Alcohol consumption and liver cancer risk: a meta-analysis.

PURPOSE: Alcohol is a confirmed risk factor of liver cancer. Yet, its dose-response function and synergistic effects with other risk factors remain unclear. METHODS: We performed a meta-analysis on publications up to May 2014. A total of 112 publications were identified. The meta-relative risk (mRR) and the dose-response trend were calculated. Tests for heterogeneity, publication bias, and sensitivity analyses were performed. The synergy index (SI) was recorded or calculated, whenever possible. RESULTS: Compared to individuals who never drank or drank at very low frequencies, the mRR for ever drinkers was 1.29 (95 % confidence interval, CI 1.16-1.42) and 1.46 (95 % CI 1.27-1.65) for case-control studies, and 1.07 (95 % CI 0.87-1.27) for cohort studies. Being a current drinker was associated with an increased liver cancer risk in case-control studies (mRR = 1.55, 95 % CI 0.38-2.73), but not in cohort studies (mRR = 0.86, 95 % CI 0.74-0.97). The dose-response relation between alcohol and liver cancer was apparent with RR = 1.08 (95 % CI 1.04-1.11) for 12 g/day (~1 drink), 1.54 (95 % CI 1.36-1.74) for 50 g/day, 2.14 (95 % CI 1.74-2.62) for 75 g/day, 3.21 (95 % CI 2.34-4.40) for 100 g/day, and 5.20 (95 % CI 3.25-8.29) for 125 g/day of alcohol consumption. There were synergistic effects of alcohol consumption with hepatitis (S = 2.14, 95 % CI 1.31-2.98) and with diabetes (S = 3.57, 95 % CI 2.29-4.84) on the risk of liver cancer, although this may be subject to publication bias. CONCLUSION: Overall, one alcoholic drink per day (~12 g/day) may be associated with a 1.1 times higher liver cancer risk. Further studies on the synergistic effects of alcohol consumption and other major risk factors are warranted.