LAUSR

Abstract Funding Acknowledgements Type of funding sources: None. Background Predicting likelihood of severe ventricular arrhythmias in Brugada syndrome (BrS) patients is challenging due to conflicting evidence. Objectives The goal of this study was to construct a prediction model of severe ventricular arrhythmias (VAs) for BrS patients. Methods Two hundred forty-two BrS patients were enrolled from 2008 to 2019 in a single-center cohort study. Those patients were followed up until December 2021. Clinical data and multiple ECG markers were collected and analyzed to determine the risk factors of severe VAs and sudden cardiac death (SCD). Multivariate logistic regression analysis was then used to develop a risk prediction model for adverse arrhythmic outcomes in BrS patients. Results During the follow-up (90.1 ± 45.8 months) of 242 BrS patients (mean age 42.3 ± 12 years; 94.6% male), 49 (20.2%) patients had syncope/pre-syncope due to VAs and 7 (2.9%) patients had aborted cardiac arrest. In multivariable analysis, epsilon wave, early repolarization sign, S wave duration in lead I, QRS duration in V2 and Tp-e duration were significant associated with adverse arrhythmic outcomes. A clinical model including these five significant predictors had a good performance to predict outcomes with balanced accuracy of 0.86 and area under curve of 0.93 (95% CI, 0.89-0.96, p < 0.001) indicating a good discriminative ability. Furthermore, the Hosmer–Lemeshow test and the calibration plot showed a good fit between the predicted and observed probabilities of the predictive model. Conclusions This study constructed a risk prediction model for severe VAs in BrS patients with a high predictive accuracy. Multivariable Logistic analysis Risk prediction model