The impact of alcohol consumption combined with genetic polymorphisms in alcohol-metabolizing enzymes on risk of pancreatic cancer in Japan

B100

Many epidemiologic studies based upon lifestyle have failed to demonstrate any impact of alcohol on pancreatic cancer (PC) risk. To assess the influence of alcohol consumption in conjunction with single nucleotide polymorphisms (SNPs) in alcohol-metabolizing enzymes, we conducted a case-control study at Aichi Cancer Center Hospital with 138 cases diagnosed as PC. To avoid chance, we used two independent non-cancer control groups [control1: n=690 and 2: n=683], and then pooled control analyses were conducted. SNPs in aldehyde dehydrogenase (ALDH)2 Glu504Lys, and alcohol dehydrogenase (ADH)1B His48Arg, and ADH1C Arg272Gln were examined with an unconditional logistic regression model adjusted for confounders to estimate odds ratios (ORs) and their 95% confidence intervals (CIs). Confounders considered in a multivariate analysis were age, sex, pack-years of smoking, current body-mass index (BMI), BMI at 20 years, history of diabetes, and family history of PC. As we found consistent findings across both control groups, we present all our results for pooled controls. Distributions of pack-years of smoking, current BMI, BMI at 20 years, family history of PC were not significantly different between cases and controls, while history of diabetes was more prevalent in cases (P<0.001). A significantly increased PC risk was observed with former drinkers [adjusted OR (95% CI), 4.21 (2.46-7.22)], but not with current drinkers [1.14 (0.73-1.80)] relative to never drinkers. As former drinkers included those who quit drinking due to PC or other alcohol-related diseases, we categorized ever drinkers into two groups according to amount of alcohol consumption; moderate drinkers drinking or having drunk less than 50g ethanol/day and heavy drinkers with an intake of 50g ethanol/day or more. Adjusted ORs (95% CI) for ethanol consumption were 1.57 (1.02-2.41) in moderate drinkers and 1.69 (0.80-3.56) in heavy drinkers, relative to never drinkers (trend P=0.051). The ALDH2, ADH1B, or ADH1C polymorphisms in themselves were not found to be significant risk factors. However, analysis of the combined impact of alcohol consumption with polymorphisms showed the adjusted ORs (95% CIs) for ADH1B His/His to be 2.51 (1.41-4.46) with moderate drinkers and 3.43 (1.27-9.28) with heavy drinkers, compared with never drinkers (trend P=0.002). The similar trend was observed in ADH1C Arg/Arg. In contrast, no significant trend was evident with the ADH1B Arg+ and the ADH1C Gln+ allele. A gene-environment interaction between the ADH1B polymorphism and alcohol consumption was suggested although without statistical significance. ORs for heavy drinkers with ALDH2 Glu/Glu and Lys+ relative to never drinkers with ALDH2 Glu/Glu were 4.05 (1.10-14.96) and 6.95 (1.50-32.15), so that the impact of alcohol drinking was enhanced with ALDH2 Lys+. Furthermore, limited to those with ADH1B His/His, this finding for the ALDH2 Lys+ allele was consistent among ever drinkers. Interestingly, the OR for the ALDH2 Lys+ allele was significantly increased relative to ALDH2 Glu/Glu even among never drinkers. In conclusion, alcohol intake has an impact on PC risk when ethanol consumption and alcohol metabolism are combined, indicating a role for acetaldehyde in the mechanisms underlying PC development.