Validation of estimates of total body water in pediatric dialysis patients by deuterium dilution.

BACKGROUND Current K-DOQI recommendations call for an assessment of dialysis adequacy that depends critically on an estimate of total body water (TBW). Such estimates are problematic in children since the range of patient size is large, and often formulas derived in normals are not validated in end-stage renal disease. Gold standard methods of TBW measurement, such as deuterium dilution ((2)H(2)O), are not appropriate in the clinical setting, yet noninvasive methods such as bioimpedance analysis (BIA) and dual energy x-ray absorptiometry (DEXA) have not been independently validated. METHODS We studied 14 stable pediatric dialysis patients on 1 to 3 occasions using (2)H(2)O dilution, BIA, DEXA, and anthropometry to measure TBW. We compared our data set to previously published formulae for TBW to determine root mean square error (RMSE) and skew of the estimate. RESULTS TBW prediction based upon the anthropometric formula proposed by the Pediatric Peritoneal Dialysis Consortium provided the best fit to our independent data set with RMSE = 2.15 L, and no skew by Bland-Altman analysis. Other formulas produced large, clinically relevant errors; obese subjects confounded many estimates. TBW calculated from hydrated lean body mass from DEXA scan was reliable with RMSE = 1.03 L and no skew. BIA-derived estimates can be useful, although the magnitude of RMSE ranged from 1.45 to 6.24 L, and one formula produced skewed results. CONCLUSION Techniques for estimating TBW in pediatric dialysis patients must be validated by independent data sets before being incorporated into clinical and research practice.