Background
A circular shunt is defined as a diversion of blood, often times intracardiac, that leads to flow recirculation to the same chamber without traversing a capillary bed. Reference Shone, Anderson, Elliott, Amplatz, Lillehei and Edwards1 Volume overload and low cardiac output are the causes of heart failure. This is a very rare condition, and it has been mostly described as a primary pre-operative condition in Ebstein’s anomaly and other CHDs involving pulmonary artery stenosis or atresia. Reference Shone, Anderson, Elliott, Amplatz, Lillehei and Edwards1–Reference Mustafa, Aghajani, Bairmani and Khalil9 However, only a case post-operative circular shunt has been published, in which IB-type tricuspid atresia underwent right cavo-pulmonary anastomosis without pulmonary trunk occlusion and developed a veno-venous fistula from the left superior vena cava to the coronary sinus with no surgical correction for the circular shunt at the moment of that publication. Reference Krishnamoorthy and Sivasankaran10
The options for definitive treatment reported so far have included surgical corrections as well as percutaneous procedures, and in the fetal life, the treatment has been transplacentary pharmacological therapy or fetal intervention. Reference Shone, Anderson, Elliott, Amplatz, Lillehei and Edwards1–Reference Mustafa, Aghajani, Bairmani and Khalil9
The aim of this manuscript is to share the first case of an “iatrogenic” circular shunt, which was resolved percutaneously, and discuss the reflections that this case provides us.
Case report
A 5-month-old girl was referred to our tertiary care hospital with a history of cyanosis and cardiac murmur. A transthoracic echocardiogram allowed diagnosis of situs solitus and levocardia, with total anomalous pulmonary venous connection to the coronary sinus. During the surgical procedure, a persistent left superior vena cava connected to the roof of the left atrium was incidentally found. The total anomalous pulmonary venous connection was corrected with unroofing of the coronary sinus, and its ostium was closed with a polyester patch with continuous suture. Due to the absence of the innominate vein, the surgical team decided to perform an end-to-lateral cavo-pulmonary anastomosis to the left branch of the pulmonary artery. She had an insidious evolution in ICU and was discharged 21 days later with loop and potassium-sparing diuretics.
Upon follow-up, at the age of 6 years, she was found to have failure to thrive, fatigue during feeding, and diaphoresis, with class II modified Ross heart failure classification. Physical examination was relevant for jugular vein distention; a collateral venous network in the neck was observed, along with hyperdynamic precordium, grade II/IV systolic heart murmur on the pulmonary valve area with fixed splitting of the second sound, normal hepatic border, and normal pulses without cyanosis.
Chest X-ray showed cardiomegaly (cardiothoracic index 0.6) and dilatation of right cavities. On electrocardiogram, there was evidence of right chamber overload. Transthoracic echocardiogram revealed an unroofed coronary sinus without obstruction of pulmonary vein flow with a mean gradient of 2 mmHg. Reversal flow in the cavo-pulmonary anastomosis was also detected, along with moderate dilatation of right chambers, an unobstructed right ventricle outflow tract, and preserved biventricular systolic function.
Cardiac catheterisation demonstrated a circular shunt through a 2.2 mm left cavo-pulmonary anastomosis with aneurysmal dilatation of the left superior vena cava and prominent collateral venous circulation draining to the right superior vena cava (Figure 1a and b). Percutaneous closure of the anastomosis was performed with an Amplatzer® ADO 8/6 device (Figure 1c), and she was discharged 24 hours after the procedure. At 7 days follow-up, symptom resolution was noted, and she was continued on aspirin solely. Non-gated cardiac CT angiography was done at 14 years of age (Figure 2) which showed an occluder device appropriately placed with multiple veno-venous collaterals between the left superior vena cava and the right superior vena cava. One year afterwards, she continues asymptomatic without medications in the last transthoracic echocardiogram; the right chamber size is normal without enlargement with conserved biventricular function.
( a ) Right ventriculography: we can see contrast medium over the left branch of pulmonary artery. ( b ) We can see the left cavo-pulmonary anastomosis (arrow) and the aneurysmatic dilatation of the LSVC. ( c ) An Amplatzer® ADO device occluding left cavo-pulmonary anastomosis. LSVC, left superior vena cava.

Non-gated cardiac CT-angiography: we can see Amplatzer® ADO device occluding left cavo-pulmonary anastomosis (arrow). ( a and b ). Anterior-lower view. ( c ). Posterior view.

Discussion
The circular shunt is a very rare condition where certain amount of the blood flow returns to the original chamber without traversing a capillary bed. In our case, the beginning chamber is the right atrium; the flow then passes to the right ventricle and to the both pulmonary branches arteries. In the left pulmonary branch, certain amount of this blood flow passes through the anastomosis into the left superior vena cava (without reaching the pulmonary capillary bed), Then by intercaval collateral circulation (left superior vena cava to right superior vena cava), this flow again reaches the right atrium (creating a circular shunt) with volume overload.
The left cavo-pulmonary shunt generated a significant circular shunt due to retrograde flow from the pulmonary artery to the left superior vena cava. Despite the fact that the anastomosis developed significant stenosis, the shunted flow was enough to cause heart failure. Percutaneous closure of the anastomosis was performed with an occluder device that allowed the resolution of the circular shunt and subsequently of heart failure symptoms with a satisfactory evolution and without complications.
Persistent left superior vena cava connected to the left atrium in the context of total anomalous connection of pulmonary veins to the coronary sinus is an extremely rare association.
The finding of a persistent left superior vena cava without an innominate vein to the roof of the left atrium with a biventricular physiology has three main approaches: (1) cavo-right atrial shunt; (2) cava-cava shunt (leaving the persistent left superior vena cava connected to the left atrium [in situ] would not be an adequate option, since it would present systemic desaturation and the long-term risk of paradoxical systemic embolism); and (3) ligation or section and suture of the left superior vena cava hoping that the collateral venous circulation from the left superior vena cava can drain to the right superior vena cava. The latter alternative is the current physiology of our case after placement of the device occluding the cavo-pulmonary anastomosis.
A detailed diagnosis of the CHD and their associated lesions, as well as the anatomical variants of the systemic venous return, is very important for the surgical team in order to be able to take the best decision and to resolve effectively these variants of the superior systemic venous return without eventualities or complications.
Despite the fact that cavo-pulmonary anastomosis is a surgical option in patients with single-ventricle or one-and-a-half ventricular physiology, our case leaves proof that it is not a good option for cases with biventricular physiology.
Unfortunately, we have some limitations in this case report: (1) the surgical note did not include any commentary about the pulmonary artery pressure before the cavo-pulmonary anastomosis was made; (2) no justification was provided in the surgical note to support this approach; (3) in the same way, any consideration and commentary about left superior vena cava physiology was not exposed in the note; and finally, (4) a closed clinical follow-up for the patient’s possible complications was not performed.
Conclusion
Circular shunt is an extremely rare condition with scarce reports involving Ebstein’s anomaly and other CHD involving pulmonary artery stenosis or atresia. Reference Shone, Anderson, Elliott, Amplatz, Lillehei and Edwards1–Reference Mustafa, Aghajani, Bairmani and Khalil9 We present the first case in the literature of a post-operative “iatrogenic” circular shunt, which was resolved by interventional catheterisation without complications.
Whenever unexpected congenital anomalies are incidentally found during surgery, it is important to try to restore blood flows as physiologically as possible. In biventricular physiology, it is important to anticipate a circular shunt whenever cavo-pulmonary anastomosis is performed and opt for other more physiological approaches.
