This investigation is one of the first studies analyzing pulmonary artery growth pattern between the time of BDCPC and TCPC focusing on the potential beneficial impact of central pulmonary artery stenting [7]. Our data show that PA stenting becomes frequently necessary, as shown by others [5,6,7,8,9,10], and may ensure symmetric PA growth comparable with the non-stented PA side branch.

Factors associated with need for central PA stenting may be predicted before BDCPC, especially for the LPA, which is predominantly affected. This includes smaller indexed diameter of LPA pre-BDCPC, larger diameter of the (neo-) ascending aorta pre-BDCPC, and cardiac anatomy consistent with HLHS-complex. These predictors align with the intraprocedural findings, according to which a compression from the (neo-) ascending aorta presumably is the main cause of LPA stenosis, whereas two of three patients that underwent RPA-stenting had a right aortic arch.

Pulsatile compression from a larger (neo-) ascending aorta is an important anatomic feature for potential postoperative LPA compression [7, 8, 11, 12]. Especially in HLHS-complex patients the position of the Damus-Kaye-Stansel anastomosis moves the ascending aorta more to the left in comparison to patients with a native aortic arch, while the reconstruction of the inner curvature of the aortic arch with xenopericardium may lead to foreshortening over time and thus decreases the space under the arch through which LPA and left bronchus pass. These could contribute to LPA—and bronchus—compression [1].

The surgical approach with a comprehensive stage I and II had been associated with higher incidence of PA-interventions [8, 9], but this was not a significant risk in our cohort. Further on, earlier BDCPC seems to affect only marginally PA growth. All other assessed anatomical and surgical factors are not associated with an increased risk for PA-stent implantation after BDCPC.

SV patients, depending on their cardiac anatomy, might have small pulmonary arteries directly after birth [2,3,4]. During the interval between stage I and II they can have a “catch-up” growth due to pulmonary overflow and pulsatile pulmonary blood flow (Sano shunt), while after stage II they normally have an inadequate development of the pulmonary vasculature, that could lead to increased pulmonary vascular resistance [2,3,4]. BDCPC creation from a shunted single ventricle leads to volume unloading, which may lead to under-filling of the branch PAs. These considerations are in line with the relative decrease of the size of PAs between BDCPC and TCPC found in our analysis and previous assessments in literature [23,24,25].

Since the growth of the PAs depends on blood flow [26], improvement in the caliber of the LPA by PA-stent implantation could potentially maximize future growth [7]. Currently, there is limited data published on distal PA growth after PA-stent implantation. One single analysis [19] on 18 SV patients showed, a similar growth of the stented and non-stented contralateral pulmonary artery branches, which is in line with our findings. In our cohort, we could also show that the growth of the PA-system is not only symmetric with the contralateral side, but also similar to the group without a PA-stent. As expected, a PA-stent implantation does not help to fully “catch-up” in growth, but ensures a PA growth comparable to non-PA stented patients [2,3,4]. However, other studies have shown direct positive implications of a LPA-stent implantation. In fact, an increase in cardiac output associated with LPA-stenting in SV patients has been demonstrated [27], while MRI-based studies have shown, that zones of narrowing within the non-pulsatile pulmonary artery circuit relate to relevant energy losses in proportion to the degree of narrowing [1].

It is well reported in literature, that early postoperative cardiac catheterization after surgery for congenital heart disease is feasible and safe, even when acting on freshly formed sutures or hemodynamic instable patients [5, 12, 14,15,16,17]. Stent implantation to treat PA stenosis in SV patients is effective and can be realized safely [7, 10]. In other cohorts [15, 27, 28], angioplasty was performed safely on fresh suture lines without reported vascular tear or suture disruption. In our cohort, one patient probably experienced a suture disruption, but this was related to the balloon-angioplasty prior to LPA-stent implantation. This is concordant with reports from Zahn et al. [29], which concluded that continuous Prolene suture lines can be expanded safely without disruption when using balloon to stenosis ratio ≤ 2.5/1 [29]. In our cohort, we used a stent-to-stenosis ratio of 2.4 (IQR 1.9–3.7). Nevertheless, a not negligible periprocedural risk profile of PA stenting early after BDCPC remains and an immediate surgical backup dealing with severe PA bleeding complications is necessary [17].

The selection of a particular stent is mainly depended on the size of the reference vessels [29, 30], the characteristics of the stenosis [13], and the size of the patient with regards to the sheath size needed for implantation. The initial diameter of the stent was chosen conservatively without overdilating the stenosed area. Since the rate of vessel complications associated with PA-stenting is low, bare metal (instead of covered) stents were used.

Since stents implantation was performed in growing vessels, repetitive stent re-dilatations were performed to accommodate the vessel diameter for somatic growth. Stent re-dilatation was always feasible [31].

When performing stent implantation stents that can be dilatated to reach the adult diameter of the pulmonary arteries (ca. 15 mm) should be preferred [32]. Until now, all available stents have limited final expansion diameters. Breakable stents, such as the Bentley BeGrow™ stents, could be a viable option in some cases since they can be implanted in small vessels (4-6 mm) with a 4 French sheath but have the potential for further re-dilations and controlled stent strut breakage to fit with the growth of the contralateral pulmonary branch [22]. Bioabsorbable stents could also potentially allow further growth of the artery and remodeling of the vessel wall [33].

A model for prediction of the necessity for an LPA-stent implantation after BDCPC would be an appealing tool for the clinician and could potentially lead to a change in the clinical management of these patients. In fact, our prediction model might estimate the probability of the necessity for PA-stent implantation and could be used to (a) perform a close follow-up or (b) plan an exit-angiography directly after BDCPC in patients at risk for pulmonary artery complications or, in high-risk patients, (c) perform a hybrid procedure for LPA stenting during BDCPC surgery.

In summary, a stent implantation (most often for external LPA compression), when clinically indicated, ensures symmetrical pulmonary flow and growth of the hilar and intrapulmonary vessels. However, side branch PA stents require multiple catheter interventions to match the child’s growth.

Our study is a retrospective study with its inherent limitations. A group comparison with patients that had a stenosis in the PA-system but did not receive a PA-stent is impossible from an ethical perspective. The growth of the pulmonary arteries is influenced by many other factors, such as the blood flow distribution to the left and right PA, that we did not assess. The direct clinical impact of the PA-implantation is hard to define and therefore not assessed in this analysis. The aim of our study focused on the time frame between BDCPC and TCPC rather than long-term outcome. Further surgical factors such as placement of the pulmonary artery confluence to the left or right of the midline [1, 34] could not been evaluated. The architecture of the aortic arch other than the dimensions of the ascending aorta has not been evaluated (34). Lastly, our prediction model is based on the results of our institution, and it might not fit for another institution with a different surgical or interventional strategy.