Rates of early AR in pediatric kidney transplantation have declined in every era but the most recent NAPRTCS cohort has shown an increase in late first AR rates. We hypothesized this was due to an increased proportion of deceased donor utilization and early steroid taper utilization. Using the NAPRTCS database, we compared the most recent three cohorts of patients transplanted between 2002-2006, 2007-2011, and 2012-2017. To determine variables that predict late first AR, we used two multivariable models: a standard Cox regression model and LASSO model. From the LASSO model, deceased donor source (P =.002), higher recipient age (P =.019), black race (P =.010), and transplant cohort 2012-17 (P =.014) were all significant predictors of more late first AR. On standard Cox regression analysis, those same variables, minus donor source, were significant, in addition to mycophenolates usage (P =.007) and lower eGFR at 12 months (P =.02). The most recent 2012-2017 cohort remains an independently significant risk factor for late first AR, suggesting unmeasured variables. Further research is needed to determine whether these higher late first AR rates will impact long-term graft survival in the most recent cohort.