BACKGROUND: Even in low doses, long-term steroid immunosuppression is known to cause serious complications. However, the safety of steroid withdrawal has not been proven in randomized clinical trials. This study examines donor-specific hyporesponsive transplant recipients before and after steroid withdrawal, to see if reduction in immunosuppression was associated with consistent changes in antidonor immunological reactivity. METHODS: Using limiting dilution assays, the circulating precursor frequency of donor and third-party-reactive helper T lymphocytes (HTLpf) were determined in 21 consecutive cadaveric renal allograft recipients on standard triple therapy, before (pre-tx) and at different time points after transplantation (post-tx). Patients were selected for steroid withdrawal by clinical criteria (stable graft function and no or only one very mild rejection episode). RESULTS: Of 21 patients studied, steroids were successfully withdrawn in nine (steroid withdrawn group, SWG) for at least 187 days (mean: 380 +/- 168.5), and were not withdrawn in 12 patients (steroid continued group, SCG). In the SWG seven of nine patients developed at least fivefold reduction of post-tx donor-reactive HTLpf (range 5-17), relative to pre-tx, as compared to two of twelve patients in the SCG, P = 0.01. In both groups, the third-party-reactive HTLpf in most of these patients remained largely unchanged throughout the study period. In the SWG, no significant difference of serum creatinine level was found before and at 6 months after steroid withdrawal (mean: 138 +/- 24 versus 132 +/- 40, P = 0.45). CONCLUSION: Patients who developed donor-specific hyporeactivity as evidenced by low donor-reactive HTLpf had stable graft function and stable HTLpf levels after complete steroid withdrawal.