The prevalence of renal artery stenosis in diabetic patients is unknown, since no noninvasive and valid screening procedures are available. We have therefore evaluated 5194 consecutive autopsy protocols from patients who died between 1980 and 1988. In addition, all available clinical records for patients with renal artery stenosis (RAS) and a random sample were evaluated. It was found that 73% of patients with RAS were hypertensive, and 53% had diabetes, all but one being Type 2 (non-insulin-dependent). Renal artery stenosis was present in 225 (4.3%) of all patients [95% confidence interval (95% CI), 3.8-4.9], and was not reported in the patients' clinical records in 93% of cases. RAS was present in 8.3% of all diabetic patients (95% CI, 6.8-9.8%), the odds ratio being 3.5 (95% CI, 2.6-4.6). The frequency of renal artery stenosis in diabetic patients with hypertension was 10.1%. Bilateral renal artery stenosis was found in 43% of patients with RAS and diabetes, and in 30% of non-diabetic patients with RAS (P = 0.059). Our results indicate that renal artery stenosis often goes undetected before autopsy. The presence of non-insulin-dependent diabetes mellitus increases the risk of renal artery stenosis. The risk of bilateral renal artery stenosis is greater in diabetic patients. These results may be of significance with regard to the diagnostic evaluation and choice of antihypertensive treatment in hypertensive non-insulin-dependent diabetic patients.