The effects of the angiotensin-converting enzyme lisinopril were compared with those of the calcium antagonist nifedipine in 162 non-insulin-dependent diabetic hypertensive patients for a 24-week period. In 83 and 79 patients, respectively, lisinopril and slow-release nifedipine produced similar reductions in blood pressure (systolic/diastolic: -16/-13 mmHg supine and -14/-11 mmHg standing after lisinopril; -15/-12 mmHg supine and -14/-11 mmHg standing nifedipine). Fasting and post-prandial plasma glucose, glycosylated haemoglobin and plasma lipids appeared to be unaffected by either agent. Also, 28% of the patients on lisinopril and 30% of those on nifedipine presented microalbuminuria. Both drugs induced a reduction in the albumin excretion rate (AER). The geometric mean x:tolerance factor of the reduction in AER among the 23 microalbuminuric patients on lisinopril (-10.0 x:1.3 micrograms/min) was greater, though not significantly so, than that observed in the 26 on nifedipine (-0.9 x 1.2 micrograms/min). Moreover, lisinopril appeared to be better tolerated than nifedipine in our study population. Microalbuminuria is an important risk factor for cardiovascular mortality in non-insulin-dependent diabetic patients as well as in the general population. To what extent a reduction in the AER could ameliorate diabetic patients is, at present, unknown. Finally, both lisinopril and nifedipine showed a similar antihypertensive effect in these patients which was not associated with significant differences in plasma glucose, insulin or lipid concentrations. The clinical consequences of the insignificant differences in AER remain unclear.