On-line mixed hemodiafiltration with a feedback for ultrafiltration control: effect on middle-molecule removal.

On-line mixed hemodiafiltration with a feedback for ultrafiltration control: Effect on middle-molecule removal. Background Increased middle-molecular uremic toxin removal seems to favorably influence survival in dialysis patients. The aim of this study was to verify if, in on-line mixed hemodiafiltration, solute removal by convection may be enhanced by forcing the ultrafiltration rate (Q UF ) and optimizing the infusion technique in order to achieve the highest possible filtration fraction (FF). Methods Removal of β 2 -microglobulin (β 2 -m), urea, creatinine, and phosphate were compared in 20 patients randomly submitted to one dialysis session (A), one postdilution hemodiafiltration session (B), and three sessions of mixed hemodiafiltration (C, D, and E) at different infusion rates (Q S ). In mixed hemodiafiltration, a newly developed feedback system automatically maintained the transmembrane pressure (TMP) within its highest range of safety (250 to 300mm Hg) at constant Q UF , while ensuring the maximum FF by splitting infusion between pre- and postdilution. Results A mean Q S of 134 ± 20mL/min (mean FF=0.65) was attained in post-HDF, and up to 307 ± 41mL/min (mean FF=0.69) in mixed hemodiafiltration. The mean dialysate clearances (K DQ ) for all tested solutes and urea eKt/V were significantly higher in all hemodiafiltration sessions than in dialysis. Only in the case of urea did the infusion mode have no significant effect. K DQ for β 2 -m was maximal in session D and significantly higher than in session B (90.2 ± 11mL/min vs. 77.5 ± 11mL/min; P = 0.02). K DQ for β 2 -m significantly correlated with Q S and the plasma water flow rate (Q PW ). The highest K DQ for β 2 -m was found at values of Q S ∼ Q PW . Beyond this value K DQ decreased. Conclusion The mixed infusion mode in hemodiafiltration, controlled by the TMP-ultrafiltration feedback, seems to improve the efficiency of hemodiafiltration by fully exploiting the convective mechanism of solute removal. The feedback automatically adjusted the infusion rate and site to the maximum FF taking into account flow conditions, internal pressures, and hydraulic permeability of the dialyzer and their complex interactions.