Intra- and inter-observer reliability using a noninvasive ultrasound cardiac output monitor in healthy anesthetized children

Abstract

Background: Accurate and reliable evaluation of cardiac index (CI) in critically ill pediatric patients can optimize their management. Although validated, noninvasive ultrasound measurement techniques have been previously shown to be unreliable because of observer variability.

Objective: To confirm intra- and inter-observer reliability when using the noninvasive USCOM(®) in healthy anesthetized children.

Methods: Prospective observational study at the Children's Hospital of Eastern Ontario, Ottawa, included newborns to 12 years of age undergoing elective surgery or magnetic resonance imaging. The USCOM(®) was used to assess CI via aortic flow with a trans-sternal approach. Two trained observers were responsible for taking two measurements of CI each at steady state in randomized succession after stable depth of anesthesia was achieved.

Results: Fifty-nine patients were included. Forty-seven (80%) were between 3 and 7 years old, with 57% male. The mean difference ± sd for repeat CI measurements by each of two observers was 0.11 ± 0.47 and 0.05 ± 0.65 l·min(-1) ·m(-2) , respectively. Intra-observer reliability for these repeat measurements by each observer determined by Lin's concordance correlation coefficient was 0.92 and 0.85, respectively. The mean difference ± sd between observers was 0.16 ± 0.59 l·min(-1) ·m(-2) , and Lin's concordance correlation coefficient was 0.87. The two observers subjectively rated measurements as 'Difficult' or 'Very difficult' only 14% (16/118) and 3% (4/118) of the time, respectively. No adverse events were reported.

Conclusion: This study confirms that the USCOM(®) is relatively easy to use and reliable in healthy children when operated by trained users.