Prevalence of left atrial appendage thrombus detected by transoesophageal echocardiography before catheter ablation of atrial fibrillation in patients anticoagulated with non-vitamin K antagonist oral anticoagulants

Abstract

Aims

There is ongoing controversy about the need for routine transoesophageal echocardiography (TOE) prior to atrial fibrillation (AF) ablation. Recently, the debate was reignited by the publication of a large series of patients showing a prevalence of left atrial appendage thrombus (LAAT) on TOE of 4.4%. We sought to assess the prevalence of LAAT on TOE before AF ablation at our institution.

Methods and results

Consecutive patients scheduled for AF ablation at our institution between January 2009 and December 2016 were included. All patients were on oral anticoagulation for at least 4 weeks prior to TOE. Transoesophageal echocardiographies were performed 3–5 days prior to scheduled AF ablation. Data were collected utilizing a prospective database. In all, 668 patients and 943 AF ablation procedures were included. Mean age was 64 ± 11 years, 72% were male, average CHADS₂ score was 1.0 ± 1.0, and 72% of the patients had paroxysmal AF. At the time of ablation, 496 (53%) were on non-vitamin K antagonist oral anticoagulants (NOACs) and 447 (47%) were on Warfarin. There were three cases with LAAT (3/943, 0.3%), all of whom had persistent AF and were on Warfarin. Two patients underwent surgical ablation and the third patient did not undergo ablation.

Conclusion

In our experience, the prevalence of LAAT in patients on anticoagulation therapy undergoing TOE before catheter ablation of AF is 0.3%, which was much lower than recently reported. None of the patients with paroxysmal AF or on NOACs were found to have LAAT. Rather than routine use of TOE prior to AF ablation, a risk-based approach should be considered.

Left atrial appendage thrombus, Left atrial appendage sludge, Transoesophageal echocardiography, Atrial fibrillation, Catheter ablation, Non-vitamin K antagonist oral anticoagulants

What’s new?

• The prevalence of left atrial appendage thrombus (LAAT) among patients anticoagulated with non-vitamin K antagonist oral anticoagulants is zero.

• None of the patients referred for catheter ablation for paroxysmal atrial fibrillation (AF) were found to have LAAT.

• Routine use of TOE prior to AF ablation warrants reconsideration.

Introduction

There is ongoing controversy about the need for routine transoesophageal echocardiography (TOE) prior to atrial fibrillation (AF) ablation to rule out left atrial appendage thrombus (LAAT). Several studies have evaluated the prevalence of LAAT on TOE among patients undergoing AF ablation who have been therapeutically anticoagulated.^1–7 The rates vary between 0.2% and 3.6%. Most of these studies were done before the widespread use of non-vitamin K antagonist oral anticoagulants (NOACs).

Non-vitamin K antagonist oral anticoagulants have been shown to reduce the risk of stroke and systemic embolism when compared to Warfarin.⁸ However, a recent report found that the rate of LAAT in patients on NOACs before AF ablation was 4.4% (8/183).⁶ Subsequently, another group reported a conflicting result with a rate of LAAT of 0.93% (6/646), however, the number of patients on NOACs was not reported.⁷ Multiple recent studies reported the safety of performing AF ablation with uninterrupted NOACs but did not perform TOE systematically.^9–11 We sought to assess the prevalence of LAAT on TOE before AF ablation amongst a large number of consecutive patients on anticoagulation, and specifically report the results of patients anticoagulated with NOACs at our institution.

Methods

We included all patients scheduled for AF ablation procedures at the University of Ottawa Heart Institute from January 2009 to December 2016. Patients were enrolled prospectively in a longitudinal registry. Patient history and baseline characteristics were systematically recorded in the database and included age, gender, CHA₂DS₂-VASc score individual components, type of AF, left ventricular (LV) function and full details of peri-procedural anticoagulation. Follow-up included a 2-week event monitor and clinical visit at 3, 6, and 12 months post-ablation. This protocol was reviewed and approved by the University of Ottawa Heart Institute Review Board.

Congestive heart failure was defined as LV systolic dysfunction (ejection fraction < 40%) or New York Heart Association (NYHA) Class ≥2. Vascular disease was defined as previous myocardial infarction, peripheral arterial disease, or aortic plaque. Paroxysmal AF was defined as AF lasting ≤ 7 days.¹² Persistent AF was defined as AF lasting >7 days. Long-standing persistent AF was defined as AF lasting >1 year. CHADS₂ and CHA₂DS₂-VASc scores were calculated for each patient.

All patients were anticoagulated for at least 4 weeks before ablation with either a NOAC or Warfarin to maintain an international normalized ratio (INR) of at least 2. Patients on Warfarin were followed in a dedicated anticoagulation clinic to ensure a therapeutic INR for at least 4 weeks pre-procedure. Warfarin was stopped 5 days prior to ablation and patients were bridged with low molecular weight heparin (LMWH) starting 3 days prior to ablation. Low molecular weight heparin was held the morning of the procedure. For patients on NOACs, the last dose was taken 2 days prior to the ablation procedure and no bridging was done (Hence for Rivaroxaban one dose was missed, and for Dabigatran/Apixaban two doses were missed).

Heparin bolus was administered in all patients after transseptal catheterization to maintain a target activated clotting time (ACT) more than 300 s. All patients received heparin after ablation until the next morning when oral anticoagulation was restarted, and patients were bridged with LMWH if they were on Warfarin prior to the procedure.

Transoesophageal echocardiography imaging

All patients underwent TOE at the University of Ottawa Heart Institute Echocardiography Laboratory (InterSocietal Accreditation Commission accredited site) within 5 days of the ablation procedure. Standard TOE acquisition was performed by a Level 3 echocardiographer. The left atrial (LA) appendage was imaged using multiple imaging planes including a continuous sweep from 0 to 180° centred on the long-axis of the LA appendage. Left atrial appendage thrombus was considered to be present if there was a well-circumscribed echo mass distinct from spontaneous contrast and the LA pectinate muscles. Left atrial appendage sludge (LAAS) or dense spontaneous contrast was defined as dynamic, viscid, layered echodensity with swirling motion throughout the cardiac cycle.¹³

Statistical analysis

Continuous variables are expressed as mean ± standard deviation and were compared using the Fisher’s exact tests. Categorical variables are reported as percentages (%) and counts and were compared using the Mann–Whitney U tests. For all comparisons, P < 0.05 was considered statistically significant. Statistical analyses were performed using SAS 9.4 (SAS Institute, Cary, NC, USA).

Results

During the study period, 668 patients were scheduled for AF ablation and 943 AF ablation procedures were performed at our institution. Of these, 928 procedures were done for AF and 15 were done for left atrial flutter/atrial tachycardia (AFL/AT). All patients underwent TOE prior to ablation. One patient developed pulmonary oedema during TOE that necessitated admission to the intensive care unit (ICU). All patients were followed for at least 1 year post-procedure.

Clinical characteristics

Patient characteristics are listed in Table 1. The mean age was 64 ± 11 years, and 72% were male. The mean CHADS₂ score was 1.1 ± 1.0 and 22% of patients had a CHADS₂ score of ≥2. In all, 72% had paroxysmal AF and 53% were on NOACs at the time of their ablation. In all, 45% were not on anticoagulation prior to the 4-week period before AF ablation.

Table 1

Baseline demographics (668 patients, 943 procedures)

Age (years)	64 ± 11
Male	478/668 (72)
CHF	149/667 (22)
HTN	250/667 (37)
Diabetes	51/666 (8)
TIA/CVA	44/667 (7)
CHADS2 score	1.0 ± 1.0
CHA2DS2-VASc score	1.9 ± 1.4
CAD	76/659 (12)
Paroxysmal AF	478/665 (72)
Persistent AF	156/665 (24)
Long-standing persistent AF	35/665 (5)
Warfarin use	447/943 (47)
NOAC use	496/943 (53)
Dabigatran	258/496 (52)
Rivaroxaban	184/496 (37)
Apixaban	54/496 (11)
LA appendage thrombus	3/943 (0.3)
LA appendage sludge	5/943 (0.5)

Values are mean (±SD) or n (%). Slight variation in denominators due to missing data.

AF, atrial fibrillation; CAD, coronary artery disease; CHF, congestive heart failure; HTN, hypertension; LA, left atrial; NOAC, non-vitamin K antagonist oral anticoagulant; TIA/CVA, transitory ischaemic attack/cerebrovascular accident.

Left atrial appendage thrombus/sludge patients

Three patients had LAAT on TOE prior to their AF ablation (0.3%), and an additional five had LAAS (0.5%). Table 2 summarizes their baseline characteristics. All patients who developed LAAT had persistent AF and were on Warfarin.

Table 2

Baseline demographics of patients with LAAT and LAAS

Patient	Age	Gender	Type of AF	Rhythm during TOE	CHADS2-VASc	CHADS2 score	OAC	LV function
LAAT (n = 3)
1	48	M	Persistent	AF	0	0	Warfarin	HCM
2	58	M	Persistent	AF	2 (TIA)	2 (TIA)	Warfarin	Normal
3	69	M	Persistent	AFL	2 (HTN)	1 (HTN)	Warfarin	Normal
LAAS (n = 5)
4	64	M	Paroxysmal	NSR	0	0	Rivaroxaban 20 mg daily	Normal
5	48	M	Persistent	AF	2 (CVA)	2 (CVA)	Dabigatran 150 mg BID	Normal
6	58	F	Persistent	NSR	1 (sex)	0	Warfarin	Normal
7	72	F	Paroxysmal	NSR	2 (age/sex)	0	Warfarin	Normal
8	70	M	Persistent	AF	1 (age)	0	Warfarin	Normal

AF, atrial fibrillation; AFL, atrial flutter; BID, twice daily; CVA, cerebrovascular accident; F, female; HCM, hypertrophic cardiomyopathy; HTN, hypertension; LAAS, left atrial appendage sludge; LAAT, left atrial appendage thrombus; M, male; NSR, normal sinus rhythm; OAC, oral anticoagulation; TIA, transitory ischaemic attack; TOE, transoesophageal echocardiography.

Of the five patients who developed LAAS only, three had persistent AF and two had paroxysmal AF. One patient was on a NOAC and four were on Warfarin.

Patient follow-up

All patients with LAAT had their AF ablation postponed. Patient no. 1 had hypertrophic cardiomyopathy (HCM), had LAAS on initial TOE then LAAT on repeat TOE and underwent surgical ablation of AF with left atrial appendage (LAA) excision. Patient no. 2 was diagnosed with IgA nephropathy and was lost to follow-up. Patient no. 3 had moderate aortic stenosis and was treated with Warfarin targeting a higher INR (2.5–3.5). A repeat TOE 5 months later revealed no LAAT.

Of the five patients with LAAS, two patients (nos 5 and 7) underwent AF ablation as scheduled at the discretion of the treating physician with no complications. Patient no. 4 was maintained on Rivaroxaban with resolution of the LAAS on a repeat TOE 5 weeks later and subsequently underwent AF ablation. Patient no. 6 had resolution of LAAS 2 months later with a higher INR target and underwent AF ablation. Patient no. 8 was maintained on Warfarin with the same INR target and repeat TOEs 2 months and 6 months later continued to show LAAS. This patient underwent surgical ablation of AF with LAA excision.

Two patients had a stroke. Both had no LAAT or LAAS and were anticoagulated with Warfarin. The first patient was not compliant with the Warfarin therapy and was admitted 10 days after the procedure with a posterior embolic stroke. The second patient presented 5 weeks after the procedure with multiple embolic strokes and was found to have an atrio-oesophageal fistula.

Comparison between patients with and without left atrial appendage thrombus/ left atrial appendage sludge

Table 3 summarizes the differences between the two groups. The type of AF was significantly different between the two groups.

Table 3

Baseline comparison between patients with and without LAAT/LAAS (668 patients, 943 procedures)

	LAAT/LAAS absent	LAAT/LAAS present	P-value
Age	64 ± 11	61 ± 9	0.340
Male	473 (72%)	5 (63%)	0.694
CHF	149 (23%)	0 (0%)	0.210
HTN	249 (38%)	1 (13%)	0.270
DM	51 (8%)	0 (0%)	1.000
TIA/CVA	43 (7%)	1 (13%)	0.422
CHADS2 score	1.0 ± 1.0	0.4 ± 0.7	0.069
PAF	476 (72%)	2 (25%)	0.008
PerAF	150 (23%)	6 (75%)	0.003
NOAC use	494 (53%)	2 (25%)	0.158

Values are mean (±SD) or n (%).

CHF, congestive heart failure; DM, diabetes mellitus; HTN, hypertension; LAAS, left atrial appendage sludge; LAAT, left atrial appendage thrombus; NOAC, non-vitamin K antagonist oral anticoagulant; PAF, paroxysmal AF; PerAF, persistent AF; TIA/CVA, transitory ischaemic attack/cerebrovascular accident.

Discussion

In this analysis, the prevalence of LAAT on TOE among anticoagulated patients going for AF ablation was low at 0.3%. To our knowledge, this is the largest series of patients on different anticoagulation regimens systematically undergoing TOE before AF ablation. A major finding of our study is that LAAT was not observed in any patients receiving NOAC therapy. Data from literature on the rate of LAAT in patients undergoing AF ablation are summarized in Table 4.

Table 4

Data from literature on the rate of LAAT in patients undergoing AF ablation

Study	Rate of LAAT	Type of AC	Thrombo-embolic risk	Type of AF in LAAT+
Khan et al.1	3/1221 (0.2%)	Warfarin	Not reported	2 PAF 1 PerAF
Puwanant et al.2	6/1058 (0.6%)	Warfarin	80% CHADS2 score <2	1 PAF 5 PerAF
Scherr et al.3	12/732 (1.6%)	Warfarin	79% CHADS2 score <2	3 PAF 3 PerAF
McCready et al.4	12/635 (1.9%)	Warfarin	68% < 2 risk factorsa	11 PAF 1 PerAF
Wallace et al.5	7/192 (3.6%)	Warfarin	Mean CHADS2 = 0.9 (SD = 0.8)	3 PAF 2 PerAF 2 LS PerAF
Frenkel et al.6	14/388 (3.6%)	Warfarin (53%) NOAC (47%)	Mean CHA2DS2-VASc = 2 (SD = 1)	4 PAF 9 PerAF 1 LS PerAF
Balouch et al.7	6/646 (0.93%)	Not reported	Mean CHA2DS2-VASc = 1.8 (SD = 1.4)	1 PAF 5 PerAF

AC, anticoagulation; AF, atrial fibrillation; LAAT, left atrial appendage thrombus; LAAT+, left atrial appendage thrombus present; LS PerAF, long-standing persistent AF; NOAC, non-vitamin K antagonist oral anticoagulant; PAF, paroxysmal AF; PerAF, persistent AF; SD, standard deviation.

a

Risk factors defined as each component of CHADS₂ score and significant valvular heart disease.

The recent study by Frenkel et al.⁶ found that the prevalence of LAAT on TOE in patients on NOACs was 4.4% (8/183). This contrasting finding could be explained, at least in part, by the higher risk population for LAAT in their study compared to ours. The mean CHA₂DS₂-VASc score for patients with LAAT in our study was 1.3 compared with 2.5 in the study by Frenkel.

Previous reports have shown that the CHADS₂ score is an independent predictor of LAAT and that the prevalence of LAAT increases with increasing CHADS₂ score.^2,3 In one study, 1058 TOEs prior to AF ablation were reviewed and a graded prevalence of LAAT based on the CHADS₂ score was found.² It is also important to note that the prevalence of persistent AF in the study by Frenkel et al. was higher compared to our study (34% vs. 28%, respectively). Persistent AF has been shown to be associated with higher prevalence of LAAT.⁴

The prevalence of LAAT in patients on Warfarin at our institution was 0.6%. This is within the range of reported prevalence of LAAT detected by TOE in patients anticoagulated with Warfarin.^2–5 None of the patients with paroxysmal AF had LAAT. This is consistent with multiple previous studies that reported no LAAT in patients with paroxysmal AF and no structural heart disease.^1,4

There are no specific guidelines regarding the use of TOE prior to AF ablation. Current guidelines are largely based on recommendations that pertain to AF cardioversion.^14,15 The 2012 HRS/EHRA/ECAS consensus statement on catheter ablation of AF suggests that the performance of TOE prior to AF ablation in patients who have been in AF for 48 h or less may be considered but is not mandatory.¹⁴ 50% of task force members routinely perform TOE prior to AF ablation. However, this was before the widespread use of NOACs. Di Biase et al.¹⁶ performed a prospective multicentre registry of patients undergoing AF ablation on uninterrupted NOACs with no prior TOE. There were no LAAT detected by intracardiac echocardiogram when the left atrial appendage was visualized. More importantly, only one thrombo-embolic event occurred out of 970 patients studied (0.1%). Although TOE is a safe procedure, it is not without risks. In a retrospective study of 2947 consecutive TOE procedures performed at our centre, there were 26 major complications (0.9%) including 1 death, 9 cases of bleeding, and 2 cases of pulmonary oedema.¹⁷ Transoesophageal echocardiography is also associated with significant costs. Gula et al.¹⁸ constructed a decision analysis model to compare the cost-effectiveness of routine TOE for detection of LAAT vs. no TOE. They found that the incremental cost-effectiveness ratio (ICER) for TOE was $226 608 per quality-adjusted life year (QALY). It is important to note that this analysis was based on an estimated LAAT prevalence of 4%, 10 times higher than what we observed at our institution suggesting that the true ICER may be significantly higher. The risk, cost and patient discomfort associated with TOE should be weighed against the potential benefit of detecting LAAT, which is extremely rare with the use of NOAC as evidenced by our results. A risk-based approach incorporating the CHADS₂ score, the type of AF and anticoagulation regimen can be used to screen patients who might benefit from TOE prior to AF ablation. This approach is even more appealing as other imaging modalities such as computed tomography angiography (CTA), which is commonly used to evaluate pulmonary vein anatomy before AF ablation, can provide good sensitivity and negative predictive values for LAAT.¹⁹

Left atrial appendage sludge is considered to be a step further along the spectrum to thrombus compared to spontaneous echo contrast (smoke). It has been found to be independently associated with subsequent thrombo-embolic events in patients with AF.¹³ There is no consensus as to whether AF ablation can be safely performed in patients with LAAS. The 2012 HRS/EHRA/ECAS consensus statement does not recognize sludge as an entity.¹⁴ Hajjiri et al.²⁰ performed a retrospective analysis of eight patients who underwent AF ablation and were found to have LAAS on TOE. There were no thrombo-embolic events that happened during or after the procedure. Increased intensity of heparin anticoagulation may reduce the risk of left atrial thrombus formation during ablation.²¹ More studies are needed to better assess the best management approach for these patients. Reassuringly, even if we consider LAAS a contraindication for AF ablation, our result shows that the prevalence in patients on NOAC is extremely low (0.4%).

Limitations

Our study has several limitations, including that it is a single centre study with a low-risk population. However, the prevalence of risk factors is comparable with recent clinical trials and registries.^22,23 Secondly, baseline characteristics were only collected at first procedure whereas anticoagulation details were collected for all procedures; as such, we were unable to compare baseline thrombo-embolic and bleeding risks between different anticoagulation agents. Lastly, the small number of patients with LAAT precludes any conclusive analysis to identify specific predictors of LAAT.

Conclusion

In our patient population, the prevalence of LAAT among patients with paroxysmal AF and those on NOAC therapy was zero. Routine use of TOE prior to AF ablation warrants reconsideration. A risk-based approach using CHADS₂ score, type of anticoagulation and type of AF should be considered in deciding whether a TOE is necessary prior to AF ablation.

Conflict of interest: none declared.