Errors in estimating peritoneal fluid by bioelectrical impedance analysis and total body electrical conductivity.

Whole-body bioelectrical impedance analysis (BIA) and total body electrical conductivity (TOBEC) have been used to estimate body composition and generalized changes in total body water (TBW). The sensitivity of these methods to measure small, rapid, localized changes in body water has not been fully evaluated. We compared the prediction of TBW by whole-body and segmental BIA and TOBEC with deuterium oxide dilution (D2O) in 10 control subjects and 7 renal failure patients receiving continuous ambulatory peritoneal dialysis (CAPD) prior to and after dialysate infusion. Using D2O as the reference method, there was no significant mean residual error between TBW predicted by BIA and TOBEC in controls (-1.2 +/- 1.5 and -0.9 +/- 1.0 kg) and CAPD patients pre-infusion (-1.0 +/- 2.0 and 0.29 +/- 2.01 kg). After infusing 1.9 +/- 0.18 kg dialysate, the mean residual error between change in body weight and the three methods was -0.44 +/- 0.53 kg for D2O (p < 0.1), -1.7 +/- 0.25 kg for BIA (p < 0.0001), and 1.2 +/- 0.4 kg for TOBEC (p < 0.001). Segmental BIA detected a 7.6% reduction in trunkal resistance with no significant change across the limbs, consistent with abdominal fluid accumulation. It is concluded that whole-body BIA underpredicts and TOBEC overpredicts small changes in peritoneal fluids.