Background: We have reported catch-up growth with hemodialysis (HD) of approximately 15 hours/week. Without an equilibrated post-treatment blood urea nitrogen, the variable-volume single-pool (VVSP) model will not account for urea rebound, inflating the estimated HD dose (K(d)t/V). A two-pool model (FVDP) predicts rebound, but requires fixed compartment volumes for the equations to be solvable in closed form, also inflating K(d)t/V.
Methods: We developed an approximate perturbation solution (WKB method) to a variable volume, two-pool (VVDP) model. Estimated model parameters were compared with the results of equilibrated kinetic studies using measured clearance K(d) (N = 17). Once the model was validated, we re-analyzed 292 kinetic studies from our earlier cohort, which was considered well-dialyzed on the basis of growth rates (N = 12, mean annual change in height standard deviation score +0.31, mean follow-up of 26 months).
Results: For the VVSP, FVDP, and VVDP models, respectively, the mean errors were (1) K(d)t/V, 0.22 +/- 0.07, 0.29 +/- 0.17, 0.06 +/- 0.07 (ANOVA, P < 0.001); (2) urea distribution volume vol/wt (%), -8.2 +/- 4.2, -9.1 +/- 3.0, -2.2 +/- 3.6 (P < 0.001). Sequential studies confirmed reproducibility, with a coefficient of variation < or = 5%. In the earlier cohort, a comparison of the VVSP and VVDP models yielded the following: (1) K(d)t/V, 1.91 +/- 0.35 vs. 1.76 +/- 0.33 (P < 0.001); (2) normalized protein catabolic rate (nPCR, g/kg/day), 1.56 +/- 0.39 vs. 1.52 +/- 0.38 (P < 0.001); and (3) K(d) (whole blood, mL/kg/min), 4.8 +/- 0.9 vs. 4.4 +/- 0.8 (P < 0.001).
Conclusion: This VVDP model yields reliable estimates of K(d)t/V and other kinetic parameters using standard blood urea nitrogen sampling. Analysis of patients previously characterized as well-dialyzed on the basis of growth rates clarifies the HD dose needed to sustain normal growth.