Alcohol and smoking habits in association with hepatocellular carcinoma risk.

Abstract

We assessed hepatocellular carcinoma (HCC) risk associated with smoking and alcohol consumption and their interactions, using both questionnaire data and objective serum biomarkers. Information on smoking and alcohol consumption was collected at baseline from 450,112 participants of the EPIC cohort, among whom 255 developed HCC after a median follow-up of 14 years. In a nested case-control subset of 108 HCC cases and 108 matched controls, known biomarkers of smoking (cotinine, nicotine) and habitual alcohol consumption (2-hydroxy-3-methylbutyric acid) were annotated from untargeted metabolomics features. Multivariable-adjusted hazard ratios (HRs) or odds ratios (ORs) with 95% confidence intervals (CIs) were computed, and multiplicative and additive interaction parameters were calculated. Compared to never smokers, current smokers had a higher HCC risk (HR = 2.46, 95% CI = 1.77-3.43) dose-dependently with the number of cigarettes smoked per day (P_trend <.001). Compared to light drinkers, HCC risk was higher in former (HR = 3.20, 95% CI = 1.70-6.03), periodically heavy (HR = 1.98, 95% CI = 1.11-3.54), and always heavy (HR = 5.51, 95% CI = 2.39-12.7) drinkers. Higher HCC risk was also observed in the highest versus the lowest tertiles of cotinine (OR = 4.88, 95% CI = 1.52-15.70), nicotine (OR = 5.80, 95% CI = 1.33-25.30) and 2-hydroxy-3-methylbutyric acid (OR = 5.89, 95% CI = 1.33-26.12). Questionnaire-assessed smoking and alcohol exposures did not demonstrate an HCC risk interaction at the multiplicative (MI = 0.88, 95% CI = 0.40-1.96) or additive (RERI = 0.71, 95% CI = -10.1 to 23.6; attributable proportion = 0.17, 95% CI = -0.52 to 1.16; synergy index = 1.27, 95% CI = 0.98-1.66) scales. Similar analyses with cotinine, nicotine, and 2-hydroxy-3-methylbutyric acid also did not show interactions between smoking and alcohol consumption on HCC risk. Smoking and alcohol consumption are strong independent risk factors for HCC and do not appear to synergistically impact its risk, but larger studies are needed.