Background: Bioelectrical impedance spectroscopy (BIS) may be more accurate in determining total body water (TBW) than bioelectrical impedance analysis (BIA). The present study compared the agreement between three TBW prediction equations developed using BIA and BIS-derived TBW in oncology outpatients.
Methods: A cross-sectional, observational study was conducted in 37 outpatients receiving radiotherapy (27 males/10 females, aged 68.3 +/- 10.2 years). TBW was estimated by BIS (TBW(BIS)) and three BIA TBW prediction equations (TBW(ca-u): underweight cancer patients; TBW(ca-n): normal-weight cancer patients; and TBW(rad): patients receiving radiotherapy). Bland-Altman analyses determined agreement between methods. BIS-derived TBW using new resistivity constants was calculated.
Results: The mean +/- SD of TBW estimated by BIS was 39.8 +/- 8.3 L, which was significantly different from the prediction equations; TBW(rad) 35.1 +/- 7.9 L, TBW(ca-u) 33.1 +/- 7.5 L and TBW(ca-n) 32.3 +/- 7.3 L, (P < 0.001). Using new resistivity constants, TBW was 36.2 +/- 8.1 L but this still differed from the equations (P < 0.001). Bias between TBW(BIS) and that predicted by the equations was in the range 4.7-7.4 L or 1.1-3.9 L using new resistivity constants.
Conclusions: TBW estimated by BIS cannot be directly compared with oncology-specific BIA equations, suggesting that BIS cannot be used at the group level in outpatients receiving radiotherapy. There was a reduced bias with BIS using new resistivity constants; however, further research should determine any advantage of BIS over BIA in this population.