In the present study we report on the predictive power of microalbuminuria for total and cardiovascular mortality in a prospective study (mean follow up 6.3 years) of treated hypertensive men, aged 50 to 72 years, with (n = 94) and without (n = 345) maturity onset diabetes mellitus. Thirty-three (35.1%) of the hypertensive patients with diabetes mellitus died during the follow-up period compared with 57 patients (16.5%) in the hypertensive group without diabetes mellitus (p <0.0002). In those with diabetes mellitus and hypertension, a log-rank test revealed a lower cardiovascular mortality in the normoalbuminuric group compared with both the microalbuminuric (p = 0.035) and the macroalbuminuric group (p = 0.002). The logarithm of urinary albumin excretion was a predictor of both total (p = 0.009) and cardiovascular (p = 0.001) mortality during the follow-up period using Cox regression analysis. This relation remained significant even after adjustment for other risk factors. HbA1c was also an independent predictor of total and cardiovascular mortality. In patients without diabetes mellitus, the small group of patients with macroalbuminuria had a markedly increased cardiovascular mortality compared with both the microalbuminuric (p <0.0001) and the normoalbuminuric groups (p <0.0001). No difference was observed between the normoalbuminuric and the microalbuminuric groups. Smoking at entry and concomitant cardiovascular disease at entry were independent predictors of cardiovascular mortality in these patients. We conclude that microalbuminuria was an independent predictor for cardiovascular mortality in treated hypertensive men with maturity onset diabetes mellitus. Macroalbuminuria, but not microalbuminuria, predicted cardiovascular mortality in nondiabetic treated hypertensive men.