Relationship of Aortopulmonary Collaterals and Pulmonary Artery Development During Staged Single Ventricle Reconstruction

The clinical significance of aortopulmonary collaterals (APCs) in patients with univentricular heart remains controversial. APCs represent an unpredictable source of pulmonary blood flow in patients with univentricular heart and are observed at each stage of surgical reconstruction toward Fontan circulation, as well as after the Fontan completion [1]. While APCs increase oxygen saturation in the short-term, they create a hemodynamic burden in single ventricle physiology by shunting up to 30–50% of total systemic blood flow to the lungs [2]. Whereas earlier studies are still discordant on the influence of APCs, more recent studies show that increased APCs are associated with adverse early post-Fontan outcomes [2, 3]. Many patients acquire a large amount of APC flow after BCPS which ultimately influences the occurrence of adverse events after TCPC [4]. There have been a number of reports of adverse events which have been attributed to the presence of APCs and the amount of APC flow. These include increased chest drain volume, prolonged hospitalization and postoperative recovery, and impaired neurodevelopmental outcome [5‐7].

While the overall effects of APCs have been widely described, the precise mechanisms leading to them are not well understood. One study suggests that the APC flow is related to small pulmonary artery size, implying that small pulmonary arteries may stimulate the development of APCs [8]. Furthermore, we have only recently been able to show that lower Nakata-Index is associated with the development of APCs [1].

However, the question remains whether APC development is influenced by pulmonary artery size or vice versa. In previous studies, we have investigated both the influence of APCs and the influence of pulmonary artery indices on the outcome after TCPC [1, 9]. The aim of this study was to investigate the relationship between these two factors and evaluate the impact of APCs on the pulmonary artery development during staged palliation.

In the present study, we demonstrated that the occurrence of APCs prior to BCPS did not influence PA index at the time of BCPS. On the contrary, the presence of APCs before TCPC showed an association with a smaller PA index at the time of TCPC. A significant change in right and left PA index between BCPS and TCPC was observed in patients with APCs. Patients who underwent intervention for APCs also demonstrated a lower PA index at TCPC, as well as a significant change in PA index between BCPS and TCPS.

Various factors have been discussed to be associated with the development of APCs. In earlier studies, we showed that the development of APCs pre-BCPS was associated with the anatomical subtype of aortic atresia and mitral atresia (AA/MA) and shunt types of right ventricle to pulmonary artery conduit (RVPAC) in patients with HLHS after the Norwood procedure [13]. Grosse-Wortmann and colleagues proposed that APC may be an adaptive mechanism during prolonged periods of cyanosis, decreased volume and velocity of flow, or promoted by low blood flow to the lungs [14]. Our data supported this hypothesis because patients who had APCs at the time of TCPC were associated with smaller PA index at the time of BCPS, compared with those did not have APCs. The short-term increase in oxygen saturation after formation of APCs suggests that APC growth may be induced by hypoxia [2, 15]. However, empirical data do not suffice to verify the hypothesis and the need for molecular analysis is called for. Sandeep et al. approached the topic of biomarkers that influence APC growth but came to the conclusion that plasma factors and angiogenic activity correlate poorly with APC severity in patients with single ventricles, suggesting complex mechanisms of angiogenesis [16].

As for the relationship between PA index and the presence of APCs in single ventricle patients, Seger et al. found that PA index was the only variable associated with APC flow (size inversely related to APC flow) in patients who underwent pre-Fontan cardiac magnetic resonance (CMR) imaging [17]. Latus et al. showed that PA index was inversely correlated with APC flow and APC flow was associated with smaller PA size in patients after the Fontan procedure, suggesting that small PA may represent a stimulus for the development of APCs in Fontan patients [8]. Our results showed that presence of APCs was associated with underdevelopment of PA before TCPC but not with PA size before BCPS. While there is no correlation between PA index and APCs prior to BCPS, this changes after the procedure is performed. We assume that before BCPS patients had shunt-dependent parallel circulation, in which PA flow might be determined by anatomical factors, such as size of aortopulmonary shunt or degree of pulmonary stenosis. Therefore, PA size might not be correlated with pulmonary blood flow. After BCPS, SVC flow becomes the only source of the pulmonary blood flow, PA size might become a more representative parameter and correlate with the development of APCs. Hydrodynamic circumstances and acting pressures change after BCPS is completed which may lead to a stronger mutual influence between PA index and APC. It is of note that patients who had APCs before TCPC had smaller PA index not only at the time of TCPC but also at the time of BCPS. We think these data are strong evidence that small PA may represent a stimulus for the development of APCs. Further studies must clarify the cause-and-effect relations taking into account other possible impact factors.

A correlation between PA index and APCs has been described several times, but to our knowledge we are the first to evaluate and compare the changes of PA index in association with occurrence of APC, and over the entire period of Fontan palliation. Schmiel, et al. showed that the presence of APCs was associated with lower PA index at the time of TCPC [1]. Wang and colleagues reported a significant negative correlation between APC flow and the PA index in patients who underwent bidirectional Glenn shunting [18]. In this current study, we analyzed the change of PA index dependent on APCs. While there was no significant progression in PA index in neither the APC group nor the group without APCs, we could demonstrate right PA index increased significantly in patients with APCs. This was even more prominent in cases where patients had to undergo interventional closure of APCs. Our results suggested that development of APCs between BCPS and TCPC may be a compensation of small pulmonary arteries, and the size of pulmonary arteries increases in spite of the appearance of APCs. In other words, we assume that the development of APCs does not impair the growth of pulmonary arteries. In patients who developed APCs before TCPC, the right PA index increased significantly more than in those who did not. The left PA index tended to decrease in both patients with and without APCs, however the extent of the decrease was milder in patients with APCs compared to those without. This is a new finding. This is a new finding, and we speculate that an increase in the total pulmonary vascular beds between BCPS and TCPC might enable the development of central PA and APCs in these patients. Powell suggests that APCs may also increase PA pressure by adding to pulmonary blood flow [19]. APC blood flow may also compete with PA flow, likely via a rise in mean PA pressure [14]. Dori et al. reported that embolization of APCs resulted in significant decreases in total pulmonary blood flow as well as the ratio of pulmonary-to-systemic flow [20]. It might be possible that the increase in PA pressure may even trigger the growth of the pulmonary arteries. It is noteworthy that patients with a sufficiently large PA index did not experience an increase in PA index and did not develop APCs, whereas in patients with small PA index; PA index increased and they also developed APCs during the period between BCPS and TCPC. Further studies are mandatory to clarify the relationship of PA development and APC development, and their effects on the outcomes after the Fontan procedure. This study was designed to gain a deeper understanding of the postoperative period after the different stages of palliation. This should make it possible to interpret individual courses and to better assess possible complications that patients might expect. This contributes to a deeper understanding of the clinical management of the Fontan patient.

This study is limited by its retrospective, non-randomized, and single-center design. The ability to identify APCs is highly dependent on technical variables and introduced detection bias. The time of the development of APCs may differ from the time of catheterization and may therefore introduce detection bias. Furthermore, quantitative analysis of APC burden was not available. Angiography is not routinely performed as a follow up after Fontan completion or APC closure. Further development of the PA index and whether the development might be influenced by interventional closure of the APCs could not be determined.

Presence of APCs did not influence the PA size by the time of BCPS. However, presence of APCs was associated with underdevelopment of PA at TCPC, especially in patients who needed interventional closure of APCs.