Androgens influence important central and peripheral mechanisms of the erectile system. The relevance of a moderate decrease of serum testosterone level for erectile dysfunction (ED) has not been clarified so far. The aim of our study was to offer an easy transcutaneous method of androgen application. A previous study on the pharmacokinetic profile of the testosterone gel applied, showed marked elevation of the serum levels of testosterone. In our study, 46 hypogonadal patients with ED and total lack of vaginal penetration applied testosterone gel (4 mg/day; supplied by Azupharma, Germany) to the penile skin twice a day over 6-8 weeks, after a run-in period with placebo gel of 2 weeks. All patients showed decreased testosterone serum levels (<3 ng/ml) in at least two morning samples over a period of 3 weeks before treatment. Psychogenic etiology was excluded by a sexual psychologist. Patient age was 37-69 years (mean 53.5). Three patients (6.5%) responded to placebo in the run-in phase and were withdrawn from further treatment. Fifteen patients (32.6%) showed improved erection, allowing penetration and sexual intercourse. Twenty-eight patients (60.9%) did not respond to therapy. Local genital skin irritation was not observed. Elevation of peripheral testosterone was not correlated to a positive therapy response. A success-rate of 32.6% in this group of patients after exclusion of psychogenic patients and placebo-responders seems to justify further investigations. A medication period of 6-8 weeks is most probably too short to induce imaginable regenerative effects of testosterone on the erectile system. We therefore suggest that future double-blind and placebo-controlled studies should be designed for a minimum of 3 months. Testosterone gel may be a cost effective form of androgen administration.