Atrial fibrillation - A multidisciplinary approach to improving patient outcomes: Part 2.pdf (Atrial fibrillation)

c h a pt e r 8 Left Atrial Appendage Excision, Ligation, and Occlusion Devices Taral K. Patel, MD, and Bradley P. Knight, MD ATRIAL FIBRILLATION AND STROKE Atrial fibrillation (AF) currently affects up to 5 million Americans and remains the most common arrhythmia encountered in clinical practice.1,2 With an aging population, the burden of AF is expected to rise 3-fold by 2050.3 Among the several downstream consequences of AF, the most feared is stroke due to thromboembolism. The primary cause of thrombus formation is mechanical dysfunction in the atria, leading to impaired blood flow and stasis. AF also promotes endothelial dysfunction, inflammation, platelet activation, and hypercoagulability, which further contribute to thrombus formation.4–6 Stroke remains the number one cause of major disability and the third leading cause of death in the United States.7 AF increases stroke risk 5-fold, leading to a 5% annual stroke rate for all-comers.7 Seen another way, the percentage of strokes attributable to AF ranges from 1.5% in those aged 50 to 59 years to an impressive 23.5% in those aged 80 to 89 years.7 While these statistics are dramatic, the influence of AF on stroke is almost certainly underestimated as AF is commonly silent and underdiagnosed.8 LEFT ATRIAL APPENDAGE Johnson and colleagues described the left atrial appendage (LAA) as “our most lethal human attachment.”9 Derived from the embryonic left atrium, the LAA forms a blind pouch 2 to 4 cm long and most commonly lies on the anterior surface of the heart. Its narrow neck forms a natural obstacle to normal blood flow. The LAA endocardial surface is highly irregular due to the presence of pectinate muscles. This is in sharp contrast to the true left atrium, which is derived from venous tissue and has a smooth endocardial surface. The LAA also has a variable number of lobes; an autopsy survey of 500 patients found that 20% had one lobe while 77% had two or three lobes.10 Atrial Fibrillation: A Multidisciplinary Approach to Improving Patient Outcomes © 2015 Joseph S. Alpert, Lynne T. Braun, Barbara J. Fletcher, Gerald Fletcher, Editors-in-Chief, Cardiotext Publishing, ISBN: 978-1-935395-95-9 109 110 Se ct io n 1: At ria l Fib rilla t io n : Ba ckg ro u n d , Eva lu a t io n , a n d Ma n a g e m e n t The LAA, because of its complex anatomy, innumerable potential spaces, and low blood flow during AF, is particularly susceptible to thrombus formation. Studies using magnetic resonance imaging (MRI) and transesophageal echocardiography (TEE) have suggested that larger LAA ostia, more lobes, and greater length all predict higher risk of stroke.11 An important review of 23 studies found that 17% of patients with nonrheumatic AF had left atrial thrombi, of which a striking 91% were located in the LAA.12 It is now well-accepted that the vast majority of strokes caused by AF represent thromboembolism originating from the LAA. LIMITATIONS OF ORAL ANTICOAGULATION Stroke prevention is the foundation of AF management. Currently, the standard of care is oral systemic anticoagulation by using the widely adopted CHADS2 stroke risk-assessment tool.13,14 The newer CHA2DS2-VASc score has helped further refine stroke risk in patients with otherwise low CHADS2 scores.15 These scoring systems balance the bleeding risk from anticoagulation with the thromboembolic risk from untreated AF. Supported by decades of data, oral anticoagulation has been unequivocally effective in reducing stroke. Warfarin, still the predominant anticoagulant, was demonstrated to reduce AF-related stroke by 64% in an extensive meta-analysis.16 However, the widespread use of systemic anticoagulation has highlighted several important limitations of this strategy. Most importantly, systemic anticoagulation unavoidably increases bleeding risk. Up to 40% of AF patients have relative or absolute contraindications to anticoagulation, usually owing to a history of pathologic bleeding or an elevated risk of falls.17,18 The HAS-BLED score has helped quantify the bleeding risk of warfarin in a manner analogous to the CHADS2 score for stroke risk. It is notable that several components of the HASBLED score—hypertension, prior stroke, and advanced age—are also found in the CHADS2 score. In other words, patients at high risk for stroke also happen to be patients at high risk for bleeding, illustrating the complexity in properly selecting patients for oral anticoagulation. Aside from bleeding risk, warfarin use is further limited by the inconvenience of frequent blood testing and extensive interactions with food and other medications. Often because of these limitations, warfarin is not utilized in up to 50% of eligible AF patients.19 Even when patients are treated with warfarin, they spend up to half of the treatment time outside the therapeutic range.20 Motivated by the challenges of using warfarin, the newer oral anticoagulants dabigatran (a direct thrombin inhibitor), rivaroxaban (a factor Xa inhibitor), and apixaban (a factor Xa inhibitor) were developed and are now in general clinical Chapte r 8 LAA Excisio n, Lig atio n, and Occlusio n De vice s use. These novel agents are comparably effective to warfarin with equivalent or lower bleeding risk.21–23 They have the advantage of minimal food and drug interactions and also eliminate the need for INR monitoring, increasing the ease of use and compliance. Unfortunately, they still suffer from the problem of elevated bleeding risk; this risk is further heightened because, unlike warfarin, the new drugs are not easily reversible with blood-product transfusion. Finally, the new agents are more costly and, at present, it is unclear whether they are truly cost effective in comparison with warfarin. Even with improved oral anticoagulation options, there remains a more fundamental issue. Because AF-related stroke appears to be largely a focal problem— thromboembolism from the LAA—a focal approach would be preferable to the currently imprecise strategy of systemic anticoagulation. Theoretically, a procedure to exclude the LAA (either by excision or by ligation or occlusion) should offer similar stroke prophylaxis while eliminating the disadvantages of systemic anticoagulation. LAA exclusion would be especially appealing for patients with either intolerance or contraindications to anticoagulation. In recent years, substantial progress has been made in developing techniques to exclude the LAA as a viable alternative for stroke prevention in AF. LEFT ATRIAL APPENDAGE EXCLUSION: SURGICAL TECHNIQUES LAA exclusion was first reported in 1949, when the surgeon Madden 24 published a case series of 2 patients who underwent LAA removal as a prophylaxis for recurrent arterial emboli. The high morbidity and mortality of the procedure prevented its widespread adoption for decades, until interest was reignited in the 1990s by the development of the Cox-Maze III procedure, which included removal of the LAA.25 Surgical techniques have evolved along two lines: LAA exclusion (using various suture techniques) and LAA excision (via surgical stapler or removal with oversew). Data for LAA surgery consist primarily of case reports and retrospective case series. Intepretation of the data is hampered by nonuniform surgical techniques and nonstandardized outcomes measurements. The use of TEE, considered the gold standard for LAA visualization, is absent in many reports. A large review of existing literature found that surgical success was highly dependent on both operator and technique; complete LAA closure rates ranged from 17% to 93%.26 Excision and oversew appeared to demonstrate the most durable results. A recent pilot trial randomized 51 patients to surgical LAA closure versus oral anticoagulation and demonstrated comparable stroke rates during follow-up.27 The results 111 112 Se ct io n 1: At ria l Fib rilla t io n : Ba ckg ro u n d , Eva lu a t io n , a n d Ma n a g e m e n t pave the way for a larger trial to answer the critical question of whether surgical LAA exclusion effectively reduces stroke risk. Current ACC/AHA guidelines limit surgical LAA exclusion as an adjunctive procedure during mitral valve or Maze surgery.13 However, two recently developed devices may rekindle interest in stand-alone surgical LAA exclusion. The first, AtriClip LAA Exclusion System (Atricure, West Chester, OH), is approved in both the United States and Europe, although it is indicated only in conjunction with other open cardiac surgical procedures in the United States. The device consists of a titanium ring covered by a woven polyester fabric. Under direct visualization, the clip is secured around the base of the LAA using a special deployment tool. In the largest trial to date, 70 patients undergoing open cardiac surgery in seven US centers had the AtriClip successfully placed.29 Of the 61 patients who underwent imaging at 3 months, 60 achieved persistent LAA exclusion. There were no device-specific adverse events reported. Although this was a small study with short-term follow-up, it demonstrated that the device could be deployed safely during open cardiac surgery. The second device involves a minimally invasive thoracoscopic approach. After left lung deflation, an endoscopic cutter (Ethicon Endo-Surgery, Cincinnati, OH) is introduced via the left lateral thorax. The cutter then simultaneously removes the LAA and staples its base closed. The procedure eliminates the need for thoracotomy, although concerns remain about the risks of lung deflation and the potential for catastrophic bleeding into a closed chest. Ohtsuka et al.30 published their experience with the technique in 30 patients with prior thromboembolism, achieving 100% procedural success and no major complications. Anticoagulation was discontinued and no recurrence of thromboembolism occurred after 18 months of follow-up. These preliminary data suggest that stand-alone surgical LAA exclusion may eventually have a place alongside the various transcatheter techniques. LEFT ATRIAL APPENDAGE EXCLUSION: TRANSCATHETER TECHNIQUES In an effort to avoid the morbidity of open surgery for LAA exclusion, minimally invasive percutaneous techniques have rapidly developed over the past decade. Of these, 4 have been tested in humans and shown promise. PLAATO Device Important for historical purposes, the Percutaneous LAA Transcatheter Occlusion (PLAATO) device (ev3 Endovascular, Plymouth, MN) became the first device of Chapte r 8 LAA Excisio n, Lig atio n, and Occlusio n De vice s 113 Fig u r e 8 .1 The PLAATO device, mounted on its delivery catheter. Source: Reprinted with permission from Syed T, Halperin J. Nat Rev Cardiol. 2007:4;428–435. its kind deployed in humans in 2001. The device consisted of a self-expanding nitinol cage covered by a blood-impermeable polytetrafluoroethylene membrane (Figure 8.1). The device was deployed in the LAA via transseptal catheterization under fluoroscopic and TEE guidance. Clinical experience with PLAATO was reported in 3 small studies. Sievert et al.31 implanted the device in 15 patients with 100% procedural success and one incident of hemopericardium. A larger international registry of 111 patients reported a 97% implant success rate and a 6% adverse event rate, including one death.32 The 10-month stroke rate of 2.2% compared favorably with the CHADS2-predicted rate of 6.3%. A North American registry of 64 patients reported 100% procedural success.33 After 5 years of follow-up, the stroke rate was 3.8%, a relative risk reduction of 42% from the expected stroke rate of 6.6%. Despite this promising clinical experience, the PLAATO device was withdrawn from development in 2006. However, its design became the inspiration for the subsequently developed WATCHMAN device. WATCHMAN Device The WATCHMAN device (Boston Scientific, Natick, MA) was first implanted in 2002. It also consists of a self-expanding nitinol frame, but is open-ended and has a permeable polyethylene membrane that only covers the part of the device exposed to the left atrium (Figure 8.2). The WATCHMAN device is also delivered via a transseptal system (Figure 8.3). Initial protocols required at least 6 weeks 114 Se ct io n 1: At ria l Fib rilla t io n : Ba ckg ro u n d , Eva lu a t io n , a n d Ma n a g e m e n t A B Fig u r e 8 .2 (A) The WATCHMAN device consists of a nitinol frame and permeable membrane. (B) Illustration of the device properly deployed in the left atrial appendage. Source: Used with permission of Boston Scientific Corporation. Chapte r 8 LAA Excisio n, Lig atio n, and Occlusio n De vice s 115 Fig u r e 8 .3 Fluoroscopic image of the WATCHMAN device (arrow) deployed in the left atrial appendage. of warfarin post-implant to prevent thrombus formation prior to device endothelialization. Warfarin was discontinued once a follow-up TEE demonstrated no flow into the LAA, signifying complete endothelialization. Subsequently, a strategy of substituting dual antiplatelet therapy for warfarin was evaluated in 150 warfarin-ineligible patients who underwent WATCHMAN implantation.34 After 14 months of follow-up, the actual ischemic stroke rate was 1.7% compared with the CHADS2-predicted rate of 7.3%, demonstrating that WATCHMAN implantation without a warfarin transition was a viable alternative for patients with contraindications to anticoagulation. Following several feasibility studies, the WATCHMAN device underwent a head-to-head trial against warfarin in the landmark PROTECT-AF trial.35 116 Se ct io n 1: At ria l Fib rilla t io n : Ba ckg ro u n d , Eva lu a t io n , a n d Ma n a g e m e n t To date, this study represents the only randomized trial comparing LAA exclusion with anticoagulation. In PROTECT-AF, 707 patients from 59 centers in the United States and Europe were randomized 2:1 to WATCHMAN versus warfarin therapy. Patients had relatively low stroke risk (68% had a CHADS2 score of 1 or 2) and no contraindications to warfarin. Overall implant success rate was 91% and at 6 months, 92% of patients in the WATCHMAN arm had discontinued anticoagulation. The trial was designed to test noninferiority of WATCHMAN to standard warfarin therapy. After 1065 patient-years, the primary efficacy end point (stroke, systemic embolism, or cardiovascular or unexplained death) was superior in the WATCHMAN arm versus the warfarin arm (3.0% vs. 4.9% per 100 patientyears), fulfilling the criteria for noninferiority. However, the primary safety end point (excessive bleeding or procedure-related complications) was worse in the WATCHMAN group (7.4% vs. 4.4%). Procedure-related complications included 22 pericardial effusions, 4 air emboli, and 3 device embolizations. On the other hand, the warfarin group had higher rates of major bleeding (4.1% vs. 3.5%) and hemorrhagic stroke (2.5% vs. 0.2%). In 2013, the 2.3-year results of PROTECT-AF were published, highlighting the durability of the initial results.36 After 1588 patient-years, the primary efficacy end point occurred in 3.0% of WATCHMAN patients and 4.3% of warfarin patients, again meeting criteria for noninferiority. With respect to the safety event rate, the WATCHMAN group continued to fare worse (5.5% vs. 3.6%), although the gap had narrowed. As expected, the adverse events in the WATCHMAN group were driven by early procedure-related complications, with relatively few events occurring in follow-up. On the other hand, adverse events continued to gradually acrue in the warfarin arm, driven primarily by warfarin-related bleeding. Despite the generally positive reception for PROTECT-AF, concerns still remain regarding periprocedural complications and thrombus formation on the device prior to endothelialization (Figure 8.4). Of note, procedure-related complications were greater in the first half of PROTECT-AF than in the second half, underscoring the learning curve involved with device implantation; adverse events continued to remain low in the Continued Access Protocol (CAP) registry of 460 patients.37 A second randomized trial of WATCHMAN versus warfarin, called PREVAIL, sought to address concerns about the high adverse-event rate from WATCHMAN implantation. The preliminary data appear promising and are currently under peer review. Another registry (Continued Access to PREVAIL) has also been created to generate more safety and efficacy data. In late 2013, the accumulated WATCHMAN data was compelling enough for an FDA advisory panel to vote strongly in favor of the device when asked if its benefits outweigh its risks, likely paving the way for eventual FDA approval. Chapte r 8 LAA Excisio n, Lig atio n, and Occlusio n De vice s Fig u r e 8 .4 Transesophageal echocardiographic image of a thrombus (arrow) on a WATCHMAN device several months after anticoagulation was discontinued. At present, the WATCHMAN device is the only LAA exclusion device with demonstrated noninferiority to warfarin for stroke prevention. There is also evidence that patients achieve improvement in quality-of-life measures after WATCHMAN implantation, likely due to discontinuation of daily warfarin, reduction in bleeding complications, and elimination of dietary and drug interactions.38 117 118 Se ct io n 1: At ria l Fib rilla t io n : Ba ckg ro u n d , Eva lu a t io n , a n d Ma n a g e m e n t AMPLATZER Cardiac Plug After the success of the AMPLATZER Septal Occluder (St. Jude Medical, Plymouth, MN) for patent foramen ovale and atrial septal defect closure, the product was redesigned specifically for the LAA and named the AMPLATZER Cardiac Plug (ACP; St. Jude Medical) (Figure 8.5). This device consists of a self-expanding nitinol mesh constructed in two parts: a distal lobe designed to prevent device migration and a proximal disk designed to occlude the LAA ostium. The lobe and disk are joined by an articulating waist that accommodates anatomic variation. The ACP is also delivered transseptally to the LAA. Three published registries summarize the worldwide data on the ACP. The initial human experience in Europe demonstrated a 96% implant success rate in 137 patients, with serious complications in 10 patients (including 3 ischemic strokes, 5 pericardial effusions, and 2 device embolizations).39 The Asian-Pacific experience, although consisting of only 20 patients, provided one-year follow-up data demonstrating no incidence of stroke or death.40 Finally, a Canadian registry of 52 patients achieved procedural success in all but one patient.41 Of note, the Canadian patients all had contraindications to anticoagulation. Two serious complications occurred (one device embolization and one cardiac tamponade). TEE at 6 months showed a disappointing 16% rate of peri-device leak, but 20-month follow-up demonstrated no incidence of device-related death or thromboembolism. Importantly, ACP implantation protocols have generally not involved periprocedural anticoagulation, instead employing dual antiplatelet therapy for one month followed by aspirin monothereapy. Concerns remain about the incidence of persistent leaks following device implantation. While achieving CE mark approval in Europe, the ACP is still in Phase I clinical trials in the United States. LARIAT Suture Delivery System Receiving FDA approval in 2009 for soft tissue approximation, the LARIAT suture delivery system (SentreHEART, Palo Alto, CA) is the newest LAA exclusion device. This hybrid system involves both epicardial and transseptal access. Epicardial and endocardial magnet-tipped guidewires meet at the tip of the LAA, forming a single rail for the delivery of an epicardial snare with a pre-tied suture loop. A balloon catheter serves as a marker for the LAA base and stabilizes the epicardial snare (Figure 8.6). Under fluoroscopic and TEE guidance, the suture is tightened around the LAA base and released from the snare. Importantly, LAA closure can be evaluated in real-time with TEE or left atrial angiography. If closure is not satisfactory, the snare can be repositioned prior to irreversible suture release (Figure 8.7).

Atrial fibrillation - A multidisciplinary approach to improving patient outcomes: Part 2

Nội dung