Equivalence of the bioimpedance and thermodilution methods in measuring cardiac output in hospitalized patients with advanced, decompensated chronic heart failure.

BACKGROUND An accurate and reliable noninvasive method for determining cardiac output/cardiac index would be valuable for patients with acutely decompensated advanced systolic heart failure. OBJECTIVES To determine whether a correlation exists for cardiac output and index determined by using bioimpedance and thermodilution in patients with acutely decompensated complex heart failure and if differences between results with the 2 methods could be explained by the patients' advanced condition. METHODS Cardiac output and index were determined by using bioimpedance and thermodilution in 33 patients. Echocardiographic and electrocardiographic data were assessed to determine if differences between results with the 2 methods could be explained by the patients' advanced condition. Concordance correlation coefficients and Bland-Altman agreement between methods were calculated. RESULTS Four patients were excluded from analysis because reliable measurements could not be obtained; the remaining 29 patients constituted the study population. Mean cardiac outputs determined by thermodilution and bioimpedance were 5.48 and 5.40 L/min, respectively (rhoc = 0.89, P < .001), and mean cardiac indexes were 2.67 and 2.65 (rhoc = 0.82, P < .001). Mean bias (limits of agreement) between data pairs was 0.08 (-0.18 to 0.35) L/min (P = .52) for cardiac output and 0.03 (-0.097 to 0.16; P = .61) for cardiac index. Six data pairs (21%) had an absolute percent difference greater than 15%. Of these, 50% had a higher thermodilution value. CONCLUSION Determinations of cardiac output and index by both methods were significantly correlated. Mean bias between the 2 methods was small, suggesting clinical utility for bioimpedance in patients with complex decompensated heart failure.