Reliability of bioelectrical impedance methods in detecting body fluids in elderly patients with congestive heart failure

OBJECTIVE: To investigate the reliability of bioelectrical impedance analysis (BIA) in estimating total body water (TBW) and extracellular water (ECW) in elderly patients suffering from congestive heart failure (CHF). MATERIAL AND METHODS: In 72 elderly subjects, 34 with CHF (aged 83.9+/-6.9 years) and 38 healthy controls (78.7+/-7.5 years), TBW and ECW values were assessed using dilution methods, and bioelectrical variables were measured using single frequency BIA (SF-BIA) at 1 and 50 kHz, and bioelectrical spectroscopy (BIS). RESULTS: In CHF patients, Ht(2)/R(1) correlated weakly with TBW (r = 0.56) and ECW (0.47). In both healthy controls and CHF patients, TBW correlated strongly with Ht(2)/R(50), Ht(2)/R(0), Ht(2)/R(8) and Ht(2)/Zc. Using multiple regression analysis and the Bland-Altmann approach, SF-BIA at 50 kHz and BIS proved similar in predicting TBW for both the explained variance (R(2)~0.89) and the limits of agreement. In all subjects, ECW was estimated best by including height, weight and Ht(2)/R(0 )(R(2) 0.75) or Ht(2)/Zc (R(2) 0.77) in multivariate models, while SF-BIA at 50 kHz did not explain more than 71 % of ECW variability. The SEE % was nonetheless about twice the SEE % for estimating TBW. CONCLUSIONS: SF-BIA at 1 kHz is unreliable in predicting body fluids in elderly people with CHF. SF-BIA at 50 kHz and BIS are useful for estimating TBW in healthy elderly people and in cases of water imbalance, but both methods are less reliable in estimating ECW, particularly in conditions of fluid overload.