Background: ARVC is one of the most common causes of SCD in the young. A recently published model predicts sustained ventricular arrhythmia (VA) in ARVC patients, but this surrogate outcome overestimates the risk of SCD.
Objective: To specifically predict life-threatening VA (LTVA: SCD, aborted SCD, sustained or ICD treated VT>250 bpm or VF)
Methods: In a retrospective cohort of definite ARVC patients from 5 registries in North America and Europe, we tested the association between 8 predictors at diagnosis [sex, age, prior sustained VA (≥30s, unstable or ICD treated VT; or aborted SCD), syncope, 24-hour premature ventricular complex (PVC) count, number of anterior and inferior leads with T-wave inversion (TWI), left and right ventricular ejection fraction] and LTVA in follow-up by Cox regression. The model was internally validated using bootstrapping.
Results: We included 864 definite ARVC patients (40±16 years; 53% male; 50% PKP2 variant carriers). Over 5.8 years [IQR 2.8, 10.6], 93 (11%) patients had LTVA. Only 4 predictors were associated with LTVA (Figure1 Panel A). Notably, prior sustained VA did not predict subsequent LTVA (p=0.85). The model had an optimism-corrected C-index of 0.74 (95%CI:0.69-0.80) and calibration slope of 0.95 (95%CI:0.94-0.98) with good concordance between predictions and observations (Figure1 Panel B). Setting an ICD use threshold at 4% risk per the model would result in a number of ICDs needed to treat one patient with LTVA of 7 at 5 years.
Conclusion: LTVA events can be accurately predicted in ARVC patients by a novel prediction model using only 4 predictors which can be developed as a clinically applicable risk score to facilitate shared decision-making for ICD use.