Comparison of three non-invasive hemodynamic monitoring methods in critically ill children

Abstract

Introduction: Hemodynamic parameters measurements were widely conducted using pulmonary artery catheter (PAC) with thermodilution as a reference standard. Due to its technical difficulties in children, transthoracic echocardiography (TTE) has been widely employed instead. Nonetheless, TTE requires expertise and is time-consuming. Noninvasive cardiac output monitoring such as ultrasonic cardiac output monitor (USCOM) and electrical velocimetry (EV) can be performed rapidly with less expertise requirement. Presently, there are inconsistent evidences, variable precision, and reproducibility of EV, USCOM and TTE measurements. Our objective was to compare USCOM, EV and TTE in hemodynamic measurements in critically ill children.

Materials and methods: This was a single center, prospective observational study in critically ill children. Children with congenital heart diseases and unstable hemodynamics were excluded. Simultaneous measurements of hemodynamic parameters were conducted using USCOM, EV, and TTE. Inter-rater reliability was determined. Bland-Altman plots were used to analyse agreement of assessed parameters.

Results: Analysis was performed in 121 patients with mean age of 4.9 years old and 56.2% of male population. Interrater reliability showed acceptable agreement in all measured parameters (stroke volume (SV), cardiac output (CO), velocity time integral (VTI), inotropy (INO), flow time corrected (FTC), aortic valve diameter (AV), systemic vascular resistance (SVR), and stroke volume variation (SVV); (Cronbach's alpha 0.76-0.98). Percentages of error in all parameters were acceptable by Bland-Altman analysis (9.2-28.8%) except SVR (30.8%) and SVV (257.1%).

Conclusion: Three noninvasive methods might be used interchangeably in pediatric critical care settings with stable hemodynamics. Interpretation of SVV and SVR measurements must be done with prudence.

Fig 1. Flow diagram showing the screening and enrollment of patients.

Fig 2

Bland-Altman analysis of overall results: Cardiac output (A), cardiac index (B). X-axis illustrated average measurements of standard from using TTE. Y-axis demonstrated mean bias using other modalities.

Fig 3

Bland-Altman analysis of preload: Stroke volume (A), Corrected flow time (B) and Stroke volume variation (C). X-axis illustrated average measurements of standard from using TTE or ICON. Y-axis demonstrated mean bias using other modalities.

Fig 4

Bland-Altman analysis of afterload: Systemic vascular resistance (A) Systemic vascular resistance index (B). X-axis illustrated average measurements of standard from using ICON. Y-axis demonstrated mean bias using USCOM.