The prevalence and natural history of severe proteinuria in mild to moderate hypertension are not completely defined. We screened 1635 men with a history of hypertension and randomized 1292 with untreated diastolic blood pressure (DBP) 95-109 mmHg to single-drug treatment with either hydrochlorothiazide, atenolol, captopril, clonidine, diltiazem-SR, prazosin, or placebo in a double-blind prospective trial. Twenty-seven of 1635 patients (1.7%) satisfying clinical criteria for primary hypertension were found to have developed proteinuria > 1000 mg/24 hours and were removed from the study. Follow-up data were obtained on 19 of these 27 patients. One patient was found to have focal segmental sclerosis and progressed to end-stage renal disease. Three other patients developed severe (serum creatinine > 3.5 mg/dl) chronic renal failure (one with diabetic nephropathy), one progressed from serum creatinine 1.4 to 2.2 mg/dl, but 14 of the 19 remained with stable serum creatinine < 2.0 mg/dl on follow-up for 6-9 years. Data were available for 1076 of 1155 (93%) treated study patients at end titration, 522/600 (87%) at one year and 322/444 (73%) at two years. There were significant associations for proteinuria with obesity and higher systolic blood pressure. There was a trend toward significant difference in mean 24-hour protein excretion rates at baseline between black (127 mg) and white (139 mg) patients (p = 0.07). There were no statistically significant changes in urinary protein excretion/24 hours between or within the different treatment groups (including placebo). Eighteen patients were removed from the study during the active treatment phase for proteinuria > 1000 mg/24 hours: hydrochlorothiazide 4, placebo 3, diltiazem 3, prazosin 3, atenolol 2, clonidine 2, and captopril 1. We conclude: (1) the prevalence of severe (> 1 g/24 hours) proteinuria in the hypertensive population is significant but does not necessarily imply a poor prognosis; (2) mean 24-hour urinary protein excretion rates did not vary in response to the different classes of antihypertensive drugs; and (3) there was no drug-specific increase in proteinuria detected in this study.