Page 1 of 29 International Journal of Cancer

Title Page:

Coffee drinking and risk of endometrial cancer: findings from a large up-to-
date meta-analysis

Authors:
Youjin Je1 and Edward Giovannucci1,2,3
1
Department of Nutrition, Harvard School of Public Health, Boston, Massachusetts
2
Department of Epidemiology, Harvard School of Public Health, Boston, Massachusetts
3
Channing Laboratory, Department of Medicine, Brigham and Women’s Hospital and Harvard
Medical School, Boston, Massachusetts

Several epidemiological studies have examined the association between coffee drinking and risk
of endometrial cancer. To provide a quantitative assessment of this association, we conducted a
meta-analysis of observational studies published up to October 2011 through a search of
MEDLINE and EMBASE databases and the reference lists of retrieved article. Pooled relative
risks (RRs) with 95% confidence intervals (CIs) were calculated using a random-effects model,
and generalized least square trend estimation was used to assess dose-response relationships. A
total of 16 studies (10 case-control and 6 cohort studies) on coffee intake with 6,628 endometrial
cancer cases were included in the meta-analysis. The pooled RR of endometrial cancer for the
highest versus lowest categories of coffee intake was 0.71 (95% CI: 0.62-0.81; p for
heterogeneity = 0.13). By study design, the pooled RRs were 0.69 (95% CI: 0.55-0.87) for case-
control studies and 0.70 (95% CI: 0.61-0.80) for cohort studies. By geographic region, the inverse
association was stronger for 3 Japanese studies (pooled RR=0.40; 95% CI: 0.25-0.63) than 5
studies from USA/Canada (pooled RR=0.69; 95% CI: 0.60-0.79) or 8 studies from Europe
(pooled RR=0.79; 95% CI: 0.63-0.99). An increment of 1 cup/d of coffee intake conferred a
pooled RR of 0.92 (95% CI: 0.90-0.95). In conclusion, our findings suggest that increased coffee
intake is associated with a reduced risk of endometrial cancer, consistently observed for cohort
and case-control studies. More large studies are needed to determine subgroups to obtain more
benefits from coffee drinking in relation to endometrial cancer risk.

Keywords: coffee, dietary factors, hyperinsulinemia, endometrial cancer; observational study,

meta-analysis

Corresponding author:
Youjin Je
Department of Nutrition
Harvard School of Public Health
665 Huntington Avenue, Boston, MA, 02115
Fax: 617-432-2435
Email: yje@hsph.harvard.edu

John Wiley & Sons, Inc.

 International Journal of Cancer Page 2 of 29

Abstract: 249 words

Text: 3,894 words
References: 50
Tables/Figures: 2 tables, 3 figures

John Wiley & Sons, Inc.

Page 3 of 29 International Journal of Cancer

Endometrial cancer is the most common gynecologic malignancy in the United States1, and the

disease has been associated with obesity, elevated levels of estrogen, and insulin which cause the

growth of endometrial cells1, 2, 3. Although regular exercise and maintenance of a healthy weight

are probably the most important ways to prevent the disease4, dietary habits may provide an

additional preventive strategy against endometrial cancer risk.

Coffee is one of the most widely consumed beverages in the world. It is a complex

mixture of more than a thousand chemicals. These constituents have potential antimutagenic and

antioxidant properties5, and thus coffee drinking may reduce total cancer incidence6. Several

studies suggested an important role of coffee consumption on hormonal modulation. It has been

reported that coffee consumption was inversely associated with circulating levels of estrogen and

C-peptide, a marker of insulin secretion, and positively associated with sex hormone-binding

globulin (SHBG) and plasma adiponectin, an insulin sensitizer7-11. In addition, coffee

consumption has been inversely associated with risk of diabetes, a disease which is preceded by

many years of increasing insulin resistance and subsequent hyperinsulinemia12. Thus, coffee

consumption has been hypothesized to play an important role in reducing endometrial cancer risk.

Several epidemiological studies have been conducted to determine the association

between coffee consumption and risk of endometrial cancer. A previous meta-analysis of this

subject combined data from 9 observational studies, 5 studies from Europe, 3 studies from Japan,

and 1 study from Canada, and found an inverse association with high coffee consumption among

all studies and case-control studies, but failed to reach statistical significance for cohort studies13.

Only two prospective cohort studies from Norway (n=84 cases)14 and Japan (n=117 cases)15 were

included in the previous meta-analysis. Although results of case-control studies suggested a

beneficial role of coffee consumption in relation to endometrial cancer risk, prospective data to

John Wiley & Sons, Inc.

 International Journal of Cancer Page 4 of 29

exclude some methodological biases that may exist in retrospective data are needed to help draw

more definitive conclusions.

Since the meta-analysis, several epidemiologic studies have examined the association

between coffee consumption and endometrial cancer risk, but the results were inconsistent. Two

case-control studies conducted in the United States showed a nonsignificant inverse association

with greater than 2 cups of coffee per day compared to no coffee consumption16, 17. Three large

prospective cohort studies showed a significant risk reduction in endometrial cancer among high

coffee drinkers18-20. The Swedish Mammography Cohort study found risk reductions in the coffee

categories of 2-3 cups/day and • 4 cups/day compared to those consuming 1 cup or less18; the

NIH-AARP Diet and Health Study in the United States found risk reductions in the coffee

category of 1 cup/day through the coffee category of > 3 cups/day compared to no

consumption19; and the Nurses’ Health Study in the United States found a risk reduction in

women who consumed 4 or more cups of coffee/day compared to those who consumed less than

1 cup/day20. Another recent cohort study conducted in Sweden found no significant inverse

association among women who consumed 1-3 or • 4 cups of coffee daily, compared to those who

consumed less than 1 cup/day21. Therefore, to provide a quantitative assessment of this

association, we systematically conducted a meta-analysis by combining all available data of both

case-control and cohort studies.

Material and Method

Search strategy and study selection

John Wiley & Sons, Inc.

Page 5 of 29 International Journal of Cancer

We searched the electronic databases MEDLINE and EMBASE to identify eligible studies

published in English through October 2011, including articles ahead of publication. The

following keywords were used in searching: “(caffeine, coffee, beverages, dietary factors, or risk

factors) combined with (corpus uterian or endometrial neoplasms).” We also performed a manual

search of references cited in the selected articles and published reviews to search for additional

relevant studies. We included those studies that met the following criteria: 1) Studies that

presented original data from case-control or cohort studies; 2) The outcome of interest was

clearly defined as endometrial cancer incidence; 3) The exposure of interest was coffee

consumption. If studies provided combined data on coffee and tea consumption, they were not

included22; and 4) Studies that provided relative risk estimates and their confidence intervals or

sufficient data to calculate them (e.g., number of cases and controls in exposure categories).

Data extraction

Two investigators (Y.J. and E.G.) extracted data independently, according to the meta-analysis of

observational studies in epidemiology (MOOSE) guidelines23, and discrepancies were resolved

via referencing the original reports and discussion. For each included study, the following

information was extracted: first author’s last name, publication year, study design, type of control,

country of origin, follow-up period or study period, number of cases and controls/subjects or

person-term, adjustment for potential confounders, relative risks, odds ratios from case-control

studies and rate ratios from cohort studies, and 95% confidence intervals for association between

coffee consumption and endometrial cancer incidence, considering various exposure levels or per

unit change in exposure (e.g., 1 cup increment per day).

John Wiley & Sons, Inc.

 International Journal of Cancer Page 6 of 29

If 95% confidence intervals were not reported, but numbers of cases and controls (or

person-time) in exposure categories were provided, these data were used to calculate the standard

error of the crude relative risk, and then approximate confidence intervals for the reported

adjusted relative risk14, 24. For studies which provided several relative risks from age-adjusted

model to different multivariate models, we used the relative risks from multivariate models with

the most complete adjustment for potential confounders. If relative risk was reported as per unit

change in coffee consumption only25, we calculated relative risk and 95% confidence interval for

the mean of the highest coffee categories from studies with same country of origin14, 18, 21, 26-28.

Statistical analysis

We combined a study-specific log (odds ratio) for case-control studies and log (rate ratio) for

cohort studies to compute a pooled relative risk (RR) and its 95% confidence interval (CI) for the

highest versus lowest category of coffee consumption from original studies, using the

DerSimonian and Laird random-effects models, which incorporates both within- and between-

study variations29. Because endometrial cancer is rare, odds ratio in case-control studies and rate

ratio in cohort studies yield similar estimates of RR. The summary measures were presented as

forest plots where the size of data markers (squares) corresponds to the inverse of the variance of

the natural logarithm of RR from each study, and the diamond indicates pooled RR. Statistical

heterogeneity among studies was estimated using Q and I2 statistics30, 31. For the Q statistic,

heterogeneity was considered for p < 0.10. We performed a sensitivity analysis in which one

study at a time was removed and the rest analyzed to evaluate whether the results could have been

affected markedly by a single study. Since various sources of heterogeneity may exist due to

international differences in coffee drinking habit or some factors that can affect the hormonal

John Wiley & Sons, Inc.

Page 7 of 29 International Journal of Cancer

modulation of coffee, we also conducted analyses stratified by geographic regions (Europe,

United States/Canada, and Japan). To test for variation in risk estimates by study design,

geographic region, or other potential source of heterogeneity, we carried out s meta-regression

analysis.

To examine dose-response relationships among different categories of coffee intake, we

used generalized least-squares trend estimation (GLST) analysis based on the method developed

by Greenland and Longnecker32-34; study-specific slopes from the correlated natural logarithm of

the RRs across coffee categories in each study were estimated14-17, 21, 27, 35, 38, 39
, and then we

combined the GLST-estimated study specific slopes with studies that reported slope estimates18-20,
25, 28
to derive an overall average slope. The highest, open-ended category was assumed to have

the same amplitude of intake as the preceding category. For a study which used units other than

cups, we converted these into cups per day as a standard measure (e.g., 240g = 1 cup)35. Two

studies that did not provide cutoff or median of coffee intake in each category24 or had only two

categories of coffee intake26 were not included for this analysis. We also examined a potential

nonlinear dose-response relationship between coffee intake and endometrial cancer by adding a

quadratic term of coffee intake in the model. A p value for nonlinearity was calculated by testing

the null hypothesis that the coefficient of the quadratic term is equal to 0.

Finally, publication bias was evaluated through visual inspection of a funnel plot (i.e., a

plot of study results against precision) and with the Begg’s36 and Egger’s tests37. A two-tailed p

value of less than 0.05 was considered statistically significant. All statistical analyses were

performed by using Stata/SE version 12.0 software (Stata Corporation, College Station, Texas).

John Wiley & Sons, Inc.

 International Journal of Cancer Page 8 of 29

RESULTS

Study characteristics

We identified a total of 16 observational studies including 6,628 incident cases of endometrial

cancer that were potentially eligible for inclusion in the meta-analysis14-21, 24-28, 35, 38, 39
. The

characteristics of the included studies are summarized in Table 1. By study design, 10 case-

control studies (n=3,484 cases)16, 17, 24-28, 35, 38, 39 and 6 prospective cohort studies (n=3,144

cases)14, 15, 18-21 of association between coffee consumption and risk of endometrial cancer were

included. Of 10 case-control studies, 6 used hospital-based controls (n=1,699 cases)16, 24- 26, 28, 38

and 4 used population-based controls (n=1,785 cases)17, 27, 35, 39. By geographic region, 8 studies

were conducted in Europe (n=2,535 cases)14, 18, 21, 24-28, 5 in the United States/Canada (n=3,640

cases)16, 17, 19, 20, 35, and 3 in Japan (n=453 cases)15, 38, 39. Most studies provided risk estimates that

were adjusted for body mass index (BMI, kg/m2) 15-21, 25-28, 35, 38, 39
and smoking14-21, 25, 27, 35, 38, 39,

which are the most likely confounder of the relationship between coffee intake and endometrial

cancer. Four studies (2 case-control and 2 cohort studies) provided data on decaffeinated coffee

consumption16, 19, 20, 24.

High versus low coffee consumption

The multivariable-adjusted RRs in each study and the pooled RR from all studies combined for

the highest versus lowest categories of coffee intake are presented in Figure 1 and Table 2. The

pooled RR of endometrial cancer for the highest versus lowest categories of coffee intake was

0.71 (95% CI: 0.62-0.81). Stratifying by study design, the pooled RRs for case-control studies

and cohort studies were 0.69 (95% CI: 0.55-0.87) and 0.70 (95% CI: 0.61-0.80), respectively. No

significant differences were found by the study design (p = 0.96) or type of controls among case-

John Wiley & Sons, Inc.

Page 9 of 29 International Journal of Cancer

control studies (p = 0.58). Stratifying by geographic region, the pooled RR was strongest for

studies conducted in Japan (pooled RR=0.40, 95% CI: 0.25-0.63) compared to the pooled RRs for

studies conducted in Europe (pooled RR=0.79, 95% CI: 0.63-0.99) or the United States/Canada

(pooled RR=0.69, 95% CI: 0.60-0.79) (p for studies in Japan vs. non-Japan = 0.03) (Figure 2 and

Table 2). There was no statistically significant heterogeneity among the studies of coffee intake

overall (p = 0.13), but there was some evidence of heterogeneity among case-control studies (p =

0.04). A sensitivity analysis of omitting one study at a time and calculating pooled RRs for the

remainder of the studies showed that the study by Levi et al.24 had the most influence on the

pooled RR. After excluding this single study, there was no study heterogeneity (p = 0.67,

I2=0.0% overall; p = 0.52, I2=0.0% for case-control studies), and pooled RRs for the highest

versus lowest category of coffee intake were 0.68 (95% CI: 0.61-0.75) overall and 0.64 (95% CI:

0.54-0.76) for case-control studies. Within the same geographic region, there was no statistically

significant heterogeneity among studies of coffee intake (Europe, p = 0.11; United States/Canada,

p = 0.93; Japan, p = 0.99).

When limited to 12 studies that had adjusted for BMI and smoking15-21, 25, 27, 35, 38, 39, the

pooled RR was 0.69 (95% CI: 0.62-0.77), which is close to the pooled RR from all studies. We

examined the possibility that different cutoffs for the highest categories of coffee intake may

affect the overall result using studies with available data. The pooled RR of studies using 3 or less

cup of coffee as a cutoff for highest category was 0.60 (95% CI: 0.48-0.74)15-17, 26, 35, 38, 39, while

the pooled RR of those using greater than 3 cups of coffee as a cutoff was a 0.70 (95% CI: 0.61-

0.79)14, 18-21, 27, 28 (p value for difference = 0.87).

Dose-response meta-analysis

John Wiley & Sons, Inc.

 International Journal of Cancer Page 10 of 29

10

Fourteen studies were included for the dose-response analysis of coffee intake and risk of

endometrial cancer14-21, 25, 27, 28, 35, 38, 39. We found no evidence of statistically significant departure

from linearity (p = 0.53). The pooled RR of endometrial cancer for a 1cup/d increment in coffee

intake was 0.92 (95% CI: 0.90-0.95), similarly for cohort studies (pooled RR= 0.94; 95% CI:

0.90-0.97) and case-control studies (pooled RR=0.90; 95% CI: 0.86-0.95) (Table 2). The pooled

RR for a 1 cup/d increment of coffee intake was strongest for studies conducted in Japan (pooled

RR= 0.76; 95% CI: 0.68-0.86) compared to pooled RRs for studies in Europe (pooled RR=0.93,

95% CI: 0.90-0.97) or United States/Canada (pooled RR=0.94, 95% CI: 0.91-0.97) (p for studies

in Japan vs. non-Japan = 0.006).

Publication Bias

There was no indication of publication bias from either visualization of the funnel plot, Begg’s

test (p = 0.34), or Egger’s test (p = 0.33) (Figure 3). For the RRs for a 1 cup/d increment in

coffee intake with 14 studies, there was no significant evidence of publication bias (Begg’s p =

0.08; Egger’s p = 0.07).

Discussion

The current meta-analysis evaluated the potential association between coffee consumption and

endometrial cancer risk, based on 16 observational studies, 10 case-control and 6 cohort studies,

with a total of 6,628 cases. The results indicate that increased intake of coffee is associated with a

reduced risk of endometrial cancer, consistently observed for cohort and case-control studies.

Overall, the risk of endometrial cancer decreased by 29% for high (median of the highest

John Wiley & Sons, Inc.

Page 11 of 29 International Journal of Cancer

11

categories: 3-4 cups/day) versus low/no coffee intake, and by 8% for a 1 cup/d increment in

coffee intake. The association seems to be stronger among the Japanese population that was

characterized by much lower endometrial cancer incidence compared to white women in North

America or Europe40.

Overall, no significant heterogeneity was found among studies. The observed

heterogeneity among case-control studies of coffee intake and endometrial cancer seemed to be

explained by 1 case-control study in Europe24. After exclusion of this single study, there was no

evidence of heterogeneity. This study was published in the earliest year, and had the least

adjustment for potential confounders.

There are several potential mechanisms through which coffee consumption may lower the

risk of endometrial cancer. Besides providing a source of caffeine, coffee contains a host of

bioactive components. Several of the components in coffee including chlorogenic acid, the most

abundant polyphenol in coffee, have strong antioxidant properties that can prevent oxidative

DNA damage, improve insulin sensitivity, and inhibit glucose absorption in the intestine41.

Caffeine and some bioactive compounds in coffee seem to increase the clearance of estradiol

overall, or even synthesize estrogen metabolites that may inhibit estradiol-mediated

carcinogenesis on endometrial cells42. Several cross-sectional studies showed that high coffee

consumption (including decaffeinated coffee) was associated with lower circulating levels of C-

peptide10 and higher levels of adiponectin11. In addition, high coffee consumption was associated

with higher levels of SHBG which is likely to decrease bioavailable estrogen7-9. These favorable

effects of coffee consumption on the hormones seemed to be stronger among obese women

relative to lean women9, 10.

John Wiley & Sons, Inc.

 International Journal of Cancer Page 12 of 29

12

Due to few studies that examined decaffeinated coffee intake in relation to endometrial

cancer16, 19, 20, 24, we did not conduct a meta-analysis of decaffeinated coffee intake. Two case-

control studies conducted in Europe24 and the United States16 showed no association with high

decaffeinated coffee intake, while two recent large prospective cohort studies in the United States

suggested an inverse association with high decaffeinated coffee intake19, 20. The NIH-AARP Diet

and Health Study showed a 34% lower risk of endometrial cancer among women who consumed

2-3 cups of decaffeinated coffee per day compared to no consumption (RR=0.66, 95% CI: 0.51-

0.85; p for trend = 0.004) excluding participants who reported drinking regular coffee19. The

Nurses’ Health Study showed a modest inverse association with 2 or more cups of long-term

decaffeinated coffee per day compared to no consumption (RR=0.78, 95% CI: 0.57-1.08), which

was slightly greater when excluding cases diagnosed during the first 2-year follow-up period

(RR=0.72, 95% CI: 0.52-1.01)20. The RRs for a 1-cup increment of decaffeinated coffee per day

were 0.93 (95% CI: 0.87-0.99)19 and 0.93 (95% CI: 0.84-1.02)20, respectively. Relatively low

frequency of decaffeinated coffee intake and its short term use in Europe or Japan may limit the

ability to evaluate its association with endometrial cancer.

The current meta-analysis had some advantages. First, the number of total cases included

in the meta-analysis was substantial (n=6,628 cases). The pooled RRs with high or increasing

coffee intake for risk of endometrial cancer were consistent with those in the earlier meta-analysis

which contained 2,409 cases13. While the previous meta-analysis, which had limited prospective

data (2 studies, n=201 cases), showed a non-significant inverse association among cohort studies,

we observed significant inverse associations with high or increasing coffee intake among cohort

studies (6 studies, n=3,144 cases). Second, although cutoffs for the highest coffee categories

varied among studies included in our meta-analysis, we found no significant difference in the

John Wiley & Sons, Inc.

Page 13 of 29 International Journal of Cancer

13

pooled RRs stratified by studies using > 3 cups/day or ” 3 cups/day as a cutoff. Furthermore, we

also found some evidence of a dose-response relationship between coffee intake and endometrial

cancer. Third, in a meta-analysis of published studies, publication bias could be of concern

because small studies with null results tend not to be published. In this meta-analysis, however,

we found little evidence of publication bias.

Interestingly, we found a strong inverse association of high coffee intake or increased

coffee intake in Japanese women compared to those in Europe or United States/Canada. This

difference may be explained by low prevalence of hormone replacement therapy (HRT) use in

Japanese population, i.e., a low exogenous estrogen environment43, 44, where the potential role of

coffee to lower endogenous estrogen levels might be more evident. Several studies on the

association between diet and endometrial cancer found a more pronounced association among

never users of HRT45-47 or women who were not currently using estrogen containing hormones48.

The large Swedish cohort study showed a slightly greater inverse association with each additional

cup of coffee among HRT non-users (RR=0.86, 95% CI: 0.75-0.99) than users (RR=0.93, 95%

CI: 0.83-1.06)18. The NIH-AARP Diet and Health study in the United States showed a

significantly greater inverse association with each additional cup of coffee among HRT non-users

(RR=0.92, 95% CI: 0.88-0.96) than users (RR=0.97, 95% CI: 0.92-1.03) (p for interaction =

0.03)19. The Nurses’ Health Study in the United States showed a significant linear trend of

decreasing risk of endometrial cancer with increasing coffee intake in women who were not

currently using HRT (p for trend = 0.03), but not in current users (p for trend = 0.68)20. In

addition, a prospective cohort study of insulin and endometrial cancer reported that the positive

association of insulin with endometrial cancer was greatest among women not using HRT at

John Wiley & Sons, Inc.

 International Journal of Cancer Page 14 of 29

14

baseline49, and demonstrated that HRT users had low levels of insulin and IGF-I due to the

hepatic first pass effect of oral HRT leading to suppression of insulin/IGF axis protein synthesis19.

In terms of BMI, there was some evidence from three large cohort studies conducted in

Sweden18 and the United States19, 20 that overweight or obese women at high risk of endometrial

cancer had more benefits from high coffee drinking than normal weight women, which is

consistent with the notion that insulin resistance and hyperinsulinemia may be involved in the

process. Since obesity was less prevalent among Japanese population compared to American or

European women, BMI may not explain the difference in results from Japan and the other

countries. Two case-control studies conducted in the United States16 and Japan38 showed some

tendency of stronger inverse association among lean women than overweight women, one of

which might be by chance because of small number of drinkers at the highest level of intake

among leaner women (n=2 cases) 38. Furthermore, we cannot exclude the possibilities that coffee

may be involved in different mechanisms of sex hormones or insulin in Asian women.

Despite these advantages, some limitations of the current meta-analysis should be

acknowledged. First, misclassification of coffee intake and thus misclassified coffee components

are inevitable due to self-reported intake and lack of specific data on type of raw coffee, roasting,

or preparation that can vary substantially within and between countries. Any remaining

misclassification, however, would have likely biased the results toward the null. Second, a meta-

analysis is unable to solve problems with confounding factors that could be inherent in the

included studies. Smoking and BMI are potentially the most likely confounders of the

relationship between coffee intake and endometrial cancer. Coffee intake tended to be positively

related to smoking and negatively related to BMI15, 18-20. Although inadequate control for those

variables might lead to an overestimation of the risk estimates, our additional analysis limited to

John Wiley & Sons, Inc.

Page 15 of 29 International Journal of Cancer

15

studies that had adjusted for smoking and BMI showed similar results. Third, although many of

studies had adjusted for important risk factors for endometrial cancer and the additional analysis

described above resulted in essentially the same estimate, unmeasured factors associated with

coffee drinking habit may also have influenced results of individual studies and thus pooled

estimates in this meta-analysis. However, coffee consumption has been associated with unhealthy

dietary habits (e.g., Western dietary pattern) rather than healthy dietary habits. Previous studies

have shown that coffee intake is positively associated with intakes of red meat, fat, and

cholesterol, and negatively associated with vitamin C, folic acid, multivitamin supplement, and

vegetable intake50. The addition of sugar and cream to coffee was not directly assessed in most of

the studies included in this meta-analysis. Since the coffee additives could contribute to insulin

resistance or weight gain which can increase the risk of endometrial cancer, our pooled estimate

would be likely to be stronger rather than weaker if the coffee additives were adjusted. These

unmeasured factors related to coffee drinking also could explain the strong inverse association

with high coffee intake among Japanese women. For example, it is possible that women in

Europe or United States are likely to add more sugar or cream to coffee, while Japanese women

add relatively small amount of the additives in their coffee. Fifth, not all studies were included for

the dose-response analysis due to lack of data. When we assessed publication bias with 14 studies

excluding 2 studies24, 26, there was still no indication of publication bias.

In conclusion, the results of this meta-analysis of 16 observational studies provide

quantitative evidence that high coffee consumption or an increased coffee consumption may

lower the risk of endometrial cancer. Sparse data on decaffeinated coffee intake in relation to

endometrial cancer precludes conducting a meta-analysis, but strong evidence of coffee

consumption among studies in Japan where coffee is not highly correlated with total caffeine

John Wiley & Sons, Inc.

 International Journal of Cancer Page 16 of 29

16

intake may support the importance of coffee components for the beneficial roles of coffee in the

disease risk38. Although it is difficult to determine causality of this relation on the basis of

observational studies (i.e., coffee is self-selected, not randomized), no significant heterogeneity of

the studies, no publication bias, large prospective data included, and a significant dose-response

relationship of coffee consumption, may suggest a real protection of coffee against endometrial

cancer. More large studies are needed to determine subgroups to obtain more benefits from coffee

drinking in relation to endometrial cancer risk.



John Wiley & Sons, Inc.

Page 17 of 29 International Journal of Cancer

17

References

1. Devivo I, Persson I, Adami HO. Endometrial cancer. In: Adami HO, Hunter D,

Trichopoulos D, eds. Textbook of cancer epidemiology 2nd ed. New York: Oxford

University Press; 2008. p. 468-93.

2. Modesitt SC, van Nagell JR Jr. The impact of obesity on the incidence and treatment of

gynecologic cancers: a review. Obstet Gynecol Surv 2005;60:683–92.

3. Friberg E, Orsini N, Mantzoros CS, Wolk A. Diabetes mellitus and risk of endometrial

cancer: a meta-analysis. Diabetologia 2007;50:1365–74.

4. Kaaks R, Lukanova A, Kurzer MS. Obesity, endogenous hormones, and endometrial

cancer risk: a systematic review. Cancer Epidemiol Biomarkers Prev 2002;11:1531–43.

5. Porta M, Vioque J, Ayude D, Alguacil J, Jariod M, Ruiz L, Murillol JA. Coffee drinking:

the rationale for treating it as a potential effect modifier of carcinogenic exposures. Eur J

Epidemiol 2003;18:289–98.

6. Xiaofeng Yu, Zhijun Bao, Jian Zou, Jie Dong. Coffee consumption and risk of cancers: a

meta-analysis of cohort studies. BMC Cancer 2011;15:11-96.

7. Ferrini RL, Barrett-Connor E. Caffeine intake and endogenous sex steroid levels in

postmenopausal women. The Rancho Bernardo Study. Am J Epidemiol 1996;144:642–4.

8. Nagata C, Kabuto M, Shimizu H. Association of coffee, green tea, and caffeine intakes

with serum concentrations of estradiol and sex hormone-binding globulin in

premenopausal Japanese women. Nutr Cancer 1998;30:21–4

9. Kotsopoulos J, Eliassen AH, Missmer SA, Hankinson SE, Tworoger SS. Relationship

between caffeine intake and plasma sex hormone concentrations in premenopausal and

postmenopausal women. Cancer 2009;115:2765–74.

John Wiley & Sons, Inc.

 International Journal of Cancer Page 18 of 29

18

10. Wu T, Willett WC, Hankinson SE, Giovannucci E. Caffeinated coffee, decaffeinated

coffee, and caffeine in relation to plasma C-peptide levels, a marker of insulin secretion,

in U.S. women. Diabetes Care 2005;28:1390–6.

11. Williams CJ, Fargnoli JL, Hwang JJ, van Dam RM, Blackburn GL, Hu FB, Mantzoros

CS. Coffee consumption is associated with higher plasma adiponectin concentrations in

women with or without type 2 diabetes: a prospective cohort study. Diabetes Care 2008;

31:504–7.

12. van Dam RM, Hu FB. Coffee consumption and risk of type 2 diabetes: a systematic

review. JAMA 2005;294:97–104.

13. Bravi F, Scotti L, Bosetti C, Gallus S, Negri E, La Vecchia C, Tavani A. Coffee drinking

and endometrial cancer risk: a metaanalysis of observational studies. Am J Obstet Gynecol

2009;200:130-5.

14. Stensvold I, Jacobsen BK. Coffee and cancer: a prospective study of 43,000 Norwegian

men and women. Cancer Causes Control 1994;5: 401–8.

15. Shimazu T, Inoue M, Sasazuki S, et al. Coffee consumption and risk of endometrial

cancer: a prospective study in Japan. Int J Cancer 2008;123:2406-10.

16. McCann SE, Yeh M, Rodabaugh K, Moysich KB. Higher regular coffee and tea

consumption is associated with reduced endometrial cancer risk. Int J Cancer

2009;124:1650–3.

17. Bandera EV, Williams-King MG, Sima C, Bayuga-Miller S, Pulick K, Wilcox H, Zauber

AG, Olson SH. Coffee and tea consumption and endometrial cancer risk in a population-

based study in New Jersey. Cancer Causes Control 2010;21:1467–73.

John Wiley & Sons, Inc.

Page 19 of 29 International Journal of Cancer

19

18. Friberg E, Mantzoros CS, Wolk A. Diabetes and risk of endometrial cancer: A

population-based prospective cohort study. Cancer Epidemiol Biomarkers Prev 2007;16:

276-80.

19. Gunter MJ, Schaub JA, Xue X, Freedman ND, Gaudet MM, Rohan TE, Hollenbeck AR,

Sinha R. A prospective investigation of coffee drinking and endometrial cancer incidence.

Int J Cancer 2011 [Epub ahead of print].

20. Je Y, Hankinson SE, Tworoger SS, DeVivo I, and Giovannucci E. A prospective cohort

study of coffee consumption and risk of endometrial cancer over a 26-year of Follow-Up.

Cancer Epidemiol Biomarkers Pre 2011;20:2487–95.

21. Nilsson LM, Johansson I, Lenner P, Lindahl B, Van Guelpen B. Consumption of filtered

and boiled coffee and the risk of incident cancer: a prospective cohort study. Cancer

Causes Control 2010;21:1533–44.

22. Salazar-Martinez E, Lazcano-Ponce E, Sanchez-Zamorano LM, Gonzalez-Lira G,

Escudero DELRP, Hernandez-Avila M. Dietary factors and endometrial cancer risk.

Results of a case-control study in Mexico. Int J Gynecol Cancer 2005;15:938-45.

23. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, RennieD, Moher D, Becker

BJ, Sipe TA, Thacker SB. Meta-analysis of observational studies in epidemiology: a

proposal for reporting. metaanalysis Of Observational Studies in Epidemiology (MOOSE)

group. JAMA 2000;283:2008–12.

24. Levi F, Franceschi S, Negri E, La Vecchia C. Dietary factors and the risk of endometrial

cancer. Cancer 1993;71:3575–81.

25. Kalandidi A, Tzonou A, Lipworth L, Gamatsi I, Filippa D, Trichopoulos D. A case-

control study of endometrial cancer in relation to reproductive, somatometric, and life-

style variables. Oncology 1996;53:354-9.

John Wiley & Sons, Inc.

 International Journal of Cancer Page 20 of 29

20

26. Petridou E, Kedikoglou S, Koukoulomatis P, Dessypris N, Trichopoulos D. Diet in

relation to endometrial cancer risk: a case-control study in Greece. Nutr Cancer

2002;44:16-22.

27. Terry P, Vainio H, Wolk A, Weiderpass E. Dietary factors in relation to endometrial

cancer: a nationwide case-control study in Sweden. Nutr Cancer 2002;42:25–32.

28. Bravi F, Scotti L, Bosetti C, Zucchetto A, Talamini R, Montella M, Greggi S, Pelucchi C,

Negri E, Franceschi S, La Vecchia C. Food groups and endometrial cancer risk: a case-

control study from Italy. Am J Obstet Gynecol 2009;200:293.e1–7.

29. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177–

88.

30. Cochran WG. The combination of estimates from different experiments. Biometrics

1954;10:101–29.

31. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-

analyses. BMJ 2003;327:557–60.

32. Berlin JA, Longnecker MP, Greenland S. Metaanalysis of epidemiologic dose-response

data. Epidemiology. 1993;4:218-228.

33. Greenland S, Longnecker MP. Methods for trend estimation from summarized dose-

response data, with applications to meta-analysis. Am J Epidemiol 1992;135:1301–9.

34. Orsini N, Bellocco R, Greenland S. Generalized least squares for trend estimation of

summarized dose response data. Stata J. 2006;6:40-57.

35. Jain MG, Howe GR, Rohan TE. Nutritional factors and endometrial cancer in Ontario.

Canada. Cancer Control 2000; 7:288–96.

36. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for

publication bias. Biometrics 1994;50:1088–101.

John Wiley & Sons, Inc.

Page 21 of 29 International Journal of Cancer

21

37. Egger M, Davey Smith G, Schneider M, Minder C. Bias in metaanalysis detected by a

simple, graphical test. BMJ 1997;315:629–63.

38. Hirose K, Niwa Y, Wakai K, Matsuo K, Nakanishi T, Tajima K. Coffee consumption and

the risk of endometrial cancer: evidence from a case-control study of female hormone-

related cancers in Japan. Cancer Sci 2007;98:411–5.

39. Koizumi T, Nakaya N, Okamura C, Sato Y, Shimazu T, Nagase S, Niikura H, Kuriyama

S, Tase T, Ito K, Tsubono Y, Okamura K, et al. Case-control study of coffee consumption

and the risk of endometrial endometrioid adenocarcinoma. Eur J Cancer Prev

2008;17:358–63.

40. Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA Cancer J Clin

2005;55:74-108.

41. Arnlov J, Vessby B, Riserus U. Coffee consumption and insulin sensitivity. JAMA

2004;291:1199–201.

42. Fotsis T, Zhang Y, Pepper MS, Adlercreutz H, Montesano R, Nawroth PP, Schweigerer

L. The endogenous estrogen metabolite 2-methoxyoestradiol inhibits angiogenesis and

suppresses tumor growth. Nature 1994;368:237–9.

43. Nagata C, Matsushita Y, Shimizu H. Prevalence of hormone replacement therapy and

user’s characteristics: a community survey in Japan. Maturitas 1996;25:201–7.

44. International agency for research on cancer. IARC monographs on the evaluation of

carcinogenic risks to humans, vol. 72: Hormonal contraception and post-menopausal

hormonal therapy. Lyon, France: IARC; 1999.

45. McCullough ML, Bandera EV, Patel R, Patel AV, Gansler T, Kushi LH, Thun MJ, Calle

EE. A prospective study of fruits, vegetables, and risk of endometrial cancer. Am J

John Wiley & Sons, Inc.

 International Journal of Cancer Page 22 of 29

22

Epidemiol 2007;166:902–11.

46. Cust AE, Slimani N, Kaaks R, van Bakel M, Biessy C, Ferrari P, Laville M, Tjonneland

A, Olsen A, Overvad K, Lajous M, Clavel-Chapelon F, et al. Dietary carbohydrates,

glycemic index, glycemic load, and endometrial cancer risk within the European

Prospective Investigation into Cancer and Nutrition cohort. Am J Epidemiol

2007;166:912–23.

47. Kasum CM, Nicodemus K, Harnack LJ, Jacobs DR, Jr., Folsom AR. Whole grain intake

and incident endometrial cancer: the Iowa Women’s Health Study. Nutr Cancer

2001;39:180–6.

48. Gannmaa D, Cui X, Feskanich D, Hankinson SE, Willett WC. Milk, dairy intake and risk

of endometrial cancer: A-26 year of follow-up. Int J Cancer 2011[Epub ahead of print]

49. Gunter MJ, Hoover DR, Yu H, Wassertheil-Smoller S, Manson JE, Li J, Harris TG,

Rohan TE, Xue X, Ho GY, Einstein MH, Kaplan RC, et al. A prospective evaluation of

insulin and insulin-like growth factor-I as risk factors for endometrial cancer. Cancer

Epidemiol Biomarkers Prev 2008;17:921-9.

50. Je Y, Liu W, Giovannucci E. Coffee consumption and risk of colorectal cancer: A

systematic review and meta-analysis of prospective cohort studies. Int J Cancer

2009:124:1662-8.

John Wiley & Sons, Inc.

Page 23 of 29 International Journal of Cancer

23

Table 1. Characteristics of case-control/cohort studies included in the meta-analysis of coffee consumption and endometrial cancer risk

First author (year) Country Study Study Case/controls Coffee consumption Relative risk Adjustment for covariates
design period or subjects (95% CI)

Levi (1993)24 Italy/ Hospital- 1988-1991 274/572 Tertile1(reference) 1.00 Age, study center
Switzland based Tertile2 1.18 (0.82-1.71)1
case-control Tertile3 1.22 (0.86-1.73)1
Decaf: Tertiles2&3 vs. Tertile1 1.26 (0.87-1.83)1
Kalandidi (1996)25 Greece Hospital- 1992-1994 145/298 1 cup per day 1.04 (0.86-1.27)2 Age, height, BMI, education, occupation, age at
based menopause, age at menarche, live births,
case-control miscarriages, induced abortions, oral
contraceptive, menopausal estrogens, smoking,
alcohol intake, and energy intake
Jain (2000)35 Canada Population- 1994-1998 552/562 0g/day (reference) 1.00 Age, body weight, university education, ever
based >0-250g/day 0.80 (0.54-1.18) smoked, history of diabetes, HRT, OC use, live
case-control >250-500g/day 1.18 (0.78-1.79) births, age at menarche, and total energy intake
>500g/day 0.68 (0.45-1.04)
Petridou (2002)26 Greece Hospital- 1999 84/84 <4 cups per week (reference) 1.00 Education, BMI, age at menarche, pregnancies
based •4 cups per week 0.47 (0.20-1.10) and abortions, alcohol drinking, and total energy
case-control intake
Terry (2002)27 Sweden Population- 1994-1995 709/2,877 Quartile1 (0.5 cups/day, 1.00 Age, BMI, smoking, physical activity, prevalence
based reference) of diabetes, fatty fish consumption, and total food
case-control Quartile2 (1.6 cups/day) 0.9 (0.6-1.3) consumption
Quartile3 (3.1 cups/day) 0.8 (0.6-1.1)
Quartile4 (4.3 cups/day) 0.7 (0.5-1.0)
Hirose (2007)38 Japan Hospital- 1990-2000 229/12,425 No intake (reference) 1.00 Age, BMI, year, motivation for consultation,
based Occasional 0.70 (0.45-1.08) parity, age at first delivery, smoking, drinking,
case-control 1-2 cups/day 0.64 (0.43-0.94) type of breakfast, fondness of salty and fatty
•3 cups/day 0.41 (0.19-0.87) foods, fruit, vegetable, beef, fish, carrot, and
exercise
Koizumi (2008)39 Japan Population- 2002-2005 107/214 ”3-4 times/week (reference) 1.00 BMI, education, smoking, age at menarche,
based 5-6 times/week-1 cup/day 0.6 (0.3-1.2) number of pregnancies, menopausal status, OC
case-control •2-3 cups/day 0.4 (0.2-0.9) use, history of diabetes, and total energy intake
Bravi (2009)28 Italy Hospital- 1992-2006 454/908 <1 cup/day (reference) 1.00 Age, BMI, education, study center, year of
based 2 cups/day 1.12 (0.81-1.56) interview, total energy intake, history of diabetes,
case-control 3 cups/day 0.95 (0.67-1.35) age at menarche, parity, menopausal status, HRT,
4 cups/day 0.83 (0.54-1.28) OC use
> 4 cups/day 0.50 (0.29-0.86)
McCann (2009)16 USA Hospital- 1982-1998 513/512 No intake (reference) 1.00 Age, BMI, education, menopausal status, HRT,
based 0.5 cup/day 0.77 (0.50-1.18) OC use, smoking status, and tea consumption
case-control 1-2 cups/day 0.89 (0.63-1.24)
>2 cups/day 0.71 (0.49-1.03)

John Wiley & Sons, Inc.

 International Journal of Cancer Page 24 of 29

24

Decaf: No intake (reference) 1.00

0.5 cup/day 0.94 (0.64-1.37)
1-2 cups/day 0.83 (0.60-1.16)
>2 cups/day 1.17 (0.74-1.84)
Bandera (2010)17 USA Population- 2001-2005 417/395 No consumption (reference) 1.00 Age, BMI, education, race, age at menarche,
based ”1 cup/day 1.05 (0.58-1.89) menopausal status and age at menopause, parity,
case-control >1-2 cups/day 1.02 (0.56-1.88) OC use, HRT use, smoking, and
>2 cups/day 0.69 (0.36-1.33) milk/cream/nondairy creamer added
Stensvold (1994)14 Norway Cohort 1977-1990 84/21,238 ”2 cups/day (reference) 1.0 Age, smoking, county of residence
3-4 cups/day 1.30 (0.62-2.72)1
5-6 cups/day 1.30 (0.62-2.74)1
•7 cups/day 0.80 (0.34-1.90)1
Shimazu (2008)15 Japan Cohort 1990-2005 117/53,724 ”2 days/week (reference) 1.00 Age, BMI, study area, menopausal status, age at
3-4 days/week 0.97 (0.56-1.68) menopause, parity, use of exogenous female
1-2 cups/day 0.61 (0.39-0.97) hormones, smoking status, consumption of green
•3 cups/day 0.38 (0.16-0.91) vegetable, beef, port, and green tea
Friberg (2009)18 Sweden Cohort 1992-2007 677/60,634 ”1 cup/day (reference) 1.00 Age, BMI, smoking, education, age at menopause,
(Swedish 2-3 cups/day 0.77 (0.63-0.94) age at menarche, OC use, PMH use, parity,
Mammo- •4 cups/day 0.75 (0.57-0.98) history of diabetes, total energy intake, tea and
graphy intake of foods correlated with coffee such as
Cohort) buns, cookies, and cakes
Nilsson (2010)21 Sweden Cohort 1987-2007 108/32,178 <1 occasion/day (reference) 1.00 Age, sex, BMI, smoking, education, and
1-3 occasions/day 0.92 (0.48-1.76) recreational physical activity
•4 occasions/day 0.88 (0.44-1.78)
Gunter (2011)19 USA Cohort 1995-2006 1,486/111,429 0 cup/d (reference) 1.00 Age, smoking, BMI, age at menarche, age at first
(NIH- <1 cup/day 0.87 (0.73-1.05) child’s birth, parity, age at menopause, HRT use,
AARP Diet 1 cup/day 0.82 (0.68-0.98) OC use, diabetes and physical activity
& Health 2-3 cup/day 0.83 (0.71-0.97)
Study) >3 cup/day 0.64 (0.51-0.80)
Decaf:
0 cup/d (reference) 1.00
<1 cup/day 0.88 (0.70-1.12)
1 cup/day 0.81 (0.62-1.06)
2-3 cup/day 0.66 (0.51-0.85)
>3 cup/day 0.81 (0.54-1.20)
Je (2011)20 USA Cohort 1980-2006 672/67,470 <1 cup/day (reference) 1.00 Age, BMI, age at menarche, age at menopause,
(Nurses’ 1 cup/day 1.04 (0.83-1.31) parity & age at last birth, duration of OC use,
Health 2-3 cups/day 0.93 (0.76-1.14) PMH use, alcohol intake, pack years of smoking,
Study) •4 cups/day 0.75 (0.57-0.97) total energy intake
Decaf:
<1 cup/month 1.00
1 cup/month-<1 cup/day 0.92 (0.75-1.14)
1 cup/day 0.96 (0.74-1.25)
• 2 cups/day 0.78 (0.57-1.08)

John Wiley & Sons, Inc.

Page 25 of 29 International Journal of Cancer

25

1
Standard error was calculated from the data.
2
We calculated RR for a 4-cup increment/d, which is the approximate mean of the highest categories for studies conducted in Europe14, 18, 21, 26-28.
Abbreviations: BMI, body mass index (kg/m2); HRT, hormone replacement therapy; PMH, postmenopausal hormone; OC, oral contraceptive.

John Wiley & Sons, Inc.

 International Journal of Cancer Page 26 of 29

26

Table 2. Pooled relative risks and 95% CI for coffee consumption and endometrial cancer risk

Study No. of No. of Relative risk Heterogeneity

studies cases (95% CI) P I 2(%)

High versus low coffee intake

All studies14-21, 24-28, 35, 38, 39 16 6,628 0.71 (0.62-0.81) 0.13 29.1
Study design
Case-control studies16, 17, 24-28, 35, 38, 39 10 3,484 0.69 (0.55-0.87) 0.04 48.5
Cohort studies14, 15, 18-21 6 3,144 0.70 (0.61-0.80) 0.62 0.0
Geographic region of study1
Europe14, 18, 21, 24-28 8 2,535 0.79 (0.63-0.99) 0.11 40.4
United States/Canada16, 17, 19, 20, 35 5 3,640 0.69 (0.60-0.79) 0.93 0.0
Japan15, 38, 39 3 453 0.40 (0.25-0.63) 0.99 0.0

Increment of 1 cup/day2
All studies14-21, 25, 27, 28, 35, 38, 39 14 6,270 0.92 (0.90-0.95) 0.18 25.6
Study design
Case-control studies16, 17, 25, 27, 28, 35, 38, 39
8 3,126 0.90 (0.86-0.95) 0.39 4.6
Cohort studies14, 15, 18-21 6 3,144 0.94 (0.90-0.97) 0.15 38.8
Geographic region of study3
Europe14, 18, 21, 25, 27, 28 6 2,177 0.93 (0.90-0.97) 0.54 0.0
United States/Canada16, 17, 19, 20, 35 5 3,640 0.94 (0.91-0.97) 0.90 0.0
Japan15, 38, 39 3 453 0.76 (0.68-0.86) 0.54 0.0

1
P value difference in relative risks for studies in Japan versus non-Japan = 0.03.
2
Levi24 and Petridou26 were not included for the dose-response analysis due to lack of data.
3
P value difference in relative risks for studies in Japan versus non-Japan = 0.006.

John Wiley & Sons, Inc.

Page 27 of 29 International Journal of Cancer

)LJXUH)RUHVWSORWRIFDVHFRQWURODQGFRKRUWVWXGLHVRIWKHULVNRIHQGRPHWULDOFDQFHUIRUWKH
KLJKHVWYHUVXVORZHVWFDWHJRULHVRIFRIIHHLQWDNH
[PP['3,

John Wiley & Sons, Inc.

 International Journal of Cancer Page 28 of 29

)LJXUH)RUHVWSORWRIREVHUYDWLRQDOVWXGLHVRIWKHULVNRIHQGRPHWULDOFDQFHUIRUWKHKLJKHVWYHUVXV
ORZHVWFDWHJRULHVRIFRIIHHLQWDNHVWUDWLILHGE\JHRJUDSKLFUHJLRQ
[PP['3,

John Wiley & Sons, Inc.

Page 29 of 29 International Journal of Cancer

)LJXUH3XEOLFDWLRQELDVLQWKHVWXGLHV%HJJ¶VIXQQHOSORWLQGLFDWHGQRSXEOLFDWLRQELDVLQWKH
VWXGLHVLQFOXGHGLQWKHPHWDDQDO\VLV1RLQGLFDWLRQRISXEOLFDWLRQELDVZDVQRWHGIURPERWK
YLVXDOL]DWLRQRIIXQQHOSORWDQG%HJJ¶VS RU(JJHU¶VWHVWS 
[PP['3,

John Wiley & Sons, Inc.