See corresponding editorial on page 787.

Coffee consumption and risk of chronic disease in the European

Prospective Investigation into Cancer and Nutrition (EPIC)Germany
study13
Anna Floegel, Tobias Pischon, Manuela M Bergmann, Birgit Teucher, Rudolf Kaaks, and Heiner Boeing

INTRODUCTION

Coffee is among the most popular nonalcoholic beverages and is

one of the most valuable export commodities in the world (1). In
2007, the top coffee consuming country was Finland with 12 kg per
capita. In Germany, the average coffee consumption in 2007 was
6.4 kg per capita, whereas in the United States it was 4.2 kg per
capita (2). After removal of the flesh from the coffee cherries, the
seeds are roasted and develop their unique bitter aroma. Coffee is a
complex chemical mixture, and its most abundant components are
chlorogenic acid and caffeine followed by other polyphenols and
methylxanthines, carbohydrates, lipids, nitrogenous compounds,
nicotinic acid, potassium, and magnesium to name a few (3). Besides its culinary value, coffee has long been suggested to affect
human health and disease (3). In the late 16th century, British
doctors observed that coffee accelerated digestion, improved

cognitive function, and increased the heart rate (4). However, the
effect of coffee consumption on chronic disease risk is still being
debated. In randomized controlled trials it was observed that acute
coffee ingestion caused an increase in blood pressure (5, 6) and led
to an elevation in plasma concentrations of total and LDL cholesterol and homocysteine (79). Case-control studies reported
a positive association (3, 10, 11), or a J-shaped relation (12), between coffee consumption and the risk of CVD4. Moreover, some
studies reported an acute elevation in the risk of MI and stroke in
the hour of coffee ingestion (13, 14). These findings contributed to
the belief that coffee is harmful (15), which is widespread among
the general public and is also supported by current dietary recommendations that favor low to moderate coffee consumption.
Prospective cohort studies, however, reported conflicting results:
some studies suggested a positive association between coffee consumption and risk of CVD (16, 17), whereas others reported no
association (18, 19) or even an inverse association (2022). Most
previous studies focused on the risk of coronary heart disease, and
studies addressing stroke risk are even fewer. An inverse association between coffee consumption and risk of T2D has been suggested in a recent case-control study in the United States (23) and
in many prospective cohort studies in Europe (2426), the United
States (27, 28), and Japan (29). Studies investigating the association between coffee consumption and total cancer risk are rarely
found in the literature because most studies focused on site-specific
cancers. Some evidence exists that coffee may reduce the risk of
liver, breast, and oral cavity/pharynx cancers (3033). In contrast,
1

From the Department of Epidemiology, German Institute of Human

Nutrition Potsdam-Rehbruecke, Nuthetal, Germany (AF, TP, MMB, and
HB); the Molecular Epidemiology Group, Max Delbruck Center for Molecular Medicine Berlin-Buch, Germany (TP); and the Division of Cancer
Epidemiology, German Cancer Research Center, Heidelberg, Germany
(BT and RK).
2
Supported by the Federal Ministry of Science, Germany (grant 01 EA
9401), the European Union (grant SOC 95 201408 05F02), and the German
Cancer Aid (grant 70-2201-Bo2).
3
Address correspondence and reprint requests to A Floegel, German Institute of Human Nutrition Potsdam-Rehbruecke, Arthur-Scheunert-Allee
114-116, 14558 Nuthetal, Germany. E-mail: anna.floegel@dife.de.
4
Abbreviations used: CVD, cardiovascular disease; EPIC, European Prospective Investigation into Cancer and Nutrition; FFQ, food-frequency questionnaire; MI, myocardial infarction; T2D, type 2 diabetes.
Received July 18, 2011. Accepted for publication December 29, 2011.
First published online February 15, 2012; doi: 10.3945/ajcn.111.023648.

Am J Clin Nutr 2012;95:9018. Printed in USA. 2012 American Society for Nutrition

901
Supplemental Material can be found at:
http://ajcn.nutrition.org/content/suppl/2012/03/22/95.4.901.D
C1.html

Downloaded from ajcn.nutrition.org by guest on April 6, 2015

ABSTRACT
Background: Early studies suggested that coffee consumption may
increase the risk of chronic disease.
Objective: We investigated prospectively the association between
coffee consumption and the risk of chronic diseases, including type
2 diabetes (T2D), myocardial infarction (MI), stroke, and cancer.
Design: We used data from 42,659 participants in the European Prospective Investigation into Cancer and Nutrition (EPIC)Germany
study. Coffee consumption was assessed by self-administered foodfrequency questionnaire at baseline, and data on medically verified
incident chronic diseases were collected by active and passive followup procedures. HRs and 95% CIs were calculated with multivariate
Cox regression models and compared by competing risk analysis.
Results: During 8.9 y of follow-up, we observed 1432 cases of
T2D, 394 of MI, 310 of stroke, and 1801 of cancer as first qualifying events. Caffeinated (HR: 0.94; 95% CI: 0.84, 1.05) or decaffeinated (HR: 1.05; 95% CI: 0.84, 1.31) coffee consumption (4
cups/d compared with ,1 cup/d; 1 cup was defined as 150 mL) was
not associated with the overall risk of chronic disease. A lower risk
of T2D was associated with caffeinated (HR: 0.77; 95% CI: 0.63,
0.94; P-trend 0.009) and decaffeinated (HR: 0.70; 95% CI: 0.46,
1.06; P-trend: 0.043) coffee consumption (4 cups/d compared with
,1 cup/d), but cardiovascular disease and cancer risk were not. The
competing risk analysis showed no significant differences between
the risk associations of individual diseases.
Conclusion: Our findings suggest that coffee consumption does not
increase the risk of chronic disease, but it may be linked to a lower
risk of T2D.
Am J Clin Nutr 2012;95:9018.

902

FLOEGEL ET AL

results from case-control studies suggest that coffee consumption

may increase the risk of pancreatic, bladder, ovarian, and prostate
cancer but the evidence is considered insufficient (3, 33, 34).
With such contrasting findings, the public debate about reducing
the risk of one disease only to increase the risk of another by
drinking coffee is very much alive. From a public health perspective,
it is most relevant to investigate whether coffee consumption affects
morbidity from major chronic diseases and whether individual
disease risks are competing with each other to optimally translate
evidence-based recommendations. Furthermore, the relation between coffee consumption and risk of MI and stroke and total cancer
risk is not resolved and deserves large-scale prospective investigations. In particular, more studies in European cohorts are desirable to support evidence from US cohorts (35). In addition,
previous studies did not generally differentiate between caffeinated
and decaffeinated coffee (36). Therefore, the current study aimed to
prospectively investigate the association between caffeinated and
decaffeinated coffee consumption and risk of chronic disease, including T2D, MI, stroke, and cancer in the EPIC-Germany cohort.

Study population
The EPIC study is a multicenter prospective cohort study in 10
European countries that aims to investigate the associations between diet, lifestyle, and chronic disease risk. In Germany, there
are 2 study centers, which are located in Potsdam and Heidelberg.
Adults aged mainly between 35 and 65 y were recruited from the
general population in Potsdam, Heidelberg, and surrounding areas
with response rates of 22.7% in Potsdam and 38.3% in Heidelberg
(37, 38). A total of 53,088 adults (30,255 women and 22,833 men)
consented to participate during the recruitment examination, which
took place between 1994 and 1998. At baseline, study participants
underwent an examination that included anthropometric and blood
pressure measurements and an interview on medical history, including prevalent diseases. A self-administered questionnaire on
sociodemographic and lifestyle factors and an FFQ were also completed. After baseline, follow-up questionnaires were administered
every 23 y to identify incident cases of chronic diseases, including
T2D, MI, stroke, and cancer. For the current analysis, we considered the data until the fourth follow-up period. Written informed
consent was obtained from all study participants a priori. The study
was approved by the ethics committees of the Medical Society of
the State of Brandenburg and of the faculty of Medicine, Heidelberg University.
For the current analysis, we excluded participants with missing
information on coffee consumption (n = 17), those with selfreported chronic diseases (T2D, MI, stroke, or cancer) at baseline
(n = 5536) or with unknown disease status (n = 2392), those with
missing information on lifestyle factors and other covariates (n =
1830), and those who never completed a follow-up questionnaire
(n = 654).
Coffee consumption
The EPIC-Germany participants were inquired about their
habitual consumption of caffeinated coffee and decaffeinated
coffee in the self-administered semi-quantitative FFQ at baseline.
In particular, the 2 items were as follows: How frequently did you

Ascertainment of incident chronic diseases

Incident cases of T2D, MI, stroke, and cancer were identified by
participants self-report of the disease occurrence, disease-relevant
medication, or change in diet or cause of death information. Potential incident diseases were further verified by contacting the
treating physician and/or clinic to obtain medical records or via
linkage to cancer and hospital registries (39). The diseases were
coded in accordance with the International Statistical Classification of Diseases, 10th Revision (ICD-10) (40), criteria: E11 for
T2D; I21 for MI; I60, I61, I63, and I64 for stroke; and C00-C97
for cancer (except C44: nonmelanoma skin cancer). In the current
analysis, in case of multimorbidity, only the first occurring incident event was counted.
Statistical analysis
Habitual caffeinated and decaffeinated coffee consumption were
considered as exposure variables. The population was divided into
5 coffee consumption categories (,1 cup/d, 1 to ,2 cups/d, 2 to
,3 cups/d, 3 to ,4 cups/d, and 4 cups/d). The categories chosen
agree well with plausible cutoff points of coffee consumption in
other studies (41). Descriptive statistics of the covariates were ageand sex-adjusted and reported across coffee-consumption categories as arithmetic means 6 SEs for continuous variables or as
percentages for categorical variables. The primary endpoint for
this analysis was defined as the first incident event occurring for
T2D, MI, stroke, or cancer, whichever came first. Cox proportional hazards analysis was used to estimate multivariableadjusted HRs as a measure of the RR and 95% CIs, with age
as the primary time-dependent variable (entry and exit time of
each participant defined as the age at recruitment and age at first
event of T2D, MI, stroke, cancer, or censoring, respectively).
Furthermore, the disease-specific HRs for T2D, MI, stroke, and
cancer across categories of coffee consumption were estimated.
Coffee consumers of ,1 cup/d were used as the reference group.
We calculated 3 different multivariate models for each exposure
and outcome. All models were stratified by integers of age at
recruitment to be less sensitive to violation of the proportional
hazards assumption. Model 1 was the crude model and stratified
by age at recruitment and center (Potsdam or Heidelberg) and
adjusted for sex. Model 2 was additionally adjusted for classic
risk factors, particularly smoking (nonsmokers; smoking intensity: ,15, 1524, or 25 cigarettes/d, other smoking; former
smokers: gave up smoking 10, 1120, or .20 y previously;
smoking duration: 10, 1120, 2130 y, 3140 y, or .40 y),
alcohol consumption (nonconsumers; women: .06, .612,
or .12 g/d; men: .012, .1224, or .24 g/d), physical activity
(average of cycling and sports during summer and winter; in
h/wk), education (none to primary school, technical to secondary
school, or higher education including university), employment

Downloaded from ajcn.nutrition.org by guest on April 6, 2015

SUBJECTS AND METHODS

consume coffee with caffeine in the past? and How frequently

did you consume coffee without caffeine in the past? Ten frequency categories were provided, ranging from never to less
than 1/mo to 5 times/d or more. In addition, the usual portion
size was assessed with categories ranging from 0.5 cups to 3 cups
(1 cup was defined as 150 mL). From the frequency and portion
size, the usual consumption of caffeinated and decaffeinated coffee
was estimated.

903

COFFEE CONSUMPTION AND RISK OF CHRONIC DISEASE

consumers of .1 cup/d, and excluding all cases that occurred

during the first 2 y of follow-up. The level of statistical significance was set at P , 0.05 for 2-sided testing. All analyses were
performed with SAS statistical software (release 9.1; SAS Institute Inc).
RESULTS

The final study population comprised 42,659 EPIC-Germany

participants (58.5% women) with a mean age of 49.7 y and mean
caffeinated and decaffeinated coffee intakes of 2.7 and 0.3 cups/d,
respectively. In this population, coffee consumption was distributed
as follows: 72.6% reported to consume caffeinated coffee only, 4.5%
reported to consume decaffeinated coffee only, 18.6% reported to
consume both, and 4.3% reported to consume neither coffee type.
The percentage of current smokers proportionally increased with
caffeinated coffee consumption from 14.4% to 30.2% in a comparison of ,1 to 4 cups/d (Table 1). With increasing caffeinated
coffee consumption, decaffeinated coffee and tea consumption
decreased on a population level, and alcohol consumption increased by trend. With increasing decaffeinated coffee consumption, the proportion of current smokers increased from 21.5% to
35.6%, and the number of cigarettes smoked per day increased
from 13.9 to 17.0 in a comparison of ,1 to 4 cups/d, respectively
(Table 2). Decaffeinated coffee consumers were less likely to have
a university degree and were more likely to have a diagnosis of
hypertension.
The consumption of caffeinated and decaffeinated coffee was
not associated with the overall risk of chronic disease in the crude
model as well as in multivariate model 2 (including known risk

TABLE 1
Baseline characteristics of the EPIC-Germany population (n = 42,659) according to caffeinated coffee consumption1
Caffeinated coffee consumption

No. of subjects
Women (%)2
Age (y)2
BMI (kg/m2)
Waist circumference (cm)
Waist-to-hip ratio
Currently employed (%)
University degree (%)
Current smokers (%)
Cigarettes smoked (no./d)3
Former smokers (%)
Sports (h/wk)
Watching television (h/wk)
Total energy intake (kcal/d)
Alcohol intake from beverages (g/d)
Decaffeinated coffee (cups/d)
Caffeinated tea (cups/d)
Taking vitamin supplement (%)
Taking mineral supplement (%)
Prevalent hypertension (%)
Systolic blood pressure (mm Hg)4
Diastolic blood pressure (mm Hg)4

,1 cup/d

1 to ,2 cups/d

2 to ,3 cups/d

3 to ,4 cups/d

4 cups/d

8689
56.1
49.9 6 8.6
25.8 6 0.04
86.4 6 0.11
0.86 6 0.00
69.0
34.6
14.4
13.2 6 0.3
32.2
3.8 6 0.0
12.3 6 0.1
2005 6 7
13.5 6 0.2
0.75 6 0.01
1.8 6 0.0
22.1
16.6
37.6
133.4 6 0.3
85.5 6 0.2

3860
58.9
49.2 6 8.5
25.8 6 0.06
86.0 6 0.17
0.85 6 0.00
74.3
36.9
14.9
11.7 6 0.4
33.3
3.4 6 0.1
12.4 6 0.1
1970 6 11
15.3 6 0.3
0.27 6 0.02
1.1 6 0.0
20.3
15.5
38.5
134.9 6 0.4
87.1 6 0.3

10,617
62.8
49.8 6 8.7
25.8 6 0.04
85.8 6 0.10
0.85 6 0.00
73.3
36.0
16.8
12.0 6 0.2
34.7
3.4 6 0.0
12.8 6 0.1
2001 6 6
15.9 6 0.2
0.15 6 0.01
0.73 6 0.02
18.8
13.4
38.2
136.0 6 0.2
87.6 6 0.2

7356
57.4
49.0 6 8.2
25.8 6 0.05
85.8 6 0.12
0.85 6 0.00
77.8
36.1
26.8
13.8 6 0.2
33.2
3.2 6 0.0
12.7 6 0.1
2045 6 8
16.8 6 0.2
0.10 6 0.01
0.54 6 0.01
19.5
14.4
35.4
134.3 6 0.3
86.8 6 0.2

12,137
56.9
50.2 6 8.4
26.2 6 0.04
86.8 6 0.09
0.85 6 0.00
73.2
33.0
30.2
15.7 6 0.2
31.0
3.2 6 0.0
13.5 6 0.1
2177 6 6
16.0 6 0.2
0.15 6 0.01
0.55 6 0.01
18.8
13.7
34.9
134.5 6 0.2
86.8 6 0.1

All values are age- and sex-adjusted means 6 SEs or percentages across coffee consumption categories, except where otherwise noted. One cup was
defined as 150 mL. EPIC, European Prospective Investigation into Cancer and Nutrition.
2
Unadjusted means 6 SDs or percentages.
3
In current smokers only.
4
n = 18,372.
1

Downloaded from ajcn.nutrition.org by guest on April 6, 2015

(yes or no), vitamin and mineral supplement use (during past 4 wk:
yes or no), total energy intake (kcal/d), and tea intake (cups/d).
Model 3 included the covariates of model 2 with additional adjustment for risk factors that could be considered intermediates,
particularly BMI (in kg/m2), waist-to-hip ratio (continuous), and
prevalent hypertension (at least one of the following: systolic blood
pressure 140 mmHg, diastolic blood pressure 90 mmHg, blood
pressurelowering medication, or self-reported: yes or no). In addition, in models 2 and 3, caffeinated and decaffeinated coffee
consumption (cups/d) was mutually adjusted. We also calculated
continuous models because a dose-dependent effect was assumed.
Furthermore, the significance of linear trends across the caffeinated
coffee consumption categories was tested by assigning each participant the median of the category and modeling this value as
a continuous variable. In addition, we conducted a competing risk
analysis for Cox regression based on multivariate model 3 and
used the likelihood ratio test to compare the disease-specific risk
associations (42, 43).
Interactions were tested between coffee consumption (continuous), chronic disease risk, and the covariates age, sex, center
(Potsdam or Heidelberg), alcohol intake (categorical), smoking
status (categorical), BMI (30, yes or no), and hypertension (yes
or no). Models with and without multiplicative interaction terms
were compared by using the likelihood ratio test. If an interaction
was observed, a stratified analysis was conducted. For sensitivity
analysis, we calculated disease-specific HRs, including subjects
with any prevalent chronic disease, except the disease under investigation. Furthermore, we conducted the analysis excluding all
nonconsumers of coffee from the reference group, excluding tea

904

FLOEGEL ET AL

ducted the analysis among those with low exposure to caffeine

from tea. Last, when we excluded all cases that occurred during the
first 2 y of follow-up, the association between high decaffeinated
coffee consumption and MI risk was no longer observed, and
a stronger inverse association was found between decaffeinated
coffee consumption and T2D risk. The proportional hazard assumption was met.
DISCUSSION

Despite a general belief that coffee may be harmful, the current

study found no association between coffee consumption and the
risk of chronic disease. Current dietary guidelines recommend
moderate coffee consumption; the American Heart Association
even recommends to restrict caffeinated coffee consumption to
12 cups/d (44). They most likely refer to early epidemiologic
studies that found a positive association between coffee consumption and CVD risk. In contrast to current dietary recommendations, we observed that coffee consumption was not
associated with risk of CVD, but that high coffee consumption
was associated with lower T2D risk.
An inverse association between coffee consumption and T2D
risk has been reported in previous studies across continents (26
28, 40, 45). In accordance to our results, Huxley et al (40) found
that T2D risk was reduced by 24% when caffeinated coffee
consumption of 35 cups/d was compared with 2 cups/d in
their meta-analysis; similar results were observed for decaffeinated coffee consumption. In addition, our findings suggest
that the association between decaffeinated coffee and risk of

TABLE 2
Baseline characteristics of the EPIC-Germany population (n = 42,659) according to decaffeinated coffee consumption1
Decaffeinated coffee consumption

No. of subjects
Women (%)2
Age (y)2
BMI (kg/m2)
Waist circumference (cm)
Waist-to-hip ratio
Currently employed (%)
University degree (%)
Current smokers (%)
Cigarettes smoked (no./d)3
Former smokers (%)
Sports (h/wk)
Watching television (h/wk)
Total energy intake (kcal/d)
Alcohol intake from beverages (g/d)
Caffeinated coffee (cups/d)
Caffeinated tea (cups/d)
Taking vitamin supplement (%)
Taking mineral supplement (%)
Prevalent hypertension (%)
Systolic blood pressure (mm Hg)4
Diastolic blood pressure (mm Hg)4

,1 cup/d

1 to ,2 cups/d

2 to ,3 cups/d

3 to ,4 cups/d

4 cups/d

38,628
58.0
49.6 6 8.5
25.9 6 0.0
86.0 6 0.1
0.85 6 0.00
74.0
36.2
21.5
13.9 6 0.1
32.8
3.3 6 0.0
12.8 6 0.0
2055 6 3
15.7 6 0.1
2.9 6 0.0
0.91 6 0.01
19.6
14.5
36.3
134.7 6 0.1
86.8 6 0.1

1273
62.9
50.3 6 8.6
26.2 6 0.1
86.6 6 0.3
0.85 6 0.00
69.0
25.6
19.1
14.2 6 0.6
32.0
3.7 6 0.1
13.0 6 0.2
2056 6 19
12.9 6 0.5
1.6 6 0.1
0.67 6 0.05
21.8
15.6
39.0
133.8 6 0.7
85.8 6 0.4

1390
64.5
51.6 6 8.6
26.8 6 0.1
88.3 6 0.3
0.86 6 0.00
65.0
24.0
19.0
14.1 6 0.6
33.4
3.7 6 0.1
13.6 6 0.2
2026 6 18
13.8 6 0.5
1.3 6 0.1
0.60 6 0.05
20.8
13.8
42.6
136.3 6 0.7
87.2 6 0.4

486
60.5
51.1 6 8.3
26.8 6 0.2
88.8 6 0.5
0.87 6 0.00
71.0
25.6
28.2
14.6 6 0.9
31.7
3.7 6 0.1
13.9 6 0.4
2038 6 30
15.2 6 0.9
1.1 6 0.1
0.65 6 0.08
21.1
17.1
40.0
133.5 6 1.1
85.3 6 0.7

882
59.9
50.6 6 8.2
27.4 6 0.1
89.6 6 0.3
0.87 6 0.00
67.0
15.7
35.6
17.0 6 0.6
32.3
3.8 6 0.1
15.1 6 0.3
2178 6 22
14.2 6 0.6
1.7 6 0.1
0.68 6 0.06
18.4
13.9
40.4
133.6 6 0.8
85.9 6 0.5

All values are age- and sex-adjusted means 6 SEs or percentages across coffee consumption categories, except where otherwise noted. One cup was
defined as 150 mL. EPIC, European Prospective Investigation into Cancer and Nutrition.
2
Unadjusted means 6 SDs or percentages.
3
In current smokers only.
4
n = 18,372.
1

Downloaded from ajcn.nutrition.org by guest on April 6, 2015

factors) and model 3 (which considered possible intermediates)

(Tables 3 and 4). For further analysis, model 3 was our model of
choice.
RRs of single endpoints showed that caffeinated coffee consumption was inversely associated with T2D risk, because consumption of 4 cups/d was associated with a 23% lower risk of
T2D compared with ,1 cup/d, but was not related to the other
endpoints. Consumption of decaffeinated coffee was inversely
associated with T2D risk (P-trend: 0.043) and was not associated
with the other endpoints. However, in the highest category of 4
cups/d compared with ,1 cup/d, decaffeinated coffee consumption was positively associated with MI risk (Table 4). The
competing risk analysis showed that there was no significant
difference between the RR associations of the individual diseases
(see Table S1 under Supplemental data in the online issue).
A significant interaction between caffeinated coffee consumption, smoking status, and T2D risk was found (P = 0.015). The
stratified analysis showed the inverse association for T2D risk to be
specific for nonsmokers at baseline (see Figure S1 under Supplemental data in the online issue). No association with T2D risk
was observed among current smokers. We further observed a significant interaction between decaffeinated coffee consumption,
prevalent hypertension, and T2D risk (P = 0.0001), which suggests
that persons with normal blood pressure may not benefit from
decaffeinated coffee consumption in respect of T2D risk.
We conducted several sensitivity analyses (data not shown). The
results were not markedly altered when we first, included subjects
with prevalent chronic diseases at baseline; second, excluded all
nonconsumers of coffee from the reference group; or third, con-

905

COFFEE CONSUMPTION AND RISK OF CHRONIC DISEASE

TABLE 3
HRs (and 95% CIs) for the association between caffeinated coffee consumption and incident chronic diseases in EPIC-Germany (n = 42,659)1
Caffeinated coffee consumption
1 to ,2 cups/d

2 to ,3 cups/d

3 to ,4 cups/d

4 cups/d

8689/78,974

3860/34,341

10,617/93,272

7356/65,881

12,137/107,290

871
1.00
1.00
1.00

336
0.96 (0.83, 1.11)
0.97 (0.84, 1.13)
0.99 (0.85, 1.14)

979
1.02 (0.92, 1.13)
1.03 (0.92, 1.15)
1.03 (0.92, 1.15)

627
0.96 (0.85, 1.08)
0.93 (0.82, 1.06)
0.95 (0.84, 1.08)

319
1.00
1.00
1.00

137
0.92 (0.72, 1.16)
0.93 (0.73, 1.19)
0.89 (0.69, 1.16)

374
0.93 (0.78, 1.11)
0.92 (0.76, 1.11)
0.92 (0.76, 1.13)

148
1.00
1.00
1.00

50
0.92 (0.64, 1.32)
0.93 (0.64, 1.35)
0.94 (0.64, 1.36)

88
1.00
1.00
1.00

P-trend

Per cup/d

1124
0.99 (0.89, 1.09)
0.93 (0.84, 1.04)
0.94 (0.84, 1.05)

0.735
0.149
0.206

1.01 (0.99, 1.02)

0.99 (0.97, 1.01)
0.99 (0.97, 1.01)

210
0.85 (0.70, 1.04)
0.83 (0.67, 1.02)
0.82 (0.65, 1.02)

392
0.89 (0.75, 1.06)
0.81 (0.67, 0.98)
0.77 (0.63, 0.94)

0.134
0.019
0.009

1.00 (0.98, 1.03)

0.99 (0.96, 1.02)
0.98 (0.95, 1.01)

164
1.07 (0.82, 1.39)
1.08 (0.82, 1.42)
1.07 (0.81, 1.42)

120
1.13 (0.85, 1.49)
1.00 (0.74, 1.36)
1.02 (0.75, 1.38)

222
1.25 (0.98, 1.59)
1.07 (0.82, 1.40)
1.10 (0.84, 1.44)

0.043
0.569
0.435

1.04 (1.00, 1.07)

1.00 (0.96, 1.04)
1.00 (0.96, 1.04)

24
0.65 (0.39, 1.10)
0.61 (0.36, 1.06)
0.62 (0.36, 1.07)

79
0.87 (0.61, 1.24)
0.89 (0.61, 1.29)
0.86 (0.59, 1.25)

70
1.07 (0.74, 1.55)
0.92 (0.61, 1.37)
0.91 (0.61, 1.36)

133
1.19 (0.87, 1.63)
0.98 (0.69, 1.39)
0.98 (0.69, 1.39)

0.114
0.760
0.765

1.04 (0.99, 1.09)

0.99 (0.94, 1.04)
0.99 (0.94, 1.04)

60
1.00
1.00
1.00

26
1.35 (0.80, 2.27)
1.35 (0.79, 2.31)
1.38 (0.81, 2.38)

85
1.41 (0.95, 2.11)
1.35 (0.89, 2.07)
1.39 (0.90, 2.12)

50
1.21 (0.78, 1.88)
1.07 (0.67, 1.73)
1.13 (0.70, 1.82)

89
1.36 (0.93, 2.01)
1.20 (0.78, 1.83)
1.28 (0.83, 1.95)

0.207
0.708
0.488

1.03 (0.98, 1.09)

1.00 (0.94, 1.06)
1.01 (0.95, 1.07)

404
1.00
1.00
1.00

149
0.99 (0.81, 1.22)
1.01 (0.82, 1.25)
1.01 (0.82, 1.25)

441
1.06 (0.91, 1.23)
1.08 (0.92, 1.26)
1.08 (0.92, 1.27)

297
0.97 (0.82, 1.15)
0.98 (0.82, 1.18)
0.98 (0.82, 1.18)

510
0.97 (0.83, 1.12)
0.97 (0.82, 1.14)
0.97 (0.83, 1.14)

0.558
0.553
0.573

0.99 (0.97, 1.02)

0.99 (0.97, 1.02)
0.99 (0.97, 1.02)

1
HRs and 95% CIs were calculated by using multivariate Cox proportional hazard regression. Model 1: stratified by age at recruitment (y) and center
(Potsdam/Heidelberg) and adjusted for sex. Model 2: adjusted as for model 1 with additional adjustment for smoking (status, duration, and intensity), alcohol
intake (nonconsumers; women: .06, .612, or .12 g/d; men: .012, .1224, or .24 g/d), physical activity (average of cycling and sports during
summer and winter; in h/wk), education (none to primary school, technical to secondary school, or higher education including university), employment (yes or
no), vitamin and mineral supplement use during past 4 wk (yes or no), total energy intake (kcal/d), tea intake (cups/d), and decaffeinated coffee intake
(cups/d). Model 3: adjusted as for model 2 with additional adjustment for BMI (kg/m2), waist-to-hip ratio, and prevalent hypertension (yes or no). One cup was
defined as 150 mL. CVD, cardiovascular disease; EPIC, European Prospective Investigation into Cancer and Nutrition.

T2D may differ between subjects with or without prevalent hypertension. We furthermore observed an interaction between caffeinated coffee consumption and smoking. Among nonsmokers, an
inverse association was found between caffeinated coffee consumption and risk of T2D; among current smokers, this dosedependent inverse association was not observed. Because
smoking is an established risk factor for T2D, the adverse effects of smoking may cancel out the potential benefits of coffee
consumption on T2D risk. Thus, we speculate that smokers may
not benefit from coffee consumption in respect of T2D risk.
We found no association between coffee consumption and
CVD risk, including MI and stroke, after multivariate adjustment.
An early meta-analysis of case-control studies found a positive
association between coffee consumption and MI risk, but only
few studies adjusted for smoking or alcohol intake (10). Results
from prospective cohort studies are inconsistent. Some studies
reported that coffee consumption was independently associated
with higher risk of MI (16, 17), whereas other studies and recent
meta-analyses support evidence that there is no harmful relation

between coffee consumption and CVD mortality (4648), risk of

coronary heart disease (19, 49, 50), or specifically MI risk (51). In
fact, more recent prospective cohort studies reported an inverse
association between coffee consumption and risk of stroke (20
22). In our crude model, we observed a positive association
between both caffeinated and decaffeinated coffee consumption
and CVD risk. This is most likely explained by the unfavorable
lifestyle of high coffee consumers, especially in respect of smoking, because smoking itself represents a strong risk factor for CVD.
In line with our findings, it was previously observed that the association between coffee consumption and CVD risk may be
confounded by smoking (52, 53). We observed a positive association between decaffeinated coffee consumption and MI risk,
in the highest category of 4 cups/d compared with ,1 cup/d.
An Italian case-control study found a positive association between decaffeinated coffee consumption and MI risk (54),
whereas other studies did not find such associations (55). Because decaffeinated coffee consumption was not common in our
population, our results have to be interpreted with caution. A

Downloaded from ajcn.nutrition.org by guest on April 6, 2015

No. of subjects/person-years
Overall chronic disease
No. of cases
Model 1
Model 2
Model 3
Type 2 diabetes
No. of cases
Model 1
Model 2
Model 3
Combined CVD
No. of cases
Model 1
Model 2
Model 3
Myocardial infarction
No. of cases
Model 1
Model 2
Model 3
Stroke
No. of cases
Model 1
Model 2
Model 3
Cancer
No. of cases
Model 1
Model 2
Model 3

,1 cup/d

906

FLOEGEL ET AL

TABLE 4
HRs (and 95% CIs) for the association between decaffeinated coffee consumption and incident chronic diseases in EPIC-Germany (n = 42,659)1
Decaffeinated coffee consumption
1 to ,2 cups/d

2 to ,3 cups/d

3 to ,4 cups/d

4 cups/d

38,628/343,062

1273/11,500

1390/12,778

486/4549

882/7868

3493
1.00
1.00
1.00

136
1.18 (0.97, 1.42)
1.13 (0.93, 1.37)
1.12 (0.92, 1.36)

154
1.10 (0.92, 1.32)
1.07 (0.89, 1.28)
0.97 (0.81, 1.17)

45
0.93 (0.68, 1.28)
0.88 (0.64, 1.21)
0.81 (0.59, 1.12)

1262
1.00
1.00
1.00

46
1.14 (0.82, 1.58)
1.04 (0.74, 1.45)
0.97 (0.68, 1.39)

71
1.51 (1.15, 2.00)
1.42 (1.08, 1.88)
1.11 (0.84, 1.48)

621
1.00
1.00
1.00

28
1.47 (0.95, 2.26)
1.39 (0.89, 2.17)
1.41 (0.91, 2.21)

342
1.00
1.00
1.00

P-trend

Per cup/d

109
1.31 (1.06, 1.63)
1.17 (0.94, 1.45)
1.05 (0.84, 1.31)

0.741
0.285
0.620

1.05 (1.02, 1.08)

1.02 (0.99, 1.06)
1.00 (0.97, 1.04)

18
1.06 (0.64, 1.78)
0.96 (0.57, 1.61)
0.70 (0.41, 1.19)

35
1.16 (0.78, 1.72)
0.94 (0.63, 1.40)
0.70 (0.46, 1.06)

0.138
0.007
0.043

1.05 (1.00, 1.11)

1.01 (0.96, 1.07)
0.95 (0.90, 1.01)

27
1.10 (0.71, 1.71)
1.06 (0.68, 1.66)
1.01 (0.64, 1.58)

6
0.76 (0.33, 1.76)
0.73 (0.32, 1.69)
0.73 (0.32, 1.67)

22
1.66 (1.03, 2.65)
1.43 (0.88, 2.32)
1.36 (0.84, 2.20)

0.043
0.338
0.300

1.07 (1.00, 1.14)

1.04 (0.97, 1.10)
1.03 (0.96, 1.10)

17
1.71 (0.98, 2.96)
1.58 (0.88, 2.83)
1.59 (0.89, 2.82)

14
1.42 (0.81, 2.49)
1.38 (0.77, 2.46)
1.28 (0.72, 2.29)

4
0.99 (0.35, 2.75)
0.93 (0.34, 2.58)
0.91 (0.33, 2.50)

17
2.26 (1.30, 3.93)
2.03 (1.15, 3.58)
1.90 (1.08, 3.34)

0.114
0.597
0.575

1.12 (1.04, 1.20)

1.09 (1.01, 1.17)
1.08 (1.00, 1.16)

279
1.00
1.00
1.00

11
1.17 (0.58, 2.35)
1.15 (0.56, 2.35)
1.15 (0.56, 2.35)

13
0.86 (0.44, 1.67)
0.84 (0.43, 1.66)
0.82 (0.42, 1.61)

2
0.52 (0.12, 2.18)
0.51 (0.12, 2.18)
0.51 (0.12, 2.17)

5
0.90 (0.36, 2.27)
0.79 (0.31, 2.01)
0.76 (0.30, 1.93)

0.207
0.476
0.403

0.95 (0.83, 1.10)

0.94 (0.81, 1.08)
0.93 (0.80, 1.07)

1610
1.00
1.00
1.00

62
1.11 (0.84, 1.47)
1.13 (0.85, 1.49)
1.12 (0.85, 1.49)

56
0.83 (0.62, 1.11)
0.82 (0.61, 1.10)
0.82 (0.61, 1.10)

21
0.92 (0.58, 1.45)
0.89 (0.56, 1.42)
0.89 (0.56, 1.42)

52
1.30 (0.96, 1.76)
1.27 (0.93, 1.73)
1.27 (0.94, 1.74)

0.558
0.973
0.950

1.03 (0.99, 1.08)

1.03 (0.98, 1.08)
1.03 (0.98, 1.08)

1
HRs and 95% CIs were calculated by using multivariate Cox proportional hazard regression. Model 1: stratified by age at recruitment (y) and center
(Potsdam/Heidelberg) and adjusted for sex. Model 2: adjusted as for model 1 with additional adjustment for smoking (status, duration, and intensity), alcohol
intake (nonconsumers; women: .06, .612, or .12 g/d; men: .012, .1224, or .24 g/d), physical activity (average of cycling and sports during
summer and winter; in h/wk), education (none to primary school, technical to secondary school, or higher education including university), employment (yes or
no), vitamin and mineral supplement use during past 4 wk (yes or no), total energy intake (kcal/d), tea intake (cups/d), and caffeinated coffee intake (cups/d).
Model 3: adjusted as for model 2 with additional adjustment for BMI (kg/m2), waist-to-hip ratio, and prevalent hypertension (yes or no). One cup was defined
as 150 mL. CVD, cardiovascular disease; EPIC, European Prospective Investigation into Cancer and Nutrition.

plausible explanation for this association is not apparent, because caffeine is considered the relevant factor for a possible
positive association between coffee consumption and MI risk.
We may not have been able to account for all possible confounders associated with high decaffeinated coffee consumption,
eg, participants who are already at high CVD risk may choose to
substitute caffeinated coffee with decaffeinated coffee. Therefore, we speculate that residual confounding may cause the association we observed between high decaffeinated coffee
consumption and increased MI risk. This assumption is supported by the fact that when we excluded all cases that occurred
during the first 2 y of follow-up in the sensitivity analysis, the
positive association between decaffeinated coffee consumption
and MI risk was attenuated and no longer significant. To summarize, previous studies and our findings support the concept
that there may be no harmful effect of coffee on CVD risk per
se. Insufficient consideration of important CVD risk factors that
are linked to coffee consumption (particularly smoking) may be

responsible for the earlier belief that coffee consumption increased CVD risk.
In previous studies, coffee consumption has been found to be
associated with reduced or increased cancer risk, depending on the
type of cancer and study design. Studies investigating the effect of
coffee consumption on total cancer risk are rare. A recent metaanalysis including data from 59 prospective cohort studies found
a 3% reduction in overall cancer risk per cup of coffee consumed
(56). In our study, there was no association between caffeinated or
decaffeinated coffee consumption and total cancer risk.
The strength of our study was that we conducted a first-event
analysis taking into account 4 major chronic diseases to determine
the overall effect of coffee consumption on chronic disease risk and
additionally tested whether the risk associations may compete with
each other. Furthermore, we considered caffeinated and decaffeinated coffee as an exposure. We also included a large subsample of the German population from 2 different regions. Last,
the prospective design of our study allowed us to investigate time-

Downloaded from ajcn.nutrition.org by guest on April 6, 2015

No. of subjects/person-years
Overall chronic disease
No. of cases
Model 1
Model 2
Model 3
Type 2 diabetes
No. of cases
Model 1
Model 2
Model 3
Combined CVD
No. of cases
Model 1
Model 2
Model 3
Myocardial infarction
No. of cases
Model 1
Model 2
Model 3
Stroke
No. of cases
Model 1
Model 2
Model 3
Cancer
No. of cases
Model 1
Model 2
Model 3

,1 cup/d

COFFEE CONSUMPTION AND RISK OF CHRONIC DISEASE

We thank all of the EPIC-Germany study participants.

The authors responsibilities were as followsAF, TP, and HB: study concept and design; MMB, BT, RK, and HB: data acquisition; AF: data analysis
and interpretation and draft of the manuscript; and TP, MMB, BT, RK, and
HB: data interpretation, intellectual input, and critical review of the manuscript. All authors read and approved the final version of the manuscript. None
of the authors declared a conflict of interest.

8.

9.

10.

11.
12.

13.

14.

15.
16.

17.

18.

19.

20.

21.
22.

23.

24.

25.

REFERENCES
1. Grigg D. The worlds of tea and coffee: patterns of consumption. GeoJournal 2003;57:28394.
2. International Coffee Organization. Historical coffee statistics. London,
United Kingdom: International Coffee Organization, 2008. Available
from: http://earthtrends.wri.org/searchable_db/results.php?years=-1&
variable_ID=294&theme=8&cID=62,70,190&ccID= (cited 4 April 2011).
3. Higdon JV, Frei B. Coffee and health: a review of recent human research. Crit Rev Food Sci Nutr 2006;46:10123.
4. Salter H. On some points in the treatment and clinical history of
asthma. Edinburgh Med J 1859;4:110915.
5. Jee SH, He J, Whelton PK, Suh I, Klag MJ. The effect of chronic coffee
drinking on blood pressure: a meta-analysis of controlled clinical trials.
Hypertension 1999;33:64752.
6. Geleijnse JM. Habitual coffee consumption and blood pressure: an
epidemiological perspective. Vasc Health Risk Manag 2008;4:96370.
7. Fried RE, Levine DM, Kwiterovich PO, Diamond EL, Wilder LB, Moy
TF, Pearson TA. The effect of filtered-coffee consumption on plasma

26.

27.

28.

29.

30.

31.

lipid levels. Results of a randomized clinical trial. JAMA 1992;267:

8115.
Nygard O, Refsum H, Ueland PM, Stensvold I, Nordrehaug JE, Kvale
G, Vollset SE. Coffee consumption and plasma total homocysteine: the
Hordaland Homocysteine Study. Am J Clin Nutr 1997;65:13643.
Panagiotakos DB, Pitsavos C, Zampelas A, Zeimbekis A, Chrysohoou
C, Papademetriou L, Stefanadis C. The association between coffee
consumption and plasma total homocysteine levels: the ATTICA
study. Heart Vessels 2004;19:2806.
Kawachi I, Colditz GA, Stone CB. Does coffee drinking increase the
risk of coronary heart disease? Results from a meta-analysis. Br Heart J
1994;72:26975.
Greenland S. A meta-analysis of coffee, myocardial infarction, and
coronary death. Epidemiology 1993;4:36674.
Panagiotakos DB, Pitsavos C, Chrysohoou C, Kokkinos P, Toutouzas P,
Stefanadis C. The J-shaped effect of coffee consumption on the risk of
developing acute coronary syndromes: the CARDIO2000 case-control
study. J Nutr 2003;133:322832.
Mostofsky E, Schlaug G, Mukamal KJ, Rosamond WD, Mittleman
MA. Coffee and acute ischemic stroke onset: the Stroke Onset Study.
Neurology 2010;75(18):15838.
Baylin A, Hernandez-Diaz S, Kabagambe EK, Siles X, Campos H.
Transient exposure to coffee as a trigger of a first nonfatal myocardial
infarction. Epidemiology 2006;17:50611.
Butt MS, Sultan MT. Coffee and its consumption: benefits and risks.
Crit Rev Food Sci Nutr 2011;51(4):36373.
Happonen P, Voutilainen S, Salonen JT. Coffee drinking is dosedependently related to the risk of acute coronary events in middle-aged
men. J Nutr 2004;134(9):23816.
LaCroix AZ, Mead LA, Liang KY, Thomas CB, Pearson TA. Coffee
consumption and the incidence of coronary heart disease. N Engl J
Med 1986;315:97782.
Sofi F, Conti AA, Gori AM, Eliana Luisi ML, Casini A, Abbate R,
Gensini GF. Coffee consumption and risk of coronary heart disease:
a meta-analysis. Nutr Metab Cardiovasc Dis 2007;17:20923.
Wu JN, Ho SC, Zhou C, Ling WH, Chen WQ, Wang CL, Chen YM.
Coffee consumption and risk of coronary heart diseases: a meta-analysis
of 21 prospective cohort studies. Int J Cardiol 2009;137(3):21625.
Lopez-Garcia E, Rodriguez-Artalejo F, Rexrode KM, Logroscino G, Hu
FB, van Dam RM. Coffee consumption and risk of stroke in women.
Circulation 2009;119:111623.
Larsson SC, Virtamo J, Wolk A. Coffee consumption and risk of stroke
in women. Stroke 2011;42(4):90812.
Larsson SC, Mannisto S, Virtanen MJ, Kontto J, Albanes D, Virtamo J.
Coffee and tea consumption and risk of stroke subtypes in male smokers.
Stroke 2008;39(6):16817.
Goto A, Song Y, Chen BH, Manson JE, Buring JE, Liu S. Coffee and
caffeine consumption in relation to sex hormone-binding globulin and
risk of type 2 diabetes in postmenopausal women. Diabetes 2011;60(1):
26975.
Schulze MB, Hoffmann K, Boeing H, Linseisen J, Rohrmann S, Mohlig
M, Pfeiffer AF, Spranger J, Thamer C, Haring HU, et al. An accurate risk
score based on anthropometric, dietary, and lifestyle factors to predict the
development of type 2 diabetes. Diabetes Care 2007;30:5105.
van Dam RM, Feskens EJ. Coffee consumption and risk of type 2 diabetes mellitus. Lancet 2002;360:14778.
Rosengren A, Dotevall A, Wilhelmsen L, Thelle D, Johansson S.
Coffee and incidence of diabetes in Swedish women: a prospective
18-year follow-up study. J Intern Med 2004;255:8995.
Pereira MA, Parker ED, Folsom AR. Coffee consumption and risk of
type 2 diabetes mellitus: an 11-year prospective study of 28 812 postmenopausal women. Arch Intern Med 2006;166:13116.
Salazar-Martinez E, Willett WC, Ascherio A, Manson JE, Leitzmann MF,
Stampfer MJ, Hu FB. Coffee consumption and risk for type 2 diabetes
mellitus. Ann Intern Med 2004;140:18.
Iso H, Date C, Wakai K, Fukui M, Tamakoshi A. The relationship between green tea and total caffeine intake and risk for self-reported type 2
diabetes among Japanese adults. Ann Intern Med 2006;144:55462.
Johnson S, Koh WP, Wang R, Govindarajan S, Yu MC, Yuan JM. Coffee
consumption and reduced risk of hepatocellular carcinoma: findings from
the Singapore Chinese Health Study. Cancer Causes Control 2011;22:
50310.
Tang N, Zhou B, Wang B, Yu R. Coffee consumption and risk of breast
cancer: a metaanalysis. Am J Obstet Gynecol 2009;200(3):290 e19.

Downloaded from ajcn.nutrition.org by guest on April 6, 2015

dependent exposure-disease associations. However, our study also

had some limitations. Coffee consumption was assessed by FFQ
at the time of recruitment and relied on participants self-report.
Nevertheless, data from the Nurses Health Study suggest that
coffee consumption is very stable in an individual over time, because it showed the best reproducibility among all food items in
repeated measures (57). As previously described (58), filtering may
alter coffees metabolic effects. The method of coffee preparation
was not assessed in our study population, and we were therefore not
able to study whether the association between coffee consumption
and chronic disease risk depends on the filtering of coffee. In
Germany, filtering is the traditional method of coffee preparation
(59); thus, our findings most likely apply to filtered coffee. Because
of the small number of decaffeinated coffee consumers in our
population, there were also fewer cases for some of the disease
endpoints. Hence, the associations found will benefit from further
investigation. Our study sample was drawn from the general
population; however, because the response rates were low, it
overrepresents health-conscious people with higher socioeconomic status (37). Because this was an observational study, we
can only show associations. Whether these associations represent
true causal effects has to be confirmed in future studies. Last, it
has to be noted that we were primarily interested in investigating
the association between coffee consumption and risk of major
chronic diseases. We cannot exclude the possibility that coffee consumption may be associated with risk of certain disease subtypes, eg,
site-specific cancers. This should be the focus of future studies.
In conclusion, our study does not support the hypothesis that
coffee is detrimental to human health. In fact, we observed an
inverse association between coffee consumption and T2D risk,
which challenges the evidence base for current recommendations.
We speculate that the unfavorable lifestyle characteristics of high
coffee consumers, especially with respect to smoking, may be responsible for the positive association between coffee consumption
and CVD risk, which was reported in early studies. In addition,
coffee and cigarettes may be an unfavorable combination, which
requires further investigation.

907

908

FLOEGEL ET AL

47.
48.
49.
50.

51.
52.

53.
54.
55.
56.
57.

58.
59.

green tea, black tea and oolong tea consumption and risk of mortality
from cardiovascular disease in Japanese men and women. J Epidemiol
Community Health 2011;65(3):23040.
Lopez-Garcia E, Rodriguez-Artalejo F, Li TY, Mukamal KJ, Hu FB,
van Dam RM. Coffee consumption and mortality in women with cardiovasculardisease. Am J Clin Nutr 2011;94(1):21824.
Lopez-Garcia E, van Dam RM, Li TY, Rodriguez-Artalejo F, Hu FB.
The relationship of coffee consumption with mortality. Ann Intern Med
2008;148(12):90414.
Willett WC, Stampfer MJ, Manson JE, Colditz GA, Rosner BA,
Speizer FE, Hennekens CH. Coffee consumption and coronary heart
disease in women. A ten-year follow-up. JAMA 1996;275:45862.
Lopez-Garcia E, van Dam RM, Willett WC, Rimm EB, Manson JE,
Stampfer MJ, Rexrode KM, Hu FB. Coffee consumption and coronary
heart disease in men and women: a prospective cohort study. Circulation 2006;113(17):204553.
Rosner SA, Akesson A, Stampfer MJ, Wolk A. Coffee consumption
and risk of myocardial infarction among older Swedish women. Am J
Epidemiol 2007;165(3):28893.
Nilsson LM, Wennberg M, Lindahl B, Eliasson M, Jansson JH, Van
Guelpen B. Consumption of filtered and boiled coffee and the risk of
first acute myocardial infarction; a nested case/referent study. Nutr
Metab Cardiovasc Dis 2010;20(7):52735.
Klatsky AL, Koplik S, Kipp H, Friedman GD. The confounded relation
of coffee drinking to coronary artery disease. Am J Cardiol 2008;101(6):
8257.
La Vecchia C, DAvanzo B, Negri E, Franceschi S, Gentile A, Tavani
A. Decaffeinated coffee and acute myocardial infarction. A casecontrol study in Italian women. Ann Epidemiol 1993;3:6014.
Sesso HD, Gaziano JM, Buring JE, Hennekens CH. Coffee and tea intake
and the risk of myocardial infarction. Am J Epidemiol 1999;149:1627.
Yu X, Bao Z, Zou J, Dong J. Coffee consumption and risk of cancers:
a meta-analysis of cohort studies. BMC Cancer 2011;11:96.
Colditz GA, Willett WC, Stampfer MJ, Sampson L, Rosner B,
Hennekens CH, Speizer FE. The influence of age, relative weight,
smoking, and alcohol intake on the reproducibility of a dietary questionnaire. Int J Epidemiol 1987;16:3928.
Jee SH, He J, Appel LJ, Whelton PK, Suh I, Klag MJ. Coffee consumption and serum lipids: a meta-analysis of randomized controlled
clinical trials. Am J Epidemiol 2001;153:35362.
Weigel T. Daten und Fakten zur Kaffeeindustrie [Data and facts about
the coffee industry]. Statista, 2011 (in German). Available from: http://
de.statista.com/statistik/faktenbuch/151/a/branche-industrie-markt/
lebensmittelindustrie/kaffeeindustrie/ (cited 17 November 2011).

Downloaded from ajcn.nutrition.org by guest on April 6, 2015

32. Turati F, Galeone C, La Vecchia C, Garavello W, Tavani A. Coffee and

cancers of the upper digestive and respiratory tracts: meta-analyses of
observational studies. Ann Oncol 2011;22(3):53644.
33. van Dam RM. Coffee consumption and risk of type 2 diabetes, cardiovascular diseases, and cancer. Symposium. Appl Physiol Nutr Metab
2008;33:126983.
34. Tavani A, La Vecchia C. Coffee and cancer: a review of epidemiological studies, 1990-1999. Eur J Cancer Prev 2000;9(4):24156.
35. Lopez-Garcia E. Coffee consumption, myocardial infarction and stroke:
what is the association? Womens Health (Lond Engl) 2011;7:2657.
36. Woodward M, Tunstall-Pedoe H. Coffee and tea consumption in the
Scottish Heart Health Study follow up: conflicting relations with coronary risk factors, coronary disease, and all cause mortality. J Epidemiol Community Health 1999;53:4817.
37. Boeing H, Korfmann A, Bergmann MM. Recruitment procedures of
EPIC-Germany. European Investigation into Cancer and Nutrition. Ann
Nutr Metab 1999;43(4):20515.
38. Boeing H, Wahrendorf J, Becker N. EPIC-Germanya source for
studies into diet and risk of chronic diseases. European Investigation
into Cancer and Nutrition. Ann Nutr Metab 1999;43(4):195204.
39. Bergmann MM, Bussas U, Boeing H. Follow-up procedures in EPICGermanydata quality aspects. European Prospective Investigation into
Cancer and Nutrition. Ann Nutr Metab 1999;43(4):22534.
40. World Health Organization (WHO). International statistical classification of diseases and related health problems (ICD-10). 10th revision.
Geneva, Switzerland: WHO, 1992.
41. Huxley R, Lee CM, Barzi F, Timmermeister L, Czernichow S, Perkovic
V, Grobbee DE, Batty D, Woodward M. Coffee, decaffeinated coffee,
and tea consumption in relation to incident type 2 diabetes mellitus:
a systematic review with meta-analysis. Arch Intern Med 2009;169(22):
205363.
42. Lunn M, McNeil D. Applying Cox regression to competing risks.
Biometrics 1995;51:52432.
43. Glynn RJ, Rosner B. Comparison of risk factors for the competing risks
of coronary heart disease, stroke, and venous thromboembolism. Am J
Epidemiol 2005;162(10):97582.
44. American Heart Association. Caffeine-AHA recommendation. Available
from: http://www.heart.org/HEARTORG/GettingHealthy/NutritionCenter/
HealthyDietGoals/Caffeine-and-Cardiovascular-Disease_UCM_305888_
Article.jsp (cited 19 July 2011).
45. van Dam RM, Hu FB. Coffee consumption and risk of type 2 diabetes:
a systematic review. JAMA 2005;294:97104.
46. Mineharu Y, Koizumi A, Wada Y, Iso H, Watanabe Y, Date C, Yamamoto
A, Kikuchi S, Inaba Y, Toyoshima H, et al; JACC study group. Coffee,