Predictors of reintervention after repair of interrupted aortic arch with ventricular septal defect

Background: Left ventricular outflow tract obstruction after neonatal repair of interrupted aortic arch with ventricular septal defect may warrant reintervention. We sought to identify clinical and preoperative echocardiographic predictors of reintervention for postoperative left ventricular outflow tract obstruction.

Methods: Retrospective data were collected on neonates with interrupted aortic arch with ventricular septal defect who underwent single-stage repair from 1995 to 2009. Univariate and multivariate analyses were performed to identify predictors of reintervention.

Results: Seventy patients underwent repair, with 16 patients requiring reintervention: 8 underwent surgical reintervention, 5 underwent percutaneous reintervention, and 3 underwent both. The median time to reintervention was 1.2 years (range, 0.2 to 7.7). All surgical reoperations involved subaortic resection, and all percutaneous reinterventions included balloon aortic valve dilation. Several preoperative echocardiographic measurements were significant by univariate analysis; however, smaller preoperative aortic root size was an independent predictor (p = 0.02) by multivariate analysis. Patients with an aortic root size less than 6.5 mm were at greater risk for reintervention compared with patients with a root size greater than 6.5 mm (reintervention rate 44% and 12%, respectively; p < 0.001). Postoperative left ventricular outflow tract gradient by echocardiogram before discharge was significantly higher in the reintervention group.

Conclusions: Preoperative aortic root size predicts reintervention for postoperative left ventricular outflow tract obstruction after single-stage repair of interrupted aortic arch with ventricular septal defect. Patients with elevated left ventricular outflow tract gradients at discharge are at higher risk of having progressive obstruction and require closer follow-up to ensure early identification and management.