Variability of M-mode versus two-dimensional echocardiography measurements in children with dilated cardiomyopathy

M-mode and 2-dimensional (2D) echocardiographic imaging are routinely used to quantify left-ventricular (LV) size and function in pediatric patients with dilated cardiomyopathy (DCM). The reproducibility of and correlation between these techniques are unknown. This analysis sought to compare interreader, intrareader, and interacquisition reproducibility of M-mode versus 2D measurements in pediatric DCM patients. The Ventricular Volume Variability study of the Pediatric Heart Network is a multicenter, prospective, observational study assessing the course of chronic DCM in children. Two sonographers performed baseline image acquisitions locally, and two readers performed measurements at the echocardiographic core laboratory. One reader repeated measurements 1 month later. These data were used to assess reproducibility and agreement between M-mode and 2D measurements. One hundred sixty-nine subjects were enrolled. M-mode had similar or greater reproducibility in both intrareader and interreader settings for LV dimensions, shortening fraction (SF), and most wall thicknesses. In contrast, 2D reproducibility was similar or better for nearly all variables in the interacquisition setting but not for SF. Interacquisition variability was approximately twice the intrareader variability. LV dimensions by either modality consistently had high reproducibility and had the highest agreement between modalities. In pediatric DCM patients, variability of linear echocardiographic assessment could be minimized by relying on a single reader and using a consistent method (M-mode or 2D) for serial measurements, preferably M-mode when SF is the primary variable of interest. Except for LV dimensions, M-mode and 2D values should not be used interchangeably due to poor agreement.