*Purpose: African-American (AA) kidney transplant recipients (KTx) have higher risk of rejection and failure. However, it is unknown to what extent the inferior outcomes in self-reported AAs are due to genetic versus environmental effects. Herein, we compared the effects of self-reported race versus genetic African admixture on graft outcomes.

*Methods: A discovery multiethnic cohort of 1,083 KTx from Columbia University and a replication cohort of 761 kidney transplant recipients from University of Pennsylvania were genotyped with high resolution SNP arrays. African admixture proportions (AFR), a genetically-derived quantitative measure of African ancestry, was estimated with ADMIXTURE software. US census tract variables matched with participants’ geocoded addresses were used to analyze the effect of neighborhood socio-economic status (SES) factors. Multivariable Cox models were used to investigate associations between African ancestry measures and time to biopsy-proven rejection and time to death-censored graft failure, with adjustments for relevant covariates. Akaike information criterion (AIC) was used to compare the models. Meta-analyses of the results in the two cohorts were computed using random-effect models.

*Results: 206 and 346 self-identified AA were included in the discovery and replication cohorts, respectively. Over a median time of 78 months, 432 patients had rejection and 193 had graft failure in the discovery cohort; during a median time of 49 months, 113 patients had rejection and 121 had failure in the replication cohort. In the discovery and replication cohorts self-reported AA ancestry and AFR were associated with acute rejection and graft failure (Figure 1). AIC values show that AFR is non-inferior or slightly better predictor than self-reported AA. The HR for measures of ancestry only mildly decreased when adjusting for SES factors in both cohorts. Meta-analyses of the results confirmed that self-reported AA and AFR were associated with rejection and failure independently of clinical covariates and SES.

*Conclusions: In conclusion, self-reported AA race and a genetically-derived continuous measure of African ancestry predict the risk of allograft rejection and failure in multiethnic and genetically diverse cohorts. A composite variable for US census-derived neighborhood SES did not confound the association between ancestry and either rejection or failure.

« Back to 2021 American Transplant Congress