Dialysate-Side Urea Kinetics. Neural Network Predicts Dialysis Dose During Dialysis

FERNÁNDEZ, ELMER ANDRÉS; VALTUILLE, RODOLFO; WILLSHAW, PETER; PERAZZO, CARLOS ALBERTO

MEDICAL AND BIOLOGICAL ENGINEERING AND COMPUTING

Año: 2003 p. 392 - 396

0140-0118

Determination of the adequacy of dialysis is a routine but crucial procedure in patient status evaluation. The total dialysis dose expressed as Kt/V has been widely recognized to be a major determinant of morbidity and mortality in hemodialyzed patients. Many different factors influence the correct determination of Kt/V such as urea sequestration in different body compartments, access and cardiopulmonary recirculation. These factors are responsible for urea rebound after the end of the Hemodialysis session, causing poor Kt/V estimation. There are many techniques which try to overcome this problem. Some of them use analysis of blood-side urea samples and, in recent years, on-line urea monitors have become available to calculate Hemodialysis dose from dialysate-side urea kinetics. All these methods require waiting until the end of the session to calculate the Kt/V dose. In this work a Neural Network (NN) method  is presented for early prediction of the Kt/V dose. Three different portions of the dialysate urea concentration time profile (provided by an on-line urea monitor Biostat 1000-Baxter Healthcare Corp.) were analyzed:  100% (A), first 50%  (B) and last 50% (C) using a NN to predict the Kt/V and compare this to that provided by the Biostat. The NN was able to predict the Kt/V at the middle of the four hour session (B data) without a significant increase in the percentage error (B data => 6.69±2.46, A data => 5.58±8.77) compared to the Biostat Kt/V. The main difference in estimation between the NN models was found at extreme values of Kt/V which could be related to multicompartment behavior which  is not present at the final stage of Hemodialysis. Model C was found to be unsatisfactory.