Author + information
- Received June 21, 2018
- Revision received August 27, 2018
- Accepted September 6, 2018
- Published online December 17, 2018.
- Mohamad Alkhouli, MD∗ (, )
- Tatiana Busu, MD,
- Kuldeep Shah, MD,
- Mohammed Osman, MD,
- Fahad Alqahtani, MD and
- Bryan Raybuck, MD
- ↵∗Address for correspondence:
Dr. Mohamad Alkhouli, Division of Cardiology, West Virginia University School of Medicine, 1 Medical Center Drive, Morgantown, West Virginia 26505-8059.
Objectives This study sought to assess the incidence and clinical impact of device related thrombus (DRT) following precautions left atrial appendage occlusion (LAAO).
Background Device-related thrombus is a known complication of LAAO. However, data on the incidence of DRT and its impact on outcomes are limited.
Methods The authors performed a meta-analysis of randomized and observational studies to calculate the pooled incidence of DRT and the pooled odds ratio (OR) of ischemic events in patients with DRT and those without DRT.
Results In the 66 included studies; the incidence of DRT was 351/10, 153 (3.8%, range 0% to 17%, I2 = 56.8). The diagnosis was made in <90, 90 to 365, and >365 days in 42%, 57%, and 1% of patients, respectively. There was no difference in DRT rates between the AMPLATZER (AMPLATZER, AGA Medical Corporation, Golden Valley, Minnesota) and WATCHMAN (WATCHMAN, Boston Scientific Corporation, Marlborough, Massachusetts) devices (3.6% vs. 3.1%, p = 0.24). In a meta-regression, age, gender, heart failure, diabetes, CHA2DS2–VASc score, previous stroke, and post-LAAO antithrombotic regimen did not explain the heterogeneity in the incidence of DRT. The pooled incidence of ischemic events in studies that compared outcomes of patients with and without DRT (32 studies; n = 7,689) was 13.2% (37 of 280) in patients with DRT and 3.8% (285 of 7,399) in those without DRT (OR: 5.27, 95% confidence interval [CI]: 3.66 to 7.59; p < 0.001, I2 = 0). In a sensitivity analysis including randomized trials and prospective multicenter registries, the incidence of DRT was 3.7%, and DRT remained associated with higher rates of ischemic events (13.5% vs. 4.4%, OR: 4.15, 95% CI: 2.77 to 6.22; p < 0.001, I2 = 0).
Conclusions DRT after LAAO is uncommon (3.8%) but is associated with a 4- to 5-fold increase in ischemic events. Further studies are needed to understand the underlying mechanisms and the optimal surveillance and management of DRT.
Percutaneous left atrial appendage occlusion (LAAO) has emerged as a feasible stroke-prevention strategy in patients with atrial fibrillation who are not optimal candidates for long-term oral anticoagulation (OAC) (1). The early safety concerns with LAAO have been allayed with a growing literature demonstrating an excellent safety profile of the procedure in real-world practice (2,3). However, this expanding literature also brought forth potential issues with LAAO, such as device-related thrombus (DRT) (4). Data from randomized trials and large registries suggested that DRT, albeit uncommon, might be associated with higher incidence of ischemic strokes (5–9). In this study, we performed a systematic review and a meta-analysis to assess the incidence, predictors, and clinical impact of DRT.
Our review protocol was conducted in accordance with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) reporting guidelines (Supplementary Protocol in the Online Appendix) (10). We conducted a literature search in PUBMED, MEDLINE, EMBASE, EBSCO, CINAHL, Web of Science, and Cochrane (June 1, 2018) to identify eligible studies using the Medical Subject Headings search terms and text word search. We manually searched reference lists of relevant studies for additional publications. When multiple publications from the same study population were found, data from the most inclusive report were used. The data were independently extracted in duplicate from full-text articles by 2 of the authors (T.B. and K.S.). Disagreements were resolved through consensus and arbitration by the senior authors (M.A. and B.R.). We included randomized clinical trials, single and multi-institutional observational registries reporting the incidence of DRT as an outcome following LAAO and published in peer-reviewed journals. We excluded case reports of DRT and duplicate reports (Figure 1). The following study characteristics were extracted: year of publication, study design, number of patients, clinical characteristics, data related to DRT (incidence, timing, treatment, and associated events). The endpoints of this study were: 1) the incidence of DRT; and 2) the rate of ischemic events (stroke, transient ischemic attack [TIA], or systemic embolization) in patients with and without DRT. We also compared the incidence of DRT among different LAAO devices and performed a meta-regression to assess the impact of certain baseline characteristics and post-LAAO antithrombotic regimen on the incidence of DRT.
Data synthesis and analysis
The meta-analyses were performed using Comprehensive Meta-Analysis version 2.0 (Biostat Solutions, Inc. [BSSI], Frederick, Maryland, www.meta-analysis.com). For each clinical endpoint, pooled estimates of odds ratio (OR) (1) with 95% confidence interval (CI) (7) were calculated, using the random effects model with the Mantel-Hansel (MH) method. Heterogeneity among individual study effect sizes was examined using the I2 index, tau-square, and the Q-test p value. Publication bias was formally assessed using funnel plots and the Egger's linear regression test of funnel plot asymmetry (Online Figure 1). All pooled estimates are displayed with a 95% CI (7); p values were considered statistically significant at <0.05. A sensitivity analysis was performed to investigate potential sources of inconsistency, including removal of single-center registries and including only randomized controlled trials (RCTs) and multicenter registries. Forest plots were generated to show the relative effect size of the individual studies on each clinical outcome.
A total of 3,437 potentially relevant citations were identified and screened. After removal of duplicate and nonrelevant studies, we retrieved 161 full-text articles for evaluation, of which 66 satisfied the selection criteria and were included in the meta-analysis (Figure 1). These studies enrolled 12,033 patients with sample sizes ranging from 12 to 1,739 patients. The patients were elderly (73 ± 9.2 years), and 59% of them had nonparoxysmal atrial fibrillation (AF). The mean CHA2DS2–VASc, and HAS-BLED score was 4.1 ± 1.6, and 3 ± 1.3, respectively. The WATCHMAN device (WATCHMAN, Boston Scientific, Marlborough, Massachusetts) was the most commonly used device (48%), followed by the AMPLATZER Cardiac Plug (AMPLATZER, Aga Medical Corporation, Golden Valley, Minnesota) (19%), and the Amulet device (14%) (Amulet, Abbott Laboratories, Abbott Park, Illinois). Procedural success was reported in 98% of patients. Antithrombotic regimen at discharge was oral anticoagulants in 20%, single or dual antiplatelet therapy in 45%, and other/none in 35%. Table 1 summarizes the pooled baseline characteristics of the patients in the included studies. Detailed baseline characteristics of individual studies included in our meta-analysis are provided in Online Table 1.
A total of 10,154 patients (84.4%) had at least 1 imaging study of the left atrial appendage (LAA) during follow-up. Among these patients, the pooled incidence of DRT was 3.8% (351 of 10,154), with significant variation in the reported incidence (range 0% to 17%; I2 = 56.8) (Figure 2). The diagnosis was made during follow-up imaging within 90 days, between 90 and 365 days and after 365 days in 42%, 57%, and 1% of patients with DRT, respectively (Figure 3). In the studies that reported the incidence of ischemic events in patients with or without DRT (32 studies, n = 7,689), the pooled incidence of ischemic events was 13.2% (37 of 280) among patients with DRT and 3.8% (285 of 7,399) among those without DRT (OR: 5.27, 95% CI: 3.66 to 7.59; p < 0.001; I2 = 0) (Figure 4). The majority of ischemic events in the DRT group were strokes (86%) (Table 1). In a sensitivity analysis including only RCTs and multicenter registries (26 studies, n = 7,552), the pooled incidence of DRT was 3.7% (268 of 7,552) (Online Figure 2), and DRT remained associated with significantly higher incidence of ischemic events (13.5% vs. 4.4%; OR: 4.15; 95% CI: 2.77 to 6.22; p < 0.001; I2 = 0) (Online Figure 3).
The incidence of DRT was similar in patients who underwent LAAO with the ACP/Amulet devices vs. the WATCHMAN device (100 of 2,744 [3.6%] vs. 138 of 4,443 [3.1%]; p = 0.24) (Online Table 2). There was a significant variability in the antithrombotic regimen used after LAAO between individual studies (Online Table 3). Nonetheless, there was no difference in incidence of DRT between patients who received short-term OAC post-LAAO and those who were treated with antiplatelet therapy (Online Tables 4 to 9). In a meta-regression assessing the impact of age, gender, diabetes, previous stroke, heart failure, CHA2DS2–VASc score, and post-procedural antithrombotic regimen on the incidence of DRT, none of these factors explained the heterogeneity in the incidence of DRT between the studies (Online Figures 4 to 10).
The treatment regimen was reported in 51% (178 of 351) of patients with established diagnoses of DRT. Among these patients, low–molecular-weight heparin was used in 37 of 178 patients (21%), and OAC with warfarin or direct thrombin inhibitors was used in 116 of 178 patients (65%). Follow-up studies were performed in 144 of 351 patients (41%), and these documented resolution of DRT in 140 of 144 patients (97%).
A limited number of studies assessed clinical and procedural predictors of DRT; in the study by Dukkipati et al. (n = 1,739), history of stroke or TIA (OR: 2.31; 95% CI: 1.26 to 4.25; p = 0.007), permanent AF (OR: 2.24; 95% CI: 1.19 to 4.20; p = 0.012), vascular disease (OR: 2.06; 95% CI: 1.08 to 3.91; p = 0.028), LAA diameter (OR: 1.06; 95% CI: 1.01 to 1.12; p = 0.019), and left ventricular ejection fraction (LVEF) (OR: 0.96 per 1% increase; 95% CI: 0.94 to 0.99; p = 0.009) were independent predictors of DRT (6). In the study by Fauchier et al. (n = 469), older age (hazard ratio [HR]: 1.07 per 1-year increase; 95% CI: 1.01 to 1.14; p = 0.02) and history of stroke (HR: 3.68; 95% CI 1.17 to 11.62; p = 0.03) were predictors of formation of thrombus on the devices, whereas dual antiplatelet therapy (HR: 0.10; 95% CI: 0.01 to 0.76; p = 0.03) and oral anticoagulation at discharge (HR: 0.26; 95% CI: 0.09 to 0.77; p = 0.02) were protective factors (7). In a smaller study by Kaneko et al. (n = 78), deep implant and higher CHA2DS2–VASc scores were the strongest predictors of DRT (OR: 24.7; p = 0.032 and OR: 2.8; p = 0.022, respectively) (11). In another small study by Pracon et al. (n = 99), patients with DRT had greater proportions of previous embolic events (7.4% vs. 30.4%; p = 0.04), deep implants (85.7% vs. 39.1%; p = 0.04), larger devices (30 mm vs. 25 mm; p < 0.01), and lower LVEF (50% vs. 60%; p < 0.01) compared with those without DRT (8).
The main findings of this present analysis are that: 1) DRT occurs in at least ∼4% of patients following LAAO, the vast majority of which were discovered at or after 6 months; 2) DRT is associated with a 4- to 5-fold increase in ischemic events; 3) although certain clinical and procedural factors predicted the occurrence of DRT in individual studies, no specific consistent predictor was identified in our meta-analysis and meta-regression; and 4) short-term anticoagulation is effective in resolving DRT in the majority of cases.
LAAO is an increasingly adopted stroke-prevention strategy in patients with AF who are not deemed suitable for long-term anticoagulation (1,2). Early in the LAAO experience, concerns arose of potential thrombus formation on the surface of the LAAO devices before their complete endothelialization. This was addressed in the RCTs by mandating OAC for at least 45 days after the procedure, a period after which complete endothelialization was believed to occur (12). However, an increasing number of studies reported the occurrence of DRT diagnosed at the time of routine follow-up transesophageal echo (TEE) both while patients are still on OAC or antiplatelet therapy or later (up to 10 years) after cessation (5-9,13). Until recently, studies investigating this worrisome complication remained limited to small case series, but several large reports from RCTs and multicenter observational registries have been published recently. Our study pooled individual study data from 66 RCTs and single and multicenter registries to provide a comprehensive analysis of the incidence or DRT and associated clinical events.
The incidence of DRT in our study was low (<4%), and this was consistent in our sensitivity analysis including only RCTs and multicenter observational registries. However, the true incidence might be difficult to assess because of the variations in surveillance protocols among the studies. Whether more frequent routine imaging will yield higher rates of DRT remains unclear. A particularly alarming finding, nonetheless, pertains to the timing and the setting during which the diagnosis of DRT was established: The majority of DRTs (85%) were discovered after more than 45 days, suggesting that the currently recommended LAA surveillance at 45 days is inadequate. In the 2 pivotal WATCHMAN RCTs and their nested registries, 29% of DRTs were detected on unscheduled TEE examinations conduced for other reasons, despite the robust LAA surveillance protocols (routine TEEs at 45, 180, and 365 days in the RCTs and at 45 and 365 days in the continuous access registries) (6). This implies that even frequent routine surveillance will likely miss a non-negligible percentage of DRTs and highlights the challenges of determining an optimal surveillance protocol following LAAO.
A significant positive association between DRT and subsequent ischemic events was observed in our meta-analysis, and this was consistent among the studies (I2 = 0), both in the overall analysis (OR: 5.27; 95% CI: 3.66 to 7.59; p < 0.001) and in the sensitivity analysis including RCTs and multicenter registries (OR: 4.15; 95% CI: 2.77 to 6.22; p < 0.001). Although a causative relationship could not be ascertained, as DRT could in theory be a manifestation of an increased global thrombotic tendency in some patients, a temporal relationship between DRT and ischemic events was noted by Dukkipati et al. (6), in which 47% of the ischemic events in patients with DRTs occurred within 1 month of the diagnosis. Yet despite this increase in ischemic events in patients with DRT, the vast majority of patients with DRT (89%) did not experience ischemic events, and most of the ischemic events reported in the included studies (88.5%) occurred in patients with no DRT. Hence, the search for an underlying mechanism for subsequent ischemic events following LAAO continues, as DRTs alone do not seem to explain the majority of these events. Nonetheless, these findings illustrate that DRT carries an increased risk of a subsequent ischemic event, and thus immediate resumption or continuation of OAC and rigorous follow-up to ensure resolution should be undertaken at the time of diagnosis. Among the studies reporting outcomes of treatment of DRT with oral or subcutaneous anticoagulants, resolution of the thrombus was documented in the overwhelming majority of patients.
Identifying predictors of DRT is key to institute effective preventive measures. However, owing to the low number of events in individual studies, the large number of plausible factors that need to be tested, and the variable antithrombotic regimens used in different studies, a conclusive evidence with regard to these predictors could not be synthesized. Specifically, device type, post-LAAO antithrombotic regimen, and key baseline characteristics (age, gender, diabetes, heart failure, previous stroke and CHA2DS2-VASc score) did not consistently predict DRT in our meta-analysis and meta-regression. Nonetheless, several clinical (previous stroke, permanent AF, older age, large LAA, low EF, vascular disease, higher CHA2DS2-VASc score, and so on), and procedural (deep implant, lack of discharge antithrombotic treatment, and so on) factors have been found to be independent predictors of DRT in individual studies. Among these factors, only 2 were modifiable (1). Deep device implantation was found to be a risk factor for DRT in 2 studies (8,11). This is intuitive, given that deep device implantation might leave significant trabeculated areas of the LAA uncovered, increasing the risk of formation of thrombus on the surface of the device. However, these studies were small, and hence this finding is yet to be confirmed in larger studies (1). The lack of OAC or antiplatelet therapy on discharge strongly correlated with the incidence of DRT in a large multicenter French registry (7). However, a clear association between post-LAAO antithrombotic regimen and DRT was not demonstrated in other studies (6,8). Until more data become available, avoiding deep device implantation and ensuring adequate antithrombotic treatment immediately after LAAO might aid in mitigating the risk of DRT. Whether the characteristics of the newer LAAO devices (less exposed knobs/pins on the atrial side in the WATCHMAN Flex and Amulet devices, more co-axial implantation with the Occlutech [Occlutech, Jena, Germany] device due to the steerable sheath), the increasing experience with the procedure, and the evolving strategies for post-LAAO antithrombotic treatment will affect risk of DRT remains to be tested in further studies.
There is no unified definition of DRT, and hence its diagnosis can be operator dependent. Also, most studies reported DRT as a binary event; quantification of the size/burden of thrombus was not undertaken.
A significant variability in follow-up imaging techniques and intervals existed among the studies. With the exception of the rigorous surveillances intervals undertaken in the pivotal WATCHMAN RCTs, the majority of patients in other studies received follow-up imaging between 1 and 3 months. This could have possibly underestimated the true incidence of DRT. In addition, although TEE was used in most studies, computed tomography imaging was also used in others. Computed tomography imaging of the LAA is not straightforward and has not been standardized, introducing another potential source of bias in the documentation of DRT (14).
To calculate the incidence of DRT, we excluded case reports/series and included only RCTs and observational registries. Hence, we could have possibly missed a small number of events. However, this was necessary to allow the calculation of a pooled incidence in this analysis.
The inclusion of single-center registries introduces the risk of reporting bias. However, we mitigated this by performing a sensitivity analysis including only RCTs and multicenter registries. In the sensitivity analysis, the incidence of DRT and the odds ratios of its association with ischemic events did not significantly change, which substantiates the findings of this meta-analysis.
DRT was diagnosed in a small percentage (∼4%) of patients who underwent LAA imaging at variable intervals following LAAO, and this was associated with 4- to 5-fold increases in ischemic events. Further studies are needed to identify the optimal surveillance protocol following LAAO to mitigate the risk of DRT and its associated ischemic events.
COMPETENCY IN MEDICAL KNOWLEDGE: Device-related thrombus occurs in ∼4% of patients after left atrial appendage occlusion and is associated with 4- to 5-fold increase in stroke. This rare but serious complication can occur late (>45 days) and at variable intervals after the procedure. No specific factors consistently predicted the occurrence of device-related thrombus including device type and post-procedural antithrombotic regimen.
TRANSLATIONAL OUTLOOK: Further studies are needed to identify modifiable predictors of device-related thrombus and its optimal prevention and screening strategies.
The authors have reported that they have no relationships relevant to 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
- left atrial appendage
- left atrial appendage occlusion
- oral anticoagulation
- device-related thrombus
- randomized controlled trial
- Received June 21, 2018.
- Revision received August 27, 2018.
- Accepted September 6, 2018.
- 2018 American College of Cardiology Foundation
- Reddy V.Y.,
- Doshi S.K.,
- Kar S.,
- et al.
- Reddy V.Y.,
- Gibson D.N.,
- Kar S.,
- et al.
- Landmesser U.,
- Tondo C.,
- Camm J.,
- et al.
- Valderrabano M.
- Kubo S.,
- Mizutani Y.,
- Meemook K.,
- Nakajima Y.,
- Hussaini A.,
- Kar S.
- Dukkipati S.R.,
- Kar S.,
- Holmes D.R. Jr..,
- et al.
- Fauchier L.,
- Cinaud A.,
- Brigadeau F.,
- et al.
- Pracon R.,
- Bangalore S.,
- Dzielinska Z.,
- et al.
- Saw J.,
- Tzikas A.,
- Shakir S.,
- et al.
- Kaneko H.,
- Neuss M.,
- Weissenborn J.,
- Butter C.
- Shamim S.,
- Magalski A.,
- Chhatriwalla A.K.,
- Allen K.B.,
- Huber K.C.,
- Main M.L.
- Cochet H.,
- Iriart X.,
- Sridi S.,
- et al.