Case report
A baby girl, a female infant, was born at term (40 weeks + 5 days) via vaginal delivery with a birth weight of 3.4 kg and Apgar scores of 8 and 10 at 1 and 5 minutes, respectively. On day 1 of life, an echocardiogram performed due to a systolic murmur revealed a significantly dilated left main coronary artery and a dilated circumflex artery. Colour Doppler evaluation identified a high-flow fistulous tract originating from the dilated circumflex and draining into the right atrium (Figures 1) ( a ) and ( b ). The initial electrocardiogram (Figure 1( c )) showed sinus rhythm without ST-segment or T-wave abnormalities. Over the following days, the patient developed tachypnoea and signs of pulmonary overcirculation. The electrocardiogram showed significant Q waves and ST segment elevation in leads I, II, and aVF (Figure 1( d )).
(a). The echocardiogram showed a significantly dilated left main coronary artery and circumflex artery, while the left anterior descending artery was of normal size. ( b ). Colour Doppler evaluation demonstrated a large flow originating from the dilated circumflex artery, draining through a large fistulous tunnel into the right atrium. ( c ). The initial ECG showed a sinus rhythm with no ST-segment or T-wave abnormalities. ( d ). Subsequent ECG findings revealed significant Q waves and ST-segment elevation localised to leads I, II, and aVF.
Figure 1. Long description
The image contains four distinct elements: two echocardiogram images and two ECG readings. The first echocardiogram image shows a significantly dilated left main coronary artery and circumflex artery, while the left anterior descending artery appears normal. The second echocardiogram image, enhanced with color Doppler, demonstrates a large flow originating from the dilated circumflex artery, draining through a large fistulous tunnel into the right atrium. The first ECG reading shows a sinus rhythm with no ST-segment or T-wave abnormalities. The second ECG reading reveals significant Q waves and ST-segment elevation localized to leads I, II, and aVF. These images are overlaid or side-by-side to provide a comprehensive view of the cardiac condition.
Laboratory tests revealed a high-sensitivity troponin level of 47 ng/mL. Due to the patient’s clinical instability, computed tomography angiography was not performed. Despite the patient’s low weight (3,440 g) and young age (5 days), the decision was made to proceed with urgent percutaneous closure.
Vascular access was obtained via 4-Fr sheaths that were positioned in the femoral artery and vein. Angiography confirmed a dilated left main coronary artery and a large, tortuous proximal coronary artery fistula with a posterior origin and a tortuous course, draining into the right atrium (Figure 2( a )).
( a ). The angiogram revealed a dilated left main coronary artery and a very large coronary artery fistula draining into the right atrium. The circumflex artery originated from the proximal section of the fistula. ( b ). Post-closure angiogram showed total occlusion of the coronary artery fistula and unobstructed flow in the circumflex artery. ( c ). A three-month follow-up angiogram showed a total occlusion of the coronary artery fistula, with optimal remodelling of the dilated coronary arteries and an unobstructed circumflex artery.
Figure 2. Long description
An angiogram image displays a dilated left main coronary artery and a very large coronary artery fistula that drains into the right atrium. The circumflex artery originates from the proximal section of the fistula. Post-closure angiogram reveals total occlusion of the coronary artery fistula and unobstructed flow in the circumflex artery. A three-month follow-up angiogram shows total occlusion of the coronary artery fistula, optimal remodeling of the dilated coronary arteries, and an unobstructed circumflex artery.
To minimise the risk of arterial vessel trauma and avoid introducing the use of large, stiff, rigid catheters into the left main coronary artery of a very low-weight patient, an antegrade approach was initially attempted. We decided to perform an antegrade closure of the coronary artery fistula.
An arteriovenous loop was created using a 0.014-inch coronary guidewire; however, the patient did not tolerate the manoeuvre and suffered a cardiac arrest. Cardiopulmonary resuscitation was initiated, and haemodynamic stability was restored promptly following the loop removal.
We speculate that the arteriovenous loop established from the femoral artery, across the left main coronary artery and the fistulous tract to the femoral vein, exerted excessive traction due to the tortuosity of the fistula, potentially causing coronary distortion, thereby triggering ventricular tachycardia and subsequent cardiac arrest. Consequently, a retrograde aortic approach was pursued.
The 4 Fr femoral artery sheath was exchanged and upsized to a 5 Fr sheath, and the left main coronary artery was engaged with using a 5 Fr Judkins Right 3.5 guiding catheter. Two 0.014-inch extra-support coronary wires were advanced through the fistula into the right ventricle to facilitate and provide sufficient support for catheter progression advancement into the distal part segment of the fistula.
With the catheter positioned in the distal portion of the fistula, an Amplatzer Piccolo Occluder 5/4 was then successfully deployed. Final angiography confirmed total complete occlusion of the fistula with preserved and unobstructed flow into the circumflex (Figure 2( b )).
After the post-procedure, the patient was started on aspirin (4 mg/kg), and systemic aspirin and heparin were initiated. After 12 days, heparin was switched to rivaroxaban (0.9 mg three times daily). The postoperative recovery was uneventful.
Follow-up echocardiography confirmed successful fistula closure, the closure of the fistula with well-preserved biventricular function, and the absence of any segmental regional wall motion abnormalities. The electrocardiogram findings subsequently also normalised.
The angiogram at the 3-month follow-up showed demonstrated persistent complete fistula occlusion with positive remodelling and an unobstructed circumflex artery without evidence of thrombosis (Figure 2( c )).
At the 12-month follow-up, the patient was growing asymptomatic without signs of ischaemia. The echocardiogram revealed and reaffirmed a complete fistula closure and good biventricular contractile function.
Rivaroxaban therapy was well-tolerated throughout the treatment period, without side effects or bleeding complications. It was discontinued after 3 months, given the positive results of the control angiogram.
Discussion
To our knowledge, this case represents one of the youngest patients with a haemodynamically significant coronary fistula successfully treated with percutaneous transcatheter intervention.
Although coronary artery fistula is a rare anomaly, Reference Loukas, Germain and Gabriel1 coronary fistulas can lead to severe significant symptoms and haemodynamic instability as early as the neonatal period or from very early childhood, as occurred and was demonstrated in our patient. The Congenital Cardiovascular Interventional Study Consortium Coronary Artery Fistulas Registry Reference Gowda, Latson and Sivakumar2 included 48 neonates and infants (under one year of age), of whom 58% presented with symptoms of heart failure, 12.5% exhibited non-specific ST-segment changes on electrocardiogram, and 6% showed ventricular dysfunction on echocardiography.
Currently, there is no universal consensus regarding the specific indications and/or the optimal timing for coronary artery fistula closure during infancy in childhood.
In the study by Gowda et al. Reference Gowda, Latson and Sivakumar2 87.5% (24/48) of patients younger than one year of age underwent coronary artery fistula closure; of these, 57% were treated via a transcatheter procedure and 43% through a surgical intervention approach, with procedural success rates of 92% and 94%, respectively.
The long-term pharmacological management remains a subject of ongoing debate. Gowda et al. Reference Gowda, Latson and Sivakumar2 emphasise that post-procedural subsequent therapy should be tailored to depend on the fistula type, as distal fistulas carry a higher greatest risk of thrombosis, and the degree and extent of vascular remodelling observed during follow-up. They propose and suggest 3–6 months of anticoagulation therapy and 6–12 months of antiplatelet therapy. We opted for a direct factor Xa inhibitor (rivaroxaban); this choice was guided by its favourable safety profile, its lack of food-drug interactions, limited minimal drug-drug interactions, and no requirement for frequent venipuncture for monitoring, a significant advantage in neonatal care Reference Chan, Sobieraj-Teague and Eikelboom3 as well as its more favourable safety profile. Reference Chan, Sobieraj-Teague and Eikelboom3
Regarding the percutaneous technique, both antegrade and retrograde approaches have been described.: antegrade and retrograde. The antegrade approach involves the creation of an arteriovenous loop, allowing for and closing the fistula closure from the venous side.
This approach is often preferred as it utilises smaller sheaths and diagnostic catheters in the arterial system, thereby minimising the risk of trauma to the femoral artery, aorta, and coronary ostia. Reference Firouzia, Alemzadeh-Ansaria and Mohebbia4 The advantages of this approach include the use of small-sized sheaths and diagnostic catheters in the arterial vessels, thereby avoiding potential trauma to the femoral artery, aorta, and coronary ostia. Reference Firouzia, Alemzadeh-Ansaria and Mohebbia4
Conversely, in the retrograde approach, which involves closure, the fistula is closed from the arterial side. While this approach does not require an it and avoids the complexity of an arteriovenous loop, it necessitates the use of larger guiding catheters and delivery systems within the arterial vessels.
In the coronary artery fistulas registry, Reference Gowda, Latson and Sivakumar2 the antegrade approach was utilised in only 45% of patients younger than one year of age. Although small-calibre in low-weight patients, the use of small French-sized materials in arterial vessels is strongly preferred in low-weight patients; the arteriovenous loop manoeuvre may not be well tolerated in neonates, as evidenced by the haemodynamic instability observed in our patient. However, as demonstrated in our patient, the arteriovenous loop may not always be well-tolerated in such small individuals.
Although there are currently no occlusive devices specifically designed for the coronary artery fistula closure, the most frequently used off-label devices are ductal occluders, vascular plugs, and coils. Reference Firouzia, Alemzadeh-Ansaria and Mohebbia4–Reference Wu, Fares and Zellers5
In our case, we chose to use the Amplatzer Piccolo Occluder 5/4 due to its excellent navigability and relative softness, which are critical features when navigating the delicate coronary anatomy of a neonate. and our familiarity with this specific device.
In conclusion, while rare, despite their low incidence, coronary artery fistulas can be a significant and important cause of heart failure in children in the neonatal period, potentially rapidly leading to rapid haemodynamic deterioration and instability. Transcatheter closure has proved to be a feasible and safe therapeutic option, even in low-weight neonates, yielding excellent procedural results. Post-procedural antiplatelet and anticoagulation therapies are crucial, although further studies are required to define their optimal duration, which remains unknown.
Acknowledgements
The authors thank Drs. Grassitelli and Ferrero for their anaesthesiological support and Dr Agnoletti for her assistance during the procedure.
Author contributions
Elena Gribaudo: concept/design, drafting the article, critical revision of the article, approval of the article, and clinical involvement. Gaetana Ferraro: performed the intervention, concept/design, drafting of the article, critical revision of the article, approval of the article, and clinical involvement. Giuseppe Alberto Annoni: critical revision of the article and approval of the article.
Financial support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Competing interests
None.
