Prospective study of alcohol consumption in the United States: Quantity, frequency, and cause-specific mortality

Background: Alcohol average volume (quantity multiplied by frequency) has been associated with mortality in drinkers. However, average volume may mask associations due to quantity or frequency alone. Methods: We prospectively assessed relationships between alcohol quantity and frequency, and mortality from all-causes, cardiovascular disease, cancer, and other-causes in a cohort created by linking the 1988 National Health Interview Survey (response rate 87%) to the National Death Index through 2002. Participants were 20,765 current drinkers age >= 18 years. At 14-year follow-up 2,547 had died. Results: For quantity, among men who consumed >= 5 drinks (compared to 1 drink) on drinking days, adjusted relative risks (RR) of mortality were: for cardiovascular disease, 1.30 [95% confidence interval (CI) 0.96-1.75; p for linear trend (p-trend) = 0.0295], for cancer, 1.53 (95% CI 1.11-2.09; p-trend = 0.0026), and for other-causes, 1.42 (95% CI 1.08-1.87; p-trend = 0.0029); among women for other-causes, 2.88 (95% CI 1.61-5.12; p-trend = 0.0010). For frequency, among men in the highest frequency quartile (compared to the lowest), RR were: for cardiovascular disease, 0.79 (95% CI 0.63-0.99; p-trend = 0.0330), for cancer, 1.23 (95% CI 0.95-1.59; p-trend = 0.0461), and for other-causes, 1.30 (95% CI 1.01-1.67; p-trend = 0.0070); among women, for cancer, 1.65 (95% CI 1.12-2.45, p-trend = 0.0031). Average volume obscured effects of quantity alone and frequency alone, particularly for cardiovascular disease in men where quantity and frequency trended in opposite directions. Conclusions: Alcohol quantity and frequency were independently associated with cause-specific mortality. Accumulating evidence of their differential effects may, in the future, be useful for clinical and public health recommendations.