Answers are grounded in this paper and cite their sources.
Fluid accumulation is a key indicator of acute decompensation in heart failure, yet body weight reflects fluid changes only indirectly. Bioelectrical impedance analysis (BIA) offers a non-invasive alternative but is difficult to interpret in isolation. This feasibility study investigated whether the Cole-model parameters R₀ and R∞, the resistances at zero and infinite frequency, can be predicted from physiological and anthropometric parameters and used to estimate extracellular, intracellular, and total body water. Multi-frequency bioimpedance spectra were acquired with an AD5940-based system and fitted to the Cole model using dynamically weighted least squares. Electrocardiogram-derived heart rate, heart rate variability, and respiration, together with oxygen saturation, blood pressure, and anthropometric data, were recorded in 21 healthy volunteers. Features were selected by least absolute shrinkage and selection operator (LASSO) regression or reduced by principal component analysis (PCA), and linear regression, random forest, and gradient boosting models were cross-validated. LASSO retained only body weight, sex, and systolic blood pressure. The random forest achieved the lowest prediction errors, and R₀ was predicted more accurately than R∞, indicating a more robust estimation of extracellular water. These results support the feasibility of signal fusion for fluid-status monitoring in heart failure; validation in larger clinical cohorts is required.
No takes yet. Share an insight, caveat, or question.
Lazik et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: