Unexpected Severe Cyanosis Following Ventricular Septal Defect Closure and Repair of Right Ventricular Outflow Tract Obstruction

Introduction

Ventricular septal defects (VSDs) are a common congenital heart defect, with an incidence of 2 per 1000 live births.^1,2 The Society of Thoracic Surgeons recognizes four types of VSDs. Type I involves the conal or outlet septum, type II includes the membranous septum, type III are inlet or atrioventricular septal defects, and type IV are muscular defects. Type II VSD, also known as perimembranous VSD, is the most common, comprising 80% of cases. Ventricular septal defects can exist in isolation or with associated defects, including Tetralogy of Fallot, double outlet right ventricle, interrupted aortic arch, aortic coarctation, or transposition of the great vessels.^1,2 Thus, when identified, it is important to have suspicion and investigate for other congenital anomalies. Although spontaneous closure is not uncommon for certain types of VSDs, surgical closure remains one of the most common congenital cardiac operations. We present the case of a significant problem during VSD repair, related to an undiagnosed intracardiac shunt.

Patients and Methods

A seven-month-old boy, weighing 7.6 kg, with a large perimembranous VSD, moderate pulmonic stenosis, and hypertrophied right ventricular muscle bundles (RVMB) was referred for surgical repair. He had previously undergone a successful balloon pulmonary valvuloplasty at six weeks of age due to downtrending arterial saturations and cyanosis. Preoperative transesophageal echocardiogram (TEE) showed moderate pulmonary stenosis (peak gradient 55 mm Hg), subpulmonary muscle bundles (peak gradient 25 mm Hg), a large perimembranous VSD measuring 8 mm, and a small patent foramen ovale (PFO) with bidirectional shunting. There was no aortic override or anterior malalignment of the conal septum, nor was there a dilated coronary sinus (CS) or persistent left superior vena cava (LSVC).

After median sternotomy and aortobicaval cannulation, cardiopulmonary bypass (CPB) was initiated, the patient was cooled to 28 °C, and the heart was arrested with cold blood Del Nido cardioplegia. The pulmonary valve was bicuspid and admitted a 6-mm probe (Z = −4). A commisurotomy was performed but resulted in only a 7-mm orifice; therefore, a transannular incision with limited ventriculotomy was performed. The VSD was closed through the ventriculotomy with an expanded polytetrafluoroethylene (ePTFE) patch. Subvalvar RVMB were resected, and the RV outflow tract was reconstructed with another ePTFE patch, easily passing a 10-mm probe. The PFO was reduced from 6 mm to 4 mm. The patient was rewarmed and weaned from CPB (CPB 193 min, cross clamp 78 min).

Immediately on weaning, we noted significant arterial O₂ desaturations (approx. 75%), requiring central venous pressures greater than 15 mm Hg to maintain systemic blood pressure. Assuming a pulmonary hypertensive crisis, inhaled nitric oxide (iNO) was started. Cardiopulmonary bypass was reinitiated and the patient was transfused to increase the hematocrit to 40; CPB time was 23 min. After weaning from CPB a second time with continued severe cyanosis, we were suspicious of a large right-to-left shunt; therefore, a TEE with agitated saline study was performed. This showed instantaneous filling of the left atrium and ventricle from the right atrium, with moderate filling of the right ventricle (Figure 1). We therefore returned to CPB, cooled the patient to 32 °C, and rearrested the heart. Our intent was to close the residual PFO; however, we were uncertain that the small PFO would explain the large atrial-level right-to-left shunt seen on the agitated saline study. We probed the CS and found an undiagnosed unroofed CS defect emptying into the left atrium (Figure 2). There was no evidence of a left SVC. The unroofed CS did not have a distinct opening in the left atrium that could be easily patched. Therefore, the orifice of the CS was closed with an autologous pericardial patch. The patient was rewarmed and weaned from CPB a third time (CPB 78 min, cross clamp 45 min). Arterial oxygen saturations were >92%. Transesophageal echocardiogram showed trivial tricuspid regurgitation, expected severe pulmonary valve insufficiency, no RV outflow tract obstruction, no residual VSD, and a small PFO with right-to-left shunting.

OPEN IN VIEWER

Figure 1.

Clip from an intraoperative transesophageal echocardiogram showing an agitated saline test with instantaneous filling of the left atrium and ventricle, confirming the presence of a large undiagnosed right-to-left shunt.

OPEN IN VIEWER

Figure 2.

Intraoperative photo, looking into the open right atrium. A 3-mm probe was passed though the coronary sinus and can be seen in the LA through an enlarged PFO. LA, left atrium; PFO, patent foramen ovale.

The patient was transferred to the intensive care unit intubated and on low dose inotropic support. He had RV dysfunction requiring iNO for several days with eventual transition to oral sildenafil. He was extubated on postoperative day 2 and was ultimately discharged home on postoperative day 14.

Comment

Overall operative mortality for VSD repair is quite low, with durable long-term outcomes. ³ In this case, we had difficulty weaning from CPB after VSD repair because of unexpected severe cyanosis, ultimately found to be due to an undiagnosed atrial septal Coronary Sinus (CS) defect. The CS is formed by the confluence of the great cardiac vein and multiple smaller venous systems, draining roughly 60% of the total coronary circulating volume. It originates along with the posterior heart between the left atrium and left ventricle and empties into the inferior aspect of the right atrium. Although rare, congenital anatomic variants of the CS exist, with a reported incidence of <1% of all atrial septal defects. Coronary sinus defects are grouped into four subtypes, based on the degree of unroofing (complete vs partial) and the presence or absence of a persistent LSVC. ⁴ A direct communication between the CS and left atrium is a rare type of atrial septal defect and is commonly associated with a persistent LSVC. ⁵ Thus, in patients in whom a persistent LSVC is found, it is important to ensure that the CS is intact by placing a left upper extremity peripheral IV and performing an agitated saline study with the preoperative TEE. If a persistent LSVC is identified in a patient with CS defect, one must be sure to investigate the presence of other defects. Raghib Syndrome is a rare disorder that involves a persistent LSVC terminating in the left atrium, absence of the CS, and an atrial septal defect located in the posteroinferior angle of the atrial septum in the position normally occupied by the right atrial ostium of the CS. ⁶ This results in a right-to-left shunt, and patients with Raghib Syndrome have an increased rate of cryptogenic stroke. ⁷ In our patient, there was neither a persistent LSVC nor CS ostial atresia, and we observed a right-to-left shunt secondary to diastolic dysfunction of the right ventricle and a patent CS os in the presence of an unroofed CS defect.

Multiple techniques exist to repair an unroofed CS. ⁸ The presence of a persistent LSVC has important implications for repair options. If the left SVC drains directly to the left atrium, the two main options are creating a left Glenn or an intra-atrial baffle directing the left SVC to the right atrium. In the absence of a persistent LSVC, the unroofed CS can be reroofed with a patch or via direct repair within the left atrium. Alternatively, the CS ostium can be closed, leaving the CS partially unroofed into the left atrium resulting in a small right-to-left intracardiac shunt that is not clinically significant. This was the reasoning for the repair done in our patient.

In the current case, the CS shunt was not diagnosed preoperatively. The patient was already desaturated due to the severe pulmonary stenosis and bidirectional shunting from the VSD, concealing the right-to-left shunt occurring at the level of the unroofed CS. It was only after the VSD closure and repair of the pulmonary stenosis that the unroofed CS shunt physiology was “unmasked.” The right-to-left shunt through the unroofed CS was secondary to a noncompliant right ventricle which was hypertrophied from seven months of pulmonary stenosis and RVOT obstruction. This led to the unexpected finding of cyanosis despite adequate hemodynamics while weaning from CPB. This is supported by an uneventful wean from CPB after the atrial-level shunt was repaired. This case highlights a strategy for evaluating unexpected cyanosis after a relatively standard congenital heart procedure.