Fluid management is a crucial issue in intensive-care medicine. This study evaluated the feasibility and reproducibility of bioimpedance spectroscopy to measure body-water composition in critically ill patients, and compared fluid balance and daily changes in total body water (TBW) measured by bioimpedance. This observational study included 25 patients under mechanical ventilation. Fluid balance and bioimpedance measurements were recorded on 3 consecutive days. Whole-body bioimpedance spectroscopy was performed with exact or ideal body weights entered into the device, and with or without ICU monitoring. Reproducibility of bioimpedance spectroscopy was very good in all conditions despite ICU monitoring and mechanical ventilation. Bioimpedance measurements using an ideal body weight varied significantly, making the weighing procedure necessary. Comparison of fluid balance and daily changes in body weight provided the best correlation (ρ = 0.74; P < 0.0001). Daily changes in TBW were correlated with fluid balance (Spearman coefficient ρ = 0.31; P = 0.003) and this correlation was improved after exclusion of patients with a SOFA score >10 (ρ = 0.36; P = 0.05) and with extracorporeal circulation (ρ = 0.50; P = 0.005). Regardless of the technique used to estimate volume status, important limits of agreement were observed. Non-invasive determination of body-water composition using bioimpedance spectroscopy is feasible in critically ill patients but requires knowledge of the patient's weight. The best method to assess volume status after fluid resuscitation and the value gained from information about body composition provided by bioimpedance techniques needs further evaluation.
Keywords: Bioelectric impedance; Body weight; Critical care; Water-electrolyte balance.