Articles

Incidence of unilateral arm lymphoedema after breast

cancer: a systematic review and meta-analysis
Tracey DiSipio, Sheree Rye, Beth Newman, Sandi Hayes

Summary
Lancet Oncol 2013; 14: 50015 Background The body of evidence related to breast-cancer-related lymphoedema incidence and risk factors has
Published Online substantially grown and improved in quality over the past decade. We assessed the incidence of unilateral arm
March 27, 2013 lymphoedema after breast cancer and explored the evidence available for lymphoedema risk factors.
http://dx.doi.org/10.1016/
S1470-2045(13)70076-7
Methods We searched Academic Search Elite, Cumulative Index to Nursing and Allied Health, Cochrane Central
See Comment page 442
Register of Controlled Trials (clinical trials), and Medline for research articles that assessed the incidence or prevalence
School of Public Health and
Social Work, Institute of Health
of, or risk factors for, arm lymphoedema after breast cancer, published between Jan 1, 2000, and June 30, 2012. We
and Biomedical Innovation, extracted incidence data and calculated corresponding exact binomial 95% CIs. We used random eects models to
Queensland University of calculate a pooled overall estimate of lymphoedema incidence, with subgroup analyses to assess the eect of dierent
Technology, Queensland,
study designs, countries of study origin, diagnostic methods, time since diagnosis, and extent of axillary surgery. We
Australia (T DiSipio PhD,
S Rye MAppSc, assessed risk factors and collated them into four levels of evidence, depending on consistency of ndings and quality
Prof B Newman PhD, and quantity of studies contributing to ndings.
S Hayes PhD)
Correspondence to: Findings 72 studies met the inclusion criteria for the assessment of lymphoedema incidence, giving a pooled estimate
Dr Tracey DiSipio, School of of 166% (95% CI 136202). Our estimate was 214% (149298) when restricted to data from prospective cohort
Public Health and Social Work,
Faculty of Health, Queensland
studies (30 studies). The incidence of arm lymphoedema seemed to increase up to 2 years after diagnosis or surgery of
University of Technology, breast cancer (24 studies with time since diagnosis or surgery of 12 to <24 months; 189%, 142247), was highest
Victoria Park Road, Kelvin Grove, when assessed by more than one diagnostic method (nine studies; 282%, 118535), and was about four times higher
QLD 4059, Australia
in women who had an axillary-lymph-node dissection (18 studies; 199%, 135282) than it was in those who had
t.disipio@qut.edu.au
sentinel-node biopsy (18 studies; 56%, 6179). 29 studies met the inclusion criteria for the assessment of risk factors.
Risk factors that had a strong level of evidence were extensive surgery (ie, axillary-lymph-node dissection, greater
number of lymph nodes dissected, mastectomy) and being overweight or obese.

Interpretation Our ndings suggest that more than one in ve women who survive breast cancer will develop arm
lymphoedema. A clear need exists for improved understanding of contributing risk factors, as well as of prevention and
management strategies to reduce the individual and public health burden of this disabling and distressing disorder.

Funding The National Breast Cancer Foundation, Australia.

Introduction Individual studies report arm lymphoedema in up to

Lymphoedema after breast cancer is characterised by
regional swelling, typically in one or both arms, due to
excess accumulation of protein-rich uid in body tissues.1
The adverse consequences of lymphoedema are well
known, and cause much morbidity. Arm lymphoedema,
and its associated symptoms, such as pain, heaviness,
tightness, and decreased range of motion, impede daily
function and adversely aect gross and ne motor skills,
with negative ramications for work, home, and personal
care functions, as well as recreational and social relation-
ships.2 The appearance of a swollen and sometimes
disgured limb provides an ever-present reminder of
breast cancer, which can contribute to anxiety, depres-
sion, and emotional distress in aected women.3 Fur-
thermore, preliminary ndings show that lymphoedema
might also lead to shortened survival.4 In view of the
increasing incidence of breast cancer worldwide,
understanding the incidence of subsequent secondary
lymphoedema and its associated risk factors is clearly of
public health importance.
94% of patients with breast cancer,5 with the wide
variation (as low as 0%) in reported results an indication
of dierences in study design, diagnostic methods and
criteria used, and timing of lymphoedema measurement
with respect to breast cancer diagnosis and treatment.6
Some estimates suggest that about 20% of women will
develop arm lymphoedema after breast cancerthis
estimation is the average incidence of studies that have
been included in several systematic reviews of
lymphoedema after breast cancer.79 However, the average
incidence of a group of studies does not take into account
factors that are known to aect detection rates, such as
study design or timing and method of lymphoedema
assessment. How common such lymphoedema is after
breast cancer is, therefore, unclear. Furthermore, our
understanding of acquired and pre-existing risk factors is
imperfect. Although more extensive treatment and a
higher body-mass index have long been thought to be the
major risk factors for the development of lymphoedema,
advances in treatment over the past 1015 years raise

500 www.thelancet.com/oncology Vol 14 May 2013


questions about whether associations between the risk of
lymphoedema and these characteristics, as well as other
personal, treatment, and behavioural characteristics,
have changed.
The body of evidence relating to the incidence of arm
lymphoedema after breast cancer has grown substantially
and has improved in quality during the past decade,
now including ndings from several prospective cohort
studies. We therefore did this systematic review and
meta-analysis to provide the most up-to-date estimate of
the incidence of arm lymphoedema after breast cancer.
Also, although the strength of treatment-related risk
factors has been assessed in a 2009 meta-analysis,10 it is
important to also consider the strength and consistency
of the association between lymphoedema and other non-
treatment-related risk factors, as well as timely to update
ndings regarding treatment-related risk factors.
Methods
Search strategy and selection criteria
We did a systematic review to identify all studies ad-
dressing the incidence of, prevalence of, or risk factors
for breast-cancer-related arm lymphoedema. We did a
comprehensive search of databases including Academic
Search Elite, Cumulative Index to Nursing and Allied
Health, Cochrane Central Register of Controlled Trials
(clinical trials), and Medline to identify studies published
between Jan 1, 2000, and June 30, 2012, that included
women who had undergone surgery for breast cancer.
The search terms included keywords for breast cancer
(breast and cancer or onco*, or neoplasm*), results from randomised controlled trials assessing an
lymphoedema (lymphoedema or lymphedema), and
the outcomes of interest (incidence, prevalence, risk intervention groups and all data (including baseline date)
factor, or prognosis).
Eligibility criteria for inclusion of studies in this review
and meta-analysis fell into six categories. Type of study:
published research articles were included; review papers,
meta-analyses, editorial or comment papers, case re-
ports, and case series were excluded, as were randomised
controlled trials that did not report lymphoedema at
baseline or lymphoedema for the control group sep-
arately. Patient characteristics: studies of female patients
with unilateral breast cancer were included; studies of
patients with bilateral breast cancer, primary lymph-
oedema, or metastatic disease were excluded. Diagnosis
of lymphoedema: self-reported swelling was the only
symptom taken as an indication of self-reported
lymphoedema; studies that reported the incidence of
lymphoedema on the basis of only multiple symptoms
(eg, do you have pain, tingling, or weakness of the
arm?) were excluded, because these symptoms are
common irrespective of lymphoedema status,11 and the
inclusion of such symptoms might therefore lead to an
overestimation of lymphoedema incidence. All objective
methods of diagnosing lymphedoema were included.
Outcome: incidence of, prevalence of, or risk factors for
secondary lymphoedema were includedin the absence
www.thelancet.com/oncology Vol 14 May 2013
Articles
of pretreatment lymphoedema status, prevalence was
thought to be a reasonable estimate of incidence because
the proportion of women with lymphoedema before
surgery for breast cancer has been reported to be very
low.12,13 Time period: outcome data measured within
3 months of diagnosis or surgery were excluded because
arm-related changes during this timeframe were
considered potentially indicative of an acute treatment-
related response. Language and origin: we included
studies available from all locations with reports written
in English; non-English-language papers, when
translations were unavailable, were excluded.
Data extraction
One investigator (TD) selected articles that potentially met
our inclusion criteria on the basis of their titles and
abstracts. Full articles were then retrieved for a more
detailed assessment. We developed a data abstraction sheet
to collect necessary information to establish the level of
evidence, study quality, and available outcome and risk
factor details. From every included study, one investigator
(TD) extracted data for study location (country), study
design, sample size, time since breast cancer diagnosis,
method of lymphoedema assessment, denition of
lymphoedema, incidence or prevalence of lymphoedema,
and any risk factor information. Study designs included
randomised controlled trials, cross-sectional, prospective
cohort, retrospective cohort, and casecontrol studies
(casecontrol studies were only included for risk factor
analysis). For our meta-analysis of incidence, we recorded
exercise intervention: we included baseline data for the
for the control groups. Lymphoedema measurement refers
to the technique used to dene the presence of
lymphoedema and included bioimpedance spectroscopy,
arm circumferences, water displacement or perometry
(optoelectronic volumeter), lymphoscintigraphy, clinician
diagnosis, and patient-reported diagnosis by a clinician or
self-reported swelling.
We categorised all studies that analysed the incidence
of arm lymphoedema into levels of evidence, on the basis
of study design, using levels of evidence (Prognosis
column) dened by the Australian National Health and
Medical Research Council (NHMRC).14 Two investigators
(TD, SR) independently categorised each study with
disagreements resolved through discussion with a third
assessor (SH) to attain consensus.
We assessed the presence of publication bias using
funnel plots by precision, Eggers Test of the Intercept,15
and Duval and Tweedies trim and ll procedure (data
not shown).16 Funnel plots were analysed for the overall
incidence and subgroup random eects models (relating
to sentinel-lymph-node biopsy compared with axillary-
lymph-node dissection) by plotting the event rate against
the inverse of the SE. The funnel plot was symmetrical
about the summary eect, with larger studies at the top
501
 Articles
Panel: Criteria for assessment of the methodological
quality of studies (risk factors)
Criteria were scored as positive (sucient information and a
positive assessment), negative (sucient information but
potential bias due to inadequate design or conduct), or
unclear (insucient information).
Study population
A Description of target or source population
B Description of relevant inclusion and exclusion criteria
C Description of study population
Response
D Response 75%
E Information on non-responders versus responders
Follow-up
F Prospective data collection
G Follow-up 6 months
H Dropouts or loss to follow-up 20%
I Information on participants who completed the trial
versus those who dropped out or were lost to follow-up
Treatment
J Treatment in study population was described
Outcome measures
K Standardised assessment of relevant outcome measures
Analysis and data presentation
L Frequencies of most important outcome measures
M Appropriate analysis techniques
N Suciently large study population (100 people per study
group)
and smaller studies at the bottom, indicating no clear
evidence of publication bias. Analysis with Eggers Test
of the Intercept suggested that the intercept of the
regression line approached the origin, which is indicative
of no or low publication bias (intercept 082; 95% CI
36 to 201; p=056). The Duval and Tweedies trim and established to assess which result to include in the overall
ll procedure also suggested no publication bias from
potentially missing studies.
Two investigators (TD, SR) independently assessed each
risk factor study for methodological quality using a 14-item
standardised checklist of predened criteria (panel)
modied from an established criteria list for systematic
reviews.1719 Similar adaptations have been used to review
studies on lymphoedema and quality of life in breast
cancer survivors.20,21 Each item of a selected study was
assigned one point, with the highest possible score for any
one study being 14. Although all 14 indicators were used
equally, criteria specic to follow-up were included in the
checklist, hence prospective cohort studies were assigned
extra points. Studies scoring more than 10 points (>70% or
more of the maximum attainable score) were rated as high
quality, studies scoring 79 points were rated as moderate
quality, and studies scoring 6 or fewer points (lower than
502
50% of maximum attainable score) were regarded as low
quality.1719 Disagreement between the investigators was
again resolved with the aid of a third assessor (SH).
Findings for each risk factor were collated into four levels
of evidence, depending on consistency (ie, association in
the same direction), quality, and quantity of studies. Each
study was also classied according to the NHMRC Levels
of Evidence; the Aetiology column was used in most cases
except for exercise or surgical intervention trials, in which
case we used the Intervention column.14
Statistical analysis
The main outcome of interest for this analysis was overall
cumulative incidence (%) of arm lymphoedema after
diagnosis of or treatment for breast cancer, which was
abstracted from the published reports of incidence or
prevalence; exact binomial 95% CIs were subsequently
calculated. When incidence was presented separately by
treatment (eg, sentinel-lymph-node biopsy vs axillary-
lymph-node dissection), we calculated an overall incidence
for the study using data from the reports. For each risk
factor, we abstracted odds ratios, hazard ratios, or risk
ratios and associated 95% CIs, based on unadjusted results
for randomised controlled trials and adjusted estimates
obtained from multivariate analyses (ie, controlled for at
least one other factor) for other study designs. Odds ratios
tend to overestimate relative risk;22 however, the clinical
interpretation of ndings remains similar unless
prevalence of the disorder under investigation is greater
than 30%.23 When study results were not appropriately
presented (such as mean arm volume change presented
without a threshold for dening lymphoedema, p values
but no odds ratios presented for risk factors, or odds ratios
presented for risk factors without CIs), we contacted the
investigators of the studies (n=5) but were unable to
provide results needed for inclusion in the meta-analysis.
In the absence of a diagnostic gold standard, and
because some studies presented more than one esti-
mation of incidence, a hierarchy of decisions was
meta-analysis. For timing, the result at or closest to
2 years after diagnosis was considered the most
appropriate because most patients usually present by this
time.6,24,25 For studies that used more than one method of
arm lymphoedema assessment, when available, results
from objective circumference measurements were
abstracted (because this is the most common method
used in clinical practice),26 followed by perometry
(because it assesses size dierence between limbs,
similar to circumferences), followed by bioimpedance
spectroscopy (an objective and validated measure27), and
then self-report. When incidence was presented for more
than one diagnostic criterion, incidence data abstracted
for inclusion in the meta-analysis showed the most
common diagnostic criteria used by the studies included
in the review (specically, 2 cm dierence between
limbs for circumferences and 10% for perometry).
www.thelancet.com/oncology Vol 14 May 2013
 Articles

We used a random eects model to produce a pooled Incidence of arm lymphoedema following breast cancer
overall estimate for lymphoedema incidence and inci- We calculated a pooled estimate for arm lymphoedema
dence estimates that allowed for comparison within incidence of 166% (95% CI 136202) using data
lymphoedema measurement (bioimpedance spectros- abstracted from 72 studies of 29 612 women with
copy, clinical diagnosis, circumferences, perometry, self- breast cancer (table 1). Incidence ranged between
reported diagnosis, self-reported swelling, and multiple 84% and 214%, with prospective cohort studies giving
methods), time since diagnosis (3 months to <6 months, the highest estimate and retrospective cohort studies
6 months to <12 months, 12 months to <24 months, giving the lowest (table 1).
2 years to <5 years, 5 years), and extent of axillary The lowest incidence was in the UK and the highest
surgery (sentinel-lymph-node biopsy or axillary-lymph- was in Australasia. The highest estimates were reported
node dissection). We used Comprehensive Meta-analysis by studies that used more than one measure to classify
(version 2) for statistical analyses.28 lymphoedema, whereas the one study31 that classied
lymphoedema according to lymphoscintigraphy reported
Role of the funding source the lowest incidence (table 1). Incidence of a clinical
The study sponsor had no role in the study design, data
collection, data analysis, data interpretation, or writing of Included Incidence (%; 95% CI)
the report. The corresponding author had full access to studies (n)
the data in the study and had nal responsibility for the Pooled estimate
decision to submit for publication. All studies 72 166 (136202)
Prospective cohort studies 30 214 (149298)
Results Randomised clinical trial 7 104 (79135)
We identied 398 potentially relevant citations, of which
Retrospective cohort 10 84 (54128)
79 were included in our analysis (gure 1).
Cross-sectional studies 25 177 (138224)
Most studies were either prospective or cross-
Location of study
sectional, but we also identied randomised control
Asia 4 180 (102298)
trials and retrospective cohort studies, nearly half of
Australasia 7 215 (150298)
studies were from North America, and the most
Europe 17 142 (109184)
common method of lymphoedema measurement was
Middle East 1 175 (139218)
arm circumference; timing of measurements varied
North America 32 210 (151285)
greatly, ranging from 3 months to 20 years after
South America 3 137 (81222)
diagnosis (tables 1 and 2).
UK 8 84 (51136)
Axillary surgery
SLNB 18 56 (6179)
398 individual studies identified and screened
ALND 18 199 (135282)
Measurement method*
286 excluded after assessment of abstracts
Lymphoscintigraphy 1 50 (16144)
and titles
Bioelectrical impedance 3 159 (46426)
Self-reported clinical diagnosis 5 125 (62236)
112 full-text articles for eligibility
Clinical diagnosis 7 126 (81193)
Circumference 38 148 (114190)
33 articles excluded Perometry 17 164 (109241)
7 duplicate or secondary analyses Self-reported swelling 19 204 (138290)
3 recored pre-operation incidence only
19 did not meet eligibility criteria More than one measure 9 282 (118535)
3 simply assessed lymphoedema Time since breast cancer diagnosis or surgery*
1 was done more than 30 years ago
3 to <6 months 8 103 (62167)
6 to <12 months 15 138 (73245)
79 articles included in review (74 incidence 12 to <24 months 24 189 (142247)
and 31 risk factors)
2 to <5 years 30 186 (136248)
5 years 16 156 (100235)
72 studies on incidence* More than one time category 6 76 (27195)
29 studies on risk factors
ALND=axillary-lymph-node dissection. SLNB=sentinel-lymph-node biopsy.
*Numbers (percentages) might not add to 72 (100%) because some studies
Figure 1: Flow diagram of study selection for inclusion in this review and
measured lymphoedema with more than one method or at more than one
meta-analysis
timepoint.
*We included 74 articles, which presented results of 72 studies (two studies were
each presented in two articles). We included 31 articles, which presented results Table 1: Incidence of breast cancer-related lymphoedema
of 29 studies (two studies were each presented in two articles).

www.thelancet.com/oncology Vol 14 May 2013 503

 Articles
Sample size and
characteristics
Prospective cohort studies
Armer et al 236 newly diagnosed patients
(2010);5 USA followed up for 5 years
Bennett 50 patients 12-months
Britton et al post-surgery and reassessed at
(2007);29 UK 3948 months
Bland et al 90 newly diagnosed patients
(2003);30 USA (stage 0IV) followed up for
3 years
Celebioglu et al 60 newly diagnosed patients
(2007);31 (grade 13) followed up for
Sweden 3 years
Clark et al 188 newly diagnosed patients
(2005);24 UK followed up for 3 years
Clough Gorr 400 newly diagnosed patients
et al (2010);32 (stage IIII) followed up for
USA 7 years
Devoogdt et al 49 patients (level I or II)
(2011);33 3-months post-surgery
Belgium (axillary dissection) and
followed up for 34 years
Francis et al 155 newly diagnosed patients
(2006);34 USA (stage 03) followed up for
12 months
Geller et al 145 initial interviews done an
(2003);35 USA average of 97 months
post-diagnosis and follow up
interviews an average of
26 months; stage local or
regional/distant recruited
Goldberg et al 600 newly diagnosed patients,
(2010);36 USA reassessed 38 years
post-surgery (median of
5 years); node-negative
recruited; T stage Tis to T2,
N stage N0 to N1
Goldberg et al 600 newly diagnosed patients,
(2011);37 USA reassessed 38 years
post-surgery (median of
5 years); T stage Tis to T2,
N stage N0 to N1
Han et al 97 newly diagnosed patients
(2012);38 South (N0N3), reassessed every
Korea 3 months
504
Measurement method
Arm circumferences;
perometry
Arm volume
Arm circumferences
Water displacement;
lymphoscintigraphy
Percentage volume dierence
(PVD): dierence in volume
between the aected and
unaected arm expressed as a
percentage of the unaected
limb volume
PVD change: baseline to
3 years adjusted for volume
change in the unaected limb
Clinical diagnosis
Self-reported swelling in the
past 4 weeks
Arm circumference
Arm circumference and
volume
Self-report arm swelling
Arm circumferences
Self-reported swelling
Arm circumference
Lymphoedema denition Incidence* Risk factors (RF) Level of evidence
and study quality
2 cm circumference; 200 mL At 6 months: 2 cm, Level II
perometry 78/178 (44%)
At 24 months: 2cm,
144/178 (81%); 200 mL,
108/193 (56%)
At 60 months: 2 cm,
167/178 (94%)
>10% relative increase in At 3948 months: Level II
volume of the ipsilateral arm 14/50 (28%)
>2 cm change at any site At 3 years: 32/90 (36%) Level II
Arm volume: operated arm Arm volume: 6/60 (10%) Level II
>10% the size of the Lymphatic dysfunction:
non-operated arm 3/60 (5%)
Lymphatic dysfunction:
clearance from the injection
site, amount of indicator in the
axilla vs in the injection site
after 180 min, and time before
1% of the injected activity
reached the axilla
PVD change 5% PVD change 5% at 3 years: Level II
Clinical diagnosis made by a 19/188 (10%)
health-care professional Any LE at 3 years:
39/188 (21%)
Persistent LE: positive response Persistent LE (over 7 years): Level II
on at least two interviews over 145/400 (36%)
the follow-up period
>10% dierence between the 3 months: 2/49 (4%) Level II
involved and the uninvolved 34 years: 9/49 (18%)
side; at least at one or both
measurement sites
Discrepancy in volume or 12 months: total, 105/155 Severe LE (>10%): Incidence: level II
circumference of 5% (68%); SLNB, 26/155 (17%); axillary surgery RF: level II
ALND, 73/155 (47%) Study quality 10
Arm or hand swelling since Follow-up: 55/145 (38%) <50 years Incidence: level II
diagnosis RF: level II
Study quality 8
>2 cm for either the upper arm Follow-up: 31/600 (5%) Level II
or the forearm
Current swelling Follow-up: 18/600 (3%) Greater number of Incidence: level II
lymph nodes RF: level II
removed during Study quality 10
surgery
2 cm 6 months: total, 1/97 (1%); Level II
SLNB, 0/14 (0%); ALND,
1/83 (1%)
(Continues on next page)
www.thelancet.com/oncology Vol 14 May 2013
 Articles

Sample size and Measurement method Lymphoedema denition Incidence* Risk factors (RF) Level of evidence
characteristics and study quality
(Continued from previous page)
Hayes et al 287 patients (grade 13) Bioimpedance spectroscopy 3 or more SDs higher than Age, income, RF level II
(2008a);39 recruited 6 months normative data surgery, physical Study quality 12
Australia post-diagnosis when recruited activity levels
Hayes et al 287 patients (grade 13) Cumulative burden SOAC measured by 6 months: SOAC (11%) Income, child Incidence: level II
(2008b);6 6 months post-diagnosis and SOAC lymphoedema >5 cm; BIA 9 months: SOAC (15%) status, side of RF: level II
Australia followed up for 18 months Bioimpedance spectroscopy measured by lymphoedema 3 or 12 months: SOAC (19%) treatment, lymph Study quality 10
post-diagnosis (BIA) more SDs higher than normative 15 months: SOAC (20%) node excision,
Self-reported arm swelling data; self-reported arm swelling 18 months: SOAC (22%); radiation, baseline
measured by lymphoedema BIA (34%); self-report (45%) symptoms
denition of yes
Hayes et al 183 patients (grade 13) Bioimpedance spectroscopy >3 SD above norm 6 year follow-up: BIA, Level II
(2011);4 6 months post-diagnosis and Self-reported swelling History of arm swelling 11/166 (7%); swelling,
Australia followed up to 6 years 40/183 (22%)
post-diagnosis
Helyer et al 137 newly diagnosed patients Arm volume >200 mL increase Arm volume at 20 months: BMI Incidence: level II
(2010);40 (grade 13), followed-up for 16/137 (12%) RF: level II
Canada 20 months (16 years) Study quality 10
Herd-Smith et 1278 newly diagnosed patients Arm circumference Increase >5% 46 years post-diagnosis: Radiotherapy; Incidence: level II
al (2001);41 (stage T1N0 to T4N+) and 203/1278 (16%) number of nodes RF: level II
Italy followed up for 46 years removed Study quality 9
Krki et al 96 patients 6 months SAQ impairments and severity Upper limb oedema 6 months: 23/96 (24%) Level II
(2005);42 post-surgery and followed up 12 months: 25/96 (26%)
Finland for 12 months
Kosir et al 30 newly diagnosed patients Arm volume 10% increase in limb volume 3 months: 1/20 (5%) Level II
(2001);43; USA (stage 03) followed up for 6 months: 2/19 (11%)
6 months
McLaughlin et 936 newly diagnosed patients Arm circumferences and >2 cm Circumferences: total, Level II
al (2008);44 (T stage Tis to T3; N stage self-report swelling Arm swelling 86/936 (9%); SLNB,
USA N0 to N3) followed up for a 31/600 (5%); SLNB+ALND:
median of 5 years post-surgery 55/336 (16%)
recruited Current arm swelling: total,
109/936 (12%)
Norman et al 631 newly-diagnosed patients Self-reported presence and Whether right and left hands Cumulative incidence: year 1, Level II
(2009);25 USA (stage in situ to III/IV) and degree of LE diered in size 160/631 (25%); year 2:
followed up for 5 years Dierence rated: none 191/631 (30%); year 3,
(score 0), mild (13), moderate/ 211/631 (33%); year 4,
severe (49) 228/631 (36%); year 5,
238/631 (38%)
Norman et al 631 newly diagnosed patients Self-reported presence and Whether right and left hands Any LE, BMI, RF level II
(2010);45 USA (stage in situ to III/IV) followed degree of LE diered in size education, ALND, Study quality 9
up for 5 years Dierence rated: none any previous
(score 0), mild (13), moderate/ anthracycline-based
severe (49) chemotherapy
Moderate/severe,
LE: ALND
Chemotherapy
Ozcinar et al 218 newly diagnosed patients Arm circumference 2 cm 12 months: total 54/218 Incidence: level II
(2012);46 (stage I, II, N0) followed up to (25%), SLNB 6/80 (8%),
Turkey 64 months post-surgery ALND 25/138 (18%); 64
(53 years) months: total 16/218 (7%)
Pain et al 70 newly diagnosed patients Volume 10% relative increase in 3 months: 11/70 (16%) Level II
(2005);47 UK followed up for 12 months volume of all or part (forearm or 12 months: 8/70 (11%)
upper arm) of the ipsilateral arm
Paskett et al 622 newly diagnosed patients SAQ: arm and hand swelling Any swelling in their arm or By 1 year: 224/622 (36%) Persistent swelling, Incidence: level II
(2007);48 USA (stage IIII) followed up for hand since surgery or in the By 3 years: 336/622 (54%) number of nodes RF: level II
3 years past 6 months removed, obese Study quality 11
Petrek et al 263 newly diagnosed patients Self-reported arm No LE: no arm swelling plus During the 20 years follow-up Level II
(2001);49 USA followed up for 20 years circumferences: standardised dierence of 127 cm LE was identied in:
mailed instrument, verbal Mild: <127 cm plus self-report of 128/263 (49%)
reports of arm swelling during arm enlargement or heaviness
telephone interviews Moderate: >127 cm
Severe: 508 cm
(Continues on next page)

www.thelancet.com/oncology Vol 14 May 2013 505

 Articles

Sample size and Measurement method Lymphoedema denition Incidence* Risk factors (RF) Level of evidence
characteristics and study quality
(Continued from previous page)
Quinlan et al 278 patients (stage IIII) Arm morbidity: volume, McGill Any of the following: >10% Some form of arm morbidity: Level II
(2009);50 612 months post-surgery Pain SAQ increase, score >0, <80 for 211/278 (76%)
Canada Range of motion rotation, <170 for abduction
Rampaul et al 1242 patients diagnosed Self-reported arm swelling Problematic arm symptoms Arm swelling: 5/1242 Level II
(2003);51 UK between 1973 and 2000 and within a 15 week period at the (<05%)
assessed at various times of clinic
follow-up
Ronka et al 83 newly-diagnosed patients Arm circumference, Mean dierence at the six 1 year post-surgery: Level II
(2005);52 (stage Tis to T4) and self-reported arm oedema measurement points (LE%) circumference: 13/83 (16%);
Finland followed up for 12 months >5%, arm oedema self-report, 35/83 (42%)
Schrenk et al 70 newly diagnosed patients (T Self-report (none, mild, Arm swelling of the operated At follow-up: total, Level II
(2000);53 stage T1a to T2) with a median moderate, severe) arm compared with the 19/70 (27%); SLNB, 0/35
Austria follow-up of 154 months in non-operated arm (mild: periods (0%), ALND 19/35 (54%)
the SNLB group and 17 months of arm swelling but no constant
in the ALND group increase in greatest dimension
and clothes t the same, to
severe: constant arm heaviness,
disability, decreased functional
activity, huge arm swelling)
Stout Gergich 196 newly diagnosed patients Perometer Volume increase 3% in treated Cumulative: 43/191 (23%) Level II
et al (2008);54 (stage 03) followed-up for upper limb compared with
USA 18 months pre-operation measure
Wilke et al 5327 newly diagnosed patients Arm circumference >2 cm change from baseline 6 months: 201/2904 (7%) Age, BMI Incidence: level II
(2006);55 USA (stage T1 to T2) followed up for compared with contralateral RF: level II
6 months arm Study quality 10
Wojcinski et al 34 newly diagnosed patients Arm circumference or 4 cm, manual compression 3 months: 12/34 (35%) Level II
(2012);56 (who had ALND surgery) subjective complaints or lymphatic massage, 6 months: 16/34 (47%)
Germany followed-up for 6 months medical treatment compression garments,
bandaging
Yang et al 191 newly diagnosed patients Arm circumference >1 cm dierence 3 months: 11/191 (6%) LE at 12 months: Incidence: level II
(2010);57 South followed up for 12 months 6 months: 16/187 (9%) ALND RF: level II
Korea 12 months: 22/183 (12%) Study quality 14
Retrospective cohort studies
Crosby et al 1117 patients (1499 breasts) Subjective or objective Data documented by 50/1499 (3%) Axillary Incidence: level III-3
(2012);58 USA 56 months post-treatment health-care providers interventions, RF: Level III-2
greater number of Study quality 9
positive lymph
nodes, radiation
therapy, BMI
Hayes et al 2579 patients (stage III) Assessed at regular intervals Severity: graded by the 464/2579 (18%) Radiotherapy eld, Incidence: level III-3
(2008c);59 USA 81 months post-treatment (every 612 months after physicians physical obesity index, RF: level III-2
radiotherapy) by the treating examination adjuvant therapy, Study quality 9
radiation oncologist number of lymph
nodes dissected
Hinrichs et al 105 patients (stage IIV) Oedema Presence of ipsilateral arm Median time to onset 391 days Radiotherapy Incidence: level III-3
(2004);60 USA post-radiotherapy oedema noted by a treating (range, 331632 days): before 1999 RF: level III-2
physician (severity based on the 28/105 (27%) Study quality 8
treating physicians impression)
Kingsmore et 2122 patients (grade 13), Objective persistent arm At least 1 year after completion Total, 136/1960 (7%); SLNB, Level III-3
al (2005);61 UK median follow-up of 8 years swelling noted by clinician of axillary treatment requiring 20/146 (5%); ALND,
since diagnosis further treatmenteg, 69/1099 (6%)
a compression sleeve
Lumachi et al 205 patients (pT1b to pT2) Arm circumferences >2 cm Total, 15/205 (7%); SLNB, Level III-3
(2009);62 Italy with median follow-up of 2/54 (4%); ALND, 13/151 (9%)
22 months post-surgery
Mathew et al 506 patient >2 years Arm circumferences >2 cm Total, 31/506 (6%); SLNB, Level III-3
(2006);63 UK post-surgery 7/312 (14%); ALND,
24/194 (12%)
Powell et al 727 newly diagnosed patients Arm circumferences assessed > 2cm Persistent arm lymphoedema, Level III-3
(2003);64 USA (stage III) given breast at regular intervals Persistent oedema: if 21/727 (3%)
conservation treatment symptoms did not resolve at
included subsequent visits
(Continues on next page)

506 www.thelancet.com/oncology Vol 14 May 2013


Sample size and
characteristics
(Continued from previous page)
Schijven et al 393 patients (stage 0III)
(2003);65 3 months to 3 years
Netherlands post-treatment
Shih et al 1877 patients; used data for
(2009);66 USA the 2 years after treatment
Wernicke et al 265 patients (stage T1 to T2)
(2011);67 USA median follow-up at 99 years
post-treatment
Randomised clinical trials
Ahmed et al 45 patients (stage DCIS to III),
(2006);68 USA 23 in intervention (weight
training) and 22 in control; 22
months post-diagnosis at
baseline
Box et al 65 newly diagnosed patients
(2002);69 (stage IIII), 32 in treatment
Australia (early physiotherapy after
axillary dissection), 33 in
control
Lee et al 61 patients (stage DCIS to III), Arm circumference
(2007);70 30 in the control group and
Australia 31 in the intervention group
(stretching after radiotherapy);
3437 months post-surgery
Lucci et al 891 newly diagnosed patients
(2007);71 USA (T stage T1 or T2 recruited; N0,
M0)
445 SLND+ALND (usual),
446 SLND alone
Newly-diagnosed
Mansel et al 1031 newly diagnosed patients Self-assessment for severity
(2006);72 UK (stage IIII); 515 SLNB,
516 standard axillary surgery
Schmitz et al 154 patients (stage in situ to
(2010);73 USA III), 77 in intervention (weight
lifting) and 77 control;
15 years post-diagnosis at
study entry
Veronesi et al 516 patients (grade 13): 257 in Arm circumference
(2003);74 Italy the axillary-dissection group
(SLNB+total axillary dissection)
and 259 in the sentinel-node
group (SLNB+axillary dissection
if SLNB contained metastases);
100 patients in each study
group was used to describe
side-eects
6 and 24 months after surgery
Casecontrol studies
Soran et al 156 patients (T stage 0 to IV):
(2006);75 USA 52 with LE, 104 without
(control)
Breast/axillary surgery between
1990 and 2000
www.thelancet.com/oncology Vol 14 May 2013
Measurement method
SAQ: treatment-specic QoL Complaint present if responded Total, 17/393 (4%); SLNB,
SAQ developed and validated
for this study
Insurance claims for LE
treatment or diagnosis on
health insurance database
Arm circumferences
Arm circumference,
self-reported clinician
diagnosis
Arm circumferences (CIRC);
arm volume; multi-frequency
bBioimpedance impedance
(MFBIA)
Arm circumferences;
subjective (risk factors)
Arm volume; clinician-dened
Arm circumferences: severity Still well: % above normal arm:
dened according to dierence slight, 1120; moderate, 2140;
in volume between aected marked, 4180; severe, >80
and unaected arms
Lymphoedema denition
23; absent if responded 01 2/180 (1%); ALND,
15/213 (7%)
Formal diagnosis listed on
health insurance claims
database
>1 cm dierence
Circumference 2 cm;
self-reported clinical diagnosis
of LE at baseline
Increase 5cm from
pre-operation, increase
200 mL from pre-operation
operated arm:unoperated arm
ratio ratio below the 95% CI of
pre-operation study group
>2 cm
>2 cm; patient self-report or
physician diagnosis (risk
factors)
Mild, moderate, severe
5%; certied LE therapists used Results for control group at
a standardised clinical
assessment, including inter-limb
dierences, changes in tissue
tone or texture, and symptoms
>2 cm
Articles
Incidence* Risk factors (RF) Level of evidence
and study quality
ALND Incidence: Level III-3
RF: III-2
Study quality 9
Within 2 years of treatment ALND; Incidence: level III-3
initiation, 180/1877 (10%); chemotherapy RF: level III-2
4 year follow-up, 12% Study quality 8
Total, 45/223 (20%); SLNB, Level III-3
5/108 (5%); SLNB+ALND,
40/115 (35%)
Baseline results for total Level IV
sample: 2 cm, 8/45 (18%);
clinical diagnosis,
13/45 (29%)
Control group at 6 months
follow-up (post-intervention):
2 cm, 1/22 (5%)
Results for control group BMI Incidence: level III-2
3 months: 5 cm, 0/33 (0%) RF: level IV
6 months, 5 cm: 1/33 (3%) Study quality 13
12 months: 5 cm: 1/33 (3%)
24 months: 5 cm: 4/33
(12%); 200 mL, 9/33 (27%);
MFBIA, 5/33 (15%)
Results for control group Level III-2
3 months: 6/61 (10%)
7 months: 4/30 (13%)
6 months: total, 50/541 (9%) ALND; age Incidence: level III-2
12 months: total, RF: level II
40/468 (9%) Study quality 9
SLND alone: 14/226 (6%)
SLNB+ALND: 26/242 (11%)
3 months: total, 81/812 (10%) ALND Incidence: level III-2
6 months: total, 90/846 (11%) RF: level II
12 months: total: 73/815 Study quality 10
(9%); SLNB, 20/412 (5%),
ALND, 53/403 (13%)
Level III-2
1 year (post-intervention)
Arm volume, 13/75 (17%);
clinician-dened, 3/69 (4%)
6 months: total, 8/200 (4%) Level IV
24 months: total,
12/200 (6%); SLNB, 0/100;
ALND, 12/100 (12%)
Infection of the RF level III-3
operated side Study quality 6
High body-mass
index, more use of
hand (occupation/
hobby)
(Continues on next page)
507
 Articles

Sample size and Measurement method Lymphoedema denition Incidence* Risk factors (RF) Level of evidence
characteristics and study quality
(Continued from previous page)
Swenson et al 388 patients: 94 with LE, Clinical diagnosis; referred for Overweight RF level III-3
(2009);76 USA 94 without (control) LE management; arm Study quality 6
Average of 6 years after circumferences
surgery
Cross-sectional studies
Ahmed et al 1287 (stage in situ to regional Self-reported LE Self-reported diagnosis 104/1287 (8%) Level IV
(2008);77 USA or distant); 81 years
post-diagnosis
Armer et al 100 patients (stage 0 to IV) Arm circumferences 2 cm Total: 36/100 (36%) Level IV
(2004);78 USA 28 months post-surgery SLNB alone: 2/9 (22%)
ALND+SLNB: 32/79 (41%)
Asim et al 193 who had ALND, Arm circumference; arm 2 cm; 10%; quite a bit or very 2 cm: 49/193 (25%) Increasing age; Incidence: level IV
(2012);79 New 56 months post-surgery volume; self-reported arm much >10%: 14/73 (19%) radiotherapy to RF: level IV
Zealand swelling Swelling: 21/193 (11%) axilla; radiotherapy Study quality 7
to breast
Infection to
operated arm
Blanchard et al 776 patients with early-stage Self-reported presence of arm Total: 70/774 (9%) Level IV
(2003);80 USA invasive node-negative breast LE SLNB-only: 39/683 (6%)
cancer (stage 0 to III); 24 years SLNB+ALND: 31/91 (34%)
post-surgery
Deo et al 299 patients (stage IIII) Arm circumferences >3 cm 100/299 (33%) Axillary irradiation; Incidence: level IV
(2004);81 India 25 years post-treatment comorbidities RF: level IV
Study quality 8
Edwards et al 201 patients 3 years Water displacement; 10% dierence between limbs; Objective: 22/201 (11%) Level IV
(2000);82 post-treatment subjective assessment swelling Subjective: 48/201 (24%)
Australia
Eversley et al 116 patients (stage IIV) Swelling from LE (110 scale) LE related swelling 73/116 (63%) Level IV
(2005);83 USA diagnosed and treated in the
past 2 years
Freitas-Silva 70 patients (stage IIII) Arm circumferences 2 cm dierence between limbs Overall prevalence of LE: Level IV
et al (2010);84 2535 years post-surgery 11/70 (16%)
Brazil
Golshan et al 125 patients (TNM stage 1 Arm circumferences >3 cm dierence between Total: 15/125 (12%) Level IV
(2003);85 USA to 3) >1 year post-surgery operated and non-operated SLNB: 2/77 (3%)
side ALND: 13/48 (27%)
Graham et al 91 patients (T stage 14) Arm circumferences; limb Arm circumferance: >2 cm; Circumferance: LE by VOL+CIR: Incidence: level IV
(2006);86 42 years post-radiotherapy volume volume >200 mL 40/89 (45%); volume axillary irridation RF: level IV
Australia 36/85 (42%) Age Study quality 7
Haddad et al 355 patients, 7984 months Arm circumferences Increase of 10% in the 62/355 (17%) Level IV
(2010);87 Iran post-surgery circumference of the arm on the
involved side compared with
the opposite arm
Hayes et al 176 Patients (grade 13) Arm circumferences; sum of SOAC: >5 cm Side of treatment; RF level IV
(2005);88 6 months post-diagnosis arm circumferences; SOAC: >10% blood pressure Study quality 10
Australia multifrequency bioimpedance MFBIA: 3 SD readings taken on
spectroscopy; self-reported Self-report: yes in previous operated arm
arm swelling 6 months
Kwan et al 467 patients , >2 years Arm volume 200 mL 14/112 (13%) Level IV
(2002);89 post-diagnosis
Canada
Leidenius et al 139 patients 3 years Arm circumferences; clinical >2 cm between operated and >2 cm: 3/139 (2%) Level IV
(2005);90 post-surgery signs; self-report non-operated wrist; swelling Clinical signs: 7/139 (5%)
Finland Swelling: 18/139 (13%)
Lopez Penha et 145 patients, >5 years Arm circumferences; >5 cm 5 cm: 23/145 (16%) Level IV
al (2011);91 post-surgery perometry; self-report >200 mL SLNB: 24/76 (32%)
Netherlands Swelling in the past year of the ALND: 37/69 (54%)
upper-limb on the treated side Total: 200mL, 11/145 (8%);
self-report, 25/145 (17%)
(Continues on next page)

508 www.thelancet.com/oncology Vol 14 May 2013

 Articles

Sample size and Measurement method Lymphoedema denition Incidence* Risk factors (RF) Level of evidence
characteristics and study quality
(Continued from previous page)
McCredie et al 809 patients, median of Self-reported arm problems: Arm problem on the treated Arm swelling: 119 years Arm swelling; BMI Incidence: level IV
(2001);92 31 years post-diagnosis shoulder stiness, arm side (excluding the rst post-diagnosis, 20%; RF: level IV
Australia swelling, arm numbness or 6 months after diagnosis) 229 years post-diagnosis, Study quality 8
pain or ache in the arm 25%; 339 years
post-diagnosis, 28%; 4 years
post-diagnosis, 29%
Meeske et al 494 (stage in situ to III-A) Self-reported arm LE Swelling due to an 4 years post-diagnosis Age at diagnosis; Incidence: level IV
(2009);93 USA <18 months post-diagnosis to accumulation of uid in their 120/494 (24%) high blood RF: level IV
4 years post-diagnosis arm, not to be confused with pressure; BMI Study quality 9
swelling that occurs after >10 nodes excised
surgery
Nagel et al 106 (stage IIII); median of Arm circumferences; volume 2 cm; >200 mL Circumference: 10/106 (9%) Level IV
(2003);94 143 months post-surgery Volume: 13/106 (12%)
Netherlands
Nesvold et al 263 Volume+circumferences 10% in volume or 2cm Volume/arm circumference: Number of Incidence: level IV
(2008);95 Median of 47 months Self-report circumference 43/263 (16%) metastatic axillary RF: level IV
Norway post-surgery; stage II recruited Arm swelling Swelling: 63/263 (24%) lymph nodes; Study quality 10
redical modied
mastectomy; BMI
Ozaslan et al 240 patients (stage IIII) Arm circumferences >4 cm 22/240 (9%) Axillary Incidence: level IV
(2004);96 1843 months post-surgery radiotherapy; BMI RF: level IV
Turkey Study quality 8
Paim et al 96 patients, 660 months Perimetry. Symptoms: 1 cm, plus 2 symptoms Total: 16/96 (17%) Level IV
(2008);97 Brazil post-surgery heaviness, swelling, tightness, SLNB: 2/48 (4%)
rmness ALND: 14/48 (29%)
Park et al 450 patients (stage IIII) Arm circumferences 2 cm 112/450 (25%) Stage; modied Incidence: level IV
(2008);98 1224 months post-surgery radical RF: level IV
South Korea mastectomy; Study quality 8
axillary dissection;
axillary
radiotherapy; BMI;
low exercise; low LE
education; low
preventive;
self-care activities
Querci della 201 patients (grade 13) Arm circumferences; >2 cm; >5% any site (risk >2 cm: 44/198 (22%) Dominant side; Incidence: level IV
Rovere et al >6 months post-treatment subjective factors); swelling noted by Subjective: 20/199 (10%) positive nodal RF: level IV
(2003);99 UK patient with clinical inspection status; Study quality 8
by nurse or doctor chemotherapy
Thomas- 347 patients (stage IIII) Arm circumferences; volume >2 cm dierence; >150 mL >2 cm: 39/347 (11%) Level IV
MacLean et al 612 months post-diagnosis dierence >150 mL: 31/347 (9%)
(2008);100
Canada
van der Veen et 245 patients, time not Arm circumferences 25 cm dierence 59/245 (24%) Level IV
al (2004);101 specied
Belgium
Velloso et al 45 patients, 213 months Arm circumferences >10% dierence 2/45 (4%) Level IV
(2011);102 Brazil post-surgery

Incidence data are n/N (%). ALND=axillary lymph node dissection. LE=lymphoedema. SLNB: sentinel-lymph-node biopsy. SOAC=sum of arm circumferences. SAQ: self-administered questionnaire. Tis=carcinoma
in situ. QoL=quality of life. BMI=body-mass index. *Timing of incidence reported is the timing lymphoedema was measured relative to diagnosis of breast cancer. Risk factors shown to be statistically signicant
(p005) in multivariable analyses or randomised clinical trials. Level of evidence (as per National Health and Medical Research Council guidelines): incidence level is the level of evidence for incidence
component; RF level is the level of evidence for risk factor component. Study quality: each item of a selected study, which met our criteria, was assigned one point (the highest possible score was 14the higher
the score the higher the study quality). Data selected as per the hierarchy of decisions and which are used in the meta-analysis pooled estimate.

Table 2: Studies reporting lymphoedema incidence or prevalence after breast cancer treatment and associated risk factors

diagnosis of lymphoedema was 126%, whereas the
incidence by all other methods of assessment was
between 148% (circumferences) and 204% (self-
reported swelling; table 1). The incidence of arm
lymphoedema seemed to increase with time up to 2 years
from diagnosis or surgery, after which incidence seemed
to decrease (table 1); incidence in women who had
axillary-lymph-node dissection was almost four times
higher than it was in those who had sentinel-lymph-node
biopsy (table 1, gure 2).
Of the 29 studies (involving 17 933 participants) that
investigated risk factors, two were low quality, 17 were of

www.thelancet.com/oncology Vol 14 May 2013 509

 Articles
A
Number of studies
Armer et al (2004)78
Blanchard et al (2003)80
Francis et al (2006)34
Golshan et al (2003)85
Han et al (2012)38
Kingsmore et al (2005)61
Lopez Penha et al (2011)91
Lucci et al (2007)71
Lumachi et al (2009)62
Mansel et al (2006)72
Mathew et al (2006)63
McLaughlin et al (2008)44
Ozcinar et al (2012)46
Paim et al (2008)97
Schijven et al (2003)65
Schrenk et al (2000)53
Veronesi et al (2003)74
Wernicke et al (2011)67
Total
B accumulate beyond this period but at a slower rate. Second,
Armer et al (2004)78
Blanchard et al (2003)80
Francis et al (2006)34
Golshan et al (2003)85
Han et al (2012)38
Kingsmore et al (2005)61 1099
Lopez Penha et al (2011)91
Lucci et al (2007)71
Lumachi et al (2009)62
Mansel et al (2006)72
Mathew et al (2006)63
McLaughlin et al (2008)44
Ozcinar et al (2012)46
Paim et al (2008)97
Schijven et al (2003)65
Schrenk et al (2000)53
Veronesi et al (2003)74
Wernicke et al (2011)67
Total
Figure 2: Incidence of arm lymphoedema after sentinel-lymph-node biopsy (A) and axillary-lymph-node
dissection (B)
510
income, age, education), disease (axillary radiotherapy,
Event rate (95% CI) stage, postoperative infection), side of treatment, and
9 0222 (00560579) lifestyle-associated factors (blood pressure, preventive
683 0057 (00420077) self-care activities, upper body symptoms, comorbidities,
155 0168 (01170235) hand use, education about lymphoedema).
77 0026 (00070098)
14 0033 (00020366)
416 0048 (00310073)
Discussion
76 0316 (02220428) Our ndings suggest that of the 138 million women
226 0062 (00370102) worldwide diagnosed with breast cancer every year,103
54 0037 (00090136) 295 320 (21%) will develop arm lymphoedema. These
412 0049 (00320074) ndings are based on data from prospective cohort studies,
312 0022 (00110046)
which are well suited for assessing incidence. Although
600 0052 (00370073)
80 0075 (00340157)
similar to reports from previous reviews,2,7,104 which
48 0042 (00100152) averaged the incidence results from included studies, the
180 0011 (00030043) subgroup analyses reported here provide additional
35 0014 (00010187) insight. First, the incidence of arm lymphoedema seems to
100 0005 (00000074) increase over time, at least up to 24 months after breast
108 0046 (00190106)
cancer diagnosis or surgery, although two prospective
0056 (00370085)
cohort studies25,48 have shown that cases continue to
79 0405 (03030516) whereas almost one in ve women with breast cancer
91 0341 (02510444) living in North America, Australasia, Asia, and the Middle
155 0471 (03940550) East develop secondary lymphoedema, less than one in six
48 0271 (01640412)
women in Europe, the UK, and South America develop the
83 0012 (00020081)
0063 (00500079)
disorder. Third, this high rate of arm lymphoedema is
69 0536 (04190650) about four times more likely when axillary-lymph-node
242 0107 (00740153) dissection is used compared with when the more
151 0086 (00510143) conservative sentinel-lymph-node biopsy procedure is
403 0132 (01020168) used. Finally, substantial evidence lends support to several
194 0124 (00840178)
risk factors for arm lymphoedema, including extensive
336 0164 (01280207)
138 0181 (01250254)
surgery, a high body-mass index, adjuvant therapy, and low
48 0292 (01810434) physical activitythese factors are potential targets for
213 0070 (00430114) future prevention strategies or for more eective
35 0543 (03790698) management of the disorder.
100 0120 (00690200) The accuracy of our estimate of arm lymphoedema
115 0348 (02670439)
incidence (214%) should be considered. Missing data
0199 (01350282)
can aect estimates of incidence. However, the direction
10 05 0 05 10 of the bias is unknown because we do not know whether
Decreased incidence Increased incidence
individuals who develop lymphoedema would be more
or less likely to remain in a study (eg, they might be more
likely to remain in a study because their lymphoedema is
being measured, or they might be less likely to remain in
the trial, because those who are more ill tend to drop out
moderate quality, and ten were high quality (table 3). of research studies). By contrast with prospective cohort
Risk factors that are lent support by a strong level of studies, retrospective studies rely on data recall or
evidence (consistent ndings in at least 75% of studies, information available from records collected for other
including at least two high-quality studies) include purposes, cross-sectional studies often show the
receipt of axillary-lymph-node dissection, and having a prevalence of lymphoedema only at one timepoint or
mastectomy, a greater number of lymph nodes dissected, include women at a wide range of times after treatment,
and a high body-mass index. Risk factors that are lent and clinical trials might recruit and follow-up study
support by a moderate level of evidence (consistent populations with inherent bias (eg, restrictive eligibility
ndings in at least 75% of studies, including at least one criteria or low recruitment rates decrease the likelihood
high-quality study) are the presence of metastatic lymph of participants being representative of the wider
nodes, receipt of chemotherapy or radiotherapy, and not population of patients). All these limitations would tend
participating in regular physical activity. We identied to yield lower lymphoedema incidence, as we saw in this
weak or inconclusive evidence to lend support to other study (table 1). Furthermore, the incidence of arm
risk factors: demographic characteristics (children, lymphoedema is likely to be underestimated irrespective
www.thelancet.com/oncology Vol 14 May 2013
 Articles

Strong evidence Moderate evidence Weak evidence Inconclusive

evidence
Demographics
Children in care aged 14 years* 1 (OR 02)
High income* 2 (OR 0205)
Age 7 (OR 0433)
Education* 1 (HR 15)
Disease and treatment
Axillary lymph node dissection 9 (OR 1367; RR 27; HR 2526)
Greater number of lymph nodes 6 (OR 1021; HR 12)
dissected
Mastectomy 3 (OR 2774)
Higher number of metastatic lymph 3 (OR 1128: RR 20)
nodes*
Radiotherapy 5 (OR 1738; HR 13)
Receiving chemotherapy* 5 (OR 1620; HR 1437)
Axillary radiotherapy 5 (OR 0177; HR 01; RR 27)
Higher stage of disease 1 (OR 25)
Postoperative infection 2 (OR 67325)
Treatment on the dominant side* 4 (OR 0247; RR 29)
Lifestyle and behaviours
Higher body-mass index 14 (OR 0155; HR 14; RR 55)
Did not participate in regular physical 2 (OR 2161)
activity*
Had blood pressure readings taken on 1 (OR 34)
the treated side*
Had not done preventive self-care 1 (OR 124)
activities
Presence of at least mild upper-body 1 (OR 2331)
symptoms*
Presence of comorbidities* 2 (OR 16; HR 01)
High blood pressure before breast 1 (OR 23)
cancer
Occupation requiring a high level of 1 (OR 2146)
hand use (eg, computer programmer)
Not receiving pretreatment 1 (OR 22)
education of lymphoedema

Data are number of articles (range of summary statistics). HR=hazard ratio. OR=odds ratio. RR=relative risk. *Evidence includes one level two (prospective cohort) study.
Evidence includes at least two level two (prospective cohort) studies. Evidence includes one randomised clinical trial. Evidence includes at least two randomised clinical trials.

Table 3: Summary of possible risk factors from multivariable analyses for the development of breast cancer-related lymphoedema

of study design. Lymphoedema is acute for some
patients, uctuating or chronic for others, with the
potential for new cases to present beyond 2 years post-
diagnosis.25,39 Therefore, prevalence data, especially when
frequency of assessment or length of follow-up are
restricted, are unlikely to capture all cases.
The type of method used to diagnose lymphoedema
also aects incidence data (table 1). Because each method
assesses dierent attributes of lymphoedema,2 its diag-
nostic sensitivity will depend on the lymphoedema
characteristics of the cohort being measured, with the
use of multiple methods of assessment leading to higher
incidence (28% compared with 20% for all other
individual methods; table 1). Incidence derived from only
one method of assessment, none of which took into
account the possibility of breast or trunk lymphoedema,
was included in the meta-analysis. Finally, characteristics
of the cohort assessed in the prospective cohort studies
also aect incidence data. Treatment advancessome
being associated with reduced lymphoedema risk (eg,
sentinel-node biopsy) and others being associated with
increased risk (eg, chemotherapy)dier between and
within countries (eg, regional vs urban areas; racial
disparities in treatment). Also, stage of disease or other
demographic characteristics (eg, level of health
insurance, access to health care) will dictate the need for
more invasive surgery and adjuvant therapy, irrespective
of subsequent risk of lymphoedema. Nearly half (13 of
30 studies) of the prospective cohort studies included in
the meta-analysis involved North American women, and
seven of these studies discussed the representativeness
of their sample; only two regarded their sample as

www.thelancet.com/oncology Vol 14 May 2013 511

 Articles
representative of the wider breast cancer population (in
terms of disease and treatment characteristics). Women
living in more rural and regional areas and women of
specic ethnic origins (eg, black women) are under-
represented in the data included in this meta-analysis. A
combination of the above issues could have aected the
incidence estimate for each country (table 1). Overall, we
consider our nding that 214% of women will develop
arm lymphoedema after breast cancer to be a conservative
estimate.
The call to integrate prospective surveillance of lymph-
oedema into standard breast-cancer care is building
momentum,105107 in part because early detection of
lymphoedema is believed to have been associated with
more eective treatment and prevention of progression.
Unfortunately, little scientic evidence exists to lend
support to this idea. Although a prospective cohort study
has shown that treating lymphoedema with compression
led to its successful resolution, the ecacy of com-
pression was not compared with a control group.54 Also,
ndings from several prospective cohort studies show
that lymphoedema, at least during the rst 1824 months
after breast cancer, might dissipate with or without
treatment in some patients.5,25,39 These results draw
attention to the risk of giving unnecessary lymphoedema
treatment, which would come at individual and public
health cost. Nonetheless, the physical, psychosocial, and
nancial burden associated with lymphoedema is well
established,66,108 lymphoedema is more likely to be under-
diagnosed than over-diagnosed (with the exception of
lymphoscintigraphy incidence, incidence was lowest
when reliant on a clinical diagnosis; table 1), and it is
likely that with more research, early detection, and
management of the disorder will prove cost-eective.109
Therefore, ndings from this meta-analysis could be
used to guide the methods (ie, measurement technique,
denition of lymphoedema, timing of assessments) of
prospective surveillance of lymphoedema in clinical
practice, to optimise early diagnosis, and to minimise the
potential for overtreatment.
Because most patients seem to present within the rst
2 years after breast cancer, more frequent surveillance
throughout this time (eg, once every 36 months) seems
reasonable. The sensitivity of diagnostic methods could
be considered when deciding which method to use and
when. Bioimpedance spectroscopy has been shown to be
sensitive to early extracellular uid changes, is fast,
reliable, and easy to administer,27 with testing being
reimbursed by some private health insurers in the
USA.109 The circumferences method is inexpensive (little
equipment needed, although it requires sta time),
reliable when used by trained assessors, and might be
the most appropriate method in the long term because it
detects size change and inter-limb size dierences
irrespective of the tissue composition of the lymph-
oedema. However, it has little sensitivity to detect
preclinical lymphoedema and needs to take into account
512
bodyweight changes over time by assessing relative
change or inter-limb dierences rather than absolute
scores.2 Self-report methods are also inexpensive, and,
unlike objective methods, can capture other symptoms
alongside arm symptoms potentially indicate lymph-
oedema over an extended period of time. However, used
on their own, self-report methods lack specicity, because
many symptoms associated with lymphoedema are also
common after breast cancer in women without
lymphoedema.11
In view of the pros and cons of the various methods, it
seems more reasonable to expect the use of multiple
methods over the use of any one method as being the
gold standard in the diagnosis of lymphoedema. However,
use of multiple measures might not be feasible in the
clinical setting. Instead, clinics could rely on specic
methods during specic periods. For example,
bioimpedance spectroscopy could be used during the rst
612 months, followed by circumferences up to 2 years
after treatment. When there is evidence of lymphoedema,
additional methods could then be used to assist in a
denitive diagnosis. Alternatively, a second measure
could be scheduled (outside the standard prospective
surveillance protocol), and only when lymphoedema is
conrmed at this additional visit (or shows signs of
further progression later), is treatment initiated. The
potential advantages of being able to change the methods
during the prospective surveillance period, for assessment
or conrmation of a diagnosis, would be best realised
when multiple methods are used to document patients
status before treatment. However, even in this ideal
situation, pretreatment assessment might not accurately
show baseline lymphoedema status; swelling can present
before treatment and the contribution of the breast cancer
disease to lymphoedema remains unknown.
One potential way to optimise the cost-eectiveness
of prospective surveillance is to stratify women into a
more or less frequent and comprehensive surveillance
protocol based on presence of lymphoedema risk factors.
Unfortunately, due to insucient raw data being re-
ported in the included risk factor studies, we were unable
to report a common statistic (notably risk ratios) for each
of the potential risk factors. Although the absence of risk
ratios is a limitation of this work, our ndings
nonetheless clearly indicate that there is strong evidence
showing that more extensive surgery (chest wall and
axilla) and being overweight or obese are associated with
increased risk of lymphoedema, and moderate evidence
that also lends support to adjuvant therapy (radiation and
chemotherapy) and sedentary lifestyles as risk factors.
Our detection of chemotherapy as a risk factor is in
contrast with the ndings of a previous meta-analysis on
breast cancer-related treatment risk factors.10 A possible
reason that we detected chemotherapy as a risk factor is
because it is a surrogate measure for more advanced
disease and thus more extensive surgical treatment, or it
might be indicative of an adverse interaction between
www.thelancet.com/oncology Vol 14 May 2013

more chemotherapeutic agents available in the past
decade and the lymph system. The detection of sedentary
lifestyle as a risk factor is, to our knowledge, a new
nding and could be as a consequence of a major lymph
ow mechanism (the muscle pump) being inactive.
There are several factors to be considered before
our risk-factor ndings are used for risk-stratication
purposes. First, these risk factors alone do not accurately
predict who will develop arm lymphoedema and who will
not. These specic factors represent six of 12 identied
risk factors in one, prospective, population-based cohort
study, which together explained only 35% of variation
between those who did and those who did not develop
lymphoedema.39 Furthermore, women can and do
present with arm lymphoedema despite having healthy
bodyweight, regular physical activity levels, and
lumpectomy surgery. Finally, little is known about the
potential contribution of the cancer itself, or the possible
importance of genetic predisposition, to lymphoedema
risk. Findings from animal models110 and one study in
human beings111 provide preliminary evidence for a
contribution of genetic susceptibility to the development
of secondary lymphoedema after breast cancer. None-
theless, current understanding of risk factors can inform
lymphoedema prevention and management strategies.
Although treatment-related risk factors are largely not
modiable (because they are generally dictated by the
type and stage of disease and available treatment
options), substantial scope exists for the modication of
patients physical activity levels and bodyweight after
breast cancer: most women with breast cancer are
insuciently active at diagnosis112114 and more than 50%
are overweight or obese.115
Further research is needed to improve our under-
standing of risk factors (including further exploration of
known risk factors, such as whether lymphedoema rates
dier between those with delayed axillary-lymph-node
dissection after sentinel-node biopsy compared with
axillary-lymph-node dissection from the outset), as well
as prevention and treatment strategies. Only with
increased knowledge will we be in a position to improve
further the lives of women with breast cancer, and reduce
the overall socioeconomic burden of this disabling,
distressing disorder.
Contributors
TD, BN, and SH participated in the conception and design of the review.
TD and SR participated in the extraction and analysis of data. TD, SR,
BN, and SH participated in the interpretation of data and writing of the
paper. All authors approved the nal version.
Conicts of interest
We declare that we have no conicts of interest.
References
1 Cheville AL, McGarvey CL, Petrek JA, Russo SA, Thiadens SR,
Taylor ME. The grading of lymphedema in oncology clinical trials.
Semin Radiat Oncol 2003; 13: 21425.
2 Hayes SC, Johansson K, Stout NL, et al. Upper-body morbidity after
breast cancer: incidence and evidence for evaluation, prevention,
and management within a prospective surveillance model of care.
Cancer 2012; 118 (suppl): 223749.
www.thelancet.com/oncology Vol 14 May 2013
Articles
3 Maunsell E, Brisson J, Deschnes L. Arm problems and
psychological distress after surgery for breast cancer. Can J Surg
1993; 36: 31520.
4 Hayes S, Di Sipio T, Rye S, et al. Prevalence and prognostic
signicance of secondary lymphedema following breast cancer.
Lymphat Res Biol 2011; 9: 13541.
5 Armer JM, Stewart BR. Post-breast cancer lymphedema: incidence
increases from 12 to 30 to 60 months. Lymphology 2010; 43: 11827.
6 Hayes S, Janda M, Cornish B, Battistutta D, Newman B.
Lymphedema secondary to breast cancer: how choice of measure
inuences diagnosis, prevalence, and identiable risk factors.
Lymphology 2008; 41: 1828.
7 Erickson VS, Pearson ML, Ganz PA, Adams J, Kahn KL. Arm
edema in breast cancer patients. J Natl Cancer Inst 2001; 93: 96111.
8 Lee TS, Kilbreath SL, Refshauge KM, Herbert RD, Beith JM.
Prognosis of the upper limb following surgery and radiation for
breast cancer. Breast Cancer Res Treat 2008; 110: 1937.
9 Petrek JA, Heelan MC. Incidence of breast carcinoma-related
lymphedema. Cancer 1998; 83 (Suppl American): 277681.
10 Tsai RJ, Dennis LK, Lynch CF, Snetselaar LG, Zamba GK,
Scott-Conner C. The risk of developing arm lymphedema among
breast cancer survivors: a meta-analysis of treatment factors.
Ann Surg Oncol 2009; 16: 195972.
11 Hayes SC, Rye S, Battistutta D, Newman B. Prevalence of upper-body
symptoms following breast cancer and its relationship with upper-
body function and lymphedema. Lymphology 2010; 43: 17887.
12 Sagen A, Kresen R, Risberg MA. Physical activity for the aected
limb and arm lymphedema after breast cancer surgery. A
prospective, randomized controlled trial with two years follow-up.
Acta Oncol 2009; 48: 110210.
13 Torres Lacomba M, Yuste Snchez MJ, Zapico Goi A, et al.
Eectiveness of early physiotherapy to prevent lymphoedema after
surgery for breast cancer: randomised, single blinded, clinical trial.
BMJ 2010; 340: b5396.
14 NHMRC. NHMRC levels of evidence and grades for
recommendations for developers of guidelines; Canberra: National
Health and Medical Research Council, 2009.
15 Egger M, Davey Smith G, Schneider M, Minder C. Bias in
meta-analysis detected by a simple, graphical test. BMJ 1997;
315: 62934.
16 Duval S, Tweedie R. Trim and ll: a simple funnel-plot-based
method of testing and adjusting for publication bias in
meta-analysis. Biometrics 2000; 56: 45563.
17 Scholten-Peeters GG, Verhagen AP, Bekkering GE, et al. Prognostic
factors of whiplash-associated disorders: a systematic review of
prospective cohort studies. Pain 2003; 104: 30322.
18 Borghouts JA, Koes BW, Bouter LM. The clinical course and
prognostic factors of non-specic neck pain: a systematic review.
Pain 1998; 77: 113.
19 Kuijpers T, van der Windt DA, van der Heijden GJ, Bouter LM.
Systematic review of prognostic cohort studies on shoulder
disorders. Pain 2004; 109: 42031.
20 Mols F, Vingerhoets AJ, Coebergh JW, van de Poll-Franse LV.
Quality of life among long-term breast cancer survivors:
a systematic review. Eur J Cancer 2005; 41: 261319.
21 Cormier JN, Askew RL, Mungovan KS, Xing Y, Ross MI, Armer JM.
Lymphedema beyond breast cancer: a systematic review and
meta-analysis of cancer-related secondary lymphedema. Cancer
2010; 116: 513849.
22 Schmidt CO, Kohlmann T. When to use the odds ratio or the
relative risk? Int J Public Health 2008; 53: 16567.
23 Davies HT, Crombie IK, Tavakoli M. When can odds ratios mislead?
BMJ 1998; 316: 98991.
24 Clark B, Sitzia J, Harlow W. Incidence and risk of arm oedema
following treatment for breast cancer: a three-year follow-up study.
Q JM 2005; 98: 34348.
25 Norman SA, Localio AR, Potashnik SL, et al. Lymphedema in breast
cancer survivors: incidence, degree, time course, treatment, and
symptoms. J Clin Oncol 2009; 27: 39097.
26 Langbecker D, Hayes SC, Newman B, Janda M. Treatment for
upper-limb and lower-limb lymphedema by professionals
specializing in lymphedema care. Eur J Cancer Care (Engl) 2008;
17: 55764.
513
 Articles
27 Cornish BH, Chapman M, Hirst C, et al. Early diagnosis of
lymphedema using multiple frequency bioimpedance. Lymphology
2001; 34: 211.
28 Borenstein M, Hedges L, Higgins J, Rothstein H. Comprehensive
Meta-analysis Version 2, Biostat. Englewood NJ; 2005.
29 Bennett Britton TM, Buczacki SJA, Turner CL, Vowler SL, Pain SJ,
Purushotham AD. Venous changes and lymphoedema 4 years after
axillary surgery for breast cancer. Br J Surg 2007; 94: 83334.
30 Bland KL, Perczyk R, Du W, et al. Can a practicing surgeon detect
early lymphedema reliably? Am J Surg 2003; 186: 50913.
31 Celebioglu F, Perbeck L, Frisell J, Grndal E, Svensson L,
Danielsson R. Lymph drainage studied by lymphoscintigraphy in
the arms after sentinel node biopsy compared with axillary lymph
node dissection following conservative breast cancer surgery.
Acta Radiol 2007; 48: 48895.
32 Clough-Gorr KM, Ganz PA, Silliman RA. Older breast cancer
survivors: factors associated with self-reported symptoms of
persistent lymphedema over 7 years of follow-up. Breast J 2010;
16: 14755.
33 Devoogdt N, Van Kampen M, Christiaens MR, et al. Short- and
long-term recovery of upper limb function after axillary lymph node
dissection. Eur J Cancer Care (Engl) 2011; 20: 7786.
34 Francis WP, Abghari P, Du W, Rymal C, Suna M, Kosir MA.
Improving surgical outcomes: standardizing the reporting of
incidence and severity of acute lymphedema after sentinel lymph
node biopsy and axillary lymph node dissection. Am J Surg 2006;
192: 63639.
35 Geller BM, Vacek PM, OBrien P, Secker-Walker RH. Factors
associated with arm swelling after breast cancer surgery.
J Womens Health (Larchmt) 2003; 12: 92130.
36 Goldberg JI, Wiechmann LI, Riedel ER, Morrow M, Van Zee KJ.
Morbidity of sentinel node biopsy in breast cancer: the relationship
between the number of excised lymph nodes and lymphedema.
Ann Surg Oncol 2010; 17: 327886.
37 Goldberg JI, Riedel ER, Morrow M, Van Zee KJ. Morbidity of
sentinel node biopsy: relationship between number of excised
lymph nodes and patient perceptions of lymphedema.
Ann Surg Oncol 2011; 18: 286672.
38 Han JW, Seo YJ, Choi JE, Kang SH, Bae YK, Lee SJ. The ecacy of
arm node preserving surgery using axillary reverse mapping for
preventing lymphedema in patients with breast cancer.
J Breast Cancer 2012; 15: 9197.
39 Hayes SC, Janda M, Cornish B, Battistutta D, Newman B.
Lymphedema after breast cancer: incidence, risk factors, and eect
on upper body function. J Clin Oncol 2008; 26: 353642.
40 Helyer LK, Varnic M, Le LW, Leong W, McCready D. Obesity is a
risk factor for developing postoperative lymphedema in breast
cancer patients. Breast J 2010; 16: 4854.
41 Herd-Smith A, Russo A, Muraca MG, Del Turco MR, Cardona G.
Prognostic factors for lymphedema after primary treatment of
breast carcinoma. Cancer 2001; 92: 178387.
42 Krki A, Simonen R, Mlki E, Selfe J. Impairments, activity
limitations and participation restrictions 6 and 12 months after
breast cancer operation. J Rehabil Med 2005; 37: 18088.
43 Kosir MA, Rymal C, Koppolu P, et al. Surgical outcomes after breast
cancer surgery: measuring acute lymphedema. J Surg Res 2001;
95: 14751.
44 McLaughlin SA, Wright MJ, Morris KT, et al. Prevalence of
lymphedema in women with breast cancer 5 years after sentinel
lymph node biopsy or axillary dissection: patient perceptions and
precautionary behaviors. J Clin Oncol 2008; 26: 522026.
45 Norman SA, Localio AR, Kallan MJ, et al. Risk factors for
lymphedema after breast cancer treatment.
Cancer Epidemiol Biomarkers Prev 2010; 19: 273446.
46 Ozcinar B, Guler SA, Kocaman N, Ozkan M, Gulluoglu BM,
Ozmen V. Breast cancer related lymphedema in patients with
dierent loco-regional treatments. Breast 2012; 21: 36165.
47 Pain SJ, Vowler S, Purushotham AD. Axillary vein abnormalities
contribute to development of lymphoedema after surgery for breast
cancer. Br J Surg 2005; 92: 31115.
48 Paskett ED, Naughton MJ, McCoy TP, Case LD, Abbott JM.
The epidemiology of arm and hand swelling in premenopausal
breast cancer survivors. Cancer Epidemiol Biomarkers Prev 2007;
16: 77582.
514
49 Petrek JA, Senie RT, Peters M, Rosen PP. Lymphedema in a cohort
of breast carcinoma survivors 20 years after diagnosis. Cancer 2001;
92: 136877.
50 Quinlan E, Thomas-MacLean R, Hack T, et al. The impact of breast
cancer among Canadian women: disability and productivity. Work
2009; 34: 28596.
51 Rampaul RS, Mullinger K, Macmillan RD, et al. Incidence of clinically
signicant lymphoedema as a complication following surgery for
primary operable breast cancer. Eur J Cancer 2003; 39: 216567.
52 Rnk R, von Smitten K, Tasmuth T, Leidenius M. One-year
morbidity after sentinel node biopsy and breast surgery. Breast
2005; 14: 2836.
53 Schrenk P, Rieger R, Shamiyeh A, Wayand W. Morbidity following
sentinel lymph node biopsy versus axillary lymph node dissection
for patients with breast carcinoma. Cancer 2000; 88: 60814.
54 Stout Gergich NL, Pfalzer LA, McGarvey C, Springer B, Gerber LH,
Soballe P. Preoperative assessment enables the early diagnosis and
successful treatment of lymphedema. Cancer 2008; 112: 280919.
55 Wilke LG, McCall LM, Posther KE, et al. Surgical complications
associated with sentinel lymph node biopsy: results from a
prospective international cooperative group trial. Ann Surg Oncol
2006; 13: 491500.
56 Wojcinski S, Nuengsri S, Hillemanns P, et al. Axillary dissection in
primary breast cancer: variations of the surgical technique and
inuence on morbidity. Cancer Manag Res 2012; 4: 12127.
57 Yang EJ, Park WB, Seo KS, Kim SW, Heo CY, Lim JY. Longitudinal
change of treatment-related upper limb dysfunction and its impact
on late dysfunction in breast cancer survivors: a prospective cohort
study. J Surg Oncol 2010; 101: 8491.
58 Crosby MA, Card A, Liu J, Lindstrom WA, Chang DW. Immediate
breast reconstruction and lymphedema incidence.
Plast Reconstr Surg 2012; 129: 789e95e.
59 Hayes SB, Freedman GM, Li T, Anderson PR, Ross E. Does axillary
boost increase lymphedema compared with supraclavicular
radiation alone after breast conservation?
Int J Radiat Oncol Biol Phys 2008; 72: 144955.
60 Hinrichs CS, Watroba NL, Rezaishiraz H, et al. Lymphedema
secondary to postmastectomy radiation: incidence and risk factors.
Ann Surg Oncol 2004; 11: 57380.
61 Kingsmore DB, Hole DJ, Gillis CR, George WD. Axillary recurrence
in breast cancer. Eur J Surg Oncol 2005; 31: 22631.
62 Lumachi F, Basso SM, Bonamini M, et al. Incidence of arm
lymphoedema following sentinel node biopsy, axillary sampling and
axillary dissection in patients with breast cancer. In Vivo 2009;
23: 101720.
63 Mathew J, Barthelmes L, Neminathan S, Crawford D. Comparative
study of lymphoedema with axillary node dissection versus axillary
node sampling with radiotherapy in patients undergoing breast
conservation surgery. Eur J Surg Oncol 2006; 32: 72932.
64 Powell S, Taghian A, Kachnic L, Coen J, Assaad S. Risk of
lymphedema after regional nodal irradiation with breast
conservation therapy. Int J Radiat Oncol 2003; 55: 120915.
65 Schijven MP, Vingerhoets AJJM, Rutten HJT, et al. Comparison of
morbidity between axillary lymph node dissection and sentinel
node biopsy. Eur J Surg Oncol 2003; 29: 34150.
66 Shih YC, Xu Y, Cormier JN, et al. Incidence, treatment costs, and
complications of lymphedema after breast cancer among women of
working age: a 2-year follow-up study. J Clin Oncol 2009; 27: 200714.
67 Wernicke AG, Goodman RL, Turner BC, et al. A 10-year follow-up of
treatment outcomes in patients with early stage breast cancer and
clinically negative axillary nodes treated with tangential breast
irradiation following sentinel lymph node dissection or axillary
clearance. Breast Cancer Res Treat 2011; 125: 893902.
68 Ahmed RL, Thomas W, Yee D, Schmitz KH. Randomized controlled
trial of weight training and lymphedema in breast cancer survivors.
J Clin Oncol 2006; 24: 276572.
69 Box RC, Reul-Hirche HM, Bullock-Saxton JE, Furnival CM.
Physiotherapy after breast cancer surgery: results of a randomised
controlled study to minimise lymphoedema. Breast Cancer Res Treat
2002; 75: 5164.
70 Lee TS, Kilbreath SL, Refshauge KM, Pendlebury SC, Beith JM,
Lee MJ. Pectoral stretching program for women undergoing
radiotherapy for breast cancer. Breast Cancer Res Treat 2007;
102: 31321.
www.thelancet.com/oncology Vol 14 May 2013

71 Lucci A, McCall LM, Beitsch PD, et al, and the American College of
Surgeons Oncology Group. Surgical complications associated with
sentinel lymph node dissection (SLND) plus axillary lymph node
dissection compared with SLND alone in the American College of
Surgeons Oncology Group Trial Z0011. J Clin Oncol 2007; 25: 365763.
72 Mansel RE, Falloweld L, Kissin M, et al. Randomized multicenter
trial of sentinel node biopsy versus standard axillary treatment in
operable breast cancer: the ALMANAC Trial. J Natl Cancer Inst
2006; 98: 599609.
73 Schmitz KH, Ahmed RL, Troxel AB, et al. Weight lifting for women
at risk for breast cancer-related lymphedema: a randomized trial.
JAMA 2010; 304: 2699705.
74 Veronesi U, Paganelli G, Viale G, et al. A randomized comparison
of sentinel-node biopsy with routine axillary dissection in breast
cancer. N Engl J Med 2003; 349: 54653.
75 Soran A, DAngelo G, Begovic M, et al. Breast cancer-related
lymphedemawhat are the signicant predictors and how they
aect the severity of lymphedema? Breast J 2006; 12: 53643.
76 Swenson KK, Nissen MJ, Leach JW, Post-White J. Case-control
study to evaluate predictors of lymphedema after breast cancer
surgery. Oncol Nurs Forum 2009; 36: 18593.
77 Ahmed RL, Prizment A, Lazovich D, Schmitz KH, Folsom AR.
Lymphedema and quality of life in breast cancer survivors: the Iowa
Womens Health Study. J Clin Oncol 2008; 26: 568996.
78 Armer J, Fu MR, Wainstock JM, Zagar E, Jacobs LK. Lymphedema
following breast cancer treatment, including sentinel lymph node
biopsy. Lymphology 2004; 37: 7391.
79 Asim M, Cham A, Banerjee S, et al. Diculties with dening
lymphoedema after axillary dissection for breast cancer. N Z Med J
2012; 125: 2939.
80 Blanchard DK, Donohue JH, Reynolds C, Grant CS. Relapse and
morbidity in patients undergoing sentinel lymph node biopsy alone
or with axillary dissection for breast cancer. Arch Surg 2003;
138: 48287, discussion 48788.
81 Deo SV, Ray S, Rath GK, et al. Prevalence and risk factors for
development of lymphedema following breast cancer treatment.
Indian J Cancer 2004; 41: 812.
82 Edwards TL. Prevalence and aetiology of lymphoedema after breast
cancer treatment in southern Tasmania. Aust N Z J Surg 2000;
70: 41218.
83 Eversley R, Estrin D, Dibble S, Wardlaw L, Pedrosa M,
Favila-Penney W. Post-treatment symptoms among ethnic minority
breast cancer survivors. Oncol Nurs Forum 2005; 32: 25056.
84 Freitas-Silva R, Conde DM, de Freitas-Jnior R, Martinez EZ.
Comparison of quality of life, satisfaction with surgery and
shoulder-arm morbidity in breast cancer survivors submitted to
breast-conserving therapy or mastectomy followed by immediate
breast reconstruction. Clinics (Sao Paulo) 2010; 65: 78187.
85 Golshan M, Martin WJ, Dowlatshahi K. Sentinel lymph node biopsy
lowers the rate of lymphedema when compared with standard axillary
lymph node dissection. Am Surg 2003; 69: 20911, discussion 212.
86 Graham P, Jagavkar R, Browne L, Millar E. Supraclavicular
radiotherapy must be limited laterally by the coracoid to avoid
signicant adjuvant breast nodal radiotherapy lymphoedema risk.
Australas Radiol 2006; 50: 57882.
87 Haddad P, Farzin M, Amouzegar-Hashemi F, et al. A multicentre
cross-sectional study of arm lymphedema four or more years after
breast cancer treatment in Iranian patients. Breast Cancer 2010;
17: 28185.
88 Hayes S, Cornish B, Newman B. Comparison of methods to
diagnose lymphoedema among breast cancer survivors: 6-month
follow-up. Breast Cancer Res Treat 2005; 89: 22126.
89 Kwan W, Jackson J, Weir LM, Dingee C, McGregor G, Olivotto IA.
Chronic arm morbidity after curative breast cancer treatment:
prevalence and impact on quality of life. J Clin Oncol 2002; 20: 424248.
90 Leidenius M, Leivonen M, Vironen J, von Smitten K. The
consequences of long-time arm morbidity in node-negative breast
cancer patients with sentinel node biopsy or axillary clearance.
J Surg Oncol 2005; 92: 2331.
91 Lopez Penha TR, Slangen JJ, Heuts EM, Voogd AC,
Von Meyenfeldt MF. Prevalence of lymphoedema more than ve years
after breast cancer treatment. Eur J Surg Oncol 2011; 37: 105963.
www.thelancet.com/oncology Vol 14 May 2013
Articles
92 McCredie MRE, Dite GS, Porter L, et al. Prevalence of self-reported
arm morbidity following treatment for breast cancer in the
Australian Breast Cancer Family Study. Breast 2001; 10: 51522.
93 Meeske KA, Sullivan-Halley J, Smith AW, et al. Risk factors for arm
lymphedema following breast cancer diagnosis in Black women and
White women. Breast Cancer Res Treat 2009; 113: 38391.
94 Nagel PH, Bruggink ED, Wobbes T, Strobbe LJ. Arm morbidity after
complete axillary lymph node dissection for breast cancer.
Acta Chir Belg 2003; 103: 21216.
95 Nesvold IL, Dahl AA, Lkkevik E, Marit Mengshoel A, Foss SD.
Arm and shoulder morbidity in breast cancer patients after
breast-conserving therapy versus mastectomy. Acta Oncol 2008;
47: 83542.
96 Ozaslan C, Kuru B. Lymphedema after treatment of breast cancer.
Am J Surg 2004; 187: 6972.
97 Paim CR, de Paula Lima ED, Fu MR, de Paula Lima A, Cassali GD.
Post lymphadenectomy complications and quality of life among breast
cancer patients in Brazil. Cancer Nurs 2008; 31: 30209, quiz 31011.
98 Park JH, Lee WH, Chung HS. Incidence and risk factors of breast
cancer lymphoedema. J Clin Nurs 2008; 17: 145059.
99 Querci della Rovere G, Ahmad I, Singh P, Ashley S, Daniels IR,
Mortimer P. An audit of the incidence of arm lymphoedema after
prophylactic level I/II axillary dissection without division of the
pectoralis minor muscle. Ann R Coll Surg Engl 2003; 85: 15861.
100 Thomas-Maclean RL, Hack T, Kwan W, Towers A, Miedema B,
Tilley A. Arm morbidity and disability after breast cancer: new
directions for care. Oncol Nurs Forum 2008; 35: 6571.
101 van der Veen P, De Voogdt N, Lievens P, Duquet W, Lamote J,
Sacre R. Lymphedema development following breast cancer surgery
with full axillary resection. Lymphology 2004; 37: 20608.
102 Velloso FS, Barra AA, Dias RC. Functional performance of upper
limb and quality of life after sentinel lymph node biopsy of breast
cancer. Rev Bras Fisioter 2011; 15: 14653.
103 Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global
cancer statistics. CA Cancer J Clin 2011; 61: 6990.
104 Hayes S. Review of Research Evidence on Secondary
Lymphoedema: incidence, prevention, risk factors and treatment.
National Breast and Ovarian Cancer Centre, 2008. http://
canceraustralia.gov.au/sites/default/les/publications/slerw-review-
research-secondary-lymphoedema_504af03e2a20c.pdf (accessedFeb
15, 2011).
105 Stout NL, Binkley JM, Schmitz KH, et al. A prospective surveillance
model for rehabilitation for women with breast cancer. Cancer 2012;
118 (Suppl): 2191200.
106 International Lymphoedema Framework. Lymphoedema Framework:
Best Practice for the Management of Lymphoedema. International
Consensus. London; 2006. http://www.mepltd.co.uk/oneos.html
(accessed April 21, 2011).
107 Australasian Lymphology Association. Position Statement:
monitoring for the early detection of breast cancer-related
lymphoedema. Surrey Hills, Victoria: ALA; 2012.
108 Rockson SG. Lymphedema. Am J Med 2001; 110: 28895.
109 Cheville AL, Nyman JA, Pruthi S, Basford JR. Cost considerations
regarding the prospective surveillance model for breast cancer
survivors. Cancer 2012; 118 (suppl): 232530.
110 Tammela T, Alitalo K. Lymphangiogenesis: molecular mechanisms
and future promise. Cell 2010; 140: 46076.
111 Newman B, Lose F, Kedda MA, et al. Possible genetic predisposition
to lymphedema after breast cancer. Lymphat Res Biol 2012; 10: 213.
112 Eakin EG, Youlden DR, Baade PD, et al. Health behaviors of cancer
survivors: data from an Australian population-based survey.
Cancer Causes Control 2007; 18: 88194.
113 Harrison S, Hayes SC, Newman B. Level of physical activity and
characteristics associated with change following breast cancer
diagnosis and treatment. Psychooncology 2009; 18: 38794.
114 Smith SG, Chagpar AB. Adherence to physical activity guidelines in
breast cancer survivors. Am Surg 2010; 76: 96265.
115 Demark-Wahnefried W, Campbell KL, Hayes SC. Weight
management and its role in breast cancer rehabilitation. Cancer
2012; 118 (suppl): 227787.
515