Renovascular hypertension is a potentially curable form of high blood pressure. However, it is unclear how best to select patients who are likely to have renovascular hypertension, what diagnostic strategy to use in these selected patients, and how to predict the hemodynamic significance of a renal artery stenosis. We determined the prevalence of renovascular hypertension in adults who exhibited suggestive clinical features. In these clinically selected patients, we then determined the test characteristics of various diagnostic and potential screening tests. Renovascular hypertension was diagnosed if correction of renal artery stenosis resulted in decreased blood pressure. Of the 66 hypertensive adults evaluated, 11 (16.7%) had renovascular hypertension. Captopril-stimulated peripheral renin activity detected renovascular hypertension with 73% sensitivity, 72% specificity, 38% positive predictive value, and 92% negative predictive value. Less optimal combinations of sensitivity and specificity were found for differential glomerular filtration rate renography, differential effective renal plasma flow renography, and selective renal vein renin ratios, each performed after a single dose of captopril. Intravenous digital subtraction renal angiography detected all patients with renovascular hypertension and was normal in 71% of patients with essential hypertension. To evaluate potential screening tests for renovascular hypertension, we calculated predictive values applied to a low prevalence population. If the observed sensitivities and specificities apply to a population with 5% prevalence of renovascular hypertension, captopril-stimulated peripheral renin would have a positive predictive value of 12% and a negative predictive value of 98%. In 16 patients with known renal artery stenosis, neither the captopril-stimulated renal vein renin ratio nor captopril-stimulated differential renography accurately predicted blood pressure response to correction of the stenosis. We conclude that clinical criteria can identify a subgroup with 16.7% prevalence of renovascular hypertension. In this high prevalence group, intravenous digital subtraction renal angiography will identify virtually all patients with renovascular hypertension, and a normal study will be sufficient to exclude renovascular hypertension. In unselected hypertensive patients, screening with captopril-stimulated peripheral renin activity may be the most useful and efficient procedure for identification of patients with renovascular hypertension. Functional tests do not accurately predict the hemodynamic significance of a renal artery stenosis.