MO654: Automatic KUF Max Determination as a Tool to Prescribe a Personalized Convection Flow in Post-Dilution Haemodiafiltration by Predicting the Transmembrane Pressure to Which the System will be Subjected

Abstract

Abstract BACKGROUND AND AIM Post-dilutional haemodiafiltration (HDF) with high convection volumes are suggested to improve survival. However, a large increase in convection flow (QUF) is associated with high transmembrane pressure (TMP) which, when exceeding limits advised by guidelines, acts as a limiting factor and prevents achieving the prescription. Our aim is to predict the TMP needed to reach any given convection flow for a patient and dialysis session. METHOD The TMP necessary to obtain a desired convection was predicted from the automatic determination of the maximum ultrafiltration coefficient (KUF max, HDF machine Dialog iQ, BBraun, Melsungen, Germany) and compared to the actual value for 16 patients and 29 dialysis sessions. KUF max was determined at the start of the session and the coefficients of the quadratic regression of the ultrafiltration coefficient (KUF) on convection flow (KUF = a*(QUF)² + b* QUF + c) were obtained and used to predict the TMP needed to obtain a convection of 30% of QB (QUF max), knowing that TMP = QUF/KUF. Then, infusion rate was set at QUF max and the observed TMP was recorded after 8 minutes stabilization. Dialyses were performed with the patient's usual dialyzer (Diacap pro for 28 sessions, Nephral 400 for 1 session) and blood flow [363 ± 30 mL/min (300–400 mL/min)]. Observed and estimated TMPs were compared. RESULTS There was a high correlation between both TMPs (Pearson's R = 0.98; P < .01). The Bland and Altman analysis showed an average difference (systematic bias) of 3.3% with 95% limits of agreement of −9 and +15% and a maximal difference of 12%, indicating a slight overestimation. CONCLUSION KUF max determination allows estimating TMP at any QUF prescription with an acceptable error (3%). Therefore, the KUF on QUF parabolic curve may be of help in the clinical setting to prescribe the most suitable convection in post-dilutional HDF by predicting the TMP the filtrating system will be submitted to. Based on this principle, automatic systems to aid personalized convection flow prescriptions can be developed.