Advanced electrocardiography predicts early cardiac involvement and incident arrhythmias in Fabry disease

Abstract

Abstract Funding Acknowledgements Type of funding sources: None. Background Fabry disease is an X-linked disorder, with cardiovascular involvement characterised by progressive myocardial sphingolipid deposition. Cardiac disease is a major contributor to morbidity and mortality. Cardiac magnetic resonance (CMR) with T1 mapping and advanced electrocardiography (A-ECG) offer both diagnostic and prognostic potential. Purpose To evaluate the predictive power of A-ECG markers in identifying: 1) early cardiac involvement defined as low myocardial T1 on CMR, and 2) adverse cardiovascular outcomes defined as any arrhythmia requiring therapy, atrial fibrillation, hospitalisation for heart failure or mortality. Methods Patients included in this longitudinal, multi-centre study underwent same-day standard resting 12-lead ECG and CMR. CMR included standard cine imaging, T1 mapping with modified Look Locker inversion recovery (MOLLI, 5s(3s)3s), and late gadolinium enhancement (LGE). ECG digital files were analysed using in-house A-ECG software. A-ECG analysis included conventional ECG measures, derived vectorcardiographic measures, and singular value decomposition measures of waveform complexity. Significant A-ECG variables were identified using stepwise forward regression and incorporated in a multivariable logistic regression A-ECG score. A Youden index was applied to identify best threshold score and bootstrapping performed to calculate the area under the receiver operating characteristics curve (AUC), sensitivity, specificity, and 95% confidence intervals (CI). Results Among included patients (n=155, 40% male, age 46±14 years, 39% on enzyme replacement therapy), left ventricular mass index was higher in males compared to females (106 vs. 59 g/m2, p<0.001), 80% of patients had myocardial native T1 below the local reference range (933 vs. 968 ms, p=0.06), and 51% (70/136) had focal LGE. Multivariable A-ECG scores for detecting low T1, any arrhythmia, or atrial fibrillation had an AUC [95%CI], sensitivity, and specificity of 0.82 [0.75-0.89], 72 [55-95]%, 85 [66-71]%; 0.89 [0.82-0.95], 82 [68-94]%, 88 [70-96]%; and 0.89 [0.80-0.96], 92 [77-100]%, 83 [76-92]%, respectively, Figure 1. No predictors of heart failure hospitalisation or mortality were found. Conclusion A-ECG analysis of the resting 12-lead ECG has good diagnostic performance for predicting early myocardial involvement and the occurrence of arrhythmias in Fabry disease. This supports the use of A-ECG both as a screening tool to diagnose early cardiac disease, and for identifying those at risk of adverse arrhythmic outcomes.