Prediction of clinically important acquired cardiac disease without an echocardiogram in large breed dogs using a combination of clinical, radiographic and electrocardiographic variables.

Introduction: Large breed (LB) dogs develop dilated cardiomyopathy (DCM) and myxomatous mitral valve disease (MMVD). Echocardiography is required for a definitive diagnosis but is not always available. Our objective was to assess the clinical utility of thoracic radiographs alone and in combination with physical examination and electrocardiography findings for the prediction of clinically important DCM or MMVD in LB dogs.
Animals: Four hundred fifty-five client-owned dogs ≥20 kg with concurrent thoracic radiographs and echocardiogram.
Materials and Methods: Medical records were reviewed and stored thoracic radiographs and echocardiographic images were measured to classify dogs as normal heart size (NHS), preclinical DCM, clinical DCM, preclinical MMVD (with cardiomegaly), clinical MMVD, or equivocal. Dogs with preclinical MMVD, without cardiomegaly, were classified as NHS. Vertebral heart size (VHS) and vertebral left atrial size (VLAS) were measured. Receiver operating characteristic curves and prediction models were derived.
Results: Prevalence of MMVD (39.3%) was higher than the prevalence of DCM (24.8%), though most MMVD dogs (67.0%) lacked cardiomegaly and were classified as NHS for analysis. The area under the curve for VHS to discriminate between NHS and clinical DCM/MMVD or preclinical DCM/MMVD was 0.861 and 0.712, respectively, while for VLAS, it was 0.891 and 0.722, respectively. Predictive models incorporating physical examination and electrocardiography findings in addition to VHS/VLAS increased area under the curve to 0.978 (NHS vs. clinical DCM/MMVD) and 0.829 (NHS vs. preclinical DCM/MMVD).
Conclusions: Thoracic radiographs were useful for predicting clinically important DCM or MMVD in LB dogs, with improved discriminatory ability when physical examination abnormalities and arrhythmias were accounted for.
Competing Interests: Conflict of Interest Statement The authors do not have any conflicts of interest to disclose.
(Copyright © 2021 The Author(s). Published by Elsevier B.V. All rights reserved.)