1. In 1984, second graft survival rates were 10% lower than first grafts, but in 1992, the survival difference was reduced to 1%. Multiple grafts in 1984 were 23% lower than first grafts, but showed only a 7% difference in 1992. In 1992, 12% of kidney grafts were performed into second graft recipients and 3% into multiple graft recipients. 2. If first grafts survived one to 12 months posttransplant, the second graft survival was less than if they had survived longer than 12 months, as seen in many previous analyses. Here we showed that patients with a first graft duration of one to 12 months had a higher incidence of sensitization than patients with a first graft duration of more than 12 months. This may indicate that immunization was the cause of failure more frequently among those patients who rejected earlier than later. 3. Since 1989, the interval between first graft rejection and second graft transplantation was not a factor in second graft survival. A strong correlation was noted between high PRA and interval to regrafting. This probably reflects the increasing difficulty in finding negative-crossmatch donors as PRA increases. 4. Repeat mismatches for HLA-DR were deleterious to second grafts, although repeat mismatches for HLA-A,B were not, confirming earlier studies (1,5). HLA-A,B,DR mismatches correlated well with second and multiple transplant outcomes. 5. Patients receiving second cadaver-donor transplants had the same graft survival regardless of whether the first graft was another cadaver donor or a living-related one. On the other hand, second living-related donor transplants had a higher graft survival rate if the first graft had also been from a living-related rather than a cadaver donor (p < 0.05). This suggests that it would be advantageous if the first graft came from a living-related donor with a cadaver donor as the second graft, rather than the reverse situation. 6. Urine production on the first postoperative day was a very strong indicator of subsequent graft survival, particularly for second and multigraft patients. Failure to diurese on the first day resulted in a second graft survival of 60% at one year compared with 80% for those that diuresed on the first day. 7. Similarly, dialysis requirements were a major factor in predicting subsequent graft survival. For second graft patients who required dialysis, one-year graft survival was 63%, compared with 84% if no dialysis was needed. 8. The fraction of patients who had insulin dependent diabetes for first grafts was 27%, 15% for second grafts, and 9% for multiple grafts.(ABSTRACT TRUNCATED AT 400 WORDS)