To determine any benefit of maintenance steroids in a cyclosporine and azathioprine immunosuppressive regimen, 112 heart transplant recipients were prospectively randomized to receive cyclosporine, azathioprine, and prednisolone (n = 59) or cyclosporine and azathioprine (n = 53). Of the 53 double-therapy patients, 47% were converted to maintenance steroids for resistant rejection or renal dysfunction. In a comparison of true double-therapy (n = 28) versus true triple-therapy (n = 59) groups, actuarial survival and systolic function did not differ. Linearized rejection during the first 3 months was lower with triple therapy than with double therapy (1.5 +/- 0.18 vs 2.3 +/- 0.23 episodes/100 patient-days; p less than 0.01) as were requirements for cytolytic therapy for rejection with hemodynamic compromise. Patients receiving triple therapy had significantly higher serum cholesterol levels and antihypertensive agent requirements at all annual time points up to 5 years. The rate of steroid-related morbidity (diabetes, bone complications, cataracts, and obesity) was low in both groups and did not differ significantly. Of the 204 patients receiving triple therapy at this unit, 45 underwent steroid withdrawal. The initial success rate was 69%, and an additional 14% of those who initially failed succeeded on the second attempt. Any rejection after steroid cessation tended to occur within 6 weeks. There were, however, no substantial short-term benefits in body weight or lipid or blood pressure control. In patients in whom infection or growth retardation was an indication for steroid withdrawal, these generally improved after cessation. Until predictive markers for the likely success of steroid withdrawal are identified, the case for steroid withdrawal, as opposed to steroid minimization, does not seem compelling.