The relation between proteinuria and mortality was investigated in 1188 patients with Type 1 diabetes and 3234 patients with Type 2 diabetes, aged 35-55 at baseline and followed up for a mean of 9.4 +/- 3.1 years in the WHO Multinational Study of Vascular Disease in Diabetes. Baseline prevalence of light or heavy proteinuria was the same (25%) in both types of diabetes after adjustment for differences in diabetes duration. Compared with patients with no proteinuria, all cause mortality ratios were 1.5 (95% confidence interval 1.1-2.0) and 2.9 (2.2-3.8) for Type 1 patients with light and heavy proteinuria, respectively, and 1.5 (1.2-1.8) and 2.8 (2.3-3.4) for Type 2 patients, after adjustment for age, duration of diabetes, blood pressure, cholesterol, and smoking. Proteinuria was associated with significantly increased mortality from renal failure, cardiovascular disease, and all other causes of death. In both types of diabetes, the association was strongest for renal deaths, and of similar magnitude for cardiovascular and all other causes of death. In conclusion, proteinuria is a common, important, and rather non-specific risk factor for increased morbidity and mortality in diabetes. The relation of proteinuria to mortality is similar for both types of diabetes. The benefits and risks of proteinuria reduction should be examined in large randomized trials with clinical endpoints.