Analysis of OPTN/UNOS registry suggests the number of HLA matches and not mismatches is a stronger independent predictor of kidney transplant survival

HLA matching and mismatching, while inversely related, are not exact opposites. Here we determined the independent effects of HLA matching and mismatching on outcomes in deceased donor kidney transplant recipients. The United Network for Organ Sharing database (1995-2012) was utilized and analyzed for delayed graft function, one-year acute rejection, and death-censored graft survival using combined multivariable models including HLA matching and mismatching. Sensitivity analyses were performed using the subgroup of deceased donor kidney transplant patients after 2003 with more uniform HLA nomenclature and resampling analyses using bootstrapping on complete data available from 96,236 recipients. Individually, both HLA matching and mismatching showed significant associations with graft survival. Adjusting the model to take into account both matching and mismatching simultaneously, the degree of HLA mismatching lost significance while matching continued to have a significant prediction for delayed graft function, the one-year acute rejection rate, and graft survival. Sensitivity analyses and bootstrapping showed similar results for all studied outcomes. Thus, analysis of this large cohort demonstrates the apparent greater association of HLA matching over HLA mismatching on both early allograft events as well as graft survival. Future analyses should preferentially utilize HLA matching as a covariate over mismatching for accurately reflecting impact on graft outcomes.