Risk factors for sensitization by blood transfusions. Comparison of the UW/Madison and UC/San Francisco donor-specific transfusion experience.

Abstract

We have tested the predictive model for risk of sensitization by donor-specific transfusions developed at the University of California, San Francisco for its applicability to the DST experience at UW/Madison. Patient sample sizes between the two groups were similar (n = 249 for UW/Madison, n = 261 for UCSF) and the two groups of patients had similar compositions in terms of mean age, ABO type, baseline panel-reactive antibody, and pregnancy rate. The two groups differed in that the UW/Madison group had a higher percentage of males, diabetic patients, previously transplanted patients, and 0 haplotype-matched (2 HLA-mismatched related and unrelated) recipients. In addition, all the UW/Madison patients received azathioprine (AZA) whereas only half the UCSF group was given AZA. Despite these differences, application of the UCSF model for prediction of sensitization by DST gave remarkably similar results in our patient population, with pregnancy, prior transplant, baseline PRA, and HLA antigen sharing giving similar odds ratios and P values. However, when female sex was included in the model along with the other variables, the significance of pregnancy risk disappeared. We have developed an alternative multivariate model using stepwise logistic regression that identifies baseline PRA greater than 40%, female sex, and prior transplant as significant risk factors for sensitization, while number of HLA (A, B, DR) antigens shared and increasing recipient age significantly decreased risk of sensitization by DST.