Noninvasive pulse contour analysis for determination of cardiac output in patients with chronic heart failure

Background Determination of cardiac output (CO) is essential in diagnosis and management of heart failure (HF). The gold standard to obtain CO is invasive assessment via thermodilution (TD). Noninvasive pulse contour analysis (NPCA) is supposed as a new method of CO determination. However, a validation of this method in HF is pending and performed in the present study. Methods Patients with chronic-stable HF and reduced left ventricular ejection fraction (LVEF = 45%; HF-REF) underwent right heart catheterization including TD. NPCA using the CNAP Monitor (V5.2.14, CNSystems Medizintechnik AG) was performed simultaneously. Three standardized TD measurements were compared with simultaneous auto-calibrated NPCA CO measurements. Results In total, 84 consecutive HF-REF patients were enrolled prospectively in this study. In 4 patients (5%), TD was not successful and for 22 patients (26%, 18 with left ventricular assist device), no NPCA signal could be obtained. For the remaining 58 patients, Bland-Altman analysis revealed a mean bias of + 1.92 L/min (limits of agreement +/- 2.28 L/min, percentage error 47.4%) for CO. With decreasing cardiac index, as determined by the gold standard of TD, there was an increasing gap between CO values obtained by TD and NPCA (r = -0.75, p < 0.001), resulting in a systematic overestimation of CO in more severe HF. TD-CI classified 52 (90%) patients to have a reduced CI (< 2.5 L/min/m(2)), while NPCA documented a reduced CI in 18 patients (31%) only. Conclusions In HF-REF patients, auto-calibrated NPCA systematically overestimates CO with decrease in cardiac function. Therefore, to date, NPCA cannot be recommended in this cohort.