In infants with TOF, palliative shunting has been performed safely and with excellent outcomes23,²⁴⁾. Even in young infants, this option still provides good outcomes similar to those of primary repair14,²²⁾. Recently, Kanter et al.¹⁴⁾ reported excellent early outcomes including shorter intensive care unit and hospital stays for the first operation and early mortality of 5.9%. According to the STS database, discharge mortality from TOF repair in neonates was 11 of 178 (6.2%) for palliation and 12 of 154 (7.8%) for primary repair²²⁾. However, outcomes of shunting in neonates have been reported inconsistently and with fluctuations. In a multicenter study of the effect of aspirin on shunting, clinical outcomes of palliative shunts in infants were poor even in the current era²⁵⁾. This study showed that event rates of shunt thrombosis and death were 5% and 15%, respectively, for TOF patients and 16% and 21%, respectively, for patients with pulmonary atresia. Recently, Guzzetta et al.²⁶⁾ reported outcomes of in-hospital shunt occlusion in 207 young infants undergoing only a modified Blalock-Taussig shunt, in which in-hospital shunt occlusion occurred in 14 patients (6.8%). Patients who had shunt occlusion had a harder postoperative course and a higher rate of in-hospital mortality (6.2% vs. 21.4%). They concluded that the risk factors of in-hospital shunt occlusion were cardiac diagnosis (pulmonary atresia) and the size of the pulmonary arteries.

With regard to the safety evaluation, interim mortality should be addressed in infants undergoing shunting13,27,²⁸⁾. Interim mortality occurs during the palliated state before repair and may account for additional attrition, which may or may not be shunt-related²⁹⁾.

For TOF neonates with pulmonary stenosis, the influence of shunting on the growth of the pulmonary annulus should be evaluated. Pulmonary regurgitation after TOF repair results in various adverse long-term outcomes³⁰⁾. Therefore, every effort should be made to preserve total pulmonary valve function as well as the pulmonary annulus. Recently, Kanter et al.¹⁴⁾ suggested that shunted patients have a greater likelihood of avoiding a transannular patch at the time of repair. Sousa Uva et al.³¹⁾ reported that initial palliation promoted the growth of pulmonary annulus and that transannular patching was less prevalent for patients who underwent initial palliation (13% vs. 56%, P=0.03). One year later, however, in a report on the same issue, the same authors revised this conclusion, stating that the observation of increased size of the pulmonary annulus after shunting could be due to chance only and that initial palliation did not allow for a reduction in incidence of transannular patching³²⁾. The incidence of transannular patching at the time of primary repair in neonates ranges from 84% to 100%10,12,14,³³⁾. Symptomatic neonates usually receive a transannular patch because of their morphology, not because of their age. The need for a transannular patch reflects the severity of the RVOT obstruction at the annular level7,10,12,¹⁸⁾. Parry et al.¹⁸⁾ reported outcomes of elective primary repair of acyanotic TOF in early infancy, in which only 28% of patients required transannular patching. Follow-up echocardiography suggested a trend towards 'catch-up' growth of the annulus. By contrast, in a disease entity distinct from TOF, Emani et al.³⁴⁾ showed progressive decrease in the pulmonary valve annular size in cases of delayed anatomic repair beyond the neonatal period. They suggested that antegrade blood flow may provide the stimulus for growth of ventricular outflow tract structures, while palliative maneuvers such as aortopulmonary shunts may lead to a reduction in blood flow through the outlet area.

Between May 2004 and December 2012, 27 consecutive neonates with TOF underwent a surgical intervention. Their cardiac diagnoses were TOF with pulmonary stenosis (n=6) or atresia (n=21) with no major aortopulmonary collateral vessels.

All patients had symptoms before surgery. Twenty-six patients were receiving an infusion of prostaglandin. The remaining patient with pulmonary stenosis suffered an anoxic spell and was treated with propranolol and mechanical ventilator care. Ten patients needed mechanical ventilator care temporarily or until repair.

Intracardiac anatomy and the RVOT were assessed preoperatively by using transthoracic echocardiography. The morphology and size of the pulmonary arteries and the presence of major collateral arteries were assessed by performing chest computed tomography (CT) scan. Preoperative chest CT scan images were available to measure the size of the branch pulmonary arteries in 24 patients. McGoon ratio⁴¹⁾ and Nakata index⁴²⁾ had median values of 1.1 (range, 0.6 to 1.5) and 121 mm²/m² (range, 27 to 189 mm²/m²), respectively (Fig. 1). The single patient who had undergone shunting for small pulmonary arteries had a McGoon ratio of 0.6 and a Nakata index of 27 mm²/m² (Fig. 1, black triangle^*).

Twenty-five neonates (93%) underwent primary repair. The median age at repair was 16 days (range, 12 to 29 days) and median body weight was 3.2 kg (range, 2.2 to 4.2 kg). Two patients (7%) had aortopulmonary shunts. One of these had a seemingly small LV size; the other had diffuse hypoplastic branch pulmonary arteries despite having no major aortopulmonary collateral vessels (Fig. 2).

The early primary repair included VSD closure (a transatrial approach in 20 and a transventricular approach in 5), resection of the hypertrophied right ventricular muscle and RVOT reconstruction with various techniques. These included transannular or transjunctional RVOT widening with an autologous pericardial patch (n=12), interposition of a conduit made with an autologous pericardial roll between the right ventricle and the pulmonary arteries (n=8, 7 to 10 mm in diameter), RVOT reconstruction using the left atrial auricle as a flap (n=3), and REV (reparation a l'etage ventricularie)-type reconstruction (native tissue to tissue anastomosis between the posterior wall of the main pulmonary artery and right ventricle, n=2). No materials for preventing pulmonary regurgitation such as artificial monocuspid or bicuspid pulmonary valves were used in any patients. Fenestration of VSD was not necessary in any patient.

All primary repairs were performed after standard median sternotomy by using full-flow cardiopulmonary bypass with moderate systemic hypothermia. Antegrade crystalloid cardioplegia was used for myocardial protection. Median cardiopulmonary bypass and aortic cross clamp times were 153 minutes (range, 95 to 257 minutes) and 75 (range 50 to 117 minutes) minutes, respectively. Total circulatory arrest was not necessary in any patient.

There were no hospital deaths. Delayed sternal closure was necessary in 6 patients and postoperative complications requiring an interventional operation occurred in 5 patients. These complications comprised left-side diaphragm palsy (n=3), postoperative bleeding (n=1) and wound infection (n=1). No patient developed junctional ectopic tachycardia or complete atrioventricular heart block.

Follow-up was performed at a median interval of 56 months (range, 2 to 105 months). There were 2 late deaths. Actuarial survival rates at 6 months, 1 year, and 5 years were 96%, 92%, and 92%, respectively. Among 23 late survivors, 15 patients (65%) have not needed further surgical intervention after the primary repair. Eight patients have undergone surgical procedures of RVOT at a median interval of 14 months after total repair (range, 6.8 to 51 months) and 3 patients have needed second re-do operations for RVOT stenosis. The first catheter intervention for RVOT obstruction was performed after primary repair in 14 patients (61%) at a median interval of 7.4 months (range, 3 to 19 months). This was performed for left pulmonary artery stenosis in 6 patients, for right pulmonary artery stenosis in 2 patients, for stenosis of both branch pulmonary arteries in 5 patients, and for pulmonary trunk stenosis in 1 patient. Eight patients needed further catheter interventions such as reballooning or placement of a stent. Reoperation-free survival rates at 6 months, 1 year, and 5 years were 100%, 81%, and 60%, respectively, and catheter plus surgery intervention-free survival rates at 6 months, 1 year, and 5 years were 79%, 50%, and 33%, respectively. At the most recent follow-up, echocardiography of right and left ventricular systolic function was normal in all patients.

In 23 late survivors, the postrepair RV/LV pressure ratio was 0.61±0.15 (range, 0.37 to 0.89), unrelated to preoperative size of the pulmonary arteries (Fig. 3). However, for patients with ratios of 0.6 or more, catheter plus surgical interventions were needed significantly more frequently (10/11, 91% vs. 5/12, 42%; P=0.027 by Fisher exact test).