Potts Shunt in a Child With End-Stage Pulmonary Hypertension After Late Repair of Ventricular Septal Defect

Introduction

Because of improvements in diagnosis and surgical interventions in patients with congenital heart defects, the number of patients with secondary pulmonary arterial hypertension (PAH) has decreased over the past decades. Patients with Eisenmenger physiology, that is, PAH with a right-to-left shunt, usually fare better than patients with idiopathic PAH, ^1–3 who have a dismal prognosis with mean survival of only a few years from diagnosis despite recent improvements in medical treatment with prostanoids, sildenafil, and endothelin antagonists. ¹ Patients with isolated, severe PAH and right heart failure will die of the disease if not treated with a lung transplantation (LTX) or heart–lung transplantation. However, the half-life of pediatric lung recipients is only 5.5 years. ²

The superior prognosis in children with Eisenmenger physiology compared to that in children with idiopathic PAH has partly been explained by the differences in right ventricle (RV) function. In a patient with Eisenmenger, the unrestrictive right-to-left shunt precludes the development of suprasystemic RV pressure and maintains a reasonable RV function. The right-to-left shunt also allows the preservation of systemic cardiac output in the presence of RV failure. ³

The Potts shunt is an anastomosis between the left pulmonary artery (LPA) and the descending aorta (Figure 1). This type of shunt was first reported in 1946 in children with cyanotic congenital heart defects, working as a systemic-to-pulmonary shunt, that is, a left-to-right shunt. ⁴

OPEN IN VIEWER

Figure 1.

Illustration of a Potts shunt.

In this case report in a patient with PAH and a surgically closed large perimembranous ventricular septal defect (VSD), we describe the use of the Potts shunt, not as a left-to-right shunt but as a right-to-left shunt. The circulation changes from that of PAH physiology to an Eisenmenger physiology, like the one found in patients with persistent arterial duct who develop right-to-left shunting.

Case Report

The boy was born prematurely on week 31 and had respiratory distress syndrome with subsequent development of bronc-hopulmonary dysplasia (BPD). During his first year of life, he had repeated pulmonary infections and had a habitual low arterial oxygen saturation (SaO₂) of 86% to 92% despite oxygen supplementation. The poor pulmonary condition of the boy was presumed to be caused by BPD alone, but at age 2 when his clinical condition started to deteriorate, it was decided to perform transthoracic Doppler echocardiography (TTE). The TTE revealed a large subaortic VSD with a right-to-left shunt, RV hypertrophy, dilatation, and pulmonary hypertension. A right heart catheterization showed a perimembraneous VSD with right-to-left shunting, systemic RV pressure, and PAH (Table 1). The findings were felt to preclude VSD closure. Treatment with endothelin receptor-selective antagonist (bosentan) was initiated but without any significant effect. After one month, a loose pulmonary artery banding (PAB) was performed with a decrease in pulmonary artery pressure from 80% of systemic pressure to 66% of systemic pressure, a drop in SaO₂ from 92% to 75%, and decrease in hemoglobin from 9.5 to 8.8 mmol/L. The TTE showed a peak gradient of 55 mm Hg across the band, which had decreased to 45 mm Hg upon discharge one week later. Bosentan treatment was continued. Further tightening of the PAB was not performed because of low SaO₂ varying from 60% to 80%. After 10 months, the RV pressure had increased to a suprasystemic level, and the VSD was starting to become restrictive (Table 1). At this point, debanding was necessary to avoid RV failure. The pulmonary artery was debanded at age 3, one year after banding, and the VSD was closed with a flap valve patch. The flap valve patch was mde of 0.4-mm thick PTFE and was sutured onto one-third of the circumference of the VSD. A loose thread was sewed on to the unsecured part of the flap thus allowing an opening of approximately 4 mm. The boy was discharged with SaO₂ of 80% to 90% on room air, with a combination therapy of sildenafil and bosentan. A lung biopsy, taken during the operation, showed no pathological changes. The function of the flap valve patch persisted only for nine months at which point no shunting through the VSD was observed on TTE. The clinical condition was stable in New York Heart Association (NYHA) class 2B during the following seven years.

OPEN IN VIEWER

Table 1.

Hemodynamic Characteristics Before and After Banding of the Pulmonary Artery.

Age	Procedure	LV Pressure, mm Hg	RV Pressure, mm Hg	PA Pressure, mm Hg	PVR (Wood Units)
1 year 10 months	Cath	110/6	87/8	84/38	23
1 year 11 months	Banding of pulmonary artery
2 years 9 months	Cath	97/7	107/7	70/36	11
3 years 1 month	Debanding of pulmonary artery and VSD closure with flap valve patch

Abbreviations: LV, left ventricle; RV, right ventricle; PA, pulmonary artery; PVR, pulmonary vessel resistance; Cath, right heart catheterization; VSD, ventricular septal defect.

At 10 years of age, the boy experienced a rapid decline in his functional status with NYHA class decreasing from 2B to 3. TTE showed a dilated right atrium and RV with a decreased function and septum deviation compromising the left ventricle (LV), moderate tricuspid and pulmonary valve insufficiency, and completely closed VSD. Prostanoid (flolan) infusion and later treprostinil (remodulin) infusion were initiated with minimal improvement. During the next two months his functional status worsened, he became confined to his bed, and experienced frequent episodes of nausea and vomiting, interpreted as signs of cardiac failure. Repetitive TTE's showed progressive RV failure and increasing pulmonary hypertension, with SaO₂ of 82% to 85%. The clinical condition was complicated by pneumonia and sepsis, and the boy was admitted to the intensive care unit where he required tracheal intubation and mechanical ventilatory support. After successful treatment of the sepsis and extubation of the trachea, his RV function was unchanged, and the N-terminal prohormone brain natriuretic peptide (NT-proBNP) values were rising from 14.7 pmol/L to a peak value of 461 pmol/L.

It was decided to create a Potts shunt to reduce the RV afterload. The shunt was made through a left thoracotomy in the fifth intercostal space. The descending aorta was exposed 10 cm distal to the subclavian artery, and the LPA was exposed from the bifurcation to a point distal to the first lobar branch. The LPA was clamped centrally, and the proximal descending aorta was isolated between two vascular clamps to make room for a large anastomosis (Figure 2). A 13- to 15-mm long opening was made in both the LPA and the descending aorta. The two vessels were anastomosed in side-to-side fashion with a running 4-0 polypropylene. The boy was weaned from mechanical ventilation 19 hours postoperatively, and the postoperative course was uneventful. The boy continued treatment with sildenafil, treprostinil, and bosentan and was discharged two weeks later. Vomiting and loss of appetite were present only during the first two months after the operation and disappeared when the hemoglobin reached levels of around 10 mmol/L. The clinical condition improved markedly, 13 months after the operation. The boy is stabile and is able to ride his bicycle to and from school every day, but his lower body becomes cyanotic when engaging in more strenuous activities. His upper body SaO₂ is 93% and lower body SaO₂ is 73%, and there have been no symptoms of upper body cyanosis since two months after the operation. His NT-proBNP level has decreased to 10 pmol/L, and hemoglobin has increased from 7.5 to 10.2 mmol/L. Improvement was also noted on TTE, where the LV was less compressed by the RV (Figure 3).

OPEN IN VIEWER

Figure 2.

Illustration showing site of side-to-side anastomosis between the left pulmonary artery and the descending thoracic aorta.

OPEN IN VIEWER

Figure 3.

Transthoracic echocardiogram showing parasternal long-axis view of the heart in end-diastolic phase. A, Before Potts shunt operation. B, One month after Potts shunt operation. LV, left ventricle; RV, right ventricle.

Discussion

This is the first case report that describes significant clinical improvement after performance of a Potts shunt in a child with end-stage PAH after late closure of a VSD. One earlier article reported similar improvement in two patients with transposition of the great arteries that had undergone arterial switch surgery, ⁵ and two other reports had similar findings in 10 children with idiopathic pulmonary hypertension. ^6,7

In our case, the boy had progressive RV failure with NT-proBNP values increasing from 14.7 to 461 pmol/L through five months in spite of maximum pharmacological treatment. We considered two different treatments: decompression of the RV by providing an excess flow valve at intra- or extracardiac level that is converting the patient’s physiology into Eisenmenger physiology or transplantation. An atrial shunt is creation of an atrial septum defect that could present a risk of uncontrolled flow of deoxygenated blood from the pulmonary circulation to the systemic circulation causing serious hypoxemia. The same scenario could be expected if a shunt was made at ventricular level. However, creating an extracardiac shunt, such as the Potts shunt, between the systemic and the pulmonary circulation would exclusively shunt cyanotic blood to the lower body and maintain normal or near-normal SaO₂ in the upper part of the body. Patients with Eisenmenger physiology are reported to live longer and to have less RV failure than patients with primary pulmonary hypertension. ^3,8 In our case, the improvement in RV function was demonstrated by both the TTE and the decrease in NT-proBNP values, after surgery. The final therapeutic option for pharmacological nonresponding pulmonary hypertension is LTX. The organ shortage and the limited life span, with a half-life of only 5.5 years, make this option a last resort, which should be postponed until there is no other alternative. ² In our case, the Potts shunt immediately reduced the afterload on the RV, thus avoiding RV failure, as shown by others. ^5–7 The subsequent option for transplantation is not prohibited, although there may be an increased risk compared to a standard transplantation, in view of the need to takedown the shunt at the time of transplantation. However, there are no reported experiences in the literature of LTX or heart–lung transplantation in patients with a Potts shunt. The survival rate for patients with Eisenmenger is 42% at 25 years and is therefore much more favorable than that of patients which is 50% at 5.5 years. ^2,3 In the present case, our patient benefited from the Potts shunt and experienced clinical improvement. There might be a future potential in treating patients with RV failure and severe PAH with the Potts shunt, thereby postponing the time of transplantation.