SUN-187 MORTALITY RISK BY DIALYSIS MODALITY OVER 1997 TO 2017 IN THE AUSTRALIAN AND NEW ZEALAND DIALYSIS POPULATION

Abstract

The relationship between mortality risk and modality can vary by era. We recently showed that improvements in mortality risk on home hemodialysis (HD) and peritoneal dialysis (PD) are outpacing those on facility HD in Australia and New Zealand (Am J Kidney Dis. 2015 Sep;66(3):489-98). Here, we extend those analyses by directly comparing mortality risk on home HD and PD relative to facility HD across different eras of the recent past. We studied all adult patients commencing renal replacement therapy in Australia and New Zealand in the 20 years prior 31-December-2017 using the Australia and New Zealand Dialysis and Transplant Registry (ANZDATA). We compared mortality risk by modality across different eras of dialysis inception using Cox proportional hazards regression, adjusting for time-varying co-morbidity, demographics, presence and type of diabetes mellitus, late referral, primary renal disease, eGFR at dialysis inception, and initial treating center as a shared frailty. For the main analysis, we used an as-treated framework with time-varying modality and a 90-day lag for the attribution of death to modality. We performed a sensitivity analysis using an intention-to-treat framework with modality assigned at 90 days post-dialysis inception and no lag for the attribution of death. Given the non-proportional hazards for mortality between PD and HD, we repeated the analyses using eras of different lengths (4 x 5-year intervals, 3 x 7-year ones, 2 x 10-year ones). We censored patients at the end of each era to ensure the same maximum duration of follow-up within each era. The study cohort contained a maximum of 52,103 subjects with 19,667 deaths over 140,183 patient-years. 61% were males, median (IQR) age was 62 (22), 46.6% had diabetes mellitus. Compared to facility HD, home HD was associated with better survival by 30-50% across all eras. Compared to facility HD, all forms of PD were associated with poorer survival in earlier eras, but with similar or better survival in cohorts initiating dialysis more recently (e.g. for 2013-2017: continuous ambulatory PD hazard ratio 0.87, 95% confidence intervals 0.78-0.99; automated PD HR 0.90, 95% confidence intervals 0.82-1.00). The sensitivity analysis using the intention-to-treat framework did not meaningfully change estimates. Our study indicates that, despite increased comorbidity, survival outcomes on dialysis are improving with time. Most recently, home HD is associated with better survival relative to facility HD, and PD with similar or better survival. Our analyses have potential for residual confounding from the limited collection of co-morbidity, and lack of socioeconomic, medication and biochemical data in analyses. The results of our study may be due to improved pre-dialysis care, patient selection (eg dialysis modality selection and dialysis versus conservative care in the elderly), improvements in general care of patient co-morbidities and advances in dialysis technology and practice.