Protein-bound uraemic toxin removal in haemodialysis and post-dilution haemodiafiltration

Background. The accumulation of larger and protein-bound toxins is involved in the uraemic syndrome but their elimination by dialysis therapy remains difficult. In the present study, the impact of the albumin permeability of recently introduced advanced high-flux dialysis membranes on the removal of such substances was tested in haemodialysis and online post-dilution haemodiafiltration. Methods. Two types of polyethersulfone membranes only differing in albumin permeability (referred as PU− and PU+) were compared in eight patients on maintenance dialysis in a prospective cross-over manner. Treatment settings were identical for individual patients: time 229 ± 22 min; blood flow rate 378 ± 33 mL/min; dialysate flow rate 500 mL/min; substitution flow rate in haemodiafiltration 94 ± 9 mL/min. Removal of the protein-bound compounds p-cresyl sulfate (pCS) and indoxyl sulfate (IS) was determined by reduction ratios (RRs), dialytic clearances and mass in continuously collected dialysate. In addition, the elimination of the low-molecular weight (LMW) proteins beta2-microglobulin, cystatin c, myoglobin (myo), free retinol-binding protein (rbp) and albumin was measured. Results. Plasma levels of the protein-bound toxins were significantly decreased by all treatment forms. However, the decreases were comparable between dialysis membranes and between haemodialysis and haemodiafiltration. The RRs of total pCS ranged between 40.4 ± 25.3 and 47.8 ± 10.3% and of total IS between 50.4 ± 2.6 and 54.6 ± 8.7%. Elimination of free protein-bound toxins as assessed by their mass in dialysate closely correlated positively with the pre-treatment plasma concentrations being r = 0.920 (P < 0.001) for total pCS and r = 0.906 (P < 0.001) for total IS, respectively. Compared to haemodialysis, much higher removal of all LMW proteins was found in haemodiafiltration. Dialysis membrane differences were only obvious in haemodialysis for the larger LMW proteins myo and rbp yielding significantly higher RRs for PU+ (myo 46 ± 9 versus 37 ± 9%; rbp 18 ± 5 versus 15 ± 5%; P < 0.05). Additionally, the albumin loss varied between membranes and treatment modes being undetectable with PU− in haemodialysis and highest with PU+ in haemodiafiltration (1430 ± 566 mg). Conclusions. The elimination of protein-bound compounds into dialysate is predicted by the level of pre-treatment plasma concentrations and depends particularly on diffusion. Lacking enhanced removal in online post-dilution haemodiafiltratio