochromator, which disperses the radiation by means of
one or two wavelength dispersive devices (either diffraction grating or prism). The
monochromator also incorporates mirrors to guide the radiation from the entrance
slit to the dispersion device and on to the exit slit, where it is incident on
(iii) a radiation detector, normally a photodiode or, for higher sensitivity, a
photomultiplier tube.
Spectroradiometry is generally considered to be the best way of specifying UV sources,
although the accuracy of spectroradiometry, particularly with respect to the UVB waveband
of terrestrial radiation, is affected by a number of parameters including wavelength cali-
bration, band width, stray radiation, polarization, angular dependence, linearity and
calibration sources. It is therefore essential to employ a double monochromator for accurate
characterization of terrestrial UVR and particularly UVB (Garrison et al., 1978; Kostkowski
eta/., 1982; Gardiner & Kirsch, 1991).
1.2.2 Wavelength-independent (thermal) detectors
General-purpose radiometers incorporate detectors that have a flat response over a
wide range of wavelengths. Such thermal detectors operate on the principle that incident
radiation is absorbed by a receiving element, and the temperature rise of the element is
measured, usually by a thermopile or a pyroelectric detector. A thermopile, which comprises
several thermocouples connected in series for improved sensitivity, must have a window
made of fused silica for measuring UVR at wavelengths down to at least 250 nm. Pyroelectric
detectors rely on a voltage generated by temperature changes in a lithium tantalate crystal.
Thermal detectors are normally used to measure the total radiant power of a source rather
than just the UV component (Moseley, 1988).
Instruments for measuring broad-band solar radiation fall into three categories: pyro-
heliometers, pyranometers and pyranometers with a shading device (Iqbal, 1983). These
types of instrument find their applications in meteorology rather than in UV photobiology.
1.2.3 Wavelength-dependent detectors
Detectors of this type have a spectral response that varies widely depending on the types
of detector and filters that may be incorporated. Detectors can be designed to have a spectral
response that matches a particular action spectrum for a photobiological end-point. The
success with which this is achieved is variable. The most widely used device, particularly for
measuring solar UVR, has been the Robertson-Berger meter (Robertson, 1972; Berger,
1976), which incorporates optical filters, a phosphor and a vacuum phototube or photo-
voltaic cell. This device measures wavelengths of less than 330 nm in the global spectrum with
a spectral response that rises sharply with decreasing wavelength. It has been used to monitor
natural UVR continuously at several sites throughout the world (Berger & Urbach, 1982;
Diffey, 1987a).
---------------- - - -- --
Detectors incorporating a photodiode or vacuum photocell in conjunction with optical
filter(s) and suitable input optics (e.g., a quartz hemispherical detector) have been produced
to match a number of different action spectra. One such detector is the International Light
Model -730 UV Radiometer, which has a spectral response close to the action spectrum
designated by the American Conference of Governmental Industrial Hygienists for eval-
uating the hazard to health of exposure to UVR, and has been used to measure irradiance
over different terrains (Sliney, 1986).
Wavelength-dependent detectors with spectral responses largely in the UVAwaveband
are used, for example, in measuring the output of irradiation units for the treatment of
psoriasis by psoralen photochemotherapy (Morison, 1983a).
A different yet complementary approach is the use of various photosensitive films as UV
dosimeters. The principle is to relate the degree of deterioration of the films, usually in terms
of changes in their optical properties, to the dose of incident UVR. The principal advantages
of the film dosimeter are that it provides a simple means of integrating exposure continuously
and allows simultaneous comparison of numerous sites that are inaccessible to bulky,
expensive instruments (Diffey, 1987a). The most widely used photosensitive film is polymer
polysulfone (Diffey, 1989a). Personal dosimeters of polysulfone film have been developed
and used in a number of dosimetric studies (Challoner et al., 1976, 1978; Leach eta/., 1978;
Holman et al., 1983a; Larko & Diffey, 1983; Diffey, 1987a; Schothorst et al., 1987a; Slaper,
1987; Rosenthal et al., 1990).
It is difficult to achieve a prescribed UVR spectral response with wavelength-dependent
detectors. Accurate results can be achieved only if the detectors are calibrated against the
appropriate source spectrum using a spectroradiometer (Gibson & Diffey, 1989). Unless this
is done, severe dosimetric errors can arise, particularly with measurements of solar UVR
(Diffey, 1987a; Sayre & Kligman, 1992).
Accurate measurement of UVB radiation is far more difficult than would appear
initially. The primary problem is that the UVB produced by most optical sources- the sun as
. well as incandescent and fluorescent lamps used for illumination-is only a very small
fraction (i.e., less than 0.3%) of the total radiant energy emitted. Additionally, biological
action spectra (e.g., for erythema and photokeratitis) typically decrease dramatically within
the same waveband in which the source spectrum increases (Diffey & Farr, 199la). This
means that either a spectroradiometer or a direct-reading filtered 'erythemal' or 'hazard'
meter must reject out-of-band radiant energy to better than one part in 10
or even 10
spectral band-width of a monochromator can also greatly affect measurement error: too
large a band-width can reduce the steepness of reported action spectra.
1.3 Sources and exposures
In the broadest sense, UVR may be produced when a body is heated (incandescence) or
when electrons that have been raised to an excited state return to a lower energy level, as
occurs in fluorescence, in an electric discharge in a gas and in electric arcs (optical plasma)
(Sliney & Wolbarsht, 1980; Phillips, 1983; Moseley, 1988). The characteristics of exposures
to both terrestrial solar radiation (an incandescent source) and artificial light sources are
discussed in the following sections.
1.3.1 Solar ultraviolet radiation
Optical radiation from the sun is modified significantly as it passes through the Earth's
atmosphere (Fig. 2), a1though about two-thirds of the energy from the sun that impinges on
the atmosphere penetrates to ground level. The annual variation in extra-terrestrial
radiation is less than 10%, but the variation in the modifying effect of the atmosphere is far
greater (Moseley, 1988). Measurements corrected for atmospheric absorption show that the
visible portion comprises approximately 40% of the total radiation received at the surface of
the Earth. While UVR comprises only a small proportion of the total radiation (approxi-
mately 5% ), this component is extremely important in various biological processes. The
principal effect of infrared radiation is to warm the earth; approximately 55% of the solar
radiation received at the surface of the earth is infrared (Foukal. 1990).
Fig. 2. p ~ t r a l irradiance from the sun outside the Earth's atmosphere (upper curve) and at
sea level (lower curve)
)( 0 2
.. E
g 01
. !::
From Moseley (1988}
1000 2000
Wavelength (nm)
On its path through the atmosphere, solar radiation is absorbed and scattered by various
constituents of the atmosphere. It is scattered by air molecules. particularly oxygen and
nitrogen (Rayleigh scattering), which produce the blue colour of the sky. It is also scattered
by aerosol and dust particles (Mie scattering) and is scattered and absorbed by atmospheric
pollution. Total solar irradiance and the relative contributions of different wavelengths vary
with altitude. Clouds attenuate solar radiation, although their effect on infrared radiation is
greater than on UVR. Reflection of sunlight from certain ground surfaces may contribute
significantly to the total amount of scattered UVR. An effective absorber of solar UVR is
ozone in the stratosphere (Moseley, 1988). An equally important absorber in the longer
wavelengths (infrared) is water vapour (Diffey, 1991); a secondary absorber in this range is
carbon dioxide. These two filter out much of the solar energy with wavelengths longer than
1000 nm (Sliney & Wolbarsht, 1980).
The quality (spectral distribution) and quantity (total UV irradiance) of UVR reaching
the Earth's surface depend on the radiated power from the sun and the transmitting
properties of the atmosphere. Although UVC exists in the extra-terrestrial solar spectrum, it
is filtered out completely by the ozone layer in the atmosphere. UVB radiation, which
represents about 5% of the total solar UVR that reaches the Earth (Sliney & Wolbarsht,
1980), has been considered to be the most biologically significant part of the terrestrial UV
spectrum. The levels of UVB radiation reaching the surface of the Earth, although heavily
attenuated, are also largely controlled by the ozone layer.
Ozone (03) is a gas which comprises approximately one molecule out of every two
million in the atmosphere. It is created by the reaction of molecular oxygen (0
) with atomic
oxygen (0), formed by the dissociation of 0
by short-wavelength UVR ( < 242 nm) in the
stratosphere at altitudes between about 25 and 100 km. Absorption of UVR at wavelengths
up to about 320 nm converts the ozone back to 0
and 0, and it is this dissociation of ozone
that is responsible for preventing radiation at wavelengths less than about 290 nm from
reaching the Earth's surface (Moseley, 1988; Diffey, 1991). Molina and Rowland (1974) first
proposed that chlorofluorocarbons and other gases released by human activity could alter
the natural balance of creative and destructive processes and lead to depletion of the stra-
tospheric ozone layer. Substantial reductions, of up to 50%, in the ozone column observed in
the austral spring over Antarctica were first reported in 1985 and may continue. There arc,
however, serious limitations in our current understanding of and ability to quantify ozone
depletion at the present levels of contaminant release and in our ability to predict the effects
on stratospheric ozone of any further increases (United Nations Environment Programme,
1989; United Kingdom Stratospheric Ozone Review Group, 1991).
A number of factors influence terrestrial UVR levels:
- Vilriations in stratospheric ozone with latitude and season (United Nations Environ-
ment Programme, 1989)
- Time of day: In summer, about 20-30% of the total daily amount of UVR is received
between 11:00 and 13:00 hand 75% between 9:00 and 15:00 h (Diffey, 1991; Table 2
and Fig. 3). Although the amount of visible light falling on the ground in the summer
may vary by only 30% between 12:00 and 15:00 h (local solar time), the short-
wavelength component of the UVB spectrum undergoes a dramatic change during
Table 2. Percentage of daily UVB and UVA radiation received during
different periods of a clear summer's day. Solar noon is assumed to
be at 12:00 h, i.e., no allowance is made for daylight saving time
Latitude ("N} UVB UVA
11:00-13:00 h 9:00-15:00 h 11:00-13:00 h 9:00-15:00 h
20 30 78 27 73
40 28 75 25 68
60 26 69 21 60
From Diffey (1991)
Fig. 3. Daily variation in ultraviolet radiation: erythemal effective irradiance falling on a
horizontal earth surface at Denver, CO, USA, on one summer's day
8 100
as] .. 10
e ... e
.J::. ......
>. 1.0
... - :1..
Clear day
--.... 7
Low dense
cloud cover _/
06.00 08.00 1.0.00 12DO 14.00 16.00 18.00
Standard time (h)
From Machta et at. (1975)
this period. At a wavelength of 300 nm, the spectral irradiance decreases by 10 fold,
from approximately 1.0 to 0.1

x nm) (Sliney, 1986).
- Season: Seasonal variation in terrestrial UV irradiance, especially UVB, at the
Earth's surface is significant in temperate regions but much less nearer the equator
(Table 3).
Table 3. 'JYpical values for ambient daily and annual UVB
radiation expressed in minimal erythema dose (MED)
Latitude ("N) Diurnal UVB (MED)
Winter Spring/ Autumn Summer Annual
20 (Hawaii, USA) 14 20 25 6000
30 (Florida, USA) 5 12 15 4000
40 (Spain) 2 7 12 2500
50 (Belgium) 0.4 3 10 1500
From Diffey (1991)
- Geographical latitude: Annual UVR exposure dose decreases with increasing distance
from the equator (Table 3).
- Clouds: Clouds reduce UV ground irradiance; changes in UVR are smaller than those
of total irradiance because water in clouds attenuates solar infrared radiation much
more than UVR. Even with heavy cloud cover, the scattered UVB component of sun-
light (often called skylight) is seldom less than 10% of that under clear sky; however,
very heavy cloud cover can virtually eliminate UVB even in summer. light clouds
scattered over a blue sky make little difference in sunburning effectiveness unless they
directly cover the sun. Complete light cloud cover prevents about 50% of UVB
energy, relative to that from a clear sky, from reaching the surface of the Earth
(Diffey, 1991}.
- Surface reflection: The contribution of reflected UVR to a person's total UVR expo-
sure varies in importance with a number of factors {Thble 4). A grass lawn scatters
about 3% of incident UVB radiation. Sand reflects about 10-15%, so that sitting
under an umbrella on the beach can lead to sunburn both from scattered UVB from
the sky and reflected UVB from the sand. Fresh snow has been reported to reflect up
to 85- 90% of incident UVB radiation, although reflectance of about 30-50% is
probably more typical. Ground reflectance is important, because parts of the body
that are normally shaded are exposed to reflected radiation {Diffey, 1990a).
Table 4. Representative terrain reflectance factors for horizontal
surfaces measured with a UVB radiometer at 12:00 h (290-315 nm)
in the USA
Lawn grass, summer, Maryland, California and Utah
Lawn grass, winter, Maryland
Wild grasslands, Vail Mountain, Colorado
Lawn grass, Vail, Colorado
Flower garden, pansies
Soil, clay/humus
Sidewalk, light concrete
Sidewalk, aged concrete
Asphalt roadway, freshly laid (black)
Asphalt roadway, two years old (grey)
House paint, white, metal oxide
Boat dock, weathered wood
Aluminium, dull, weathered
Boat deck, wood, urethane coating
Boat deck, white fibreglass
Boat canvas, weathered, plasticized
Chesapeake Bay, Maryland, open water
Chesapeake Bay, Maryland, specular component of reflection
at Z = 45 N
Atlantic Ocean, New Jersey coastline
Sea surf, white foam
Atlantic beach sand, wet, barely submerged
Atlantic beach sand, dry, light
Snow, fresh
Snow, two days old
From Sliney (1986)
2.0- 3.7
3.0- 5.0
10- 12
4.1- 5.0
5.0- 8.9
15- 18
- Altitude: In general, each 300-m increase in altitude increases the sunburning
effectiveness of sunlight by about 4%. Conversely, places on the Earth's surface
below sea level have lower UVB exposures than nearby sites at sea level (Diffey,
- Air pollution: Tropospheric ozone and other pollutants can decrease UVR, parti-
cularly in urban areas (Frederick, 1990).
(a) Measurements of te"estrial solar radiation
Since UVR wavelengths between about 295 and 320 nm (UVB radiation) in the terres-
trial solar spectrum are thought to be those mainly responsible for adverse health effects, a
number of studies have concentrated on measuring this spectral region (Sliney, 1986}.
Accurate measurements of UVR in this spectral band are difficult to obtain, however,
because the spectral curve of terrestrial solar irradiance increases by a factor of more than
five between 290 and 320 nm (Fig. 4). Nevertheless, extensive measurements of ambient
Fig. 4. Action spectrum designated by the American Conference of Governmental Industrial
Hygienists (ACGIH) for assessing the hazard of ultraviolet radiation (very similar to erythe-
mal action spectrum from 300-230 nm) and the solar spectrum. The ACGIH action spec
trum, which is unitless, is closely tit by some radiometers; however, because of the small over-
lap of the terrestrial solar spectrum with the action spectrum, problems of stray light must
be dealt with by constant checks with a filter that blocks wavelengths of less than 320 nm
::I. lG
0 CD
. ~
a ...
<D <
1 ~
~ ~ ~ ~ ~ ~ ~ ~ 1 i f
300 320 340 360 380 400
Wavelength (nm)
Adapted from Sliney et a/. (1990)
UVR in this spectral band have been performed worldwide (Schulze, 1962; Schulze & Grafe,
1969; Henderson, 1970; Sundararaman eta/., 1975; Garrison el a/., 1978; Doda & Green,
1980; Mecherikunnel & Richmond, 1980; Kostkowski eta/., 1982; Ambach & Rehwald,
1983; Blumthaler eta/., 1983; Uvingston, 1983; Blumthaler et al., 1985a,b; Kolari et al., 1986;
Hietanen, 1990; Sliney eta/., 1990). Longer-wavelength UVR (UVA) was measured at the
same time in many of these studies. Measurements of terrestrial solar UVA radiation are less
subject to error than measurements of UVB, since the spectrum does not vary widely with
zenith angle and the spectral irradiance curve is relatively flat.
Maps of annual UVR exposure, such as that shown in Figure 5, have been compiled for
epidemiological studies of skin cancer and other diseases (Schulze, 1962, 1970; Scotto eta/.,
1976). Despite the large numbers of measurements, their interpretation in relation to human
exposure has been complicated by three factors: (i) the considerable variation in UVB
spectral irradiance with solar position throughout the day and with ~ e a s o n (ii) the effect of
the geometry of exposure of individuals; and (iii) variation between humans in outdoor
exposure and the parts of their bodies that are exposed.
Fig. 5. Global distribution of ultraviolet radiation
From Schulze (1970); WHO (1979)
The total solar radiation that arrives at the Earth's surface is termed 'global radiation',
and measurements of terrestrial UVR most frequently pertain to this quantity, i.e., the
radiant energy falling upon a horizontal surface from all directions (both direct and scattered
radiation). Global radiation comprises two components, referred to as ' direct' and 'diffuse'.
Approximately 70% of the UVR at 300 nm is in the diffuse component rather than in the
direct rays of the sun (Fig. 6). The ratio of diffuse to direct radiation increases steadily from
less than 1.0 at 340 nm to at least 2.0 at 300 nm (Garrison eta/., 1978).
Fig. 6. Diffuse and direct solar spectral irradiance (solar zenith angle, 45)
From Garrison et al. (1978)
Spectral irradiance IJWI(cm'x nm)
..... .. .
t .. : ... .:
':: :t .. : .
.. :. ..

Ratio diffuse: direct
"' 0
UVR reflected from the terrain (the albedo) may also be important; however,essentially
all measurement programmes have been limited to the direct and total diffuse components
of sunlight. While such measurements are of interest in calculating the exposure dose of
UVR of a prone individual, they are of very limited value in estimating exposure of the eye
and shaded skin surfaces (e.g., under the chin), where the UVB radiation incident upon the
body from terrain reflectance and horizon sky is of far greater importance. Sliney (1986) and
Rosenthal eta/. (1988) reported measurements of outdoor ambient UVR that included the
reflected component to the eye. Exposure data for different anatomical sites is of value in
developing biological dose- response relationships (Diffey eta/., 1979). The fact that ocular
exposure differs significantly from cutaneous exposure is emphasized by the finding that
photokeratitis is seldom experienced during sunbathing yet the threshold for UV photo-
keratitis is less than that for erythema of the skin (Sliney, 1986).
Measurements of the angular distribution of UVR relative to solar position and cloud
distribution have been reported (Sliney, 1986; Fig. 7). A cloud obscuring the sun had no
effect upon the UV radiance of open blue sky or the horizon sky; however, when the sun was
'out' (i.e., in an open sky), clouds near the horizon opposite the sun apparently reflected more
UVR than would otherwise be present from the blue sky. This confirms the findings of
studies of photographs of the sky taken through a narrow-band filter at 320 nm (Livingston,
1983), which revealed that the sky looks almost uniformly bright even when clouds are
present and the clouds disappear into a uniformly hazy sky. Only the sun stands out, as would
be expected from the plots on Figure 7. When the sun is near the horizon and can be looked
at without great discomfort (i.e., at Z = 75- 90 ), the effective UV irradiance is again of the
order of 0.3 J!W/cm2, e.g., about 0.08-1.1 JlW/cm
at an elevation angle of 12-15 (Sliney,
Fig. 7. Semilogarithmic plots of the angular dependence of skylight for 290-315 nm
ultraviolet radiation (UVR) with the sun at zenith angle of about 45 o. A narrow field-of-view
detector was scanned from zenith to the horizon. Uppermost curves show that direct UVR
from the sun is more than 10 times greater than scattered UVR normally incident upon the
eye at near-horizon angles where the zenith angle Z = 70-90 o . Most surprising is the simi
larity of blue sky and cloudy sky UV irradiances at zenith or near the horizon.
- 10 -
~ 10 -
Sun' s position
Hazy- ,.. .... - ... .., Sunny-
toward / . .,_toward
sun ' '. sun
' , .
::: Clear sky-away.
UJ from sun
Cloudy bright-toward sun
' .
1 0
~ ~ ~ ~ ~ ~ ~ ~ ~ ~
0 30 60 90
Zenith angle (degrees)
Adapted from Sliney (1986)
(b) Personal exposures
The exposure of different anatomical sites to solar UVR depends not only on ambient
UVR and orientation of sites with respect to the sun but also on cultural and social
behaviour, type of clothing arid whether spectacles are worn.
Measurements of ambient UVR are useful in that they provide upper limits on human
exposure (Scotto et al . 1976), They are of lesser value for assessing exposure doses received
by groups of individuals. Polysulfone film has been used to monitor personal exposure to
solar UVR (see p. 49). The wide variations in recorded exposure doses reflect diversity of
behaviour and. in most cases. the small numbers ( < 30) of subjects monitored. Nevertheless.
it can be estimated that recreational (excluding vacations) exposure to the sun of people in
northern Europe (where most of these studies were carried out) results in an annual solar
exposure dose to the face of 20-100 MED. depending on the propensity for outdoor pursuits.
The annual weekday UV exposure dose of indoor workers is around 30 MED; as a two-week
outdoor vacation can result in a further 30- 60 MED. the total annual exposure dose to the
face of most indoor workers is probably in the range 40-160 MED. Outdoor workers at the
same latitudes receive about two to three times these exposure doses, typically around 250
MED (Diffey, 1987b; Slaper, 1987).
An alternative approach to estimating personal exposure is to combine measured data
on ambient UVR with a behavioural model of exposure. This approach was applied to a
group of more than 800 outdoor workers in the USA (40 oN) by Rosenthal et al. (1991).
These investigators estimated annual facial exposure doses of 30-200 MED, which are
considerably lower than those estimated for outdoor workers in northern Europe, perhaps
because Rosenthal et al. assumed facial exposure to be about 5-10% of ambient. A number
of researchers have used polysulfone film badges on both human subjects (Holman et a/.,
1983a; Rosenthal et al., 1990) and mannequins (Diffeyet al., 1977, 1979; Gies et al., 1988) to
measure solar UVR exposure on the face relative to ambient exposure. The results vary
considerably, reflecting factors such as positioning of film badges, behaviour of individuals,
solar l t i t u e and the influence of shade. Examination of the data suggests, however. that the
exposure of an unprotected face is probably close to 20% of the ambient. Using this estimate,
the annual facial exposure doses in the outdoor worker group studied by Rosenthal et a/.
(1991) would be about 80- 500 MED. These data demonstrate clearly the current
uncertainties associated with estimates of population exposure doses.
1.3.2 Exposure to artificial sources of ultraviolet radiation
(a) Sources
Six artificial sources that often produce UVR incidental to the production of visible light
(Sliney & Wolbarsht, 1980; Phillips. 1983; Moseley, 1988) are described below.
(i) Incandescent sources
Optical radiation from an incandescent source appears as a continuous spectrum. Incan-
descent sources are usually ascribed a certain 'colour temperature', defined as the tempe-
rature of a black body that emits the same relative spectral distribution as the source. UVR is
emitted in significant quantity when the colour temperature exceeds 2500 OK (2227 oc).
Thngsten-halogen lamps in a quartz envelope (colour temperature, 3000 K (2727 C]) may
emit significant UVR, whereas the UVR emission of an ordinary tungsten light bulb is
(ii) Gas discharge lamps
Another method of producing optical radiation is to pass an electric current through a
gas. The emission wavelengths are determined by the type of gas present in the lamp and
appear as spectral lines. The width of the lines and the amount of radiation in the interval
between them (the continuum) depend on the pressure in the lamp. At low pressures. fine
lines with little or no continuum are produced; as pressure is increased, the lines broaden and
their relative amounts alter. Low-pressure discharge lamps, commonly containing mercury,
argon, xenon, krypton or neon, are useful for spectral calibration. Medium-pressure mercury
lamps operate at an envelope temperature in the region of 600-800 o C.
(iii) Arc lamps
Arc lamps operate at high pressures (20-100 atm (2020-10133 kPa]) and are very
intense sources of UVR. Commonly available lamps contain xenon, mercury or a mixture of
the two elements, which are effective sources of UVR. Xenon arc lamps operate at a colour
temperature of 6000 oK (5727 oC); they are often used as the light source in solar simulation
or are combined with a monochromator in spectral illumination systems. Deuterium arc
lamps provide a useful source of UVC radiation and find their main use in spectro-
photometers and as a calibration source for spectroradiometers.
(iv) Fluorescent lamps
The primary source of radiation in a fluorescent lamp arises from a low-pressure
mercury discharge, which produces a strong emission at 254 nm, which in turn excites a
phosphor-coated lamp to produce fluorescence. By altering the composition and thickness of
the phosphor and the glass envelope, a wide variety of emission spectral characteristics can
be obtained. The output is thus chiefly the fluorescent emission spectrum from the coating,
with a certain amount of breakthrough of UVB mercury lines at 297, 303 and 313 nm, as well
as those in the UVA and visible regions (WHO, 1979).
(v) Metal halide lamps
The addition of other metals (as halide salts) to a mercury discharge lamp allows for the
addition of extra lines to the mercury emission spectrum. Most such tubes are basically
medium-pressure discharge lamps with one or more metal halide additives, usually iodide.
Advantage has been taken of the strong lead emission lines at 364, 368 and 406 nm in the lead
iodide lamp, in which there is a 50% increase in output in the region between 355 and 380 nm
compared to a conventional mercury lamp. Antimony and magnesium halide lamps provide
spectral lines in the UVB and UVC regions.
(vi) Electrodeless lamps
A type of lamp recently introduced on a large scale is the electrodeless lamp. In this
design, the discharge tube absorbs microwave energy fed, via waveguides, irito a microwave
chamber containing the tube. TWo 1500-W magnetrons generate microwave energy at 2450
MHz. The life of such lamps is longer than that of electrode lamps, and a greater range of
metal halides is available. Electrodeless lamps are used extensively for UV curing of inks and
coatings, particularly when a short lamp length is adequate for the area to be irradiated. They
have often been the first choice for curing prints on containers such as two-piece cans, plastic
pots and bottles, and tubes.
(b) Human exposure
Although the sun remains the main source of UVR exposure for humans, the advent of
artificial UVR sources has increased the opportunity for both intentional and unintentional
Intentional exposure is most often to acquire a tanned skin, frequently using sunbeds
and solaria emitting principally UVA (315-400 nm) radiation (Diffey, 1987c). Another
reason for intentional exposure to artificial UVR is the treatment of skin diseases, notably
Unintentional exposure is most often the result of occupation, and workers in many
industries (see p. 66) may be exposed to UVR from artificial sources. The general public is
exposed to low levels of UVR from sources such as fluorescent lamps used for indoor lighting
and may be exposed in shops and restaurants where UVA lamps are employed in traps to
attract flying insects.
(i) Cosmetic use
To some individuals, a tanned skin is socially desirable. A 'suntanning industry' has
grown up, particularly in northern Europe and North America, in which artificial sources of
UVR supplement exposure to sunlight.
Description of UVR sources used for tanning: Prior to the mid-1970s, the source of UVR
was usually an unfiltered, medium- or high-pressure mercury arc lamp which emitted a broad
spectrum of radiation, from UVC through to visible and infrared radiation (Diffey & Farr,
1991b). The units often incorporated one or more infrared heaters and were commonly
called 'sunlamps' or 'health lamps' (Anon., 1979). One disadvantage of this type of unit was
that the area of irradiation was limited to a region such as the face and so whole-body tanning
was tedious. By incorporating several mercury arc lamps into a 'solarium', whole body
exposure was achieved. Tanning devices based on mercury arc lamps emit relatively large
quantities of UVB and UVC radiation, resulting in a significant risk of burning and acute eye
damage. Solaria that incorporate unfiltered mercury arc lamps are therefore now less
popular (Diffey, 1990a).
So-called UVB fluorescent lamps (e.g., Westinghouse FS Sunlamp, Philips TL12) emit
approximately 55% of their UV energy in the UVB and approximately 45% in the UVA
regions (Diffey & Langley, 1986). They were often used in tanning booths, more commonly
in the USA than in Europe.
Sunbeds, incorporating high-intensity UVA fluorescent lamps, were developed in the
1970s. These devices consist of a bed and/or canopy incorporating 6- 30 fluorescent lamps
150-180 em in length. The earliest type of UVA lamp used in sun beds is typified by the Philips
TL09, Wotan Ll00/79 and Wolff Solarium lamps (Diffey, 1987c). The spectral power
distribution from this type of lamp is shown in Figure Sa. The emission spectrum comprises
the fluorescence continuum, extending from about 315 to 400 nm and peaking at 350-355
nm, together with the characteristic lines from the mercury spectrum down to 297 nm (UVB)
(Diffey & McKinlay, 1983). The UVA irradiance at the skin surface from a typical sunbed or
suncanopy containing these lamps is between 50 and 150 W/m
(Bowker & Longford, 1987;
Bruyneel-Rapp et al., 1988).
---- ------- - ----- - - -- --
Fig. 8. Spectral emissions of different lamps used for cosmetic tanning: (a) Philips TL09
(Diffey, 1987c); (b) Philips TLlOR (Diffey, 1987c); (c) Wolff BelJarium S (B.L. Diffey,
unpublished data); (d) optically filtered high-pressure metal halide lamp (Diffey, 1987c)
0.01 -
.. 0.001
g_ 0.0001
310 340 370 400
Wavelength (nm)
280 310 340 370 400
Wavelength (nm)
~ 0.0001 .
0 280
~ (d)
(; 0.1
310 340 370 400
Wavelength (nm)
280 310 340 370 400
Wavelength (nm)
In the mid-1980s, another type of UVA fluorescent lamp (Philips TLIOR) was
introduced especially for cosmetic tanning. The principal features of this type of lamp were a
reflector intrinsic to the lamp envelope and a fluorescence spectrum extending from about
340 to 400 nm, peaking at 370 nm (Fig. 8b ); note also the presence of characteristic mercury
lines in the UVB region. The skin surface irradiance from a sunbed or suncanopy
incorporating Philips TLlOR lamps is typically around 250 W/m2 (Diffey, 1987c).
Another type of UV fluorescent lamp that has been used in sunbeds is the so-caJJed 'fast
tan' tube. This type of lamp is typified by the Wolff Bellarium S, the spectral power
distribution of which is shown in Figure 8( c). The spectrum extends from about 290 to 400 nm
and peaks at around 350 nm (Diffey & Farr, 1987).
Optically filtered, high-pressure mercury lamps doped with metal halide additives are
also used in cosmetic tanning. The spectral emission lies entirely within the UVAwaveband
(Fig. 8d), and irradiances at the skin surface of more than 1000 W 1m
can be achieved. The
best known of this type of unit is probably the UVASUN (Mutzhas, 1986).
A summary of the physical and photobiological emissions from these different types of
lamps is given in Table 5 (Diffey & Farr, 1991a).
Table 5. Characteristics of different ultraviolet (UV) lamps used for tanning
Lamp Radiation emission (%)
Mercury arc sunlamp
Simulated sunlight lamp
l}'pe I UV A lamp
Type II UVA lamp
Optically filtered high-pressure lamp
Summer UV sunlightb
> 99.9
< 0.1
From Diffey & Farr (199lb) unless otherwise specified
OFrom Mutzhas (1986)
bFrom Sliney & Wolbarsht (1980)
Contnbution to tanning(%)
0 35 65
20 80 0
60 40 0
>90 < 10 0
100 0 0
20 80 0
Exposure to UVR sources used for tanning: Telephone surveys carried out in the
Netherlands (Bruggers et al., 1987) and in the United Kingdom (Anon., 1987) in the
mid-1980s showed that 7-9% of the adult population in each country had used sunbeds in the
previous one to two years. A more recent market survey in the United Kingdom
(R. McLauchlan, personal communication), with a sample size of 5800, gave a slightly higher
figure, with 10% of the population having used a sun bed during the previous year (1988) and
19% of the sample admitting to having used a sunbed at some time in the past. In these and
other surveys in the United Kingdom (Diffey, 1986) and the USA (Dougherty eta/., 1988),
women accounted for 60-85% of users, about half of the subjects being young women aged
between 16 and 30. The commonest reason given for using tanning equipment was to acquire
a pre-holiday tan (Anon., 1987; R. McLauchlan, personal communication); other reasons
included perceived health benefits, reduction of stress and improved relaxation, protection
of the skin before going on holiday, sustaining a holiday tan and treatment of skin diseases
such as psoriasis and acne (Diffey, 1986; Dougherty eta!., 1988).
In the Dutch survey (Bruggers et a!., 1987), about half of the users interviewed used
tanning equipment at home and the other half used facilities at commercial premises, such as
tanning salons, hairdressers, sports clubs and swimming pools. Most people had used UVA
equipment; 24% had used either UVB mercury arc sunlamps or solaria incorporating these
lamps. A more recent survey in the United Kingdom (McLauchlan, 1989) confirmed the
Dutch finding that the amount of use at home and at commercial premises was approximately
the same. A survey carried out at commercial establishments in the United Kingdom
indicated that all the equipment used emitted primarily UVA radiation, mostly from
fluorescent UVA lamps and 10% from optically filtered high-pressure metal halide lamps
(Diffey, 1986). Sales of tanning appliances in the United Kingdom increased rapidly during
the 1980s, but by the end of the decade there appeared to be a steady, or possibly reduced,
level of sales (Diffey, 1990a).
The mean number of tanning sessions per year in the Dutch study was 23 (Bruggers eta/.,
1987). In the United Kingdom, half-hour sessions were the most popular (Diffey, 1986). Each
tanning session with UVA equipment normally results in an erythemally-weighted exposure
of about 0.8 MED (150 J/m
), whereas exposure to mercury arc lamps results in about
2 MED per session ( 400 J/m
). In the Dutch survey, it was estimated that the median annual
exposure was 24 MED (4.8 kJ/m
) (Bruggers et al., 1987).
(ii) Medical and dental applications
UVR has both diagnostic and therapeutic applications in medicine and dentistry. The
diagnostic uses are confined largely to fluorescing of skin and teeth, and the UVR source is
normally an optically filtered medium-pressure mercury arc lamp producing radiation
mainly at 365 nm (so-called 'Wood's lamps') (Caplan, 1967). Radiation exposure is limited to
small areas ( < 15 em in diameter), and the UVA radiation dose per examination is probably
no more than 5 J/cm
. The therapeutic uses of UVR, which result in considerably higher
doses, are mainly in the treatment of skin diseases and occasionally the symptomatic relief of
Phototherapy: The skin diseases that are most frequently treated with UVR are psoriasis
and eczema. Phototherapy of psoriasis at hospital may include the use of tar and related
derivatives and other substances, such as anthralin, on the skin (Morison, 1983a; see also
IARC, 1987a).
The first treatment of psoriasis with an artificial source of UVR is credited to
Sardemann, who used a carbon arc lamp of the type developed by Finsen at around the turn
of the century. These lamps were unpopular in clinical practice because they emitted noise,
odour and sparks, and they were superseded by the development of the medium-pressure
mercury arc lamp. In the 1960s, a variety of metal halides were added to mercury lamps to
improve emissions in certain regions of the UV and visible spectra. Fluorescent lamps were
developed in the late 1940s; since then, a variety of phosphor and envelope materials have
been used to produce lamps with emissions in different regions of the UV spectrum, such
that, today, there exists a wide range oflamps for the phototherapy of skin diseases (Diffey &
Farr, 1987).
Lamp systems can be classified into one of five categories in terms of suitability for
phototherapy (Diffey, 1990b):
Type A: a single, medium-pressure mercury arc or metal halide lamp;
Type B: one or more vertical columns containing five or six optically filtered
high-pressure metal halide lamps;
Type C: a canopy or cubicle containing fluorescent sunlamps which emit predominantly
UVB but also significant amounts of radiation at wavelengths below 290 nm
(e.g., Westinghouse FS sunlamp, Philips TL12 and Sylvania UV21lamps);
TypeD: a canopy, sunbed or cubicle incorporating fluorescent lamps which emit
predominantly UVB radiation and negligible amounts of radiation at wave-
lengths below 290 nm (e.g., the Wolff Helarium);
Type E: a newly developed fluorescent lamp that emits a narrow band of radiation
around 311- 312 nm (Philips TLOl).
The spectral power distributions characteristic of each of these five types of lamp are
shown in Figure 9. The therapeutic radiation for psoriasis lies principally within the UVB
waveband (Parrish & Jaenicke, 1981), and the cumulative UVB dose required for clearing
Fig. 9. Spectral power distributions of different types of phototherapy lamp (Diffey, 1990b).
l)rpe A: unfiltered medium-pressure mercury arc lamp; type B: optically filtered iron iodide
lamp; type C: fluorescent sunlamp (Philips TL12); type D: Wolff Helarium lamp; type E:
narrow-band UVB fluorescent lamp (Philips TLOl)
10 0 10 D r-------------r----.
; 10 I
"iii ral
Gl l
a. I ()
~ Ia
a: 10 -s
10 i
200 250
; 10 I
"iii 10 -z
~ 10-
~ 10 .,
a: 10 -s
10 6
200 250
Wavelength (nm)
300 3SD
Wavelength (nm)
~ 10 I
(ij 10 -z
~ I0 -
~ 10_,
a: 10 -s
Qj 10 I
~ 10-
.:: 10 _,
a: 10 -s
10 ...
200 250 300 350
Wavelength (nm)
~ 10-l
Q. .
(ij 10 ~
4l . 1
g. 10
~ 10 _,
a: 10 -s
10 -
200 250 100 3SD
Wavelength (nm)
10 - L---.....I...U.U....----------1
200 250 300 350
Wavelength (nm)
psoriasis is typically 100-200 MED (Diffey, 1990a), usually delivered over a course
consisting of 10-30 exposures over 3-10 weeks (van der Leun & van Weelden, 1986).
Annual doses received by 90% of patients given UVB phototherapy for psoriasis range
from about 60 to 670 MED, with a typical dose in a single course being between 200 and 300
MED (Slaper, 1987).
Psora/en pholochemotherapy (see also IARC, 1980, 1986a, 1987b ): This form of treat-
ment, known colloquially as PUVA, involves the combination of photoactive drugs, pso-
ralens (P), with long-wave UVR (UVA) to produce a beneficial effect. Psoralen photo-
chemotherapy has been used to treat many skin disease in the past decade, although its
principal success has been in the management of psoriasis (Parrish eta/., 1974), a disorder
characterized by an accelerated cell cycle and rate of DNA synthesis. Psoralens may be
applied to the skin either topically or systemically; the latter route is generally preferred, and
the psoralen most commonly administered is 8-methoxypsoralen. The patient is usually
exposed to UVA radiation from banks of fluorescent lamps with the spectral power
distribution shown in Figure Sa. Values for UVA irradiance in clinical treatment cubicles
have been found to range from 16 to 140 W/m
(Diffey eta/., 1980; Diffey, 1990b), although
an irradiance of 80 W 1m
is probably typical. The UVA dose per treatment session is usually
in the range 1-10 J/cm
(Diffey et al. , 1980).
Generally, approximately 25 treatments over a period of 6-12 weeks, with a cumulative
UVA dose of 100-250 J/cm
, are required to clear psoriatic lesions (Melski eta/., 1977;
Henseler eta/., 1981). PUVA therapy is not a cure for psoriasis, and maintenance therapy is
often needed at intervals of between once a week to once a month to prevent relapse (Gupta
& Anderson, 1987).
Neonatal phototherapy for hyperbilirubinaemia: Phototherapy is sometimes used in the
treatment of neonatal jaundice or hyperbilirubinaemia. The preferred method of treatment
is to irradiate the baby for several hours a day for up to one week with visible light,
particularly blue light (Sisson & Vogl, 1982). The lamps used for phototherapy, although
intended to emit only visible light, may also have a UV component: One commercial
neonatal phototherapy unit was found to emit not only visible light and UVA but also
radiation at wavelengths down to 265 nm (Diffey & Langley, 1986).
Fluorescence in cutaneous and oral diagnosis: Wood's light-a source of UVA obtained
by filtering optically a mercury arc lamp with 'blackglass' -is used by dermatologists as a
diagnostic aid in skin conditions that produce fluorescence (Caplan, 1967; Diffey, 1990a). As
irradiation of the oral cavity with a Wood's lamp can produce fluorescence under certain
conditions, this has been used in the diagnosis of various dental disorders, such as early
dental caries, the incorporation of tetracycline into bone and teeth, dental plaque and
calculus (Hefferren et a!., 1971 ).
Polymerization of dental resins: Pits and fissures in teeth have been treated using an
adhesive resin polymerized with UVA. The resin is applied with a fine brush to the surfaces to
be treated and is hardened by exposure to UVA radiation at a minimal irradiance of 100
for 30 s or so (Eriksen et al., 1987; Diffey, 1990a).
(iii) Occupational exposures
Artificial sources of UVR are used in many different ways in the working environment.
In some cases, the UV source is well contained within an enclosure and, under normal
circumstances, presents no risk of exposure to personnel. In other applications of UVR, it is
inevitable that workers are exposed to some radiation, normally by reflection or scattering
from adjacent surfaces. Occupational exposure to UVR is also a consequence of exposure to
general lighting in the workplace.
lndustn"al photoprocesses: Many industrial processes involve a photochemical com-
ponent. The large-scale nature of these processes often necessitates the use of high-power
(several kilowatts) lamps such as high-pressure metal halide lamps (Diffey, 1990a).
The principal industrial applications of photopolymerization include the curing of
protective coatings and inks and photoresists for printed circuit boards. The curing of
printing inks by exposure to UVR is now widespread; as the cure takes only a fraction of a
second, UV drying units can be installed between printing stations on a multicolour line, so
that each colour is dried before the next is applied. Another major use ofUV curing has been
for metal decorating in the packaging industry (Phillips, 1983). UVA is also used to inspect
printed circuit boards and integrated circuits in the electronics industry (Pauw & Meulemans,
Artificial sources of UVR are used to test the weathering capability of materials such as
polymers. Xenon-arc lamps are often the light source because their emission spectra is
similar to the spectrum of terrestrial sunlight, although some commercial weathering
chambers incorporate carbon-arc lamps, high-pressure metal halide lamps or fluorescent
sunlamps (Davis & Sims, 1983).
Sterilization and disinfection: Radiation with wavelengths in the range 260-265 nm is the
most effective for this use, since it corresponds to a maximum in the DNA absorption
spectrum. Low-pressure mercury discharge tubes are thus often used as the radiation source,
as more than 90% of the radiated energy lies in the 254 nm line. These lamps are often
referred to as 'germicidal lamps', 'bactericidal lamps' or simply 'UVC lamps' (Diffey, 1990a).
UVC radiation has been used to disinfect sewage effluents, drinking-water, water for the
cosmetics industry and swimming pools. Germicidal lamps are sometimes used inside
microbiological safety cabinets to inactivate airborne and surface microorganisms (Diffey,
1990a). The combination of UVR and ozone has a very powerful oxidizing action and can
reduce the organic content of water to extremely low levels (Phillips, 1983).
Welding (see also IARC, 1990): Welding equipment falls into two broad categories: gas
welding and electric arc welding. Only the latter process produces significant levels of UVR,
the quality and quantity of which depend primarily on the arc current, shielding gas and
metals being welded (Sliney & Wolbarsht, 1980).
Welders are almost certainly the largest occupational group with exposure to artificial
sources of UVR. It has been estimated (Emmett & Horstman, 1976) that there may be as
many as half a million welders in the USA alone. The levels ofUV irradiance around electric
arc welding equipment are high; effective irradiance (relative to the action spectrum of the
American Conference of Governmental Industrial Hygienists) at 1 m at an arc current of 400
A ranged from 1 to 50 W/m
(Thble 6), and the unweighted UVAirradiance ranged from 3 to
70 W 1m
, depending on the type of welding and the metal being welded (Cox, 1987; Mariutti
& Matzeu, 1987). It is not surprising therefore that most welders at some time or another
experience 'arc eye' or 'welder's flash' (photokeratitis) and skin erythema. The effective
irradiance at 0.3 m from many types of electric welding arcs operating at 150 A is such that
the maximum permissible exposure time for an 8-h working period on unprotected eyes and
skin varies from a few tenths of a second to about 10 s, depending on the type of welding
process and the material used (Cox, 1987).
Table 6. Limits of exposure to ultraviolet radiation and
radiation effectiveness
Exposure limit
Relative spectral
effectiveness (Sx_)
1.0 X 10
1.3 X 104
1.5 X 10
1.9 X 10
25 X 104
Table 6 (contd)
Wavelength Exposure limit Relative spectral
(nm) (J/m2) effectiveness (S>..)a
320 2.9 X 104 0.0010
322 4.5 X 104 0.00067
323 5.6 X 104 0.00054
325 6.o x 10
328 6.8 X 104 O.IXXl44
330 7.3 X 104 0.00041
333 8.1 X 104 0.00037
335 8.8 X 104 0.00034
340 1.1 X lQS
345 1.3 X lOS 0.00024
350 1.5 X 105 0.00020
355 1.9 X 105 0.00016
360 2.3 X 1QS 0.00013
2.7 X lOS 0.00011
370 3.2 X lOS 0.000093
375 3.9 X 105 0.000077
380 4.7 X lOS 0.000064
385 5.7 X lOS 0.000053
390 6.8 X l()S 0.000044
395 8.3 X lQS 0.000036
400 1.5 X lo6 0.000030
From American Conference of Governmental Industrial Hygienists
(1991); . wavelengths chosen are representative, and other values
should be interpolated at intermediate wavelengths.
a'f'or explanation, see pp. 46-47
bEmission lines of a mercury discharge spectrum
In a survey of electric arc welders in Denmark, 65% of those questioned had experienced
erythema; however, as no indication of the frequency of skin reactions was reported, it is not
possible to estimate annual exposure (Eriksen, 1987). Monitoring of the exposure to UVR of
non-welders working in the vicinity of electric arc welding apparatuses showed that their
daily exposure dose exceeded the maximum permissible exposure limits by almost an order
of magnitude (Barth eta/., 1990).
Phototherapy: Although there is a trend to the use of enclosed treatment cubicles, some
of the lamps used to treat skin disease (see the section on medical and dental applications)
are unenclosed, emit high levels of UVR and can present a marked hazard to staff; at 1 m
from these lamps, the recommended 8-h occupational exposure limits can be exceeded in less
than 2 min (Diffey & Langley, 1986).
In a study of the exposure of staff in hospital phototherapy departments (Larko & Diffey,
1986), annual exposure to UVR could be estimated from the number of occasions per year
on which staff had experienced at least minimal erythema (Diffey, 1989b ). Estimated annual
occupational exposures to UVR were 15, 92 and 200 MED, corresponding to a frequency of
erythema of once per year, once per month and once per week, respectively.
Operating theatres: VVC lamps have been used since the 1930s to decrease the levels of
airborne bacteria in operating theatres (Berg, 1987). The technique requires complete
protection of the eyes and skin of staff and patients; for this and other reasons, filtered air
units are often preferred.
Research laboratories: Sources of UVR are used by most experimental scientists engaged
in aspects of photobiology and photochemistry and in molecular biology. These applications,
in which the effect of UV irradiation on biological and chemical species is of primary interest
to the researcher, can be differentiated from UV fluorescence by absorption techniques
where the effect is of secondary importance (Diffey, 1990a).
UV photography: There are two distinct forms of UV photography: reflected or
transmitted UV photography and UV fluorescence photography. In both applications, the
effective radiation lies within the UVA waveband (Lunnon, 1984).
UV lasers: High-power lasers which emit in the UV region, used in nuclear and other
research laboratories, are far less common than those that emit in the visible or infrared
regions of the electromagnetic spectrum.
Nitrogen lasers emit at a wavelength of337 nm (Phillips, 1983), and instruments with a
peak power output of up to 2.3 MW per pulse are available. Nitrogen lasers can be used in
conjunction with fluorescent dyes to produce spectral emissions of360-900 nm, with a power
pulse of 200-480 kW. If frequency doubling crystals are used in conjunction with a nitrogen
laser, UV emissions down to 260 nm are possible.
An alternative laser source of UVR is the excimer laser. (The term 'excimer' denotes a
homonuclear molecule which is bound in an electronically excited state but is dissociative in
the ground state [Phillips, 1983 ]. ) The wavelength of the pulsed UVR from this type of laser
depends on the excimer molecules, such as ArF, F
, XeCl and KrF, which emit at 193, 157,
308 and 248 nm, respectively (Phillips, 1983; Bos & de Haas, 1987). On the basis of worst-
case assumptions, the estimated annual risk for skin cancer for workers exposed to UV lasers
in medical applications is equivalent to about one additional day of sunbathing, and that for
workers exposed to UV lasers in laboratories is comparable to the risk for outdoor workers
(Sterenborg et al., 1991).
Quality assurance in the food industry: Many contaminants of food products can be
detected by UV fluorescence techniques. For example, the bacterium Pseudomonas
aeruginosa, which causes rot in eggs, meat and fish, can be detected by its yellow-green fluo-
rescence under UVA irradiation. One of the longest established uses ofUVA fluorescence in
public health is to demonstrate contamination with rodent urine, which is highly fluorescent
(Ultra-Violet Products, Inc., 1977).
Insect traps: Many flying insects are attracted by UVA radiation, particularly in the
region around 350 nm. This phenomenon is the principle of electronic insect traps, in which a
UVA fluorescent lamp is mounted in a unit containing a high-voltage grid. The insect,
attracted by the UVA lamp, flies into the unit and is electrocuted in the air gap between the
high-voltage grid and a grounded metal screen. Such units are commonly found in areas
where food is prepared and sold to the public (Diffey, 1990a).
Sunbed salons and shops: The continuing popularity of UVA sunbeds and suncanopies
for cosmetic tanning has resulted in the establishment of a large number of salons and shops
selling sunbeds for use at home. Some shops may have 20 or more UVA tanning appliances,
all switched on, thus exposing members of the public and staff to high levels ( > 20 W /m
) of
UVA radiation (Diffey, 1990a).
Discotheques: UVA 'blacklight' lamps are sometimes used in discotheques to induce
fluorescence in the skin and clothing of dancers. The levels of UVA emitted are usually low
( < 10 W/m2) (Diffey, 1990a).
Offices: Signatures can be verified by exposing a signature obtained with colourless ink
to UVA radiation, under which it fluoresces. UVA exposure of office staff is normally to
hands, and irradiance is low ( < 10 W/m
) (Diffey, 1990a).
(iv) Genera/lighting
Fluorescent lamps used for general lighting in offices and factories emit small quantities
of both UVA and UVB. A UVA irradiance of 30 mW/m
(Diffey, 1990a) and a UVB irra-
diance of3 m W 1m
(McKinley & Whillock, 1987) were found for bare fluorescent lamps with
a typical illuminance of 500 lux. These UV levels give rise to an annual exposure of indoor
workers to no more than 5 MED, and this dose can be reduced appreciably by the use of
plastic diffusers (McKinlay & WhiiJock, 1987). A study of the personal doses of UVR
received by workers in the car manufacturing industry who were engaged in inspecting
paintwork of new cars under bright fluorescent lamps indicated a similar annual exposure
(Diffey et al., 1986). Most plastic diffusers reduce erythemally effective irradiance to 0.2% or
less of that of the bare lamp. An exception is clear acrylic diffusers, which absorb only about
20% of the erythemally effective radiation. The absorption of UVA radiation by diffusers is
less effective, transmission ranging from 1% for opal polycarbonate to 74% for clear acrylic
(McKinlay & Whillock, 1987). Spectroradiometric measurements of the UV levels from
indoor fluorescent lamps carried out in the USA, however, indicated much higher annual
doses for people exposed occupationally for 2000 h per year: The annual estimated exposure
dose ranged from 8 to 30 MED for an illuminance level of 500 lux from bare lamps (Cole et
al., 1985).
Desk-top lights which incorporate tungsten- halogen (quartz) lamps may result in
exposure to UVR of the hands and arms, if the lamps are used in excess of recommended
occupational exposure levels (McKinlay eta/., 1989). Experimental studies have shown that
erythema can be induced in susceptible individuals after a 15-min exposure at 10 em from a
100-W tungsten-halogen source, principally by the UVB component of the emission
(Cesarini & Muel, 1989). Tungsten- halogen lamps are also used for general lighting (e.g.,
spotlights, indirect lighting, floor lamps) in some countries.
(c) Regulations and guidelines
(i) Cosmetic use
The most comprehensive guidelines for the use of sunlamps and sunbeds in cosmetic
tanning are those published by the International Electrotechnical Commission (1987, 1989).
The guidelines classify tanning appliances into one of four types according to the effective
irradiance at short (A. < 320 nm) and long (320 < A.< 400 nm) UV wavelengths (Thble 7).
Table 7. Classification of tanning appliances
Type Effective irradiance{Wfm2)
X ~ 320 nm
< 0.0005
< 0.15
~ 0.1
320 nm < X. ~ 400 nm
~ 0.15
~ 0.15
< 0.15
< 0.15
From International Electrotechnical Commission
Effective radiance is defined as:
where Ex. is the spectral irradiance (W/m
x nm) at wavelength A. (nm) at the shortest
recommended exposure distance; ~ x is the wavelength interval used in the summation; and
Sx. is the relative erythemal effectiveness recently adopted by the Commission lnternationale
de l'Eclairage (McKinlay & Diffey, 1987), specified as shown in Table 8. The guidelines
recommend that the exposure time for the first session on untanned skin should correspond
to an effective dose not exceeding 100 J/m
; this is approximately equivalent to 1 MED for
subjects with sun-reactive skin type I. The annual exposure should not exceed an effective
dose of 25 kJ/m
(International Electrotechnical Commission, 1989).
Table 8. Specifications of relative erythemal effectiveness
Wavelength (X.; nm)
X. < 298
298 < X. < 328
328 < X. S. 400
Relative erythemal effectiveness (Sx.)
(weighting factor)
100.094(298- X.)
1Q0.015(139- A)
From McKinlay & Diffey (1987); International Electrotechnical
Commission (1989)
Although these guidelines form the basis of several national standards on sunlamp and
sunbed use, it should be noted that variations exist; for example, in the Netherlands, Norway
and Sweden, certain UV appliances are not permitted. Regulations concerning the use of
tanning appliances are in force in only a few countries, but many others have published
advice on sunbed use, including information on adverse effects, as well as guidelines on
manufacturing standards. ,
(ii) Occupational exposure
Guidance on the maximal limits of exposure to UVR as a consequence of occupation is
given by the International Non-ionizing Radiation Committee of the International Radiation
Protection Association. These exposure limits, which apply only to incoherent (i.e., non-
laser) sources, represent conditions under which it is expected that nearly all individuals may
be repeatedly exposed without adverse effects and are below levels which would be used for
medical or cosmetic exposure to UVR. The limits for occupational exposure to UVR
incident upon the skin or eye were considered separately for the UVA spectral region
(315-400 nm) and the actinic UVspectral region(UVC and UVB, 180-315 nm). In 1984, the
limit provided an equal spectral weighting between 315 and 400 nm, a maximal 1000-s
radiant exposure of 10 KJ/m
and a maximal irradiance of 10 W /m
for longer periods (Inter-
national Non-ionizing Radiation Committee of the International Radiation Protection Asso-
ciation, 1985). Studies of skin and ocular injury resulting from exposure to UVA led the
Committee to issue revised exposure limits in 1988: For the UVA spectral region (315-400
nm), the total radiant exposure incident upon the unprotected eye should not exceed 1.0
(10 kJ/m
) within an 8-h period, and the total 8-h radiant exposure incident upon the
unprotected skin should not exceed the values given in Table 6. Values for the relative
spectral effectiveness S ,._are given up to 400 nm to expand the action spectrum into the UVA
region for determining the exposure limit for skin exposure. For the actinic UV spectral
region (UVC and UVB, 180-315 nm), the radiant exposure incident upon the unprotected
skin or eye within an 8-h period should not exceed the values given in Table 6 (International
Non-ionizing Radiation Committee of the International Radiation Protection Association,
The effective irradiance (Eerr) in W 1m
of a broad-band source weighted against the
peak of the spectral effectiveness curve (270 nm) is determined according to the formula:
where E>.. is the spectral irradiance (W /m
x nm) from measurements, S>.. is the relative
spectral effectiveness (Table 6) and is the band-width (nm) of the calculation or
measurement interval (International Non-ionizing Radiation Committee of the Interna-
tional Radiation Protection Association, 1985).
The maximal permissible exposure time in seconds for exposure to UVR incident on the
unprotected skin or eye within an 8-h period is computed by dividing 30 J/m
by the value of
Eerr in W 1m
(American Conference of Governmental Industrial Hygienists, 1991 ). A worker
receiving the maximal permissible exposure of 30 J/m
per 8-h day will, in the course of a
working year, have a cumulative dose of 60-70 MED (Diffey, 1988), a value comparable with
the naturaJ exposure of non-occupationally indoor workers (Diffey, 1990a).
Occupational exposure limits to lasers were also defined by the International Non-
Ionizing Radiation Committee of the International Radiation Protection Association in
1989, at 3 mJ/cm
and 40 mJ/cm
over 8 h for argon- fluoride and xenon-chloride lasers,
respectively (Sliney, 1990).
2. Studies of Cancer in Humans
2.1 Solar radiation
2.1.1 Nonmelanocytic skin cancer
Nonmelanocytic skin cancer is classified into two major histological types: basal-cell
carcinoma and squamous-cell carcinoma. Basal-cell carcinoma is the commoner type in
white populations. No information was available to the Working Group on other types of
nonmelanocytic skin cancer.
(a) Case reports
In general, case reports were not considered, owing to the availability of more infor-
mative data.
(i) Studies of xeroderma pigmentosum patients
Xeroderma pigmentosum is a rare autosomal-recessive genetic disease in which there is
an excision repair defect, as observed in cultured skin fibroblasts damaged by UVR (Cleaver,
1968). Patients display cellular and clinical hypersensitivity to UVR (Kraemer, 1980). The
disease is present in about one in 250 000 people in the USA and Europe (Cleaver &
Kraemer, 1989), and as many as 1 in 100 000 (Take beet a/., 1987) or even 1 in 40 000 (Cleaver
& Kraemer, 1989) people may be affected in Japan.
In a survey of 830 cases located through published case reports (Kraemer eta!., 1987),
45% had malignant skin neoplasms. Most of the patients were young, and the median age of
development of the first skin cancer in the 186 patients for whom information was available
was eight years; this observation presumably represents a substantial excess over the ex-
pected number. Only 259 neoplasms were specifically categorized as basal- or squamous-
cell carcinoma in the published reports. Of these, 97% were on constantly exposed sites (face,
head and neck) by comparison with 80% of similar tumours in the US general population.
[The Working Group recognized that data collected from previously published case reports
is not uniform and may not be typical of a true incidence or prevalence series.]
(ii) Studies of transplant recipients
Australian renal transplant recipients were reported to have an increased risk for non-
melanocytic skin cancer (Hardie eta/. , 1980). Among 875 male and 669 female Australasian
recipients, aged 35-64, 47 squamous-cell carcinomas and 27 basal-cell carcinomas were
observed among males and 27 squamous-cell and 15 basal-cell carcinomas were observed
among females (Kinlen eta/., 1979). The rates/10
person-years for squamous-ce1l carcinoma
were 2680 in males and 1710 in females, or 3.0 and 5.9 times the rates observed among resi-
dents of the same age distribution surveyed in Geraldton. Western Australia (Kricker et al.,
1990). For basal-cell carcinoma. the rates for 1540 (males) and 940 (females) were 1.154 and
1.150 times the Geraldton rates, respectively.
By February 1980, a registry in Denver, Colorado (USA), had received data on 906
organ transplant recipients who had developed 959 types of cancer: 42% arose iri the skin, of
which 47% were squamous-cell carcinomas (Penn, 1980). While several studies from areas
with lower solar radiation are available (Boyle eta/., 1984), neither singly nor collectively do
they contain enough observations to permit a comparable calculation.
(b) Descriptive studies
Nonmelanocytic skin cancer is often not recorded in cancer registries (e.g., in the USA
and in most parts of Australia), and when it is registered case ascertainment is likely to be
incomplete since many patients are treated in consulting rooms, frequently without histo-
logical verification (Doll et al., 1970). Thus, descriptive studies of the incidence of non-
melanocytic skin cancer can be difficult to perform because of the absence of routinely
collected data or difficult to interpret because of incomplete registration. Studies in Australia
and the USA have relied upon special surveys, while in the United Kingdom and the Nordic
countries data from cancer registries have been used. Studies of mortality rates are also
difficult to interpret because nonmelanocytic skin cancer is rarely fatal, and many deaths are
incorrectly attributed to skin cancer (Muir et al., 1987).
A number of features of the occurrence of nonmelanocytic skin cancer as revealed by
descriptive studies have been taken as evidence that exposure to the sun is a major cause of
the disease. These include features presumed to be related to sun exposure such as sex,
anatomical site, latitude of residence (or annual dose of UVB radiation), migration from
places of low insolation to places of high insolation, occupation and features related to
sensitivity to the sun such as race (i.e., degree of skin pigmentation).
(i) Host factors
The occurrence of nonmelanocytic skin cancer according to host factors such as race
provides indirect evidence that sunlight is a cause. In most white populations, non-
melanocytic skin cancer occurs more commonly in men than in women (Muir et al., 1987).
The highest incidence rates have been recorded among Australians, who are largely of
British (Celtic) descent (Giles et al., 1988). Populations with greater skin pigmentation have
low rates of nonmelanocytic skin cancer, for instance, in South Africa (Oettle, 1963) and
Singapore (Shanmugaratnam et al., 1983).
Albinism is an inherited disorder of melanin metabolism, with a decrease or complete
absence of melanin. Large numbers of skin cancers (mostly squamous-cell carcinomas) have
been reported in albinos (Luande et al., 1985; Kromberg eta/., 1989).
(ii) Anatomical distribution
The majority of cases of skin cancer recorded in cancer registries (Haenszel, 1963
[USA]; Whitaker et al., 1979 [United Kingdom]; Swerdlow, 1985 [United Kingdom]; Levi et
al., 1988 [Switzerland]; 0sterlind et al., 1988a [Denmark]; Moan eta/. , 1989 [Norway]) and in
special surveys in the USA (Haenszel, 1963; Scotto eta/., 1983) occurred on the head and
neck. In contrast, in two studies in Australia-one of incidence (Giles eta/., 1988) and the
other of prevalence (Kricker eta/., 1990)-the proportions of cancers on the head and neck
were lower. [The Working Group noted that the contrasting results may be due to time
differences.] In the incidence survey, 43% of squamous-cell carcinomas and 66% of
basal-cell carcinomas were on the head and neck. In the prevalence survey, about one-third
of all basal-cell carcinomas were on the head and neck, whereas the trunk accounted for
about half of these lesions. The density of tumours was five times greater in men and eight
times greater in women on usually exposed sites than on sites which were sometimes exposed.
Squamous-cell carcinomas occurred almost exclusively on exposed sites. The site distri-
butions of both types of nonmelanocytic skin type are generally similar in the two sexes
(0sterlind eta/., 1988a; Moan eta/. , 1989; Kricker eta/., 1990).
A distinctive feature of the site distribution of basal-cell carcinoma is a virtual absence
on the dorsa of the hands and infrequent occurrence on the forearms, compared with the
distribution of squamous-cell carcinoma (Haenszel, 1963; Silverstone & Gordon, 1966; Levi
eta/., 1988; Magnus, 1991). Basal-cell carcinoma also occurs frequently on parts of the face
that receive comparatively little sun exposure (Urbach eta/., 1966).
[The Working Group noted that cancers on the head and neck may be more likely to be
diagnosed than cancers at other sites.]
(iii) Geographical variation
Nonmelanocytic skin cancer incidence and mortality have long been known to increase
with increasing proximity to the equator. Gordon and Silverstone (1976) demonstrated a
negative correlation between incidence of nonmelanocytic skin cancer in various countries
and latitudes by tabulating the incidence according to latitudinal zones. Much of the early
evidence came from surveys conducted in the USA. In the first of these, Dorn (1944a,b,c)
reported the results of the US First National Cancer Survey conducted in 10 urban areas in
1937-38. [Nonmelanocytic] skin cancer incidence was greater among whites living in the
south than in the north of the country. Blum (1948) subsequently reanalysed these data,
substituting latitude for place of residence, and showed a strong inverse relationship between
incidence of mostly nonmelanocytic skin cancer and latitude. No other cancer, with the
exception of the buccal cavity (including the lip), showed a similar latitude gradient.
Auerbach (1961), using data from the US Second National Cancer Survey conducted in
1947-48 in the same areas as the previous survey, calculated that the age-adjusted rates for
skin cancer doubled for each 3 o 48 ' (approximately 265 miles) of latitude towards the
equator; similar gradients were seen for men and women and in all age groups. Haenszel
(1963) reanalysed data from this survey for four southern and four northern cities. The
inverse gradient with latitude was present for both basal-cell and squamous-cell carcinoma.
In addition, there was some evidence that the gradient was strongest for head, neck and
upper limbs (sites which are usually exposed).
A similar latitude gradient was seen in the US Third National Cancer Survey (Scotto
eta/., 1974). Inverse latitude gradients have also been reported in Australia (Silverstone &
Gordon, 1966; Giles et al., 1988) and in the Nordic countries (Teppo eta/., 1980; Moan eta/.,
1989; Magnus, 1991).
Several authors have correlated nonmelanocytic skin cancer incidence (or mortality)
with estimates of UVR. Green eta/. ( 197 6) reported a positive correlation be tween estimates
of annual UV dose and of incidence rates in the USA, the United Kingdom, Canada and
Australia. Estimates of UV dose were derived from models relating latitudinal and seasonal
ozone distributions, adjusted for cloud cover. [The Working Group noted that no allowance
was made in the analysis for different methods of case ascertainment. It is not clear how well
the predicted values were correlated with actual levels of UVR.]
A positive correlation, stated to be stronger than that for latitude, was seen between
UVR, as measured by Robertson-Berger meters, and the incidence of nonmelanocytic skin
cancer in four cities in the US Third National Cancer Survey (Scotto et al., 1982). Scotto et al.
(1983) examined incidence data collected in eight cities in 1977-78 and again showed an
inverse relationship with latitude and a positive correlation with measurements of UVR. The
gradient was steeper for squamous-cell than for basal-cell carcinoma.
Moan eta/. (1989) examined nonmelanocytic skin cancer incidence in six regions of
Norway from 1976 to 1985, excluding the area around Oslo to reduce bias due to possible
differences in reporting and diagnosis. Two measures of UVR, one weighted according to the
action spectrum for erythema and the other according to the action spectrum for mutagenesis
in cells in the basal layer of the skin, were derived from atmospheric models. Similar, positive
relationships between UVR and nonmelanocytic skin cancer incidence were obtained with
each method.
Elwood eta/. (1974) conducted a study of mortality from nonmelanocytic skin cancer in
the contiguous states of the USA and in all of the provinces of Canada in 1950-67. The
correlation between latitude and mortality was as strong as that between mortality and an
index of UVR derived from a model relating erythemal dose according to latitude with
adjustments for cloud cover.
(iv) Migration
Studies of migrants to Australia (and other countries with high exposure to the sun) offer
the opportunity to examine, indirectly, the effect of exposure to the sun. Most migrants to
Australia come from higher latitudes which have lower levels of exposure to the sun than
Australia. The effect of exposure to the sun is most readily examined in migrants from the
British Isles to Australia, from whom most Australians are descended.
Armstrong eta/. (1983) found that the age-adjusted mortality rate among men born in
England or Wales was 0.55 (95% confidence interval (CI), 0.43-0.71) times that in
Australian-born men. There was little evidence that rates in migrants increased with duration
of residence in Australia, although the numbers of deaths were small and the rates unstable.
Giles et al. (1988) found age-adjusted incidence rates of 402 per 100 000 person-years
among immigrants from the British Isles and 936 in the Australian-born population.
(v) Occupation
Death certificates for 1911-44 in England and Wales were used in an analysis of cancer
of the skin, excluding melanomas, in male agricultural workers, miners and quarriers and
professionals (Atkin eta/., 1949). During part of the period (1911-16), cancers ofthe penis,
scrotum and skin were classified together, and the numbers of cancers of the skin alone were
estimated from the proportions occurring in the later period. The standardized mortality
ratios (SMRs) were greater for those engaged in agriculture (142.4 [137.4-147.6]) than for
those in mining (94.4 (88.8-100.3]), and lowest of all for professionals (47.5 (42.6-52.9]).
Whitaker eta/. (1979) examined occupations among cases of squamous-cell carcinoma
reported to the Manchester Regional Cancer Registry, United Kingdom, in 1967-69. The
occupations of 23% of cases were not ascertained. In men, standardized registration ratios
(SRRs) were elevated for textile workers (238;p < 0.001) and farmers (243;p < 0.001). The
SRR was also high for female farmers (690; p < 0.001). Male fishermen, chemical workers
and paper/printing workers had high SRRs for squamous-cell carcinoma of the arm, and
building workers for squamous-cell carcinoma of the ear.
The association between occupation and nonmelanocytic skin cancer was examined in
England and Wales in 1970-75 in a 10% sample of all male incident cases for which occu-
pation was recorded (Beral & Robinson, 1981 ). Individuals were assigned, on the basis of
stated occupation, to one of three groups: outdoor workers, indoor office workers and other
indoor workers, according to the classification of occupations of the Office of Population
Censuses and Surveys. The SRRs for men aged 15-64 were 110 [95% CI, 109-116) for
outdoor work, 97 [92-103] for office work and 92 (86-89] for other indoor work. Since place
of work may be confounded with social class, the analyses were repeated for men aged 15-64
years in social class III; the SRRs were 112 (102-122) for outdoor work, 111 [100-123) for
office work and 85 [78-92] for other indoor work.
Vagero eta/. (1986) linked cancer incidence data in Sweden from 1961 to 1979 with
census data from 1960 to determine the occupations of cases of nonmelanocytic skin cancer.
Occupations were classified into three main groups: office workers, other indoor workers
and outdoor workers. SRRs standardized for age, county of residence and social class, were
slightly higher for outdoor workers (106; 95% CI, 101-112) than for office workers (103;
96-110) and other indoor workers (95; 91- 100). The authors noted that registration may
have been more complete among high socioeconomic groups.
(c) Cross-sectional studies
Design features of cross-sectional studies of exposure to the sun are summarized in
Table 9, and the results are shown in Table 10.
A population-based survey of the prevalence of nonmelanocytic skin cancer [types not
separated] was conducted in County Galway, Ireland (O'Beirn eta/., 1970). Exposed areas of
skin were examined for the presence of cancers. In the 26 cases found, there was no signi-
ficant association with frequent severe sunburn for basal-cell or squamous-cell skin cancer;
among males, there was a positive relationship between cumulative hours of exposure to
sunlight and the prevalence of nonmelanocytic skin cancer.
Silverstone and Gordon (1966) and Silverstone and SearJe ( 1970) reported the results of
three surveys in Queensland, Australia. Exposed areas of the skin were examined, and
subjects were asked to report previously treated nonmelanocytic skin cancer [types not
separated]. Women performing home duties were classified as indoor workers. Outdoor
occupation showed a weakly positive association with past and present incidence in men and
a negative association in women.
Holman eta/. (1984a) conducted a population-based survey of 1216 subjects in western
Australia. After controlling for age, cutaneous sun damage (as assessed by microtopography)
was strongly related to a past history of nonmelanocytic skin cancer.
Engel et al. (1988) analysed data on basal-cell epithelioma (carcinoma) from the First
National Health and Nutrition Examination Survey in the USA (1971-74). Dermatologists
diagnosed skin cancers and assessed actinic skin (solar) damage, but histological con-
firmation of the diagnosis was not obtained routinely. Strong associations between the
prevalence of basal-cell epithelioma and solar skin damage were seen in both men and
Green eta/. (1988a) conducted a survey of the prevalence of nonmelanocytic skin cancer
(types not separated for calculation of RR] in Queensland, Australia. Information about
exposure to the sun was obtained from questionnaires; dermatologists diagnosed skin
cancers and assessed signs of actinic damage (solar lentigines, telangiectasia of the face, solar
elastosis of the neck and solar keratoses). After adjustment for age, sex, skin colour and
ability to tan, outdoor occupation and number of sunburns were both weakly associated with
increased prevalence. Stronger associations were seen for cutaneous indicators of sun
exposure, particularly for solar lentigines on the hands and telangiectasia on the face.
Recreational exposure was not associated independently with nonmelanocytic skin cancer.
In a later report (Green, 1991), the occurrence of nonmelanocytic skin cancer was posi-
tively correlated with grade of cutaneous microtopography.
In a subsequent study (Green & Battistutta, 1990), subjects were asked to report
nonmelanocytic skin cancer treated between 1 December 1985 and 30 November 1987,
around the survey in 1986. Medical records were searched to confirm the diagnoses. Subjects
who had had a skin cancer diagnosed at the prevalence survey were excluded. Outdoor occu-
pation, outdoor leisure activities and number of sunburns showed little association with
basal-cell carcinoma in an analysis including past history of skin cancer. All three variables
were related to incidence of squamous-cell carcinoma. [The Working Group noted that the
exclusion of subjects found to have skin cancer during the prevalence survey makes inter-
pretation of these results difficult. The inclusion of past history of skin cancer in the analysis
would have weakened any association with exposure to the sun. ]
Vitasa et al. (1990) conducted a survey of the occurrence of nonmelanocytic skin cancer
among men engaged in traditional fishing practices ('watermen') in Maryland, USA Subjects
were examined by dermatologists and interviewed about their history of exposure to the sun.
Estimates of individual annual and lifetime doses of UVB radiation were made by weighting
the ambient UVR by a history of occupation and outdoor activities and by taking into
account relative doses recorded by film dosimeters on the face. Patients with squamous-cell
carcinoma aged 15-60 had had an 11% higher annual dose ofUVB radiation and those with
basal-cell carcinoma had had an 8% lower annual dose than that of age-matched watermen
without cancers. The effect of cumulative UVB radiation was examined after adjustment for
age, eye colour, childhood freckling and skin reaction to sunlight, all of which were positively
associated with occurrence of both types of nonmelanocytic skin cancer. Cumulative UVB
radiation dose was not associated with basal-cell carcinoma but was positively associated
with squamous-cell carcinoma. The latter association was significant in a comparison of the
top quarter of cumulative UVB versus the bottom three-quarters but not in a comparison of
exposures above and below the median. (The Working Group noted that the results for the
two types of cancer are not necessarily incompatible, both because of the small number of
cases and the fact that the diagnosis was confirmed histopathologically in only 62%.]
Table 9. Design features of cross-sectional studies of sun exposure and nonmelanocytic skin cancer
Reference Place Period of Population Sample Response Cases Histological
size rate
O'Beim et al. (1970) County 1960s Population-based 1338 Approx. 13 BCC; 13 sec on Incomplete; 57%
81% exposed sites only had biopsies
t:J Ireland
Silverstone & Gordon Queensland, 1961- 63 Population-based About 2200 87% 221 BCC or SCC on Incomplete
(1966); Silverstone & Australia
exposed surfaces 0
Searle (1970) "r1
Holman eta/. (1984a) Busselton, 1981 Population-based 1216 102, type not stated No
El)gel et al. (1988) USA 1971-74 Population -based 20 637 74% BCC, number not Incomplete (small

- stated proportion]
Green el al. (1988a) Nambour, 1986 Population-based 2095 70-78% 42 sec or sec (90% Yes :I:
of subjects examined
on head/neck/ hands/
forearms only]
Green & Nambour, 1985-87 Population-based 1770 84% 66 BCC; 21 SCC self- Incomplete
Battistutta (1990) Australia
reported (confirmed
from medical records)
Vitasa et at. (1990) Maryland, 1985-86 Male fishermen 838 70% 33 Bee; 35 sec Incomplete
USA > 30 years old
BCC, basal-cell carcinoma; SCC, squamous-cell carcinoma

Table 10. Summary of results of cross-sectional studies of nonmelanocytic skin cancer
Reference Index of exposure Categories Odds ratio (95% CI) Comments
O'Beirn et a!., (1970) Sunlight hours (lifetime) < 30000h 1.00 Mean aged > 60 years;
> 50000h [8.10 (1.2-348.2)] calculated from raw data
(p = 0.02]
Silverstone & Searle Occupation Indoors 1.0 Men, chi-square = 1.4
(1970) Outdoors (1.29] (p > 0.1]; calculated from raw
data, no adjustment
Occupation Indoors 1.0 Women, chi-square = 0.3
- Outdoors [0.6} (p > 0.1J; calculated from raw
data, no adjustment
Holman et al. (1984a) Cutaneous microtopography Grades 1-3 1.0 p = 0.004, trend adjusted for age
Grade 4 3.9
Grade 5 3.6 z
Grade 6 9.2
Engel et al. (1988) Solar skin damage None 1.0 BCC, men, age-adjusted pre-
Any [8.0) valence ratio, p < 0.01
None 1.0 BCC, women, age-adjusted pre-
Any [6.0) valence ratio, p < 0.01
Green et al. (1988a) Occupational exposure Indoors 1.00 Adjusted for age, sex, skin colour
Indoors and outdoors 1.01 (0.44:..2.31) and propensity to sunburn
Outdoors 1.76 (0.77-4.05)
Painful sunburns None 1.00 Adjusted for age, sex, skin colour 1:!1
1 0.77 (0.22-2.61) and propensity to sunburn
2- 5 1.09 (0.41-2.95)
~ 1.66 (0.59-4.64)
Solar lentigines on hands None 1.00 Adjusted for age, sex and other
1-10 1.61 (0.78-3.35) signs of actinic damage
11-20 1.43 (0.43-4.77)
~ 21 3.78 (1.06-13.41)
Telangiectasia on face None 1.00 Adjusted for age, sex and other
Mild 1.63 (0.58-4.57) signs of actinic damage
Moderate 2.74 (0.89-8.40)
Severe 3.67 (0.79-17.11)
Table 10 (contd)
Index of exposure
Odds ratio (95% CI) Comments
Green et a/. (1988a)
Actinic elastosis on neck
Adjusted for age, sex and other
Mild to moderate
1.42 (0.53-3.80) signs of actinic damage
1.75 (0.56-5.45)
Solar keratoses on face
Adjusted for age, sex and other
1.55 (0.67-3.59)
signs of actinic damage
1.86 (0.69-5.04)
3.00 (0.54-16.69)
2.72 (0.73-10.15)
V> Green & Battist utta BCC
Occupational exposure
Mainly indoors
Adjusted for age, sex, skin colour
Indoors and outdoors 1.5 (0.8-2.9)
and past history of skin cancer (')
Mainly outdoors
1.3 (0.6-2.8)
z Leisure exposure
Mainly indoors
Adjusted for age, sex, skin colour ("')
Indoors and outdoors 1.0 (0.4-2.2)
and past history of skin cancer m
Mainly outdoors
0.6 (0.3-1.3)
No. of painful sunburns
Adjusted for age, sex, skin colour
0.5 (0.2-1.4)
and past history of skin cancer
c: 2- 5
0.6 (0.3- 1.5)
6 1.0 (0.4-2.5)
Occupational exposure
Mainly indoors
Adjusted for age, sex, skin colour
Indoors and outdoors 4.4 (0.9- 20.9)
and past history of skin cancer
Mainly outdoors
5.5 {1.1-28.2)
Leisure exposure
Mainly indoors 1.0
Adjusted for age, sex, skin colour
Indoors and outdoors
2.0 (0.2- 19.9)
and past history of skin cancer
Mainly outdoors
3.9 (0.5-30.9)
No. of painful sunburns (}-1
Adjusted for age, sex, skin colour
3.3 (0.9-12.3)
and past history of skin cancer
6 3.0 (0.7-12.2)
Table 10 (contd)
Reference Index of exposure Categories Odds ratio (95% CI) Comments
Vitasa et aJ. (1990) sec
Cumulative UVB dose to face Below median 1.0 Proportionate odds ratios; ad
Above median 2.05 (0.84-5.01) justed for age, eye colour. freck-
Below 75 percentile 1.0 ling and sunburn reaction
Above 75 percentile 2.53 (1.18-5.40)
Cumulative UVB dose to face Below median 1.0 Proportionate odds ratios; ad-
Above median 0.69 (0.31- 1.53) justed for age, eye colour, freck-
Below 75 percentile 1.0 ling and sunburn reaction
Above 75 percentile 1.11 (0.50-2.44)
BCC, basal-cell carcinoma; SCC, squamous-cell carcinoma; unless otherwise specified, all analyses are for the two types together





(d) Case-control studies
Design features of the case-control studies of exposure to the sun and the occurrence of
nonmelanocytic skin cancer are summarized in Table 11. Most of the studies employed
hospital- or clinic-based controls, which introduces potential for selection bias. The results
are summarized in Thble 12. The methods of analysis and of measurements of exposure to the
sun, particularly in the earlier studies, were crude. Neither sensitivity to the sun, usually
measured as the ability to tan or propensity to burn, nor pigmentary characteristics (such as
skin colour and hair colour), which are likely to be confounding variables, were taken into
account in most of the analyses.
The hospital-based study of Lancaster and Nelson (1957) in Sydney, Australia, was
primarily a case-control study of melanoma (described in detail on p. 100). It can also be
considered to be a case-control study of nonmelanocytic skin cancer, however, because it
included two control groups-one of patients with basal-cell carcinoma, squamous-cell
carcinoma or solar keratosis and the second of patients with leukaemia or cancer at a site
other than the skin. All groups were matched by age and sex. Among males, long duration of
occupational exposure to the sun was associated with an increased risk for nonmelanocytic
skin cancer or solar keratosis. A summary of total exposure to the sun was devised by
assigning scores to a number of factors considered to be related to exposure to the sun. Risk
was highest among subjects judged to have excessive exposure to the sun. [The Working
Group noted that the proportion of cases who had a solar keratosis is not stated, that no
account was taken of matching in the analyses, and that the effect of exposure to the sun was
not adjusted for sensitivity to the sun.]
Gellin et al. (1965) conducted a study in a single hospital in New York, USA, on 861
patients with basal-cell carcinoma and 1938 non-cancer dermatological patients attending
the same clinic. Since 95% of cases and 43% of controls were 40 years old and over, the study
was limited to these patients, resulting in 771 cases and 783 controls. The skin cancer patients
spent more time outdoors per day than did control patients and were significantly more likely
than controls to have light hair, fair complexion, blue eyes and an inability to tan. [The
Working Group noted that the analyses were not adjusted for age, sex or sensitivity to the sun,
and that confounding by age is likely because controls were younger than cases.]
Urbach et a/. (1974) conducted a hospital-based study in Philadelphia, USA, and
compared exposure to the sun of 392 patients with histologically confirmed basal-cell carci-
noma, 59 patients with histologically confirmed squamous-cell carcinoma and 281 out-
patients receiving treatment for a skin disease other than cancer. Controls were matched to
cases by age and sex. Among male patients, those with basal-cell or squamous-cell carcinoma
had more cumulative hours of exposure than did controls. Skin cancer patients also reported
more sunburns. [The Working Group noted that the analyses were not adjusted for ability to
tan, age or sex (apart from the sex-specific analysis).]
Vitaliano (1978) subsequently reanalysed the data of Urbach et al. (1974) and showed
that, after adjustment for complexion (dark versus pale), ability to tan and age ( < 60, > 60),
the cumulative time spent outdoors was related to both types of nonmelanocytic skin cancer.
For basal-cell carcinoma, the odds ratio for> 30 000 h of exposure relative to < 10 000 h was
3.19; for squamous-cell carcinoma it was 22.8. [The Working Group noted that confi-
dence intervals were not given. Part of the apparently stronger effect for squamous-cell
carcinoma could be due to confounding by age: the controls were matched by age to the
basal-cell carcinoma cases, who were younger than the squamous-cell carcinoma cases.]
A hospital-based case-control study was conducted in Montreal, Canada (Aubry & <