Résumé : Previous studies from our center have shown that donor-recipient HLA-DR mismatches (MM), characterized by the presence of the DR antigen in the donor but not in the recipient or vice versa, are associated with differential effects on graft survival (GS): some of them are beneficial (BEN) with results similar to those of HLA-DR identical or compatible pairs (85% 18 months GS) and some are detrimental (DET) (64% 18 months GS), whereas the other MM, neither BEN nor DET (neutral [NEU]) yield intermediate results (78% 18 months GS). The aim of the present study was to update the results at a longer follow-up time and to assess whether they are influenced or not by prophylactic administration of anti-CD3 mAb (OKT3). The analysis of 234 transplantations performed from 1980 to 1994 with only 1 HLA-DR MM confirmed the BEN effects of HLA-DR5 in either the donor or the recipient and the DET effects of HLA-DR1 or -DR2 in the donor and of HLA-DR2 or -DRW6 in the recipient. These effects were independent of those exerted by other, HLA-DR not related, prognostic factors. The transplants with 1 HLA-DR MM were then compared with those with zero HLA-DR MM (n = 378) and 4 groups were formed according to 2 levels of HLA-DR MM (zero or BEN MM vs. NEU or DET MM) and immunosuppression (with vs. without OKT3 prophylaxis). GS at 5 years was 63% in the group with either zero or BEN MM as compared with 41% in the group with either NEU or DET MM in the absence of OKT3 prophylaxis (P < 0.02); in comparison, with OKT3 prophylaxis, GS at 5 years was 73% in the group with either zero or BEN MM as compared with 58% in the group with either NEU or DET MM (P = 0.07). We conclude that the differential effects of HLA-DR MM on GS are still observed under OKT3 prophylaxis, that the effects of HLA-DR and immunosuppression on graft outcome are additive, and that OKT3 induction therapy is superior to therapy without OKT3. These observations could have important implications for the allocation policy and management of renal transplants.