The assessment of an erectile dysfunction (ED) includes the history, a clinical examination and blood tests. There is some confusion about which basic hormonal tests are needed at the beginning of clinical evaluation. We feel that with the results from our patients we could help to answer this question. From 1 January 1990 until the December 31 1993 we evaluated 1134 patients for ED. Those who favoured a surgical correction of their ED were fully evaluated by nocturnal penile tumescence testing, penile arteriography, intracavernosal injection of vasoactive agents and dynamic pharmaco-cavernosometry. The results from these tests were correlated with luteinizing hormone (LH), follicle-stimulating hormone (FSH), testosterone and prolactin. 183 (16.1%) of our patients with a mean age of 45 +/- 14 were fully evaluated. From these patients 76 were excluded because their ED was posttraumatic, undoubtedly psychogenic or could not be proven by the tests mentioned above. From the 107 patients finally included in this study, 90 had normal endocrine parameters. 17 patients had low testosterone. 14 of these patients had otherwise completely normal hormonal tests without evidence of secondary hypogonadism. Three patients had their low testosterone levels confirmed by repeated measurements. In addition, prolactin was significantly increased, and FSH and LH were near or below the lower reference value. When evaluating patients for the first time because of an erectile dysfunction, the measurement of testosterone as a single endocrine test is adequate. If testosterone is low, repeated measurements, combined with LH, FSH and prolactin, will identify patients with an ED due to an endocrine disease.