Comparing cardiac output measurements using electrical cardiometry versus phase contrast cardiac magnetic resonance imaging

Background: There is a growing interest in using non-invasive electrical cardiometry (EC) for continuous cardiac output (CO) measurements in neonates and children. There is a lack of sufficient data to validate CO measured by EC with currently accepted noninvasive techniques such as cardiac MRI (CMRI) in children. Objectives: To assess the degree of agreement between CO measured by EC and CMRI in children. We hypothesized that there is a good degree of agreement between CO measured by EC and CMRI in children. Methods: A retrospective chart review was performed to include patients who had noninvasive CO measured using EC (ICON (R)) and CMRI at our children's hospital between Feb 2019 and May 2020. Anthropometric, clinical and CO data were collected and CMRI parameters such as flow measurements in the aorta and pulmonary artery were analyzed. Data was analyzed using a Bland-Altman plot to compare the agreement between the CO obtained by EC vs. CMRI in this cohort of patients. Results: A total of 52 patients with cardiac disease who had data available from CMRI and ICON (R) were included. The median (IQR) age was 174 (99,201) months and male: female ratio was 1.2:1. A total of 34/52 (60 %) patients had congenital heart disease and 22 had history of surgical interventions prior to noninvasive CO measurements. The Bland-Altman analysis for CO values by both EC and CMRI produced a mean bias (SD) of 0.12 (2.73)1/min for ICON (R) and the 95 % limits of agreement were -5.229 to 5.473 1/min. The precision error was only 1.86 %. The Bland- Altman plot for reproducibility for CO values by EC produced a mean bias (SD) of 0.15 (0.41) 1/min with limits of agreement of -0.66 to 0.97 1/min. The percentage error was 2.34 %. Conclusions: In our cohort of cardiac patients, there was a good degree of agreement between noninvasive CO as measured by Electric Cardiometry and Cardiac MRI.