Two-sample hemodialysis urea kinetic modeling: validation of the method

Urea kinetic modeling (UKM) by a two-sample method (UKM2) was compared with the classical three-sample method (UKM3) and direct quantification. Assuming the patient to be in a weekly steady state and using an approximate treatment schedule, UKM2 can compute urea generation rate (G), distribution volume (V), Kt/V, and normalized protein catabolic rate (NPCR). Twenty-three stable patients were examined. The results obtained by UKM2 and UKM3 differed little (G -4.2%, V -1.0%, Kt/V 0.9%, NPCR -2.7%), and the correlations between them were high (r > or = 0.96). The differences between UKM2 and direct quantification were greater, but also highly correlated. G determined midweekly by UKM2 was highly correlated with G found directly from 1-week collection of dialysate and urine (r = 0.96). Repeating analysis over a 5-week period, the results obtained by UKM2 varied no more than those obtained by UKM3 (around 8% for all four kinetic variables). In conclusion, UKM2 produces reliable results requiring less data to be entered than using UKM3.