Heart, Lung and Circulation (2014) 23, 407–413 1443-9506/04/$36.00 http://dx.doi.org/10.1016/j.hlc.2013.12.008
Percutaneous Transcatheter Left Atrial Appendage Closure Devices: Role in the Long-Term Management of Atrial Fibrillation Chieh Howe (Gary) Gan, MBBS, BMedSc*, Aditya Bhat, MBBS, BMedSc, Lloyd Davis, PhD, FRACP, FCSANZ, A. Robert Denniss, MD, MSc, FRACP, FCSANZ Department of Cardiology, Blacktown and Mount Druitt Hospital, Blacktown Road, Blacktown, Sydney NSW 2148, Australia Received 17 October 2013; received in revised form 23 November 2013; accepted 23 December 2013; online published-ahead-of-print 22 January 2014
Atrial fibrillation remains the most common cardiac rhythmic disorder worldwide and is associated with significant health hazards, most notably an increase in the rate of cerebrovascular events. Stroke prevention in atrial fibrillation has traditionally been managed with warfarin therapy which is encumbered by risk of bleeding, drug administration logistics and interactions as well as issues of non-compliance. Occlusion of the left atrial appendage has recently been explored as an alternative method of stroke prevention. The aim of this article is to evaluate the history, efficacy and draw-backs associated with percutaneous left atrial appendage occlusion devices in the management of atrial fibrillation.
The current literature and clinical experience was used to summarise the history and evaluate the efficacy of percutaneous left atrial appendage occlusion devices.
Percutaneous left atrial appendage occlusion devices are effective novel therapies for stroke prevention in atrial fibrillation, with proven reductions in thromboembolic events in comparison with placebo and noninferiority with warfarin therapy. Pericardial effusions and embolic strokes are primary peri-procedural adverse reactions. The rates of adverse reactions reduce with operator experience.
Percutaneous left atrial appendage occlusion is an exciting and novel therapy of stroke prevention in atrial fibrillation. Whilst further trials and long-term data are required prior to widespread implementation of this procedure, trials so far have highlighted the clinical efficacy of the procedure.
Atrial Appendage Atrial Fibrillation Stroke Warfarin Thrombosis
Atrial fibrillation: The looming epidemic Atrial fibrillation remains the most common cardiac rhythmic disorder worldwide, with an expected exponential increase in incidence in our aging population.  The condition has associated health hazards, with an increase in rates of ischaemic cerebrovascular accidents (CVAs), congestive
cardiac failure and ischaemic myocardial events. [2,3] In addition to the health burden of the disease, there is an associated economic reality to the condition, with an estimated cost of $26.0 billion US dollars per year. [1,4] Stroke prevention in atrial fibrillation has traditionally been based on oral anticoagulant therapy, primarily with the Vitamin K Antagonists (VKA). This standard of therapy was derived from many trials conducted within the last two
*Corresponding author. 118/3, Stromboli Strait, Wentworth Point, NSW 2128, AUSTRALIA Tel.: +040 272 9772., Emails: [email protected]
, [email protected]
, [email protected]
, [email protected]
© 2014 Australian and New Zealand Society of Cardiac and Thoracic Surgeons (ANZSCTS) and the Cardiac Society of Australia and New Zealand (CSANZ). Published by Elsevier Inc. All rights reserved.
decades which compared the efficacy of warfarin therapy against placebo and oral anti-platelet agents.  In the meta-analyses by Hart et al. comparing 29 of these trials, warfarin therapy reduced the rate of CVAs by 65% in comparison with placebo, and by 40% relative to oral antiplatelet agents.  Despite the caveat of increased absolute risk of intra-cerebral haemorrhage and major bleeding events by 0.2% per year, warfarin therapy has been shown to be associated with a statistically significant reduction in overall mortality in comparison to placebo, highlighting its clinical potency.  Despite this clear advantage however, a significant proportion of patients with atrial fibrillation do not receive warfarin therapy. [8,9] Whilst this is largely medically driven due to the risk of bleeding, a considerable extent can also be attributed to the logistics of warfarin itself. The drug, with a multitude of food and drug interactions, and the need for frequent laboratory monitoring,  can be best described as ‘cumbersome’. Furthermore, even for patients identified as ‘suitable’, studies involving controlled cohorts on warfarin
Figure 1 The Left Atrial Appendage.
C.H. (Gary) Gan et al.
therapy have shown that patients remain in therapeutic range only 50-70% of monitored days.  Precipitated by these drawbacks with warfarin therapy, newer novel anticoagulants have come to the forefront of clinical research. Agents such as Dabigatran, a direct thrombin inhibitor, and Rivaroxaban and Apixaban, Factor Xa inhibitors, have become the hope and focus of many physicians worldwide as a potential replacement for warfarin therapy. Early studies comparing these novel anticoagulants to warfarin have shown non-inferiority with regards to prevention of thrombo-embolism. Unfortunately, as with warfarin therapy, however, the risk of major bleeding with these new agents remains prominent, [11–14] heralding interest in alternative device therapies for stroke prevention in atrial fibrillation.
The left atrial appendage A trabeculated remnant of the embryonic left atrium, the left atrial appendage (LAA) is a vestigial structure located in the postero-lateral aspect of the left atrium (Figure 1). It has a
Percutaneous Transcatheter Left Atrial Appendage Closure
finger-like appearance, averaging 2-4 cm in length and contracts rhythmically with its smooth-walled counterpart, the left atrium, during sinus rhythm. The LAA was first implicated in the pathogenesis of strokes in the 1940s, whereby its excision was explored as a form of therapy for recurrent arterial emboli. [15,16] With the advent of warfarin in the 1950’s, however, the hypothesis was left relatively untested for subsequent decades until the late 1990’s, whereby renewed interest was sparked by the observation that the LAA was the site of the majority (>90%) of thrombi in patients with atrial fibrillation undergoing trans-oesophageal echocardiographic studies.  In atrial fibrillation, it is thought that the loss of atrial contraction results in atrial stretch and dilatation that, in turn, precipitates stasis and increased probability of thrombus formation [17,18] This pathophysiological precept is supported by the findings of the Stroke Prevention in Atrial Fibrillation (SPAF) III trial whereby reduced LAA peak flow velocity (an early indicator for stasis) was identified as an independent predictor of increased thrombo-embolic risk. 
Percutaneous Transcathether LAA Occlusion - How Far Have We Come? The first technology for percutaneous transcatheter LAA occlusion was the PLAATO (Percutaneous Left Atrial Appendage Transcatheter Occlusion) device.  Manufactured by Appriva Medical Inc. (Sunnyvale, CA), it consisted of a self expanding nitinol cage covered with a polymeric membrane with three rows of anchors along the struts to help stabilise its position within the left atrial appendage (Figure 2). The device is delivered trans-septally under transoesophageal echocardiography and fluoroscopic guidance.  Intravenous heparin infusion is administered intra-procedurally with dual anti-platelet (aspirin 300 mg and clopidogrel 75 mg) cover 48 hours prior to the procedure and for up to four to six weeks post-procedure while the device
Figure 2 PLAATO Device.
endothelialises.  Single anti-platelet therapy is utilised for life following this period. [21,22] The clinical efficacy of the PLAATO device was first studied in a multi-centre observational trial by Ostermayer et al. between August 2001 and November 2003. The study, which focussed on patients with non-valvular atrial fibrillation with contraindications to anti-anticoagulation therapy and high thrombo-embolic risk, looked at the occurrence of major adverse events (defined as new strokes, myocardial infarction, cardiac and/or neurological death as well as requirement for cardiovascular surgery related to the implantation procedure) that occurred within 30 days of implantation of the PLAATO device. [21,22] End points of device performance were that of (i) implantation success, defined as successful delivery and deployment of the PLAATO device into the LAA in the absence of major adverse events and, (ii) treatment success, defined as successful implantation of the device with angiographically-visualised LAA occlusion immediately after placement of the implant.  Of a cohort of 111 participants, successful implantation of the PLAATO device was achieved in 108 subjects. Three patients did not receive the device because of presence of LAA thrombus at the time of the procedure, perforation of the right femoral artery in attempt to access the right femoral vein and haemodynamically significant cardiac tamponade after trans-septal puncture requiring emergent pericardiocentesis and open thoracotomy. Amongst the 108 patients in whom the device was successfully implanted, four developed pericardial effusion 30 days post-implantation, two of whom required inpatient pericardiocentesis due to development of tamponade. One patient suffered a haemothorax 24 hours post-index procedure that was treated with thoracocentesis without any sequaelae. The average follow up duration was 9.8 months, at which point two patients were noted to have experienced strokes despite good positioning and the absence of mobile thrombus on the device during transoesophageal echocardiogram pre- and post-stroke. The cause of the strokes remains uncertain but was thought to be of non-cardiac aetiology. At one- and six-month transoesophageal echocardiography, no device migration or mobile thrombus was noted in all 108 patients. [21,22] Though non-randomised and limited by a small and relatively select patient population, the trial was the first in highlighting the feasibility of percutaneous transcatheter LAA occlusion as an alternative form of stroke prevention in atrial fibrillation. Allowing for the assumption that the patient cohorts were compliant on aspirin, the estimated stroke rate post implantation of the PLAATO device was 2.2% compared to the expected 6.3% prior to device implantation, a relative event reduction of 4.1%.  Coupled with the procedural success rate of 97% in the cohort and estimated serious adverse event rate of 8%, the prospect was appealing. Long term data regarding the PLAATO device was unfortunately scarce as the PLAATO study was regrettably halted during the follow-up phase due to financial considerations. In 2009, however, Block et al. reviewed the
Figure 3 WATCHMAN Device.
five-year follow-up of 64 participants of the PLAATO trial who received the device. In this study, the annualised stroke/transient ischaemic attack rate was 3.8%, substantially lower than anticipated stroke/transient ischaemic attack risk (6.6%) utilising the CHADS (2) scoring system, affirming the efficacy of the PLAATO system.  Following the success of the PLAATO, The WATCHMAN Left Atrial Appendage device (Atritech, Plymouth, MN, USA) was introduced in 2005. Like the PLAATO device before it, the design is based on a self expanding nitinol frame but with fixation barbs and a permeable polyester fabric cover (Figure 3). It is delivered trans-septally under transoesophageal echocardiographic and fluoroscopic guidance with intravenous heparin boluses to achieve an activated clotting time of >250 seconds pre-procedurally and intra-procedurally.  Post-procedurally, warfarin is administered until device endothelialisation is achieved and stable positioning of device is confirmed on transoesophageal echocardiography. Warfarin is then substituted for clopidogrel, and subsequently aspirin after a period of 4.5 to 6 months. [24,25] The PROTECT-AF (WATCHMAN Left Atrial Appendage System for Embolic Protection in Patients with Atrial Fibrillation) trial was performed between February 2005 and June 2008 assessing the clinical efficacy of the WATCHMAN device. A prospective, multi-centre, non-inferiority randomised controlled trial, The PROTECT-AF study looked at the occurrence of stroke, cardiovascular or unexplained death and systemic emboli as the primary end-point. Primary safety end-points were that of excessive bleeding, strokes and peri-procedural adverse events.  From an initial sample size of 4998 participants, 707 were found to be eligible and were randomised in a 2:1 ratio to percutaneous LAA closure with subsequent discontinuation of warfarin therapy (463 subjects) or long-term warfarin therapy with a target international normalised ratio [INR] between 2.0 to 3.0 (244 subjects). Patients in the intervention arm were followed up with transoesophageal echocardiography at 45 days, six months and one year post-implantation. Warfarin therapy was continued for at least 45 days until confirmation of device stability on transoesophageal echocardiogram whereby it was substituted for clopidogrel for a
C.H. (Gary) Gan et al.
period of six months with subsequent aspirin therapy indefinitely. Patients in the control group were commenced on warfarin therapy and received INR monitoring no less than every two weeks for the first six months and monthly thereafter.  Of the 463 patients in the percutaneous intervention cohort, successful device implantation was achieved in 408 (88%). Fourteen patients declined implantation of the device and 41 patients had unsuccessful attempts at implantation. Pericardial effusion was the most common peri-procedural adverse event, occurring in a total of 22 cases (5%). Device embolisation occurred in three patients and peri-procedural ischaemic strokes occurred in five patients (0.9%). There were no fatalities due to pericardial effusions; 15 patients were treated with pericardiocentesis and seven with surgical drainage without any complications. Of the five patients with peri-procedural strokes, three had no residual neurological deficits whilst two were discharged to high level care facilities where they died. A vascular snare was used to retrieve the device in one patient with device embolisation while the remaining two required surgery for device retrieval. At 45 days, warfarin therapy was discontinued in 349 of the 408 patients (86%). The primary reason for continuation of warfarin in the remaining patients was due to device leak detected by transoesophageal echocardiography or on advice by the treating physician. Within the control group, all patients, bar three, received warfarin, and the time in therapeutic INR range was 66% based on the Rosendaal method. [25,26] Overall, the investigators of the PROTECT-AF study found non-inferiority of LAA occlusion in comparison with warfarin therapy. After a median 18-month follow-up period, the primary event rate (composite of stroke, systemic embolism and cardiovascular death) was found to be 3.0% per year in the percutaneous group compared to 4.3% per year in the control group. This was consistent across subgroups (distinguished by sex, age, pattern of atrial fibrillation and morphology of the LAA) and was driven by a 94% lower incidence of haemorrhagic stroke and a 74% reduction in cardiovascular mortality. [25,26] Unsurprisingly, there was a higher incidence of primary safety events (major bleeding, pericardial effusion and device embolisation) in the interventional group, but this appeared to decrease with operator experience as indicated by the analyses of the non-randomised Continued Access Protocol (CAP) registry. Data from the registry, which included 460 patients after the completion of the PROTECT-AF study, noted a significant improvement in the incidence of peri-procedural complications (2.2% rate of pericardial effusion, 0% peri-procedural stroke) when implantation of the device was performed by experienced operators.  The clinical efficacy of the WATCHMAN device was further supported by sub-studies of the PROTECT-AF trial which highlighted the absence of any significant interaction between residual peri-device flow postimplantation and clinical outcome.  Iatrogenic atrial septal defects frequently observed after transeptal puncture were also noted to have a very high spontaneous closure
Percutaneous Transcatheter Left Atrial Appendage Closure
rate and were not associated with an increased rate of stroke or systemic embolisation in the long term.  Despite promising results of the PROTECT AF and CAP registry, concerns with the early safety results of these trials remained high, prompting a second randomised trial of the WATCHMAN device to confirm safety and efficacy profiles. The trial, entitled Evaluation of the WATCHMAN LAA Closure Device in Patients With Atrial Fibrillation Versus Long Term Warfarin Therapy (PREVAIL), enrolled 407 patients from 41 centres in the United States, with randomisation of participants in a 2:1 ratio of LAA occlusive device implantation versus control (warfarin). In fashion with the PROTECT-AF study, patients receiving device implantation received warfarin therapy for at least 45 days until confirmation of device stability on transoesphageal echocardiogram.  Preliminary data indicate that the device had a 95.1% implant success rate, an improvement from the PROTECTAF trial (91%) The primary safety end point [acute (sevenday) occurrence of death, ischaemic stroke, systemic embolism and procedure, or device-related complications requiring major cardiovascular or endovascular intervention] occurred in 2.2% of participants (6 out of 269). This was considered a success, with the upper confidence interval (2.62%) being lower than the pre-specified criterion for success (2.67%). A second, broader, safety end point [rates of cardiac perforation, pericardial effusion with tamponade, ischaemic stroke, device embolisation, and other vascular complications] occurred in 4.4% of patients with device implantation in the PREVAIL trial, compared with 8.7% in PROTECT-AF. No procedure-related deaths were noted.  In terms of efficacy end points, the primary end point comparing composite of stroke, systemic embolism, and cardiovascular/unexplained death in the PREVAIL STUDY did not demonstrate non-inferiority relative to composite data from PROTECT-AF, with the upper 95% confidence limit higher than allowed to meet the success criterion (95% CI upper bound