Cardiac Output Measurement in Neonates and Children Using Noninvasive Electrical Bioimpedance Compared With Standard Methods: A Systematic Review and Meta-Analysis*

Objective: To systematically review and meta-analyze the validity of electrical bioimpedance-based noninvasive cardiac output monitoring in pediatrics compared with standard methods such as thermodilution and echocardiography. Data Sources: Systematic searches were conducted in MEDLINE and EMBASE (2000–2019). Study Selection: Method-comparison studies of transthoracic electrical velocimetry or whole body electrical bioimpedance versus standard cardiac output monitoring methods in children (0–18 yr old) were included. Data Extraction: Two reviewers independently performed study selection, data extraction, and risk of bias assessment. Mean differences of cardiac output, stroke volume, or cardiac index measurements were pooled using a random-effects model (R Core Team, R Foundation for Statistical Computing, Vienna, Austria, 2019). Bland-Altman statistics assessing agreement between devices and author conclusions about inferiority/noninferiority were extracted. Data Synthesis: Twenty-nine of 649 identified studies were included in the qualitative analysis, and 25 studies in the meta-analyses. No significant difference was found between means of cardiac output, stroke volume, and cardiac index measurements, except in exclusively neonatal/infant studies reporting stroke volume (mean difference, 1.00 mL; 95% CI, 0.23–1.77). Median percentage error in child/adolescent studies approached acceptability (percentage error less than or equal to 30%) for cardiac output in L/min (31%; range, 13–158%) and stroke volume in mL (26%; range, 14–27%), but not in neonatal/infant studies (45%; range, 29–53% and 45%; range, 28–70%, respectively). Twenty of 29 studies concluded that transthoracic electrical velocimetry/whole body electrical bioimpedance was noninferior. Transthoracic electrical velocimetry was considered inferior in six of nine studies with heterogeneous congenital heart disease populations. Conclusions: The meta-analyses demonstrated no significant difference between means of compared devices (except in neonatal stroke volume studies). The wide range of percentage error reported may be due to heterogeneity of study designs, devices, and populations included. Transthoracic electrical velocimetry/whole body electrical bioimpedance may be acceptable for use in child/adolescent populations, but validity in neonates and congenital heart disease patients remains uncertain. Larger studies in specific clinical contexts with standardized methodologies are required.