This study compared the measurements of total body water (TBW) by 58% body weight (TBW58%), the Watson equation (TBWWV) and bioelectric impedance (TBWBIA) with the gold standard, Deuterium oxide (TBWD2O) dilution method in twenty continuous ambulatory peritoneal dialysis (CAPD) patients. TBW volumes were highest when calculated as TBW58% (42.6 +/- 9.4 liter) and lowest when calculated from TBWWV (34.6 +/- 6.8 liter). TBWBIA underestimated TBW when compared to TBWD2O, although the difference was not statistically significant (37.1 +/- 9.8 liter and 38.8 +/- 9.3 liter, respectively). In fact, TBWBIA correlated strongly with TBWD2O (r = 0.8, P < 0.0001). These discrepancies resulted in significant differences when Kt/V week(-1) derived from the four methods were compared. To determine the effect of percent fat mass on the estimation of TBW by each method, we compared TBW and Kt/V week(-1) derived from the four methods in nine CAPD patients who had normal percent fat mass (Non-Obese) and 11 CAPD patients who had greater than normal % fat mass (Obese). In the Non-Obese group, there was close correlation of TBWBIA, TBWWV and TBW58% when compared with TBWD2O (r = 0.93, P < 0.001, r = 0.89, P < 0.01 and R = 0.86, P < 0.01, respectively). Also, Kt/V week(-1) derived from TBWBIA, TBWWV and TBW58% correlated strongly with Kt/V week(-1) from TBWD2O (r = 0.93, P < 0.0005, r = 0.83, P < 0.01 and r = 0.8, P < 0.01, respectively). In the Obese patients, the correlations between TBWD2O and TBWBIA and TBWWV were only moderate in strength (r = 0.67, P = 0.02 and r = 0.61, P = 0.04, respectively). There was no significant correlation between TBW58% and TBWD2O in the Obese group (r = 0.44, P = NS). Kt/V week(-1) derived from TBWBIA; TBWWV and TBW58% had only moderate correlation with Kt/V week(-1) from D2O. Similar findings in TBW measurements and Kt/V week(-1) were obtained in CAPD patients (N = 9) who had normal BMI (< 26 kg/m(2); termed, Normal Weight) and those who had BMI > 26 kg/m(2) (N = 11; termed, Overweight; data not shown). Our results suggest that the discrepancies in TBW and Kt/V in CAPD patients with different body habitus were a result of relative body fat mass. Thus, in CAPD patients who were not obese, all four methods gave good correlations of TBW and Kt/V. Even though TBWBIA had the smallest bias and the best correlation with TBWD2O, the wide limits of agreement between BIA and D2O dilution suggest that BIA is inappropriate at the present time for assessing TBW in CAPD patients, especially in the obese. In the case of discrepancy in Kt/V and clinical outcome, adequacy of dialysis in CAPD patients should be assessed by accurate measurement of V using the D2O dilution method.
