Predicting Suboptimal Outcomes Following Initially Acceptable Aortic Valve Repair in Children

BACKGROUND Aortic valve repair is preferred for congenital aortic valve disease in children, but its durability is limited. Although acceptable immediate repair results improve outcomes, late failures still occur. This study aimed to identify risk factors for suboptimal outcomes after an initially acceptable repair. METHODS We retrospectively reviewed 287 patients undergoing primary aortic valve repair (2014-2023). Acceptable repair was defined as less than moderate aortic stenosis (AS) and mild or less aortic insufficiency (AI) after bypass. Suboptimal outcomes were defined as more than moderate AS/AI or reintervention after discharge. Cox analysis identified predictors among patients with acceptable repair, with a nomogram developed for pure AS patients. RESULTS Median age at repair was 3.5 years (interquartile range, 1.0-8.1 years). Acceptable repair was achieved in 183 of 287 (64%). Median follow-up was 2.5 years (interquartile range, 0.7-4.7 years). Acceptable repair yielded superior 4-year freedom from suboptimal outcomes vs residual lesions (75% vs 51%, P < .001). Among 121 pure AS patients with acceptable repair, independent predictors for suboptimal outcomes were smaller preoperative annulus z-score (hazard ratio HR, 0.69; P = .007), higher preoperative peak gradient (HR, 1.02; P = .034), and prior balloon aortic valvuloplasty (HR, 5.92, P < .001). The nomogram showed good discrimination (C index = 0.78). For mixed/pure AI patients, leaflet extension/augmentation was associated with suboptimal outcomes on univariate but not multivariable analysis. CONCLUSIONS Acceptable initial repair is crucial but does not guarantee long-term success. Smaller annulus z-score, higher gradient, and prior balloon aortic valvuloplasty in pure AS independently predict subsequent valve failure despite adequate initial repair. These factors may aid risk stratification and inform surveillance intensity.