Unifocalization cannot rely exclusively on native pulmonary arteries: the importance of recruitment of major aortopulmonary collaterals in 249 cases

OBJECTIVES: We sought to define the early and late outcomes of unifocalization based on a classification of the native pulmonary artery (nPA) system and major aortopulmonary collateral arteries (MAPCAs) with a policy of combined recruitment and rehabilitation and to analyse the role of unifocalization by leaving the ventricular septal defect (VSD) open with a limiting right ventricle-pulmonary artery (RV-PA) conduit in borderline cases. METHODS: An analysis of 271 consecutive patients assessed for unifocalization at a single institution between 1988 and 2016 was performed. Patients were classified according to the pulmonary blood supply: group A, unifocalization based on nPA only; group B, based on nPA and MAPCAs; group C, MAPCAs only (absent nPAs). RESULTS: Unifocalization was achieved in 249 (91.9%) cases with an early mortality of 2.8%. Group A included 72 (28.9%) patients, group B 119 (47.8%) patients and group C 58 (23.3%) patients with no difference in early survival between groups. Survival at 5, 10 and 15years was 90.0% (85.9-94.3), 87.2% (83.5-91.2) and 82.3% (75.2-89.9), respectively. Late survival in groups A and B was similar but 10- and 15-year survival in group C decreased to 79.2% (68.2-92.1) and 74.3% (61.1-90.4) (P=0.02), respectively. A mean of 1.9 (0.6) MAPCAs were recruited per patient (range 0-6). The VSD was left open with a limiting RV-PA conduit in 97 (39.0%) cases, but subsequently closed in 48 patients, giving a total of 200 (80.3%) patients achieving VSD closure (full repair). Delaying VSD closure was not associated with increased risk for early or late survival. A central shunt to rehabilitate the nPAs was used in 56 (22.5%) cases. This was associated with a reduction in the number of MAPCAs recruited, but still required a mean of 1.8 (+/- 0.5) MAPCAs recruited per patient to achieve unifocalization. In multivariate risk analysis, those suitable for single-stage full repair had the best long-term outcomes. Group C anatomy was associated with poor late survival compared to groups A and B (hazard ratio 2.7). CONCLUSIONS: Survival is maximized by a combined approach of rehabilitation and recruitment. MAPCAs should always be recruited if they supply areas with absent nPA supply. A strategy of leaving the VSD open with a limiting RV-PA conduit is a safe and effective way of managing borderline cases.