Abstract 4367391: Predictors of Clinically Relevant Genetic Variants Identification in Dilated Cardiomyopathy within a Brazilian Population Sample

Dilated cardiomyopathy (DCM) has an underlying cause associated with clinically relevant variants (CRVs) in up to 40% of cases. The Madrid Score optimizes genetic testing yield by evaluating five variables: absence of left bundle branch block (LBBB), absence of hypertension, low ECG voltage, family history of DCM, and skeletal myopathy. However, it was developed in a population genetically distinct from Brazilians. This study aims to: evaluate the effectiveness of the Madrid Score, and identify predictors of CRVs in a Brazilian cohort of DCM patients. This retrospective observational study evaluated clinical characteristics and complementary test results from DCM patients who underwent genetic testing as part of the MAPA Genoma Brasil project. Inclusion criteria were: left ventricular ejection fraction <50%, and diagnosis after 14 years of age. Participants were stratified into two groups based on presence or absence of CRVs. The Madrid Score was applied to all patients with complete data for its required variables. Univariate and multivariate regression analyses were performed to identify statistically significant predictors. The study evaluated 749 patients, including 441 (58.9%) males. CRVs were identified in 168 (22.4%) patients. Application of the Madrid Score in this population showed increasing probability of CRV detection with higher scores (Image 1A), though at lower rates than in the original validation population. Multivariate regression identified the following independent predictors of CRVs: non-black (OR: 2.1; 95% CI: 1.02-4.2), NYHA class III-IV (OR: 1.8; 95% CI: 1.01-3.3), absence of LBBB (OR: 3.9; 95% CI: 2.1-7.4), presence of atrial fibrillation (OR: 1.9; 95% CI: 1.03-3.86) (Image 2). The ROC curve demonstrated a predicted probability of 0.697 (Image 1B). Based on these predictors, we developed the "DCM Genetic Score" (Image 3A), which showed increasing yield for CRV identification with higher scores in this population (Image 3B). The Madrid Score improved prediction effectiveness for identifying CRVs in a Brazilian DCM population sample, though with lower accuracy than in the original validation study. We identified four independent predictors of CRVs: absence of LBBB, advanced heart failure symptoms (NYHA class III-IV), presence of atrial fibrillation and being non-black. The newly developed DCM Genetic Score incorporated these predictors. These findings suggest potential genetic differences between Brazilian and European DCM populations.