Extracardiac Fontan With Direct Inferior Vena Cava to Main Pulmonary Artery Connection Without Cardiopulmonary Bypass

Introduction

Despite significant improvements in the outcomes of the Fontan operation for functionally univentricular hearts, various total cavopulmonary connection (TCPC) designs are fraught with imperfections such as potential thromboembolic complications, lack of somatic growth, and development of stenosis relative to the vessel growth and hemodynamics in the graft. ¹

In the past years, many surgeons have used a variety of materials or techniques in extracardiac TCPC (extracardiac Fontan [ECF]) such as Dacron, expanded polytetrafluoroethylene, homograft, autologous pericardium, and xenopericardium. Several reports have described TCPC involving an anastomosis between the inferior vena cava (IVC) and the main pulmonary artery (MPA), with or without the use of additional pericardium or right atrial wall. ^{2 -6} Of these, only one patient was described as having undergone an MPA to IVC anastomosis along with a bilateral bidirectional Glenn (BDG) that was performed under cardiopulmonary bypass (CPB). ² To the best of our knowledge, there is no report of a direct IVC to MPA connection without using extracorporeal circulation. We describe the application of this technique in an appropriately selected patient and discuss its potential advantages.

Case Report

A six-year-old boy with a diagnosis of double outlet right ventricle with a ventricular septal defect that was clearly unroutable to the aorta and severe infundibular and valvar pulmonary stenosis was scheduled for a primary Fontan operation as the hemodynamics were considered favorable (mean pulmonary artery [PA] pressure 12 mm Hg on cardiac catheterization) and the pulmonary arteries were well developed. A particular point of interest was that the great arteries were malposed, with the MPA positioned behind and to the right of the aorta. There were equally sized right and left superior caval veins without any connecting vein between the two. Prior to sternotomy, monitoring lines were placed in the left- and right-sided internal jugular veins to monitor the pressure in each superior vena cava (SVC), and near-infrared reflectance spectroscopy (NIRS) was used to measure the adequacy of cerebral perfusion. A triple lumen catheter was placed in the right femoral vein to monitor the IVC pressure and for inotropic infusions. Two large bore intravenous lines were placed in the upper extremity veins for rapid volume infusion.

After sternotomy, bilateral SVC and a right-sided IVC were noted. The MPA was behind and to the right of the aorta and was around 2 cm long. The atrial appendages were juxtaposed to the left. The MPA was dissected free from the aorta and both branch pulmonary arteries were dissected free to the hilum of the lungs. The intra-abdominal IVC was mobilized up to the point of entry of the hepatic veins. Both the SVCs were also widely mobilized and dissected free. Heparin (3 mg/kg) was administered, so that if needed CPB could be established rapidly. Each SVC was clamped and divided and was anastomosed to the ipsilateral branch PA in a standard end-to-side fashion, with oversewing of the cardiac end of the SVC. No cannula was used in either SVC for cerebral decompression, as the superior cavopulmonary anastomoses were performed sequentially, and it was anticipated that the other SVC would provide adequate decompression while one SVC was being clamped. This was confirmed by the absence of a significant rise in the SVC pressure as monitored by the pressure line in the SVC and by equal and acceptable NIRS saturations in both cerebral hemispheres.

Following the completion of the bilateral bidirectional superior cavopulmonary anastomoses, an immediate rise in the systemic saturation from a baseline level of 68% to 90% after the BDG was noted by SpO₂ monitoring. The MPA was then divided between clamps placed as close to the pulmonary valve as possible and the proximal stump was closed, taking deep bites to incorporate the leaflets of the pulmonary valve and avoid the creation of a blind cul-de-sac above the valve, which could theoretically increase the risk of thrombus formation.

Optimal exposure of the IVC was achieved by widely opening the left pleura and reflecting the heart into the left pleural cavity after applying a stay suture at the acute margin of the heart. The IVC was now clamped with the intention of observing the degree of rise in IVC pressure and overall hemodynamic response to IVC occlusion. The IVC pressure rose from a baseline of 14 to 20 mm Hg. The systolic blood pressure dropped from 120 to 90 mm Hg. However, there was no hemodynamic compromise, as adequate preload was promoted by administering crystalloids through the upper extremity veins. Vasopressors were kept ready but were not needed. Therefore, without placing a cavoatrial shunt, the IVC was divided between clamps taking care to avoid a circumferential incision in the crista terminalis. The cardiac end of the IVC was oversewn. Then the distal MPA stump was brought down and anastomosed to the transected caudal end of the IVC using a continuous 5-0 polypropylene suture (Figures 1 and 2). One half of the usually calculated dose of protamine was administered to reverse the heparin. The SVC clamp times was 13 minutes on the right side and 15 minutes on the left side, and the IVC clamp time was 9 minutes. There was no hemodynamic instability during clamping of either SVC or during clamping of the IVC. No inotropic drugs were administered, and the patient was uneventfully extubated on the operating table with SVC pressure 14 to 15 mm Hg, IVC pressure 16 mm Hg, and systemic oxygen saturation of 98% on room air.

OPEN IN VIEWER

Figure 1.

Operative sketch. A, External cardiac anatomy. B, Completed reconstruction. L indicates left superior vena cava; LBG, left bidirectional Glenn; MPA, main pulmonary artery; R, right superior vena cava; RBG, right bidirectional Glenn.

OPEN IN VIEWER

Figure 2.

Operative photograph. IVC indicates inferior vena cava; L, left bidirectional Glenn; MPA, main pulmonary artery; P, proximal main pulmonary artery stump; R, right bidirectional Glenn.

Postoperative recovery was uneventful without any suggestion of adverse neurological sequelae and without major pleural effusions. Chest tubes were removed on the eight postoperative day. Serial blood tests indicative of hepatic and renal functions were normal, and liver ultrasound did not demonstrate hepatic congestion. There was no gradient in the Fontan pathway on postoperative echocardiography. Computed tomography angiogram (Figure 3) prior to hospital discharge demonstrated laminar flow through the entire Fontan circuit.

OPEN IN VIEWER

Figure 3.

Postoperative computed tomography (CT) angiogram. 1 indicates right bidirectional Glenn; 2, left bidirectional Glenn; 3, main pulmonary artery; 4, inferior vena cava.

Discussion

This particular modification of the ECF may confer several advantages. First and foremost, it obviates the need for lifelong anticoagulation. Because it is constructed entirely of native tissue with naturally endothelialized surface throughout, it may prevent the formation of an intimal peel that is not uncommonly observed in prosthetic extracardiac conduits. ³ No interposition graft was used, and therefore, the growth potential is preserved with this kind of connection that may obviate the need for a future conduit replacement. In addition, as we have demonstrated earlier, ⁷ performing this procedure completely off CPB may avoid the harmful effects of CPB and lead to substantial savings in hospital stay and costs.

In addition, this modification avoids intra-atrial suture lines which may be related to the cumulative risk of arrhythmia. ⁴ Utilizing the MPA as the site for IVC connection may contribute to balanced distribution of blood flow to both lungs and may reduce the incidence of pulmonary arteriovenous malformations. This is contrary to the conventional TCPC designs where it is claimed that the energy efficiency is better ⁸ ; however, balanced flow to both lungs is also an important consideration.

A detailed literature search identified only five reports describing IVC to MPA connection for TCPC. ^{2 -6} In all these reports, the TCPC was performed with CPB support, and in two of these, ^2,4 pericardium and atrial wall were used to gain length. ^{2 -6} However, the specific anatomy of our patient and extensive mobilization of the PAs and the IVC allowed us to achieve a complete autologous tissue direct IVC-to-PA connection without using any extra material. Hemodynamic instability during the period of IVC clamping was avoided by optimizing the preload by fluid administration through intravenous lines placed in the upper limb veins. In addition, the short period of IVC clamping (9 minutes) obviated the need to administer any vasopressor drugs.

However, caution must be exercised with respect to broader application of this procedure. It may be difficult to achieve a direct anastomosis between the MPA and IVC without using additional material when the MPA is not of adequate length and in reoperations where extensive mobilization of the pulmonary arteries is less feasible. The favorable points in this patient were the length and location of the MPA and the presence of juxtaposed atrial appendages, which provided room for manipulation. A theoretical concern in the long-term follow-up is the distensibilty of the MPA, as it is not a rigid tube. Whether this would lead to the loss of energy in the Fontan circuit (as may occur in the setting of an atriopulmonary Fontan connection) is a matter of speculation. We plan to perform computational fluid dynamic studies to assess this further.