European Journal of Nutrition

https://doi.org/10.1007/s00394-018-1863-2

ORIGINAL CONTRIBUTION

Consumption of a dark roast coffee blend reduces DNA damage

in humans: results from a 4-week randomised controlled study
Dorothea Schipp4   · Jana Tulinska1 · Maria Sustrova1 · Aurelia Liskova1 · Viera Spustova1 ·
Miroslava Lehotska Mikusova1 · Zora Krivosikova1 · Katarina Rausova2 · Andrew Collins3 · Vaineta Vebraite3 ·
Katarina Volkovova3 · Eva Rollerova2 · Magdalena Barancokova1 · Sergey Shaposhnikov3

Received: 15 July 2018 / Accepted: 14 November 2018

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Abstract
Purpose  To determine the DNA protective effects of a standard coffee beverage in comparison to water consumption.
Methods  The single-blind, randomised controlled study with parallel design included healthy women (n = 50) and men
(n = 50) recruited from the general Central European population. The subjects were randomised in a coffee and a control
group, with stratification for sex and body mass index. The study comprised two periods of 4 weeks: a preconditioning
period, with daily consumption of at least 500 ml water but no coffee, nor tea, nor any other caffeine-containing product.
During the subsequent intervention period the coffee group consumed 500 ml of freshly brewed dark roast coffee blend per
day, the control group consumed water instead. On the last day of each period, blood was drawn and analysed by comet assay
(single-cell gel electrophoresis) to assess the level of DNA damage (strand breakage).
Results  At the end of the intervention period the mean level of DNA strand breaks in the coffee group has decreased in
comparison to the control group [difference in means 0.23% TI (tail intensity), p = 0.028]. The mean change from baseline
(delta value) was − 23% in the coffee group (p = 0.0012). Effects of coffee intake were similar for men and women. During
intervention, neither group showed any significant change in body weight or calorie intake.
Conclusions  Our results indicate that regular consumption of a dark roast coffee blend has a beneficial protective effect on
human DNA integrity in both, men and women.

Keywords  Coffee · Comet assay · Human intervention study · DNA strand breaks · DNA damage

Abbreviations WRST Wilcoxon rank sum test

BMI Body mass index WSRT Wilcoxon signed rank test
CQA Caffeoylquinic acid
ITT Intention to treat
NMP N-Methylpyridinium Introduction
PP Per protocol
SD Standard deviation It is known that a range of disorders are associated with an
TI Tail intensity impairment of cell DNA integrity due to increased damage
to DNA. Among these disorders, there are certain cancers,
cardiovascular diseases, diabetes and inflammatory bowel
* Dorothea Schipp diseases. Numerous epidemiological and experimental
kontakt@ds‑statistik.de data suggest that coffee consumption is associated with
1 a reduced risk of developing these diseases [1–6]. Previ-
Medical Faculty, Slovak Medical University, Bratislava,
Slovakia ous investigations have shown DNA-protective effects of
2 coffee constituents with different chemical compositions
Faculty of Public Health, Slovak Medical University,
Bratislava, Slovakia [7–10]. It was assumed that a probable mechanism respon-
3 sible for protection of DNA is the induction of a resistant
Norgenotech AS, Skreia, Norway
state towards toxic compounds by pre-incubation with cof-
4
Statistical Consulting, Pirnaer Str. 1, fee constituents. For instance, several coffee constituents
01824 Rosenthal‑Bielatal, Germany

13
Vol.:(0123456789)

have antioxidant properties that would protect against
DNA oxidative damage. Also, nuclear translocation of
Nrf2 is induced which binds to antioxidant or electrophile
response elements (ARE/EpRE) located in the promoters
of several cell defence genes, such as those coding for
Phase II detoxifying enzymes (glutathione-S-transferase,
haeme oxygenase and others). As a consequence, there is
increased expression of a large number of such protec-
tive enzymes [11]. When comparing caffeic acid, catechol,
hydroxyhydroquinone, trigonelline and the alkylpyri-
dinium compounds (found as degradation products of
trigonelline during roasting), by far the strongest Nrf2-
inducing effects were noted for the alkylpyridinium com-
pounds N-methylpyridinium (NMP) (significant effects
at 0.1 µmol/l) and 1,2-dimethyl pyridinium (significant
effects at 1 fmol/l) [11].
In a previous, single arm human intervention study with
33 healthy males, a significant decrease in DNA damage
was observed in white blood cells after 4 weeks of daily
consumption of a dark roast coffee blend [12]. In another
uncontrolled intervention study of two different coffee
blends with a cross-over design, a protective effect on DNA
integrity (level of strand breakage) in blood cells of 84 vol-
unteers (males and females) was observed after consumption
of either coffee blend. Since the design of that study did
not include a water consuming control group, the results
were based on a comparison of the levels of DNA strand
breaks after treatment to those prior to treatment, i.e. after
the coffee-free preconditioning phase. Because statistical
analyses indicated a carry-over effect from first to cross-
over treatment period, the latter was not analysed in detail
[13]. These preliminary findings were verified, by testing a
dark roast coffee blend also in a randomised controlled study
with 90 male volunteers [14].
The current parallel-design human intervention trial was
carried out in order to investigate whether consumption of
a similar coffee type, at a lower daily dose, would mediate
protection against DNA damage in another European region
Fig. 1  Study design
13
European Journal of Nutrition
(Slovakia instead of southwestern Germany). Further, the
trial included women in addition to men.
Materials and methods
Study design and study population
The study was a single-blind, randomised controlled, inter-
vention trial, with parallel design. The study included 50
healthy men and 50 healthy women, recruited from the
general Central European population (Fig. 1). Volunteers
responded to local announcements of the study via info
sheets, flyers and posters. Participants were randomly allo-
cated to two groups, coffee and control, each containing 25
men and 25 women. To avoid an imbalance of body mass
index (BMI) distribution between groups, a block randomi-
sation procedure was employed. Participants were divided
into two blocks according to their BMI (< 25 kg/m2 and
≥ 25 kg/m2). Subsequently, participants of each block were
randomly assigned to control and coffee group, respec-
tively. Inclusion criteria for volunteer recruitment were: age
19–50 years, BMI 19–32 kg/m2, habitual coffee drinkers (by
self-assessment), healthy non-smokers, non-vegetarians.
The exclusion criteria were high alcohol consumption (by
self-assessment), high-performance sport (by self-assess-
ment), pregnancy, presence of metabolic disorders or other
diseases, as well as taking pharmaceutical drugs and food
supplements. After informed written consent, the volunteers
were subjected to a standard medical health check including
measurements of body weight and height. The study was
approved by the local Ethics Committee at the Slovak Medi-
cal University in Bratislava (No. E-46/16).
The study consisted of two periods, preconditioning and
intervention (4 weeks preconditioning, 4 weeks interven-
tion). During the preconditioning period both groups were
not allowed to drink any coffee or black/green tea nor any
caffeine-containing beverages. During the intervention
European Journal of Nutrition
period, participants of the coffee group consumed daily four
cups (in total 500 ml) of the study coffee; each cup (125 ml)
was freshly prepared. Participants of the control group con-
sumed cups of warm water instead (also 4 × 125 ml). Dietary
restrictions (avoidance of caffeine containing foods, supple-
ments, polyphenol-rich foods such as berries or cocoa, etc.)
applied during both study periods.
A total of 110 volunteers were assessed for eligibility.
Five did not meet inclusion criteria, 2 declined to partici-
pate before entering the study, and 3 had to be excluded for
other reasons. The remaining 100 volunteers were allocated
to control and coffee intervention group. In the end, 99 vol-
unteers completed the study in total. One person allocated
to the test group discontinued the study early due to acute
illness (not related to coffee consumption). One person allo-
cated to the control group showed no compliance for coffee
abstention at visit 2.
ITT Population: the intention-to-treat population con-
sisted of all participants initially assigned to treatment, 100
participants in total.
PP Population: the per protocol population was restricted
to all compliant participants who finished study regularly
with no missing end of trial outcomes. Finally, the PP popu-
lation consisted of 49 volunteers allocated to the control
group and 49 volunteers allocated to the coffee group. One
participant in the coffee group was excluded because the
DNA sample obtained after intervention was lost to analysis.
On the 3 days allotted to sampling (first visit for health
check, and at the end of preconditioning and intervention
period, respectively), volunteers visited the study site in a
fasting state for venous blood drawing (16 ml). Urine (10 ml)
was collected and stored at − 80 °C (2nd and 3rd visit) for
assessment of compliance: NMP and creatinine measure-
ments were carried out by the group of Thomas Hofmann,
TU Munich, with HILIC–HPLC–MS/MS according to Lang
et al. [15].
Indices of health, compliance and nutrition
Standard health checks were carried out during the recruit-
ment phase (visit 1) and on visits 2 and 3 of the trial. Medi-
cal examinations included measurement of blood pressure,
as well as body weight and height for BMI calculation. Med-
ical history evaluation for known diseases that could lead
to exclusion from the study was also carried out. Standard
pregnancy tests were offered to female participants at all
three visits.
Haematological and clinical chemistry analyses were
performed during the recruitment phase to exclude persons
with acute or chronic disease. Analyses included erythro-
cyte count, haematocrit, haemoglobin, mean corpuscular
volume, mean corpuscular haemoglobin concentration,
leukocyte count, differential leukocyte count (lymphocytes,
neutrophils, eosinophils, basophils and monocytes), and
platelet count. Clinical chemistry measures included
C-reactive protein, glucose, creatinine, gamma-glutamyl
transferase, glutamic oxaloacetic transaminase, glutamate-
pyruvate transaminase, and cystatin C.
Participants were asked to complete 7-day food records
reporting about the last week of each study period, prior to
blood sampling. Individual food intake was evaluated using
the free nutrition software package ‘Chudnutie’ (in Slovak—
https​://www.chudn​utie-ako.sk/potra​viny/energ​etick​e-hodno​
ty-potra​vin.php) to obtain an estimate of intake of kcal and
nutrients.
Study coffee description
The study coffee (designated C21) was a blend of pure Ara-
bica dark roast coffee (Coffea arabica L.). The conditions
of roasting (temperature and time) were adjusted in such a
way that after roasting and blending the final product had a
composition of key ingredients caffeoylquinic acids (sum of
3-, 4- and 5-CQA) (9.40 mg/g), NMP (1.08 mg/g) and trigo-
nelline (4.10 mg/g). The caffeine content was 12.4 mg/g.
The coffee was manufactured by Tchibo GmbH Hamburg,
Germany. The composition of coffee C21 was comparable
to the coffee type analysed in previous clinical studies of
the DNA-protective effect of a dark roast coffee blend [12,
14]. Data on coffee constituents were kindly provided by
Tchibo GmbH.
The study coffee was supplied in coffee pads containing
7.5 g (± 0.1 g) each, whichwere packed in neutral packaging
labelled as “Arabica Coffee” and stored under dry condi-
tions, avoiding temperatures above 25 °C. Participants of
the intervention group used commercial ­Senseo®-type pad
machines for the preparation of a 125-ml beverage.
Comet assay analysis of DNA damage
Alkaline single cell gel electrophoresis (the comet assay)
was carried out on whole blood according to Akor-Dewu
et al. with slight modifications [16]. Briefly, 6 µl of blood
was mixed with 140 µl of 0.8% agarose and distributed on
glass slides (two 70 µl gels per slide), kept at 4 °C for solidi-
fication, lysed in cold 2.5 M NaCl, 0.1 M EDTA, 10 mM
Tris, 1% Triton X-100, pH 10 for 1 h, incubated in cold 0.3M
NaOH, 1 mM EDTA for 20 min, and then electrophoresed
in the same solution for 20 min at 0.8 V/cm. If breaks are
present, DNA supercoiling is relaxed, and the DNA moves
towards the anode. After neutralisation with PBS, and stain-
ing with SYBRGold (Invitrogen), the comet-like structures
were visualised by fluorescence microscopy; the greater the
relative intensity of the comet tail, the more DNA breaks are
present. Comet Assay IV image analysis software (Percep-
tive Instruments) was used to score 100 comets per gel, 200
13

comets per sample. Results are expressed as % DNA tail
intensity (TI), which is linearly related to break frequency
over the range of damage levels expected [17]. All comet
assay analyses were conducted by the same investigator who
was unaware of the participants’ allocation to either group.
Statistical methods
The sample size estimation was based on the assumption
of normally distributed data. The significance level was set
at 5% and the power of the test was required to be 80%.
The expected effect size was derived from the results of
Bakuradze et al. [14], who examined the effect of coffee
on DNA damage in 84 male subjects in a parallel design
with two groups, coffee and water consumption. Standard
deviations of 0.16% TI in the control group and 0.15% TI
in the intervention group were observed for DNA damage.
Expecting a difference between groups of 0.1% TI and simi-
lar standard deviations, for the present study, an appropriate
sample size of 39 per group was estimated. To account for
possibly higher standard deviations induced by the inclusion
of women 50 subjects per group seemed appropriate.
The single primary endpoint of the study was defined as
the comparison of coffee and control groups with respect to
the levels of DNA strand breaks in blood cells. The hypoth-
esis of no differences between study groups at the end of
the intervention period baseline adjusted (null hypothesis)
was subjected to a two-sided test setting the level of signifi-
cance at 5%. Repeated measurements were summarised by
the arithmetic mean. The normal distribution hypothesis was
verified by the Shapiro–Wilk test. Since DNA strand break
data were not normally distributed the Wilcoxon rank sum
test (WRST) was applied. In order to account for varying
baseline levels, delta levels for the change before/after inter-
vention were calculated. Relative changes were assessed by
relating delta values to individual baseline levels. Age and
BMI were taken into account as possible confounding fac-
tors on changes in DNA damage but correlations were weak
(less than 10% in absolute size of correlation coefficients).
Since only two samples were compared, a p value adjust-
ment for multiple comparisons was not necessary. Analy-
ses were based on both ITT and PP population. Data from
subjects with missing outcome values were excluded for PP
analysis. In contrast to PP analysis, baseline values were
carried forward for ITT analysis in case of missing values.
Homogeneity of groups, specified as BMI, age, weight,
energy and nutritional intake, was examined additionally
per visit. In general t tests for independent samples were
applied for group comparisons and paired t test to assess
changes from baseline to post-intervention by group. In case
of rejection of the hypothesis of normally distributed data
WRST was used for independent samples and WSRT for
paired samples.
13
European Journal of Nutrition
Results
Participants
A hundred persons, 50 males and 50 females, who fulfilled
the inclusion criteria were randomly allocated to two treat-
ment groups after stratification for BMI and sex. Coffee and
control groups were comparable for mean BMI, age, and
height, p > 0.05% for all comparisons, t test (see Table 1).
Assessing DNA strand breakage
The trial tested for a possible DNA-protective effect of
4 weeks of coffee consumption in comparison to 4 weeks
of water consumption, both preceded by a coffee/tea-free
preconditioning period of 4 weeks. At the end of the pre-
conditioning and at the end of the intervention period, DNA
strand breakage was measured in samples of whole blood
with the comet assay. Figure 2 shows median DNA strand
breaks expressed as % tail intensity (TI) and interquartile
ranges. Mean ± standard deviation (SD) for the control group
were 0.92 ± 0.34% TI at the beginning and 0.92 ± 0.33%
TI at the end of the intervention period, p > 0.05, WSRT.
Mean ± SD for the coffee group were 1.05 ± 0.36% at the
beginning and 0.83 ± 0.28% TI at the end of the intervention
period, resulting in a significant decrease of − 0.23 ± 0.46%
TI, p = 0.0012, paired t test. The two test groups were com-
pared with respect to DNA strand breaks at the beginning
and after the intervention. At the beginning of intervention
mean DNA damage was significantly higher (+ 0.16% TI,
p = 0.029, t test) in the test group versus control. In order to
control for the subject’s individual initial disposition, base-
line adjustment was performed by analysing the change from
baseline (delta values). Evaluation of delta values by WRST
shows a significant difference between groups at the end of
the intervention period (difference in means of delta values
− 0.23 ∆TI%, p = 0.028, for PP analysis, difference − 0.22
∆TI%, p = 0.029 for ITT analysis).
The result is confirmed by analysis of relative changes,
i.e. individual delta values divided by individual baseline
values. The transformation to individual relative changes
generated some extremes and thus medians are reported
Table 1  Demographic data at visit 2 (means ± SD)
Control group Coffee group (n = 50)
(n = 50)
Age (years) 32 ± 8 33 ± 7
BMI (kg/m2) 23.6 ± 3.0 24.0 ± 3.0
Height (cm) 174 ± 10 173 ± 10
Body weight (kg) 72 ± 13 72 ± 13
European Journal of Nutrition

Fig. 2  DNA damage expressed

as comet tail intensity (TI)
for control and coffee groups,
measured in blood samples
taken at end of precondition-
ing phase (visit 2) and end
of intervention (visit 3). The
height of the box corresponds
to the interquartile range and
the bold line to the median.
All values within the range of
box ± 1.5*interquartile range are
covered by the whiskers. Values
above or below this range are
marked as separate dots

instead of means. The median of relative changes in the con-
trol group is − 7%, and in the test group − 22% (difference in
medians − 15%, p = 0.035, WRST). Analysis of ITT popula-
tion data led to similar results (difference − 14%, p = 0.035,
WRST).
The level of DNA strand breakage was not significantly
different between women and men in both treatment groups,
before or after the intervention period (p > 0.05, WRST,
Table 2). When changes from baseline were analysed, DNA
strand breaks were found reduced in women by 0.26 ∆TI%
and in men by 0.20 ∆TI% in the coffee group in comparison
to the control group.
Compliance
After the preconditioning period (baseline visit 2), concen-
trations of the coffee constituent NMP in urine exceeded the
upper level expected for abstention from coffee consumption
(0.2 nmol/µmol creatinine) in one sample from the control
group, and the related data were excluded from statistical
analysis (Fig. 3). At baseline, the difference in median NMP/
creatinine-values between groups was not statistically sig-
nificant (p = 0.173, WRST). After the intervention period, all
samples of the control group had NMP levels close to zero,
while study participants of the coffee intervention group
showed values from 2.1 to 94.2 nmol/µmol NMP/creatinine
(p < 0.0001; WRST, for the comparison between groups).
The reasons for the large variation of NMP/creatinine values
after drinking coffee have not been well researched but have
also been observed earlier [14].
Body weight and nutritional intake
Weekly intakes of energy and nutrients of the study partici-
pants were calculated from 7-day food records for precon-
ditioning and intervention periods. Body weight was ana-
lysed by gender. Mean body weight did not change during
the course of the study in both, women and men (Table 3).
Control and coffee groups also did not differ with regard to
the mean weekly intakes of total energy, proteins, fats or
saccharides, and intakes did not change over the course of
the trial (p > 0.05 for all comparisons, t tests). The highest
individual decrease in body weight observed at study end
was (−) 3 kg, the highest increase was (+) 6 kg.
Caffeine intake
The caffeine content of C21 ground coffee was 1.24%.
Assuming 100% caffeine transfer from each coffee pad to
the beverage, the daily consumption of four pads of C21

Table 2  DNA strand breaks (% Group Blood sampling Women Men

tail intensity) in women versus
men n Mean SD n Mean SD
a
Control Visit 2 25 0.92 0.35 24 0.93 0.33
Visit 3 25 0.89 0.28 24a 0.95 0.38
Coffee Visit 2 25 1.11 0.31 25 1.05 0.41
Visit 3 23b 0.84 0.31 25 0.87 0.25

In both groups mean changes were not significantly different between women and men (p > 0.05, WRST)
a
 Data of one participant were excluded from analysis due to non-compliance
b
 Data of one participant were missing due to a sample lost to analyses, another participant discontinued the
study

13
 European Journal of Nutrition

Fig. 3  Compliance: NMP/cre-
atinine concentrations (nmol/
µmol) of individual subjects, a
coffee group, b control group.
The negative values are due
to lower signals in study urine
samples than in the “negative”
calibration sample, which is due
to lower background signals
in study urine compared to the
calibration solution

Table 3  Body weight and Control group Coffee group

nutritional intakes per week
(means ± SD, PP population) Preconditioning Intervention Preconditioning Intervention

Body weight (kg)

 Total 72 ± 13 (n = 49) 72 ± 13 (n = 49) 73 ± 13 (n = 50) 72 ± 13 (n = 50)
 Women 63 ± 9 (n = 25) 63 ± 9 (n = 25) 63 ± 8 (n = 25) 63 ± 8 (n = 25)
 Men 81 ± 9 (n = 24) 81 ± 9 (n = 24) 82 ± 10 (n = 25) 82 ± 9 (n = 25)
Nutritional intakes (per week)
 Energy (kcal) 13,900 ± 3300 13,700 ± 3700 13,800 ± 3400 12,900 ± 3000
 Proteins (g) 550 ± 130 530 ± 120 570 ± 150 540 ± 170
 Fats (g) 580 ± 170 550 ± 180 570 ± 170 520 ± 160
 Saccharides (g) 1580 ± 460 1590 ± 470 1500 ± 390 1400 ± 340
 Caffeine (mg)a ≤ 100 ≤ 100 ≤ 100 ~ 2700 ± 140
a
 As estimated from 7-day food records, and from C21 caffeine content. A fixed maximum intake of 100 mg
caffeine per week with other foods than coffee was added as domain of uncertainty. The deviation of caf-
feine consumption in the coffee group during intervention is calculated from the coffee consumed plus
100 mg caffeine from possibly other food sources. The variation of 140 mg accounts for deviations in cof-
fee pad content (± 0.1 g)

coffee was 372 mg (± 5.0 mg), resulting in ~ 2600 mg/week.
Participants in the coffee group were asked to refrain from
any caffeine-containing food or beverage apart from the
study coffee. As a safety margin, the additional consump-
tion of 100 mg caffeine per week was added, resulting in
total consumption of 2700 mg caffeine per week (Table 3).
The individual caffeine consumption thus may amount up
to 386 mg/day within this group. This quantity observes the
current EFSA caffeine safety evaluation (≤ 400 mg/day) for
adults, except for women during pregnancy [18].
Discussion
In this human intervention trial, with parallel groups of well-
matched, randomised healthy adults, DNA damage (basal
levels of strand breaks and alkali-labile sites) measured with
the comet assay in whole blood was significantly lower in
coffee drinkers compared with non-coffee drinkers at the end
of the 4-week intervention period. In order to better recog-
nise the effects of coffee consumption, all study participants
had to avoid coffee and other caffeine containing beverage
during a period of 4 weeks prior to the intervention. Assum-
ing that DNA damage measured by the comet assay reflects
exposure of the individual to DNA-damaging, potentially
mutagenic agents in the body (such as endogenous free radi-
cals) or the environment (such as passive smoking, benzene
vapours at petrol stations, UV or high energy space radia-
tion), a decrease in DNA breakage can reasonably be inter-
preted as increased cell protection from DNA damage, possi-
bly and favourably affecting cancer risk. Previous studies of
basal DNA damage using the comet assay indeed indicated
significantly increased damage in leukocytes from persons
exposed to air pollution, traffic borne ultrafine particles,

13
European Journal of Nutrition
tobacco smoke or poor health practices. Compared to non-
exposed controls, 11–58% more DNA strand breaks were
reported in studies reviewed by Valverde and Rojas and in
the study of Lu et al. [19, 20]. The decrease of strand breaks
by 15% after coffee consumption compared with water, as
observed in the present study, falls into this range and there-
fore may be clinically relevant.
A possible beneficial effect of coffee consumption on
health outcomes is supported by a large number of epide-
miological studies. Poole et al. [21] recently published an
‘umbrella review’, evaluating the evidence from as many
as 201 meta-analyses of observational studies (135 articles,
67 unique outcomes), and 17 meta-analyses of randomised
controlled trials (6 articles, 9 unique outcomes). Overall, sig-
nificant, consistent negative associations were seen between
coffee consumption and all-cause mortality, cardiovascular
mortality and cancer incidence, as well as with various indi-
vidual diseases; the only harmful effects of coffee were in
relation to pregnancy (low birth weight, preterm birth and
pregnancy loss).
The current study focused on the measurement of DNA
strand breaks as a commonly used biomarker of DNA dam-
age in population-based studies of environmental and occu-
pational exposure [19]. We observed a basal level of dam-
age at around 1% DNA in tail (TI); this is in the range of
previous studies of coffee consumption (0.3–1.1%) [12, 14,
22]. The present study avoided the isolation of peripheral
mononuclear cells by density gradient centrifugation and
subsequent washing but used unprocessed whole blood,
as previously described [12, 14]. The level of damage, at
around 1% DNA in tail intensity (TI) is low by comparison
with levels of around 5% typically seen in isolated periph-
eral blood mononucleocytes [16]. Whether this indicates an
under-estimation of damage in whole blood, or whether the
isolation procedure of mononuclear cells by density gradi-
ent centrifugation and washing causes additional damage, is
currently under investigation.
Among several previous studies of the effects of coffee
consumption on DNA damage, only two were randomised
controlled trials, with DNA breakage rather than oxidative
modification of DNA bases analysed [23]. Bakuradze et al.
[14] also studied the effects of coffee C21 consumption, after
4 weeks of intervention, and found a decrease of DNA strand
breaks by 27% in comparison to the control group with water
consumption (p < 0.001). Misik et al. [22] studied a much
shorter intervention period (5 days) and a different coffee
type and used isolated lymphocytes. In comparison to C21,
the coffee had an approximately twofold content of caffeoyl
quinic acids and trigonelline and about half of N-methylpyri-
dinium. Therefore, the coffee used by Misik et al. is a light/
medium roast type while we used here a dark roast blend.
We assume that the different degree of roasting accounts
for the different outcome observed. In animal studies, it was
shown that dark roast coffee is particularly potent in induc-
ing oxidative stress defence genes via the nuclear factor
Nrf2/EpRE and in promoting anti-inflammatory reactivity
by inhibiting NF-kB activation [24].
A recent randomised-controlled trial of coffee consump-
tion analysed DNA strand breaks after 8 weeks of inter-
vention and did not observe a protective effect [25]. Again,
a light/medium roast coffee was used in the intervention
period, characterised by an almost twice as high content of
caffeoyl quinic acids and three times lower N-methylpyri-
dinium as compared to the coffee used in the present study
or reported by Bakuradze et al. [14].
The present study extends the previous report on the
DNA-protective effect of a well-characterised dark roast cof-
fee blend by including female participants and by recruiting
study participants from a different, central European region.
Between men and women, we found neither a difference in
baseline DNA strand breaks, nor a difference in DNA strand
breaks at the end of the study. The protective effect of coffee
is further supported by the location of the trial in a country
of a different European region than the location of the pre-
vious study (Slovakia versus southwestern Germany [14]).
In the current study, a lower dose of coffee was consumed
(500 instead of 750 ml/day). This indicates that the DNA-
protective effect of dark roast coffee is not limited to the
consumption of a defined volume per day. Further studies on
the chemical composition and biological effects of dark roast
versus light/medium roast coffee seem warranted.
Acknowledgements  We thank Thomas Hofmann and Roman Lang
(Technical University of Munich) for NMP, trigonelline and creati-
nine measurements (compliance), and Lubica Prochazkova, Helena
Nagyova, Edita Mrvikova and Zuzana Krchnava from Slovak Medical
University for technical help.
Funding  This study has been supported by Tchibo GmbH, Hamburg,
Germany.
Compliance with ethical standards 
Conflict of interest  D. Schipp is a self-employed statistician, who has
been appointed and financed by Tchibo GmbH for this and other pro-
jects.
Ethical standards  The study has been performed in accordance with
the ethical standards laid down in the 1964 Declaration of Helsinki and
its later amendments.
References
1. Hamer M, Witte DR, Mosdøl A, Marmot MG, Brunner EJ (2008)
Prospective study of coffee and tea consumption in relation to risk
of type 2 diabetes mellitus among men and women: the White-
hall II study. Br J Nutr 100(5):1046–1053. https:​ //doi.org/10.1017/
S0007​11450​89441​35
13

2. Zhang W, Lopez-Garcia E, Hu FB, van Dam RM (2009) Cof-
fee consumption and risk of cardiovascular diseases and all-
cause mortality among men with type 2 diabetes. Diabetes Care
32(6):1043–1045. https​://doi.org/10.2337/dc08-2251
3. Bidel S, Tuomilehto J (2012) The emerging health benefits of
coffee with an emphasis on type 2 diabetes and cardiovascular dis-
ease. Eur Endocrinol 9(2):99–106. https​://doi.org/10.1002/97811​
19949​893.ch9
4. Malerba S, Turati F, Galeone C, Pelucchi C, Verga F, La Vec-
chia C, Tavani A (2013) A meta-analysis of prospective studies
of coffee consumption and mortality for all causes, cancers and
cardiovascular diseases. Eur J Epidemiol. https:​ //doi.org/10.1007/
s1065​4-013-9834-7
5. von Ruesten A, Feller S, Bergmann M, Boeing H (2013) Diet and
risk of chronic diseases: results from the first 8 years of follow-up
in the EPIC-Potsdam study. Eur J Clin Nutr 67(4):412–419. https​
://doi.org/10.1038/ejcn.2013.7
6. Ding M, Satija A, Bhupathiraju SN, Hu Y, Sun Q, Han J, Lopez-
Garcia E, Willett W, van Dam RM, Hu FB (2015) Association of
coffee consumption with total and cause-specific mortality in 3
large prospective cohorts. Circulation 132(24):2305–2315. https​
://doi.org/10.1161/circu​latio​naha.115.01734​1
7. Bakuradze T, Lang R, Hofmann T, Stiebitz H, Bytof G, Lantz
I, Baum M, Eisenbrand G, Janzowski C (2010) Antioxidant
effectiveness of coffee extracts and selected constituents in cell-
free systems and human colon cell lines. Mol Nutr Food Res
54(12):1734–1743. https​://doi.org/10.1002/mnfr.20100​0147
8. Bichler A, Cavin C, Simic T, Chakraborty A, Ferk F, Hoelzl C,
Schulte-Hermann R, Kundi M, Haidinger G, Angelis K, Knasmül-
ler S (2007) Coffee consumption protects human lymphocytes
against oxidative and 3-amino-1-methyl-5H-pyrido[4,3-b]indole
acetate (Trp-P-2) induced DNA-damage: results of an experimen-
tal study with human volunteers. Food Chem Toxicol 45(8):1428–
1436. https​://doi.org/10.1016/j.fct.2007.02.001
9. Esposito F, Morisco F, Verde V, Ritieni A, Alezio A, Caporaso
N, Fogliano V (2003) Moderate coffee consumption increases
plasma glutathione but not homocysteine in healthy subjects.
Aliment Pharmacol Ther 17(4):595–601. https​://doi.org/10.104
6/j.1365-2036.2003.01429​.x
10. Hoelzl C, Knasmüller S, Wagner K-H, Elbling L, Huber W, Kager
N, Ferk F, Ehrlich V, Nersesyan A, Neubauer O, Desmarchelier
A, Marin-Kuan M, Delatour T, Verguet C, Bezençon C, Besson
A, Grathwohl D, Simic T, Kundi M, Schilter B, Cavin C (2010)
Instant coffee with high chlorogenic acid levels protects humans
against oxidative damage of macromolecules. Mol Nutr Food Res
54(12):1722–1733. https​://doi.org/10.1002/mnfr.20100​0048
11. Boettler U, Sommerfeld K, Volz N, Pahlke G, Teller N, Somoza
V, Lang R, Hofmann T, Marko D (2011) Coffee constituents
as modulators of Nrf2 nuclear translocation and ARE (EpRE)-
dependent gene expression. J Nutr Biochem 22(5):426–440. https​
://doi.org/10.1016/j.jnutb​io.2010.03.011
12. Bakuradze T, Boehm N, Janzowski C, Lang R, Hofmann T,
Stockis J-P, Albert FW, Stiebitz H, Bytof G, Lantz I, Baum M,
Eisenbrand G (2011) Antioxidant-rich coffee reduces DNA dam-
age, elevates glutathione status and contributes to weight control:
results from an intervention study. Mol Nutr Food Res 55(5):793–
797. https​://doi.org/10.1002/mnfr.20110​0093
13
European Journal of Nutrition
13. Bakuradze T, Parra GAM, Riedel A, Somoza V, Lang R,
Dieminger N, Hofmann T, Winkler S, Hassmann U, Marko
D, Schipp D, Raedle J, Bytof G, Lantz I, Stiebitz H, Richling
E (2014) Four-week coffee consumption affects energy intake,
satiety regulation, body fat, and protects DNA integrity. Food
Res Int 63(Part C):420–427. https​: //doi.org/10.1016/j.foodr​
es.2014.05.032
14. Bakuradze T, Lang R, Hofmann T, Eisenbrand G, Schipp D, Galan
J, Richling E (2015) Consumption of a dark roast coffee decreases
the level of spontaneous DNA strand breaks: a randomized con-
trolled trial. Eur J Nutr 54:149–156. https:​ //doi.org/10.1007/s0039​
4-014-0696-x
15. Lang R, Wahl A, Stark T, Hofmann T (2011) Urinary N-meth-
ylpyridinium and trigonelline as candidate dietary biomarkers of
coffee consumption. Mol Nutr Food Res 55(11):1613–1623. https​
://doi.org/10.1002/mnfr.20100​0656
16. Akor-Dewu MB, El Yamani N, Bilyk O, Holtung L, Tjelle TE,
Blomhoff R, Collins AR (2014) Leucocytes isolated from simply
frozen whole blood can be used in human biomonitoring for DNA
damage measurement with the comet assay. Cell Biochem Funct
32(3):299–302. https​://doi.org/10.1002/cbf.3016
17. Collins AR, Oscoz AA, Brunborg G, Gaivão I, Giovannelli L,
Kruszewski M, Smith CC, Štětina R (2008) The comet assay: topi-
cal issues. Mutagenesis 23(3):143–151. https​://doi.org/10.1093/
mutag​e/gem05​1
18. EFSA Panel on Dietetic Products Nutrition and Allergies (2015)
Scientific opinion on the safety of caffeine. EFSA J 13(5):4102.
https​://doi.org/10.2903/j.efsa.2015.4102
19. Valverde M, Rojas E (2009) Environmental and occupational
biomonitoring using the Comet assay. Mutat Res Rev Mutat Res
681(1):93–109. https​://doi.org/10.1016/j.mrrev​.2008.11.001
20. Lu Y, Morimoto K, Nakayama K (2006) Health practices and leu-
kocyte DNA damage in Japanese hard-metal workers. Prev Med
43(2):140–144. https​://doi.org/10.1016/j.ypmed​.2006.03.008
21. Poole R, Kennedy OJ, Roderick P, Fallowfield JA, Hayes PC,
Parkes J (2017) Coffee consumption and health: umbrella review
of meta-analyses of multiple health outcomes. BMJ 359:j5024.
https​://doi.org/10.1136/bmj.j5024​
22. Misík M, Hoelzl C, Wagner K-H, Cavin C, Moser B, Kundi M,
Simic T, Elbling L, Kager N, Ferk F, Ehrlich V, Nersesyan A,
Dusinská M, Schilter B, Knasmüller S (2010) Impact of paper
filtered coffee on oxidative DNA-damage: results of a clinical
trial. Mutat Res Fundam Mol Mech Mutagen 692:42–48. https​://
doi.org/10.1016/j.mrfmm​m.2010.08.003
23. Martini D, Del Bo C, Tassotti M, Riso P, Del Rio D, Brighenti
F, Porrini M (2016) Coffee consumption and oxidative stress: a
review of human intervention studies. Molecules 21(8):979. https​
://doi.org/10.3390/molec​ules2​10809​79
24. Paur I, Balstad TR, Blomhoff R (2010) Degree of roasting is the
main determinant of the effect of coffee on NF-kB and EpRE. J
Free Radic Biol Med 48(9):1218–1227. https​://doi.org/10.1016/j.
freer​adbio​med.2010.02.005
25. Shaposhnikov S, Hatzold T, Yamani NE, Stavro PM, Lorenzo
Y, Dusinska M, Reus A, Pasman W, Collins A (2018) Coffee
and oxidative stress: a human intervention study. Eur J Nutr
57(2):533–544. https​://doi.org/10.1007/s0039​4-016-1336-4