Left ventricular assist device (LVAD) implantation is considered an important treatment option for patients with advanced heart failure, and can lead to recovery or be a bridge to transplantation. However, mechanical circulatory support with an LVAD may be associated with complications, including haemolysis, ventricular arrhythmias, thromboembolism and hypoxaemia.1 Due to abnormal physiological reactions precipitated by the implantation of an LVAD, right-to-left atrial shunting in patients with pre-existing patent foramen ovale (PFO) or iatrogenic atrial septal defect (ASD) may cause critical hypoxaemia.
In most cases, transcatheter closure of an ASD under echocardiographic guidance using fluoroscopy is the standard technique. Transcatheter ASD closure using transoesophageal echocardiography (TOE) without fluoroscopy has been discussed as an alternative in efforts to minimise the risks of radiation.2
We report on LVAD placement complicated by critical hypoxaemia secondary to a large iatrogenic ASD that was successfully treated bedside using TOE-guided iatrogenic ASD closure without fluoroscopy.
Case Report
A 72-year-old man presented to the emergency department for worsening shortness of breath on exertion and bilateral lower extremity oedema. On examination, he was dyspnoeic with jugular vein distension and pulmonary rales. His extensive medical history included permanent AF, moderate pulmonary hypertension, end-stage heart failure secondary to idiopathic cardiomyopathy, severe left ventricular systolic dysfunction (ejection fraction 25–29%), severe mitral regurgitation after placement of a MitraClip (Abbott Vascular), biventricular ICD and long-term infusion of inotropic milrinone.
Two months previously, the patient had been admitted to the same hospital for acute-on-chronic systolic heart failure and diuresed approximately 25 l of fluid. At the time, the patient successfully underwent two MitraClip placements at the A2 and P2 leaflets with improvements in functional capacity and the severity of mitral regurgitation. During the current admission, transthoracic echocardiography showed a severely dilated left ventricular cavity with normal wall thickness, an ejection fraction of <20%, normal wall motion, Doppler findings consistent with decreased left ventricular diastolic compliance and elevated left atrial pressure and mild-to-moderate mitral regurgitation with well-seated clips. The patient was stabilised on 0.3 mg/kg/min intravenous milrinone and a bumetanide drip.
This patient was presented to the medical review board of HCA Houston Medical Center for LVAD placement because he was not a candidate for transplantation due to his age and significant comorbidities. On day 13 after presentation, the patient had a femoral Swan-Ganz catheter placed in preparation for the upcoming LVAD implantation, and measurements were compatible with left-sided heart failure (pulmonary capillary wedge pressure 18 mmHg), with mildly elevated pulmonary artery systolic pressure (26 mmHg) and right atrial pressure (9 mmHg).
On day 14 after presentation, the patient underwent implantation of an Abbott HeartMate 3 LVAD via cannulation of the left ventricular apex for inflow and ascending thoracic aorta for outflow, along with primary closure of the ASD. Intraoperatively, the quality of the atrial septal tissue was found to be extremely poor, with a tear in the atrial septum extending from the inferior aspect of the fossa ovalis terminating a few centimetres from the coronary sinus, likely related to the transseptal puncture performed during the previous MitraClip procedure. Primary suture closure was attempted intraoperatively, but the patient remained hypoxic, and postoperative findings showed an improved but persistent iatrogenic ASD.
The patient was transferred to the cardiovascular intensive care unit for ventilator and vasopressor support with an LVAD flow rate of 2.8 l/min. Overnight, the patient became hypoxic, with a PaO2 of 41.5 mmHg on 100% FiO2 and oxygen saturation ranging from 80% to 88%. TOE was performed and revealed a mid-ASD measuring 6–7 mm with continuous right-to-left atrial shunting (Figures 1 and 2).
Due to concerns of paradoxical embolisation and worsening hypoxia, emergency bedside ASD closure using TOE was performed. Using right femoral access, a 9 Fr sheath was placed and the catheter was advanced into the atrial septum under 3D TOE guidance without fluoroscopy. A 10 mm Amplatzer atrial septal occluder was advanced through the sheath, and left and right atrial discs were deployed with no residual shunt noted (Figure 3). The procedure was deemed successful, with SaO2 improving from 60% to 100% and complete abolition of the shunt. Following closure, PaO2 increased markedly to 267.5 mmHg on 100% FiO2. The patient was stabilised with resolving hypoxaemia. However, due to the patient’s significant comorbidities and prolonged hospitalisation, his family eventually elected for hospice care. Figure 4 presents a visual timeline summarising the patient’s hospital stay, beginning at admission and highlighting key clinical events and interventions through to discharge.
Discussion
LVAD implantation has substantial survival benefits, with improved quality of life compared with medical therapy.3 However, postoperative complications can occur, one of which is the development of worsening right-to-left atrial shunting, especially in patients with pre-existing PFO or iatrogenic ASD.
The physiological change induced by LVAD implantation is secondary to increased unloading of the left ventricle via the outflow cannula, thereby decreasing left atrial pressures.1 This change may exacerbate shunting, causing increasing hypoxaemia in patients already in a hypoxaemic state. Other causes of hypoxia when using LVAD support may include primary pulmonary aetiologies (e.g. pulmonary oedema) and right ventricular failure.4
Right-to-left shunting can be encountered in patients with a previously undiagnosed PFO. However, iatrogenic ASD may be another infrequent cause of the right-to-left atrial shunt. Iatrogenic ASD is a known sequela of transcatheter edge-to-edge repair of the mitral valve with MitraClip because the procedure requires left atrial access through a transseptal puncture. Although often transient, iatrogenic ASD has been noted in 43–82% of patients 1 month after MitraClip placement.5 Persistent iatrogenic ASD after MitraClip is associated with more challenging anatomy, with a higher degree of residual mitral regurgitation and higher left atrial pressures.6 Recognition of the presence of iatrogenic ASD or PFO is an important step in the management of patients in whom LVAD implantation is planned. Generally, an intraoperative bubble study via surface or intracardiac echocardiography is an important assessment for the presence of PFO or iatrogenic ASD during LVAD placement.5 Management in such cases generally includes primary closure of the iatrogenic ASD intraoperatively when implanting the LVAD.5 In our patient, severe hypoxaemia and haemodynamic instability after LVAD implantation prompted immediate percutaneous ASD closure; the persistent shunt likely reflected extremely poor atrial septal tissue despite attempted primary closure.
Immediate ASD closure is rarely recommended but should be considered in patients who become acutely hypoxic secondary to right-to-left shunting.5 In such cases, percutaneous closure guided by intracardiac echocardiography or TOE, with the use of fluoroscopy, is common practice. Unlike PFO, iatrogenic ASDs are generally more acute in onset and the risk of acute decompensation following closure appears to be lower.6 In our case, the patient was haemodynamically unstable and, due to the challenges of transporting him to the cardiac cath lab, percutaneous ASD closure was performed at the bedside using TOE guidance alone, without fluoroscopy. Typically, the combined use of TOE and fluoroscopic imaging provides complementary information that facilitates ASD closure.
Several reports have studied the safety of TOE-guided percutaneous ASD closure without fluoroscopy. These reports have been mostly in the setting of paediatric patients to minimise the risk of radiation.7,8 In a single-centre retrospective study (n=130) comparing TOE-guided ASD closure with versus without fluoroscopy, Xu et al. reported closure without fluoroscopy to be safe and effective in the adult population.9 Postoperative complications were similar between the two groups, and no residual shunt, occlusion device shedding or pericardial effusion was noted in any patient.9
The present case report highlights how TOE-guided ASD closure without fluoroscopy can be performed successfully as an alternative intervention in an emergency setting. Because this is a single case, generalisability is limited. In practice, such procedures would be expected to require specialised centres with experienced operators, access to multidisciplinary support and the resources to manage potential complications.
Conclusion
LVAD implantation in the presence of an ASD can lead to significant right-to-left shunting, predisposing patients to critical hypoxaemia. Transcatheter closure of an ASD under echocardiographic guidance using fluoroscopy is considered the standard of care in such cases. However, transcatheter ASD closure using TOE without fluoroscopy has been reported to be a successful and an effective alternative in previous studies.2
We report the first successful TOE-guided ASD closure without fluoroscopy at a patient’s bedside in an emergency and unstable setting. This case demonstrates that fluoroscopy-free, bedside TOE-guided ASD closure can be life saving and should be considered in unstable patients when immediate access and patient transfer to the cath lab is not possible. Future research and real-world data could help validate this as a standard option in specific emergency care settings in which access to conventional fluoroscopy may be limited.