Antithymocyte globuline (ATG) and OKT3 have been used for treatment of severe biopsy confirmed acute renal allograft rejection (BCAR). We report results on graft and patient survival including 399 subjects diagnosed with BCAR treated with either ATG or OKT3. Multivariable analyses including Banff scores were performed following three different strategies to account for confounding variables. Fifty per cent of subjects in the OKT3 group had a functioning graft 6.3 years after diagnosis of BCAR, but 74% of ATG patients' grafts were still functioning at that time point (log rank P = 0.006). Median actual graft survival was only 4.6 years in the OKT3 subjects, but 9.5 years for ATG-treated patients (log rank P = 0.004). Multivariable analysis revealed that the risk for functional graft loss was significantly elevated in the OKT3 compared to ATG patients (HR = 1.79, 95% CI 1.06-3.02, P = 0.029). The risk for actual graft loss, counting death as event, was also significantly elevated in the OKT3 patients (HR = 1.73, 95% CI 1.09-2.74, P = 0.019). The hazard of death was not different between the groups (HR = 1.55, 95% CI 0.87-2.77, P = 0.137). These data suggest that rejecting renal allografts treated with ATG exhibit longer graft survival than OKT3 treated transplant kidneys. Causal inference, however, cannot be drawn from this associational study.