Choice of primary palliation

The mechanism underlying progressive bulboventricular foramen restriction is multifactorial. Reduction in ventricular volume load after pulmonary artery banding, bidirectional cavopulmonary shunt, and Fontan completion may contribute to altered ventricular geometry and increased wall thickness, predisposing to progressive narrowing at the bulboventricular foramen ^{Reference Rychik, Jacobs and Norwood7,Reference Alsoufi, McCracken, Kanter, Shashidharan and Kogon8,Reference Alsoufi, Slesnick and McCracken11,Reference Mosca, Hennein and Kulik12} . Our observation that bulboventricular foramen restriction occurred disproportionately in the pulmonary artery banding + arch reconstruction group is consistent with prior reports that systemic outflow tract obstruction is more likely in patients with concomitant arch hypoplasia or after arch reconstruction. However, this association may also reflect confounding by indication, as patients requiring arch reconstruction represent a more complex anatomical subset with a higher baseline risk. A key clinical challenge is the selection of the optimal palliative pathway. Although some institutions achieve excellent outcomes with primary Norwood reconstruction in non-HLHS single-ventricle anatomies ^{Reference Alsoufi, McCracken, Kanter, Shashidharan and Kogon8,Reference Alsoufi, Slesnick and McCracken11,Reference Mosca, Hennein and Kulik12} , the Norwood operation remains a technically complex neonatal procedure associated with non-negligible mortality in many settings. ^{Reference Karl, Watterson, Sano and Mee5,Reference Mosca, Hennein and Kulik12–Reference McElhinney, Reddy, Silverman and Hanley19} Consequently, in patients judged to be at lower risk of systemic outflow tract obstruction, strategies such as pulmonary artery banding with or without arch reconstruction remain attractive due to lower early operative risk.

Ruzmetov et al. compared pulmonary artery banding ± arch reconstruction with the Norwood procedure and reported no significant difference in mortality between the groups, while freedom from heart block was better in the Norwood group. ^{Reference Ruzmetov, Geiss and Fortuna20} Tchervenkov et al. reported mortality rates of 43% in the pulmonary artery banding + coarctation repair group and 27% in the Norwood group, with a lower incidence of reintervention in the Norwood cohort. ^{Reference Tchervenkov, Shum-Tim, Béland, Jutras and Platt21} In a study by Kalustian et al. comparing palliative arterial switch operation, pulmonary artery banding ± arch reconstruction, and Norwood–Damus–Kaye–Stansel strategies in 71 patients, no significant differences were observed among groups with respect to overall survival, Fontan completion rate, reintervention-free survival during the first interstage period, freedom from pulmonary artery reintervention, or ventricular dysfunction. ^{Reference Kalustian, Spigel and Greenleaf22} In our series, the Norwood procedure was performed in three patients (17.6%), one of whom died.

An alternative technique is the palliative arterial switch operation; ^{Reference Karl, Watterson, Sano and Mee5,Reference Lacour-Gayet, Serraf and Fermont23} however, this approach is less favoured due to difficulties in controlling pulmonary blood flow. ^{Reference Lacour-Gayet24} The combination of pulmonary artery banding with coarctation repair and bulboventricular foramen enlargement has also been described, but it is rarely used today. ^{Reference Serraf, Conte and Lacour-Gayet25}

When bulboventricular foramen restriction develops after pulmonary artery banding + arch reconstruction, performing a Damus–Kaye–Stansel anastomosis as a second-stage procedure represents another option. In our cohort, three patients (17.6%) developed bulboventricular foramen restriction before bidirectional cavopulmonary shunt and underwent Damus–Kaye–Stansel at the second stage. When bulboventricular foramen restriction emerges late—particularly after division of the main pulmonary artery at the time of bidirectional cavopulmonary shunt—Damus–Kaye–Stansel may no longer be feasible, leaving bulboventricular foramen enlargement as the primary surgical option. Bulboventricular foramen enlargement is associated with restenosis risk, atrioventricular block, and ventricular dysfunction. ^{Reference Alsoufi, McCracken, Kanter, Shashidharan and Kogon8,Reference Ruzmetov, Geiss and Fortuna20,Reference Rothman, Lang and Lock26} Rijnberg et al. reported heart block and ventricular dysfunction rates of 4% each, and a restenosis rate of 30% in a series of 23 patients undergoing bulboventricular foramen enlargement. ^{Reference Rijnberg, Sojak, Blom and Hazekamp27} In our cohort, three patients developed bulboventricular foramen restriction after the main pulmonary artery division and underwent bulboventricular foramen enlargement; none developed heart block, although one patient (33.3%) required reintervention due to restenosis.

Another strategy is routine Damus–Kaye–Stansel anastomosis at the time of bidirectional cavopulmonary shunt, even in the absence of bulboventricular foramen restriction. Clarke et al. reported a primary palliation mortality of 4% and an overall mortality of 12% with this approach and suggested it as an alternative to the primary Norwood procedure. ^{Reference Clarke, Kasahara and Andrews28} Although Damus–Kaye–Stansel can be performed with lower mortality at the bidirectional cavopulmonary shunt stage, it adds surgical complexity. Therefore, selectively performing Damus–Kaye–Stansel only in patients at high risk for bulboventricular foramen restriction may represent a balanced strategy. An important question is whether predictors of bulboventricular foramen restriction reported in the literature remain applicable before bidirectional cavopulmonary shunt.

Prediction of BVF restriction

Several studies have attempted to predict bulboventricular foramen restriction before primary palliation. Matitiau et al. reported that a bulboventricular foramen area index <2 cm²/m² was a risk factor for bulboventricular foramen restriction. ^{Reference Rychik, Jacobs and Norwood7} A bulboventricular foramen/Aortic annulus ratio<1 has also been identified as a predictor of bulboventricular foramen restriction. ^{Reference Franklin, Sullivan and Anderson4,Reference Lacour-Gayet, Serraf and Fermont23} Park et al. reported that a low single-ventricle outflow tract area index and the presence of aortic arch hypoplasia were risk factors for outflow tract obstruction. ^{Reference Park, Baek and Kwon29} Lan et al. suggested that patients with a bulboventricular foramen area index between 1 and 2 cm²/m² could be managed with pulmonary artery banding or pulmonary artery banding + arch reconstruction. ^{Reference Lan, Chang and Drant30} pulmonary artery banding reduces ventricular volume load and alters ventricular geometry, while myocardial hypertrophy leads to increased wall thickness, all of which may contribute to bulboventricular foramen stenosis. ^{Reference Jensen, Williams, Laks, Drinkwater and Kaplan9,Reference Donofrio, Jacobs, Norwood and Rychik31,Reference Miura, Kishimoto, Kawata, Hata, Hoashi and Nakajima32} Several studies have also reported pulmonary artery banding + arch reconstruction as a risk factor for bulboventricular foramen restriction. ^{Reference Matitiau, Geva and Colan3,Reference Karl, Watterson, Sano and Mee5} In our study, pre-palliation echocardiographic parameters, including bulboventricular foramen area index and bulboventricular foramen/aortic annulus ratio, did not differ significantly between patients who did and did not develop bulboventricular foramen stenosis, although values tended to be higher in patients without subsequent restriction.

In addition, the anatomical and hemodynamic mechanisms underlying the association between a lower bulboventricular foramen-to-aortic annulus ratio and subsequent restriction deserve further consideration. A relatively small bulboventricular foramen in relation to the aortic annulus may reflect a limited effective systemic outflow orifice at baseline. During staged palliation, progressive reductions in ventricular preload after pulmonary artery banding, bidirectional cavopulmonary shunt, and Fontan circulation, together with myocardial hypertrophy and geometric remodelling of the dominant ventricle, may further decrease the effective cross-sectional area of the bulboventricular foramen. In parallel, continued growth of the aortic annulus, in contrast to limited or absent growth of the bulboventricular foramen, may exacerbate this mismatch over time, leading to a relative and eventually functional restriction of the bulboventricular foramen. Data regarding the prediction of bulboventricular foramen restriction in the pre-bidirectional cavopulmonary shunt period are limited. Matitiau et al. suggested that the initial size of the bulboventricular foramen, rather than its growth rate, is the critical determinant of later restriction. ^{Reference Matitiau, Geva and Colan3} In our study, we specifically evaluated echocardiographic and catheterisation data obtained before the bidirectional cavopulmonary shunt to identify risk factors at this stage. While the aortic annulus enlarges over time, bulboventricular foramen growth may be minimal, absent, or even negative, ^{Reference Matitiau, Geva and Colan3} leading to a progressive decline in the bulboventricular foramen/Aortic annulus ratio. However, the literature does not provide clear evidence regarding when this decrease becomes a risk factor for bulboventricular foramen restriction. In our cohort, the median pre-bidirectional cavopulmonary shunt bulboventricular foramen/Aortic annulus ratio was 0.7, and values were <1 in both groups who developed bulboventricular foramen stenosis and those who did not (0.6 versus 0.8). Our receiver operating characteristic analysis demonstrates that a pre-bidirectional cavopulmonary shunt bulboventricular foramen/aortic annulus ratio ≤0.65 identifies patients at risk for subsequent bulboventricular foramen restriction with strong discriminatory performance (AUC 0.87). This finding is particularly clinically relevant because it provides a practical pre-bidirectional cavopulmonary shunt echocardiographic parameter that may support selective performance of Damus–Kaye–Stansel at the time of bidirectional cavopulmonary shunt, aiming to prevent late bulboventricular foramen enlargement and its associated complications.