Effects of Cardiovascular Comorbidity Adjustment on SRTR Risk-Adjusted Cox Proportional Hazard Models of Graft Survival

Abstract

The Scientific Registry of Transplant Recipients (SRTR) and the Centers for Medicare and Medicaid Services (CMS) determine expected graft survivals to identify potentially underperforming transplant centers. There has been recent interest in evaluating adjustments for comorbidities when performing these calculations. This study was performed to determine the influence that adjustment for pre-transplant cardiovascular disease comorbidity can have on risk-adjusted Cox models, such as those used by SRTR and CMS. We analyzed Cox proportional hazards models for 1-year and 3-year graft survival for kidney recipients from a single center where cardiovascular disease covariates were added to a baseline model derived by using the SRTR calculated risk scores and including all standard SRTR parameters. Living and deceased donor recipient 1-year and living donor 3-year Cox models that included all seven covariates demonstrated 8% to 13% improved discrimination. Only the 1-year deceased donor recipient Cox model demonstrated significantly improved calibration (likelihood ratio test P=0.038). The expected graft losses increased by >30% for living donor recipients at 1 and 3 years and decreased by 2% to 4% for deceased donor recipients at 1 and 3 years. SRTR and CMS use of pre-transplant cardiovascular comorbidity adjustment might impact center performance evaluations.