Mid-Term Follow-Up of Transcatheter Closure for Coronary Cameral Fistula in Pediatrics

Abstract

Background

Coronary cameral fistula (CCF), a rare abnormal coronary communication to cardiac chambers, may lead to coronary steal phenomenon and increase cardiac overload. We investigated the clinical and cardiovascular characteristics in children before and after transcatheter closure.

Methods

We retrospectively reviewed pediatric patients with CCFs diagnosed by echocardiography in a tertiary medical center between 1998 and 2019. Basic information, echocardiogram, catheterization and interventional procedures were obtained from medical charts.

Results

A total of 12 pediatric subjects were included. The median ages at diagnosis and catheterization were 0.2 and 2.8 years, respectively. All CCFs were unilateral and single with varying degrees of coronary artery dilatation and aneurysm formation and diagnosed by echocardiography. The median follow-up periods before and after catheterization were 2.5 and 7.3 years, respectively. Seven of the CCFs originated from the left side. The drainage sites were all right hearts. Before catheterization, the median size of the proximal end of the fistula was 3.1 mm, concomitant with enlargement of conduit coronary arteries. Eleven of the 12 patients underwent transcatheter closure using coils in six and vascular plugs in five. Only one patient had a significant increase in pulmonary-to-systemic flow ratio. The size of conduit coronary artery gradually decreased and the size of ipsilateral coronary branch increased after closure.

Conclusions

Transcatheter occlusion for CCFs in children is safe and effective. The morphology of CCFs varies with the degrees of dilation, tortuosity, and aneurysmal formation. After occlusion, alterations in the size of coronary arteries may be a prognostic indicator.

INTRODUCTION

Coronary cameral fistula (CCF) is an abnormal communication between one or more coronary arteries and cardiac chambers.¹ It is seen in approximately 7% of heart transplants because of surgery or post-transplant myocardial biopsy,² in some hearts with hypertrophic cardiomyopathy after septal myectomy,³ and secondary to infection⁴ or injury after a percutaneous coronary artery intervention.⁵ It is very rare in native hearts, accounting for less than 0.1% in patients who undergo diagnostic coronary angiography.⁶ Most patients with CCFs are asymptomatic, but a few have complications such as congestive heart failure, myocardial ischemia, fistula aneurysm rupture and sudden death.^6,7 Treatment options for CCFs include surgical closure and transcatheter closure.^8,9 However, few studies have compared the clinical and cardiovascular characteristics before and after device closure for congenital CCFs. In children, the majority of coronary fistulas are seen in those who are healthy or have Kawasaki disease, which mostly drain into the main pulmonary artery but not into cardiac chambers.¹⁰ With advances in interventional catheterization, the clinical characteristics and treatment strategies for CCF in children warrant further investigation owing to the particular hemodynamic status in such patients. In this study, we examined the clinical manifestations of patients who underwent cardiac catheterization in our institute for CCFs and analyzed the cardiovascular characteristics before, during and after the interventional catheterization.

METHODS

We performed a retrospective review of medical records of pediatric patients (aged less than 18 years) who had undergone cardiac catheterization to occlude CCFs between 1998 and 2019 at Kaohsiung Chang Gung Memorial Hospital, Kaohsiung, Taiwan. This study was approved by the Institutional Review Board (Permit number, no: 201901039B0) of Chang Gung Memorial Hospital, Kaohsiung, Taiwan and conducted in accordance with the principles of the Declaration of Helsinki.

Basic information was collected from the medical records including sex, vital signs, body weight, body height and age at diagnosis, at the time of transcatheter intervention and during follow-up, as well as chest x-ray, electrocardiogram and echocardiogram data before and after the catheterization. The standard size of coronary arteries on echocardiography was calculated as z score according to the patient’s body weight and body height.¹¹ Catheterization studies revealed the origin, course, and drainage site of fistulas, and hemodynamic studies included pulmonary to systemic blood flow ratio (Qp/Qs) and mean pulmonary artery pressure. The type of devices, method of approach and device placement of occlusion, which was defined as proximal, middle and distal third of the fistula, were also recorded.

RESULTS

Twelve pediatric patients with CCFs were recruited in this study, eight (66.67%) of whom were female. The median age at diagnosis was 0.2 years (range from 0.01 to 12.6 years). All of these cases had heart murmur as the initial presentation; ten of whom (83.3%) had systolic murmurs, and the other two had continuous murmurs. None had signs of heart failure or failure to thrive when they were initially diagnosed. Four (33.3%) had borderline or mild cardiomegaly as assessed by chest x-ray. Electrocardiography (EKG) did not show any significant ST-T change, but incomplete right bundle branch block was noted in two, right ventricle hypertrophy in one, and left ventricle hypertrophy in one patient.

All of the cases had undergone catheterization, and transcatheter closure for CCFs was performed in all cases except for spontaneous closure found in one CCF. The median age when undergoing catheterization was 2.8 years (range from 1.1 to 12.8 years). The median period between initial diagnosis and catheterization was 2.5 years (range from 0.5 months to 5.1 years) (Table 1).

Table 1. Anatomical characteristics of cameral fistula(s) by echocardiography during follow-up.

No.	Age		Origin (conduit CA)	Drainage	Fistula size*		Follow-up		Coronary artery
	At Dx	At Cath			At Dx (mm)	Before Cath (mm)	Period# (years)		Before closure (mm/z score)	After closure† (mm/z score)
1	1M	4Y7M	LCA	RV	2	3	4.5/11.8	LCA	5.0/+4.89	3.18/+0.74
								LAD	3.07/+2.24	2.34/+0.36
2	1M	1Y1M	LCA	RV	3.6	3.6	0.9/9.9	LCA	5.0/+6.02	4.4/+2.67
								LAD	3.89/+3.67	3.96/+2.17
3	2M	3Y	LCA	RA	2.9	4	2.8/1.1	LCA	4.40/+4.79	4.81/+5.03
								LAD	1.20/-0.46	2.2/+1.36
4	2Y3M	2Y4M	RCA	RA	3.1	3.1	0.04/16.5	RCA	6.60/+6.41	6.04/+4.05
5	1Y1M	2Y7M	RCA	RA	Difficulty in measurement	1.5/21.3	RCA	5.28/+5.31	3.08/+1.44
6	< 1M	3Y5M	RCA	RA	2.6	5	3.4/0.5	RCA	10.0/+8.18	9.54/+7.93
7	4M	2Y1M	RCA	RV	2	2.9	1.8/7.3	RCA	5.68/+5.58	3.71/+2.3
8	< 1M	2Y6M	RCA	RA	2.4	2.9	2.5/6.1	RCA	6.10/+5.86	2.57/+1.32
9	12Y7M	12Y8M	LAD	RV	5.5	5.5	0.3/9.4	LCA	8.60/+5.92	4.96/+2.39
								LAD	-	-
10	3M	5Y1M	LAD	RV	3.2	2.4	4.8/0.4	LCA	5.19/+4.72	3.53/+2.48
								LAD	4.01/+2.89	3.47/+2.33
11	8M	5Y9M	LCX	RV	3.5	3.7	5.1/6.2	LCA	5.49/+4.92	3.33/+0.76
								LAD	2.17/+0.83	2.96/+0.99
12	3M	2Y5M	LCX	RV	2.2	0	2.2/1.4	LCA	4.50/+4.93	4.75/+4.82
								LAD	1.54/+0.36	2.24/+1.34

CA, coronary artery; Cath, catheterization; Dx, diagnosis; LAD, left anterior descending artery; LCA, left main coronary artery; LCX, left circumflex artery; M, month-old; RA, right atrium; RCA, right coronary artery; RV, right ventricle; Y, year-old.

* The size was measured at the proximal end of CCF. ^# Follow-up period between the initial diagnosis and catheterization/and duration from the catheterization to the latest follow-up. ^† The data was obtained from the echocardiograms taken in the latest follow-up.

Echocardiography study

All CCFs were diagnosed by echocardiography performed due to heart murmurs. Echocardiograms revealed that each patient had a single and unilateral CCF (Figure 1). The median diameter of fistula was 2.9 mm (range from 2.0 to 5.5 mm) at initial diagnosis and 3.1 mm (range from 0 to 5.5 mm) before catheterization (Table 1). Over time, seven (58%) fistulas became larger, three (25%) remained stationary, and the other two seemed to become smaller or even spontaneously close. The conduit coronary arteries proximal to the CCFs were measured. Five of the CCFs (41.7%) originated from the right coronary artery (RCA), three (25%) from the left main coronary artery (LCA), two (16.7%) from the left anterior descending coronary artery (LAD), and two (16.7%) from the left circumflex coronary artery (LCX). The drainage chambers were the right ventricle (RV) in seven fistulas (58.3%) and right atrium (RA) in the other five (41.7%). Before catheterization, six of the patients had a normal chamber size, one (8.3%) had RA dilatation, two (16.7%) had left ventricle (LV) dilatation, and the other three (25%) had dilatations of both left atrium (LA) and LV. The median z score of the RCA was +5.86 (range from +5.31 to +8.18) in the CCFs drained from the RCA, and the median z score of the LCA was +4.92 (range from +4.72 to +6.02) in the CCFs drained from the left side (Table 1). After closure, all of the cases except one had decreased sizes of the conduit arteries. At the latest follow-up with a median period of 7.3 years after transcatheter device closure, the median z score of the RCA in the CCFs from the RCA decreased to +2.30 (range from +1.32 to +7.93) and those of the LCA in the CCFs from the left side coronary artery decreased to +2.48 (range from +0.74 to +5.03).

Figure 1.

Representative echocardiograms of coronary cameral fistulas, showing (A) the origin of left main coronary artery (LCA) and (B) right coronary artery (RCA) both draining into right atria (RA). Yellow arrows indicated the coronary cameral fistula (CCF).

CCF morphology, catheterization, and interventional procedures

Angiography studies showed that all CCFs had varying degrees of tortuosity, dilation and aneurysmal formation of coronary arteries (Figure 2). All of them had diffuse ectasia of the original conduit arteries and/or CCFs, and seven (58.3%) had a very tortuous course of the CCFs. Ten (83.3%) had aneurysmal formation; six (50%) had giant aneurysm and seven (58.3%) had multiple aneurysms. Hemodynamic studies showed that the median shunt ratio of pulmonary-to-systemic blood flow (Qp/Qs) was 1.1 (range from 1.0 to 2.5) and the median mean pulmonary artery pressure was 17 mmHg (range from 12 to 26 mmHg).

Figure 2.

Representative coronary angiograms, showing morphology of coronary cameral fistulas before (PRE) and after (POST) transcatheter closure, using (A) Gianturco coils to occlude a coronary cameral fistula (CCF) from right coronary artery (RCA) to right atria (RA) and (B) an Amplatzer vascular Plug II to occlude the CCF from LCA to RA. Yellow arrows indicated the CCFs (left panels) before closure and coils and the device (right panels) after closure.

Transcatheter closures were performed in all patients except for one with spontaneous closure (Table 2). One patient underwent a second closure 3 months after the first closure because of significant heart murmur and the residual shunt remained. Among the 12 procedures in 11 patients, six (6/11, 54.5%) used coils and vascular plugs were used in five patients (45.5%) to occlude the CCFs. One patient underwent closure twice using one vascular plug each time. The antegrade approach was used in nine patients (81.8%) and the retrograde approach in the other two. All of the patients had successful device or coil closure for CCFs without complications. The CCFs were occluded at distal sites in seven (63.6%), at proximal sites in another two (18.2%), and at middle sites in the other two (18.2%). The residual shunts were detected in five patients (45.5%) by angiography immediately after the closure, and by echocardiography 1 week after the closure. Seven (63.6%) of the 11 patients took low-dose aspirin for a median duration of 6.3 months (range from 5.5 to 9 months), including those with no residual shunt (n = 3), small residual shunt (n = 3) and tiny shunt (n = 1) after the procedure. Four did not take aspirin; three of them had no residual shunts and one had a small residual shunt after closure. None of the five residual shunts were detected by echocardiography after a median period of 9 months of follow-up.

Table 2. Catheterization and interventional procedures.

Age at catheterization (Y)	2.8 (1.1-12.8)
Device, n*
Amplatzer vascular plug I	4/11 (36.4%)#
Amplatzer vascular plug II	1/11 (9.1%)
Gianturco coil	5/11 (45.5%)†
VortX-18 coil	1/11 (9.1%)‡
Delivery sheath, n
Heartrail catheter	4 (36.4%)
JR-4	4 (36.4%)
JL-4	2 (18.2%)
Multipurpose catheter	1 (9.1%)
Obliterated site, n
Proximal	2 (18.2%)
Middle	2 (18.2%)
Distal	7 (63.6%)
Qp/Qs	1.1 (1.0-2.5)
MPP, mmHg	17 (12-26)
Residual shunt§	5 (45.5%)

JL, Judkins left; JR, Judkins right; MPP, mean pulmonary artery pressure; Y, year-old.

* The number of patients received the certain type of devices. ^# One patient underwent transcatheter closure twice due to significant residual remained after the first closure. An 8-mm and 12-mm Amplatzer vascular plug I was used for the first and the second time of closure, respectively. ^† Three of 5 patients received 2 coils for one procedure. ^‡ Three VortX-18 coils were used in this patient. ^§ Residual shunt was detected immediately by angiograms after transcatheter closure or by echocardiograms one week after closure.

Electrocardiography

All of the cases had an EKG survey when they were initially diagnosed, and before and after the catheterization. None of the cases had ischemic heart findings such as ST-T change or inverted T wave. In addition, none had chest pain, ischemic heart disease, arrhythmia or mortality.

DISCUSSION

In this study, a total of 12 pediatric patients with CCFs were recruited. All underwent catheterization and follow-up at our hospital. Eleven patients were successfully treated with transcatheter closures.

Most patients with small CCFs are asymptomatic and do not need treatment. However, further management such as surgery or transcatheter device closure may be performed if the patient has significant left-to-right shunt with volume overload, myocardial ischemia, congestive heart failure and to prevent endarteritis or rupture of an aneurysm.¹² There are currently no guidelines for the specific management of pediatric CCF due to its scarcity. According to the guidelines for adults with congenital heart diseases, adult patients with large coronary artery fistulas, regardless of symptomatology, or adult patients with small to moderate fistulas concomitant with documented myocardial infarction or unexplained ventricular dysfunction, the fistula should be closed either by a surgical procedure or a transcatheter intervention.¹³ For pediatric patients with CCFs, transcatheter occlusion is indicated for those with symptomatic coronary artery fistulas, or possibly for asymptomatic patients with moderate or large fistulas.¹⁴ In our study, all patients had large CCFs and coronary ectasia. One half of the patients underwent catheterization due to chamber dilatation, which may suggest significant left-to-right shunt, and the other half due to progressive coronary artery dilatation, including the spontaneous closure of one CCF. Persistent or progressive dilatation of conduit coronary arteries may suggest an increase in significant blood flow or post-stenotic dilatation secondary to the fistula.¹⁵ Over time, varying degrees of steal phenomenon and subclinical myocardial ischemia may develop, and the fistula may also become larger and complicated with aneurysm rupture.⁷ Additionally, the need for endocarditis prophylactics is necessary for such significant shunt. However, the relatively higher heart rate and poor cooperation of children, the cost and X-ray exposure mean that both physicians and parents of small children hesitate to use coronary CT angiography for diagnostic purposes. Therefore, transcatheter closures were arranged. Surprisingly, Qp/Qs shunts were not high for most of the CCFs despite large fistulas and/or conduit coronary arteries with giant aneurysms. The small opening at the site of fistulous termination in most cases may account for this finding. None of the patients experienced clinical signs of heart failure, failure to thrive or ischemic change over resting EKG before and after the closure. Some studies have reported that about 30 to 55% of adult patients with coronary artery fistulas had ischemia as assessed by myocardial perfusion scans.^16,17 However, no pediatric data have been established because of the rarity of the disease and an inability of young children to cooperate with myocardial scans or exercise EKG. In our study, stressed myocardial scans were performed in two patients before transcatheter closure and one after the closure. The results revealed no evidence of ischemia. However, we found that the size of the ipsilateral coronary branches from the same origin of the CCFs might increase after device closure. In patient no. 3, the z score of the LAD was -0.46 before catheterization but increased to +2.01 (data not shown) 6 months after closure. Cases no. 2, 11, and 12 had similar increases. This may imply steal phenomenon and a mild degree of myocardial ischemia, which could not be detected by resting EKG or even myocardial scan.

There are different opinions about the alteration of coronary size after device closure. Some reports have shown that coronary artery size may remain enlarged even after an intervention,¹⁸ whereas others have reported opposite results.¹⁹ Interestingly, we found that, over time, most (10/11, 90.9%) of our patients had decreases in conduit coronary size after closure except for case (no. 3) and another case whose CCF spontaneously closed for more than 1 year. More time may have been needed to observe continuing alterations for these previously dilated coronary arteries. Accordingly, we suggest that a decrease in conduit coronary size could be a non-invasive parameter for complete occlusion in addition to the absence of residual CCF shunt. Echocardiography is important in the diagnosis and follow-up for CCF.

Previous reports have shown that transcatheter closure for CCFs may be less invasive than surgery and a safe choice.²⁰ With advances in device materials and designs, new types of devices and coils may be used instead of Gianturco coils according to CCF morphology. Multidetector computed tomography and coronary angiography are superior to echocardiography in delineating CCF morphology for the device choice and to determine the optimal site of obliteration. In the future, three-dimensional bioprinting methods may provide more suitable biomaterials or devices to fit the geometry of CCFs.²¹ All CCFs undergoing transcatheter closures in our series had either a very tortuous course or varying degrees of aneurysm formation. The only CCF which spontaneously closed had a non-tortuous course and no aneurysm formation. It remains uncertain whether there is any association between CCF morphology and its rate of spontaneous closure. A previous study on adults reported that CCFs arose more from the left side and had more signs and complication than in our study.²² In our study, more than half of the CCFs originated from the left side and drained into the right-side chamber of the heart, with no LV or LA drainage chamber. The complications of transcatheter closure include myocardial ischemia, device embolization and device migration.^23,24 Our patients did not have any complicated thromboembolic or myocardial ischemic events during follow-up. Unlike long-term aspirin in conjunction with coumadin use reported in a previous study,²⁴ approximately 60% (7/11) of our patients took low-dose aspirin only for a short term. Small and tiny residual shunts remained in two of them at the time of aspirin discontinuation, which spontaneously closed years later. Among four patients who did not take aspirin, the one who initially had a residual shunt did not develop signs of myocardial ischemia, and the shunt spontaneously closed later on. Thus, it remains unclear how long aspirin should be taken to prevent such complications.

CONCLUSION

Transcatheter occlusion may be a good choice for pediatric CCF closure owing to its safety and efficacy. Morphology of CCFs varies with degrees the of dilation, tortuosity, and aneurysmal formation. During follow-up, echocardiography may be helpful in predicting the prognosis.

CONFLICTS OF INTEREST

The authors declare no conflict of interest.