Feasibility of Left Atrial Appendage Closure in Atrial Fibrillation Patients with a History of Intracranial Bleeding: A Systematic Review of Observational Studies

Background Left atrial appendage occlusion (LAAO) is performed in patients with nonvalvular atrial fibrillation to reduce the risk of ischemic stroke. The patients with a history of intracranial hemorrhage were excluded from the pioneer randomized controlled trials. The purpose of this systemic review was to evaluate the data from observational studies reporting the efficacy and safety of LAAO in patients with a history of intracranial hemorrhage. Methods Ovid MEDLINE, Embase, Cochrane Central Register of Controlled Trials, Web of Science Core Collection, Scopus, Global Index Medicus, and ClinicalTrials.gov data sources were utilized for data collection. Results A total of 12 studies met the inclusion criteria that included seven retrospective observational and five prospective observational studies. A total of seven hundred and twenty-seven patients with a history of intracranial hemorrhage underwent percutaneous left atrial appendage occlusion. There were 11 events of recurrent intracranial hemorrhage, 12 ischemic strokes, 4 transient ischemic attacks, and 26 all-cause deaths. The duration of follow-up varied from 3 months to 3.6 years in the included studies. Conclusion Left atrial appendage occlusion can potentially be an effective and relatively safe treatment option to reduce the risk of ischemic stroke in selected patients with nonvalvular atrial fibrillation patients and history of intracranial hemorrhage. Future prospective randomized trials are needed to validate this approach.


Introduction
Atrial fibrillation is associated with a high arterial thromboembolic risk, leading to increased morbidity and mortality. e use of anticoagulants (warfarin or direct oral anticoagulants) effectively reduces the risk of thrombosis [1,2]. However, the side effect profile of oral anticoagulants restricts their use in patients who have a history of major bleeding or if they are at increased risk of bleeding.
Approximately, 90% of thrombus in atrial fibrillation are formed in the left atrial appendage. erefore, mechanical left atrial appendage occlusion (LAAO) is a novel therapeutic modality used to reduce the risk of atrial thromboembolism in patients with nonvalvular atrial fibrillation (AF) and increased risk of bleeding [3,4]. e two pivotal trials, PROTECT AF (Watchman Left Atrial Appendage System for Embolic Protection in Patients with Atrial Fibrillation) and PREVAIL (Evaluation of the Watchman LAA Closure Device in Patients with Atrial Fibrillation Versus Long Term Warfarin erapy), evaluated the effectiveness of LAAO with Watchman device (Boston Scientific, St. Paul, Minnesota) compared with warfarin [3,5]. e results of these two trials demonstrated that LAAO is noninferior to warfarin in preventing ischemic stroke and is superior in preventing cardiovascular and all-cause mortality. Subsequently, the Watchman device was approved by the United States Food and Drug Administration (FDA) in 2015 [5]. In 2019, American Heart Association/American College of Cardiology/Heart Rhythm Society recommended LAAO for patients who are at increased risk for thromboembolism and have contraindications to anticoagulation [6].
In PROTECT AF and PREVAIL trials, individuals with a history of an intracranial hemorrhage were excluded due to the perceived risk of recurrent intracranial hemorrhage perioperatively [5]. However, in real-world, LAAO is being used in AF patients with a history of intracranial hemorrhage and there are emerging data from observational studies and registries that this modality might be safe to use [7].
Despite the increasing use of LAAO procedure in AF patients with a history of intracranial hemorrhage, the data regarding efficacy and safety are limited. erefore, we conducted a systematic review of literature to evaluate the efficacy and safety of mechanical left atrial appendage occlusion in nonvalvular atrial fibrillation patients with a history of intracranial hemorrhage.

Study Selection.
is systematic review is reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [8]. Articles were screened and selected by 2 independent reviewers using 3-step approach. First, all the articles mentioning left atrial appendage closure were screened, and there were a total of 10379 articles; after removing duplicates, the remaining articles were 7201. en out of these, articles reporting left atrial appendage closure in intracranial hemorrhage and/or high-risk patients were reviewed for relevance, and 26 articles were assessed for eligibility. After that, articles with incomplete characteristics or the ones not reporting separate bleeding or thromboembolic risk scores for the intracranial hemorrhage cohort were removed, and the final 12 observational studies were extracted for final analysis.
ese steps are shown in the PRISMA file ( Figure 1).

Outcomes.
e outcomes of interest included devicerelated thrombus (DRT), ischemic stroke/TIA, recurrence of intracranial hemorrhage, and all-cause mortality.

Data Extraction and Quality Assessment.
Two reviewers independently extracted data on baseline characteristics, type of intracranial hemorrhage, duration between intracranial hemorrhage and device implantation, type of devices used, antithrombotic and their duration after device implant, follow-up and outcomes of device-related thrombus, ischemic stroke, recurrent intracranial hemorrhage, and mortality. Quality of the studies was assessed by two reviewers using National Heart, Lung and Blood Institute (NIH) Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies [9].

Data Synthesis and
Analysis. Data were summarized using descriptive statistics, with means, medians, and ranges for continuous variables and frequencies and percentages for dichotomous variables.

Study Selection.
We identified 7201 articles, and out of these, 12 observational studies (n � 727) met the inclusion criteria. ere were seven retrospective and five prospective studies. e study characteristics are summarized in Table 1.

Patient Characteristics.
A total of 727 patients underwent LAAO, in which 65.1% were males and 34.9% were females. Mean age was 74.1 ± 2.2. CHA2DS2-VASc risk score and HAS-BLED risk scores were reported by mean and median in different studies as shown in Table 1. e type of intracranial hemorrhage was not reported in 3 studies that contributed 358 patients from a total cohort of 727 patients. Out of 365 patients, 71% had intracranial hemorrhage, 19% had subdural hemorrhage, and 7.5% had subarachnoid hemorrhage, while 2.5% had microhemorrhages, ocular bleeding, or other bleedings. e data for four patients were missing. e duration between intracranial hemorrhage and LAAO procedure varied among studies, but the average duration was about two months, as shown in Table 1. All three major devices were used in all the procedures which included Watchman, Amplatzer Cardiac Plug, and Amplatzer Amulet.

Outcomes.
Primary outcomes were device-related thrombus (DRT), ischemic stroke/TIA, recurrence of intracranial hemorrhage, and mortality. e results of primary outcomes are summarized in Table 2. DRT was not reported in 3 studies which included 264 patients, and in the remaining 463 patients, there were 6 DRT (1.3%). e ischemic stroke and TIA were reported in all studies which were 12 (1.6%) and 4 (0.5%), respectively. e recurrence of intracranial hemorrhage was reported in all 12 studies with a total of 11 events (1.5%). e overall mortality rate was 3.6% (26/727). Twenty-two deaths were reported in one study of 104 patients with a median follow-up of 3.6 years by Pouru et al. (Table 2). is cohort included a total of 75% arterial thromboembolic events, 64% recurrent intracranial hemorrhage, and 85% deaths reported in our study.
ere were limited data on the day of occurrence of the event and which antithrombotic regimen patients were on at the time of primary outcomes. Follow-ups were reported in absolute days or months, mean with standard deviation (SD), or median with interquartile range (IQR) in different studies, as shown in Table 2.
e duration of follow-up varied from 3 months to 3.6 years depending upon the study.

Quality Assessment.
e quality of the studies was lowto-moderate based on the NIH quality assessment tool, which is reported in Supplementary Material 3. Titles of all the articles were relevant to the study question, and baseline characteristics were well described in all the studies. 9 out of 12 studies reported the type of intracranial hemorrhage, and most of the studies reported duration between intracranial hemorrhage and device implant. Antithrombotic regimen after the procedure and device-related thrombus (DRT) were reported in most of the studies. Primary outcomes of recurrence of intracranial hemorrhage, transient ischemic attack (TIA)/stroke, and mortality were reported in all the studies.

Discussion
e findings from our study suggest that left atrial appendage occlusion can offer an effective treatment option with a satisfactory safety profile to reduce the risk of ischemic stroke in selected patients with nonvalvular atrial fibrillation and history of intracranial hemorrhage. ere are several challenges in establishing the efficacy and safety of available treatment options for this high-risk cohort of atrial fibrillation. Based on the average CHADSD2-VASc and HAS-BLED risk scores, the inherent predisposition to a higher risk for arterial thromboembolism from atrial fibrillation and major bleeding from antithrombotic therapies are the most important considerations. In addition, this cohort mostly represents the geriatric population who are more likely to be frail and high risk for falls and injuries and, subsequently, higher risk of adverse intracranial hemorrhagic events [10].
e results of our study are supporting the emerging evidence indicating the safety and efficacy of LAAO with the perioperative short-term use of anticoagulation in patients with contraindications to anticoagulation including a history of intracranial hemorrhage. Barakat et al. reported no recurrence of spontaneous bleeding in  Journal of Interventional Cardiology 20 consecutive patients with contraindications to anticoagulation including 7 of whom had a history of intracranial hemorrhage [11]. Antithrombotic regimen included anticoagulation for 45 days, followed by 4.5 months of dual antiplatelet with aspirin and clopidogrel and then lifelong aspirin. J.R. Lopez-Minguez studied 598 patients who underwent LAA occlusion at 13 tertiary care centers across the Iberian Peninsula, which included 160 patients with previous intracranial hemorrhage. At a mean follow-up of 22.9 months, there was a 0.8% recurrence of intracranial hemorrhage and, at a follow-up of >24 months, recurrence of intracranial hemorrhage was 0.4% and the expected recurrence of bleeding based on the HAS-BLED score was 0.9% and 0.8%, respectively [12]. e antithrombotic regimen consisted of 600 mg loading of clopidogrel with aspirin 300 mg on the first day, followed by 100 mg aspirin daily for at least 6-12 months along with clopidogrel 75 mg daily for 3-6 months.
ASAP (ASA Plavix Feasibility Study with Watchman Left Atrial Appendage Closure Technology) is a nonrandomized, prospective study, which evaluated the safety and efficacy of LAAO in high-risk population. is cohort of 150 patients had contraindications to anticoagulation and was at high risk for ischemic stroke. ese patients underwent LAAO and received dual antiplatelets with aspirin and clopidogrel for 6 months without receiving warfarin and then lifelong aspirin [13]. e recurrence of intracranial hemorrhage and thromboembolism reported in the high-risk population of ASAP study, although the design and patients' characteristics were different, was 0.7% and 2.6%, while our systemic review containing 727 patients reports recurrent intracranial hemorrhage of 1.5% and thromboembolism (TIA and stroke) in 2.2% of patients. e review of 4 prospective clinical trials reporting device-related thrombus (DRT) in the clinical trials of left atrial appendage occlusion including 1739 patients by Dukkipati et al. found that DRT was 3.7% compared to 1.3% in 464 patients in our study [14].
After comparing the results of our study to other studies of the high-risk population (population at high risk for stroke and bleeding), we propose that left atrial appendage occlusion in patients with nonvalvular atrial fibrillation and a history of intracranial hemorrhage is safe and effective. We also propose that the timing between the intracranial bleeding event and LAAO should be individualized and decided by a multidisciplinary approach. Regarding the antithrombotic regimen, we suggest that strategy of anticoagulation for 6 weeks after LAAO followed by 4.5 months of DAPT and then lifelong aspirin or DAPT (aspirin and clopidogrel) without the use of anticoagulation for 6 months after LAAO followed by lifelong aspirin can be considered.
is review provides the most up to date evidence of left atrial appendage occlusion in patients with nonvalvular atrial fibrillation and a history of intracranial hemorrhage. ere are several limitations. e conclusions are primarily drawn from the observational studies with no comparator group, which poses a potential risk of confounding and bias. ere is a potential for a higher likelihood of invalidity because of the lack of standardized approach for the assessment and treatment, and the potential for loss of follow-up might have affected the interpretation of the results. ere was also a lack of generalized approach for the antithrombotic regimen postprocedure compared to the pioneer trials, and the timing of the procedure after intracranial hemorrhage was also not uniform and we cannot generalize one approach to all the patients. e included studies did not report safety and efficacy comparison among different left atrial appendage devices, and we cannot conclude the superiority of any device. Pouru et al. (Table 2) reported more safety events compared to the other studies, although we can speculate that it might be related to the longer duration of follow-up with a mean follow-up of 3.6 years compared to the other studies but the author is unable to conclude any significant reason for the higher safety events, as the baseline characteristics in this study were comparable to other studies. Finally, the duration of follow-up in the included studies was relatively short, and longer follow-up might lead to different outcomes.
In summary, the findings from our study indicate that LAAO may be considered in selected patients with nonvalvular atrial fibrillation and a history of intracranial hemorrhage. Although the   Journal of Interventional Cardiology