Abstract
Background
Direct Oral Anticoagulants (DOACs) are the first line anticoagulation for patients with non-valvular atrial fibrillation (NVAF). Percutaneous Left Atrial Appendage Occlusion (LAAO) has emerged as a new therapy and its safety and effectiveness compared with DOACs are still controversial.
Methods
A systematic review of randomized controlled trials and observational studies was conducted, focusing on patients with NVAF. Outcomes analyzed included: (1) all-cause mortality; (2) cardiovascular (CV) mortality; (3) thromboembolic events; (4) stroke or transient ischemic attack (TIA); (5) bleeding events; and a (6) composite of death, hemorrhagic, and thromboembolic events. We performed a subgroup analysis of major bleeding according to different definitions: (1) Bleeding Academic Research Consortium (BARC); (2) International Society on Thrombosis and Haemostasis (ISTH); and (3) other definitions.
Results
Ten studies involving 18,507 patients were included, with 42.35 % undergoing LAAO. In pooled analysis, LAAO was associated with lower rates of all-cause mortality (HR 0.63; 95 % CI 0.50-0.80), cardiovascular mortality (HR 0.56; 95 % CI 0.45-0.70), and of the composite outcome (HR 0.73; 95 % CI 0.58-0.92). A trend towards lower stroke/TIA events was observed but not statistically significant. Overall bleeding events did not significantly differ between groups; using the ISTH definition, LAAO showed significantly lower incidence of bleeding events (HR 0.63; 95 % CI 0.43-0.91). No difference was found in thromboembolic events.
Conclusion
LAAO was associated with a significantly lower all-cause mortality and cardiovascular mortality, as well as the composite of death, hemorrhagic or thromboembolic events, as compared with DOACs.
Introduction
Atrial fibrillation (AF) stands as the most prevalent cardiac arrhythmia and a major risk factor for ischemic stroke, particularly among the elderly It is projected that by 2050, 6-12 million people will be affected in the United States and 17.9 million people in Europe by 2060.
Warfarin historically served as the primary therapeutic approach for patients with AF requiring anticoagulation. Regarding stroke prevention for non-valvular atrial fibrillation (NVAF), direct oral anticoagulants (DOACs) have emerged as the first-line anticoagulation choice due to their heightened effectiveness and safety compared to vitamin K antagonists DOACs exhibit lower rates of mortality and major bleeding, maintaining noninferiority in preventing stroke and embolic events.
More recently, percutaneous left atrial appendage occlusion (LAAO) has emerged as a novel alternative for stroke prevention in AF. The randomized controlled trial (RCT) by Holmes et al. has demonstrated noninferiority in preventing cardioembolic strokes and bleeding risks when compared to long-term warfarin anticoagulation However, there is some controversy about the safety and effectiveness of the LAAO when compared with DOACs. The only RCT about this topic shows no inferiority of the method however, newest observational studies have shown conflicting data regarding rates of mortality, cardiovascular mortality and major bleeding events. ,
Methods
This systematic review and meta-analysis were performed and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) Statement guidelines and the Cochrane Collaboration Handbook for Systematic Review of Interventions guidelines. ,
Eligibility criteria
Inclusion criteria in this meta-analysis was restricted to studies that met all the following eligibility criteria: (1) randomized trials or observational studies; (2) comparing LAAO to DOACs; and (3) enrolling patients with atrial fibrillation. Furthermore, studies were only included if they reported at least one of the outcomes of interest. We excluded studies with (1) no control group; (2) non-observational studies (e.g. non-peer-reviewed studies, systematic reviews, letter to editors) and (3) overlapping patient populations.
Search strategy and data extraction
The prospective pre-specified protocol, along with its search was registered on PROSPERO on December 16, 2023, under protocol #494097. We systematically searched Medline, Embase and Cochrane Central Register of Controlled Trials databases from inception to November 2023 with the following search terms: “atrial fibrillation”, “Afib”, “amulet”, “watchman”, “cardiac plug”, “amplatzer”, “lambre”, “LAAO”, “left atrial appendage occlusion”, “LAA occlusion”, “left atrial appendage closure”, “LAA closure”, “LAAC”, “NOAC”, “DOAC”, “noacs” “doacs”, “non vitamin k anticoagulants”, “direct oral anticoagulants”, “new oral anticoagulants”, “edoxaban”, “dabigatran”, “apixaban”, “rivaroxaban”, “oral anticoagulants”, “xarelto”, “eliquis”, and “betrixaban”.
The references from all included studies, previous systematic reviews, and meta-analyses, were also searched manually for any additional studies and we stayed updated in case new studies were published about the topic. Two authors (J.F. and V.P.) independently extracted the data following predefined search criteria and quality assessment. For some studies, WebPlotDigitizer version 4.6 was used to extract data from Kaplan Meier curves. A complete electronic search strategy is reported in the Supplemental appendix.
Endpoints and subgroup analysis
The outcomes considered were: (1) all-cause mortality; (2) cardiovascular (CV) mortality; (3) thromboembolic events; (4) stroke or transient ischemic attack (TIA); (5) bleeding events; and a (6) composite of death, hemorrhagic, thromboembolic events. A subgroup analysis of major bleeding was made, separating the groups according to the study definition: (1) Bleeding Academic Research Consortium (BARC); (2) International Society on Thrombosis and Haemostasis (ISTH); and (3) other definitions. All definitions are described in the Supplemental Appendix.
Quality assessment
We evaluated the risk of bias in randomized studies using version 2 of the Cochrane Risk of Bias (ROB2) assessment tool Non-randomized studies were assessed with the Risk of Bias in Non-randomized Studies-of Interventions tool (ROBINS-I) Two authors independently completed the risk of bias assessment (J.F. and R.P.). Disagreements were resolved through a consensus after discussing reasons for discrepancy. Publication bias was investigated by funnel-plot analysis of point estimates in relation to study weights; the Egger’s test was not performed because it is only recommended by Cochrane if the analysis includes more than 10 studies.
Statistical analysis
Statistical analysis employed risk ratio (RR) and hazard ratio (HR) with 95 % confidence intervals to compare treatment effects for all the outcomes. Heterogeneity was assessed with I² statistics and Cochran Q test; p-values <0.10 and I²>25 % were considered significant for heterogeneity. RR and HR were computed using the Der Simonian and Laird random-effects model. R for Statistical Computing (2022, Vienna, Austria) was used for statistical analysis.
Results
Study selection and baseline characteristics
The diagram of study selection is demonstrated in Fig. 1 . The initial search identified 1,588 studies (Medline [n = 371], Embase [n = 1,097], and Cochrane [n = 120]). After title and abstract screening and removal of duplicates, 16 studies were selected to be fully reviewed according to inclusion and exclusion criteria. Of these, 9 studies were included, and a recently published article was added later after the search. This resulted in 10 studies, comprising 18,507 patients from 9 observational studies , , and 1 RCT
Study characteristics are in Table 1 . A total of 7,838 (42.3 %) patients underwent LAAO and 10,669 (57.6 %) were treated with DOACs. The included subjects were mostly male (53.8 %), with mean age of 75.4 years, and mean CHADSVASC score of 4.76 (LAAO group: 4.70; DOACs group: 4.81). The follow up ranged between 1 year to 5 years, with a mean of around 3 years (33.6 months).
| Authors | Country | Type of study | Follow-up, m | Subjects, n | Chosen therapy, n (LAAO | NOAC) | Age † , y | Female, n ( %) | NOACs, % | CHA2DS2-VASc † (LAAO | NOAC) | HAS-BLED † score (LAAO | NOAC) | Previous stroke/TIA, % (LAAO | NOAC) | Previous bleeding history, % (LAAO | NOAC) | Device type, % |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Ding 2022 | Global registry | Observational | 24 | 1,322 | 661 | 661 | 69.6 | 450 (34.0) | A: 51.4 D: 15.0 E: 1.5 R: 34.8 | N/A | N/A | 6.8 | 7.0 | ICB: 1.8 | 2.4 GIB: 6.4 | 6.8 | N/A |
| Kei-Yan Ng 2023 | China | Observational | 35 | 2,350 | 874 | 1,476 | 75.9 | 1,001 (24.6) | N/A | 4.40 | 4.50 | 4.6 | 4.6 | 38.7 | 39.5 | 43.1 | 37.9 | N/A |
| Khalid 2023 | USA | Observational | 60 | 4,496 | 2,248 | 2,248 | 75-79 1 | 2,156 (48.0) | N/A | 4.81 | 4.81 | N/A | 21.0 | 22.0 | N/A | Watchman (100) |
| Melillo 2023 | Italy | Observational | 60 | 192 | 96 | 96 | 74.6 | 60 (31.3) | N/A | 4.30 | 4.30 | 3.50 | 3.50 | 35.4 | 29.2 | 65.6 | 49.0 | Watchman (33.7) Amplatzer Cardiac Plug (22.3) Amplatzer Amulet (44) |
| Nielsen-Kudsk 2021 | Europe | Observational | 24 | 2,258 | 1,071 | 1,184 | 75.1 | 844 (37.4) | N/A | 4.20 | 4.30 | 3.3 | 3.4 | 31.1 | 31.8 | 74.1 | 75.0 | Amulet (100) |
| Noseworthy 2022 | USA | Observational | 12 | 4,410 | 2,205 | 2,205 | 75.8 | 966 (45.2) | A: 67.3 | D: 4.3 | E: 0.2 | R: 28.2 | 5.70 | 5.70 | N/A | 33.6 | 33.6 | ICB: 12.2| 12.2 | N/A |
| Osmancik 2022 | Czeck republic | RCT | 48 | 402 | 201 | 201 | 73.3 | 138 (34.3) | A: 95 | D: 4.0 | R: 0.5 | 4.70 | 4.70 | 3.10 | 3.00 | 32.8| 31.3 | N/A | Amulet (61.3) Watchman (35.9) Watchman-FLX (2.8) |
| Paiva 2021 | Portugal | Observational | 36 | 240 | 91 | 149 | 76.6 | 112 (46.7) | N/A | 4.30 | 5.30 | 3.00 | 4.00 | 34.1 | 30.9 | N/A | N/A |
| Tiosano 2023 | Israel | Observational | 12 | 456 | 114 | 342 | 77.3 | 184 (40.4) | N/A | 4.17 | 3.96 | 4.14 | 3.95 | 43.0 | 37.1 | 74.6 | 74.9 | Amulet or Watchman. c |
| Magnocavallo 2024 | Italy and Belgium | Observational | 25 | 554 | 277 | 277 | 79.1 | 160 (57.4) | R: 59 | A: 135 | E: 17 | DTI:197 | 5.8 | 5.8 | 3.0 | 3.0 | 50.9 | 45.5 | 40.4 | 38.6 | Watchman 2.5 (63.2) FLX device (37.8) |
† : Mean; LAAO: Left Atrial Appendage Occlusion; NOAC: Novel Oral Anticoagulation; TIA: Transient Ischemic Attack; RCT: Randomized Clinical Trial; A:Apixaban D: Dabigatran; E: Edoxaban; R: Rivaroxaban; DTI: Direct thrombin inhibitor;
1 : Most representative age group.
c Decision based on the operating physicians’ preference and clinical characteristics.
Regarding the patients who underwent LAAO therapy, three aspects must be highlighted: (1) the device used; (2) the antithrombotic therapy; and (3) the periprocedural complications. According to available data, the Watchman device was the most commonly used (2,530 patients; 65 %), followed by the Amulet (1,236 patients; 31.7 %); most studies included both devices. Different generations of these devices were also used, such as the previous Amplatzer Cardiac Plug (ACP) (19 patients; 0.4 %) and the newer Watchman FLX (108 patients; 2,8 %). The antithrombotic therapy post LAAO is demonstrated in Table 2 . Up to 6 months after the LAAO procedure dual antiplatelet therapy (DAPT) was the most common regimen (48.12 %), followed by single antiplatelet therapy (15 %), which included aspirin or P2Y12 inhibitors. Anticoagulants were used by 445 patients (4.4 %) after 6 months; which included DOACs (4.4 %), warfarin (0.8 %) and low molecular weight heparin (0.3 %). Regarding the procedural complications ( Table 3 ), the overall rate was low (3.4 %; considering the studies with available data) being the most common pericardial effusion (0.8%).
| Nielsen-Kudsk 2021, n ( %) | Noseworthy 2022, n ( %) * | Paiva 2021, % | Tiosano 2023 | Khalid 2023, n ( %) | Melillo 2023 * , n ( %) | Osmancik 2022, % | Kei-Yan Ng 2023, n ( %) | Magnocavallo 2024, n ( %) | Ding 2022 | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Time period | 1-3 months | 12 months | 24 months | 1 month | 3 months | 6 months | 12 months | 1-6 months | 6 weeks | Lifelong | 6 months | 1-2 months | 1-6 months | 1-3 months | At discharge | At 6 months | Until 6 months | >6 months | N/A |
| Anticoagulant + P2Y12 inhibitor | – | – | – | 738 (9.2) | 666 (9.4) | 61 (1) | 8 (0.2) | – | – | – | – | – | – | – | – | – | – | – | |
| Anticoagulant + aspirin | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 236 (85.2) | – | |
| Dual antiplatelet therapy | 318 (31.3) | 79 (8.3) | 66 (7.6) | – | – | – | – | – | Most patients | – | – | – | 141 (64) | 100 | 793 (90.7) | 515 (58.9) | – | – | |
| Aspirin | 532 (52.2) | 614 (64.5) | 543 (62.8) | – | – | – | – | 100 | – | Most patients | – | – | 12 (6) | – | 813 (93) | 777 (88.9) | – | 132 (47.6) | |
| P2Y12 inhibitor | 653 (8.2) | 4,272 (60.2) | 3,005 (50.4) | 499 (12.6) | – | – | – | – | – | 844 (96.6) | 554 (63.4) | – | – | ||||||
| Low-dose NOAC | – | – | – | 383 (4.8) | 63 (0.69) | 37 (0.3) | 21 (0.5) | – | – | – | – | – | – | – | – | – | – | 145 (52.3) | |
| Standard-dose NOAC | 39 (3.8) | 38 (4) | 41 (4.7) | 1,069 (13.4) | 159 (2.2) | 103 (1.7) | 83 (2.1) | – | – | – | – | 2 (1) | – | – | 220 (25.2) | 45 (5.2) | 41 (14.8) | – | |
| Warfarin only | – | – | – | 3153 (39.4) | 303 (4.3) | 116 (2) | 53 (1.3) | – | – | – | – | 7 (3.6) | – | – | 192 (22) | 9 (1) | – | – | |
| Low molecular weight heparin | – | – | – | – | – | – | – | – | – | – | – | 31 (16) | – | – | – | – | – | – | |
| No antithrombotic therapy | 76 (7.5) | 182 (19.2) | 186 (21.5) | 2,007 (25.1) | 1,629 (23) | 1,640 (44.3) | 3,284 (83.2) | – | – | – | 2,112 (94) | – | – | – | – | – | – | – | |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
