Noninvasive cardiac output monitoring using bioreactance-based technique in pediatric patients with or without ventricular septal defect during anesthesia: in comparison with echocardiography

Background: We evaluated the use of bioreactance-based noninvasive cardiac output (CO) monitoring technique (NICOM(™), CO(NICOM)) in pediatric patients with or without ventricular septal defect (VSD) during anesthesia induction to determine its agreement with the measurements assessed by echocardiography (echo, CO(ECHO)).

Methods: Twenty-eight pediatric patients with normal heart anatomy (group NHA) and 32 with isolated ventricular septal defects (group VSD) were included in this study. The cardiac output was measured simultaneously in minute-by-minute using NICOM and echo (Simpson's rule) during anesthesia induction and intubation. Linear regression and revised Bland-Altman analyses were performed to evaluate the agreement by comparing the paired CO results. The mean percent error ((CO(ECHO)-CO(NICOM))/CO(ECHO) × 100%) was used to assess the impact of congenital heart disease on the agreement.

Results: The measurements of CO by NICOM and echo techniques were highly correlated in group NHA (γ = 0.96, P < 0.005) and VSD (γ = 0.84, P < 0.005). The mean bias (CO(ECHO) - CO(NICOM)) between the two methods was 0.03 and 0.31 l·min(-1) with the limits of agreement (LOA) -0.29 to +0.35 l·min(-1) and -0.44 to +1.05 l·min(-1), which include 96.9% (31/32) and 89.3% (25/28) of all patients' different data in group NHA and VSD, respectively. The median percent errors were significantly lower at all time points in group NHA than those in group VSD (all P < 0.05).

Conclusion: In children without heart defects, the CO measured by NICOM shows a good agreement with the echo during anesthesia induction. The NICOM technique underestimates echo although a strong correlation exists between two methods in children with ventricular septal defect.