This study examines the magnitude of alcohol-related premature death in the French population, which still has the highest average alcohol intake in the world and a relatively low coronary heart disease mortality rate. Two data sources were used: the national mortality data in 1990 and a prospective mortality experience in a cohort of 2,687 middle-aged working men examined in 1980-1985 and followed-up during an average of 9.3 years. In the general population study, alcohol-related premature mortality (35-64 years) was calculated using alcohol-attributable fractions (AAFs) derived from studies of alcohol involvement in deaths from various causes. In the cohort, it was estimated from AAFs and attributable risk using both alcohol exposure prevalence and relative risks of death according to alcohol intake categories. In 1990, estimates of 19.1% and 13.0% of all premature deaths in French men and women were attributed to alcohol. Digestive diseases followed by malignant neoplasms in men and by unintentional injuries in women were major contributors to the total number of alcohol-related premature deaths. In the cohort study, 90 deaths occurred during the follow-up period. The estimation of alcohol-related premature mortality using AAFs was 24.4%. The relative risk of total mortality (adjusted for age, smoking habits, and body mass index) for men who consumed > 60 ml/day of alcohol, compared with those who consumed 0-25 ml/day was 1.9 (95% confidence interval: 1.1-3.5). The consumption of 26-60 ml/day was not associated with reduced risk of mortality, and the adjusted relative risk for this group was 1.3 (95% confidence interval: 0.7-2.5). The estimated attributable risk of premature mortality caused by consumption of > 25 ml/day of alcohol was 29.9%. This study reports the persisting high alcohol-related premature mortality in the French general population, as well as in middle-aged working men. The results suggest that efforts should be paid to reduce further the consumption of alcohol in France.