Comparison of three two-dimensional echocardiographic methods of assessing left ventricular size in Doberman Pinschers

Introduction: Compare three methods of obtaining linear left ventricular dimensions within the same Doberman Pinscher (DP). Animals: One hundred and thirty-nine client-owned DP. Materials and Methods: Linear left ventricular dimensions were measured using two-dimensional short-axis (Sx-2D), motion-mode short-axis (Sx-MM), and motion-mode long-axis (Lx-MM) methods, then left ventricular volumes were obtained using monoplane Simpson's method of discs (SMOD). A Friedman test with Dunn's multiple comparisons was used to compare differences between methods. Bias and correlation were evaluated via Bland-Altman and Spearman's correlation. Sensitivity and specificity for diagnosing occult dilated cardiomyopathy (DCM) compared to SMOD were determined. Coefficients of variation (CVs) were calculated for intra- and inter-observer measurement variability. Results: There were significant differences between all linear dimensions in diastole and systole. Short-axis 2D measurements had significant bias compared with Sx-MM (diastole +1.19 mm, systole +1.65 mm) and Lx-MM (diastole +4.36 mm, systole +3.87 mm) as did Sx-MM compared with Lx-MM (diastole +3.17 mm, systole +2.22 mm). All linear dimensions had a moderate positive correlation with SMOD. The sensitivity and specificity of linear measurements to detect DCM were: Sx-2D (sensitivity 72.0%, specificity 88.5%), Sx-MM (sensitivity 52.0%, specificity 92.0%), and Lx-MM (sensitivity 37.5%, specificity 99.1%). All methods had acceptably low CV for intra- and inter-observer measurement variability. Conclusions: Results of this study suggest that linear measurements are repeatable and correlate with reference standard; however, there is a significant bias between measurements, and they should not be used interchangeably.