Fifty patients with mild to moderate essential hypertension were randomized to receive either 20 mg fosinopril daily for 16 weeks or placebo for 4 weeks followed by 12 weeks of 50 mg atenolol daily. Prior to these 16 weeks there was a placebo wash-out period of 2-6 weeks. Blood pressure measurements, euglycaemic, hyperinsulinaemic glucose clamps, and intravenous glucose tolerance tests (IVGTT) were performed at baseline and after 4 and 16 weeks. Blood lipid status was evaluated at baseline and 16 weeks. The insulin sensitivity index (M/I) increased by 12% during the prolonged placebo period, and subsequently decreased by 12% during treatment with atenolol in that group. A post-hoc analysis of covariance indicated that the increase in insulin sensitivity during the initial 4 weeks may have been due to carry-over effects from previous antihypertensive treatment. Fosinopril increased glucose disappearance during IVGTT at 4 and 16 weeks (k values 1.46 and 1.33 vs 1.10 at baseline) but had no effect on insulin sensitivity. The change in insulin sensitivity and serum triglycerides during treatment with fosinopril was related to angiotensin-converting enzyme inhibition in serum. In conclusion, carry-over effects from previous antihypertensive medication were indicated in this study, probably because of an insufficient wash-out period in many patients. Therefore, 4 weeks of placebo wash-out in all patients is advisable in this kind of investigation.