Background. The modeled volume of urea distribution (Vm) in intermittently hemodialyzed patients is often compared with total body water (TBW) volume predicted from population studies of patient anthropometrics (Vant). Methods. Using data from the HEMO Study, we compared Vm determined by both blood-side and dialysate-side urea kinetic models with Vant as calculated by the Watson, Hume-Weyers, and Chertow anthropometric equations. Results. Median levels of dialysate-based Vm and blood-based Vm agreed (43% and 44% of body weight, respectively). These volumes were lower than anthropometric estimates of TBW, which had median values of 52% to 55% of body weight for the three formulas evaluated. The difference between the Watson equation for TBW and modeled urea volume was greater in Caucasians (19%) than in African Americans (13%). Correlations between Vm and Vant determined by each of the three anthropometric estimation equations were similar; but Vant derived from the Watson formula had a slightly higher correlation with Vm. The difference between Vm and the anthropometric formulas was greatest with the Chertow equation, less with the Hume-Weyers formula, and least with the Watson estimate. The age term in the Watson equation for men that adjusts Vant downward with increasing age reduced an age effect on the difference between Vant and Vm in men. Conclusion. The findings show that kinetically derived values for V from blood-side and dialysate-side modeling are similar, and that these modeled urea volumes are lower by a substantial amount than anthropometric estimates of TBW. The higher values for anthropometry-derived TBW in hemodialyzed patients could be due to measurement errors. However, the possibility exists that TBW space is contracted in patients with end-stage renal disease (ESRD) or that the TBW space and the urea distribution space are not identical.
