1. The transfusion effect, which had its greatest impact in 1985, disappeared in 1989. Since then, transfused and nontransfused groups have had identical 1-year graft survival rates. 2. The percentage of nontransfused recipients reached nearly 50% in 1992. The amount of pretransplant transfusion blood has dropped from 7 units in 1982 to below 5 units in 1992. Whole blood and packed red cells were the preferred products, comprising 98% of total transfusions in 1992. 3. The transfusion effect was observed in both sensitized and nonsensitized transfused recipients until 1988, although the effect was greater in the nonsensitized group. Since 1989, the sensitized group of transfused recipients had an even lower 1-year graft survival rate, while the nonsensitized group had a slightly better survival rate than those with no pretransplant transfusions. 4. Although washed red cells, frozen red cells, and frozen plasma constituted only a small percentage of total transfusions, any pretransplant transfusion of these blood products still produced a transfusion effect. Since 1982, patients who received washed red cells, frozen red cells, or frozen plasma had 8% higher 1-year graft survival rates since 1982 than those who had no transfusions, a difference still patient in 1990. 5. The percentages of broadly sensitized recipients decreased significantly in the last decade. Even recipients who had very specific antibodies (PRA 1-10%) had significantly lower graft survival rates than those with no antibodies (PRA = 0). Patients with PRA 1-10% and those with PRA 11-50% had almost identical graft survival curves and the same half-life, whether first or regrafted recipients. Only 0-PRA patients could be distinguished as "nonsensitized." 6. Once patients were sensitized, any time pretransplant, the current PRA values had less meaning for predicting graft survival than the peak (historical) PRA values. Those with peak PRA 51-100% had the same graft survival rates regardless of their current PRA values. Peak PRA was more important than current PRA for predicting graft status. 7. The deleterious effect of a positive flow cytometry crossmatch (FCXM) was obvious in both first and regrafted recipients. The positive group had about a 10% lower 1-year graft survival rate than the FCXM-negative group. 8. The significance of the positive FCXM outweighed the current PRA values, especially in regrafted recipients and first transplant recipients with peak PRA values. 9. FCXM seemed to work only on patients who had a chance of being sensitized by HLA stimulation. There was no positive FCXM effect in nontransfused recipients.(ABSTRACT TRUNCATED AT 400 WORDS)