Diabetes is a well-established cause of cardiovascular disease (CVD) and all-cause mortality. The burden of death attributable to diabetes in the United States is not well quantified, particularly with regard to age. The authors analyzed data from the Second National Health and Nutrition Examination Survey (NHANES II) (1976-1980) and the NHANES II Mortality Study, in which a nationally representative cohort of 9,250 adults aged 30-75 years was followed for 12-16 years, to determine all-cause and cause-specific mortality. Overall, between 1976 and 1980, the prevalence of diagnosed diabetes was 4.3%. By 1992, the relative hazard of all-cause mortality was 1.9 (95% confidence interval: 1.5, 2.3), and the population attributable risk (PAR) was 3.6%. The relative hazard of CVD mortality was 2.3 (95% confidence interval: 1.8, 2.8), and the PAR was 5.2%. Including participants with undiagnosed diabetes in the estimates increased the PAR for all-cause mortality to 5.1% and that for CVD mortality to 6.8%. Women had a higher prevalence of diagnosed diabetes than men and a greater relative hazard of death than nondiabetic women, leading to a higher PAR for women (3.8% for all causes and 7.3% for CVD) versus men (3.3% for all causes and 3.8% for CVD). These data suggest that diabetes accounts for at least 3.6% of all deaths and 5.2% of CVD deaths in US adults. Improvements in diabetes prevention and treatment should produce noticeable effects on US life expectancy.