Subgroup analysis
The treatment effect was distinct between the patients with supravalvular and subvalvular EVA sites. For multiple-comparison, either strategy was superior in IA success rate to treat its ipsilateral EVA sites as compared with contralateral ones (P<0.0083) (Table 2). In the logistic regression model, patient sex, age, presence of VT, PVC frequency, sharp potential reversal at target, EVA site (supravalvular or subvalvular), initial ablation strategy, RVOT diameter, and PA diameter were included to adjust clinical confounders (Table 2). The EVA site was the only baseline characteristic with a remarkable multiplicative treatment interaction (Pinteraction<0.001). The AERI was calculated as 95.0%+85.7%-46.7%-28.6%=105.2%, suggesting a large magnitude of super-addictive interaction between the EVA site and its ipsilateral strategy. In patients with supravalvular EVA sites, supravalvular strategy was associated with a lower risk of IA failure (HR 0.021, 95% CI 0.002-0.214; P=0.001). Whereas in patients with subvalvular EVA sites, the supravalvular strategy was associated with a higher risk of IA failure (HR 7.000, 95% CI 1.098-44.607; P=0.039). The supravalvular strategy had fewer ablation applications in treating supravalvular EVA sites than subvalvular EVA sites, as well as the subvalvular strategy (P<0.0083) (Table 3).