IJC International Journal of Cancer

Transcription

1 IJC International Journal of Cancer Coffee intake and the risk of colorectal adenoma: The colorectal adenoma study in Tokyo Sanjeev Budhathoki, Motoki Iwasaki, Taiki Yamaji, Shizuka Sasazuki and Shoichiro Tsugane and Prevention Group, Research Center for Cancer Prevention and Screening, National Cancer Center, Tokyo, Japan Coffee is a commonly consumed beverage which contains several potential anticarcinogenic and chemopreventive compounds, and has been hypothesized to have protective effects in colorectal neoplasia. However, the limited available data on coffee consumption in relation to colorectal adenoma (CRA), a precursor lesion to most colorectal cancers, remain largely inconsistent. In this study, we evaluated the association of coffee intake with the risk of CRA in a middle-aged Japanese population. Study subjects were selected from examinees who underwent total colonoscopy as part of a cancer screening program and responded to self-administered dietary and lifestyle questionnaires. A total of 738 patients with adenoma and 697 controls were included in the study. Coffee intake was assessed with a food frequency questionnaire, and divided into quartiles based on the distribution among controls. Unconditional logistic regression models were used to estimate odds ratio (OR) and 95% confidence interval (CI) of CRA, with adjustment for potential confounding factors. High coffee consumption was associated with a reduced risk of CRA, with a multivariate-adjusted OR for the highest versus lowest quartile of coffee intake of 0.67 (95% CI ; p trend ). The inverse association of coffee intake was limited to proximal (OR ; 95%CI ; p trend ) and distal colon adenoma (OR ; 95%CI ; p trend ), and appeared to be more evident with small (OR ; 95%CI ; p trend ) and single adenomas (OR ; 95%CI ; p trend ). Green tea intake was not found to be associated with CRA risk. This study provides support for the protective effect of coffee drinking on colon adenomas, a precursor of colon cancer. Colorectal cancer (CRC) is one of the most common cancers, and accounts for 10% of all cancer cases worldwide. 1 Considerable geographic variation in the incidence of CRC has been observed, with almost 60% of cases occurring in developed regions, suggesting that environmental factors, including dietary factors, are implicated in the etiology of CRC. 1,2 Among dietary factors, however, only intake of red and processed meat and alcohol (in men) are classified as convincing risk factors, while foods containing dietary fiber are classified as protective factors. 2 Key words: coffee, colorectal adenoma, epidemiology Grant sponsor: Ministry of Health, Labour and Welfare of Japan (Grant-in-Aid for the 3rd Term Comprehensive 10-Year-Strategy for Cancer Control and Grant-in-Aid for Cancer Research 17-9); Grant sponsor: Ministry of Education, Culture, Sports, Science, and Technology of Japan (Grants-in-Aid for Scientific Research on Innovative Areas: 221S0001); Grant sponsor: Japan Society for the Promotion of Science, and Foundation for Promotion of Cancer Research in Japan (Grant-in-Aid for Scientific Research C and Grant-in-Aid for Scientific Research C ) DOI: /ijc History: Received 12 Aug 2014; Accepted 4 Dec 2014; Online 10 Dec 2014 Correspondence to: Motoki Iwasaki, and Prevention Group, Research Center for Cancer Prevention and Screening, National Cancer Center, Tsukiji, Chuo-ku, Tokyo , Japan, Tel.: , Fax: , moiwasak@ncc.go.jp Coffee is among the most commonly consumed beverages worldwide, and its potential role in the etiology of CRC has drawn considerable interest. 3,4 A meta-analysis of 13 casecontrol studies found that regular coffee drinkers had an 17% lower risk of CRC than non/occasional drinkers. 5 Although a pooled 6 and a meta-analysis 7 of prospective cohort studies including >5000 cancer cases found no measurable association between coffee and colon or colorectal cancer, a recent large prospective study 8 with more cases than the pooled or meta-analysis reported a significant inverse association of coffee intake with CRC risk. Coffee contains several bioactive compounds, including chlorogenic acid, caffeic acid, cafestol and kahweal, which have shown anticarcinogenic effects in in vitro and in vivo studies, and which may play a protective role in colorectal carcinogenesis Coffee drinking may also protect against colon neoplasia by reducing the synthesis and secretion of bile acids, which are the potential promoters of colon carcinogenesis, and by increasing colonic motility, thereby reducing exposure of the epithelium to potential carcinogens. 12 Further, studies have also reported an inverse association between coffee consumption and biomarkers of inflammation and insulin resistance, which have been positively associated with CRC As an established precursor lesion of CRC, colorectal adenomas (CRAs) are an informative endpoint in the process of colon carcinogenesis, and studying adenoma will allow the evaluation of risk factors early in the colorectal neoplastic process among asymptomatic individuals. 17,18 However, only a few studies have evaluated the effect of coffee intake in Int. J. Cancer: 137, (2015) VC 2014 UICC

2 464 Coffee intake and colorectal adenoma What s new? The presence of cancer-fighting compounds in coffee is of considerable interest to the understanding of colorectal cancer etiology. But data on the relationship between coffee intake and colorectal adenoma is limited, and studies have arrived at varying conclusions. In the present analysis of the coffee-drinking habits of 1,435 individuals, coffee intake was found to be inversely associated with the risk of proximal and distal colon adenoma, precursors of malignant disease. The association was evident particularly in individuals with single and small adenomas. The findings support the idea that coffee consumption helps defend against colorectal carcinogenesis. relation to CRA risk; of these, two small studies suggested beneficial effects of coffee in CRA, 19,20 while other studies did not support this effect. Thus, evidence for the effect of coffee on CRA remains inconsistent. Here, we evaluated the association between coffee intake and CRA risk in a large group of asymptomatic men and women who underwent total colonoscopy as part of as cancer screening program. We also investigated whether the association differed by adenoma site, size or number. Material and Methods Study population The study participants were selected from examinees who underwent magnifying colonoscopy with dye spreading as part of a cancer screening program provided by the Research Center for Cancer Prevention and Screening, National Cancer Center, Tokyo, Japan. Details of the study have been described previously. 27,28 Men aged years and women aged years undergoing total colonoscopy from the anus to the cecum and without a history of CRA, any malignant neoplasm, ulcerative colitis, Crohn s disease, familial adenomatous polyposis, carcinoid tumor or colectomy were defined as eligible for inclusion in the study. From February 2004 to February 2005, a total of 3,212 examinees underwent total colonoscopy, among whom 2,234 satisfied the above conditions and were eligible for inclusion in the study. Based on the pit pattern classification of colorectal lesions, 782 individuals (526 men and 256 women) were found to have at least one or more adenomas and were included as adenoma cases. Of the remaining 1,452 individuals, we identified 482 men and 721 women as potential controls who were also free from other benign lesions (e.g., hyperplastic polyps, inflammatory polyps, and diverticula). We included all potential male controls in the study, as there were fewer male controls than cases. Female controls were randomly selected from potential controls and frequency-matched to the female cases in 5 age categories (40 49, 50 54, 55 59, and 65 years of age) and two screening periods (first and second halves). The screening period was matched because standard operating procedures were improved during the first half period after the establishment of the Research Center, which might have influenced, for example, the accuracy of diagnosis. 29 All examinees gave written informed consent to participate in the study. The study protocol was approved by the institutional review board of the National Cancer Center, Tokyo, Japan. Lifestyle and dietary factor assessment Data on lifestyle and behavioral factors, as well as personal and family medical history and supplemental drug use were collected by a self-administered questionnaire survey prior to the screening procedure. Information on dietary intake factors was assessed by a 138-item food frequency questionnaire with pre-specified standard portion sizes and nine intake frequency categories for most food items. Information on consumption of coffee and canned coffee was obtained by two closed-ended questions with nine precoded frequency categories (<1 cup/ week, 1 2 cups/week, 3 4 cups/week, 5 6 cups/week, 1 cup/ day, 2 3 cups/day, 4 6 cups/day, 7 9 cups/day and >10 cups/ day). The amount of coffee or canned coffee consumed was calculated by multiplying the frequency of intake by the standard portion size of each beverage (120 ml per cup for regular coffee and 250 ml per can for can coffee), and the sum of these two were taken as total coffee intake. Similarly, the combination of sencha (portion size 200 ml per cup) and bancha/ genmaicha (portion size 120 ml per cup) was taken as green tea. The validity of food and nutrient intake estimated from the questionnaire was evaluated by comparison with estimates from 4-day dietary records provided by a subsample of the screening examinees. 30 The Spearman rank correlation coefficient between estimates from the questionnaire and diet record was 0.80 (men) and 0.82 (women) for coffee, and 0.67 (men) and 0.42 (women) for green tea consumption. Statistical analysis For the present analysis, we excluded 74 individuals with extreme energy intake (top and bottom 2.5% of total energy intake distribution) and 11 individuals with missing values for covariates, leaving a total of 738 cases and 697 controls for final analysis. All participants were divided into quartile categories based on coffee intake among controls. Logistic regression models were used to compute the odds ratio (OR) and 95% confidence interval (CI) with the first quartile as reference group. Adjustment was made for age (40 49, 50 54, 55 59, 60 64, 65 years), sex, screening period (first or second half), body mass index (<21.0, , , 25.0 kg m 22 ), physical activity (MET-hours/day, quartile), cigarette smoking (never, 20, and 40 pack-years), alcohol consumption (never, past, <150, , 300 g/ week), parental CRC history, nonsteroidal anti-inflammatory drug use, history of diabetes, and intakes of total energy, folate, fiber, isoflavone and red and processed meat (all quartiles). Ordinal scores of 0 3 were assigned to quartile

3 Budhathoki et al. 465 categories of coffee consumption to assess the trend of an association. Two-sided p values of < 0.05 were considered statistically significant. All statistical analyses were performed using SAS version 9.3 (SAS Institute, Cary, NC). Results Selected characteristics of adenoma cases and controls have been previously reported. 27,28 In brief, the number of males was 498 (67.5%) in the case group and 453 (65.0%) in the control group. Mean age was 60.8 and 59.9 years in cases and controls, respectively (p ). Compared with the controls, cases were more likely to be heavy smokers and alcohol drinkers, and to have a higher BMI and higher frequency of family CRC history and lower NSAID use. Cases had lower intake of fiber, folate and isoflavone than controls but higher total energy intake. Distribution of selected characteristics of controls according to quartile category of coffee intake is summarized in Table 1. Individuals with higher coffee intake were more likely to be cigarette smokers and alcohol users, and tended to be younger than individuals with low coffee intake. In comparison to the first quartile, individuals in the fourth quartile of coffee intake also had higher mean BMI and lower physical activity. While intake of total energy increased, intake of folic acid, isoflavone and vegetables decreased with increasing coffee intake. 48.5% of the control participants consumed 1 or more cups of coffee daily, and the overall average amount of coffee intake was 157 ml day 21. In the age-, sex-and screening period-adjusted model (Table 2), higher coffee intake was not associated with the risk of CRA (p trend ). However, when we further adjusted for smoking, alcohol drinking and other lifestyle and dietary factors, the OR of CRA decreased with increasing coffee intake, and was statistically significant in the highest intake group (OR for highest intake group compared to reference group , 95% CI , p trend ). Green tea intake, on the other hand, was not related to CRA risk in either model. Table 3 shows the relation between coffee intake and CRA by adenoma site, size, and number. The site-specific analysis was conducted in 384 proximal, 261 distal and 81 rectal adenoma cases, following exclusion of 12 cases with missing information. The inverse associations between coffee intake and adenoma risk were observed only in proximal (Q4 vs. Q1 OR , 95% CI , p trend ) and distal colon adenoma (Q4 vs. Q1OR5 0.62, 95% CI , p trend ). The association between coffee intake and adenoma risk appeared to be independent of adenoma size (Q4 vs. Q1OR5 0.68, 95% CI , p trend for small adenoma, and OR , 95% CI , p trend for large adenoma) and adenoma number (Q4 vs. Q1OR5 0.65, 95% CI , p trend for single adenoma, and OR , 95% CI , p trend for multiple adenoma), although the p value for trend was not significant for large and multiple adenoma. We further evaluated the association of coffee consumption and CRA risk stratified by sex, smoking and drinking habits, body mass index, physical activity and intake of red and processed meats (Table 4). Overall, the risk estimates were below 1 across most of the strata of these factors. Although the p value for interaction was not significant, the inverse association of higher coffee consumption with CRA tended to be stronger among never smokers (Q4 vs. Q1 OR , 95% CI , p trend ), never drinkers (Q4 vs. Q1 OR5 0.45, 95% CI , p trend ), physically active persons (Q4 vs. Q1 OR5 0.58, 95% CI , p trend ) and those with lower red and processed meat consumption (Q4 vs. Q1 OR5 0.50, 95% CI , p trend ). The inverse association of coffee with CRA also tended to be stronger among individuals with a BMI < 25 kg m 22. Discussion In this colonoscopy-based case-control study of a middleaged Japanese population, we found a decreased risk of CRA in the higher compared to lower coffee consumption group. On site-specific analysis, the inverse association of coffee consumption was limited to colon adenoma, and although the interaction effect was not significant, the association was stronger in never smokers, never drinkers, physically active persons, and in those with lower meat consumption. Few studies to date have investigated the relation of coffee intake with the risk of CRA occurrence or recurrence. 31 A case-control study in Denmark reported a marked decrease in the risk of large bowel adenoma with higher coffee intake (8 cups/day vs. <3 cups/day OR 5 0.3, 95%CI ), 19 while in their colonoscopy-based case-control study in Japan, Kato et al. observed a significant decrease in the relative risk of proximal (RR , 95% CI ) and distal (RR , 95% CI ) adenoma, but not rectal adenoma, associated with daily coffee intake compared to non-drinkers. 20 However, coffee intake was unrelated to CRA in other studies in Japan 21,22 and the USA, 25 while another US study reported a suggestive but non-significant increase in risk of CRA with coffee consumption of six cups/day or more. 24 Our present findings are partly in accordance with those of the findings of the studies in Denmark 19 and Japan. 20 Nevertheless, evidence from these earlier studies should be interpreted with caution as most had a small number of adenoma cases, 19,21,23,24 controls which were not all verified by colonoscopy, 20,21 and even failure to account for important confounders such as smoking and alcohol drinking. 19,20,26 Coffee drinking is often positively correlated with unhealthy behaviors like smoking, alcohol drinking, physical inactivity and high red and processed meat consumption, and tended to be positively associated with these factors in the present study (Table 1). Indeed, the inverse association of coffee with CRA we identified was visible only after adjusting for these factors, implicating them as confounders. We therefore stratified our analysis by smoking, alcohol drinking,

4 466 Coffee intake and colorectal adenoma Table 1. Characteristics of controls according to quartile of coffee intake Quartile (Q) of coffee intake (ml day 21 ) Q1 (<25.7) Q2 ( ) Q3 ( ) Q4 (>290.7) Men, n (%) 86 (62.3) 130 (65.0) 93 (61.6) 144 (69.2) Age (years), mean (SD) 61.1 (5.4) 60.8 (6.1) 59.8 (5.8) 58.3 (6.0) BMI (kg m 22 ), mean (SD) 22.9 (3.0) 23.0 (2.7) 22.7 (2.7) 23.4 (2.9) Physical activity 1, mean (SD) 37.3 (9.8) 36.8 (8.1) 35.9 (6.2) 36.0 (6.6) Smoking status, n (%) Never 89 (64.5) 110 (55.0) 81 (53.6) 82 (39.4) Past 44 (31.9) 79 (39.5) 56 (37.1) 86 (41.3) Current 5 (3.6) 11 (5.5) 14 (9.3) 40 (19.2) Alcohol use, n (%) Never 41 (29.7) 52 (26.0) 37 (24.5) 44 (21.2) Past 6 (4.3) 5 (2.5) 11 (7.3) 8 (3.8) Current 91 (65.9) 143 (71.5) 103 (68.2) 156 (75.0) Diabetes mellitus history, n (%) 6 (4.3) 11 (5.5) 7 (4.6) 18 (8.7) Parental colorectal cancer, n (%) 8 (5.8) 17 (8.5) 10 (6.6) 18 (8.7) NSAID use, n (%) 7 (5.1) 19 (9.5) 17 (11.3) 10 (4.8) Dietary intake 2, mean (SD) Total energy (kcal day 21 ) 1823 (465) 1899 (491) 1936 (476) 2013 (514) Total fiber (g day 21 ) 13.9 (5.5) 13.4 (5.0) 13.9 (4.7) 13.1 (4.7) Folate (mg day 21 ) (162.3) (158.5) (140.3) (142.8) Isoflavone (mg day 21 ) 52.7 (59.7) 44.8 (28.9) 44.8 (25.6) 42.6 (25.9) Vegetable (g day 21 ) (132.6) (122.7) (107.1) (124.0) Fruit (g day 21 ) (161.4) (162.4) (160.7) (165.2) Red and processed meat (g day 21 ) 34.5 (23.8) 37.3 (24.9) 39.9 (27.0) 35.2 (20.9) 1 MET-hours day Energy-adjusted values except for total energy. Table 2. Logistic regression models of the effect of coffee intake on colorectal adenoma Quartile (Q) category Q1 Q2 Q3 Q4 p trend Coffee Case/controls 163/ / / /208 OR (95% CI) (reference) 0.91 ( ) 0.93 ( ) 0.86 ( ) 0.38 OR (95% CI) (reference) 0.85 ( ) 0.83 ( ) 0.67 ( ) 0.02 Green tea Case/controls 166/ / / /206 OR (95% CI) (reference) 1.01 ( ) 0.89 ( ) 0.97 ( ) 0.69 OR (95% CI) (reference) 1.05 ( ) 0.99 ( ) 1.50 ( ) 0.20 Cutoff ranges for the first to fourth quartile were <25.7, , and >290.7 ml day 21 for coffee, and <194.3, , and >960.0 ml day 21 for green tea, respectively. 1 Adjusted for age, sex and screening period. 2 Adjusted for age, sex, screening period, cigarette smoking, alcohol drinking, body mass index, physical activity, parental colorectal cancer history, history of diabetes mellitus, non-steroidal anti-inflammatory drug use and total energy, folate, fiber, isoflavone and red and processed meat intake. physical activity, and red and processed meat consumption, and found that the inverse association between coffee and CRA was evident only in the never smoker, never drinker, physically active group and low meat consumers. Although we adjusted for smoking, drinking, physical activity, BMI and red and processed meat consumption, residual confounding,

5 Budhathoki et al. 467 Table 3. Logistic regression models of the effect of coffee intake on colorectal adenoma according to site, size and number Quartile (Q) of coffee intake (ml day 21 ) Q1 (<25.7) Q2 ( ) Q3 ( ) Q4 (>290.7) p trend Controls Site of adenoma Proximal colon Case OR (95% CI) (reference) 0.75 ( ) 0.75 ( ) 0.64 ( ) 0.04 Distal colon Case OR (95% CI) (reference) 0.77 ( ) 0.76 ( ) 0.62 ( ) 0.06 Rectum Case OR (95% CI) (reference) 1.96 ( ) 1.59 ( ) 1.09 ( ) 0.61 Size of adenoma <10 mm Case OR (95% CI) (reference) 0.82 ( ) 0.86 ( ) 0.68 ( ) mm Case OR (95% CI) (reference) 0.98 ( ) 0.75 ( ) 0.55 ( ) 0.06 Number of adenomas 1 adenoma Case OR (95% CI) (reference) 0.83 ( ) 0.67 ( ) 0.65 ( ) 0.02 >1 adenoma Case OR (95% CI) (reference) 0.86 ( ) 1.15 ( ) 0.70 ( ) Adjusted for age, sex, screening period, cigarette smoking, alcohol drinking, body mass index, physical activity, parental colorectal cancer history, history of diabetes mellitus, non-steroidal anti-inflammatory drug use and total energy, folate, fiber, isoflavone and red and processed meat intake. if present, could have attenuated the association between coffee intake and CRA risk in the smoker, alcohol user, and physically inactive groups, and higher meat consumers. On site-specific analysis, the inverse association of higher coffee intake with CRA risk in the present study was confined to the proximal and distal colon. Kato et al. also observed a similar association of coffee intake with proximal and distal colon adenoma, but not with rectal adenoma. 20 In general, such differences in the findings of colorectal neoplasia according to colorectal subsite have also been reported for coffee 8 or other risk factors The subsite analysis is of interest because there are distinct differences within subsites of the colorectum with regard to embryonic origin, histology, physiological functions and genetic mutations, and it is likely that different etiological factors are involved in carcinogenesis at different sites of the colorectum. 35,36 Nonetheless, considering the small number of rectal adenoma cases in the present (81 cases) and an earlier study, 20 further large studies are required to verify these findings. We were also interested to note that the inverse association of coffee in our subjects was more evident with small (<10 mm) and single adenomas than with large (10 mm) and multiple adenomas. In a study of sigmoid adenoma in men, Kono et al. similarly observed a tendency towards a decrease in odds ratio with increasing coffee intake with small but not large adenomas. 23 The reason for the lack of inverse association in larger and multiple adenomas is unknown and might merely be a chance finding, given findings that coffee drinking was uniformly associated with both single and multiple adenoma in another study. 20 In the present study, 12.5% had large (>10 mm) adenomas and 42.4% had multiple adenomas, and 9.3% had both these conditions. The associations by number and size of adenoma in the present study might therefore be a chance finding due to the small sample sizes in subgroup analysis. Compared to the situation with adenoma, the association of CRC with coffee intake has been extensively studied, although our understanding of whether the putative

6 468 Coffee intake and colorectal adenoma Table 4. Logistic regression models of the effect of coffee intake on colorectal adenoma stratified by the selected factors Quartile (Q) of coffee intake (ml day 21 ) Q1 (<25.7) Q2 ( ) Q3 ( ) Q4 (>290.7) p trend Sex Men Case/controls 113/86 132/ /93 149/144 OR (95% CI) (reference) 0.73 ( ) 0.81 ( ) 0.68 ( ) 0.12 Women Case/controls 50/52 80/70 56/58 54/64 OR (95% CI) (reference) 1.11 ( ) 0.75 ( ) 0.62 ( ) Cigarette smoking Never smoker Case/controls 91/89 105/110 74/81 61/82 OR (95% CI) (reference) 0.97 ( ) 0.89 ( ) 0.62 ( ) 0.06 Ever smoker Case/controls 72/49 107/90 86/70 142/126 OR (95% CI) (reference) 0.80 ( ) 0.85 ( ) 0.75 ( ) 0.31 Alcohol drinking Never drinker Case/controls 48/41 43/52 37/37 33/44 OR (95% CI) (reference) 0.76 ( ) 0.92 ( ) 0.45 ( ) 0.07 Ever drinker Case/controls 115/97 169/ / /164 OR (95% CI) (reference) 0.91 ( ) 0.82 ( ) 0.73 ( ) 0.08 Body mass index (kg m 22 ) <25 Case/controls 111/ / / /154 OR (95% CI) (reference) 0.95 ( ) 0.82 ( ) 0.76 ( ) Case/controls 52/25 57/44 46/27 62/54 OR (95% CI) (reference) 0.54 ( ) 0.80 ( ) 0.50 ( ) 0.14 Physical activity (MET-hours day 21 ) < Median (<34.6) Case/controls 85/68 104/100 88/77 112/103 OR (95% CI) (reference) 0.79 ( ) 0.84 ( ) 0.74 ( ) 0.30 Median (34.6) Case/controls 78/70 108/100 72/74 91/105 OR (95% CI) (reference) 0.87 ( ) 0.79 ( ) 0.58 ( ) 0.02 Red and processed meat intake (g day 21 ) < Median (<32.0) Case/controls 93/74 99/97 74/72 82/105 OR (95% CI) (reference) 0.76 ( ) 0.77 ( ) 0.50 ( ) 0.01 Median (32.0) Case/controls 70/64 113/103 86/79 121/103 OR (95% CI) (reference) 0.97 ( ) 0.87 ( ) 0.86 ( ) Adjusted for age, sex, screening period, cigarette smoking, alcohol drinking, body mass index, physical activity, parental colorectal cancer history, history of diabetes mellitus, non-steroidal anti-inflammatory drug use and total energy, folate, fiber, isoflavone and red and processed meat intake.

7 Budhathoki et al. 469 chemopreventive effect of coffee is more relevant during tumor initiation, progression or both is not clear. Our present finding of an inverse association between coffee intake and small adenoma suggests that coffee intake may be protective from the early stages of adenoma formation. After coffee, another large source of polyphenol intake in Japan is green tea, 37 a rich source of antioxidant polyphenols called catechins. Although experimental studies have consistently demonstrated that catechins have a protective effect in the development of colon cancer, 38 findings of four observational studies of green tea intake and CRA, all conducted in Japan, have been less consistent: one study found that daily hot green tea intake was associated with a reduced risk of distal and rectal adenoma 20 ; two found no association between green tea intake and CRA risk in men, women, or both 21,22 ; and the fourth reported no clear overall association between green tea intake and male sigmoid adenoma, but a non-significant tendency of a decreased risk of small adenomas. 23 Our present study had a larger number of cases than these four studies, and found no evidence for a protective association between green tea drinking and CRA risk. Among the strengths of the study, the provision of total colonoscopy to all study subjects likely reduced the chance of misclassification between cases and controls. Dietary habits and other lifestyle information were ascertained prior to colonoscopy procedures that is, before the determination of cases and control status, which likely reduced concerns regarding recall bias. Also, the number of subjects was considerably larger than in previous studies of coffee and CRA risk. One major limitation of the study was its crosssectional design, and the observed associations might be due to reverse causality. However, it is unlikely that participants changed their coffee drinking habits because of the presence of adenoma, as these lesions are generally asymptomatic. Adenoma cases were not histologically confirmed in our study, and we might therefore have included subjects with an early cancer or non-neoplastic lesion. However, our preliminary survey reported that the accuracy of diagnosis by magnifying chromoendoscopy was high (90%), and misclassification with this technique is likely to be minimal. 39 Coffee intake and other dietary factors were self-reported and therefore could have suffered some non-differential misclassification. Against this, however, our earlier study showed fairly good validity for coffee intake. 30 One of the putative protective compounds in colorectal neoplasia is caffeine, 12 almost 95% of which is derived from the intake of coffee and green tea in Japan. 40 Although we lacked information on coffee type (caffeinated or decaffeinated), which may be important in understanding the potential mechanisms of caffeine, it is likely that most of the participants consumed caffeinated coffee, because decaffeinated coffee is not common in Japan. 41 Finally, although we were able to adjust for known or potential confounders, the possibility of residual confounding cannot be ruled out. In summary, our study suggests that coffee drinking may have a protective role in the risk of CRA. However, given the limited and inconsistent evidence of coffee and CRA from earlier studies and the inherent limitations of our study design, verification of these findings awaits further large studies. Acknowledgements The authors are grateful to all the participants of the Colorectal Adenoma Study in Tokyo, and to all the doctors, nurses, and administrative staff at the Research Center for Cancer Prevention and Screening who assisted in the implementation of the study. References 1. Ferlay J, Shin HR, Bray F, et al. Estimates of worldwide burden of cancer in 2008: GLOBO- CAN Int J Cancer 2010;127: World Cancer Research Fund/American Institute for Cancer Research. Food, nutrition, physical activity, and the prevention of cancer: a global perspective. Washington, DC: American Institute for Cancer Research, Dorea JG, da Costa TH. Is coffee a functional food? Br J Nutr 2005;93: Thomson CA, Martınez ME. Coffee, tea, what beverage for me? Associations between beverage intake and colorectal neoplasia risk. J Natl Cancer Inst 2010;102: Galeone C, Turati F, La Vecchia C, et al. Coffee consumption and risk of colorectal cancer: a meta-analysis of case-control studies. Cancer Causes Control 2010;21: Zhang X, Albanes D, Beeson WL, et al. Risk of colon cancer and coffee, tea, and sugar-sweetened soft drink intake: pooled analysis of prospective cohort studies. J Natl Cancer Inst 2010;102: 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: Sinha R, Cross AJ, Daniel CR, et al. Caffeinated and decaffeinated coffee and tea intakes and risk of colorectal cancer in a large prospective study. Am J Clin Nutr 2012;96: Nkondjock A. Coffee consumption and the risk of cancer: an overview. Cancer Lett 2009;277: Kang NJ, Lee KW, Kim BH, et al. Coffee phenolic phytochemicals suppress colon cancer metastasis by targeting MEK and TOPK. Carcinogenesis 2011;32: Grubben MJ, Van Den Braak CC, Broekhuizen R, et al. The effect of unfiltered coffee on potential biomarkers for colonic cancer risk in healthy volunteers: a randomized trial. Aliment Pharmacol Ther 2000;14: BøhnSK,BlomhoffR,PaurI.Coffeeandcancer risk, epidemiological evidence, and molecular mechanisms. Mol Nutr Food Res 2014;58: Rebello SA, Chen CH, Naidoo N, et al. Coffee and tea consumption in relation to inflammation and basal glucose metabolism in a multi-ethnic asian population: a cross-sectional study. Nutr J 2011;10: Lopez-Garcia E, van Dam RM, Qi L, et al. Coffee consumption and markers of inflammation and endothelial dysfunction in healthy and diabetic women. Am J Clin Nutr 2006;84: Huxley R, Lee CM, Barzi F, et al. 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: Wu T, Willett WC, Hankinson SE, et al. Caffeinated coffee, decaffeinated coffee, and caffeine in relation to plasma C-peptide levels, a marker of insulin secretion, in US women. Diabet Care 2005;28: Leslie A, Carey FA, Pratt NR, et al. The colorectal adenoma-carcinoma sequence. Br J Surg 2002;89: Courtney ED, Melville DM, Leicester RJ. Review article: chemoprevention of colorectal cancer. Aliment Pharmacol Ther 2004;19: Olsen J, Kronborg O. Coffee, tobacco and alcohol as risk factors for cancer and adenoma of the large intestine. Int J Epidemiol 1993;22:

8 470 Coffee intake and colorectal adenoma 20. Kato I, Tominaga S, Matsuura A, et al. A comparative case-control study of colorectal cancer and adenoma. Jpn J Cancer Res Gann 1990;81: Hoshiyama Y, Kono S, Sasaba T, et al. Relation of cigarette smoking, alcohol use, and dietary habits to colon adenomas: a case-control study in Saitama, Japan. Asian Pac J Cancer Prev 2000;1: Nagata C, Shimizu H, Kametani M, et al. Diet and colorectal adenoma in Japanese males and females. Dis Colon Rectum 2001;44: Kono S, Imanishi K, Shinchi K, et al. Relationship of diet to small and large adenomas of the sigmoid colon. Jpn J Cancer Res Gann 1993;84: Lee WC, Neugut AI, Garbowski GC, et al. Cigarettes, alcohol, coffee, and caffeine as risk factors for colorectal adenomatous polyps. Ann Epidemiol 1993;3: Kahn HS, Tatham LM, Thun MJ, et al. Risk factors for self-reported colon polyps. J Gen Intern Med 1998;13: Almendingen K, Hofstad B, Vatn MH. Dietary habits and growth and recurrence of colorectal adenomas: results from a three-year endoscopic follow-up study. Nutr Cancer 2004;49: Otani T, Iwasaki M, Ikeda S, et al. Serum triglycerides and colorectal adenoma in a case-control study among cancer screening examinees (Japan). Cancer Causes Control 2006;17: Yamaji T, Iwasaki M, Sasazuki S, et al. Methionine synthase A2756G polymorphism interacts with alcohol and folate intake to influence the risk of colorectal adenoma. Cancer Epidemiol Biomarkers Prev 2009;18: Yamaji T, Iwasaki M, Sasazuki S, et al. Association between plasma 25-hydroxyvitamin D and colorectal adenoma according to dietary calcium intake and vitamin D receptor polymorphism. Am J Epidemiol 2012;175: Takachi R, Ishihara J, Iwasaki M, et al. Validity of a self-administered food frequency questionnaire for middle-aged urban cancer screenees: comparison with 4-day weighed dietary records. J Epidemiol 2011;21: Baron JA, Greenberg ER, Haile R, et al. Coffee and tea and the risk of recurrent colorectal adenomas. Cancer Epidemiol Biomarkers Prev 1997;6: Ben Q, Sun Y, Chai R, et al. Dietary fiber intake reduces risk for colorectal adenoma: a meta-analysis. Gastroenterology 2014;146: e Hjartåker A, Aagnes B, Robsahm TE, et al. Subsite-specific dietary risk factors for colorectal cancer: a review of cohort studies. J Oncol 2013;2013: Isomura K, Kono S, Moore MA, et al. Physical activity and colorectal cancer: the Fukuoka colorectal cancer study. Cancer Sci 2006;97: Potter JD. Nutrition and colorectal cancer. Cancer Causes Control 1996;7: Iacopetta B. Are there two sides to colorectal cancer? Int J Cancer 2002;101: Fukushima Y, Ohie T, Yonekawa Y, et al. Coffee and green tea as a large source of antioxidant polyphenols in the Japanese population. J Agric Food Chem 2009;57: Higdon JV, Frei B. Tea catechins and polyphenols: health effects, metabolism, and antioxidant functions. Crit Rev Food Sci Nutr 2003;43: Fu KI, Sano Y, Kato S, et al. Chromoendoscopy using indigo carmine dye spraying with magnifying observation is the most reliable method for differential diagnosis between non-neoplastic and neoplastic colorectal lesions: a prospective study. Endoscopy 2004;36: Yamada M, Sasaki S, Murakami K, et al. Estimation of caffeine intake in Japanese adults using 16 d weighed diet records based on a food composition database newly developed for Japanese populations. Public Health Nutr 2010;13: All Japan Coffee Association. A basic survey for monitoring trends in the demand for coffee (Japanese). Tokyo: All Japan Coffee Association, 2003.