Abstract
Transcatheter techniques are emerging for left atrial (LA) decompression under venoarterial extracorporeal membrane oxygenation (VA-ECMO). We aimed to assess whether balloon atrioseptostomy (BAS) is a safe and efficient strategy.
All patients who underwent percutaneous static BAS under VA-ECMO at four tertiary institutions were retrospectively reviewed.
From 2000 to 2014, BAS was performed in 64 patients (32 adults and 32 children). Indications for ECMO support included acute myocarditis (31.3%) and non-myocarditis cardiac disease, mostly end-stage dilated cardiomyopathy (32.8%). BAS was required because of pulmonary oedema/haemorrhage and left ventricular (LV) distension. The mean balloon diameter was 21.8 ± 8.4mm. Adequate LA decompression was achieved in all patients. Mean LA pressure fell from 24.2 ± 6.9 mmHg to 7.8 ± 2.6 mmHg (p < 0.001). The left-to-right atrial pressure gradient fell from 17.2 ± 7.1 mmHg to 0.09 ± 0.5 mmHg (p < 0.001). Echocardiography showed an unrestrictive left-to-right atrial shunting in all patients. Improvement of day 1 chest X-ray was observed in 76.6% of patients, clinical status in 98.4% of patients and pulmonary haemorrhage in 14 out of 14 patients. Complications occurred in 9.4% of patients, representing pericardial effusion, fast atrial fibrillation, ventricular fibrillation requiring defibrillation, transient complete heart block and femoral venous dissection requiring covered stent placement. In the 37 (57.8%) patients who were successfully decannulated, the median ECMO duration was 9 (range: 4–24) days. After a median follow-up of 12.3 (range: 0.1–142) months, 35.9% patients died, 17.2% received a LV assist device as a bridge to transplantation, 31.2% were transplanted and 56.2% were home discharged and alive.
Percutaneous BAS may be a safe and efficient strategy for discharging the LA in both adults and children supported by VA-ECMO.
Introduction
Venoarterial extracorporeal membrane oxygenation (VA-ECMO) is widely used for oxygenation and circulatory support for neonates, children and adults whose hearts and lungs can no longer provide adequate physiologic support, despite conventional therapies.1–5 However, persistently high left ventricular (LV) filling pressures and volumes can lead to high myocardial oxygen demand,6 such that VA-ECMO alone may not significantly reduce wall stress.7,8 This may compromise myocardial recovery and prolong the resulting lung injury unless the left heart is vented or unloaded. Several techniques have been described in order to decompress the left heart, including intra-aortic balloon pumps (IABPs), axial flow catheters (Impella®; Abiomed, Inc., Danvers, MA, USA) and left atrial (LA)-to-femoral artery bypass (TandemHeart®; Cardiac Assist, Inc., Pittsburgh, PA, USA).9–12 LA decompression has also been achieved by the transcatheter creation of an atrial septal defect (ASD) using different approaches, including vent placement, static balloon dilation and stent implantation.13–18
In this retrospective multicentre study, we aimed to assess the safety and efficiency of percutaneous balloon atrial septostomy (BAS) for left heart decompression in patients on VA-ECMO support.
Methods
All patients who underwent percutaneous BAS for LA decompression at the four participating institutions (Royal Brompton Hospital in London, UK; Marie-Lannelongue Hospital in Paris, France; Haut Lévèque Hospital in Bordeaux, France; and Columbia University Medical Center in New York, NY, USA) from 2000 to 2014 were included in the study. Demographic variables included age, weight, cardiac diagnosis and other significant medical problems. Infants were defined as patients less than 2 years of age and children as patients between 2 and 18 years of age.
The study’s primary end point was defined as procedural outcome. Secondary end points were short- and long-term outcomes. Procedural success was defined as: (a) reduction in the left-to-right atrial pressure gradient to less than 5 mmHg; and (b) unrestrictive left-to-right shunting through the created ASD on echocardiography. Short- and long-term outcomes included post-BAS clinical status, chest X-ray progression, ECMO duration, need of a LV assist device (LVAD), heart transplantation and death. A composite outcome was defined as LVAD, orthotopic heart transplantation (OHT) or death. LV function recovery was defined as LV ejection fraction >40% or fractional shortening >25% in paediatric patients after ECMO.
Our research was conducted according to the World Medical Association Declaration of Helsinki. The Institutional Review Boards at each participating institution approved the study with a consent waiver.
ECMO and BAS procedure
ECMO indications and characteristics were recorded. Cannulation sites, cannulation techniques and cannula sizes were at the discretion of the cardiothoracic surgeon or the interventional cardiologist. Anticoagulation during ECMO support was managed by the ECMO team according to each centre’s protocol. The decision and timing regarding performing percutaneous LA decompression were set by the multidisciplinary heart team managing the patient. Indications for BAS included refractory cardiogenic pulmonary oedema and the presence of LA and LV distention on transthoracic echocardiography, despite maximal pharmacologic support. Pulmonary oedema was defined as respiratory distress with or without pulmonary haemorrhage, bilateral moist rales at chest auscultation, chest X-ray signs of alveolar oedema and need of either non-invasive ventilation support or endotracheal intubation with mechanical ventilation.19
BAS procedure characteristics including type of balloon and maximum diameter used, radiation dose and procedural complications were recorded. Pre-BAS mean right atrial (RA) pressure as well as the mean pre-BAS LA pressure immediately after the trans-septal puncture and after balloon inflation were recorded. The BAS technique was similar in the four participating institutions. All procedures were performed in the catheterisation laboratory under fluoroscopy with or without transthoracic or transoesophageal echocardiography guidance. BAS procedures were performed via a femoral venous approach using trans-septal puncture with a Brockenbrough needle followed by static balloon dilation of the atrial septum.20 Static balloon dilation was performed over a guidewire parked in the LA or a pulmonary vein. The mid-portion of the balloon was positioned across the atrial septum and inflated with diluted contrast until disappearance of the balloon waist on fluoroscopy. The type of balloon and maximum balloon diameter used were at the discretion of the interventionist. Additional interventions such as endomyocardial biopsies or angiography were performed when deemed necessary.
Statistical analysis
Clinical and biological data are expressed as a median (range) for continuous variables or as a number (percentage) for categorical variables. Non-parametric Mann–Whitney tests were used for quantitative variables (as quantitative variables do not fit with the normal distribution according to the Shapiro–Wilk test). χ2 tests and Fisher’s exact tests were used for qualitative variables in order to assess differences between subgroups when appropriate. Multivariate analyses were performed for each outcome using a forward stepwise logistic regression. Statistical significance was set at p < 0.05. Data analysis was performed using the STATA® software (StataCorp LP, College Station, TX, USA).
Results
Patient characteristics
BAS was performed in 64 patients supported by ECMO, of which 32 (50%) were adults. The median age and weight for the adult population were 37.0 years (range: 18–72 years) and 65.0 kg (range: 44–86 kg), respectively. Among the 32 children, including three infants, the median age was 8.0 years (range: 0.3–17.9 years) and the median weight was 29.0 kg (range: 5.3–69.9 kg). VA-ECMO was instituted because of progressive haemodynamic deterioration/cardiogenic shock in 57 (89.1%) patients and during or following cardiopulmonary resuscitation for in-hospital cardiac arrest in seven (10.9%) patients.
The cardiac conditions leading to ECMO cannulation included myocarditis or suspected myocarditis in 20 (31.3%) patients, including two adults (6.2%) and 18 children (56.2%), and non-myocarditis in 44 (68.7%) patients (Table 1). Patients were cannulated via neck vessels (15.6%), femoral vessels (70.3%) or both neck and femoral vessels (14.1%). ECMO was placed percutaneously (at the bedside or under fluoroscopic guidance) in 44 (68.7%) patients. In 19 (29.7%) spontaneously breathing patients, BAS was performed with local anaesthesia. No patient required concomitant IABP or Impella support. Associated medical problems were found in 10 (15.6%) patients, including atrial fibrillation (n = 4), diabetes mellitus (n = 3), chronic thromboembolic pulmonary hypertension (n = 1), asthma (n = 1), tuberculosis (n = 1), breast neoplasia (n = 1), thyrotoxicosis (n = 1) and stroke (n = 1).
Cardiac conditions leading to extracorporeal membrane oxygenation cannulation
| Condition | Patients (n = 64) |
|---|---|
| Myocarditis or suspected myocarditis, n (%) | 20 (31.3) |
| Non-myocarditis cardiac disease, n (%) | 44 (68.7) |
| End-stage dilated cardiomyopathy, n (%) | 21 (32.8) |
| Ischaemic cardiomyopathy, n (%) | 8 (12.5) |
| Postcardiotomy shock, n (%) | 4 (6.2) |
| Miscellaneous condition with left ventricular dysfunction, n (%) | 11 (17.2) |
| Rheumatic heart disease, n (%) | 3 (4.7) |
| Repaired congenital heart disease, n (%) | 2 (3.1) |
| Previous orthotopic heart transplantation, n (%) | 6 (9.4) |
| For hypoplastic left heart syndrome | 2 |
| For pulmonary atresia with intact ventricular septum | 1 |
| For restrictive cardiomyopathy | 1 |
| For primitive dilated cardiomyopathy | 1 |
| For peripartum cardiomyopathy | 1 |
| Condition | Patients (n = 64) |
|---|---|
| Myocarditis or suspected myocarditis, n (%) | 20 (31.3) |
| Non-myocarditis cardiac disease, n (%) | 44 (68.7) |
| End-stage dilated cardiomyopathy, n (%) | 21 (32.8) |
| Ischaemic cardiomyopathy, n (%) | 8 (12.5) |
| Postcardiotomy shock, n (%) | 4 (6.2) |
| Miscellaneous condition with left ventricular dysfunction, n (%) | 11 (17.2) |
| Rheumatic heart disease, n (%) | 3 (4.7) |
| Repaired congenital heart disease, n (%) | 2 (3.1) |
| Previous orthotopic heart transplantation, n (%) | 6 (9.4) |
| For hypoplastic left heart syndrome | 2 |
| For pulmonary atresia with intact ventricular septum | 1 |
| For restrictive cardiomyopathy | 1 |
| For primitive dilated cardiomyopathy | 1 |
| For peripartum cardiomyopathy | 1 |
Cardiac conditions leading to extracorporeal membrane oxygenation cannulation
| Condition | Patients (n = 64) |
|---|---|
| Myocarditis or suspected myocarditis, n (%) | 20 (31.3) |
| Non-myocarditis cardiac disease, n (%) | 44 (68.7) |
| End-stage dilated cardiomyopathy, n (%) | 21 (32.8) |
| Ischaemic cardiomyopathy, n (%) | 8 (12.5) |
| Postcardiotomy shock, n (%) | 4 (6.2) |
| Miscellaneous condition with left ventricular dysfunction, n (%) | 11 (17.2) |
| Rheumatic heart disease, n (%) | 3 (4.7) |
| Repaired congenital heart disease, n (%) | 2 (3.1) |
| Previous orthotopic heart transplantation, n (%) | 6 (9.4) |
| For hypoplastic left heart syndrome | 2 |
| For pulmonary atresia with intact ventricular septum | 1 |
| For restrictive cardiomyopathy | 1 |
| For primitive dilated cardiomyopathy | 1 |
| For peripartum cardiomyopathy | 1 |
| Condition | Patients (n = 64) |
|---|---|
| Myocarditis or suspected myocarditis, n (%) | 20 (31.3) |
| Non-myocarditis cardiac disease, n (%) | 44 (68.7) |
| End-stage dilated cardiomyopathy, n (%) | 21 (32.8) |
| Ischaemic cardiomyopathy, n (%) | 8 (12.5) |
| Postcardiotomy shock, n (%) | 4 (6.2) |
| Miscellaneous condition with left ventricular dysfunction, n (%) | 11 (17.2) |
| Rheumatic heart disease, n (%) | 3 (4.7) |
| Repaired congenital heart disease, n (%) | 2 (3.1) |
| Previous orthotopic heart transplantation, n (%) | 6 (9.4) |
| For hypoplastic left heart syndrome | 2 |
| For pulmonary atresia with intact ventricular septum | 1 |
| For restrictive cardiomyopathy | 1 |
| For primitive dilated cardiomyopathy | 1 |
| For peripartum cardiomyopathy | 1 |
Balloon atrial septostomy
All patients had pulmonary oedema at the time of the BAS. Macroscopic pulmonary haemorrhage was present in 14 (21.9%) patients. LA distention was reported on echocardiogram in 56 (87.5%) patients. BAS was performed at a median time of 1.46 days (range: 0–12 days) from ECMO implantation. In eight (12.5%) patients, ECMO placement and subsequent static BAS were performed during the same procedure, as part of resuscitation manoeuvers.
Trans-septal puncture was guided by transoesophageal echocardiography in nine of 64 (14.1%) patients. Static balloon dilation of the atrial septum was performed using an Inoue balloon in 41 of 64 (64.1%) patients, whereas different over-the-wire angioplasty balloons were used in 23 of 64 (35.9%) patients, with a mean of 2.3 ± 0.8 balloons per procedure. The maximum balloon diameters used were 18 mm for infants, 28 mm for children and 30 mm for adults. Cutting balloons were not used in any patient. The median fluoroscopy time and radiation dose were 13.75 minutes (range: 3.0–27.2 minutes) and 1460 μGy/m2 (range: 248.4–8033 μGy/m2), respectively. Additional procedures were performed in 29 (45.3%) patients: 36 endomyocardial biopsies leading to the diagnosis of viral myocarditis in 14 (38.9%) patients; and aortic angiograms in 12 paediatric patients in order to rule out a potential anomalous left coronary artery from the pulmonary artery.
Procedural and early clinical outcomes
Procedural success was achieved in all patients. The mean LA pressure fell from 24.2 ± 6.9 mmHg to 7.8 ± 2.6 mmHg (p < 0.0001). The left-to-right atrial pressure gradient fell from 17.2 ± 7.1 mmHg to 0.09 ± 0.5 mmHg (p < 0.0001). At the end of the procedure, echocardiography showed unrestrictive left-to-right shunting through the ASD with no residual LA distension in all patients. Improvement on chest X-ray was observed in 49 of 64 (76.6%) patients (Figures 1 and 2), and improvement of clinical status was observed in 63 of 64 (98.4%) patients. Pulmonary haemorrhage resolved in 14 of 14 patients within 48 hours. No patient required reintervention for either redilate ASD or repeat BAS.
Primitive end-stage dilated cardiomyopathy (A) complicated by cardiogenic shock and refractory pulmonary oedema (B) in a 12-year-old, 48-kg boy supported by percutaneous femoral venoarterial extracorporeal membrane oxygenation (C). Balloon atrial septostomy using various over-the-wire angioplasty balloons (D and E) enabled quick haemodynamic and clinical improvement (day 1 chest X-ray; F)
Primitive end-stage dilated cardiomyopathy complicated by cardiogenic shock and refractory pulmonary oedema in a 55-year-old, 62-kg patient supported by percutaneous femoral venoarterial extracorporeal membrane oxygenation. Trans-septal puncture with a Brockenbrough needle (A) and balloon atrial septostomy using a 28-mm Inoue balloon (B and C) enabled quick haemodynamic improvement. Examination of the explanted heart at the time of heart transplantation 4 months later showed an unrestrictive atrial septal defect (view from the right atrium; D)
Peri-procedural complications occurred in six (9.4%) patients, representing pericardial effusion without haemodynamic compromise that resolved spontaneously due to needle perforation of the LA in two patients, fast atrial fibrillation requiring emergent electrical cardioversion, ventricular fibrillation requiring defibrillation, transient complete heart block and iliac vein perforation requiring covered stent placement in one patient each.
Long-term outcomes
During a median follow-up period of 12.3 months (range: (0.1–142 months), 23 (35.9%) patients died, 11 (17.2%) patients received a LVAD as a bridge to transplantation and 20 (31.2%) patients were transplanted. Sudden death occurred in an 11-year-old girl with primitive dilated cardiomyopathy 3.2 years after ECMO support and LVAD during 4 months as a bridge to recovery. At the latest follow-up, 35 (54.7%) patients were home discharged and alive. In-hospital course, long-term clinical outcomes and complications related to ECMO are summarised in Table 2.
Clinical outcomes of balloon atrioseptostomy patients
| Outcomes | BAS patients (n = 64) |
|---|---|
| Extracorporeal membrane oxygenation complications | |
| Patients, n (%) | 16 (25.0) |
| Mechanical complication | 10 (15.6) |
| Compartment syndrome, fasciotomy, n (%) | 6 (9.4) |
| Leg necrosis, amputation, n (%) | 2 (3.1) |
| Vascular injury requiring surgery, n (%) | 2 (3.1) |
| Neurologic complication | 6 (9.4) |
| Intra-cerebral haemorrhage, n (%) | 5 (7.8) |
| Stroke, n (%) | 1 (1.6) |
| In-hospital course | |
| Extracorporeal membrane oxygenation duration, days, median (minimum–maximum) | 9 (4–24) |
| Extracorporeal membrane oxygenation duration after balloon atrioseptostomy, days, median (minimum–maximum) | 6 (3.3–24) |
| In-hospital mortality, n (%) | 22 (34.4) |
| Multiorgan failure, n (%) | 9 (14.1) |
| Neurologic injury, n (%) | 7 (10.9) |
| Infection, n (%) | 4 (6.2) |
| Cardiac failure, n (%) | 2 (3.1) |
| Long-term follow-up | |
| Follow-up duration, median (minimum–maximum) | 12.3 (0.1–142) |
| Death, n (%) | 23 (35.9) |
| Late death, n (%) | 1 (1.6) |
| Ventricular assist device, n (%) | 11 (17.2) |
| Orthotopic heart transplantation, n (%) | 20 (31.2) |
| Left ventricular ejection fraction >40%, n (%) | 18 (28.1) |
| Home discharged and alive, n (%) | 36 (56.2) |
| Outcomes | BAS patients (n = 64) |
|---|---|
| Extracorporeal membrane oxygenation complications | |
| Patients, n (%) | 16 (25.0) |
| Mechanical complication | 10 (15.6) |
| Compartment syndrome, fasciotomy, n (%) | 6 (9.4) |
| Leg necrosis, amputation, n (%) | 2 (3.1) |
| Vascular injury requiring surgery, n (%) | 2 (3.1) |
| Neurologic complication | 6 (9.4) |
| Intra-cerebral haemorrhage, n (%) | 5 (7.8) |
| Stroke, n (%) | 1 (1.6) |
| In-hospital course | |
| Extracorporeal membrane oxygenation duration, days, median (minimum–maximum) | 9 (4–24) |
| Extracorporeal membrane oxygenation duration after balloon atrioseptostomy, days, median (minimum–maximum) | 6 (3.3–24) |
| In-hospital mortality, n (%) | 22 (34.4) |
| Multiorgan failure, n (%) | 9 (14.1) |
| Neurologic injury, n (%) | 7 (10.9) |
| Infection, n (%) | 4 (6.2) |
| Cardiac failure, n (%) | 2 (3.1) |
| Long-term follow-up | |
| Follow-up duration, median (minimum–maximum) | 12.3 (0.1–142) |
| Death, n (%) | 23 (35.9) |
| Late death, n (%) | 1 (1.6) |
| Ventricular assist device, n (%) | 11 (17.2) |
| Orthotopic heart transplantation, n (%) | 20 (31.2) |
| Left ventricular ejection fraction >40%, n (%) | 18 (28.1) |
| Home discharged and alive, n (%) | 36 (56.2) |
Clinical outcomes of balloon atrioseptostomy patients
| Outcomes | BAS patients (n = 64) |
|---|---|
| Extracorporeal membrane oxygenation complications | |
| Patients, n (%) | 16 (25.0) |
| Mechanical complication | 10 (15.6) |
| Compartment syndrome, fasciotomy, n (%) | 6 (9.4) |
| Leg necrosis, amputation, n (%) | 2 (3.1) |
| Vascular injury requiring surgery, n (%) | 2 (3.1) |
| Neurologic complication | 6 (9.4) |
| Intra-cerebral haemorrhage, n (%) | 5 (7.8) |
| Stroke, n (%) | 1 (1.6) |
| In-hospital course | |
| Extracorporeal membrane oxygenation duration, days, median (minimum–maximum) | 9 (4–24) |
| Extracorporeal membrane oxygenation duration after balloon atrioseptostomy, days, median (minimum–maximum) | 6 (3.3–24) |
| In-hospital mortality, n (%) | 22 (34.4) |
| Multiorgan failure, n (%) | 9 (14.1) |
| Neurologic injury, n (%) | 7 (10.9) |
| Infection, n (%) | 4 (6.2) |
| Cardiac failure, n (%) | 2 (3.1) |
| Long-term follow-up | |
| Follow-up duration, median (minimum–maximum) | 12.3 (0.1–142) |
| Death, n (%) | 23 (35.9) |
| Late death, n (%) | 1 (1.6) |
| Ventricular assist device, n (%) | 11 (17.2) |
| Orthotopic heart transplantation, n (%) | 20 (31.2) |
| Left ventricular ejection fraction >40%, n (%) | 18 (28.1) |
| Home discharged and alive, n (%) | 36 (56.2) |
| Outcomes | BAS patients (n = 64) |
|---|---|
| Extracorporeal membrane oxygenation complications | |
| Patients, n (%) | 16 (25.0) |
| Mechanical complication | 10 (15.6) |
| Compartment syndrome, fasciotomy, n (%) | 6 (9.4) |
| Leg necrosis, amputation, n (%) | 2 (3.1) |
| Vascular injury requiring surgery, n (%) | 2 (3.1) |
| Neurologic complication | 6 (9.4) |
| Intra-cerebral haemorrhage, n (%) | 5 (7.8) |
| Stroke, n (%) | 1 (1.6) |
| In-hospital course | |
| Extracorporeal membrane oxygenation duration, days, median (minimum–maximum) | 9 (4–24) |
| Extracorporeal membrane oxygenation duration after balloon atrioseptostomy, days, median (minimum–maximum) | 6 (3.3–24) |
| In-hospital mortality, n (%) | 22 (34.4) |
| Multiorgan failure, n (%) | 9 (14.1) |
| Neurologic injury, n (%) | 7 (10.9) |
| Infection, n (%) | 4 (6.2) |
| Cardiac failure, n (%) | 2 (3.1) |
| Long-term follow-up | |
| Follow-up duration, median (minimum–maximum) | 12.3 (0.1–142) |
| Death, n (%) | 23 (35.9) |
| Late death, n (%) | 1 (1.6) |
| Ventricular assist device, n (%) | 11 (17.2) |
| Orthotopic heart transplantation, n (%) | 20 (31.2) |
| Left ventricular ejection fraction >40%, n (%) | 18 (28.1) |
| Home discharged and alive, n (%) | 36 (56.2) |
The rate of freedom from death, OHT or LVAD was 26.6% (95% confidence interval: 16.0–39.0). In the univariate analysis, age ⩾10 years, acute myocarditis, end-stage dilated cardiomyopathy, ECMO duration ⩾96 hours prior to BAS and total ECMO duration were identified as risk factors for a composite outcome. However, in multivariate analysis, only age ⩾10 years and acute myocarditis remained significant for a composite outcome (Table 3).
Risk factors analysis for death, ventricular assist device or orthotopic heart transplantation
| Patient characteristics | VAD/OHT or death (n = 47) | Freedom from VAD, OHT or death (n = 17, 26.6%) | p-value univariate | p-value multivariate, OR (95% CI) |
|---|---|---|---|---|
| Male, n (%) | 26 (55.3) | 7 (41.2) | 0.317 | – |
| Age, years | 26 (0.25–72) | 8.8 (1.6–52) | 0.0058 | – |
| Age <10 years, n (%) | 9 (19.1) | 10 (58.8) | 0.002 | 0.073, 0.29 (0.07–1.12) |
| Weight, kg | 58 (5.3–85) | 32 (10–86) | 0.0153 | 0.981, 0.99 (0.94–1.06) |
| Weight <30 kg, n (%) | 9 (19.1) | 7 (41.2) | 0.073 | – |
| BSA, m2 | 0.95 (0.08–1.82) | 0.36 (0.06–1.51) | 0.0104 | 0.225, 2.70 (0.54–13.4) |
| Pulmonary haemorrhage, n (%) | 10 (21.3) | 4 (23.5) | 0.547 | – |
| Cardiogenic shock, n (%) | 43 (91.5) | 14 (82.3) | 0.269 | – |
| Cardiopulmonary resuscitation, n (%) | 4 (8.5) | 3 (17.6) | 0.269 | – |
| Acute myocarditis, n (%) | 39 (83%) | 5 (29%) | <10−3 | 0.002, 0.12 (0.03–0.45) |
| End-stage dilated cardiomyopathy, n (%) | 20 (42.5) | 1 (5.9) | 0.006 | 0.261, 4.21 (0.43–41.1) |
| Postcardiotomy shock, n (%) | 4 (8.5) | 0 (0.0) | 0.281 | – |
| Ischaemic cardiomyopathy, n (%) | 7 (14.9) | 1 (5.9) | 0.313 | – |
| ECMO–BAS time, hours | 24 (0.0–337.5) | 12 (0.3–193.2) | 0.1781 | – |
| ECMO–BAS time <96 hours, n (%) | 33 (70.2) | 16 (94.1) | 0.041 | 0.296, 0.30 (0.03–2.86) |
| Pre-BAS mLA pressure, mmHg | 24 (15–45) 24.8 ± 7.5* | 22 (15–30) 22.1 ± 4.4* | 0.3009 | – |
| Pre-BAS mean LA >30 mmHg, n (%) | 11 (23.4) | 2 (11.7) | 0.259 | – |
| Pre-BAS LA–RA gdt, mmHg | 16 (6–43)* 18.0 ± 7.5 | 17 (8–22)* 15.9 ± 4.9 | 0.4462 | – |
| Pre-BAS LA–RA gdt >20 mmHg, n (%) | 17 (36.2) | 6 (35.3) | 0.949 | – |
| Post-BAS mLA pressure, mmHg | 8 (2–14) 7.9 ± 2.9* | 7 (5–10) 7.4 ± 1.6* | 0.6091 | – |
| LV function recovery, n (%) | 1 (2.1) | 17 (100.0) | <10−3 | – |
| ECMO duration, days | 11 (2–29) 12.0 ± 5.8* | 6.5 (3.7–19.2) 7.9 ± 3.8* | 0.0081 | 0.123, 1.12 (0.97–1.30) |
| ECMO complications, n (%) | 14 (29.8) | 2 (11.7) | 0.124 | – |
| Patient characteristics | VAD/OHT or death (n = 47) | Freedom from VAD, OHT or death (n = 17, 26.6%) | p-value univariate | p-value multivariate, OR (95% CI) |
|---|---|---|---|---|
| Male, n (%) | 26 (55.3) | 7 (41.2) | 0.317 | – |
| Age, years | 26 (0.25–72) | 8.8 (1.6–52) | 0.0058 | – |
| Age <10 years, n (%) | 9 (19.1) | 10 (58.8) | 0.002 | 0.073, 0.29 (0.07–1.12) |
| Weight, kg | 58 (5.3–85) | 32 (10–86) | 0.0153 | 0.981, 0.99 (0.94–1.06) |
| Weight <30 kg, n (%) | 9 (19.1) | 7 (41.2) | 0.073 | – |
| BSA, m2 | 0.95 (0.08–1.82) | 0.36 (0.06–1.51) | 0.0104 | 0.225, 2.70 (0.54–13.4) |
| Pulmonary haemorrhage, n (%) | 10 (21.3) | 4 (23.5) | 0.547 | – |
| Cardiogenic shock, n (%) | 43 (91.5) | 14 (82.3) | 0.269 | – |
| Cardiopulmonary resuscitation, n (%) | 4 (8.5) | 3 (17.6) | 0.269 | – |
| Acute myocarditis, n (%) | 39 (83%) | 5 (29%) | <10−3 | 0.002, 0.12 (0.03–0.45) |
| End-stage dilated cardiomyopathy, n (%) | 20 (42.5) | 1 (5.9) | 0.006 | 0.261, 4.21 (0.43–41.1) |
| Postcardiotomy shock, n (%) | 4 (8.5) | 0 (0.0) | 0.281 | – |
| Ischaemic cardiomyopathy, n (%) | 7 (14.9) | 1 (5.9) | 0.313 | – |
| ECMO–BAS time, hours | 24 (0.0–337.5) | 12 (0.3–193.2) | 0.1781 | – |
| ECMO–BAS time <96 hours, n (%) | 33 (70.2) | 16 (94.1) | 0.041 | 0.296, 0.30 (0.03–2.86) |
| Pre-BAS mLA pressure, mmHg | 24 (15–45) 24.8 ± 7.5* | 22 (15–30) 22.1 ± 4.4* | 0.3009 | – |
| Pre-BAS mean LA >30 mmHg, n (%) | 11 (23.4) | 2 (11.7) | 0.259 | – |
| Pre-BAS LA–RA gdt, mmHg | 16 (6–43)* 18.0 ± 7.5 | 17 (8–22)* 15.9 ± 4.9 | 0.4462 | – |
| Pre-BAS LA–RA gdt >20 mmHg, n (%) | 17 (36.2) | 6 (35.3) | 0.949 | – |
| Post-BAS mLA pressure, mmHg | 8 (2–14) 7.9 ± 2.9* | 7 (5–10) 7.4 ± 1.6* | 0.6091 | – |
| LV function recovery, n (%) | 1 (2.1) | 17 (100.0) | <10−3 | – |
| ECMO duration, days | 11 (2–29) 12.0 ± 5.8* | 6.5 (3.7–19.2) 7.9 ± 3.8* | 0.0081 | 0.123, 1.12 (0.97–1.30) |
| ECMO complications, n (%) | 14 (29.8) | 2 (11.7) | 0.124 | – |
Note: Bracketed ranges represent median (min-max).
*These numbers refer to mean +/- standard deviation.
BSA: body surface area; mLA: mean left atrial; gdt: gradient; OHT: orthotopic heart transplant; VAD: ventricular assist device; OR: odds ratio; CI: confidence interval; LA: left atrial; RA: right atrial.
Risk factors analysis for death, ventricular assist device or orthotopic heart transplantation
| Patient characteristics | VAD/OHT or death (n = 47) | Freedom from VAD, OHT or death (n = 17, 26.6%) | p-value univariate | p-value multivariate, OR (95% CI) |
|---|---|---|---|---|
| Male, n (%) | 26 (55.3) | 7 (41.2) | 0.317 | – |
| Age, years | 26 (0.25–72) | 8.8 (1.6–52) | 0.0058 | – |
| Age <10 years, n (%) | 9 (19.1) | 10 (58.8) | 0.002 | 0.073, 0.29 (0.07–1.12) |
| Weight, kg | 58 (5.3–85) | 32 (10–86) | 0.0153 | 0.981, 0.99 (0.94–1.06) |
| Weight <30 kg, n (%) | 9 (19.1) | 7 (41.2) | 0.073 | – |
| BSA, m2 | 0.95 (0.08–1.82) | 0.36 (0.06–1.51) | 0.0104 | 0.225, 2.70 (0.54–13.4) |
| Pulmonary haemorrhage, n (%) | 10 (21.3) | 4 (23.5) | 0.547 | – |
| Cardiogenic shock, n (%) | 43 (91.5) | 14 (82.3) | 0.269 | – |
| Cardiopulmonary resuscitation, n (%) | 4 (8.5) | 3 (17.6) | 0.269 | – |
| Acute myocarditis, n (%) | 39 (83%) | 5 (29%) | <10−3 | 0.002, 0.12 (0.03–0.45) |
| End-stage dilated cardiomyopathy, n (%) | 20 (42.5) | 1 (5.9) | 0.006 | 0.261, 4.21 (0.43–41.1) |
| Postcardiotomy shock, n (%) | 4 (8.5) | 0 (0.0) | 0.281 | – |
| Ischaemic cardiomyopathy, n (%) | 7 (14.9) | 1 (5.9) | 0.313 | – |
| ECMO–BAS time, hours | 24 (0.0–337.5) | 12 (0.3–193.2) | 0.1781 | – |
| ECMO–BAS time <96 hours, n (%) | 33 (70.2) | 16 (94.1) | 0.041 | 0.296, 0.30 (0.03–2.86) |
| Pre-BAS mLA pressure, mmHg | 24 (15–45) 24.8 ± 7.5* | 22 (15–30) 22.1 ± 4.4* | 0.3009 | – |
| Pre-BAS mean LA >30 mmHg, n (%) | 11 (23.4) | 2 (11.7) | 0.259 | – |
| Pre-BAS LA–RA gdt, mmHg | 16 (6–43)* 18.0 ± 7.5 | 17 (8–22)* 15.9 ± 4.9 | 0.4462 | – |
| Pre-BAS LA–RA gdt >20 mmHg, n (%) | 17 (36.2) | 6 (35.3) | 0.949 | – |
| Post-BAS mLA pressure, mmHg | 8 (2–14) 7.9 ± 2.9* | 7 (5–10) 7.4 ± 1.6* | 0.6091 | – |
| LV function recovery, n (%) | 1 (2.1) | 17 (100.0) | <10−3 | – |
| ECMO duration, days | 11 (2–29) 12.0 ± 5.8* | 6.5 (3.7–19.2) 7.9 ± 3.8* | 0.0081 | 0.123, 1.12 (0.97–1.30) |
| ECMO complications, n (%) | 14 (29.8) | 2 (11.7) | 0.124 | – |
| Patient characteristics | VAD/OHT or death (n = 47) | Freedom from VAD, OHT or death (n = 17, 26.6%) | p-value univariate | p-value multivariate, OR (95% CI) |
|---|---|---|---|---|
| Male, n (%) | 26 (55.3) | 7 (41.2) | 0.317 | – |
| Age, years | 26 (0.25–72) | 8.8 (1.6–52) | 0.0058 | – |
| Age <10 years, n (%) | 9 (19.1) | 10 (58.8) | 0.002 | 0.073, 0.29 (0.07–1.12) |
| Weight, kg | 58 (5.3–85) | 32 (10–86) | 0.0153 | 0.981, 0.99 (0.94–1.06) |
| Weight <30 kg, n (%) | 9 (19.1) | 7 (41.2) | 0.073 | – |
| BSA, m2 | 0.95 (0.08–1.82) | 0.36 (0.06–1.51) | 0.0104 | 0.225, 2.70 (0.54–13.4) |
| Pulmonary haemorrhage, n (%) | 10 (21.3) | 4 (23.5) | 0.547 | – |
| Cardiogenic shock, n (%) | 43 (91.5) | 14 (82.3) | 0.269 | – |
| Cardiopulmonary resuscitation, n (%) | 4 (8.5) | 3 (17.6) | 0.269 | – |
| Acute myocarditis, n (%) | 39 (83%) | 5 (29%) | <10−3 | 0.002, 0.12 (0.03–0.45) |
| End-stage dilated cardiomyopathy, n (%) | 20 (42.5) | 1 (5.9) | 0.006 | 0.261, 4.21 (0.43–41.1) |
| Postcardiotomy shock, n (%) | 4 (8.5) | 0 (0.0) | 0.281 | – |
| Ischaemic cardiomyopathy, n (%) | 7 (14.9) | 1 (5.9) | 0.313 | – |
| ECMO–BAS time, hours | 24 (0.0–337.5) | 12 (0.3–193.2) | 0.1781 | – |
| ECMO–BAS time <96 hours, n (%) | 33 (70.2) | 16 (94.1) | 0.041 | 0.296, 0.30 (0.03–2.86) |
| Pre-BAS mLA pressure, mmHg | 24 (15–45) 24.8 ± 7.5* | 22 (15–30) 22.1 ± 4.4* | 0.3009 | – |
| Pre-BAS mean LA >30 mmHg, n (%) | 11 (23.4) | 2 (11.7) | 0.259 | – |
| Pre-BAS LA–RA gdt, mmHg | 16 (6–43)* 18.0 ± 7.5 | 17 (8–22)* 15.9 ± 4.9 | 0.4462 | – |
| Pre-BAS LA–RA gdt >20 mmHg, n (%) | 17 (36.2) | 6 (35.3) | 0.949 | – |
| Post-BAS mLA pressure, mmHg | 8 (2–14) 7.9 ± 2.9* | 7 (5–10) 7.4 ± 1.6* | 0.6091 | – |
| LV function recovery, n (%) | 1 (2.1) | 17 (100.0) | <10−3 | – |
| ECMO duration, days | 11 (2–29) 12.0 ± 5.8* | 6.5 (3.7–19.2) 7.9 ± 3.8* | 0.0081 | 0.123, 1.12 (0.97–1.30) |
| ECMO complications, n (%) | 14 (29.8) | 2 (11.7) | 0.124 | – |
Note: Bracketed ranges represent median (min-max).
*These numbers refer to mean +/- standard deviation.
BSA: body surface area; mLA: mean left atrial; gdt: gradient; OHT: orthotopic heart transplant; VAD: ventricular assist device; OR: odds ratio; CI: confidence interval; LA: left atrial; RA: right atrial.
All ASD remained patent and unrestrictive at the time of heart transplant, LVAD implantation or last follow-up echocardiogram (Figure 2). Two (11.1%) patients underwent transcatheter ASD closure with a 24-mm and a 26-mm Amplatzer™ Septal Occluder (Saint Jude Medical, St Paul, MN, USA) at 18 and 24 months after transcatheter LA decompression, respectively.
Discussion
To our knowledge, the present study describes the largest reported cohort of patients supported by peripherally cannulated VA-ECMO to undergo successful left heart decompression using percutaneous static BAS (Table 4).14,15,18,21,22
Literature summary of balloon atrioseptostomy procedures in venoarterial extracorporeal membrane oxygenation patients
| Publication | Year | Study design | nb static BAS/nb LA decompression | Adult/paediatric | Age (years) | Weight (kg) | ECMO–BAS timing | Type of balloon used | Size of balloon used (mm) | Procedural adverse events (%) | Adequate LA decompression (%) |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Koenig et al.14 | 1993 | Monocenter retrospective | 4/4 Blade and BAS in 1 patient | Paediatric | 0.03 (0.01–5) | N/A | 2.5 days (0.5–3) | Angioplasty balloons | N/A | 0 | 100 |
| Seib et al.15 | 1999 | Monocenter retrospective | 9/10 Blade and BAS in all 9 patients | Both | 3 (1–24) | 15.3 (9–77) | N/A | Angioplasty balloons | 18 (10–20) | 40 | 100 |
| Kotani et al.21 | 2012 | Monocenter retrospective | 4/23 | Paediatric | 0.48 (0.03–10.8) | 7 (4–34) | 8 hours (4–12) | Angioplasty balloons | 16.5 (13–18) | 0 | 100 |
| O’Byrne et al.22 | 2015 | Monocenter retrospective | 37/37 | Paediatric | 6 (0.01–17) | 18.7 (2.2–90) | 0 days (0–2) | Angioplasty balloons | N/A | 0 | 100 |
| Eastaugh et al.18 | 2015 | Monocenter retrospective | 17/44 Blade and BAS in 3 patients | Both | 12.2 (0.02–18.9) | 35 (3–124) | N/A | Angioplasty balloons | 12 (10–24) | 4.5 | 100 |
| Baruteau et al. (present series) | 2016 | Multicentre retrospective | 64/64 | Both | 18 (0.3–72) | 54.5 (5.3–86) | 1.5 days (0–12) | Angioplasty balloons or Inoue balloons | 28 (6–30) | 9.4 | 98.4 |
| Publication | Year | Study design | nb static BAS/nb LA decompression | Adult/paediatric | Age (years) | Weight (kg) | ECMO–BAS timing | Type of balloon used | Size of balloon used (mm) | Procedural adverse events (%) | Adequate LA decompression (%) |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Koenig et al.14 | 1993 | Monocenter retrospective | 4/4 Blade and BAS in 1 patient | Paediatric | 0.03 (0.01–5) | N/A | 2.5 days (0.5–3) | Angioplasty balloons | N/A | 0 | 100 |
| Seib et al.15 | 1999 | Monocenter retrospective | 9/10 Blade and BAS in all 9 patients | Both | 3 (1–24) | 15.3 (9–77) | N/A | Angioplasty balloons | 18 (10–20) | 40 | 100 |
| Kotani et al.21 | 2012 | Monocenter retrospective | 4/23 | Paediatric | 0.48 (0.03–10.8) | 7 (4–34) | 8 hours (4–12) | Angioplasty balloons | 16.5 (13–18) | 0 | 100 |
| O’Byrne et al.22 | 2015 | Monocenter retrospective | 37/37 | Paediatric | 6 (0.01–17) | 18.7 (2.2–90) | 0 days (0–2) | Angioplasty balloons | N/A | 0 | 100 |
| Eastaugh et al.18 | 2015 | Monocenter retrospective | 17/44 Blade and BAS in 3 patients | Both | 12.2 (0.02–18.9) | 35 (3–124) | N/A | Angioplasty balloons | 12 (10–24) | 4.5 | 100 |
| Baruteau et al. (present series) | 2016 | Multicentre retrospective | 64/64 | Both | 18 (0.3–72) | 54.5 (5.3–86) | 1.5 days (0–12) | Angioplasty balloons or Inoue balloons | 28 (6–30) | 9.4 | 98.4 |
Case reports are not included. Age, weight, ECMO–BAS timing and size of balloon used are expressed in median (range).
BAS: balloon atrioseptostomy; LA: left atrial; ECMO: extracorporeal membrane oxygenation; N/A: not available; nb: number.
Literature summary of balloon atrioseptostomy procedures in venoarterial extracorporeal membrane oxygenation patients
| Publication | Year | Study design | nb static BAS/nb LA decompression | Adult/paediatric | Age (years) | Weight (kg) | ECMO–BAS timing | Type of balloon used | Size of balloon used (mm) | Procedural adverse events (%) | Adequate LA decompression (%) |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Koenig et al.14 | 1993 | Monocenter retrospective | 4/4 Blade and BAS in 1 patient | Paediatric | 0.03 (0.01–5) | N/A | 2.5 days (0.5–3) | Angioplasty balloons | N/A | 0 | 100 |
| Seib et al.15 | 1999 | Monocenter retrospective | 9/10 Blade and BAS in all 9 patients | Both | 3 (1–24) | 15.3 (9–77) | N/A | Angioplasty balloons | 18 (10–20) | 40 | 100 |
| Kotani et al.21 | 2012 | Monocenter retrospective | 4/23 | Paediatric | 0.48 (0.03–10.8) | 7 (4–34) | 8 hours (4–12) | Angioplasty balloons | 16.5 (13–18) | 0 | 100 |
| O’Byrne et al.22 | 2015 | Monocenter retrospective | 37/37 | Paediatric | 6 (0.01–17) | 18.7 (2.2–90) | 0 days (0–2) | Angioplasty balloons | N/A | 0 | 100 |
| Eastaugh et al.18 | 2015 | Monocenter retrospective | 17/44 Blade and BAS in 3 patients | Both | 12.2 (0.02–18.9) | 35 (3–124) | N/A | Angioplasty balloons | 12 (10–24) | 4.5 | 100 |
| Baruteau et al. (present series) | 2016 | Multicentre retrospective | 64/64 | Both | 18 (0.3–72) | 54.5 (5.3–86) | 1.5 days (0–12) | Angioplasty balloons or Inoue balloons | 28 (6–30) | 9.4 | 98.4 |
| Publication | Year | Study design | nb static BAS/nb LA decompression | Adult/paediatric | Age (years) | Weight (kg) | ECMO–BAS timing | Type of balloon used | Size of balloon used (mm) | Procedural adverse events (%) | Adequate LA decompression (%) |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Koenig et al.14 | 1993 | Monocenter retrospective | 4/4 Blade and BAS in 1 patient | Paediatric | 0.03 (0.01–5) | N/A | 2.5 days (0.5–3) | Angioplasty balloons | N/A | 0 | 100 |
| Seib et al.15 | 1999 | Monocenter retrospective | 9/10 Blade and BAS in all 9 patients | Both | 3 (1–24) | 15.3 (9–77) | N/A | Angioplasty balloons | 18 (10–20) | 40 | 100 |
| Kotani et al.21 | 2012 | Monocenter retrospective | 4/23 | Paediatric | 0.48 (0.03–10.8) | 7 (4–34) | 8 hours (4–12) | Angioplasty balloons | 16.5 (13–18) | 0 | 100 |
| O’Byrne et al.22 | 2015 | Monocenter retrospective | 37/37 | Paediatric | 6 (0.01–17) | 18.7 (2.2–90) | 0 days (0–2) | Angioplasty balloons | N/A | 0 | 100 |
| Eastaugh et al.18 | 2015 | Monocenter retrospective | 17/44 Blade and BAS in 3 patients | Both | 12.2 (0.02–18.9) | 35 (3–124) | N/A | Angioplasty balloons | 12 (10–24) | 4.5 | 100 |
| Baruteau et al. (present series) | 2016 | Multicentre retrospective | 64/64 | Both | 18 (0.3–72) | 54.5 (5.3–86) | 1.5 days (0–12) | Angioplasty balloons or Inoue balloons | 28 (6–30) | 9.4 | 98.4 |
Case reports are not included. Age, weight, ECMO–BAS timing and size of balloon used are expressed in median (range).
BAS: balloon atrioseptostomy; LA: left atrial; ECMO: extracorporeal membrane oxygenation; N/A: not available; nb: number.
Indications for BAS
The two primary goals of extracorporeal life support systems in cardiogenic shock are: (a) provide circulatory support by increasing mean arterial pressure and vital organ perfusion; and (b) reduce LV wall stress, stroke work and myocardial oxygen consumption by reducing ventricular pressure and volume. However, while VA-ECMO can effectively provide circulatory support, it does not completely unload the LV in cases of cardiogenic shock secondary to LV dysfunction,7 leading us to consider adjuvant strategies in order to unload the left heart, such as IABP, Impella and TandemHeart.9,10,23,24 Prolonged left heart distension and LA hypertension with pulmonary oedema and/or haemorrhage may impair pulmonary protection and myocardial recovery.18 We showed that percutaneous static BAS may be a safe and effective alternative strategy in order to decompress the left heart in VA-ECMO paediatric and adult patients.
BAS technique
Percutaneous insertion of a trans-septal LA or LV vent has been described in different studies.18,25–28 Although these approaches have shown encouraging results, static BAS carries the advantage of avoiding the placement of intracardiac cannulas in patients who are at a high risk of thromboembolic events.29–31 Indeed, several technical issues related to the management of indwelling catheters have been reported in case of trans-septal venting, such as kinking, poor flow, movement of the catheter and risk of catheter dislodgement.18,27 Other groups evaluated stent implantation across the atrial septum; although this was effective as well, it carried a higher procedural risk.18 A minimally invasive strategy has also been described, involving placing a small cannula antegrade in the pulmonary artery trunk, from either the pulmonary or the jugular vein,32 thereby minimising the risk of pericardial effusion. Eastaugh et al. showed that the decompression technique used does not significantly influence survival to hospital discharge.18 In our study, static balloon dilation alone with low-profile, high-pressure balloons was effective at creating a suitable, long-standing ASD, reaching a good simplicity/efficiency ratio for the procedure.
Technical perspectives
In our series, BAS carries a reasonable rate of procedural complications of 9.4%. The Brockenbrough puncture, which was necessary in order to access the LA, was performed safely, with only two procedures being complicated by LA perforation. Advancing a 0.014’ wire through the tip of the Brockenbrough needle into the LA or pulmonary vein as soon as the atrial septum is punctured decreases the risk of inadvertently perforating the LA wall as the needle and/or the trans-septal sheath are advanced further into the LA. Different radiofrequency energy techniques using either the Nykanen wire or the Baylis trans-septal needle have also been described in an attempt to decrease the risk of this technical step.33,34 In the future, the non-invasive bedside creation of an ASD by histotripsy might also be an exciting approach in VA-ECMO patients.35 A transcatheter, transvenous interatrial shunt device (IASD® System II, Corvia Medical, Inc., Tewksbury, MA, USA) has recently been evaluated with very promising results in adult patients suffering from chronic heart failure with preserved ejection fraction,36 and might be a valuable approach in VA-ECMO patients in the future.
BAS short-term outcomes
All procedures were successful at achieving the primary objective of decompressing the left heart and no patient required reintervention. LA decompression led to a significant reduction in LA pressure, improvement on chest X-ray, improvement in clinical status and cessation of all cases of pulmonary haemorrhage, confirming the beneficial response to prompt left heart decompression when clinically indicated. In our patient cohort, age ⩾10 years and acute myocarditis were found to be independent predictors of death, OHT or LVAD. However, given the complexity of these patients, residual left-to-right interatrial shunting has been reported in survivors with varying proportions, ranging from 24%18 to 76%.22 In our cohort, all ASDs remained patent and unrestrictive at the time of transplant or at the latest follow-up. The septal anatomy and thickness can vary substantially among patients; thus, there is not a linear relationship between balloon size, short-term ASD diameter and patency. The decision is relatively simple in patients with ‘unrecoverable’ heart dysfunction in whom large-diameter balloons can be used. However, patients with potentially recoverable conditions may pose a challenge, as a significant residual ASD may not be well tolerated in the long term, particularly in patients with residual LV dysfunction. The presence of unrestrictive ASD in all of our patients at the latest follow-up suggests that operators may have used large-diameter balloons in order to prevent the ASDs from becoming restrictive in the short term. The fact that no patient required reintervention supports this finding. Smaller, non-restrictive ASDs can likely be achieved with stent implantation. However, proper stent implantation can be challenging in large patients, and it adds a source for thrombus formation in addition to the potential risk of stent dislodgement by the nearby cannulas. Only two patients required ASD closure during their follow-up in our group. Since BASs are aimed at the centre of the atrial septum, most of these ASDs are technically suitable for percutaneous closure,22 such that the long-term haemodynamic consequences of the BAS should not be a matter of concern.
Limitations
This feasibility study of uncontrolled observations represents the 14-year experience from four large referral centres. The indications for ECMO support and subsequent timing of the BAS were assessed on a case-by-case basis, without specific guidelines. Institutional or operator factors may have evolved over this time period. The retrospective data collection at multiple medical centres may have underestimated the incidence of adverse events during follow-up.
Conclusion
Percutaneous static BAS is a safe and effective alternative strategy in order to decompress the left heart in adults and children in cardiogenic shock supported by VA-ECMO. A prospective, randomised controlled trial in order to assess the different LA decompression techniques would be beneficial for comparing the indications and roles of each approach in this multifaceted group of patients.
Acknowledgement
The authors are grateful to Professor Emile A Bacha (New York, NY, USA) and Professor Emre Belli (Paris, France) for their contributions to the management of these patients.
Conflict of interest
The authors declare that there is no conflict of interest.
Funding
A-E Baruteau is supported by a research grant from the French Federation of Cardiology and a research grant from the European Society of Cardiology.
Comments