Growing evidence supports left atrial appendage closure (LAAC) in selected patients with nonvalvular atrial fibrillation.1,2 The adoption of the procedure has expanded at a remarkable rate, reflecting the major unmet need for an effective stroke prevention strategy in patients for whom long-term oral anticoagulation (OAC) is unacceptable.3 Recent refinements in procedural planning and technique have optimized the LAAC procedure and minimized complications.4,5 Device-related thrombus (DRT) is one of the remaining major concerns associated with LAAC, as it has been linked to higher rates of stroke and ischemic events.6,7 Postimplant antithrombotic therapy (AT) aims to mitigate the risk of DRT during the sealing process of the atrial surface of the device.8,9 The type and intensity of AT after LAAC have been arbitrarily proposed as either a temporary OAC regimen for 6 to 12 weeks,10 or dual antiplatelet therapy for 12 weeks followed by single antiplatelet therapy for up to 1 year.11 However, nonprocedural bleeding associated with intensive AT, particularly dual antiplatelet therapy, after left atrial appendage occlusion (LAAO) is not uncommon and has been linked to an increased risk of mortality,12 potentially diminishing the benefit of the procedure. More recently, less intensive protocols using single antiplatelet therapy or even no AT have been reported in selected high-risk bleeding patients undergoing LAAC, with an acceptable occurrence of ischemic adverse events.13,14

We have recently shown that LAAC device implant depth is an independent risk factor for DRT.15 Large proximal uncovered areas and uncovered pulmonary ridge (PR) observed with a deep implant have been considered equivalent to a neoappendage, thereby increasing the risk of postimplant DRT. In addition, peridevice leakage (PDL) may cause turbulence in blood flow adjacent to the device and stagnant blood in a residual neo-cavity, which is a well-known predisposing factor for DRT.16 Accordingly, an optimal device implant, in terms of minimizing the risk of postprocedure DRT, might be one with a proximal implant and without significant PDL. In contrast, a suboptimal device implant may result from a deeper device position and/or the presence of significant PDL.15 Although several studies have highlighted the negative impact of distal implants and significant PDL, none have evaluated these factors in combination. Furthermore, the role of intensive AT following optimal or suboptimal LAAC implants remains uncertain. Therefore, the aim of the present study was to evaluate the incidence of DRT in patients receiving either intensive or nonintensive AT after achieving an optimal or suboptimal LAAC device implant.

A total of 1225 patients (35.0% female) undergoing successful transcatheter LAAC were included in the study (table 1 and figure 1). Among them, the achieved procedural result was optimal in 757 patients (61.8%) and suboptimal in 468 (38.2%) (figure 2). The AT in the study population is described in table 1 of the supplementary data. The baseline characteristics of the 2 groups are summarized in table 2 of the supplementary data. The proportion of patients with permanent nonvalvular atrial fibrillation was lower in the optimal implant group (48% vs 56%, P=.01). The baseline and clinical characteristics of the 2 groups according to postimplantation AT are summarized in table 1. Patients with nonintensive AT had a more frequent history of severe bleeding in both the optimal (93.0% vs 81.2% for those with intensive AT) and suboptimal implant groups (89.4% vs 85.1%, respectively). Compared with patients with intensive AT, the proportion of patients without intensive AT was higher among patients with a prior intracranial bleeding event and was lower among those with a prior gastrointestinal bleeding event (P=.027).

Figure 1.

Central illustration. Incidence and predictors of device-related thrombus after left atrial appendage closure in patients with an suboptimal or optimal implant receiving nonintensive or intensive AT. Left panel: main baseline and procedural characteristics and outcomes of the study population. Mid panel: incidence of device-related thrombus in patients with an optimal or suboptimal implant who were prescribed intensive or a nonintensive AT. Right panel: predictors of device-related thrombus in the study population. Exploratory analysis in patients with an optimal implant, showing no impact of AT intensity or device type. AT, antithrombotic treatment; BARC, Bleeding Academic Research Consortium; DRT, device-related thrombus; ICH, intracranial hemorrhage; LAAC, left atrial appendage closure; SE, systemic embolic events.

(0.64MB).

Figure 2.

Study flowchart. A total of 1225 patients with successful LAAC were stratified by implant features into the suboptimal LAAC device implant group (distal implant and/or moderate or greater peridevice leak) or optimal LAAC device implant group (proximal implant and/or less than moderate peridevice leak) and nonintensive antithrombotic therapy (none or aspirin) or intensive antithrombotic therapy (other than previously described) at discharge. AT, antithrombotic treatment; LAAC, left atrial appendage closure; PDL, peridevice leak.

(0.34MB).

Clinical and device outcomes

As shown in table 3, DRT was detected in 79 patients (6.7%). Overall, 81% of DRTs were detected <180 days post-LAAO and 19%>180 days after the procedure. Compared with patients with a suboptimal device implant, patients with an optimal implant showed a lower rate of DRT (22/757, 2.9% vs 57/468, 12.1%, respectively, P<.001). Moreover, compared with patients with nonintensive AT, the incidence of DRT was slightly but not significantly decreased with intensive AT in both groups with optimal (2.6% vs 3.7% for patients with nonintensive AT, P=.38) and suboptimal implants (11.2% vs 15.5%, respectively, P=.19), as shown in table 3, figure 1, and figure 3. Of note, the incidence of DRT was higher among patients with a suboptimal device implant than among those with an optimal implant, irrespective of the AT intensity. This association was independent of the type of implanted device (figure 1 of the supplementary data).

Table 3.

Device outcomes at maximum follow-up

Device outcomes	Total(n=1225)	Suboptimal implant+nonintensive AT(n=160)	Suboptimal implant+intensive AT(n=308)	Optimal implant+nonintensive AT(n=272)	Optimal implant+intensive AT(n=485)	P
Median first imaging FU, d	58 [44-88]	68 [47-101]	56 [45-81]	62 [44-96]	54 [44-82]	.01
Imaging type first FU						.23
TEE	740 (60)	95 (58)	190 (61)	155 (56)	300 (62)
CT	485 (40)	65 (42)	118 (39)	118 (44)	185 (38)
Second imaging FU	454 (37.1)	54 (34.0)	120 (38.8)	94 (34.5)	186 (38.4)	.32
Device migration	1 (0.1)	0	0	0	1 (0.2)	.68
Device-related thrombus (n=1185)	79 (6.7)	24 (15.5)	33 (11.2)	10 (3.7)	12 (2.6)	<.001
DRT diagnosis						.97
<180 d	64 (81)	20 (83)	26 (79)	8 (80)	10 (83)
>180 d	15 (19)	4 (17)	7 (21)	2 (20)	2 (17)
Peridevice leak	261 (22.3)	62 (41.3)	118 (41.6)	21 (7.9)	60 (12.7)	<.001
Peridevice leak						<.001
1-3 mm	115 (9.4)	10 (6.3)	26 (8.4)	21 (7.7)	58 (11.9)
3-5 mm	114 (9.3)	41 (25.6)	73 (23.7)	0	0
>5 mm	32 (2.6)	12 (7.5)	20 (6.5)	0	0

AT, antithrombotic treatment; CT, computed tomography; DRT, device-related thrombus; FU, follow-up; ICH, intracranial hemorrhage; TEE, transesophageal echocardiography.

Values are expressed as No. (%), or median [interquartile range].

Figure 3.

Device-related thrombus rates according to implant features and antithrombotic therapy at discharge. The proportion of patients with a DRT was not significantly different with intensive or nonintensive AT after a suboptimal (11.2% vs 15.5%, P=.19, respectively) or optimal device implant (2.6% vs 3.7%, P=.38, respectively). AT, antithrombotic treatment; DRT, device-related thrombus.

(0.08MB).

Four strokes were observed at 3 months in the suboptimal implant and intensive AT group (vs none in the other groups, P=.014). At 12 months of follow-up, clinical outcomes, including stroke, systemic embolism, major bleeding, and all-cause mortality, did not differ between the groups (table 3 of the supplementary data). The cumulative incidence of all-cause death, ischemic stroke and major bleeding at maximum follow-up showed no difference between the groups (figure 4).

Figure 4.

Clinical outcomes at follow-up. Clinical outcomes stratified by implant features and AT at discharge. A: all-cause death; B: ischemic stroke, and C: major bleeding events. 95%CI, 95% confidence interval; AT, antithrombotic therapy; HR, hazard ratio; LAAO, left atrial appendage occlusion.

(0.47MB).

Table 4 and figure 1 show the independent predictors of DRT. After multivariate analysis, suboptimal implantation was identified as the sole independent predictor of DRT. Importantly, device type and intensive AT did not emerge as independent DRT predictors. Optimal implantation was significantly associated with lower DRT, with no interaction of AT or device type as shown in figure 1 and figure 5.

Table 4.

Predictors of device-related thrombus

Characteristic	Univariate analysis		Multivariate analysis
	HR (95%CI)	P	HR (95%CI)	P
Age	0.99 (0.97-1.01)	.45	-	-
Female sex	1.35 (0.86-2.12)	.19	-	-
Prior ischemic stroke	0.86 (0.49-1.50)	.59	-	-
Permanent AF	1.19 (0.76-1.86)	.45	-	-
CKD (<60 mL×min – 1 per 1.73 m2)	0.98 (0.63-1.54)	.94	-	-
Suboptimal implant	4.72 (2.84-7.84)	<.001	4.51 (2.70-7.54)	<.001
Peridevice leak	0.82 (0.38-1.74)	.60	-	-
Iatrogenic pericardial effusion	1.21 (0.43-3.47)	.71	-	-
Intensive AT	0.67 (0.42-1.08)	.099	0.66 (0.40-1.07)	.09
Disc/lobe device	0.62 (0.39-0.98)	.04	0.70 (0.43-1.13)	.15
CHADS-VASc score ≥3	2.56 (0.80-8.26)	.12	-	-

95%CI, 95% confidence interval; AF, atrial fibrillation; AT, antithrombotic treatment; CKD, chronic kidney disease; HR, hazard ratio; OR, odds ratio.

Variables identified following univariable logistic regression of P values <.10 included in the final multivariable logistic regression model. Multivariable analysis data are presented as ORs with corresponding 95%CIs.

Figure 5.

Exploratory subgroup analysis for device-related thrombus according to optimal implantation. 95%CI, 95% confidence interval; AT, antithrombotic treatment; DRT, device-related thrombus; HR, hazard ratio.

(0.18MB).

The main findings of the present study assessing the impact of intensive vs nonintensive AT in patients with optimal and suboptimal implants after transcatheter LAAC were as follows: a) suboptimal implants were associated with a higher incidence of DRT after LAAC, which was not mitigated by intensive AT; b) in patients with optimal implants, the use of intensive AT might not be necessary, as the incidence of DRT was low and similar between the intensive and nonintensive AT groups; and c) suboptimal device implantation, but not intensive AT, emerged as an independent predictor of DRT.

DRT remains one of the main concerns after LAAC, as it is associated with a significant increase in ischemic stroke and systemic embolization.25 Considering the growing procedural volumes and the expansion of LAAC to lower-risk patients, predicting and preventing DRT is of the utmost importance. Risk factors for DRT are mostly nonmodifiable and related to patient characteristics, such as prior transient ischemic attack and stroke, permanent AF, advanced age, heart failure, and renal insufficiency.19 In contrast to these conditions, the depth of device implantation seems to be the only modifiable factor associated with varying DRT rates.15,16,18–22

Although the rationale for AT after LAAC is to prevent DRT by facilitating adequate device endothelialization, the association between post-LAAC AT and DRT remains contentious. Some studies showed a strong correlation between the type of AT after the procedure and the incidence of DRT,26 while others have found no significant association.7,19,22,27 These conflicting data are reflected in the significant heterogeneity of post-LAAC AT in current practice and result from a lack of evidence supporting a specific AT intensity after LAAC.28

Using data from the largest cohort to date, including both disc/lobe and single- lobe occluders, we classified patients according to achieved implant as either optimal (proximal implant and the absence of significant PDL) or suboptimal (distal implant and/or the presence of significant PDL). According to the latest evidence, different cutoffs were selected to define distal implants: ≥1mm for disc/lobe devices and>10mm for single lobe devices.15,22,24,27 In addition, the cutoff value for a large PDL was primarily considered to be>5mm, but recent data indicate that PDL 3 to 5mm may be associated with an increased rate of thromboembolic events after LAAC.28 Accordingly, PDL>3mm was considered significant for both devices and entered the definition of a suboptimal implant.

Our study reinforces the lack of additional benefits from intensive AT in preventing DRT, since both intensive and nonintensive AT were associated with comparable incidences of DRT in patients with either suboptimal or optimal implants. In other words, the present study suggests that in patients with an optimal device implant, the use of intensive therapies after LAAC might not be justified and would only translate into a potential higher risk of bleeding. Moreover, the intensity of AT did not seem to offset the deleterious effects of a distal implant. This finding is clinically relevant as several studies have previously reported a higher risk of DRT after deep implants.15,19,22 Thus, at hospital discharge, physicians were tempted to prescribe intensive AT in patients with a suboptimal device implant to decrease the incidence of DRT.

The incidence of PDL and DRT varies considerably among published studies due to the variability of imaging tools and the lack of standardized post-LAAC surveillance imaging. Because of higher spatial resolution, CT has been shown to be more sensitive than TEE in detecting PDL.24 However, while PDL detected by TEE has been associated with adverse events, residual leaks identified by coronary computed tomography angiography (CCTA) are more frequent but lack prognostic significance.21 These findings underscore the need for further research to characterize and identify high-risk PDLs using CCTA.

Regarding DRT, TEE and CCTA offer similar diagnostic capabilities, although CCTA provides higher sensitivity in detecting hypoattenuated thickening (HAT), a commonly observed imaging finding in all transcatheter and surgical devices.24 Although the presence of DRT has largely been considered a binary finding in the literature, HAT is now increasingly recognized on post-LAAC CT. The question of whether this subclinical finding should be considered a precursor to DRT remains open.

In the present study, initial imaging follow-up was performed using either TEE (60%) or CT (40%), and DRT was defined according to classical criteria as a thrombus adherent to the left atrial surface of the device. This definition excluded low grade HAT (grades 0 and 1). The first follow-up imaging study was performed between the 6th and 12th weeks after the procedure, at a median of 58 [44-88] days, when most patients were still receiving the AT prescribed at discharge. Consequently, most teams consider a DRT-free examination to be a prerequisite for AT de-escalation. As proposed by expert recommendations,17 a second follow-up imaging study was performed in 37% of patients between the 6th and 12th months after the index procedure, and it is notable that DRT were mostly diagnosed (81%) during the initial examination. This is not in accordance with previous reports showing a balanced detection between early and late timings.

For instance, in the PROTECT-AF and PREVAIL trials, 51% of DRTs detected during scheduled TEE were observed at the 1-year examination, and 62% of those detected during unscheduled examinations were observed at>1 year.10 Interestingly, in the AMULET IDE trial, most DRTs associated with the AMULET device (61%) occurred ≤45 days, whereas most DRTs associated with the WATCHMAN device (74%) were identified>45 days.29 This discrepancy has raised concerns about possible device-specific mechanisms underlying the timing of DRT formation, although our data showed consistency for both devices. Irrespective of our findings, the nature and duration of the optimal AT still remain an open and challenging question.

The incidence of DRT after LAAC was higher in patients with suboptimal implants, which was not mitigated by intensive AT. Compared with nonintensive AT, intensive AT was not associated with a lower incidence of DRT after either an optimal or suboptimal implant. Suboptimal implants, but not intensive AT, emerged as an independent predictor of DRT. Further studies are warranted to confirm these results.

This study received no funding.

The study was conducted in accordance with the institutional ethics committee of each participating center, and all patients provided signed informed consent for the procedures. The study conformed to the guiding principles of the Declaration of Helsinki. In accordance with the SAGER guidelines, the information was disaggregated in table 1 with no difference between groups.

The authors declare that they have not used any type of generative artificial intelligence for the preparation of this manuscript.

Each author contributed significantly to the submitted work. Substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work: P. Garot, P. Cepas-Guillén, X. Freixa. Drafting the work or revising it critically for important intellectual content: E. Flores-Umanzor, N. Leduc, V. Bajoras, N. Perrin, A. McInerney, A. Lafond, J. Farjat-Pasos, X. Millán, G. ÓHara, S. Zandjebil, R. Ibrahim, P. Salinas, O. de Backer, J.E Nielsen-Kudsk, I. Cruz-González, D. Arzamendi, L. Sanchis, L. Nombela-Franco, A. Aminian, J. Rodés-Cabau. All authors approved the final version to be published. All authors agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

P. Garot reports receiving speaker/advisory/proctor fees from Abbott, Biosensors, Boston Scientific, Edwards Lifesciences, General Electric HealthCare, and Terumo. He is medical director and shareholder of the Cardiovascular European Research Center in Massy, France. O. de Backer received institutional research grants and consulting fees from Abbott and Boston Scientific. R. Ibrahim reports receiving speaker/advisory/proctor fees from Abbott, Boston Scientific, Edwards Lifesciences, and Medtronic. L. Sanchis is proctor for Abbott Medical and is associate editor of Rev Esp Cardiol. The journal's editorial procedure to ensure impartial handling of the manuscript has been followed. A. Aminian is a proctor and consultant for Abbott and Boston Scientific. J.E. Nielsen-Kudsk is a proctor and consultant for Abbott and Boston Scientific. X. Freixa is a proctor for Abbott and Boston Scientific. The remaining authors report no potential conflicts of interest with respect to the content of this article.