The Bilateral Bidirectional Glenn Operation as a Risk Factor Prior to Fontan Completion in Patients With Complex Congenital Heart Disease

Keizman and colleagues have published a series of 98 patients who underwent Fontan palliation. ¹ Eighteen patients of those patients underwent prior bilateral bidirectional Glenn (BilatBDG) palliation and had worse outcomes, both short- and long-term, versus patients with prior unilateral bidirectional Glenn (UnilatBDG) procedures. They propose that an etiology for these worse outcomes may be differential pulmonary vascular resistance (PVR) between the lungs. They note that PVR for both lungs is ordinarily calculated as a single unit. Patients with prior BilatBDG palliation often have central pulmonary artery (PA) stenosis, thereby subjecting the venous return from each superior vena cava to a different vascular resistance. This is in distinction to the superior vena cava return in patients with prior UnilatBDG palliation, where the systemic venous return is subjected to the total effective PVR without hindrance from central PA stenosis. Failure to account for this differential PVR may be an etiology for the worse outcomes seen in Fontan procedures in patients with prior BilatBDG palliation.

I found their explanation interesting but not compelling. The assumption presented is that UnilatBDG patients can distribute systemic venous return freely between the two lungs in a ratio that is inversely proportional to the PVR of each lung without limitation from the native or surgically reconstructed anatomy. On the other hand, central PA stenosis in BilatBDG patients limits this flow distribution so each vena cava is subjected to a different PVR. This line of reasoning is not persuasive. BilatBDG patients have two excellent mechanisms to equilibrate flow to each lung based on PVR, similar to the distribution physiology for UnilatBDG patients. One mechanism is the robust intracerebral venous plexus, which is adequate to allow ligation of an internal jugular vein without apparent ipsilateral venous hypertension. This could allow more upper body systemic venous blood to return via the vena cava contralateral to the Fontan, thereby reducing the quantity of flow necessary via the central PA’s. The central PAs, though relatively hypoplastic, represent the other mechanism. In a UnilatBDG patient after Fontan palliation, 50% of the venous return must traverse the central PA’s to distribute flow to each lung (not accounting for size differences between lungs, PVR, etc). On the other hand, the central PA in a BilatBDG patient after Fontan palliation must only carry 30%, assuming 40% upper body venous return. The presence of the intracerebral venous plexus would make this number even smaller. In their own data, the mean ratio of diameters of the central PA’s to the branch PA’s is 0.72. This should be adequate to allow flow to equilibrate adequately based on PVR. The authors do not address the subject of postop transcatheter stent placement in patients with markedly hypoplastic central PA’s, which would be the expectation at my institution, to help to insure that adequate Fontan flow be able to get to each lung.

The concern over differential PVR stated by Keizman and colleagues also does not identify why they would expect this differential PVR to exist. Their patients have had Glenn physiology for nearly 4 years prior to completion Fontan. Presumably, most patients would have little reason to have ever developed differential PVR. Even if there were some reason for differential PVR to have existed at the time of the BilatBDG procedure, for example, a Blalock-Taussig shunt with asymmetrical flow, 4 years of Glenn physiology would likely cause these differences to disappear. The authors did identify one patient, not included in the series of 98 patients, who did have a higher overall PVR by calculating PVR separately for each lung using MRI. Additional information would be needed in order to predict how MRI data would likely affect the cohort overall.

Even if the total volume of systemic venous return can distribute itself between each of the two lungs, there is the question of satisfactory distribution of hepatic venous return to the contralateral lung in order to prevent AV malformations. The presence of a contralateral SVC may direct the hepatic venous flow very much to the ipsilateral lung in some patients. This could potentially explain the greater percentage of patients with mild desaturation in the long run. It does not, however, explain short-term instability.

The authors did point out that there was a marked difference in the incidence of heterotaxy between the unilateral and BilatBDG groups. The presence of heterotaxy as a substrate is more compelling than differential PVR as an etiology for worse outcomes. The increased risk of Fontan patients with heterotaxy is a consistent finding in the literature. Ono and colleagues found heterotaxy as a risk factor for late mortality in a series of 434 Fontan operations BD performed at a single institution between 1994 and 2015. ² In another large single institution series of 500 patients, Nakano and colleagues found that heterotaxy was the only risk factor for mortality in a multivariate analysis. ³ Alsoufi and colleagues reported a series of 67 heterotaxy patients over a ten-year period of time who underwent single ventricle palliation. In-hospital mortality for the first-stage (neonatal) palliation was higher (27% vs 10%) than a matched contemporaneous cohort of nonheterotaxy patients. They did not find that mortality was higher with the Fontan procedure, but only 23 (47%) of 49 patients who underwent second-stage (Glenn) palliation were considered candidates for Fontan. ⁴ Cao and colleagues had similar findings, with only 27 of 70 single ventricle patients with heterotaxy being considered candidates for the Fontan procedure. ⁵ There is evidence that ciliary function is impaired in patients with heterotaxy, ⁶ which is felt to impact results. ⁷ This represents only one noncardiac mechanism for increased risk in patients with heterotaxy. It is difficult to account for other subtle disorders that have not been described or hypothesized.

There are other reasons to take the findings with some caution. The UnilatBDG and BilatBDG study groups were median 4.3 and 4.8 years of age, respectively; their weights were 15.0 kg and 17.5 kg, respectively. This represents an older/larger population than what is experienced at many centers. ^2,3 It is possible that referral patterns within this very particular region of the world may impact surgical timing and the ability to bring patients back for transcatheter interventions postoperatively, though this concern is speculative.

This study does reinforce the additional risk of performing Fontan completion in patients with prior BilatBDG palliation, regardless of the etiology to which one attributes this increased risk. This study should remind the pediatric cardiac community to temper its enthusiasm regarding the long-term outcomes for single ventricle patients and how precarious is the margin of error for patients with this physiology. Some reports imply an optimistic view of the intermediate- to long-term outcomes of Fontan patients, but these may focus on the selected patients who have survived to a successful Fontan. ⁸ Other reports offer a more sober viewpoint. The Single Ventricle Reconstruction Trial demonstrates a transplant-free survival of 62% at six years in a cohort of 549 patients with hypoplastic left heart syndrome. ⁹ Even these results exclude patients with congenital or acquired abnormalities which would independently affect the likelihood of survival. ¹⁰ The Mayo Clinic reported a 62% 20-year survival in patients who had either a lateral tunnel or extracardiac Fontan in a patient population where >50% of patients had LV-dominant morphology. ¹¹

The general treatment algorithm for patients with single ventricle physiology has been largely unchanged for nearly 25 years. It is hard to know what the next “great advance” will be in the surgical treatment of single ventricle physiology, that is, long-term mechanical support, ideal valve replacements, better assessment/treatment of lymphatic permeability, or medical therapy for heart failure. Until that time, however, we will continue to manage these patients with precious little margin for error. We will also continue to find, as did Keizman and colleagues, that very minimal changes in Fontan risk factors will significantly impact results.