Background: We examined the associations of alcohol consumption and liver holidays with all-cause mortality and with mortality due to cancer, heart disease, cerebrovascular disease, respiratory disease, and injury using a large-scale prospective study in Japan.
Methods: We followed 102,849 Japanese who were aged between 40 and 69 years at baseline for 18.2 years on average, during which 15,203 deaths were reported. Associations between alcohol intake and mortality risk were assessed using a Cox proportional hazards model, with analysis by the number of liver holidays (in which a person abstains from drinking for several days a week).
Results: A J-shaped association was observed between alcohol intake and total mortality in men (nondrinkers: reference; occasional drinkers: hazard ratio [HR] 0.74; 95% confidence interval [CI], 0.68-0.80; 1-149 g/week: HR 0.76; 95% CI, 0.71-0.81; 150-299 g/week: HR 0.75; 95% CI, 0.70-0.80; 300-449 g/week: HR 0.84; 95% CI, 0.78-0.91; 450-599 g/week: HR 0.92; 95% CI, 0.83-1.01; and ≥600 g/week: HR 1.19; 95% CI, 1.07-1.32) and in women (nondrinkers: reference; occasional: HR 0.75; 95% CI, 0.70-0.82; 1-149 g/week: HR 0.80; 95% CI, 0.73-0.88; 150-299 g/week: HR 0.91; 95% CI, 0.74-1.13; 300-449 g/week: HR 1.04; 95% CI, 0.73-1.48; and ≥450 g/week: HR 1.59; 95% CI, 1.07-2.38). In current drinkers, alcohol consumption was associated with a linear, positive increase in mortality risk from all causes, cancer, and cerebrovascular disease in both men and women, but not heart disease in men. Taking of liver holidays was associated with a lower risk of cancer and cerebrovascular disease mortality in men.
Conclusions: Alcohol intake showed J-shaped associations with the risk of total mortality and three leading causes of death. However, heavy drinking increases the risk of mortality, which highlights the necessity of drinking in moderation coupled with liver holidays.
Keywords: Japan/epidemiology; adult; alcohol; cardiovascular diseases/mortality; follow-up studies; mortality; neoplasms/mortality; proportional hazards models.