Atrial fibrillation (AF) is the most common cardiac arrhythmia in adults with a current prevalence estimated at 1.5% to 2% of the general population. With the aging population in the United States, AF prevalence is projected to increase steadily from approximately 6 million cases in 2010 to 15.9 million by 2050.¹ AF is a major cause of stroke, being responsible for 15% of all strokes and 30% of strokes in patients over age 80.² Unfortunately, stroke is the leading cause of long-term disability and the fourth leading cause of death in the United States.^3,4 The presence of AF is associated with a 4- to 5-fold risk of ischemic stroke,⁵ and the incidence increases significantly with advancing age. Moreover, strokes associated with AF are more severe; AF-related stroke victims have a 50% greater likelihood of becoming disabled or handicapped and a >50% likelihood of death.^6,7 Accordingly, stroke prevention with anticoagulation is one of the main pillars of AF management, and guidelines for anticoagulation have become more stringent recently. The Canadian Cardiovascular Society had lowered their threshold for recommending oral anticoagulation (OAC) for CHADS₂ (congestive heart failure, hypertension, age 75 years, diabetes mellitus, stroke) score ≥1, the European Society of Cardiology (ESC) for CHA₂DS₂-VASc (congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, stroke/transient ischemic attack, vascular disease, age 65-74 years, sex category) score ≥1, and the American College of Cardiology (ACC) for CHA₂DS₂-VASc score ≥2.^1,8-10

Anticoagulation for AF and Limitations

Several randomized placebo-controlled trials have demonstrated that OAC is highly effective in preventing thromboembolism with AF, and landmark meta-analysis with warfarin demonstrated a 64% relative stroke reduction and 26% relative mortality reduction.^11,12 Although OAC is the current gold standard for AF stroke prevention, a significant proportion (30%-50%) of eligible patients do not receive therapy due to absolute contraindications or perceived risks of bleeding.^13,14 Furthermore, OAC increases the incidence of major bleeding despite the use of novel OAC (NOAC); a recent meta-analysis of all 4 NOAC trials showed major bleeding rates of 6.2% with warfarin and 5.3% with NOAC.¹⁵ Even though the risk of intracranial hemorrhage is consistently lower with NOAC, overall major bleeding is not diminished with dabigatran or rivaroxaban compared with warfarin.^16,17 Apixaban and edoxaban were the only 2 agents that lowered major bleeding compared to warfarin.^18,19

There are also other concerns with OAC therapy, including patients with renal and liver dysfunction (excluding use with NOAC), high risk of falls, noncompliance, and patients requiring dual antiplatelet therapy after stent placement. For warfarin, there are additional issues with drug and diet interaction, the need for anticoagulation monitoring, and narrow therapeutic window, with time in therapeutic range of only 50% to 60% in community practices.^20,21 There are also high discontinuation rates with OAC even in the setting of clinical trials, with discontinuation rates of 16% to 34% for warfarin and 21% to 37% with NOAC at the end of study follow-up.^16-19

Therefore, even though OAC is effective for thromboembolic prevention, there remains a large proportion of eligible patients not on therapy for a multitude of reasons. Moreover, there is a residual stroke risk of 2% to 5% annually despite optimal OAC.²² These challenges have led to the investigations of device-based therapies for nonvalvular AF. In fact, the ESC recently gave a Class IIB recommendation for percutaneous left atrial appendage (LAA) closure for patients with high stroke risk and contraindications to long-term OAC.⁹

Rationale for LAA Occlusion

In the presence of AF, the left atrium loses active contraction, leading to decreased blood velocity in the LAA, which results in stasis and increased risk of thrombus formation.²³ Abnormalities in the thrombosis cascade in overloaded LAA may additionally increase the risk of thrombus formation with AF.²⁴ Transesophageal echocardiography (TEE), autopsy, and surgical reports confirmed that >90% of nonrheumatic AF-related left atrial thrombi were isolated to or originated from the LAA.²⁵ Atrial thrombi are detected on TEE in approximately 15% of AF patients not receiving long-term OAC.²⁶ The presence of LA thrombus, spontaneous echo contrast, and low LAA velocities (≤20 cm/s) are independent predictors of stroke and thromboembolism on TEE.^27-29 Thus, local mechanical approaches to exclude the LAA from systemic circulation to limit embolization in nonvalvular AF patients have been developed. Early attempts of surgical removal or ligation of LAA developed over 60 years ago were limited by the invasiveness of major cardiac surgery and by significant rates of incomplete exclusion that were associated with increased risks (~2.5-fold) of stroke.^30,31 Minimally invasive percutaneous approaches have since been developed over the past 2 decades and can be broadly divided into endocardial and epicardial devices (Table 47-1). This chapter reviews contemporary percutaneous LAA closure, with in-depth discussions of the available leading devices with regard to procedural techniques and clinical outcomes.

The current most commonly adopted indication for percutaneous LAA closure is for patients ineligible for long-term OAC, although this practice varies geographically. In Europe, the majority of LAA closures are done according to the ESC recommendation for patients with high stroke risk and contraindications to long-term OAC.⁹ However, there are high-volume centers (eg, in Switzerland and Germany) where LAA closures may be performed for anticoagulation-eligible patients, and this is likely related to the full reimbursement of this procedure in these countries. In Canada, LAA closure is generally restricted to patients with high stroke risk and contraindications to long-term OAC. In the United States, LARIAT (SentreHEART, Redwood City, CA) and WATCHMAN [Boston Scientific, Natick, MA] received approval by the US Food and Drug Administration [FDA]) and may be performed for patients with high stroke risk with or without contraindications to OAC.

WATCHMAN

The WATCHMAN device was the second dedicated LAA device manufactured after PLAATO (Appriva Medical, Sunnyvale, CA), which was the first device manufactured but was subsequently removed from the market for commercial reasons. WATCHMAN was originally designed and manufactured by Atritech (Plymouth, MN) and was acquired by Boston Scientific in 2011. The current second-generation WATCHMAN LAA Closure Technology is constructed of a self-expanding nitinol frame covered by a permeable 160-μm polyethylene terephthalate (PET) membrane (Fig. 47-1). There are 10 fixation anchors at the perimeter of the nitinol frame that are designed to engage the LAA tissue for device stability. The PET membrane covers approximately 50% of the proximal outer aspect of the nitinol frame, which blocks thrombus embolization from the LAA and promotes endothelialization. The spherical contour of WATCHMAN can accommodate most LAA anatomy. There are 5 device sizes available (21, 24, 27, 30, and 33 mm), and all are delivered through dedicated 14-Fr sheaths (double curve, single curve, and anterior curve). The fourth-generation device was evaluated in a European registry, and the fifth-generation (WATCHMAN-FLEX) device is awaiting first-in-man evaluation. The WATCHMAN received Conformité Européene (CE) mark in 2005 and FDA approval in March 2015. To date, >10,000 WATCHMAN devices have been implanted worldwide.

Amplatzer Cardiac Plug and Amulet

The Amplatzer Cardiac Plug (ACP) (St. Jude Medical, Plymouth, MN) is the third dedicated LAA device to be manufactured and is specifically designed to occlude the proximal segment of the LAA. In Europe, operators initially used nondedicated Amplatzer devices for LAA closures in a small series of patients.^32,33 However, the incidence of embolization was high (12% in the 32-patient series in Bern), and these nondedicated devices are no longer used for this indication.

This device consists of a self-expanding nitinol mesh, which forms a lobe and a disc connected by a central articulating waist (Fig. 47-2). The lobe is intended to be implanted at 10 mm inside the orifice (proximal neck of LAA) and, along with the 6 pairs of stabilizing wires at the distal lobe, serves as the key anchoring mechanism. The disc is intended to be deployed in the left atrium and pulled in under traction against the LAA orifice by the central waist, which helps seal the orifice. Both the lobe and disc have polyester mesh sewn in manually. The ACP comes in 8 different sizes according to the lobe dimension accommodating LAA diameters of 12.6 to 28.5 mm (Fig. 47-3). The second-generation ACP device, Amulet, is similar in design but has a wider lobe, longer waist, recessed proximal end screw, and more stabilizing wires. These features improve device stability and may theoretically reduce the risk of thrombus formation on the atrial side of the device. Amulet comes in 8 different sizes and can accommodate larger LAA (up to 32 mm diameter). The ACP device has to be manually loaded onto the delivery cable, but Amulet comes preloaded on the delivery cable for ease of setup. ACP/Amulet is delivered through the TorqueVue 45 × 45 sheath (see Fig. 47-2), which has a 3-dimensional distal tip allowing anterior and superior angulation for coaxial positioning at the landing zone. The TVLA1 and TVLA2 sheaths are no longer being manufactured. The access sheath size varies according to device size (see Fig. 47-3).

The ACP received CE mark approval in December 2008, and the Amulet was approved in January 2013. The delivery system of Amulet has been redesigned, and the updated system will be relaunched in the latter half of 2014. To date, >10,000 ACP/Amulet devices have been implanted worldwide.

LARIAT

The LARIAT LAA closure system requires both an endocardial and epicardial approach for implantation. LARIAT has FDA and CE mark approvals for suture and knot tying during surgical applications but is not approved specifically for stroke reduction with AF. There was a prior surge in interest and procedural volume in the United States before the approval of WATCHMAN, due to the availability of this device for patients who were not candidates for anticoagulation. LARIAT consists of a snare with a pretied suture that is magnetically guided over the LAA in the epicardium. The device consists of 3 components: (1) 15-mm compliant occlusion balloon catheter (EndoCATH), (2) 0.025- and 0.035-inch magnet-tipped guide wires (FindrWIRZ), and (3) 12-Fr LARIAT suture delivery device.³⁴

WaveCrest

The WaveCrest occluder by Coherex Medical (Salt Lake City, UT) is made of a nitinol frame with retractable anchors to enable optimal device positioning. The occluder is covered by a polytetrafluoroethylene (PTFE) membrane on the left atrial side of the device and a foam substrate on the inside surface to minimize peridevice leak. In contrast to other devices, the anchoring system can be operated independently of the lobe, allowing repositioning before anchoring. Contrast can be injected through the delivery sheath or on the appendage side of the occluder to evaluate occlusion. It comes in 3 sizes (22, 27, and 32 mm) and can be implanted through 4 different 15-Fr delivery sheaths. The device was first implanted in June 2012, and CE mark was obtained in 2013.

Preprocedural Imaging

Baseline TEE is important to exclude preexisting LAA thrombus and to assess LAA anatomy for device closure suitability. A full 0° to 180° sweep of the LAA is useful to appreciate the anatomy and for accurate widest and deepest dimension measurements. For WATCHMAN, the widest LAA ostium (from the level of the circumflex artery to ~2 cm within the pulmonary vein ridge) is measured at 0°, 45°, 90°, and 135°, and the available depth (from ostium to apex of LAA) is also measured at the same angles (Fig. 47-4). For ACP, measurements at both the short-axis (30-60°) and long-axis (120-150°) angles of the landing zone and orifice are performed. The LAA orifice for ACP/Amulet is measured from the pulmonary vein ridge to the circumflex artery. The landing zone is measured at 10 mm inside the orifice for ACP (~13-18 mm for Amulet), at an angle that is perpendicular to the neck axis. The LAA depth is measured from the orifice to the back wall along the axis of the neck. The LAA measurements are usually wider at the long axis (corresponding to caudal projection on fluoroscopy) compared to the short axis (corresponding to right anterior oblique [RAO] cranial).

Preprocedural cardiac computed tomography angiography (CCTA) is also useful given its superior spatial resolution and 3-dimensionality as a noninvasive alternative to or in addition to TEE. CCTA has excellent sensitivity for ruling out LAA thrombus. Although the specificity and negative predictive value for LAA thrombus have traditionally been suboptimal with CCTA due to difficulty differentiating thrombus from inadequate contrast mixing (eg, spontaneous echo contrast), additional protocols using double contrast injections, delayed imaging, dual energy, and prone positioning have optimized the results.³⁵ CCTA also offers superior anatomic visualization and measurements of the LAA compared to TEE. Patients do not need to be fasting prior to CCTA; furthermore, saline infusion is typically administered before scans, providing more accurate measurements at euvolemic states. Thus, CCTA may become a preferred alternative to TEE in experienced CCTA centers and is increasingly being performed routinely prior to LAA closures.

Procedural Imaging

Procedural TEE is the routine standard in the majority of centers and is typically accompanied by general anesthesia. There are a few centers adept at procedural intracardiac echocardiography (ICE) to guide LAA closure, obviating the need for general anesthesia. However, obtaining adequate ICE LAA images can be challenging, and proceduralists may overcome this problem by advancing the ICE probe into the left atrium through another transseptal puncture. Although a few centers rely on fluoroscopy alone during LAA closure, this is not advised for the average operator.

Transseptal Puncture

Venous access is preferred through the right femoral vein for more direct transseptal access. Access site subcutaneous tissues should be well prepared and separated with scalpel and forceps to ease advancement of large 13- to 14-Fr sheaths. Manual compression, “figure-of-8” suture, and preclosing with 6-Fr Perclose (Abbott Vascular, Chicago, IL) are commonly used for venous hemostasis.