Echocardiographic Characterization of Left Ventricular Radial and Circumferential Wall Motion in Horses Using Strain, Strain Rate, and Displacement by 2D Speckle Tracking: Methodology and Reliability
Echocardiographic assessment of left ventricular (LV) function in horses is currently limited to subjective evaluation and calculation of two-dimensional ejection phase indices. The goal of this study was to show the feasibility, describe the techniques, and determine the reliability of 2D speckle tracking (2DST) for characterization of LV radial and circumferential wall motion that could be useful for assessment of regional and global LV systolic and diastolic function in horses.
Six healthy, adult horses were included. Repeated echocardiographic examinations were performed by two independent observers in unsedated horses, using a GE Vivid 7 echocardiograph. A right-parasternal short-axis view of the LV at the level of the chordae tendineae was imaged in 2D mode at a frame rate between 59 and 73 FPS. 2DST analyses were performed blinded and in random order using the GE EchoPac Software (v6.1.2). Measurements included circumferential and radial peak-systolic strain (SC, SR); circumferential and radial strain rate during systole (SrC-S; SrR-S), early diastole (SrC-E; SrR-E), and late diastole (SrC-A; SrR-A); and maximum systolic radial displacement (DR). All measurements were reported by the software as average over each of six LV wall segments. Global strain, global strain rate, and global displacement were then calculated as averages over all segments. Reliability of 2DST was assessed by calculating measurement variability, within-day interobserver variability, between-day intraobserver variability, and between-day interobserver variability of all variables. Variability was reported as coefficient of variation (CV) in per cent.
2DST analyses were possible for all recordings in which at least one complete cardiac cycle was available. The automated tracking seemed accurate during systole but inaccurate during early diastole based on visual assessment by the operator. Generally, reliability was higher for SR and SrR compared to SC and SrC. For segmental SR, SrR-S, and DR, measurement error, within-day interobserver variability, between-day intraobserver variability, and between-day interobserver variability, respectively, were 2.4-12.4% [range of CV], 7.6-33.1%, 11.1-22.4%, and 8.8-26.5%. For global SR, SrR-S, and DR they were 4.1-6.3%, 12.6-15.3%, 12.0-16.1%, and 12.0-15.7%.
We conclude that global SR, SrR-S, and DR can be reliably characterized in horses by use of 2DST. Circumferential strain and strain rate measurements are less reliable. Diastolic strain rate measurements may not be valid due to inaccurate tracking, possibly related to under-sampling. The use of longitudinal strain and strain rate by 2DST, determination of the clinical value of 2DST for assessment of LV systolic function and LV synchrony at rest and during stress-testing, and its applicability to horses with cardiac disease will require further investigation.