074 - Reliability of two-dimensional mode versus M-mode echocardiography for measurement of left atrium/aortic diameter ratio and fractional shortening in extreme preterm infants

Publication Number: 74.6311

Uthaya Kumaran Kanagaraj, University of British Columbia Faculty of Medicine, Vancouver, BC, Canada; Michael P.. Castaldo, University of British Columbia Faculty of Medicine, Vancouver, BC, Canada; Souvik Mitra, University of British Columbia Faculty of Medicine, Vancouver, BC, Canada

Background: In extremely preterm neonates, left atrium to aortic root ratio (LA:Ao) and fractional shortening (FS) are widely measured by motion mode (M-mode) for assessment of left heart volume overload in patent ductus arteriosus (PDA) and left ventricle (LV) function respectively as per most targeted neonatal echocardiography (TNE) protocols. This however contradicts the American Society of Echocardiography (ASE) guidelines on pediatric echocardiography that recommend measurement of FS by two-dimensional (2D) mode. Further, the ASE guidelines do not provide any guidance on choice of modality for LA:Ao assessment. There is a lack of studies comparing M-mode with two-dimensional (2D) mode for measurement of LA: Ao and FS, specifically in preterm infants.

Objective: To determine the inter observer reliability between 2D vs M-mode echocardiography for (a) LA:Ao ratio and (b) FS assessment in extremely preterm infants.

Design/Methods: We conducted a retrospective observational study between Oct 2020 and Sep 2023 involving infants born < 29 weeks of gestational age (GA) who underwent TNE during their NICU stay. The primary outcomes were the measurement of the left atrium to aortic root ratio (LA:Ao) and fractional shortening (FS). We assessed interobserver reliability using the intraclass correlation coefficient (ICC) (calculated using average of observers, absolute agreement, two-way random-effects models) for each measurement and mode (2D and M-mode). With three interpreters and assuming an ICC of ≥ 0.80, and an 80% probability of obtaining a 95% CI with width of 0.15, we calculated a sample size of 79.

Results: A total of 79 scans [N=31 infants, median (IQR) GA at birth 25.7 (24.8, 26.9) was assessed independently by 3 observers (2 TNE-trained neonatologists, 1 echo technician). Median (IQR) day of life at scan was 25.0 days (12.0, 57.0). PDA assessment was the primary indication for scanning in 48 scans (60.8%). The ICC for LV FS was 0.72 (95% CI: 0.40–0.85) for M-mode and 0.73 (95% CI: 0.56–0.83) for 2D-mode. The ICC for LA/Ao ratio was 0.87 for both M-mode (95% CI: 0.73–0.93) and 2D-mode (95% CI: 0.80–0.92).

Conclusion(s): Both 2D and M-modes are reliable for measuring the LA/Ao ratio in extremely preterm infants, while neither appear to provide sufficient reliability for LV FS measurements. Therefore, TNE protocols may include LA:Ao measurements either in M-mode or 2D-mode as a marker of pulmonary overcirculation in PDA assessment. FS as a marker for LV systolic function should be interpreted with caution.

Histogram illustrating LV FS measurements across three interpreters for both M-mode and 2D-mode
X-axis – LV FS (%), Y-axis – number of measurements

Histogram illustrating LA/Ao ratio measurements across three interpreters for both M-mode and 2D-mode
X-axis – LA/Ao ratio, Y-axis – number of measurements

Table displaying ICC values and 95% confidence intervals for LV FS and LA/Ao ratio measurements using M-mode and 2D-mode