A total of 632 cyclosporin (CyA)-treated primary renal allograft recipients with a functioning graft at 6 months were retrospectively evaluated for risk factors correlated with long-term allograft function. Mean follow-up after the 6th month was 68.4 +/- 40.6 months. One hundred twenty-one of these patients (19%) were lost: 29 died (23/29 with a functioning graft), 77 of the remaining 92 (83%) lost their graft because of chronic allograft dysfunction, 9 due to recurrence of glomerulonephritis, 5 due to renal artery thrombosis, and 1 due to chronic CyA toxicity. At univariate analysis, factors correlated with a better renal (R) and pure renal (PR) allograft survival were: dialysis duration of less than 5 years, fewer than 2 rejections within the 6th post-Tx month, immediate graft function recovery, plasma creatinine below 1.5 mg/dl at the 6th month, age at Tx above 15 years, and receiving a living donor graft. Cox's regression analysis was also performed to obtain relative risks for the same parameters. Long-term dialysis patients had more frequent late recoveries (P = 0.002) and reductions in therapy (P = 0.01) in order to reduce the side effects of steroids. In young patients receiving an initial oral CyA dose of 17 mg/kg per day, steroids were stopped at the 6th month in order to achieve catch-up growth: only one such patient lost his graft. In contrast, 72% of the young patients who lost their grafts received an initial oral CyA dosage of 13 mg/ kg per day. Thus, young patients did worse not because of steroid withdrawal, but because of inadequate initial CyA dosage. These results suggest that although we cannot exclude alloantigen-independent mechanisms as factors that stimulate progression of chronic allograft dysfunction, it would appear that the initial lesions are induced by events mostly mediated by immunological mechanisms.