Quantitating hemodialysis: a comparison of three kinetic models.

Three urea kinetic analyses were applied to hemodialysis and their conformity assessed. Sixteen patients underwent 50 measurements of dialyzer clearance (K), protein catabolic rate (PCR), and dialysis quantification (Kt/V) using the urea kinetic model (UKM) of Gotch and Sargent, Malchesky's direct dialysis quantification (DDQ), and the graphic technique of urea reduction analysis (URA) devised by Keshaviah. Additionally, the equations proposed by Jindal (percent urea reduction), and Daugirdas were used to calculate Kt/V values for each study. Dialyzer performance determined by whole blood urea clearance consistently exceeded simultaneous dialysate urea removal and was 55% greater than the clearance calculated by DDQ. Despite these variations, dialysis adequacy (Kt/V) and normalized protein catabolic rate (nPCR) were remarkably constant when derived by fixed-volume single-pool analyses (ie, UKM, DDQ, and URA). Application of variable-volume corrections increased Kt/V and nPCR, but caused DDQ values to diverge from those of UKM and URA. During rapid high-efficiency dialysis (RHED), the UKM predicted urea removal in excess of that documented by DDQ. During this trial (low-level RHED with K = 2.98 mL.kg-1 per min), urea dysequilibrium across blood-cell interfaces was sufficient to cause UKM to overestimate protein catabolism by 5%. The basic assumption of single-pool kinetics may be inappropriate during RHED, and further increases in dialyzer clearance will increase the discrepancy between projected and actual urea removal. Future comparisons of RHED prescriptions should employ mass balance data, or redesigned kinetic analyses.