OBJECTIVE To compare the efficacy of angiotensin converting enzyme inhibition with calcium antagonism in diabetic patients with microalbuminuria. METHODS Randomised study of diabetic patients with microalbuminuria treated with perindopril or nifedipine for 12 months and monitored for one or three months after stopping treatment depending on whether they were hypertensive or normotensive. Patients were randomised separately according to whether they were hypertensive or normotensive. METHODS Diabetic clinics in three university teaching hospitals. METHODS 50 diabetic patients with persistent microalbuminuria. In all, 43 completed the study: 30 were normotensive and 13 hypertensive; 19 had type I diabetes and 24 had type II diabetes. METHODS For 12 months 20 patients were given perindopril 2-8 mg daily and 23 were given nifedipine 20-80 mg daily. METHODS Albumin excretion rate, blood pressure, and glomerular filtration rate. RESULTS Both perindopril and nifedipine significantly reduced mean blood pressure. During treatment there was no significant difference between those treated with perindopril and those treated with nifedipine with respect to albuminuria or mean blood pressure. Stopping treatment with both drugs was associated with a sustained increase in albuminuria and mean blood pressure. There was a significant correlation between mean blood pressure and albuminuria and also between the reduction in mean blood pressure and the decrease in albuminuria during treatment with both drugs. In hypertensive patients both drugs caused significant decreases in mean blood pressure and albuminuria. In normotensive patients there was no significant reduction in albuminuria with either regimen. CONCLUSIONS In diabetic patients with microalbuminuria blood pressure seems to be an important determinant of urinary albumin excretion. Perindopril and nifedipine have similar effects on urinary albumin excretion, both preventing increases in albuminuria in normotensive patients and decreasing albuminuria in hypertensive patients.