Biventricular repair after Norwood palliation

BACKGROUND: In general, neonates with severe left ventricular outflow tract obstruction, aortic valvar stenosis or atresia, and arch hypoplasia with either interruption or coarctation, and a small left ventricle undergo Norwood palliation followed classically by a bidirectional cavopulmonary shunt and eventual modified Fontan. However, a subset of patients, usually neonates with a ventricular septal defect, may have adequate left ventricle and mitral valve sizes making them candidates for future biventricular repair (BVR). In view of the long-term advantage of BVR, the feasibility and outcome of this approach was studied. Additionally, echocardiographic data were reviewed in an attempt to develop objective prognostic criteria for selection of patients suitable for BVR. METHODS: During a 4-year period, 8 of 58 infants undergoing Norwood palliation were identified as potential two-ventricle candidates. Their mean age was 6 days. Diagnoses included aortic atresia (n = 1), or aortic valve stenosis and subaortic stenosis (n = 7), with an interrupted aortic arch in 3 and coarctation in 4. All patients had a ventricular septal defect and a left ventricle that was considered to be apex forming. Mean mitral valve size was 11 mm (z-score = -1.7). Mean aortic valve size was 4.1 mm (mean z-score = -8.4). RESULTS: All 8 patients survived Norwood palliation. Six subsequently underwent BVR with ventricular septal defect closure and a right ventricle to pulmonary artery conduit at a mean age of 7 months. One patient is awaiting repair, and 1 underwent a cavopulmonary shunt. At the time of BVR, mean mitral valve z-score was essentially unchanged at -1.4 (14 mm). No early deaths or late deaths occurred during a mean follow-up of 32 months. CONCLUSIONS: A small subset of patients requiring Norwood palliation as newborns may be candidates for eventual BVR with low risk. In general, patients suitable for BVR have a mitral valve z-score of more than -3 and a normal-sized left ventricle. Recognition of neonatal BVR candidates enables consideration of complete neonatal repair. However, single-stage repair needs to be compared with the excellent results obtainable with the staged approach.