Cardiac Output Monitoring in Children, Adolescents and Adults Based on Pulse Contour Analysis: Comparison with Echocardiography-Derived Data and Identification of Factors Associated with Their Differences

Abstract

Through pulse contour analysis (PCA) devices would enable to obtain non-invasive operator-independent cardiac output (CO) measurements [CO(PCA)]. The agreement between CO(PCA) and data from two-dimensional [CO(2D)] or Doppler [CO(VTI)] echocardiography (references) remains controversial. To analyze: (1) CO(PCA), CO(2D) and CO(VTI) agreement, (2) determinants of methods' differences in measured CO values. Simultaneous echocardiography and PCA records (Mobil-O-Graph/Germany) were obtained in 130 subjects (age: 29 ± 17years). Hemodynamic (e.g., heart rate [HR]), arterial (e.g., arterial stiffness, augmentation index [AIx]) and cardiac structural-functional (e.g., left ventricle end-diastolic diameters [LVEDD]) parameters were obtained. Data from the entire group (all; 10-85years), children (≤ 16years), adolescents (17-24years) and adults (> 24years) were separately analyzed. The highest Lin's concordance correlation coefficient (CCC) were obtained when analyzing CO(PCA)/CO(2D) association (0.672, 0.785, 0.721, 0.487 for all, children, adolescents and adults, respectively); CCC levels were higher at younger ages. Bland-Altman's systematic errors between CO(PCA)/CO(2D) were 0.12, 0.17, 0.07 and 0.14L/min, for all, children, adolescents and adults, respectively (non-significant). CO(VTI)/CO(PCA) systematic error only reached significance in adults (0.34L/min, p = 0.002). Bland-Altman's proportional errors were not statistically significant when CO(PCA)/CO(2D) differences were analysed in children and adolescents. Higher AIx and LVEDD levels associated greater CO(2D)/CO(PCA) differences; higher AIx and HR levels were associated to differences between CO(VTI) and CO(PCA). CO(PCA) had systematic (< 0.17L/min) and percent (≤ 30%) errors, which allow us to postulate that its use allows reaching levels comparable to those of echocardiography. Differences in CO-data between methods were associated with arterial and cardiac properties.