Renal allograft biopsy is the accepted gold standard for investigating episodes of graft dysfunction in the early posttransplant period. The situation is less clear in late transplant biopsies. Later renal biopsies performed for graft dysfunction or as part of a routine investigative protocol have not been subjected to detailed critical evaluation. Two hundred sixty-three consecutive renal allograft biopsies in a single center were evaluated. They were arbitrarily divided into three groups based on interval after transplantation: group 1, up to 3 months (n=117); group 2, 4-12 months (n=60); and group 3, greater than 12 months after transplantation (n=86). There were no significant differences in demographic factors among the groups. The mean interval after transplantation was 0.8+/-0.1 months in group 1, 6.1+/-0.3 months in group 2, and 40.1+/-3.4 months in group 3. There were six principal diagnostic categories: acute rejection (AR), chronic rejection (CR), cyclosporine (CsA) nephrotoxicity, acute tubular necrosis (ATN), normal, and others. A statistically significant decrease in the frequency of AR (P<0.001) was seen in group 3 (3%) compared with groups 1 (43%) and 2 (37%). In contrast, the frequency of CR was significantly higher (P<0.001) in group 3 (71%) compared with groups 1 (0) and 2 (10%). ATN was seen almost exclusively in group 1. All but one of the 37 patients with ATN were in this group. CsA nephrotoxicity remained an important cause of graft dysfunction in all three groups, with no significant difference in incidence among the three groups. The differences between groups with other histological types were not significant. Patient management was changed based on the biopsy report in 84 patients in group 1 (72%), 45 patients in group 2 (75%), and only 16 patients in group 3 (19%) (P<0.001). In only seven patients in group 3 did the change in management result in a significant change in serum creatinine. All of these seven patients had CsA nephrotoxicity on biopsy and also had a significantly higher level of CsA compared with those with AR or CR. Thus, the diagnosis might have been possible without the need for biopsy. We conclude that late renal allograft biopsies are only rarely helpful in patient management and as such should be an investigation of last resort.