Effect of ultrafiltration profiling on outcomes among maintenance hemodialysis patients: a pilot randomized crossover trial

Background More rapid fluid removal during hemodialysis is associated with adverse cardiovascular outcomes and longer dialysis recovery times. The effect of ultrafiltration (UF) profiling, independent of concomitant sodium profiling, on markers of intradialytic hemodynamics and other outcomes has been inadequately studied. Methods Four-phase, blinded crossover trial. Participants (UF rates > 10 mL/h/kg) were assigned in random order to receive hemodialysis with UF profiling (constantly declining UF rate, intervention) vs. hemodialysis with conventional UF (control). Each 3-week 9-treatment period was followed by a 1-week 3-treatment washout period. Participants crossed into each study arm twice (2 phases/arm); 18 treatments per treatment type. The primary outcomes were intradialytic hypotension, pre- to post-dialysis troponin T change, and change from baseline in left ventricular global longitudinal strain. Other outcomes included intradialytic symptoms and blood volume measured-plasma refill (post-dialysis volume status measure), among others. Each participant served as their own control. Results On average, the 34 randomized patients (mean age 56 years, 24% female, mean dialysis vintage 6.3 years) had UF rates > 10 mL/h/kg in 56% of treatments during the screening period. All but 2 patients completed the 15-week study (prolonged hospitalization, kidney transplant). There was no significant difference in intradialytic hypotension, troponin T change, or left ventricular strain between hemodialysis with UF profiling and conventional UF. With UF profiling, participants had significantly lower odds of light-headedness and plasma refill compared to hemodialysis with conventional UF. Conclusions Ultrafiltration (UF) profiling did not reduce the odds of treatment-related cardiac stress but did reduce the odds of light-headedness and post-dialysis hypervolemia. Trial registration Clinicaltrials.gov identifier: NCT03301740 (registered October 4, 2017)