Abstract Background and Aims Recently, the positive results of the CONVICE trial have raised the question whether the improved outcome of postdilution high volume HDF (oHDF-p) compared to high flux HD (hfHD) requires increased water consumption. From mathematical modelling it can be shown that oHDF-p always leads to higher low molecular clearance than hfHD at equal total dialysate consumption Qd_tot = Qd_dial + Qs (Qd_dial: dialysate flow through the dialyzer, Qs: substitution flow). This has also been shown in a clinical trial (Mesic et al, HDI 15, p. 522-529). However, mathematical modelling might not be intuitive for everyone, and clinical trials are always influenced by multiple confounders. To add a more intuitive way of reasoning, a simple educational bench model was designed based only on equipment easily available in a standard dialysis clinic which allows to compare dialysis efficiencies based on the urea concentrations as surrogate parameter. Since the fundamental task of dialysis is to cleanse blood from uremic solutes, the treatment modality which at equal resource utilization reduces solute concentration most is the more efficient one! Method Dialysis treatments were simulated in single pass setup using a Fresenius 5008S machine (Fig. 1). Blood inlet was simulated as a fixed urea concentration of 170 mg/dL in dialysate from a 110 L reservoir “cbi”. Treatment parameters were set such that they correspond to effective values in in-vivo treatments: As the mass transfer coefficient is always much higher in vitro with water than in vivo with blood, a small high-flux dialyzer (Fresenius FX40) was used. Blood flow Qb was set to 300 mL/min, corresponding to the in vivo blood water flow for Qb_set ∼ 350 mL/min. Alternating experiments in oHDF-p and hfHD were carried out in triplicate with Qd_tot=400 mL/min and 300 mL/min. This resulted in the settings HDF400, HD400, HDF300 and HD300 (cf. Fig. 1). Qs was set to 100 ml/min in oHDF-p. In each setting the fluid from blood and dialysate outlet were collected (recipients “cbo” and “cdo”, resp.) for 8 min. To be independent of the machine accuracy of set flow rates, sample volumes were measured by weighing, and effective flow rates of blood Qb_eff and total dialysate Qd_tot_eff were calculated. Samples were taken into monovettes from “cbi”, “cbo” and “cdo”. Urea concentrations cbi, cbo and cdo were determined by sending the samples to two commercial clinical labs who delivered consistent results. Blood and dialysate side urea clearances were calculated as KUB= Qb_eff*(cbi-cbo)/cbi and KUD=Qd_tot_eff*cdo/cbi. Results The table in Fig. 1 includes the measure values of Qb_eff and Qd_tot_eff for each setting. Blood and dialysate outlet concentrations (cbo and cdo, resp.) and calculated blood and dialysate side clearances are shown in Fig. 2. Lab results in repeated settings were highly reproducible with a SD < 2 mg/dL. Switching from hfHD to oHDF-p at equal Qd_tot always decreased cbo, corresponding to a higher clearance in oHDF-p. Reducing total dialysate flow from 400 to 300 ml/min caused a slight decrease in clearance in both modalities. KUB and KUD were in good agreement, showing that urea mass balance was well controlled in the experiment. Conclusion The educational bench model demonstrates the superior low molecular clearance of oHDF-p compared to hfHD at equal Qd_tot, reflected by lower urea cbo values in oHDF-p. Although the clinical outcome advantage of oHDF-p is rather related to middle molecular clearance, the model shows that the improved treatment quality of oHDF-p can be provided with no extra resource utilization compared to hfHD, and that dialysis adequacy as measured by urea clearance is even increased despite of the fact that a part of Qd_tot is deviated as substitution fluid. As moderately reducing Qd_tot in many cases only results in acceptable clearance reductions, the automated adjustment of both dialysate and substitution flow rates for oHDF-p could improve patient outcome at even reduced resource utilization.
Read full abstract