Author + information
- Received August 29, 2018
- Revision received November 19, 2018
- Accepted November 21, 2018
- Published online March 18, 2019.
- Karim Abdur Rehman, MDa,
- Oussama M. Wazni, MDa,
- Amr F. Barakat, MDa,
- Walid I. Saliba, MDa,
- Shailee Shah, MDb,
- Khaldoun G. Tarakji, MDa,
- John Rickard, MDa,
- Mohamed Bassiouny, MDa,
- Bryan Baranowski, MDa,
- Patrick J. Tchou, MDa,
- Mandeep Bhargava, MDa,
- Thomas J. Dresing, MDa,
- Thomas D. Callahan, MDa,
- Daniel J. Cantillon, MDa,
- Mina Chung, MDa,
- Mohamed Kanj, MDa,
- Samuel Irefin, MDa,
- Bruce Lindsay, MDa and
- Ayman A. Hussein, MDa,∗ ()
- aSydell and Arnold Miller Family Heart and Vascular Institute, Robert and Suzanne Tomsich Department of Cardiovascular Medicine, Section of Cardiac Pacing and Electrophysiology, Cleveland Clinic Foundation, Cleveland, Ohio
- bDepartment of Internal Medicine, Cleveland Clinic Foundation, Cleveland, Ohio
- ↵∗Address for correspondence:
Dr. Ayman Hussein, Department of Cardiovascular Medicine, Section of Cardiac Pacing and Electrophysiology, Cleveland Clinic Foundation, 9500 Euclid Avenue, J2-2, Cleveland, Ohio 44195.
Objectives This study sought to assess the incidence and outcomes of life-threatening complications from atrial fibrillation ablations in a high volume center.
Background With increasing rates of atrial fibrillation ablation procedures, an increase in life-threatening procedure-related complications has been reported despite improvements in technology and ablation strategies.
Methods Between 2000 and 2015, 10,378 patients underwent atrial fibrillation ablation at our institution and were enrolled in a prospectively maintained data registry. We identified all patients who had life-threatening cardiac, neurological, respiratory, or vascular complications to the ablation resulting in death or requiring emergent intervention.
Results Major life-threatening complications occurred in 100 patients (0.9%). The most common was pericardial effusion requiring pericardiocentesis (0.5%), with 7 (0.07%) requiring emergent surgical repair for cardiac perforation. Stroke occurred in 27 patients (0.3%) with a vast majority having an ischemic stroke (93%) followed by hemorrhagic (3.5%) and ischemic stroke with hemorrhagic conversion (3.5%). The yearly incidence of stroke decreased from an average of 1.1% per year in the first tertile (2000 to 2004) to 0.2% per year in the last 2 tertiles (2005 to 2015). Permanent neurological deficits occurred in 23 patients. Vascular complications causing hemorrhagic shock occurred in 7 patients (0.06%), 5 of whom required urgent surgical intervention. Acute coronary syndrome requiring urgent percutaneous coronary revascularization occurred in 2 patients whereas 1 developed a right coronary artery air embolus. No procedural death or atrio-esophageal fistulae occurred.
Conclusions In a large quaternary care center, the incidence of life-threatening complications is low. Experienced operators, high volume, continuous quality improvement initiatives, and efficient back-up support have allowed exemplary safety profiles and 0 procedure-related deaths over 16 years.
Atrial fibrillation (AF), the most common tachyarrhythmia in clinical practice, has become a public health problem; one that is affecting an increasing number of patients (1). This may reflect aging of the population and improved overall survival in patients with cardiac disease and comorbid conditions (1). AF is associated with increased morbidity and mortality and is a cause of significant reduction in quality of life.
Catheter ablation, targeting isolation of the pulmonary veins (PV), has become an effective treatment strategy and nowadays is frequently offered early in the disease process. Though considered generally safe, major and life-threatening complications still occur and remain a concern (2–7). The ablation procedures often involve complex interventions (8–10), which predispose to higher rates of complications with AF ablation than ablations targeting less complex arrhythmias (3,4).
With increasing rates of AF ablation procedures in clinical practice, data from the U.S. NIS (Nationwide Inpatient Sample) showed an increasing trend in rates of cardiac complications and steady persistent rates of procedure-related deaths despite improvement in technology and ablation strategies (2,11). This may reflect wider adoption and possibly the increasing trend of performing AF ablation in low-volume centers by low-volume operators.
In this study, we assessed the rates and outcomes of major and life-threatening complications in a large prospective cohort at our center over a 16-year period. The study’s setting has an established structure for quality monitoring and improvement including monthly meetings to discuss complications and interventions to reduce them, adoption of new technology, and cumulative experience over time and importantly proper surgical and interventional back-up to deal with complications.
All consecutive patients undergoing radiofrequency catheter-based AF ablation procedures at the Cleveland Clinic were enrolled in a prospectively maintained data registry (N = 10,378 procedures between 2000 and 2015). Procedure-related complications were prospectively entered into the complication database. The database records pre-procedural, peri-procedural, and post-procedural data; follow-up visits; and any phone calls or communication between patients and a dedicated registry nurse. The current report included all consecutive patients who experienced life-threatening complications related to the ablation procedure. These were defined as complications that represent an imminent threat to life, result in death or permanent disability/damage, or require acute surgical intervention.
In addition to data from the prospectively maintained registry, individual chart review was performed to further characterize the procedure-related complications and patient outcomes. The study was approved by the Cleveland Clinic Institutional Review Board.
AF ablation protocol
Our AF ablation protocol and peri-procedural anticoagulation strategies were previously reported in detail (12,13). Between 2005 and 2007, we transitioned in our practice from interruption of warfarin with bridging anticoagulation peri-procedurally to performing all AF ablation procedures with uninterrupted therapeutic anticoagulation. Patients on non-vitamin K antagonist oral anticoagulants (NOAC) were instructed to hold only 1 dose on the day of the procedure, whereas those on warfarin continued treatment and procedures were performed under therapeutic international normalized ratios. Transesophageal echocardiography was obtained in all patients who presented in AF on the day of ablation but had interruption of therapeutic anticoagulation in the 3 weeks preceding their procedure with either subtherapeutic international normalized ratios or missing a dose of NOAC. Procedures were aborted when a left atrial or atrial appendage thrombus were detected.
The type of sedation used was up to the treating electrophysiologist, but we have generally transitioned our practice from performing AF ablation procedures under moderate sedation to general anesthesia in the last 10 years.
Femoral vascular access with ultrasound guidance was adopted between the years 2006 and 2007 and has since been used in all patients. Intracardiac echocardiography, with an intravascular ultrasound catheter positioned in the right atrium, was used in all procedures to assist with transseptal punctures, assess left atrial and pulmonary venous anatomy, monitor for complications during the procedures, and ensure catheter-tissue contact during ablation.
A circular mapping catheter (LASSO, Biosense Webster, California) was used to guide ablations with real-time monitoring with intracardiac echocardiography to ensure antral radiofrequency applications. In addition, 3-dimensional navigation systems were used for further guidance of ablation (CARTO, Biosense- Webster Inc., Irvine, California, or Ensite NavX, St. Jude Medical Inc., St. Paul, Minnesota). Radiofrequency energy was generally 35 to 40 W for irrigated catheters, but lower energy or shorter duration was used at the posterior wall as needed. Impedance and esophageal temperature were closely monitored to avoid excessive heating. During the study period, clinical practice transitioned from use of nonirrigated catheters to irrigated catheters and more recently to use of contact force–sensing catheters.
All PV were targeted for isolation guided by Lasso and intracardiac echocardiography. Additional substrate modification including ablations at the posterior wall, septal to the right PV, and left atrial flutter were performed in most cases and were at the discretion of the operating electrophysiologist. The superior vena cava was ablated in the absence of phrenic nerve stimulation, at the discretion of the operator. In patients with concomitant atrial flutter, activation mapping and entrainment were performed to locate and ablate the critical isthmus. Esophageal temperature monitoring was performed using a single point probe that was continuously adjusted to be in proximity to the site of ablation.
The complications of interest were procedural-related death, stroke, pericardial effusion requiring pericardiocentesis or surgery, atrio-esophageal fistula, and cardiac or vascular injury resulting in shock or requiring emergent surgical intervention. Events occurring peri-procedurally, during hospital stay, and up to 3 months after ablation were included in the study.
Pericardial effusion requiring pericardiocentesis was defined as the occurrence of pericardial effusion that eventually required to be drained due to the size of effusion, rate of accumulation, clinical symptoms, or hemodynamic instability. Early pericardial effusion was defined as pericardial effusion that was drained peri-procedurally or within 24 h of the procedure, whereas pericardial effusion requiring drainage beyond this time was defined as late pericardial effusion. Pericardial effusions were drained in the electrophysiology lab under fluoroscopic and echocardiographic guidance. A drain was typically left in the pericardial space until no further pericardial fluid accumulated. Patients were admitted to the coronary intensive care unit for close hemodynamic monitoring overnight. Patients with late cardiac tamponade were identified based on hemodynamic instability and echocardiographic evidence. Patients with late small pericardial effusions without hemodynamic instability were typically managed conservatively with serial echocardiograms.
Stroke was defined as a neurological deficit lasting more than 24 h with imaging, confirmed by a neurologist. Transient ischemic attacks were not included in this analysis.
Post-procedural care and follow-up
Anticoagulation was generally resumed as soon as patients were awake and able to swallow in the anesthesia recovery unit, unless contraindicated due to procedure-related complication.
Patients were monitored for thromboembolic and hemorrhagic complications throughout the procedure and overnight. At discharge, all patients were evaluated by a physical exam for any new neurological deficits and vascular and cardiac complications. Additional diagnostic testing was performed as needed. Patients received rhythm transmitters, and for the first 3 months after ablation, the devices transmitted electrocardiographic tracings on a weekly basis and whenever symptomatic. In addition, telephone calls were conducted in the first 3 months post-discharge by dedicated AF-electrophysiology registered nurses to assess progress of recovery and symptoms suggestive of any complications. All patients were also instructed to call our center for AF if any symptoms developed. All clinical documentation, including telephone encounters or other forms of communication with the patients or their referring physicians, such as letters, were documented electronically. All patients had scheduled follow-up appointments with their electrophysiologist 3 to 4 months following the ablation procedures or earlier if symptoms or complications occurred. At the follow-up appointments all patients were scheduled for cardiac computed tomography or magnetic resonance imaging to assess for PV stenosis.
Descriptive statistics were obtained by using the statistical software JMP Pro version 10.0 (SAS Institute, Cary, North Carolina). Continuous variables are presented as mean ± SD or median (interquartile range [IQR]), as appropriate. Categorical variables are presented as frequencies.
Between 2000 and 2015, 10,378 AF ablation procedures were performed at our institution and 100 of them (0.96%) resulted in major complications. Those 100 patients were included in the current report. Their mean age was 63.2 ± 10.4 years and 62% of them were male. Their baseline characteristics are summarized in Table 1.
Life-threatening complications included pericardial effusion requiring pericardiocentesis or emergent surgical intervention (57%, 12.2% of whom required emergent cardiac surgery), stroke (27%), vascular access complications with hemorrhagic shock (7%), respiratory failure (7%), and myocardial infarction (3%). Of these, 5% of patients went into cardiac arrest as a result of the complication (Table 2, Figure 1).
Pericardial effusion requiring pericardial drainage
The overall incidence of pericardial effusion requiring pericardial drainage was 0.5% (n = 57) (Table 2). The yearly incidence decreased from an average of 0.6% per year between 2000 and 2013 to 0.1% per year over the last 2 years of the study period.
Pericardial effusion was identified during the ablation procedure in 45 patients and 2 to 36 days later in the remaining 12 patients. Of these patients, 3 developed cardiac tamponade and received successful urgent pericardiocentesis. Peri-procedural anticoagulants were uninterrupted warfarin (n = 51), NOAC (n = 3), or enoxaparin-bridging strategy (n = 3). All patients underwent pericardiocentesis and 7 (0.07%) eventually required emergent surgical repair for cardiac perforation. Surgical findings revealed injury to left atrial appendage (n = 1), dome of left atrium (n = 2), interatrial groove in the right atrium (n = 2), coronary sinus (n = 1), and the ridge between the left superior PV and left appendage (n = 1). In patients requiring surgical repair, the anticoagulants used at the time of ablation were warfarin (n = 5), apixaban (n = 1), and enoxaparin (n = 1), but none of them had bleeding complications from the surgical repair. Three of the 57 patients experienced a cardiac tamponade requiring cardiopulmonary resuscitation and were successfully resuscitated.
All but 5 of the 57 patients (8.9%) were discharged on anticoagulants and these were eventually restarted in the remaining 5 patients within 14 days. All patients were followed up for at least 3 months post-procedure and recovered with no long-term disability or surgery-related complications.
The incidence of procedure-related stroke was 0.3% (n = 27). The incidence rate decreased from an average of 0.9% per year during the years 2000 to 2006 to 0.1% per year from 2007 to 2015 as practices evolved toward uninterrupted anticoagulation during ablation and an increase in the use of irrigated catheters. Ischemic stroke occurred in 25 patients. Cases of ischemic stroke with hemorrhagic conversion were reported in 2 patients, additionally. Stroke occurred after a median of 2 (IQR: 1 to 5) days in these patients and was diagnosed during the ablation procedure in 6 of the 27 patients (22%). Only 1 patient was receiving NOAC at the time of the procedure, whereas 16 were on warfarin and 10 on enoxaparin-bridging anticoagulation. Three patients (11%) were successfully treated with thrombolysis/thrombectomy, whereas 2 (7%) underwent emergent craniectomy and 1 (4%) required stenting of the left middle cerebral artery. The remaining patients (78%) were managed conservatively. The median length of hospital stay was 4 (IQR: 4 to 12) days with a median of 1 (IQR: 0 to 5) day in the intensive care unit. Only 4 of 27 patients (15%) were discharged off anticoagulation, of which only 1 eventually restarted anticoagulation on follow-up. Residual neurological deficits persisted in 23 patients (85%), but no deaths occurred.
Major vascular complications
Major vascular access bleeding causing hemorrhagic shock occurred in 7 patients, 5 of whom required urgent surgical intervention: external iliac vein repair (n = 1); profunda femoris artery repair (n = 1); femoral artery pseudoaneurysm repair (n = 2); and internal iliac artery repair (n = 1); whereas 2 were managed conservatively with hemodynamic support in the intensive care unit. Six of these occurred between 2000 and 2008, but only 1 occurred after this period.
Respiratory failure requiring intubation occurred in 6 patients and was secondary to pulmonary edema (n = 2), aspiration pneumonia (n = 1), mediastinal hemorrhage (n = 1), hemothorax (n = 1) requiring surgical intervention, and a neck hematoma (n = 1) requiring emergent evacuation followed by a tracheostomy to relieve airway compression. The patient with mediastinal hemorrhage underwent emergent thoracotomy to repair a superior vena cava leak and restore hemostasis. The patient did not receive superior vena cava ablation; however, the patient had internal jugular venous access, along with femoral access for the procedure. Note that the current practice at our institution does not involve jugular venous access for AF ablation procedures.
Coronary artery-related complications occurred in 3 patients: 1 experienced a cardiac arrest due to left anterior descending artery thrombus in dissection requiring emergent percutaneous coronary intervention after successful cardiopulmonary resuscitation; and 1 had a thrombotic inferior wall ST-segment elevation myocardial infarction peri-procedure requiring urgent percutaneous coronary intervention to the distal right coronary artery. The third patient had an air embolism to the right coronary artery that was associated with hypotension requiring pressor support.
Cardiac arrest, procedure-related mortality, and atrio-esophageal fistula
No cases of atrio-esophageal fistulae were reported. In total, 5 patients experienced a cardiac arrest: 3 patients with cardiac perforation requiring emergent pericardiocentesis; 1 patient as a result of respiratory failure; and 1 from a left anterior descending artery thrombus complicated by ventricular fibrillation. No procedure-related mortality occurred during the 16-year period.
In a large prospective cohort of patients undergoing 10,378 AF ablation procedures at a tertiary care center and spanning the course of 16 years, the incidence of major life-threatening complications was <1%. The most commonly encountered life-threatening complication from AF ablation procedures was pericardial effusion requiring pericardiocentesis (0.5%), whereas emergent open heart surgery was needed in 0.07% of patients to repair cardiac perforation. The yearly incidence of pericardial effusion requiring pericardiocentesis remained unchanged between the years 2000 and 2013 with an average of 0.6% per year and appeared to have decreased to about 0.1% per year over the last 2 years of the study period, which correlates in clinical practice with the introduction of contact force–sensing catheters. Stroke, the second most common complication, occurred at a relatively low incidence rate of 0.3% over the study period; however, stroke was associated with morbidity leaving most of the acute stroke patients with permanent neurological deficits. Nevertheless, the yearly incidence of stroke decreased significantly with evolving practices of uninterrupted anticoagulation and ablation with irrigated catheters. Most importantly, there was not a single procedure-related death or atrio-esophageal fistula identified.
AF ablation has been established as an effective treatment strategy resulting in arrhythmia-free survival and improvement in quality of life and as such is recommended by the leading scientific societies worldwide (1,14). The procedure is considered to be generally safe based on multiple reports. A recent study compared a cohort of patients that received an ablation from 2009 to 2011 with a cohort from 2014 to 2015 to find fewer complications in the modern cohort (5% vs. 2.3%, p = 0.007) (15). Similarly, data from a large single-center experience of 2,750 patients outside the United States demonstrated major complications in 0.84% cases and no reported mortality (16). That being said, in recent years, the number of operators and training programs have increased. In the United States, the number of AF ablation procedures performed every year has increased by about 10-fold between the years 2000 and 2013 (11). Along the widespread adoption of AF ablation procedures, especially in lower volume centers, the field has witnessed an alarming increase in complication rates (11). A worldwide survey on AF ablation safety revealed lowest success and highest complication rates in low-volume centers having performed under 30 procedures (4). Data from the NIS reporting on AF ablation complication rates in the United States between 2000 and 2013 showed that 7.21% of patients experienced at least 1 procedure-related complication with a 2.02% incidence of pericardial complications and 1.15% incidence of cardiac complications (11). The procedure-related death in the United States between 2000 and 2013 was estimated to have occurred at a rate of 0.24%. The rates of cardiac, pericardial, and hemorrhagic complications from ablation procedures appeared to have increased over time in the United States, whereas the rates of stroke and procedural-related deaths have not decreased over time despite improvement in ablation techniques and strategies (2,11).
To the best of our knowledge, this is the first study to report on major life-threatening complications of AF ablation in a prospective cohort at a tertiary care center and spanning the course of more than 15 years. In contrast to the data that suggest an increase in AF ablation complications in the United States, the current study shows no increasing trends and importantly, the risk of life-threatening complications has remained very low with even lower rates in recent years. The clinical characteristics of patients who developed complications in this report were comparable to the overall population of patients undergoing AF ablation at our center (13,17,18).
This is most likely due to cumulative experience over the years in a high-volume center, but it may also reflect ongoing efforts to improve patient safety and quality of care including monthly meetings to discuss complications and interventions to reduce them, adoption of technologies that have improved the safety of the procedures, and the availability of proper surgical and interventional backup to deal with complications (Online Table 1).
Some procedural and technical considerations may have improved the outcomes. A report from our institution has demonstrated a significant decrease in the risk of vascular complications with use of ultrasound for vascular access and ultrasound has been the standard in all of our ablation procedures since then (19). The use of intracardiac echocardiography enhances the safety of transseptal access and allows monitoring and early detections of complications during the procedure. The transition in clinical practice from nonirrigated to irrigated catheters has probably contributed to the reduction in procedure-related stroke risk. Moreover, we but concomitantly transitioned to performing AF ablation with uninterrupted anticoagulation (12,20). Ablations with contact force–sensing catheters have resulted in shorter procedure times and shorter fluoroscopy time compared with non-contact force–sensing catheters as reported from our center; this may reflect more efficient ablation due to ability to deliver effective and safe lesions, which in turn may translate into an improved safety profile (21). For the purpose of stroke prevention, heparin infusion is started before transseptal access and maintained during the procedure as long as sheaths and catheters are in the left atrium to target an activated clotting time of 350 s. In addition, heparinized bubble-free saline infusion is maintained in all sheaths, especially long sheaths, to reduce the risk of thrombus formation.
Another typically lethal complication (22) is the development of an atrio-esophageal fistulas, which did not occur in this cohort despite the fact that we have generally followed an ablation strategy that targets the posterior wall in addition to PV isolation. Proper esophageal temperature monitoring and caution about position of the probe tip relative to the ablation site with fluoroscopy confirmation is key to preventing esophageal injuries. Training for contact-force catheters has focused on adjustment of power, force, and duration while ablating the posterior wall of the left atrium, resulting in lesser power and greater catheter stability and tissue contact than can be obtained using conventional catheters (23). Finally, the technical aspects discussed herein, early recognition of complications, and proper surgical and interventional backup perhaps allowed us to keep a zero death rate from AF ablation in a large cohort over 16 years, while mortality rates from the procedure in the United States have ranged between 0.24% and 0.45% (2,11).
This study is limited by its observational nature. However, data was obtained from a prospectively maintained AF ablation database that records complications and procedural data in real time, in addition to monthly meetings, to adjudicate complications. The data provided in this report is from a tertiary care referral center with large volume and experienced operators, and as such cannot be generalized to low-volume centers. In fact, the increasing trends in AF ablation complications in the United States appear to be mostly related to low-volume centers and less experienced operators (2,11), which emphasizes the importance of experience, volume, and most importantly the proper infrastructure.
In a large prospective cohort at a tertiary care center, the risk of major complications from AF ablation is very low. In contrast to the recently published data regarding increasing complication rates from AF ablation in the United States, this report suggests that experience, volume, a quality improvement program, and proper backup to deal with complications have allowed excellent safety profiles and zero procedure-related death in more than 10,000 ablations over 16 years.
COMPETENCY IN MEDICAL KNOWLEDGE: With increasing rates of AF ablation procedures, a surge in life-threatening procedure-related complications has been reported. This study evaluates the incidence of life-threatening complications over the course of 16 years at a tertiary care center with a quality improvement initiatives and surgical backup. The study shows that in more than 10,000 patients, no procedure-related deaths or atrio-esophageal fistula occurred. Both tamponades and strokes appeared to decrease over the course of the study.
TRANSLATIONAL OUTLOOK: The findings highlight the effect of novel technology and changing practices in decreasing the annual incidence of such complications.
Drs. Wazni and Cantillon have received consulting honoraria from Biosense Webster. Dr. Saliba is an Advisory Board member of Boston Scientific. Dr. Tarakji is an Advisory Board member of Medtronic and AliveCor; and has received consulting honoraria from Medtronic and AliveCor. Dr. Rickard is a consultant for Abbott and Medtronic. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- atrial fibrillation
- interquartile range
- non-vitamin K antagonist oral anticoagulants
- pulmonary vein(s)
- Received August 29, 2018.
- Revision received November 19, 2018.
- Accepted November 21, 2018.
- 2019 American College of Cardiology Foundation
- January C.T.,
- Wann L.S.,
- Alpert J.S.,
- et al.
- Deshmukh A.,
- Patel N.J.,
- Pant S.,
- et al.
- Cappato R.,
- Calkins H.,
- Chen S.A.,
- et al.
- Mugnai G.,
- Irfan G.,
- De Asmundis C.,
- et al.
- Haissaguerre M.,
- Sanders P.,
- Hocini M.,
- et al.
- Takahashi A.,
- Iesaka Y.,
- Takahashi Y.,
- et al.
- Nademanee K.,
- McKenzie J.,
- Kosar E.,
- et al.
- Hosseini S.M.,
- Rozen G.,
- Saleh A.,
- et al.
- Bassiouny M.,
- Saliba W.,
- Rickard J.,
- et al.
- Camm A.J.,
- Lip G.Y.,
- De Caterina R.,
- et al.
- Muthalaly R.G.,
- John R.M.,
- Schaeffer B.,
- et al.
- Voskoboinik A.,
- Sparks P.B.,
- Morton J.B.,
- et al.
- Hussein A.A.,
- Saliba W.I.,
- Martin D.O.,
- et al.
- Hussein A.A.,
- Saliba W.I.,
- Barakat A.
- Tanaka-Esposito C.C.,
- Chung M.K.,
- Abraham J.M.,
- Cantillon D.J.,
- Abi-Saleh B.,
- Tchou P.J.
- Wazni O.M.,
- Beheiry S.,
- Fahmy T.,
- et al.
- Hussein A.A.,
- Barakat A.F.,
- Saliba W.I.,
- et al.
- Cappato R.,
- Calkins H.,
- Chen S.,
- et al.
- Mansour M.,
- Lakkireddy D.,
- Packer D.,
- et al.