Supplementary Material for: Survival differences in patients with high-flux hemodialysis versus expanded hemodialysis: A cohort study

Introduction
Survival is a core outcome of hemodialysis therapy. Expanded hemodialysis enabled by Theranova dialyzer increases clearance of medium-sized molecules and was reported to improve various patient-reported outcomes compared to hemodialysis using high-flux membranes but the impact on survival is unclear. Herein, we evaluated the risk of death for expanded hemodialysis versus high-flux hemodialysis over a 4- year follow-up period.
Methods
This a multicenter, retrospective cohort study in adult prevalent chronic hemodialysis at Renal Care Services Colombia starting between September 1, 2017, and November 30, 2017, with follow-up for up to 4 years. The sociodemographic and clinical characteristics of all patients were summarized descriptively. The Fine-Gray sub distribution hazard model was employed to evaluate the sub-hazard ratio of factors associated with time to death from causes other than COVID-19; death attributable to COVID-19 was considered as a competing risk. Sensitivity analyses included Cox proportional-hazards model, extended Cox regression, and Cox regression after adjusting for imbalances between cohorts using the inverse probability of treatment weighting method.
Results
We evaluated 1092 patients, 559 in expanded hemodialysis cohort and 533 in high-flux hemodialysis cohort; the mean age 61 years; 42% had diabetes, and 19% had cardiovascular disease. Expanded hemodialysis using the Theranova dialyzer had a protective effect with reduction of mortality risk when controlling for confounding variables, SHR = 0.79 [95% CI: 0.62 to 0.98, p = 0.035]. Sensitivity analyses also showed a statistically significant beneficial effect of expanded hemodialysis versus high-flux hemodialysis.
Conclusions
Expanded hemodialysis enabled by Theranova dialyzer was associated with an approximately 21% reduction in mortality risk compared to treatment with high-flux hemodialysis in patients followed for up to 4 years. It would be desirable for these very promising results to be corroborated by a randomized controlled trial, with sufficient follow-up time to investigate the effect described in this study