Background: Knowledge of urea volume of distribution (Vurea) in patients with acute renal failure (ARF) is critical in order to prescribe and monitor appropriate dialytic treatment. We have recently shown that in ARF patients, Vurea estimation by urea kinetic modeling is significantly higher than total body water (TBW) by anthropometric estimation. However, these estimates of Vurea and TBW have not been validated by isotopic methods, considered as reference measurement standards.
Methods: In this study, we measured Vurea by [13C]urea and TBW by deuterium oxide (D2O) in 21 patients with ARF (14 males, 7 females, age 62.0 +/- 10.6 years old, 83% Caucasian, 17% African American) at three different centers. These measurements were compared to TBW estimates from anthropometric and bioelectrical impedance (BIA) measurements.
Results: Our results show that Vurea by [13C]urea (51.0 +/- 11.7 L) is significantly higher than TBW estimated by all other methods (TBW by D2O: 38.3 +/- 9.8 L, P < 0.001; TBW by BIA: 45.7 +/- 15.7 L, P= 0.08; TBW by Watson formula: 38.3 +/- 7.3 L, P < 0.001; TBW by Chertow formula: 39.3 +/- 7.8 L, P= 0.002, all versus Vurea). Despite significant overestimation of the absolute value and considerable variation, Vurea significantly correlated with TBW by BIA (r= 0.66, P < 0.01) and TBW by D2O (r= 0.5, P= 0.04). There was also significant correlation between D2O and BIA determined TBW (r= 0.8, P < 0.001).
Conclusion: In terms of useful guidelines to prescribe a specific dose of dialysis in patients with ARF, conventional estimates of TBW as surrogates for Vurea should be used with caution. We propose that these conventional estimates of TBW should be increased by approximately 20% (a factor of 1.2) to avoid significant underdialysis.