Intensive Hemodialysis and Mortality Risk in Australian and New Zealand Populations

Abstract

Intensive hemodialysis (HD) is characterized by increased frequency and/or session length compared to conventional HD. Previous analyses from Australia and New Zealand did not suggest benefit with intensive HD, although recent research suggests that relationships have changed. We present updated analyses. Observational cohort study using marginal structural modeling to adjust for changes in renal replacement modality and time-varying medical comorbid conditions. Adults initiating renal replacement therapy since March 31, 1996, followed up through December 31, 2012; this analysis included 40,842 patients over 2,187,689 patient-months. Time-varying renal replacement modality: conventional facility HD (≤3 times per week,≤6 hours per session), quasi-intensive facility HD (between conventional and intensive), intensive facility HD (≥5 times per week, any hours per session), conventional home HD, quasi-intensive home HD, intensive home HD, peritoneal dialysis, deceased donor kidney transplantation, and living donor kidney transplantation. Patient mortality, with a 3-month lag in primary analyses and 6- and 12-month lags in sensitivity analyses. Conventional facility HD was the reference group. Conventional home HD had a similar mortality risk. For quasi-intensive home HD, mortality risk was lower (HR, 0.56; 95% CI, 0.44-0.73). For intensive home HD, mortality risk was nonsignificantly lower in primary analyses and significantly lower using a 6-month lag (HR, 0.41; 95% CI, 0.20-0.85), but not using a 12-month lag. For quasi-intensive facility HD, mortality risk was nonsignificantly lower in primary analyses, although significantly lower using 6- (HR, 0.41; 95% CI, 0.20-0.85) and 12-month lags (HR, 0.59; 95% CI, 0.44-0.80). Mortality risk was similar between intensive and conventional facility HD. For peritoneal dialysis, mortality risk was greater than for conventional facility HD (HR, 1.07; 95% CI, 1.03-1.12). Kidney transplantation had the lowest mortality risk. Potential residual confounding from limited collection of comorbid condition, socioeconomic, and medication data. There is an emerging HD dose-effect in Australia and New Zealand, with lower mortality risks associated with some of the more intensive HD regimens in these countries.