| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2005;80:1604-1609
© 2005 The Society of Thoracic Surgeons
a Department of Cardiology, Children's Hospital, Boston, Massachusetts
b Department of Cardiac Surgery, Children's Hospital, Boston, Massachusetts
c Department of Pediatrics, Harvard Medical School, Boston, Massachusetts
d Department of Surgery, Harvard Medical School, Boston, Massachusetts
Accepted for publication May 9, 2005.
* Address correspondence to Dr McElhinney, Department of Cardiology, Children's Hospital, 300 Longwood Ave, Boston, MA 02115 (Email: doff.mcelhinney{at}cardio.chboston.org).
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsCommentReferences |
|---|
METHODS: We studied 11 patients with PAVMs who underwent creation of a brachial AV fistula a median of 11 years after CPA.
RESULTS: Eight patients had discontinuous pulmonary arteries or unilateral flow of a bidirectional CPA and were not considered good candidates for Fontan completion; the other 3 patients had polysplenia and unilateral hepatic venous streaming after Fontan completion. Three patients died of progressive complications of their heart disease 4 to 18 months after AV fistula creation. Pulmonary arteriovenous malformations resolved after creation of a brachial AV fistula in 4 of 5 surviving patients with unilateral flow of a superior CPA, but in none of 3 patients with polysplenia who had unilateral hepatic venous streaming after Fontan completion and PAVMs in the contralateral lung.
CONCLUSIONS: These findings are consistent with the "hepatic factor" hypothesis, according to which the development of PAVMs is facilitated when an unidentified factor produced or metabolized in the liver does not reach the pulmonary circulation before traversing another capillary bed. Patients with unilateral superior CPA flow and PAVMs who are not considered candidates for Fontan completion may benefit from a brachial AV fistula.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
Although the cause of PAVMs is unknown, there is increasing evidence to support the theory that the development of PAVMs is facilitated when an unidentified factor produced or metabolized in the liver does not reach the pulmonary circulation before traversing another capillary bed [3, 4, 9, 12]. The "hepatic factor" hypothesis is supported by reports of PAVMs apparently resolving in patients with a bidirectional CPA after diversion of hepatic venous effluent to the pulmonary arteries, including a recent series from our center [4, 7–9]. The hepatic factor hypothesis is also consistent with the observation that PAVMs may persist in one lung after Fontan completion in patients with a prior Kawashima circulation if there is streaming of hepatic venous blood to the other lung, a finding that suggests resolution of PAVMs only in the lung receiving hepatic venous effluent [4, 13].
Creation of an upper extremity systemic arteriovenous (AV) fistula (usually brachial or axillary) to augment pulmonary blood flow and systemic oxygenation in patients with a CPA was first proposed by Glenn in 1972 [14]. Since that report, the clinical effects of this procedure have been described only in a few small series and case reports, which have demonstrated modest improvement in hypoxemia after AV fistula creation in patients with a CPA who are not considered candidates for Fontan completion [2, 15–17]. However, there is no published information on the effect of an upper extremity systemic AV fistula on PAVMs in patients with a CPA. In the Toronto series reported by Magee and colleagues [15], for example, patients with PAVMs after CPA were specifically excluded from creation of an axillary AV fistula owing to concerns that augmentation of blood flow to the lung with PAVMs might lead to increased intrapulmonary right-to-left shunting [15]. We have created a brachial AV fistula in selected patients with a CPA and hypoxemia, in most cases as an attempt to deliver hepatic venous blood to the affected lung in patients with PAVMs. Our hypothesis has been that an upper extremity systemic AV fistula will allow systemic arterial blood containing the putative hepatic factor to bypass the systemic capillary network and perfuse the PAVM-containing lung supplied by the CPA.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
Cavopulmonary Circulation at the Time of AV Fistula Creation
All patients had undergone prior unidirectional or bidirectional CPA, at a median age of 1.4 years (range, 4 months to 13 years). The original CPA was unidirectional in 7 patients and bidirectional in 4. Three of the patients who originally underwent bidirectional CPA had polysplenia syndrome, interruption of the inferior vena cava with azygous continuation to the SVC, and bilateral SVCs. These patients underwent bilateral bidirectional CPA, with a Kawashima procedure on the side of azygous return [5]. Three patients with an original unidirectional CPA had been converted to a bidirectional CPA, but CPA flow was primarily or exclusively unilateral in all 3, 2 of whom also had a systemic-to-PA shunt to the lung not receiving CPA flow. One of these patients originally had a unilateral CPA, underwent a failed attempt at Fontan completion followed by acute take-down of the Fontan to a bidirectional CPA, and had hypoplasia of the left PA and almost completely unilateral CPA flow to the right PA. One patient with an original bidirectional CPA had developed discontinuous PAs. All 3 patients with polysplenia syndrome and a Kawashima procedure had undergone Fontan completion, with incorporation of hepatic venous effluent into the cavopulmonary circulation by means of an intraatrial lateral tunnel or an extracardiac hepatic vein-to-PA conduit. In all 3 of these patients, there was documented unilateral streaming of hepatic venous flow contralateral to the lung with PAVMs (Fig 1).
|
In 6 patients, systemic-to-pulmonary venous collaterals were identified and embolized at the time of catheterization before AV fistula creation.
Diagnosis of PAVMs
The diagnosis of PAVMs was made at catheterization a median of 10.5 years (range, 6 months to 29 years) after CPA. Pulmonary arteriovenous malformations were diagnosed on the basis of the following criteria in all patients: (1) rapid pulmonary AV transit (fewer than 3 heart beats) of contrast on proximal PA angiography; (2) a typical reticular or spongy pattern in the peripheral pulmonary vasculature on PA angiography (Figs 1 and 2);
and (3) pulmonary venous desaturation (
92%) in angiographically affected lung segments.
|
Prior Interventions for PAVMs
Interventions for PAVMs before AV fistula creation had been performed in 7 patients, including coil embolization of PAVMs in 4 patients, attempted reconstruction of central PA continuity in 4 patients with discontinuous PAs, hepatic vein inclusion (Fontan completion) in 2 patients with polysplenia and PAVMs after a Kawashima procedure [4], and stenting of the central PAs in 2 patients with polysplenia syndrome who had unilateral streaming of hepatic venous flow after Fontan completion. The 2 patients who underwent Fontan completion after a Kawashima procedure had bilateral PAVMs before hepatic vein incorporation, and resolution of PAVMs in only 1 lung, presumably because of hepatic venous streaming to that lung [4].
Brachial AV Fistula Creation
The median age at AV fistula creation was 14 years (range, 5 to 32), and the median duration from CPA to AV fistula creation was 11 years (range, 5 to 31). Angiography of the upper extremity vasculature was performed at the time of catheterization before AV fistula creation. The AV fistula was created in the right arm in 7 patients and the left arm in 4, and was ipsilateral to the lung with PAVMs in the 3 patients with bilateral bidirectional CPA and unilateral streaming of hepatic venous blood. The brachial artery and basilic vein were accessed through a small incision just beneath the biceps muscle. The vessels were dissected free, isolated, and controlled with vessel loops. After administration of heparin, the artery and vein were clamped. Longitudinal arteriotomy and corresponding venotomy incisions approximately 4 mm in length were made, and a side-to-side anastomosis was created with 7-0 nonabsorbable continuous suture. The vascular clamps were removed and flow in the artery, vein, and through the AV fistula assessed. The basilic vein was not ligated distal to the fistula. Hemostasis was achieved and the incision was closed.
Survival
There were no early deaths. Three patients died during follow-up, 4 to 18 months after AV fistula creation, owing to progressive complications of their heart disease.
Status of PAVMs
The 8 surviving patients were followed up for a median of 4.3 years (range, 1.5 to 7.1). Seven of these patients underwent cardiac catheterization during follow-up, 2.0 to 5.7 years (median 4.0) after AV fistula creation. In 4 of these 7 patients, all with unidirectional CPA flow (either discontinuous PAs or unilateral flow of a bidirectional CPA), the previously documented PAVMs had resolved (normal angiography and pulmonary venous saturation; Fig 2). The single patient with discontinuous PAs who had documented persistence of PAVMs after AV fistula creation had multiple large PAVMs in the affected lung. In the other 3 catheterized patients, including both surviving patients with polysplenia and unilateral hepatic venous streaming after Fontan completion, the PAVMs were still present. The eighth surviving patient did not undergo follow-up catheterization, but remained hypoxemic and required frequent erythropheresis for symptomatic polycythemia. At most recent follow-up, systemic arterial oxygen saturation ranged from 76% to 92% among the 4 patients in whom PAVMs had resolved, and 80% to 84% in the 3 patients with documented persistence of PAVMs.
Of the 3 patients who died, 1 (with polysplenia and hepatic venous streaming) underwent catheterization during follow-up, with confirmed persistence of PAVMs, whereas the other 2 did not undergo follow-up catheterization or autopsy.
Status of AV Fistula
The AV fistula was taken down (n = 4) or documented to have occluded spontaneously (n = 2) in 6 patients, 1 of whom has since died. In 3 patients, the AV fistula was taken down surgically at the time of Fontan completion (n = 2) or at the time of a hepatic vein-left PA conduit (see below). In 1 patient, the AV fistula was occluded with a polytetrafluoroethylene-covered self-expanding stent (WallGraft; Boston Scientific, Natick, Massachusetts) before placement of a hepatic vein-right PA conduit to compensate for unilateral hepatic venous streaming (Fig 3). In 2 patients, the AV fistula was found at follow-up catheterization (2 and 5.7 years after the AV fistula) to have occluded spontaneously. Pulmonary arteriovenous malformations had resolved in both of these patients. In patients with a patent AV fistula on follow-up catheterization, the axillary artery was dilated, tapering to normal caliber beyond the fistula. The venous system receiving fistula flow was dilated and often tortuous (Fig 3). The AV fistula remains patent in 2 of the 8 surviving patients: 1 of these patients has persistent PAVMs, the other has not been studied by catheterization, but requires frequent erythropheresis for symptomatic polycythemia.
|
Reinterventions
Aside from AV fistula take-down, interventions during follow-up included Fontan completion in 2 patients in whom PAVMs resolved after AV fistula creation, and placement of hepatic vein-to-PA conduits in 2 patients with polysplenia and unilateral hepatic venous streaming after prior cavopulmonary incorporation of the hepatic veins (16 and 27 months after AV fistula creation). Before AV fistula creation, the 2 patients who ultimately underwent Fontan completion had unilateral flow of a bidirectional CPA, but PA flow was normally distributed by the time PAVM resolution was documented. Of the 2 patients who underwent hepatic vein-to-PA conduits, 1 subsequently died and the other has been followed up for only 3 months and has not been recatheterized.
|
Comment |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
These findings are consistent with the hepatic factor hypothesis, according to which the development of PAVMs is facilitated when an unidentified factor produced or metabolized in the liver does not reach the pulmonary circulation before traversing another capillary bed [3]. In patients with unilateral superior CPA flow, hepatic venous effluent returns to the heart, is ejected to the systemic circulation, and flows through the brachial AV fistula to the SVC, the CPA, and the PAVM-containing lung without crossing a capillary bed. Pulmonary arteriovenous malformations resolved after AV fistula creation in 4 of 5 surviving patients with this circulation. In contrast, in patients with unilateral streaming of hepatic venous blood after incorporation into the cavopulmonary circulation, hepatic venous blood traverses the capillary bed of the lung it perfuses before returning to the single ventricle and being ejected to the systemic circulation. Thus, blood traveling through the AV fistula has already crossed a capillary bed (ie, the contralateral pulmonary capillary bed) before returning to the SVC and the PAVM-containing lung. Pulmonary arteriovenous malformations did not resolve after AV fistula creation in any of the 3 patients with this circulatory arrangement. These observations support the following inferences: (1) development of PAVMs is facilitated by exclusion from the pulmonary circulation of a substance produced or metabolized in the liver; (2) resolution of PAVMs after CPA requires delivery of hepatic venous blood, and the putative hepatic factor, to the affected lung; and (3) the putative hepatic factor either has a short circulating half-life or is eliminated/inactivated during passage through a systemic or pulmonary capillary bed.
The support these findings lend to the hepatic factor hypothesis complements prior reports demonstrating resolution of PAVMs after hepatic vein inclusion in patients with post-Kawashima PAVMs [4, 7–9], after heart transplant [12], and after liver transplant in patients with the hepatopulmonary syndrome [18]. These clinical studies, however, shed no light on the specific mechanisms by which exclusion of hepatic venous blood predisposes to PAVM formation, or by which readdition of hepatic venous effluent may cause regression of PAVMs.
Several recent studies in humans and animal models have characterized pulmonary vascular changes and functional alterations in the pulmonary vasculature after CPA, which may lead to an understanding of the mechanisms of PAVM formation. In a lamb model of right unidirectional CPA, Malhotra and coworkers [19–21] demonstrated downregulation of angiotensin-converting enzyme, upregulation of angiotensin II receptors, and upregulation of oxidative stress-related genes in the PAs, alterations that were independent of decreased pulmonary blood flow. The same group [22] found that CPA leads to upregulation and intimal-medial redistribution of cMet, a receptor for the hepatocyte growth factor that activates the antiapoptotic factor Bcl2, which was interpreted as part of a genetic response to CPA that promotes pulmonary endothelial cell survival. In a separate study, Ikai and colleagues [23] observed attenuation of hypoxic pulmonary vasoconstriction in an animal model of CPA, which they suggested may be due to altered vasoactive signaling pathways, similar to those implicated in the formation of PAVMs. Another group has described a rat model of CPA that reproduces the alterations in pulmonary microvessel density and dilation that they observed in humans with a bidirectional CPA and PAVMs, in whom alterations in angiogenic protein levels were also detected [9, 24–26].
In patients with a bidirectional CPA and unilateral PAVMs, the resistance in the lung with PAVMs is likely to be substantially lower than in the contralateral lung, although the magnitude of this difference may be difficult to assess. Unilateral or unbalanced flow of blood from the bidirectional CPA in this situation may be due to low resistance in the lung with PAVMs, high resistance in the contralateral lung, or a combination thereof. Among the patients in this study, 3 had unilateral PAVMs in the setting of a bidirectional CPA. All three patients originally had a unidirectional CPA, which was converted to a bidirectional CPA 1 to 30 years later. These patients also had unidirectional CPA flow. In 2 of these cases, creation of a brachial AV fistula led to resolution of the PAVMs and to normalization of pulmonary blood flow distribution, suggesting that low flow to the contralateral lung was due primarily to PAVMs. This is an encouraging finding, as it demonstrates that patients with maldistributed pulmonary blood flow and PAVMs may not have elevated resistance in the lung receiving less flow, and may become better Fontan candidates if the PAVMs can be treated effectively.
Technical and Therapeutic Considerations
With respect to creation of an AV fistula in patients with PAVMs after CPA, there are several therapeutic and technical considerations worth noting. First, although most prior reports have described creation of an axillary AV fistula [2, 14, 15], we prefer to use the brachial artery, which is easier to access, and may be more comfortable for the patient if venous dilation results. An AV fistula using the brachial artery may also be subject to less potential distortion than an axillary AV fistula, which is located closer to the shoulder joint. Second, insofar as both subclavian veins drain to the SVC, the sidedness of the AV fistula only matters if there are bilateral SVCs and bilateral CPA (as in the 3 patients in this series with unilateral HV streaming bilateral bidirectional CPA), and it may be worthwhile to consider patient handedness when planning a brachial AV fistula. Third, surgical closure of an AV fistula can entail a difficult and bloody dissection. An alternative means of closing an AV fistula is transcatheter placement of a self-expanding covered stent, which we found to be straightforward and effective. Fourth, it is important to assess for and occlude systemic-to-pulmonary venous collaterals before creation of an AV fistula. In particular, venous collaterals from the venous system ipsilateral to the AV fistula might allow prepulmonary right-to-left shunting of AV fistula flow.
In conclusion, in patients with unilateral superior CPA flow who have PAVMs in the lung receiving CPA flow, but who are poor candidates for a Fontan operation, creation of an upper extremity systemic AV fistula may serve as an effective treatment for PAVMs and may improve their candidacy for successful Fontan completion. In contrast, a brachial AV fistula does not facilitate resolution of PAVMs in patients with unilateral hepatic venous streaming and contralateral PAVMs after a Kawashima procedure and subsequent Fontan completion, and we no longer perform brachial AV fistulas in such patients.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
Related Article
This article has been cited by other articles:
|
|
 |
|
  S. Nayak and P.D. Booker The Fontan circulation CEACCP, February 1, 2008; 8(1): 26 - 30. [Full Text] [PDF]
|
|
|
|
 |
|
 C. S.D. Almond, J. E. Mayer Jr, R. R. Thiagarajan, E. D. Blume, P. J. del Nido, and D. B. McElhinney Outcome After Fontan Failure and Takedown to an Intermediate Palliative Circulation Ann. Thorac. Surg., September 1, 2007; 84(3): 880 - 887. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. B. McElhinney, J. Kreutzer, P. Lang, J. E. Mayer Jr, P. J. del Nido, and J. E. Lock Incorporation of the Hepatic Veins Into the Cavopulmonary Circulation in Patients With Heterotaxy and Pulmonary Arteriovenous Malformations After a Kawashima Procedure Ann. Thorac. Surg., November 1, 2005; 80(5): 1597 - 1603. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
