A Primary Prevention Clinical Risk Score Model for Patients With Brugada Syndrome (BRUGADA-RISK)
Shohreh Honarbakhsh, Rui Providência, Jorge García-Hernández, Claire A Martin, Ross J. Hunter, Wei Yao Lim, Claire Kirkby, Adam Graham, Ardalan Sharifzadehgan, Victor Waldmann, Éloi Marijon, Carmen Muñoz‐Esparza, Javier Lacunza, Juan R. Gimeno, Bénédicte Ankou, Philippe Chevalier, Natália António, Luís Elvas, Silvia Castelletti, Lia Crotti, Peter Schwartz, Maurício Scanavacca, Francisco Darrieux, Luciana Sacilotto, Johanna Mueller-Leisse, Christian Veltmann, Alessandro Vicentini, Andrea Demarchi, Nuno Cortez‐Dias, Pedro Silverio Antonio, João de Sousa, Pedro Adragão, Diogo Cavaco, Francisco Morosco Costa, Ziad Khoueiry, Serge Boveda, Mario João Sousa, Zeynab Jebberi, Patrick M. Heck, Sarju Mehta, Giulio Conte, Tardu Özkartal, Angelo Auricchio, Martin Löwe, Richard J. Schilling, David Prieto‐Merino, Pier D. Lambiase
Abstract
OBJECTIVES: The goal of this study was to develop a risk score model for patients with Brugada syndrome (BrS). BACKGROUND: Risk stratification in BrS is a significant challenge due to the low event rates and conflicting evidence. METHODS: A multicenter international cohort of patients with BrS and no previous cardiac arrest was used to evaluate the role of 16 proposed clinical or electrocardiogram (ECG) markers in predicting ventricular arrhythmias (VAs)/sudden cardiac death (SCD) during follow-up. Predictive markers were incorporated into a risk score model, and this model was validated by using out-of-sample cross-validation. RESULTS: A total of 1,110 patients with BrS from 16 centers in 8 countries were included (mean age 51.8 ± 13.6 years; 71.8% male). Median follow-up was 5.33 years; 114 patients had VA/SCD (10.3%) with an annual event rate of 1.5%. Of the 16 proposed risk factors, probable arrhythmia-related syncope (hazard ratio [HR]: 3.71; p < 0.001), spontaneous type 1 ECG (HR: 3.80; p < 0.001), early repolarization (HR: 3.42; p < 0.001), and a type 1 Brugada ECG pattern in peripheral leads (HR: 2.33; p < 0.001) were associated with a higher risk of VA/SCD. A risk score model incorporating these factors revealed a sensitivity of 71.2% (95% confidence interval: 61.5% to 84.6%) and a specificity of 80.2% (95% confidence interval: 75.7% to 82.3%) in predicting VA/SCD at 5 years. Calibration plots showed a mean prediction error of 1.2%. The model was effectively validated by using out-of-sample cross-validation according to country. CONCLUSIONS: This multicenter study identified 4 risk factors for VA/SCD in a primary prevention BrS population. A risk score model was generated to quantify risk of VA/SCD in BrS and inform implantable cardioverter-defibrillator prescription.