The relationship of cardiovascular disease (CVD) to the control of blood pressure (BP) and serum cholesterol levels was studied in 686 treated, middle-aged hypertensive men whose condition was followed up for 12 years. Both mean in-study BP (P less than .001) and serum cholesterol levels (P less than .001) were better predictors for CVD than the respective entry levels (both not significant) in multivariate analysis. Surprisingly, for mean in-study systolic and diastolic BP there seemed to be a level (approximately 150 and 85 mm Hg, respectively) below which further reduction of BP had no additional benefit from treatment. Analyses of CVD morbidity in relationship to changes in BP and serum cholesterol levels clearly showed that a combined reduction of both risk factors was necessary to achieve a substantial reduction in morbidity. These study findings indicate that (1) it may be unfavorable to reduce BP below a certain level in middle-aged hypertensive men; and (2) other risk factor interventions must be improved to improve the patient's prognosis.