To evaluate the metabolic effects of two anti-hypertensive agents with different actions, nifedipine 20 mg twice daily and furosemide 60 mg twice daily, 23 patients with untreated essential hypertension performed a double-blind, cross-over study in treatment periods of 5 months. Metabolic effects were evaluated by serum lipoprotein determinations, the intravenous glucose tolerance test and the hyperinsulinaemic euglycaemic clamp technique. Nifedipine and furosemide reduced blood pressure to the same extent (-14 to -15 mm Hg for supine SBP and -9 to -10 mm Hg for supine DBP, both P < 0.0001). Whereas both drugs significantly increased the levels of glycolysated haemoglobin (HbA1c, +0.24%, P < 0.005 for nifedipine and +0.43%, P < 0.001 for furosemide), only furosemide increased fasting blood glucose (+0.3 mmol/L, P < 0.01) and fasting insulin (+2.2 mU/L, P < 0.05) but impaired the early insulin response to i.v. glucose (-15 mU/L, P < 0.05). Insulin sensitivity on the other hand was significantly impaired by nifedipine treatment only (-1.6 mg/kg/min, P < 0.01). Whereas treatment with nifedipine did not change serum lipids, furosemide caused an increase in serum cholesterol (+0.2 mmol/L, P < 0.05) because of a rise in the LDL fraction (+0.32 mmol/L, P < 0.001). The insignificant change in heart rate induced by nifedipine treatment correlated with the change in HbA1c (r = 0.50, P = 0.05) and was inversely related to the change in insulin sensitivity (r = -0.56, P < 0.05). In conclusion, both furosemide and nifedipine caused abnormalities in glucose metabolism. In the nifedipine group the effects on glucose metabolism were related to the occurrence of tachycardia suggesting that sympathetic nerve activation could be involved in the metabolic impairments.