In forty six hypertensive patients in whom a high level of clinical suspicion for renovascular hypertension was present on the basis of clinical clues, a captopril test was performed with either 25 mg of captopril or placebo on 2 separate days to determine prospectively the value of the captopril test. Blood pressure (BP) and peripheral renin were used as response variables. All patients had discontinued their anti-hypertensive medication and were not salt depleted. In all patients selective renal angiography was performed irrespective of the results of the captopril test. Twenty patients proved to have uni- or bilateral renal artery stenosis (RAS) giving a prevalence of 43%. After the placebo and after captopril there were no significant changes (absolute or proportional) in BP values between patients with essential hypertension or RAS, either for all measurements or if only the fall in BP was taken into account. The receiver operator characteristic curves of both baseline and post-captopril peripheral renin levels indicate that the captopril test does not discriminate appropriately between patients with essential hypertension and RAS. Therefore, we would not advise the use of the captopril test as a screening test for RAS in hypertensive patients in whom a high level of clinical suspicion for RAS is already present.