In this double blind cross-over study, 20 infertile men, who had sperm antibodies detected by the mixed antiglobulin reaction (MAR test) in the ejaculate and by the tray agglutination test (TAT) in serum, were treated with 40 mg/day prednisolone or placebo from days 1 to 10 of the partners' menstrual cycle. Patients were randomly allocated to different treatment groups. While group 1 started with placebo followed by verum for three consecutive cycles, group 2 began with verum and continued with placebo. All patients had regular intercourse (n = 19) or intra-cervical insemination at ovulation (n = 1). A post-coital test or a sperm penetration test was performed during verum and placebo regimes. Blood samples were drawn from the male partner at this time to control the efficacy of prednisolone treatment by checking the TAT titre. No pregnancy occurred during prednisolone or placebo treatment. In eight of 12 patients, post-coital testing showed little improvement and antibody titres decreased in seven of 16 patients. Side-effects from medication were reported by eight patients (seven verum and one placebo cycle) and caused treatment to be discontinued in two cases. Five patients' partners conceived at a later stage by intrauterine insemination with spermatozoa prepared by 'swim up' (n = 3) or by in-vitro fertilization (IVF n = 2). Thus high dose corticosteroid therapy was ineffective in achieving pregnancies induced by infertile men positive for antisperm antibodies. Since side-effects of corticosteroids should not be underestimated in otherwise healthy men, other reproductive techniques such as intrauterine insemination or IVF should be offered to such couples.