Author + information
- Received November 14, 2010
- Revision received March 25, 2011
- Accepted March 30, 2011
- Published online July 1, 2011.
- Abdi Jama, MD⁎,
- Michel Barsoum, MD⁎,
- Haraldur Bjarnason, MD†,
- David R. Holmes Jr, MD⁎ and
- Charanjit S. Rihal, MD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Charanjit S. Rihal, Division of Cardiovascular Disease, College of Medicine, Mayo Clinic, 200 First Street South West, Rochester, Minnesota 55905
Objectives This study sought to assess clinical and angiographic outcomes in a series of 29 patients who underwent transcatheter closure of coronary artery fistulae (CAF).
Background Transcatheter closure of CAF has become an alternative to surgical closure, but the reported experience is relatively limited.
Methods Medical records of all patients with CAF who underwent transcatheter closure at the Mayo Clinic, Rochester, Minnesota, between 1997 and 2010, were reviewed. Patients with other complex cardiac lesions and those requiring surgery were excluded.
Results Twenty-nine patients with CAF underwent 36 transcatheter closure procedures. The most were women (55%), and the median age at the time of transcatheter closure was 49 years. Twenty-three patients had a single CAF. The most common presenting symptom was chest pain (52%). Thirty devices were deployed antegrade into 1 or more arterial feeders, 3 using an arteriovenous wire loop and 3 retrograde at the fistulous connection. Successful closure occurred immediately in all patients with no residual flow in 89% and with trivial flow in 11%. Four complications occurred including 2 device migrations, 1 coronary spasm, and 1 coronary thrombosis. A follow-up angiogram was obtained in 18 (62%) patients with a median time to follow-up angiography of 1.5 years. Ten patients (56%) of the 18 patients with follow-up angiography had no recanalization of embolized vessel; 4 patients (22%) had trivial recanalization, and 4 patients (22%) had large recanalization. A repeat closure procedure was performed in all 4 patients of the latter.
Conclusions Transcatheter closure of CAF is feasible and should be considered in carefully selected patients. Recanalization of the treated coronary fistulae can occur, so follow-up angiography or other imaging modality should be performed in these patients.
Coronary artery fistulae are abnormal connections between a coronary artery and a cardiac chamber (coronary—cameral) or major intrathoracic vessel (arteriovenous). Most coronary fistulae are congenital in nature; however, they can also be acquired (1,2). Coronary artery fistulas are rare with a 0.2% to 0.6% incidence in angiographic series and 0.002% overall incidence in the general population (3,4). Although first reported by the German anatomist Wilhelm Krause in 1865, the clinical significance of coronary artery fistulae has not been well understood. Most patients are asymptomatic. When symptomatic, the most common findings are heart failure secondary to volume overload resulting from left to right shunting, ischemia secondary to coronary steal, arrhythmia, fistula rupture or thrombosis, and infective endocarditis (5).
Because of these potential complications, some advocate closure of coronary artery fistulae even in asymptomatic patients (5). The American College of Cardiology/American Heart Association (2008) guidelines for the management of adults with congenital heart disease recommend closure of all large coronary artery fistulae regardless of symptomology using transcatheter or surgical techniques. Guidelines call for closure of small to moderate fistulae only in the presence of symptoms (including ischemia, arrhythmia, and unexplained systolic or diastolic dysfunction). The American College of Cardiology/American Heart Association guidelines do not recommend closure of small, asymptomatic coronary artery fistulae (6). Surgical closure has been the mainstay of treatment; however, Reidy et al. (7), in 1983, first reported transcatheter closure, which has become an alternative treatment in patients with suitable anatomy. Transcatheter closure has several theoretical advantages over open heart surgery, including shorter hospitalization and quicker recovery. A number of transcatheter studies have reported acute procedural success rates but longitudinal follow-up is lacking. The aim of this study was to evaluate acute and late clinical and angiographic outcomes in patients treated with percutaneous transcatheter closure.
The medical records of 29 adult patients with coronary artery fistulae who underwent transcatheter closure at the Mayo Clinic, Rochester, Minnesota, between 1997 and 2010, were retrospectively reviewed. The 29 patients were largely referred to our catheterization laboratory after diagnosis of coronary artery fistulae. Patients were included in the study if they were symptomatic or had large fistulae. Patients with small incidentally found fistulae and those with complex additional cardiac lesions or who required surgery were excluded. Institutional Review Board approval was obtained, and only patients who consented for the procedure and for medical records research were included. Thirty-six transcatheter closure procedures in the 29 patients were undertaken. All patients underwent cardiac evaluations including history, physical examination, laboratory, and electrocardiogram. A sizable minority of patients had echocardiography and stress testing before intervention.
Access was obtained with femoral arterial and venous catheters. Once access was secured, all patients received an intravenous heparin bolus (60 to 100 U/kg). After obtaining hemodynamic data, the coronary arteries were engaged selectively using various diagnostic guides and catheters. Hand injection of contrast allowed visualization of coronary anatomy, and location of fistulae including origin and drainage sites. To access the fistula, guides, microcatheters, and coronary guidewires were used. Multiple different closure devices were used including Tornado coils (Cook Medical, Bloomington, Indiana), Nester coils (Cook Medical), Amplatzer ductal occluder (AGA Medical Corporation, Golden Valley, Minnesota), Amplatzer Vascular Plug (AGA Medical Corporation), Contour Embolization Particles (Boston Scientific, Natick, Massachusetts), and iCast covered stent (Atrium Medical, Hudson, New Hampshire). Devices were selected by the operator based on the size and other characteristics of the fistulae. Coils were primarily used in small- to medium-sized fistulae, whereas the Amplatzer devices were used to occlude larger fistulas. Details regarding the devices and their specific delivery systems have been reported elsewhere (7–11). Devices were deployed either via the femoral artery or femoral vein, using an arteriovenous wire loop (via femoral artery and femoral vein or jugular vein) when necessary. Selective coronary angiography was performed immediately after device deployment to assess presence of residual flow.
Continuous variables were expressed as median (interquartile range) and categorical data as percentage.
Medical records and follow-up tests, if available, including electrocardiograms, echocardiograms, and stress tests were reviewed. Follow-up angiograms were performed in selected patients and were reviewed for presence, rate, and size of fistula recanalization. Subsequent transcatheter closure was performed if the recanalized fistula was deemed significant.
The clinical characteristics of the cohort are summarized in Table 1. Twenty-nine patients with coronary fistulae underwent 36 transcatheter closure procedures. Twenty-three patients had single fistulas, and 6 had multiple fistulae. Most patients were women (55%), and the median patient age at the time of the procedure was 49 (interquartile range: 19 to 71) years. Of the 29 patients, 27 had symptoms. Although chest pain was the most common presenting symptom, occurring in 15 patients (52%), dyspnea was also common and was present in 7 (24%) patients. One patient presented with fevers and was subsequently found to have endocarditis. Two patients were asymptomatic but had large fistulae. A continuous systolic-diastolic murmur was present in 9 patients (31%); and no murmur was heard in 20 patients. Fourteen patients (48%) were receiving antiplatelet therapy before intervention. The most common electrocardiogram findings were nonspecific ST- and T-wave changes, which were present in 24% of the patients. Thirteen patients had echocardiography before the procedure and 9 of those 13 had Doppler and 2-dimensional evidence of coronary artery fistulae, including chamber (right ventricle or right atrium) volume overload. Seventeen patients underwent functional stress testing with either echocardiography or nuclear imaging before the procedure. Of those 17 patients, 10 had documented ischemia on their stress tests corresponding to the area supplied by the feeding artery.
Coronary artery fistulae origin and drainage site is shown in Figure 1. Origin from the left anterior descending artery (LAD) was the most common origin, followed by the right coronary artery (RCA) and circumflex artery. One patient had fistulae originating from the aortic arch. The fistulas drained most commonly into the pulmonary artery, in 19 cases. Of those, 11 came from the LAD, 5 from the RCA, 2 from the left circumflex, and 1 from the aortic arch.
In the 36 closure procedures performed, coils were placed in 31 procedures, an Amplatzer ductal occluder in 1 procedure, an Amplatzer vascular plug in 2 procedures, a covered stent in 1, and Contour embolization particles were used in 1 procedure (Fig. 2). Thirty devices were deployed via an arterial approach and 3 using an arteriovenous wire loop. Three devices were delivered transvenously (via the femoral or jugular vein) (Fig. 3). Immediate post-deployment angiography (Fig. 4) documented complete occlusion in 32 (89%) procedures and trivial residual flow in 4 (11%) procedures.
Four complications (14%) occurred and included coronary spasm in 1 patient that was successfully treated with intracoronary and sublingual nitroglycerin. Two patients, 1 with circumflex-to–coronary sinus fistulae and 1 with right coronary artery–to–pulmonary artery fistulae had embolization of Tornado coils into the coronary sinus and left lower lobe branch pulmonary artery. The coils were successfully removed with a 4-mm snare without sequelae. One patient with multiple LAD-to-right ventricle fistulae treated with a covered stent developed chest pain the day after the procedure. Coronary angiography showed complete occlusion of the LAD due to thrombosis of the covered stent. Thrombectomy, coronary angioplasty, and stenting with a bare-metal stent were performed and the patient subsequently did well. There were no deaths, strokes, fistulae dissection, or significant arrhythmias.
Clinical follow-up was available in 19 patients at median of 1.2 years (range: 1 month to 5 years). Eight of the 19 patients were symptomatic at clinical follow-up. Three patients were dyspneic, 3 had chest pain, and 2 had palpitations secondary to atrial arrhythmias. Follow-up echocardiography was obtained in 12 patients at median of 5 months (range: 1 month to 7 years) and none had 2-dimensional or Doppler evidence of abnormal flows. Eight patients underwent nuclear stress testing, but only 2 of those patients had ischemia. Both were later found to have recanalization of fistulae.
A follow-up angiogram was obtained in 18 (62%) patients at a median of 1.5 years (range: 3 months to 4.5 years) after the initial procedure. Ten patients (56%) had no recanalization, 4 patients (22%) had trivial recanalization, and 4 patients (22%) had significant recanalization (Fig. 5). Of the 14 patients with none or trace recanalization, 4 had symptoms compared with 3 of the 4 with large recanalization. All 4 of the patients with large recanalization underwent reclosure. Three required recoiling of previously closed fistulas (RCA to pulmonary artery [PA] and LAD to PA). Another patient underwent direct ligation of fistulae during coronary artery bypass grafting for symptomatic coronary artery disease. One patient who had an RCA-to-PA fistula closed 14 months earlier had growth of a separate LAD-to-PA fistula on follow-up angiogram, and it was closed with coils. Examples depicting the types of devices used to close coronary fistulae in this patient series is shown in Figures 6, 7, 8, and 9.⇓⇓⇓⇓
Percutaneous approaches to closure of coronary artery fistulae are an alternative to traditional surgical closure. We document acute success and complication rates for this rare procedure. Although the guidelines recommend closure of all large fistulae and small to moderate ones in the presence of symptoms, elective closure of all fistulae even in childhood is advocated by some to prevent future complications (12–16). Defining the size of coronary artery fistulae is controversial and has not been done systematically. Many definitions have been proposed including calling coronary artery fistulae “small” if fistula is not larger than twice the proximate, “medium” if bigger than twice but smaller than 3× the size of proximate, and “large” if bigger (17). The prognostic significance of clinically silent coronary artery fistulae have been investigated by Sherwood et al. (18). In 31 patients with echocardiographic and catheterization findings of clinically silent coronary fistulae, at mean follow-up of 2.6 years, none developed symptoms and 7 (23%) were later found to have spontaneous closure based on either echocardiogram or angiography.
Before the advent of percutaneous approaches, surgical closure of coronary fistulae was the only option. The first successful surgical closure was reported by Biorck et al. (16) in 1947. Since then, there have been many published series with up to 56 patients. Overall, mortality is low (∼1%); single ligation is common; excision of fistula is rare; and cardiopulmonary bypass is used in approximately one-half of the cases (19,20). Reported complications are rare and include arrhythmias, myocardial infarctions, stroke, and ST-/T-wave changes on the electrocardiogram. Approximately a 10% rate of incomplete closure or recurrence of fistulae is reported in the surgical literature (21). However, in the Cheung et al. series of 41 patients (21), 21 agreed to undergo follow-up cardiac catheterization. Among the 21, 4 (19%) had residual or recurrent fistulae. They noted that the incidence of recurrence was higher in patients who underwent external plication versus those who had intracardiac closure.
Transcatheter closure is an emerging alternative in patients with suitable anatomy and those without other complex cardiac lesions such as concomitant valvular heart disease. It was first reported by Reidy et al. (7) in 1983. Over the years, there have been a series of transcatheter closure studies in up to 33 patients. Overall, the results have been comparable to surgery with low morbidity and mortality. Multiple devices have been used including coils, detachable balloons, covered stents, vascular plugs, and atrial septal defect devices. Although, most of the series report excellent procedural success rate, there is limited information about the short- and long-term recanalization rate of transcatheter treated patients with coronary artery fistulae.
Armsby and colleagues reported short-term findings in 33 patients who underwent transcatheter closure (22). Follow-up echocardiograms (median: 2.8 years) were obtained in 27 patients. There was no flow in 22 patients and 5 had small residual shunts. None had follow-up angiograms. More recently, Zhu et al. (23) reported transcatheter closure results of 24 patients. All had clinical follow-up ranging from 3 months to 10 years. No follow-up angiogram was obtained. One patient had a still murmur on examination at 6-month follow-up and subsequent echocardiography revealed residual flow. Cardiac catheterization demonstrated large residual flow, which was closed with coils. Review of 45 patients reported in the transcatheter closure of coronary artery fistulae literature between 1982 and 2002 demonstrated that follow-up (mean: 12 months) imaging studies were obtained in 33 patients. There were 27 follow-up echocardiograms and only 6 angiograms. Complete closure at follow-up was reported in 91% of cases (22).
In our study, we report short-term findings of one of the largest transcatheter closure series and the largest number of patients with follow-up coronary angiography after transcatheter closure of coronary fistulae to date. The immediate closure success rate in our study was equal to or better than what has been reported in prior transcatheter closure studies with similar patient numbers (22,23) and similar complication rates. The origin and drainage site of coronary fistulae in our study is similar to those reported in the transcatheter closure and surgical literature with origin from the left coronary system and drainage to the pulmonary artery being the most common. Coils were most commonly used in our study. All but 6 of the 36 devices were delivered antegrade (via femoral artery) in our study. In our series, follow-up angiograms were obtained on 18 patients. Of the 14 patients without or with trace recanalization, 10 were asymptomatic at the time of follow-up angiogram. Among the 4 patients who were found to have significant recanalization flow on follow-up angiograms, 3 were symptomatic.
The recanalization rate in our study is similar to that of Cheung et al. (21), who evaluated 41 patients with surgically closed coronary artery fistulae. Because there is a paucity of follow-up angiograms in the transcatheter closure literature, we were unable to compare our recanalization rate with any prior transcatheter closure study. In the surgical literature, there are reports of multiple other abnormalities noted on follow-up angiograms including persistent fistulae and coronary artery dilations (21). Latson (17) reported 4 patients who underwent recent follow-up evaluation 4 to 41 years after surgical closure at his institution. Three of the 4 were found to have thrombosis of the parent coronary artery. The risk of parent coronary artery thrombosis after closure either surgically or percutaneously is unknown as there is little data with the exception of few case reports (17,24). The mechanism of thrombus formation after fistulae closure is unknown, but potential causes include reduction of blood flow in the aneurysmal artery and trauma. In our series, 1 patient with multiple LAD-to–right ventricle fistulae treated with a covered stent developed stent thrombosis the day after his procedure. Although a large number of patients (48%) in our series were on aspirin before catheterization, whether these patients should be receiving either an antiplatelet agent or anticoagulant such as warfarin before or after the procedure is unclear. Certainly, stasis of flow can occur with closure of long-standing coronary artery fistulae and consideration should be given to oral anticoagulation. The duration of antiplatelet or anticoagulant therapy after the procedure is unclear but it is reasonable to continue at least 4 weeks and perhaps longer.
Given the presence of recanalization in our series as well as the other post–fistulae closure abnormalities noted in the surgical literature, it is important to routinely follow these patients' after fistulae closure even if they remain asymptomatic. When compared with other modalities such as an echocardiogram or computed tomography scan, angiography has the advantage of allowing direct device visualization and sizing of recanalization. Finally, 4 patients in our series had periprocedural complications. This finding underscores the importance of having experienced operators perform this procedure in appropriately selected patients.
Our study has several limitations including its small size and retrospective nature, which inherently makes it susceptible to certain biases and limitation to a single center. Most patients included in our study were referred to our institution, thus referral bias does not allow for an accurate assessment of the prevalence of incidentally found coronary artery fistulae. The numbers of patients with small fistulae who were not referred for evaluation is unknown. Not all patients in the study had clinical or angiographic follow-up, and this is an important limitation.
Transcatheter closure of CAF is feasible and an effective alternative therapy in patients with coronary artery fistulae with suitable anatomy. Procedure-related complications are not inconsequential; therefore, transcatheter closure of coronary artery fistulae should only be undertaken by experienced operators. A minority of patients treated with percutaneous transcatheter closure will experience recanalization. Routine clinical follow-up in all and imaging surveillance after closure in symptomatic patients should be considered after transcatheter coronary artery fistulae closure.
The authors have reported that they have no relationships to disclose. Bernhard Meier, MD, served as Guest Editor for this paper.
- Abbreviations and Acronyms
- left anterior descending artery
- pulmonary artery
- right coronary artery
- Received November 14, 2010.
- Revision received March 25, 2011.
- Accepted March 30, 2011.
- American College of Cardiology Foundation
- Sandhu J.S.,
- Uretsky B.F.,
- Zerbe T.R.,
- et al.
- Warnes C.A.,
- Williams R.G.,
- Bashore T.M.,
- et al.
- Reidy J.F.,
- Sowton E.,
- Ross D.N.
- Liberthson R.R.,
- Sagar K.,
- Berkoben J.P.,
- Weintraub R.M.,
- Levine F.H.
- Biorck G.,
- Crafoord C.
- Armsby L.R.,
- Keane J.F.,
- Sherwood M.C.,
- Forbess J.M.,
- Perry S.B.,
- Lock J.E.,
- et al.
- Loo B.,
- Cox I.D.,
- Morgan-Hughes G.J.,
- Marchbank A.J.