Measurement of cardiac function by transthoracic echocardiography: day to day variability and repeatability in normal Thoroughbred horses

In order to assess the reliability and repeatability of transthoracic echocardiography for detecting serial changes in cardiac function in horses, day to day variability of a number of echocardiographic indices of ventricular function were studied. The variables investigated were, from 2-dimensional (2-D) and M-mode echocardiography - aortic diameter in systole (Aos), pulmonary artery diameter in systole (Pas), left ventricular internal diameter in systole (LLVIDs) and diastole (LLVIDd), and left ventricular fractional shortening (%FS) and estimated ejection fraction (EF). From pulsed Doppler echocardiography - maximum velocity (Vmax) and acceleration (dv/dtmax) of aortic and pulmonary arterial blood flow, left and right ventricular velocity time integral (AoVTI and PAVTI) and estimated cardiac output (COLV, and CORV), left and right ventricular ejection time (LVET and RVET) and pre-ejection period (LVPEP and RVPEP), and heart rate. Maximum velocity of early rapid right ventricular filling (ETV) and late right ventricular filling due to right atrial contraction (ATV) were measured. Maximum deceleration and time for deceleration of early rapid filling (dv/dtTV and dt TV) were also determined. Seven healthy mature Thoroughbred horses weighing 490-600 kg were studied. The horses' management regimes and environment were standardised and an echocardiographic examination was carried out on each horse at the same time every day for 6 consecutive days. Two statistical measures of repeatability were calculated for each variable: the intraclass correlation coefficient (rI) and the 95% confidence intervals for error free value of a single measurement. In general, the dimensions and indices of cardiac function derived from 2-D and M-mode echocardiography showed better repeatability for individual horses than did the indices derived from Doppler echocardiography. Overall the 95% confidence intervals for an error-free value of all variables derived from Doppler echocardiography were wide. This must be appreciated when attributing serial measured changes of Doppler echocardiographic variables in an individual animal to effects of a treatment or disease.