Influence of chest wall conformation on reproducibility of main echocardiographic indices of left ventricular systolic function

BACKGROUND: The possible influence of chest wall conformation, as noninvasively assessed by Modified Haller Index (MHI, the ratio of chest transverse diameter over the distance between sternum and spine), on reproducibility of both left ventricular ejection fraction (LVEF) and global longitudinal strain (GLS) has never been previously investigated.
METHODS: Two equal groups of healthy individuals, matched by age, sex, and cardiovascular risk factors and categorized according to MHI in those with concave-shaped chest wall (MHI>2.5) and those with normal chest shape (MHI <= 2.5), who underwent transthoracic echocardiography implemented with echocardiographic deformation imaging between June 2018 and May 2019, were retrospectively analyzed. LVEF and GLS were measured twice by the two echocardiographers in a double blinded manner. Intra-class correlation coefficients (ICCs), bias and limits of agreement determined with Bland-Altman analysis were calculated for repeated measurements of both LVEF and GLS.
RESULTS: Thirty-four healthy individuals with MHI>2.5 (54.9 +/- 6.4 years, 58.8% females) and 34 matched controls with MHI <= 2.5 (52.5 +/- 8.1 years, 50% females) were separately analyzed. In comparison to MHI <= 2.5 group, the MHI>2.5 group was found with significantly smaller cardiac chambers and significantly lower GLS magnitude (-15.8 +/- 2.5 vs. -22.2 +/- 1.3%, P<0.001), despite similar LVEF (61.3 +/- 6.4 vs. 61.1 +/- 3.6%, P=0.87). In the MHI>2.5 group, intra-rater and inter-rater ICCs were <= 0.5 for both LVEF and LV-GLS, whereas in the MHI <= 2.5 group intra-rater and inter-rater ICCs values indicated good reliability for LVEF and excellent reliability for GLS. The greatest bias and largest limits of agreement were detected for LVEF assessment (bias ranging from -1.09 to 2.94%, with the 95% limits of agreement ranging from -13.9 to 21.3%) in individuals with MHI>2.5. On the other hand, the smallest bias and narrowest limits of agreement were obtained for GLS measurement (bias ranging from -0.26 to 0.09%, with the 95% limits of agreement ranging from -1.4 to 1.4%) in participants with normal chest wall conformation (MHI <= 2.5).
CONCLUSIONS: The test reliability of LVEF and GLS is strongly influenced by the chest wall conformation. MHI might represent an innovative approach for selecting the best echocardiographic method for LV systolic function estimation in the individual case.