Early Mortality Among Peritoneal Dialysis and Hemodialysis Patients Who Transitioned With an Optimal Outpatient Start

Kidney Int Rep. 2018 Oct 16;4(2):275-284. doi: 10.1016/j.ekir.2018.10.008. eCollection 2019 Feb.


Introduction: Lower early mortality observed in peritoneal dialysis (PD) compared with hemodialysis (HD) may be due to differential pre-end-stage renal disease (ESRD) care and the stable setting of transition to dialysis where PD starts are more frequently outpatient rather than during an unscheduled hospitalization. To account for these circumstances, we compared early mortality among a matched cohort of PD and HD patients who had optimal and outpatient starts.

Methods: Retrospective cohort study performed among patients with chronic kidney disease (CKD) who transitioned to ESRD from 1 January 2002 to 31 March 2015 with an optimal start in an outpatient setting. Optimal start defined as (i) HD with an arteriovenous graft or fistula or (ii) PD. Propensity score modeling factoring age, race, sex, comorbidities, estimated glomerular filtration rate (eGFR) level, and change in eGFR before ESRD was used to create a matched cohort of HD and PD. All-cause mortality was compared at 6 months, 1 year, and 2 years posttransition to ESRD.

Results: Among 2094 patients (1398 HD and 696 PD) who had optimal outpatient transition to ESRD, 541 HD patients were propensity score-matched to 541 PD patients (caliper distance <0.001). All-cause mortality odds ratios (OR) in PD compared with HD were 0.79 (0.39-1.63), 0.73 (0.43-1.23), and 0.88 (0.62-1.26) for 6 months, 1 year, and 2 years, respectively. Time-varying analysis accounting for modality switch (19% PD, 1.9% HD) demonstrated a mortality hazard ratio of 0.94 (0.70-1.24).

Conclusion: Among an optimal start CKD cohort that transitioned to ESRD on an outpatient basis, we found no evidence of differences in early mortality between PD and HD.

Keywords: hemodialysis; mortality; optimal end-stage renal disease transition; peritoneal dialysis; propensity-matched cohort.