Abstract
INTRODUCTION: While medically induced end-stage renal disease (m-ESRD) has been well-studied, outcomes in patients with surgically induced ESRD (s-ESRD) are unknown. We sought to quantitatively compare the non-oncological outcomes for s-ESRD and m-ESRD in a large, population-based cohort. METHODS: Medicare patients >65 years old initiating hemodialysis were identified using the U.S. Renal Data System database (2000-2012). Metastatic cancer, prior transplant history, and nephrectomy for polycystic kidney disease were exclusion criteria. Patients were classified as having s-ESRD or m-ESRD based on hospital and physician claims for nephrectomy within a year preceding the onset of maintenance hemodialysis. Outcomes included non-cancer mortality (NCM), overall survival (OS), cardiovascular event (CVE), and renal transplantation. Time-to-event analyses were performed using Kaplan-Meier and cumulative incidence curves, and multivariable Cox and Fine-and-Grey regression models. RESULTS: The cohort included 312 612 patients, of whom 1648 (0.53%) had s-ESRD. Compared to m-ESRD patients, s-ESRD patients had a significantly lower five-year cumulative incidence of NCM (68% vs. 80%; p<0.001) and CVE (62% vs. 68%; p<0.001), with a correspondingly higher probability of OS (22% vs. 17%; p<0.001) and rate of renal transplantation (3.6% vs. 2.0%; p<0.001). On multivariable analyses, s-ESRD remained associated with lower risks of NCM (p<0.001) and CVE (p<0.001), improved OS (p<0.001), and higher chance of renal transplantation (p<0.001). CONCLUSIONS: While outcomes for s-ESRD appear more favorable than m-ESRD, s-ESRD is still associated with a substantial risk of NCM and CVE, and a low incidence of renal transplantation in Medicare patients >65 years old. These non-oncological outcomes are worth considering in patients potentially facing postoperative ESRD.