External Validation of a Risk Prediction Model for Ventricular Arrhythmias in Arrhythmogenic Right Ventricular Cardiomyopathy

Abstract

The new 5-year ventricular arrhythmia (VA) occurrence risk model is a major breakthrough for arrhythmia risk stratification in the challenging population of patients with arrhythmogenic right ventricular cardiomyopathy (ARVC). In the original study, the model resulted in a 20.6% reduction in implantable cardioverter-defibrillator (ICD) placement compared with the 2015 consensus, for the same protection level. However, only internal validation was performed, limiting generalisation. We externally validated the model in a European tertiary care cohort of 128 patients with ARVC with restrictive indications for primary prevention ICD placement. Overall, 74% were men, none had VA history, and a single patient had an ICD at baseline. Median age at diagnosis was 38 years (interquartile range [IQR] 28-50). During a median follow-up of 7.8 years (IQR 6.1-9.7), 15 patients (12%) experienced VA. The model provided good discrimination, with a C-index for 5-year VA risk prediction of 0.84 (95% confidence interval 0.74-0.93). However, the model led to an overestimation of the 5-year VA risk when applying thresholds < 50%. With a < 10% predicted risk, no patient showed VA. With a 7.5% predicted risk, the ICD:VA ratio was 6.3 vs 3.4 in the original study. The model still outperformed the 2015 International Task Force Consensus. Overall, in a relatively large European ARVC cohort with restrictive indications for ICD placement, the ARVC model for VA prediction successfully identified ARVC patients with VA during follow-up. Yet, our study underscores the need for careful threshold selection, considering the model's associated risk overestimation in low- to intermediate-risk patients.