Estimation of total body water from foot-to-foot bioelectrical impedance analysis in patients with cancer cachexia - agreement between three prediction methods and deuterium oxide dilution

Introduction Bioelectrical impedance analysis (BIA) is a useful bedside measure to estimate total body water (TBW). The aim of this study was to determine the agreement between three equations for the prediction of TBW using BIA against the criterion method, deuterium oxide dilution, in patients with cancer cachexia. Methods Eighteen measurements of TBW using foot-to-foot BIA in seven outpatients with cancer cachexia (five male and two female, age 56.4 +/- 6.7 years) at an Australian hospital. Three prediction formulae were used to estimate TBW - TBWca-radiotherapy developed in patients with cancer undergoing radiotherapy, TBWca-underweight and TBWca-normal weight developed in underweight and normal weight patients with cachexia. TBW was measured using the deuterium oxide dilution technique as the gold standard. Results Mean measured TBW was 39.5 +/- 6.0 L. There was no significant difference in measured TBW and estimates from prediction equations TBWca-underweight and TBWca-radiotherapy. There was a significant difference in measured TBW and TBWca-normal weight. All prediction equations overestimated TBW in comparison with measured TBW. The smallest bias was observed with TBWca-underweight (0.38 L). The limits of agreement are wide (> 7.4 L) for each of the prediction equations compared with measured TBW. Conclusions At a group level, TBWca-underweight is the best predictor of measured TBW in patients with cancer cachexia. For an individual however, the limits of agreement are wide for all prediction equations and are unsuitable for use. Practitioners need to be aware of the limitations of using TBW prediction equations for individuals.