This prospective multicenter study included 1,205 patients, who were referred for difficult-to-treat hypertension or analysis of possible secondary hypertension. After a standardized selection protocol based on sharply defined drug-resistant hypertension or renal function impairment during angiotensin-converting enzyme inhibition, patients underwent renal scintigraphy and a captopril-renin challenge test. A set of clinical characteristics was also recorded. Sensitivity and specificity of renal scintigraphy for diagnosing renal artery stenosis were 0.72 and 0.90 and of the captopril-renin test 0.77 to 0.91 and 0.69 to 0.75 depending on the criterion used. The clinical characteristics were used to construct a clinical prediction rule for renal artery stenosis, which had a sensitivity of 0.68 and a specificity of 0.87 at a cut-off level of 30% predicted probability. However, with the prediction rule a sensitivity of 0.90 could be reached by performing arteriography only in patients with a predicted probability of stenosis of > or =10%, resulting in a considerable reduction of the number of arteriograms to be made. A diagnostic strategy is advocated starting with drug-resistant hypertension and continuing to renal arteriography only in patients with increased probability of stenosis. Patients with atherosclerotic renal artery stenosis were then randomized to balloon angioplasty (n = 56) versus antihypertensive medication (n = 50). Three months after randomization 22 patients from the medication group underwent balloon angioplasty in second instance. In an intention-to-treat analysis, no difference in blood pressure was found between the groups after 3 months, nor after 12 months of follow-up, although there was a small medication-sparing effect of balloon angioplasty. The lack of a beneficial effect of balloon angioplasty compared with medication could not be attributed to the high stenosis recurrence rate after angioplasty, nor to the fact that the inclusion criterion was set at a stenosis level of > or =50% so that patients with relatively mild stenosis were also included. Renal function after angioplasty was slightly better in the angioplasty group than in the medication group, and improvement of the renal scintigram occurred more often after angioplasty. Apart from the treatment of patients with specific characteristics, the presented therapeutic approach starts with extending the antihypertensive drug therapy to control blood pressure. Only if blood pressure cannot be controlled or if renal function deteriorates, balloon angioplasty (with stent placement) is indicated.