Author + information
- Received May 20, 2016
- Revision received September 6, 2016
- Accepted September 15, 2016
- Published online April 17, 2017.
- Muhammad Balouch, MD,
- Esra Gucuk Ipek, MD,
- Jonathan Chrispin, MD,
- Rizma J. Bajwa, MD,
- Tarek Zghaib, MD,
- Ronald D. Berger, MD, PhD,
- Hiroshi Ashikaga, MD, PhD,
- Hugh Calkins, MD,
- Saman Nazarian, MD, PhD,
- Joseph E. Marine, MD and
- David D. Spragg, MD∗ ()
- ↵∗Address for correspondence:
Dr. David D. Spragg, Johns Hopkins Heart and Vascular Institute, The Johns Hopkins Hospital, 600 N. Wolfe Street, Baltimore, Maryland 21287.
Objectives This study assessed trends in transesophageal echocardiography (TEE) use, rate of left atrial appendage (LAA) thrombus detection, and incidence of periprocedural cerebrovascular accident (CVA) since transitioning to a strategy of uninterrupted warfarin or briefly interrupted novel oral anticoagulant therapy in 2010.
Background TEE is routinely performed before ablation for atrial fibrillation (AF) to ensure absence of LAA thrombus.
Methods Patients with AF ablation presenting between January 2010 and September 2015 at Johns Hopkins Hospital were enrolled in an AF ablation registry; TEE and ablation outcomes were retrospectively analyzed. Presence of LAA thrombus, dense spontaneous echo contrast (SEC), or patent foramen ovale (PFO) were recorded. CVA incidence from procedure onset to 30 days post-procedure was evaluated using electronic medical record review.
Results Pre-procedure TEE was performed in 646 of 1,224 AF ablation cases (52.8%). There was a decline in pre-procedure TEE use from 86% in 2010 to 42% in 2015 (p < 0.001). CVA incidence was 4/1,224 (0.33%) cases, and did not change during the study period. TEE findings included LAA thrombus (n = 6; 0.93%), PFO (n = 23; 3.6%), and dense spontaneous echo contrast (n = 99; 15.3%). Both SEC and LAA thrombus were associated with persistent AF, higher CHA2DS2VASC score, increased LA size, reduced LAA flow velocity, and decreased left ventricular ejection fraction. PFO was not associated with prior AF ablation, and SEC was not associated with increased CVA incidence.
Conclusions CVA is a rare complication of AF ablation in patients with minimally interrupted anticoagulation. Pre-ablation TEE may be reasonably avoided in patients without high-risk features.
Catheter ablation is well-established as a safe and effective therapy for symptomatic atrial fibrillation (AF), but can be complicated by periprocedural systemic thromboembolism and cerebrovascular accident (CVA) (1). Pre-existing left atrial (LA) thrombus, most frequently located in the LA appendage (LAA), is an absolute contraindication to catheter ablation for AF, and can be screened for by pre-ablation LAA imaging. Transesophageal echocardiography (TEE) is the gold standard for screening for LAA thrombus (1,2). However, there is great variability in the use of TEE before ablation.
Since 2010, a range of studies have suggested that uninterrupted periprocedural anticoagulation with either warfarin or novel oral anticoagulants (NOACs) is an effective approach to minimize CVA risk (3–6). Whether adoption of minimally interrupted anticoagulation obviates the need in some patients to screen for LAA thrombus with TEE or other imaging is not clear. In the current study, we evaluate trends in TEE use, findings, and CVA incidence in 1,224 patients undergoing AF ablation in the era of minimally interrupted periprocedural anticoagulation.
All patients undergoing AF ablation at Johns Hopkins from January 2010 to September 2015 were enrolled prospectively in an electronic database, with demographics, clinical history, imaging data, procedural data, complications, and outcomes recorded for each procedure. Retrospective analysis of that data was performed for this study. Two groups were identified based on use of pre-procedure TEE. Cases in which a TEE was attempted without success (n = 3) were excluded from the analysis. All patients signed written informed consent at the time of the procedure. The study was approved by the Johns Hopkins Institutional Review Board.
Paroxysmal AF was defined as AF that terminates spontaneously or with intervention within 7 days of onset, whereas persistent AF was defined as AF sustained for >7 days (2). Periprocedural CVA was defined as acute onset of new neurological deficit between the start of the procedure and up to 30 days after ablation. Stroke was defined as an acute infarct on imaging with neurological deficit; transient ischemic attack (TIA) was defined as absence of an acute infarct on imaging with neurological deficit that resolved within 24 h (7). Major procedure-related bleeding complications were defined as pericardial effusions with tamponade or any hematoma that required transfusion or surgical intervention.
Anticoagulation was initiated at least 3.5 weeks before ablation in 94.3% of patients, either with warfarin or a NOAC. Patients treated with warfarin continued uninterrupted therapy throughout the ablation period. Patients treated with NOACs were treated at the discretion of the operating physician, with cessation of NOAC therapy for 0 to 23 h pre-procedure, and with resumption 4 h post-procedure. Adherence to a prescribed anticoagulation regimen was assessed by international normalized ratio (INR) for warfarin-treated patients or by verbal confirmation (for NOAC-treated patients) pre- and post-procedure. During the ablation procedure itself, activated clotting time levels were maintained above 350 s with intravenous heparin during LA access.
All TEEs were performed immediately before the planned ablation procedure. LA size and LAA velocities were assessed, and the LA and LAA were examined for presence of thrombus or spontaneous echo contrast (SEC). In addition, the presence of patent foramen ovale (PFO) or atrial septal defect was assessed with color Doppler and/or injection of agitated saline. TEEs were interpreted independently by experienced cardiologists specializing in echocardiography.
Post-procedure CVA follow-up
All patients were monitored overnight in the hospital and clinically evaluated before discharge the following day. Patients in whom neurological deficits were reported or observed underwent brain magnetic resonance imaging for evaluation of CVA. Clinical follow-up was performed routinely at 3 months, 6 months, and 1 year or earlier if the patient reported any symptoms.
All cases (index and repeat procedures) were included in the analysis. Continuous variables are reported as mean ± SD, whereas categorical variables are reported as percentages. The univariable association of independent variables with dependent variables was examined using generalized estimating equations models, clustered by patient, using binomial family, logit link, and exchangeable working correlation matrix. A p value of <0.05 was considered significant. All statistics were done using Stata version 12 (StataCorp LP, College Station, Texas). Multivariable analysis for LAA thrombus and CVA predictors was not performed because of very low event rates.
A total of 1,224 AF ablations were performed in 999 patients between January 2010 and September 2015. Three additional cases were excluded from analysis after aborted attempt at TEE. Baseline characteristics of the study cohort are summarized in Table 1. Average age was 60.8 ± 10.2 years, and 267 (26.7%) patients were female. There were 415 (33.9%) repeat ablation procedures done in 376 (37.6%) patients. Overall 428 (42.8%) patients were ablated for treatment of persistent AF, with 378 (37.9%) patients presenting for ablation in AF. Eighty patients (8.0%) had a history of prior CVA. Pre-procedure anticoagulation was used in 942 (94.3%) patients, with either uninterrupted warfarin therapy (n = 477; 47.8%), or minimally interrupted dabigatran (n = 232; 23.2%), rivaroxaban (n = 219; 21.9%), or apixaban (n = 64; 6.4%) (Figure 1). Average INR for warfarin-treated patients was 2.4 ± 0.8 at the time of procedure.
Temporal trends in TEE use and CVA
A pre-procedural TEE was performed in 576 (57.7%) patients for 646 (52.8%) cases, whereas in 497 (49.8%) patients undergoing 578 (47.2%) cases, ablation was performed without prior TEE. Comparison of the 2 procedural groups is summarized in Table 2. TEEs were requested more frequently among patients with a history of persistent AF (p < 0.001), patients presenting in AF (p < 0.001), and patients with a history of hypertension and congestive heart failure (p = 0.024 and 0.005, respectively). There was no significant difference in age, redo procedure rate, prior CVA history, and diabetes mellitus in the 2 groups.
From 2010 to 2015, rates of pre-ablation TEE use fell significantly from 86% to 42% (p < 0.001) (Figure 2). During the same interval, the incidence of CVA was 4 of 999 patients (0.40%) or 4 of 1,224 cases (0.33%), including 2 TIAs and 2 strokes, with no appreciable change in CVA rate over time (CVA incidence of 0.30 ± 0.35%/year; p = NS for change over time) (Figure 2). There were no episodes of noncerebral thromboembolism. Clinical characteristics and anticoagulation profiles of the CVA patients are summarized in Table 3. All were anticoagulated with warfarin (n = 1; pre-procedure INR of 2.7; INR at time of CVA of 2.7) or NOACs (n = 3). Pre-procedure TEE was performed in 3 cases in which the presenting rhythm was AF or atrial flutter. One patient had a prior history of TIA, presented in sinus rhythm on warfarin, with TEE deferred at the discretion of the operator.
TEEs positive for LAA thrombus occurred in 6 of 646 (0.93%) cases (6 of 591 [1.04%] patients), leading to cancellation of those ablation procedures (Table 4). Patients with LAA thrombus had a history of persistent AF in 5 of 6 cases (83%), presented in AF in 3 of 6 cases (50%), had an average CHA2DS2VASc score of 2.5 ± 1.6, and a reduced ejection fraction (48.3 ± 13.3%). All were systemically anticoagulated at the time of TEE. A total of 3 of 6 patients with TEE positive for LAA thrombus underwent revision of anticoagulation strategy (2 patients transitioned from warfarin to a NOAC; 1 patient transitioned from dabigatran to warfarin), whereas 3 of 6 maintained ongoing anticoagulation. All 6 patients proceeded to eventual pulmonary vein isolation; no patients with LAA thrombus on initial TEE had subsequent CVA. Statistical comparison with the patients with negative TEEs was limited by sample size; of note, patients with negative TEEs had a history of persistent AF in 221 (47.1%) patients, had an average CHA2DS2VASC score of 1.7 ± 1.4, and an average ejection fraction of 55 ± 9. No thrombi were seen in the RA or nonappendage regions of the LA.
Other TEE findings: LAA velocity, SEC, and PFO
In the 646 cases undergoing TEE, LAA flow velocities could be assessed in 621. Flow velocity in patients with normal TEEs (n = 469) was 48.2 ± 18 cm/s, whereas in patients with dense SEC (n = 91) or frank thrombus (n = 6), LAA flow velocities were 30.3 ± 1.7 cm/s and 24.2 ± 4.5 cm/s, respectively (p < 0.001 vs. TEE-negative patients). LA diameter in TEE-negative patients was 4.4 ± 0.7 cm, whereas in patients with positive TEE for SEC or thrombus, LA diameters were 4.9 ± 0.8 cm and 4.8 ± 0.4 cm, respectively (p < 0.001 vs. TEE-negative patients). Dense SEC was seen in 99 (15.3%) cases on TEE in 91 (15.9%) patients. The remaining 515 cases (79.7%) in 469 (85.6%) patients had no dense SEC or thrombus on TEE (Table 4). Patients with dense SEC, when compared with patients without abnormal TEE findings, were more likely to have persistent AF (p < 0.001), presentation in AF (p = 0.004), elevated CHA2DS2VASC score (p < 0.001), reduced EF (p = 0.001), and lower LAA flow velocities and larger LA diameters (p < 0.001). Of the 99 cases with SEC, all underwent ablation and none had CVA.
Twenty-three of 646 cases (3.6%) screened by TEE were found to have a PFO. Of these, 18 of 23 (78.3%) were index ablations, whereas 5 of 23 (21.7%) were repeat ablations. Of the 431 patients undergoing TEE screening before index ablation, 18 (4.2%) had a detectable PFO, whereas 5 of 204 patients (2.5%) undergoing TEE before repeat ablation were found to have PFO or iatrogenic atrial septal defect (p = 0.31).
Pericardial effusion with tamponade requiring drainage occurred in 8 of 1,224 (0.65%) cases. There was 1 of 1,224 (0.08%) case of large pelvic hematoma that required surgical intervention. Minor access site bleeding and small hematoma occurred in another 3 of 1,224 (0.25%) cases.
We investigated trends in pre-procedure TEE acquisition before AF ablation at our institution since transitioning to an anticoagulation strategy of uninterrupted warfarin or minimally interrupted NOAC therapy. We report 3 principle findings: 1) TEE use before ablation has fallen dramatically, with no appreciable rise in either CVA incidence or LAA thrombus (when TEE was performed) over the last 5 years; 2) dense SEC is more likely on univariate analysis to associate with persistent AF, presentation in AF, increased LA size and reduced LAA flow velocity, and clinical factors captured in CHA2DS2VASC scoring, but are not associated with adverse events following ablation; and 3) PFO is rare, and does not correlate with prior ablation procedures.
Emerging anticoagulation strategies
Several studies have suggested that continuation of systemic anticoagulation throughout the ablation period with warfarin or NOAC therapy is a reasonable strategy for CVA risk reduction (3–6), although not all investigations have been as reassuring (8). In 2010 we adopted a strategy of continuous warfarin anticoagulation or minimal NOAC interruption, and have not seen appreciable changes in embolic or bleeding complications since that transition (9). The safety of continuous systemic anticoagulation is the subject of ongoing investigation with randomized controlled studies, and definitive conclusions about the safety and efficacy of that approach should be available in the near future.
Targeted TEE use in patients undergoing ablation
We have observed that, as an institution, we have shifted away from routine TEE use before AF ablation as we perform more procedures with patients continuously anticoagulated. This study verifies that observation; TEE use has dropped from 86% in 2010 to 42% in 2015. During that period, we have not seen an appreciable increase in CVA incidence in 1,224 cases performed. Indeed, we reported CVA rates of 1.4% in patients undergoing ablation from 2001 to 2008 (10); that rate has fallen to 0.33% in the current study. Similarly, rates of TEE positive for thrombus have fallen from 1.6% (11) to 1% currently. We attribute these declines in part to the strategy of minimally interrupted anticoagulation, and (in the case of CVA reduction, despite fewer screening TEEs performed) to improvements in equipment and institutional experience.
In the 4 patients with CVA, 3 had TEE studies immediately before ablation that were negative. In the same interval, we cancelled 6 cases because of positive TEE for LAA thrombus. There are several clinical characteristics that appear consistently in the population still referred for TEE and the cohort of patients with positive TEEs: persistent AF, AF at the time of presentation, and increased CHA2DS2VASC risk profile. Our practice in 2016 is that TEE before AF ablation is most reasonable in patients with these clinical risks. Patients presenting in normal sinus rhythm with low CHA2DS2VASC scores typically are not undergoing TEE before ablation at our institution; patients with persistent AF, even in the setting of a relatively low CHA2DS2VASC score, typically are referred for TEE, based in part on the sobering finding that 2 of our CVA/TIA patients had a CHA2DS2VASC score of 0, but persistent AF. Additionally, our findings that CVAs occurred in 3 patients for whom screening TEE was negative and that all patients with positive screening TEE for LAA thrombus avoided subsequent CVA underscores that particular care must be given to intraprocedural and post-procedural anticoagulation.
We found that presence of LAA thrombus on TEE was associated with increased CHA2DS2VASC score. Whether a history of prior CVA alone is enough to warrant mandatory TEE remains uncertain. However, Hussein et al. (12) recently investigated the risk of periprocedural CVA in patients with prior cerebrovascular events, reporting a 0% CVA risk in 247 patients following AF ablation. Importantly, only 58 (23.5%) of the patients who presented in AF and had subtherapeutic INR had pre-procedure TEE done. The very low CVA risk reported by Hussein et al. is unusual; other large series of patients undergoing pulmonary vein isolation typically report a thromboembolic incidence between 0.5% and 4% (4,9,13).
Other investigators have reported on foregoing TEE before AF ablation, regardless of CHA2DS2VASC score or prior CVA history. In a study of 970 patients anticoagulated with either apixaban or rivaroxaban without interruption before AF ablation, Di Biase et al. (14) found that CVA risk was minimal (0.1%) even when TEE screening was entirely eliminated. Of note, more than 70% of this cohort was screened for LAA thrombus with intracardiac echocardiography (ICE) catheter imaging, which limits the general conclusions that can be drawn about elimination of LA imaging altogether.
What is gained by avoiding TEE?
TEE is the recommended screening tool in patients with sustained AF for detection of LA/LAA thrombus (2) before restoration of sinus rhythm, with a sensitivity and specificity ranging from 83% to 100% (1,15–17). However, TEE is somewhat invasive and carries its own risk of potential complications including oropharyngeal, laryngeal, esophageal, and gastric injury (17,18). Because TEE is often performed under anesthesia at the time of the ablation procedure, it also increases procedure duration and cost; accordingly, there is a wide range in use at different high-volume centers (19).
An attractive alternative to pre-ablation TEE is the intraprocedural use of ICE. This modality has been shown to provide complementary information to that from pre-ablation TEE (20), allowing for real-time assessment of LAA thrombus formation, pericardial effusion, and immediate post-ablation atrial septal defects after transseptal puncture. ICE use is predicated on operator ability to acquire and interpret images correctly, which may limit use in some centers.
Other findings: SEC and PFO
Frequently the presence of dense SEC on TEE before ablation gives rise to concern about the potential for thrombus formation in the near future. In our cohort of 99 patients with dense SEC who proceeded with AF ablation, we did not see increased rates of thromboembolism.
PFO in our cohort was uncommon. In a study involving 197 repeat AF ablations, Anselmino et al. (21) reported a 21.8% incidence of PFO on patients referred for repeat ablation. We found a total of 23 cases with PFO at baseline (3.5% of the population screened with TEE), a rate somewhat lower than that reported from a comparable series assessed by ICE (22). In our series, PFOs were more commonly seen in patients undergoing initial ablation, rather than redo procedures. The generation of iatrogenic atrial septal defects following single- and double-transseptal puncture for AF ablation have been well characterized by Ren et al. (22–24), and by others (25). Septal defects are frequently detectable immediately post-procedure; subsequent closure rates may be related in part to the size of the transseptal sheath (25). Our relatively low rate of PFO in patients undergoing repeat ablation may be caused in part by our frequent use of double-transseptal puncture with smaller sheaths.
This study is a single-center retrospective analysis of patients undergoing AF ablation. Intrinsic limitations include changes in clinical practice and clinical volume. The anticoagulation protocol may be different than other centers and so may be the method of deciding on pre-procedure TEE. In terms of data collection, there may have been some subclinical cases of thromboembolic events that have not been identified given the fact that not all patients underwent post-procedural cardiac magnetic resonance imaging evaluation.
Avoiding TEE use in low-risk patients for CVA, including those with low CHA2DS2VASC scores, presenting in sinus rhythm, and systemically anticoagulated, is not associated with increased risk of periprocedural CVA. Whether this strategy can be safely extended to higher-risk patients requires further investigation.
COMPETENCY IN MEDICAL KNOWLEDGE: Providing care to AF patients that is effective and safe is at the core of 2 basic clinical competencies: medical knowledge, patient care and procedural skills. We believe that our investigation allows readers to supplement these 2 competencies, by providing data about whether ancillary procedures (i.e., TEE) around AF ablation is needed, in whom it may be avoided, and what that ancillary procedure is likely to find in a particular patient population.
TRANSLATIONAL OUTLOOK: Making ablation for AF patients a more safe and effective procedure is the subject of active investigation worldwide. This investigation raises important questions yet to be tested. Although we found that TEE was of limited value in low-risk patients, the rather obvious question raised by our investigations (and other similar studies) is whether higher-risk groups can forego TEE, and what other strategies (e.g., rigorous anticoagulation monitoring, ICE-catheter assessment, pre-procedure computed tomography or magnetic resonance imaging) might need to be in place to ensure patient safety in the absence of TEE. We anticipate that as pre-procedural and intraprocedural imaging options continue to evolve, and as periprocedural uninterrupted anticoagulation becomes a more widely adopted strategy, investigations to determine how best to reduce periprocedural CVA will increase.
Funding for this research was provided in part by the Edward St. John Foundation for AF Research, Roz and Marvin H. Weiner and Family Foundation, Dr. Francis P. Chiaramonte Foundation, Marilyn and Christian Poindexter Arrhythmia Research Fund, and Norbert and Louise Grunwald Cardiac Arrhythmia Research Fund. Dr. Nazarian is a consultant for Biosense Webster, Cardiosolv, and Spectranetics; and has received a research grant from Biosense Webster. 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
- cerebrovascular accident
- intracardiac echocardiography
- international normalized ratio
- left atria
- left atrial appendage
- novel oral anticoagulants
- patent foramen ovale
- spontaneous echo contrast
- transesophageal echocardiography
- transient ischemic attack
- Received May 20, 2016.
- Revision received September 6, 2016.
- Accepted September 15, 2016.
- 2017 American College of Cardiology Foundation
- Calkins H.,
- Kuck K.H.,
- Cappato R.,
- et al.
- January C.T.,
- Wann L.S.,
- Alpert J.S.,
- et al.
- Di Biase L.,
- Burkhardt J.D.,
- Santangeli P.,
- et al.
- Kernan W.N.,
- Ovbiagele B.,
- Black H.R.,
- et al.
- Hussein A.A.,
- Gadre A.,
- Wazni O.M.,
- et al.
- Di Biase L.,
- Briceno D.F.,
- Trivedi C.,
- et al.
- Klein A.L.,
- Murray R.D.,
- Grimm R.A.,
- et al.
- Kumar D.,
- Linq L.H.,
- Halloran K.,
- et al.
- Cappato R.,
- Calkins H.,
- Chen S.A.,
- et al.
- Ren J.F.,
- Callans D.J.,
- Marchlinski F.E.
- Ren J.F.,
- Marchlinski F.E.,
- Callans D.J.,
- et al.
- Ren J.F.,
- Callans D.J.,
- Marchlinski F.E.