Association of early initiation of dialysis with all-cause and cardiovascular mortality: A propensity score weighted analysis of the United States Renal Data System

Background Early initiation of maintenance hemodialysis has been associated with excess mortality in some studies, but the effects on cardiovascular (CV) mortality has not been studied. Moreover, whether the increased mortality is due to co-morbidities or early initiation of dialysis is unclear. We used a propensity score weighted analysis of the United States Renal Data System (USRDS) to examine how the estimated glomerular filtration rate (eGFR) at initiation of dialysis affects total and CV mortality. Methods Association between tertiles of eGFR at initiation of hemodialysis and all-cause and CV mortality were assessed in 676,196 adult patients who initiated hemodialysis between 2006 and 2014, using inverse probability of treatment weighting (IPTW) weighted multivariable regression models. Results The intermediate (eGFR 8.7 to <13.0 mL/min) and early start groups (eGFR >= 13.0 mL/min) had a 42% and 93% increased all-cause mortality, respectively compared to late (eGFR < 8.7), start group (unadjusted hazard ratio (HR) = 1.42; 95% CI, 1.41-1.43 and HR = 1.93; 95%CI, 1.91-1.94, respectively). This association was attenuated but remained significant in propensity weighted multivariable analysis (adjusted HR = 1.13; 95%CI, 1.12-1.14 for intermediate and HR = 1.37; 95%CI, 1.36-1.39, for early start, respectively). The CV mortality was similarly increased (adjusted HR = 1.08; 95%CI, 1.07-1.10 and HR = 1.23; 95%CI, 1.21-1.24, for intermediate and early start, respectively). In patients with cystic kidney disease, all-cause mortality was increased with early start, but there were no differences in CV mortality between groups. Conclusions Early initiation of dialysis is associated with increased all-cause and CV mortality. Our observations support delaying hemodialysis according to the eGFR values.