Correspondence Address: Himanshu Mahla Sri Jayadeva Institute of Cardiovascular Sciences and Research, Jayanagar 9th Block, BG Road, Bengaluru - 560 069, Karnataka India
Source of Support: None, Conflict of Interest: None
Background: Transcatheter closure of coronary artery fistula has emerged as an alternative to surgery. The management of coronary artery fistula is complicated, and recommendations have been based on anecdotal cases or very small retrospective series. Objective: The objective of this study is to determine the safety, feasibility, and immediate and intermediate follow-up results of the transcatheter closure of coronary arterial fistula. Methods: We reviewed the records of all patients presenting with significant coronary arterial fistulae between January 2011 and May 2014. Those with additional complex cardiac disease requiring surgical management were excluded from the study. A total of nine patients aged 11 months to 58 years with congenital coronary arterial fistulae underwent percutaneous transcatheter closure using various devices between January 2011 and March 2014. The immediate closure results and clinical follow-up were reviewed using information from a database or telephone calls to the center. Results: Fistulae originated from the right coronary artery in 4 patients (44.4%), left circumflex coronary artery in 3 patients (33.3%), and left anterior descending coronary artery in 2 patients (22.2%). The drainage site was the right atrium in 5 patients (55.6%), right ventricle in 2 patients (22.2%), left ventricle in 1 patient (11.1%), and coronary sinus in 1 patient (11.1%). All of the patients (100%) underwent successful transcatheter closure using the various devices. Angiography after device deployment revealed complete occlusion in 7 patients (77.7%) and trivial to mild residual flow in 2 patients (22.2%). Two patients (22.2%) had transient ST-T wave changes after the procedure. The Amplatzer Vascular Plug II was used in four patients (44.4%), Amplatzer Duct Occluder II was used in three patients (33.3%), and Lifetech Duct Occluder was used in two patients (22.2%). The left ventricular end-diastolic volume decreased from a mean baseline value of 82.77 ± 4.55 ml/m 2 to 77.22 ± 3.49 ml/m 2 at 1 month after the procedure (P = 0.001) and had normalized in all of the patients at 2 months postprocedure. The cardiothoracic ratio decreased from a mean baseline value of 0.57 ± 0.035 to 0.53 ± 0.02 at the 1-month follow-up and further decreased to 0.50 ± 0.007 at 2 months (P < 0.001). Follow-up was 100% complete and ranged from 2 months to 3 years (mean = 1.44 ± 0.79 years). There were no early or late deaths. All of the patients were asymptomatic at 1 month postclosure, except one patient (11%) who had a mild residual shunt until 3 months of follow-up and dyspnea on exertion that was medically managed. One patient (11%) had an associated ostium secundum atrial septal defect that was percutaneously closed. Conclusion: Transcatheter closure of coronary arterial fistula is feasible and safe in anatomically suitable vessels and is a promising alternative to surgery in most patients.
How to cite this article: Mahla H, Mahimarangaiah J, Mandikal Kodanda Rama Sastry U, Chikkaswamy SB, Bhat P, Manjunath CN. Transcatheter closure of coronary arterial fistula: A single-center experience. Nig J Cardiol 2016;13:111-8
How to cite this URL: Mahla H, Mahimarangaiah J, Mandikal Kodanda Rama Sastry U, Chikkaswamy SB, Bhat P, Manjunath CN. Transcatheter closure of coronary arterial fistula: A single-center experience. Nig J Cardiol [serial online] 2016 [cited 2022 Sep 26];13:111-8. Available from: https://www.nigjcardiol.org/text.asp?2016/13/2/111/187709
Coronary artery fistulae are rare anomalies that are commonly diagnosed during the evaluation of a continuous murmur in an asymptomatic child. Elective closure in childhood has been recommended ,, after reports that patients with coronary artery fistulae can develop complications such as myocardial ischemia, congestive heart failure, endocarditis, or aneurysmal dilation during or after the second decade of life. , Transcatheter closure of coronary fistulae has been performed since 1983. ,,,,,,,,,, Although surgical closure began in our institution in 1995, transcatheter closure began later (in 2011) due to our stringent criteria for device closure only of isolated fistulae with symptoms or chamber dilatation. Thus, following these criteria, only nine cases underwent device closure in our center, whereas the remainder underwent surgical closure because they did not meet the institutional criteria for device closure. We present our experience herein to describe the techniques used in the transcatheter closure of coronary arterial fistulae, report our results with this procedure, compare our findings with those described in other transcatheter literature, and determine the best approach for this procedure.
Patient identification and selection
We searched our medical databases for all patients with coronary arterial fistulae who underwent a catheterization procedure between January 2011 and May 2014. After the exclusion of patients with additional complex cardiac disease requiring surgical management and those with small fistulae found incidentally during echocardiography or catheterization, we identified ten patients with clinically significant fistulae (those producing symptoms or a typical murmur). A total of nine patients successfully underwent transcatheter closure.
Inclusion required the presence of symptoms of heart failure or ischemia and/or a continuous murmur of Grade III. In addition, one of the following criteria was required: Electrocardiographic criteria indicating volume overload, increased cardiothoracic ratio on a posteroanterior (PA) view chest X-ray, cardiac chambers showing volume overload in echocardiography, color Doppler evidence of jet of >3 mm, or the presence of coronary dilatation >4 mm.
Patients were excluded due the following criteria: Presence of multiple drainage sites on angiography or presence of significant comorbidities, for example, severe anemia, respiratory tract infection with sepsis, active infective endocarditis, or more than mild ventricular dysfunction.
All of the patients underwent standard 12-lead electrocardiography. The criteria for chamber enlargement were applied as previously described. 
All patients who were diagnosed with coronary artery fistula on transthoracic echocardiography underwent a detailed assessment for chamber dilatation, left ventricular end-diastolic volume (LVEDV), and left ventricular function. Echocardiography was performed using an EnVisor HD 7 machine (Philips Medical Systems, Andover, MA, USA). In patients younger than 3 years old, LVEDV was measured using the "bullet" method as described previously.  In patients older than 3 years, end-diastolic volume was measured as recommended by the American Society of Echocardiography.  LVEDV was indexed to the body surface area, which was measured using the DuBois formula. 
Chest X-ray posteroanterior view
All the patients underwent chest X-ray to obtain a PA view, with the cardiothoracic ratio taken as the ratio between the maximal horizontal cardiac diameter and the maximal horizontal thoracic diameter. 
Heparin (100 units/kg) and a single dose of an intravenous antibiotic were administered after femoral venous and arterial accesses were established. After obtaining hemodynamic data, aortic root angiography and selective coronary angiography were performed to demonstrate the anatomy of the fistula and its drainage site and to identify the distal coronary branches. Computed tomography (CT) angiography was performed in selected cases if a major branch origin was distal to the fistula origin, if the anatomy was not clear after conventional angiography, or if the fistula origin was very near the coronary ostium. In addition, CT angiography was performed after conventional coronary angiography. To facilitate catheter positioning in the distal region of the fistula, high-torque floppy wires or microcatheters were used as required. The feasibility of closure using a device was determined according to the number and location of drainage sites, ability to cannulate the distal fistula, and proximity of the coronary branches to the optimal occlusion site. Special care was taken to avoid device interference with flow into any visible coronary artery. An Amplatzer Vascular Plug II (AVP II) (St. Jude Medical, Inc., St. Paul, MN, USA) was used in 4 patients; an Amplatzer Duct Occluder was used in 3 patients (St. Jude Medical, Inc., St. Paul, MN, USA), and a Lifetech Duct Occluder was used in 2 patients (LifeTech Scientific Corporation, Shenzhen, China). The selection of the occlusion device was based on the anatomic features of the fistula and operator convenience. To occlude the distal fistula immediately proximal to the drainage site, we selected vascular plugs 40% larger than the maximum diameter of the vessel. Devices were deployed as far as possible from the origin site but not too close to the distal narrowing (if one was present) near the drainage site.
Devices were deployed using either a venous or an arterial route. The typical approach involved making an arteriovenous (AV) loop and delivering the device through the venous side. Six patients had device delivery through the femoral vein, one through the left internal jugular vein (because the persistent left superior vena cava [SVC] and fistula were communicating with the coronary sinus), and two through arterial routes. After making an AV loop with the typical approach, an exchange length of guidewire was snared and exteriorized through the venous end, over which a Cook Sheath (Cook Medical (Bloomington, IN) was anchored at the desired site of device deployment. All of the patients had a minimum hospital stay of 3 days, and all received heparin infusions at 10-15 units/kg/h for 12 h. Postclosure, all of the patients were treated with aspirin 5 mg/kg for 6 months.
Computed tomography angiography
Two of our patients underwent CT angiography after conventional angiography to delineate the anatomy.
Medical records, including echocardiograms, electrocardiograms, and chest X-rays, were reviewed. The presence and degree of residual flow after closure were determined based on the most recent echocardiographic data.
Medical records, including clinical evaluations, catheterization, and electrocardiographic and echocardiographic follow-up data, were reviewed in all of the patients. Approval to conduct the medical record review was obtained from our Institutional Review Board.
Descriptive and inferential statistical analyses were performed. The results for continuous measurements are presented as the mean ± standard deviation (minimum-maximum), and the results for categorical measurements are presented as percentages. Significance was assessed at the 5% level. The following assumptions about the data were made: Dependent variables should be normally distributed, and samples drawn from the population should be random (the cases of the samples should be independent). Student's t-test (two-tailed, dependent) was used to determine the significance of the study parameters on a continuous scale within each group. SAS version 9.2 (Cary, NC), SPSS version 15.0 (IBM Corporation, USA), Stata version 10.1 (Texas, USA), and MedCalc version 9.0.1 (Ostend, Belgium) were used to analyze the data, and Microsoft Word and Excel were used to generate the tables.
The baseline characteristics and procedural details of the study patients are shown in [Table 1] and [Table 2]. Nine patients aged 11 months to 58 years (mean 14.88 ± 18.75 years) underwent device closure of coronary fistulae from January 2011 to March 2014. Six patients were female, and three were male (male:female = 1:2). Eight patients (88.9%) had fistulae without any associated congenital heart disease. One patient (11.1%) had an associated ostium secundum atrial septal defect that was closed percutaneously. The most common primary indication for closure was volume overload of the cardiac chambers, which was present in seven patients (77.8%). Venous approaches were used to deliver the device in seven patients (77.8%). Two patients (22.2%) had ST-T changes during the procedure or immediately postclosure; one of these had associated ischemic chest pain. Neither required specific treatment. No arrhythmic complications or any incident related to embolization/device migration occurred, and no patients experienced device thrombosis. In seven patients (77.8%), no residual flow occurred on the immediate postclosure angiogram. Trivial to mild residual flow was present in two patients (22.2%). One of these patients had trivial residual flow and showed no residual flow at 2 months postclosure (as documented via echocardiogram); the other had mild residual flow that persisted up to 3 months after closure (extending beyond the study period). Therefore, at 2 months, the immediate closure rate was 77.8% and the closure rate was 88.9%. The follow-up (minimum 3 months and maximum 3 years, mean = 1.44 ± 0.79 years) rate was 100%. All of the patients were compliant with the antiplatelet regimens for the first 6 months following closure. Eight patients (88.9%) were completely asymptomatic at the 1-month follow-up. One patient was symptomatic (although with decreased symptoms) from 3 months until the end of the study period. Cardiothoracic ratio and LVEDV were normalized in all of the patients at 2 months postprocedure [Table 3]. There was no early or late mortality or episode of infective endocarditis.
Elective closure of significant coronary arterial fistulae in childhood has been advocated to prevent later complications. ,,, Since its introduction in 1983,  transcatheter closure of these lesions has increasingly been utilized as an alternative to surgical closure. Our institution has adopted the following strategy for coronary fistulae: Patients with fistulae and additional complex heart disease requiring surgery are referred for surgical repair. Patients with clinically significant fistulae (i.e., producing symptoms or a typical murmur) are referred for catheterization during which the fistula anatomy is further defined. If doubt remains regarding the feasibility of percutaneous closure (as in cases of multiple drainage sites, origin of a major branch from the distal site of the fistula or the exact distance from the coronary ostia), CT angiography can be performed to delineate the anatomy (based on operator choice). Finally, patients with suitable anatomy undergo percutaneous closure, whereas those who are unsuitable are offered surgical ligation. In our study, the most common indication for closure was volume overload of cardiac chamber/s.
The clinical presentation of coronary artery fistula is primarily dependent on the severity of the left-to-right shunt. The majority of adult patients with coronary fistulae are asymptomatic. However, their natural history can be variable, and fistulae can cause symptoms at any age. The type of fistula, shunt volume, site of the shunt, and presence of other cardiac conditions have been associated with the clinical characteristics of coronary fistulae.
In our experience, the most common site of origin was the right coronary artery (RCA), which is same as that described in the literature.  The drainage site is most commonly on the right side of the heart.  Based on our study, fistulae were more common in female patients (male:female = 1:2), which has been a point of controversy in previous studies. A study by Ata etal.  found that male patients were more commonly affected than female patients, with a ratio of 2.3:1. Another study by Chiu etal.  showed no sex bias in coronary fistulae. However, their patients were than those of our study (59 ± 13); in addition, our study had a small cohort (nine total), of whom 7 (77.8%) were <20 years old. We hypothesize that coronary fistulae are more easily detected in pediatric groups due to murmurs found at school health checkups, failure to thrive or respiratory tract infections; in adults, fistulae are primarily detected incidentally when angiography is performed for other reasons (most fistulae that escape detection in childhood are asymptomatic). In some adults, fistulae are detected if they are severely symptomatic due to ischemia or heart failure.
Small coronary artery fistulae are usually asymptomatic due to small shunt flow; however, large coronary artery fistulae are primarily symptomatic, causing heart failure, pulmonary arterial hypertension, myocardial infarction, arrhythmias, endocarditis, or rupture. The symptoms of coronary fistula in adult patients usually begin in the fifth or sixth decade of life. 
The route for delivering the device can be either arterial or venous. Antegrade (venous) deployment avoids potential damage to the femoral artery, enables the use of larger catheters, and allows a straighter catheter course. Thus, the venous route is preferred by most operators. Other factors contributing to the choice of approach for closure are determined by the number and the location of drainage sites, location of the proximal coronary branches, and ability to cannulate the distal region of the fistula. We used arterial routes in only two patients because the sheath was backing out and unstable in these patients during venous approaches. Of these two patients, one had a left circumflex coronary artery to right atrial fistula [Figure 1], [Figure 2] and Video 1], and the other had an RCA to left ventricle fistula [Video 2].
Figure 1: Anteroposterior view showing device delivery through the arterial route in a patient with left circumflex artery to right atrium fistula
An AV loop can cause a tamponade-like effect if attempted in a small child. Procedural success was defined as stable positioning of the device with no more than mild residual flow. Residual flow can be observed postclosure and can decrease or persist on follow-up [Figure 2], [Figure 3] and Videos 3, 4]. In one patient, a left internal jugular venous approach was used because the fistula was communicating via the left circumflex artery with the coronary sinus; there was also a persistent left SVC, which provided easy access to the fistula and facilitated device positioning [Figure 3] and Videos 5, 6]. Balloon occlusion before device closure can be performed, especially if a major branch compromise is anticipated or if the fistula origin is near the ostia [Figure 4], [Figure 5] and Video 7]. The device is usually held immediately at or beyond the site of balloon occlusion. 
Figure 3: Device closure through the left internal jugular vein in a patient with left circumflex artery to coronary sinus fistula and persistent left superior vena cava in anteroposterior cranial view
Various devices can be used for closure. Gianturco coils were previously very popular; however, the lack of a controlled release mechanism eventually rendered them obsolete. AVPs were introduced to the field of pediatric cardiology in 2003.  The AVP II is a modification of the original plug; it contains three segments instead of only one and provides better immediate occlusion rates in high-flow situations.  The choice of the device in our institution depends on fistula anatomy and operator experience. All of the devices used were 20-40% larger than the fistula diameter.
Related procedural complications include transient ischemic changes, device embolization, fistula dissection, transient T-wave changes, transient bundle branch block, myocardial infarction, transient atrial arrhythmia, and death. These complications are rare, except inadvertent device migration, which can occur as a result of high flow in a large fistula or an undersized device. Notably, devices can be retrieved with snares if they migrate. 
Catheter-based closure of coronary fistulae has better clinical outcomes, and the operative mortality is 0-1% in the current transcatheter closure era. , Intimal dissection of the coronary artery, thrombosis, or pericardial effusion can also occur. However, related morbidity and mortality rates are generally considered to be low. Pericardial effusion can be associated with increased hydrostatic pressure of the pericardial vessels and pericardial inflammation; pericardial effusion can also be similar to postpericardiotomy syndrome following open-heart surgery.
One of our patients had persistent mild flow through the fistula. This patient had a left circumflex to right atrial fistula that had been closed with an AVP II. The fistula had a very high flow rate. The patient is currently in follow-up, with a plan of revision on an as-needed basis if symptoms increase. Previous studies described a rate of residual flow of approximately 10-15%.  We used two additional criteria for efficacy: The closure-cardiothoracic ratio via PA view chest X-ray and the LVEDV at 1 and 2 months of follow-up, both of which showed significant decreases (P ≤ 0.001). A previous study demonstrated the same outcomes. 
The role of CT angiography for diagnosis and management has been demonstrated by others. , Overweight patients present a particular difficulty due to their insufficient acoustic windows for echocardiography. In this group, multidetector CT (MDCT) may enable excellent adjunctive anatomical delineation with high resolution, notably at the origin of the intervening anatomy and distal entry sites of the fistula. Distal vessel entry depiction on MDCT enables assessments of the presence or absence of obstruction, which determines the likelihood of a stealth presentation in the coronary artery. Contrast opacification of the receiving chamber/vessel is useful in confirming the fistula entry site and the patency of the shunt. Radiation exposure represents the only drawback.
Coronary arterial fistula has been classified as proximal or distal.  Most cases in our series had proximal fistula. Patients with proximal coronary artery fistula are believed to be at relatively low risk for adverse events from thrombosis and remodeling of the residual fistula segment following closure because there are no normal coronary branches arising from the fistula. Proximal fistulae can be closed proximally or distally. Thrombosis develops in the residual fistula segment up to the proximal run-off major epicardial artery without progression of the thrombus to the normal major coronary artery. All of the patients in this series had a distal closure; thus, we performed heparin infusion for 12 h postprocedure. Although there is a risk of thrombosis and embolization to a normal segment, previous studies revealed a few related coronary events.  Some authors recommended proximal device closure or surgical ligation for proximal fistulae;  however, this cannot be taken as a rule because many factors (e.g., branch origin) contribute to determining the best approach.
Distal fistulae can only undergo distal closure at the drainage site or immediately distal to the normal coronary branch. These cases are at risk for adverse coronary events from thrombosis and unfavorable remodeling from discrete intimal stenosis. The dilated conduit coronary artery undergoes a significant size reduction to conform to adjacent normal coronary branches due to a substantial reduction in flow following closure. The relative size of the distal coronary artery fistula, size of the conduit coronary artery, proximity of adjacent runoff vessels, and age at intervention can determine whether a reduction in the conduit coronary artery size involves favorable diffuse uniform intimal thickening versus discrete intimal stenosis. 
No strict guidelines exist for anticoagulation in coronary fistula, whether managed conservatively or following device closure. Only expert opinions and consensuses are available.  We apply a strategy of IV heparin infusion for 12 h in the hospital with activated partial thromboplastin time at 2-3-fold, the normal level followed by oral aspirin for 6 months. No patient had any event with this strategy throughout the study period.
One limitation of this study was that it enrolled a small number of patients. However, our results clearly show the benefits and efficacy of the transcatheter closure of coronary arterial fistulae.
With increased experience and improved devices and techniques, percutaneous closure of coronary arterial fistulae is emerging as a successful therapeutic strategy. The safe and effective results of transcatheter closure support the current convention of elective closure of clinically significant fistulae in childhood. The optimal approach depends on the anatomy of the fistula, presence or absence of associated defects, and experience of the interventional cardiologists and surgeons.
This was a single-center, nonrandomized, retrospective study that does not represent all age groups. It primarily included a pediatric population. In addition, because surgery is the gold standard, the results would have been more reliable if a comparison had been performed. Finally, limitations arose from calculating the LVEDV using echocardiography alone.
We are thankful to all of our patients and our institution for providing us with the opportunity to serve humankind.
Sunder KR, Balakrishnan KG, Tharakan JA, Titus T, Pillai VR, Francis B, et al. Coronary artery fistula in children and adults: A review of 25 cases with long-term observations. Int J Cardiol 1997;58:47-53.
Liberthson RR, Sagar K, Berkoben JP, Weintraub RM, Levine FH. Congenital coronary arteriovenous fistula. Report of 13 patients, review of the literature and delineation of management. Circulation 1979;59:849-54. [PUBMED]
McElhinney DB, Burch GH, Kung GC, Villegas MD, Silverman NH, Moore P. Echocardiographic guidance for transcatheter coil embolization of congenital coronary arterial fistulas in children. Pediatr Cardiol 2000;21:253-8.
Hancock EW, Deal BJ, Mirvis DM, Okin P, Kligfield P, Gettes LS, et al. AHA/ACCF/HRS recommendations for the standardization and interpretation of the electrocardiogram: Part V: Electrocardiogram changes associated with cardiac chamber hypertrophy: A scientific statement from the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society: Endorsed by the International Society for Computerized Electrocardiology. Circulation 2009;119:e251-61. [PUBMED]
Lytrivi ID, Bhatla P, Ko HH, Yau J, Geiger MK, Walsh R, et al. Normal values for left ventricular volume in infants and young children by the echocardiographic subxiphoid five-sixth area by length (bullet) method. J Am Soc Echocardiogr 2011;24:214-8.
Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA, et al. Recommendations for chamber quantification: A report from the American Society of Echocardiography's Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiogr 2005;18:1440-63.
Olivotti L, Moshiri S, Santoro G, Nicolino A, Chiarella F. Percutaneous closure of a giant coronary arteriovenous fistula using free embolization coils in an adult patient. J Cardiovasc Med (Hagerstown) 2008;9:733-6.
Dodd JD, Ferencik M, Liberthson RR, Nieman K, Brady TJ, Hoffmann U, et al. Evaluation of efficacy of 64-slice multidetector computed tomography in patients with congenital coronary fistulas. J Comput Assist Tomogr 2008;32:265-70.
Lee CM, Leung TK, Wang HJ, Lee WH, Shen LK, Chen YY. Identification of a coronary-to-bronchial-artery communication with MDCT shows the diagnostic potential of this new technology: Case report and review. J Thorac Imaging 2007;22:274-6.
Gowda ST, Forbes TJ, Singh H, Kovach JA, Prieto L, Latson LA, et al. Remodeling and thrombosis following closure of coronary artery fistula with review of management: Large distal coronary artery fistula - To close or not to close? Catheter Cardiovasc Interv 2013;82:132-42.