Bioimpedance based determination of cardiac index does not show enough trueness for point of care use in patients with systolic heart failure

Cardiac output (CO) is a key parameter in diagnostics and therapy of heart failure (HF). The thermodilution method (TD) as gold standard for CO determination is an invasive procedure with corresponding risks. As an alternative, thoracic bioimpedance (TBI) has gained popularity for CO estimation as it is non-invasive. However, systolic heart failure (HF) itself might worsen its validity. The present study validated TBI against TD. In patients with and without systolic HF (LVEF ≤ 50% or > 50% and NT-pro-BNP < 125 pg/ml, respectively) right heart catheterization including TD was performed. TBI (Task Force Monitor©, CNSystems, Graz, Austria) was conducted semi-simultaneously. 14 patients with and 17 patients without systolic HF were prospectively enrolled in this study. In all participants, TBI was obtainable. Bland-Altman analysis indicated a mean bias of 0.3 L/min (limits of agreement ± 2.0 L/min, percentage error or PE 43.3%) for CO and a bias of -7.3 ml (limits of agreement ± 34 ml) for cardiac stroke volume (SV). PE was markedly higher in patients with compared to patients without systolic HF (54% vs. 35% for CO). Underlying systolic HF substantially decreases the validity of TBI for estimation of CO and SV. In patients with systolic HF, TBI clearly lacks diagnostic accuracy and cannot be recommended for point-of-care decision making. Depending on the definition of an acceptable PE, TBI may be considered sufficient when systolic HF is absent. Trial registration number: DRKS00018964 (German Clinical Trial Register, retrospectively registered)