Potts Shunt and Pulmonary Hypertension
Typically, creation of a post-tricuspid shunt involves an anastomosis between the descending aorta and the left pulmonary artery, a surgical technique first described in 1946 by Dr. Willis Potts.14 The original goal of his eponymous shunt was to provide pulmonary blood flow to a child with tetralogy of Fallot at a time when cardiopulmonary bypass was not available. The procedure fell out of favor once primary repair became standard. However, in 2004, a group from Hôpital Necker-Enfants Malades in Paris, France resurrected the Potts shunt to treat two boys suffering from severe, suprasystemic right ventricular systolic pressures.15 Their insightful theory postulated that by creating an unrestrictive pulmonary-to-systemic shunt, the afterload on the right ventricle would be reduced to systemic levels and thereby recapitulate the pathophysiology seen in patients with Eisenmenger’s syndrome. By doing so, the researchers hoped to confer upon these children the relatively superior life expectancy purported for patients with Eisenmenger’s syndrome.16,17 One immediate consequence of the shunt would be to send deoxygenated blood to the lower half of the body, while preserving the highest oxygenated blood for the upper half and the brain in particular. This contrasts with an AS that leads to global deoxygenation. The two children experienced significant improvement both clinically and hemodynamically with follow-up at 6 and 18 months respectively.
This initial success led the way for their publication of 24 consecutive children with PH treated with a Potts shunt in 2012.18In this series, children’s ages ranged from 1.5 to 17 years with the majority classified as severely compromised with WHO-FC IV. Nineteen of the children had a traditional shunt via a left thoracotomy and direct anastomosis of the pulmonary artery and the aorta. Five had transcatheter shunts through the stenting of a restrictive patent ductus arteriosus (PDA). There were 3 procedural deaths (12.5%). Children that went home had a median follow-up of 2.1 years with 7 children having over 8 years follow-up post-procedure. There was rapid clinical and hemodynamic improvement with better WHO-FC (from a median FC IV to II), longer 6 minute walk distances, lower brain natriuretic peptide (BNP) levels and improved echocardiographic findings. Perhaps most remarkable, these improvements occurred while weaning PH medications. Indeed, 6 of the children that were on intravenous epoprostenol pre-operatively were weaned completely off post-shunt. This contrasts with studies of patients undergoing AS for palliation where medical regiments often increased post-procedure.10,13 Overall, this series demonstrated that a child with severe PH, WHO-FC IV on intravenous prostacyclin therapy, could experience a sustained, prostacyclin-free, improved quality of life by means of a Potts shunt.
In 2016, a group from Washington University in St Louis, USA published a series of five consecutive children who had undergone a surgical Potts shunt for severe PH.19 Their findings mirrored that of the Parisian group, demonstrating in 4 of the children, with less follow-up (mean ~10 months), an improvement in functional class and BNP levels. Since that publication, the St Louis group has performed a total of 23 Potts shunts in children with a mean age of 10.3 years (range 0.3-20 years) with diagnoses including idiopathic PH, heritable PH, PH associated with congenital heart disease and PH associated with lung disease (personal observation). Twenty had surgical shunts, either through 1) direct anastomosis, 2) the use of a short inter-arterial graft or 3) with a valved conduit placed via a sternotomy. Three children had catheter-based interventions, two with PDA stenting and one with reopening a previously device-closed duct. Four of the children died (17%) within 30 days of the procedure. Three of these early mortalities were children who presented in extremis, requiring veno-arterial ECMO prior to their shunt. The fourth child died from progressive hypoxia due to interstitial lung disease. Nineteen of the children were discharged home with a mean follow-up of 2.5 years (range 0.5-6 years). Seventeen of these children showed improvement clinically with lower WHO-FC scores, lower BNP levels and a reduction in overall PH medications. There were four late deaths and two lung transplants post-shunt. Reasons for death or transplant included worsening right ventricular failure, acute infection and one due to accidental tracheostomy tube dislodgement.
The data from Paris and St Louis demonstrate, with a combined mean follow-up of 2-3 years, that a Potts shunt can provide symptomatic improvement in children with severe PH. The question of longer-term benefits, however, remains open. The original idea behind utilizing a Potts shunt was to “convert” a child to Eisenmenger-like physiology and possibly confer the purported survival benefits seen with Eisenmenger patients. This conferral of benefit assumes that a child with PH due to an etiology distinct from Eisenmenger’s syndrome would respond similarly to right ventricular decompression. In adult series, median survival with Eisenmenger syndrome has been estimated as long as 40-60 years, far better than that reported in children with PH.2,3,16,17 The survival rate for children with Eisenmenger’s syndrome, however, may not be as robust as seen in adults, with one estimate of 77% survival at four years.20Nonetheless, even this reduced survival projection would still compare favorably to that of children who typically present for Potts shunt consideration, specifically, those who are WHO FC IV despite PH specific therapy. These compromised children have a purported median survival of less than two years.21
To address survival benefits and other aspects regarding the use of a Potts shunt in treating children with PH, an international registry with at least fourteen centers has begun to compile data. Even without registry results, however, there are relevant observations and discussions that can be made regarding the Potts shunt and PH. From the St. Louis experience, three children were sick enough prior to their shunt procedure to require veno-arterial ECMO (personal observation). All three had surgical shunts and all three died in house within 30 days of their procedure. These results would suggest that in the severely decompensated child, a Potts shunt might not be an effective option. In fact, if these three children were excluded, acute mortality in the St Louis series would drop from 17% to less than 5%. Thus, as with most procedures, increased experience in patient selection can lead to improved surgical outcomes. Certainly, if a child’s anatomy allows, dilating and stenting a restrictive PDA would be the route of choice for creating a shunt, avoiding the potential morbidity from surgery. Unfortunately, this option is seldom available. Several centers have reported on creating a transcatheter shunt through direct puncture from the aorta to the pulmonary artery with subsequent placement of a covered stent.22,23 Between the two series, a total of ten patients underwent this transcatheter procedure with four deaths, an early mortality more than double of that seen in the St Louis and Paris experience. This relatively high risk may preclude the transcatheter Potts shunt from gathering widespread acceptance.
Not all children with severe PH will benefit from a Potts shunt. In theory, because the shunt is an unrestrictive communication between the pulmonary and systemic circulations, it will only be effective in reducing right ventricular afterload in those whose right ventricular systolic pressures are higher than the left, i.e., suprasystemic. Documenting suprasystemic right ventricular pressures can be done with either echocardiography or cardiac catheterization. Relying on right ventricular ejection fraction alone, however, is not sufficient. In the St Louis series, one child had a calculated right ventricular ejection fraction by MRI of 16% yet the echocardiogram still demonstrated suprasystemic pressures. Post-procedure this child did well clinically (personal observation). While a Potts shunt can successfully lower suprasystemic right ventricular pressures, subsequent pressures will still remain at abnormally high systemic levels. Consequently, a shunt placed in a child whose right ventricle has lost the ability to generate systemic to suprasystemic pressures may prove ineffectual. As a case in point, one child, early on in the St Louis series, was diagnosed initially with suprasystemic right ventricular pressures.19 At the time of his shunt however, echocardiograms demonstrated at most systemic right-sided pressures. The child failed to improve post-procedure and eventually died three months later from progressive right ventricular failure. Thus, intervening with a Potts shunt should occur before the onset of frank right ventricular failure, i.e., falling, sub-systemic, right ventricular systolic pressures with rising RVEDP.
In theory, a post-tricuspid shunt like a Potts has the potential to intercede relatively early in a child’s disease course, while systolic function is still preserved at suprasystemic levels and diastolic pressures are comparatively low. This contrasts with an AS which relies upon an elevated RVEDP to be effective, a situation more typical of end-stage disease. This potential advantage of a Potts shunt over AS, leads to the interesting question as to how early in a child’s disease course does one consider a shunt. Current recommendations suggest considering a Potts shunt in a child with severe, symptomatic disease who has failed aggressive medical therapy, including prostacyclin infusion.11 However, does waiting for such a child to develop symptoms reduce a shunts potential benefits and increase its risks? Those answers are currently unclear. What is clear, however, is that managing a 24-hour prostacyclin infusion places a substantial burden upon a child and family and that a Potts shunt offers the hope of easing that onus. Thus, perhaps intervention should be considered in a child with severe PH who is still WHO FC I-II rather than III-IV. A technical development that might provide impetus for employing a Potts shunt earlier in a child’s course is the use of a valved conduit to create the inter-arterial anastomosis. A shunt that allows for only right-to-left flow might benefit the child who at rest has systemic to even sub-systemic right ventricular pressures yet with exertion develops suprasystemic pressures with resulting symptoms. A recently published series demonstrates the feasibility of such an approach.24 Managing the inevitable development of valve incompetency and a conduit’s limitations in a growing child would need to be considered.
For children with severe, progressive, medically recalcitrant PH, bilateral lung transplantation remains the ultimate palliative option. Whether a Potts shunt can provide a similar survival rate as lung transplantation requires more data. A more immediate question, however, is whether having a Potts shunt precludes further consideration for lung transplantation. There is one reported case from the Paris series of a child who had a Potts shunt and subsequently underwent successful lung transplantation.25 This child initially did well post-shunt with clinical improvement and weaning from her intravenous epoprostenol. However, six years after her shunt, she developed severe, persistent hemoptysis and was listed for transplant. On the day of her transplant, her shunt was closed with a covered stent prior to her operation. From the St Louis experience, three children post-shunt came to transplant, two from progressive RV failure and one for hemoptysis (personal observation). At the time of transplant, two were on relatively high doses of IV prostacyclins and one was on veno-venous ECMO. All three transplants were marked by severe bleeding which resulted in two of the children dying post-operatively. The one survivor recovered fully. Reasons for the profound bleeding are likely multifold and include the known risk that prior thoracotomies have on chest wall neo-vascularization and lung adhesions, failure to adequately control the shunt at the time of transplant, and clotting derangements due to medications such as prostacyclins. One technical modification that might reduce these complications is to place a shunt through a sternotomy. This approach could potentially reduce lung adhesions, provide better control of the shunt at the time of transplant, and eliminate the need for any additional peri-operative procedures.24Overall, while having a Potts shunt does not prevent future consideration of lung transplantation, it does appear to increase the risk and technical challenge of the operation. Lung transplantation in of itself is a complex procedure. Consequently, performing one on a child who has had a Potts shunt is not something every transplant center is willing to consider. Therefore, before proceeding with a Potts shunt, frank discussions with families about possible future limitations and risks would obviously be important in helping them make informed decisions.