Current guidelines, based on several randomized controlled trials, suggest to treat patient with AF, elevated stroke risk, and bleeding risk not prohibitive, with chronic anticoagulation – traditionally with warfarin and, more recently, with direct thrombin and factor Xa inhibitors. Although warfarin is highly effective, it requires a frequent INR monitoring and dose adjustment due to its narrow therapeutic range. As a result, bleeding risk and variable compliance result into the exclusion of 14–44 % of treatable patients from oral anticoagulant therapy [7]. Moreover, only 55 % of treated and well-adherent patients remain in therapeutic range during warfarin therapy [8].

Novel oral anticoagulants (NOAC) do not require INR monitoring and therefore are associated with a priori higher patient compliance. Furthermore, they have a favorable stroke/bleeding risk profile, driven by reductions in the rates of stroke and intracranial hemorrhage [9]. However, higher risk of GI bleeding observed with NOAC counterbalances the overall risk of bleeding, resulting to similarity between warfarin and novel molecules.

In terms of patient compliance, randomized controlled trials comparing NOAC vs. warfarin, such as the RELY (dabigatran vs. warfarin) [10], ROCKET-AF (rivaroxaban vs. warfarin) [11], ARISTOTLE (apixaban vs. warfarin) [12], and ENGAGE AF-TIMI 48 (edoxaban vs. warfarin) [13] study, NOAC interruption, owing to side effects or low compliance, occurred in 10 %, 24 %, 25 %, and 34 %, respectively, during the study period. Therefore, a sizable proportion of patients still remain at high risk of stroke because of undertreatment. Moreover, the increased bleeding risk, especially among older patients and/or those who are on antiplatelet medications, and the required lifelong pill compliance prompted the development of alternative strategies for stroke prevention.

Percutaneous LAA occlusion is a novel therapy to prevent stroke in patient with atrial fibrillation. The LAA, due to its trabeculae and blood stasis, is the most common site of thrombus generation in patients with AF [14]. Surgical exclusion of LAA has been previously used to reduce the risk of embolic events in patients with AF [15]. Percutaneous left atrial appendage occlusion (LAA) has been developed as a less invasive procedure to prevent stroke in patient with AF who cannot tolerate oral anticoagulation [9, 10]. This catheter-based technique provides an atrial endoluminal mechanical orifice obstruction. In 2002, the first case of percutaneous LAA closure was reported with the use of the PLAATO system (percutaneous left atrial appendage), a nitinol self-expandable cage with lateral hook, covered by ePTFE membrane. This device was implanted in few non-randomized series which showed good clinical results in terms of embolic event reduction: at 5 years stroke rate was 3.3 % compared to an expected rate of 6.6 % based on CHADS₂ score [16]. However, it was recalled from the market due to its complex and risky implantation technique, which resulted in high rate of procedural complication such as vessel perforation, cardiac tamponade, and device embolization [16–21].

Currently, three percutaneous LAA closure devices are available on the market: the Watchman™ (Boston Scientific, Natick, MA, USA), the Amplatzer™ Cardiac Plug (St. Jude Medical, Minneapolis, MN, USA), and WaveCrest.

The Watchman is a nitinol cage partially covered by a polyethylene terephthalate (PTFE) membrane, with lateral small barbs which anchor the device to LAA minimizing the risk of dislodgment. It is the only device studied in a randomized fashion. The Watchman Left Atrial Appendage System for Embolic Protection in Patients with Atrial Fibrillation (PROTEC-AF) trial was the first randomized study to investigate a LAA occlusion device. It included 707 patients with non-valvular AF and at least 1 CHADS₂ risk factor, who were randomized 2:1 to LAA closure with the use of the Watchman device or warfarin [22]. This study found that LAA closure was non-inferior to warfarin with respect to the combined risk of death, stroke, or embolism (8.4 % vs. 13.9 %) and the single risk of overall (12.3 % vs. 18.5) and cardiovascular death (3.7 % vs. 9.0 %) at early- and long-term follow-up [23].

The Prospective Randomized Evaluation of the Watchman LAA Closure Device in Patients with Atrial Fibrillation Versus Long-Term Warfarin Therapy (PREVAIL) trial was a confirmatory randomized study which demonstrates an improved safety compared to PROTECT-AF trial [24]. The most comprehensive study on Watchman was a recent individual patient-level meta-analysis of 2,406 patients including two randomized and two non-randomized trial (PROTECT-AF, PREVAIL, and their respective registries) with a mean follow-up of 2.7 years. The analysis found that the LAA occlusion with Watchman device provides similar benefit to warfarin for the composite efficacy end point of stroke, embolism, or cardiovascular death. In terms of single end points, the device provides an advantage for cardiovascular death (1.1 % vs. 2.3 % per year, hazard ration 0.48, p = 0.006) and bleeding events (hemorrhagic strokes 0.15 % vs. 0.96 % per year, hazard ratio 0.22; p = 0.004; and non-procedural bleeding 6.0 % vs. 11.3 %, hazard ratio 0.51; p = 0.006) compared with warfarin. This is balanced by a slight increase in ischemic stroke rate in device-treated group (1.6 % vs. 0.9 % per year, hazard ratio 1.95, p = 0.05) [25].

The Amplatzer Cardiac Plug – Amulet – is a second-generation device made of a nitinol lobe and disk connected by a short flexible waist, laid of polyester fabric. Such device is based on pacified principle: the main lobe is placed into the LAA body, and the disk covers the LAA orifice. The safety and efficacy were reported in few registries, which demonstrate good clinical performance, at least comparable to PROTECT-AF results. The procedural success varied from 95 to 100 %; embolization and pericardial effusion were 1–2 %. At the follow-up, the death rate was 6–9 % and stroke rate was 2–3 % [3, 26–29].

The Coherex WaveCrest (Coherex Medical Inc., Salt Lake City, UT, USA) device is made by a nitinol structure without exposed metal; it has a foam layer facing the LAA to promote the rapid endothelialization and PTFE layer facing the LAA to reduce thrombus formation. The device can be delivered and then fixed by actionable lateral hooks once the correct position is obtained. CE Marking was granted in 2013, and the device is now available in Europe. Currently, no clinical data have been published, but a phase 2 clinical study has been recently completed (clinicalTrials.gov Identifier: NCT02239887).

The indication for LAA occlusion is summarized in Fig. 15.1. Based on European Society of Cardiology guidelines, the LAA occlusion is recommended (class IIb with level of evidence B) when high risk of stroke is present (CHA₂DS₂-VASc more than two or CHADS₂ more than one), and oral anticoagulation is not possible, as well as previous life-threatening bleeding events or high risk of bleeding under oral anticoagulation (increased HAS-BLED score, patient with coronary artery disease requiring prolonged dual antiplatelet therapy plus oral anticoagulation, or patient with end-stage renal failure) [6].