Document Lpp53zd24rzZDxkjRJ3X5GzDz

American Journal of Epldemidogy Copyright 02001 by The Johns Hopkns UruverSny schod d Hygtene and PuMk Health All nghts reserved Alcohol Consumption and Bladder Cancer Risk: Results from the Netherlands Cohort Study Vd. 153. NO. 1 Pnnted in U.S.A. Maurice t? A. Zeegers,' Alex Volovics,2Elisabeth Dorant,' R.Alexandra Goldbohm? and Piet A. van den Brandt' Although several epidemiologic studies have been conducted on alcohol consumption and bladder cancer risk, the risk according to quantity and type of alcohol consumed is not clear. The authors investigatedthese associations in a large pr&pective cohort study on diet and cancer among 120,852 subjects in the Netherlands aged 55-69 years at baseline (1986). Subjects completed a questionnaire on risk factors for cancer, including alcohol consumption. Follow-up for incident cancer was established by record linkage to cancer registries. The case-cohort analysis was restricted to a follow-up periodof 6.3 years and was based on 594 caseswith bladder cancer and 3,170 subcohort members. The authors corrected for age and smoking in multivariable analyses. The incidence rate ratios for men who consumed 4,5-15, 15-30, and 230 grams of alcohol per day were 1.49, 1.52, 1.16, and 1.63 compared with nondrinkers, respectively (pfor trend = 0.13). Alcohol consumedfrom beer, wine, and liquor was associated with moderately elevated risks, although most were not statistically significant.The incidence rate ratios for women varied around unity.The results of this study do not suggest an important association between alcohol consumption and bladder cancer risk. Am J pidemrbl2001;153:38-41. alcohol drinking; alcoholic beverages: bladder neoplasms; urologic neoplasms Although several epidemiologic studies have been conducted to investigate the association between alcohol consumption and bladder cancer risk, the risk according to quantity and type of alcohol consumed is stlll not clear. In a recent meta-analysis of these studies, a slightly elevated risk of bladder cancer was demonstrated for men currently drinking alcohol compared with abstainers (1). In this metaanalysis, however, no data were available to explore the influence of quantity and types of alcoholic beverages consumed. These results prompted us to investigate these associations in more detail in the ongoing Netherlands Cohort Study. MATERIALS AND METHODS cohort The study design has been described in detail previously (2). The study population originated from 204 municipal population registries throughout the Netherlands, and the cohort includes 58,279 men and 62.373 women who were aged 55-69 years at baseline (1986).We used the casecohort approach for data processing andanalysis (3).A subcohort of 3,500 subjects was randomly sampled from the cohort after baseline exposure measurement and was followed up to obtain vital status infomation. No subcohort members were lost to follow-up during the follow-up period. FOIIOW-UP Follow-up for incident cancer was established by record linkage to cancer registries and the Dutch database of pathology reports (4). and follow-up was more than 95 percent complete (5). The present analysis was restricted to 6.3 years of follow-up. After we excluded prevalent cases. 3.346 subcohort members and 619 incident cases with microscop i d y confirmed carcinomas of the bladder, ureters, renal pelvis, or urethra were identified. Because the overwhelming majority of tumors occurred in the bladder, and because the renal pelvis and ureter are covered by the same urothebum. the term bladder cancer was used as a synonym for these neoplasms. Received for publication December 6, 1999. and accepted tor publicationApril 10. 2000. 'Abbreviation: CI. confidence interval. Department of Epidemiology, Maastricht University, Maastricht, the Nelherlands. Oepartment of Methodology and Statistics, Maastricht Univentty, Maastricht, the Netherlands. 'Department of NutritionalEpidemiology,TNO Nutrition and Food Research Institute, Zeist. the Netherlands. Correspondence to Maurice P. A. Zeegers, Department of Epidemiology. Maastricht University, PO Box 616. 6200 M D Maastricht, the Netherlands (e-mail: mpazeegersOepid.unimaas.nl). Questionnaire All subjects completed a self-administered questionnaire on risk factors for cancer. The food-frequency section concentrated on habitual consumption during the year before the study began. Consumption of alcoholic beverages was addressed by questions on beer, red wine, white wine, sherry, other fortified wines, liqueur, and liquor. The questionnaire data were keyed twice and were processed in a standardized manner blinded with respect to case-subcohort 38 t % i f s!sAieue nea .. 40 Zeegers et al. TABLE 2. Adjusted Incidence rats ratio1s- ( Md a U n w Maty-8) for bladder cancer In -men, according to alcohol consumption trwn rpecm~-, Alcohd -m Nod No.d N&-,S WdbaOe Cohort Study, 1986-1992 Adjusted forage and smoking' incohort vlsubcohort R R t 95% Clt RR 95% CI No alcohol intake 62 1,446 1.00 Refer- 1.00 Reference Alcohol from beer <5 #day 174 3.383 1.32 0.95, 1.85 1.35 0.94.1.95 g/day 89 1,607 1.49 1.02,2.18 1.44 0.95,2.18 1 5 - a #W 22 331 1.94 1.07,3.52 1.70 0.90,3.23 130 g/day 10 170 1.45 0.66.3.20 1.09 0.46.2.57 p value for linear trend 4.01 0.12 Alcohol increment 10 @day 1.07 0.96,1.19 1.03 0.91,1.16 Adjusted from wine* <5W Y 5-d5 9w 15-a @day 230 !#day p value for linear trend Alcohol increment 10@day 151 67 25 - -11 2.688 1.417 581 135 1.38 1.19 1.13 2.08 0.24 1.00 0.98,1.94 0.8o.i.n 0.66, 1.93 0.94,4.61 0.68, 1.15 1.54 1.23 1.14 1.73 0.46 0.99 1.06.2.23 0.80.1.90 0.65,2.00 0.74.4.05 0.07,1.13 Alcohol from r i r e5 $day 114 2,207 1.30 0.91,l.S 1.44 0.98.2.11 3-<15@daY 89 1.603 1.37 0.94,2.00 1.38 0.92.208 15-<3ogldaY 230 dday 70 1.155 1.41 0.95.2.10 1.25 0.81,1.91 50 512 2.48 1.57,3.92 1.94 1.17,3.22 p value for linear trend 4.01 0.03 Alcohol increment 10 @day 1.14 1.05,1.23 1.06 0.97,1.16 Smokingseatus.amount. and duration. t FIR. indderoerate ratio; CI.confidence interd. $ Among mak, wine drinkers in the subcohort. 63%drank red wine, 58% drank white wine, 48% drank sherry, and 16% drank liqueur. wine had an increased incidencerateratio of 1.73 (95 percent CI: 0.74,4.05) compared with nondrinkers. The highest risk was found for men who drank 230 g/day from liquor comp a d with nondrinkers. The comsponding incidence rate ratio was 1.94 (95 percent CI: 1.17. 3.22). Lower intakes from liquor did not seem to be related to an elevated risk. Adjustment for age only or simultaneousinclusionof alcohol consumption frombeer. wine, and liquor in one regression model did not change the results substantially. Although some point estimates suggested increased risks for alcohol consumers, practically none of the incidence rate ratios or dose-responsetrends was statistically significant (table 2). DISCUSSION The results of this prospective study do not suggest an important association between alcohol consumption and bladder cancer risk. If any, the association between alcohol consumption and male bladder cancer is probably small. Some authors have suggested that residual confounding due to tobacco smoking could explain an increased risk as a result of alcohol drinking (IO, 11). We attempted to model cigarette smoking habits such that they best explained bladder cancer by using smoking status, amount. and duration. However, correction for smoking did not change the incidence rate ratios substantially. Therefore, the association observed between alcohol consumption and bladder cancer risk did not seem to be entirely due to residual confounding by smoking, although some influence cannot be excluded. A recent meta-analysis based on predominantly case- control studies concluded that alcohol consumption slightly increases male bladder cancer risk (summary odds ratio = 1.35, 95 percent CI: 0.91, 2.02). an estimate that might not be of practiai importance (1). We npeattdthis meta-analysis to evaluate whether the summary odds ratio for male alcohol consumers compared with nondnnkers remained stable after the present study was included (1). We found that the new age- and srnoking-adjustedsummary odds ratio was 1.35 (95 percent CI: 0.96, 1.91). Associations between specific alcoholic beverages and bladder cancer risk wen reported in nine studies without consistent results (12-20). The literature to date does not support a causal role for alcohol consumption in bladder cancer etiology. Several mechanisms have been postulated to explain ethanol-related carcinogenesis. Ethanol slows down protein synthesis. One consequence is that cell repair mechanisms might be inhibited, which could lead to malignant changes (21. 22). Furthermore, ethanol might improve permeability of mern- branes to carcinogens and might enhance carcinogenic activ- -f- ity (22-24). Other explanations include the effect of ethanol . . 2;on cell proliferation. possibly caused by acetaldehyde (22, fj 23. 25). Animal experiments have shown that nitrosamines 4'Am J Epidemiol Vol. 153,No. 1, 2001 ''L81Z: '92 32 'PZ 'EZ -'1Z2Z .oz '61 '8I 'LI '9I 'SI 'PI '1 'II '11 '0I '6 8