Hemodialysis catheter rewiring vs. removal and replacement: post hoc analysis of a national stepped wedge cluster randomized trial.

Abstract

Background: Tunnelled haemodialysis catheters are widely used but often fail prematurely. The optimal replacement strategy is unknown. Rewiring is less burdensome for patients but may predispose to infection. We aimed to quantify variation in tunneled haemodialysis catheter rewiring practices among Australian nephrology services, and to determine whether rewiring was associated with infection. Method(s): In a post-hoc analysis of the national stepped wedge cluster randomized REDUCCTION trial, encompassing 37 nephrology services and 6399 adult patients, we examined variation in the service-wide proportion of tunnelled catheters that were replaced by rewiring, for infectious or non-infectious reasons. Given the findings, we compared the absolute risks of, and time to infectious removal between rewired and nonrewired (new exit site) catheters that replaced failing non-infected catheters. Marginal Cox and competing risk proportional hazard models, including catheter, patient, and service-level covariates were used. Confirmed bloodstream infections were assessed in sensitivity analysis. Competing risks included removal for dysfunction and death before removal. Result(s): Services universally avoided rewiring infected catheters but varied widely in rewiring non-infected failing catheters (range = 0 - 90% rewired). Among new catheters that replaced failed non-infected catheters, 36 of 480 rewired (7.5%), and 36 of 372 nonrewired (9.7%) were removed for infection. At three months the cumulative incidence of premature infectious and mechanical removals was 5% and 21% for rewired and 7% and 23% for non-rewired respectively. The hazard of infectious removal did not differ between rewired and non-rewired catheters (adjusted HR 0.81, 95% CI 0.46, 1.42). The incidence of confirmed catheter-related bloodstream infections and catheter dysfunction requiring removal were similar between groups. Conclusion(s): The practice of rewiring non-infected failing catheters varied widely between services, was not associated with catheter infection, and did not appear to affect the high rate of subsequent catheter dysfunction in this population.