In a pilot study, successful implantation of the Watchman device was observed in 66 of 75 patients. Due to device migration after placement, the original device and delivery system have been modified [15]. The PROTECT AF trial was the first randomised controlled trial testing LAA closure by the Watchman device versus long-term warfarin therapy [14]. 707 patients with non-valvular AF from 59 centres worldwide were randomised (2:1) to receive the Watchman device or control treatment (warfarin). Patients in the device group took warfarin for 45 days to allow endothelialisation of the device. Warfarin was discontinued in case of complete closure or significantly decreased flow around the device as assessed by transoesophageal echocardiography. Patients in the device group were then treated with aspirin and clopidogrel for 6 months followed by aspirin indefinitely. The main results of the study were published in 2009. During a median follow-up of 18 months, the device group proved to be non-inferior with respect to occurrence of stroke, systemic thromboembolism and cardiovascular mortality [5]. Later on the 4-year results were published which even showed superiority of the device group with respect to the composite primary endpoint as well as all-cause mortality [14].

A major concern of the PROTECT AF trial was the substantial number of safety events. Serious pericardial effusions occurred in 22 of the 463 patients randomised to LAA occlusion. Of these 7 underwent surgical intervention. Consequently, the PREVAIL trial (Prospective Randomized Evaluation of the Watchman LAA Closure Device in Patients With Atrial Fibrillation versus long-Term Warfarin Therapy) was designed to address efficacy and safety of the Watchman device compared with long-term warfarin therapy [6]. Early safety events (defined as 7-day occurrence of death, ischaemic stroke, systemic embolism, procedure- or device-related complications requiring major cardiovascular or endovascular intervention) occurred in 2.2 % of patients in the Watchman arm, significantly lower than in early results of the PROTECT AF trial.

The majority of data about LAA occlusion have compared devices versus warfarin therapy. However, the patient population which would benefit the most of an LAA occluder has not been studied extensively: patients with contraindications to long-term oral anticoagulation. The ASA Plavix Feasibility Study (ASAP) focused on patients considered ineligible for warfarin to determine whether implantation of a Watchman device could be safely performed without a transition period with warfarin [13]. After device implantation, patients were treated with clopidogrel for 6 months and ASA indefinitely. Of the 150 enrolled patients which were followed up for a mean of 14.4 months, three of four strokes that occurred were ischaemic (1.7 % per year) reflecting a 77 % reduction from the expected rate of 7.3 % based on the respective CHADS₂ scores of the patient cohort.

A recent meta-analysis included 2406 patients (5931 patient years of follow-up) from the PROTECT AF and PREVAIL trials and their respective registries (Continued Access to PROTECT AF registry and Continued Access to PREVAIL registry) [4]. In patients with non-valvular AF at increased risk for stroke and bleeding, LAA closure with the Watchman device resulted in improved rates of haemorrhagic stroke (hazard ratio [HR]: 0.22; p = 0.004), cardiovascular/unexplained death (HR: 0.48; p = 0.006) and non-procedural bleeding (HR: 0.51; p = 0.006). While rates of all-cause stroke or systemic embolism (SE) were similar between the device and control group, more ischaemic strokes or SE could be observed in the device group (HR: 1.95; p = 0.05).

The procedural feasibility and safety of ACP implantation was initially evaluated in a retrospective pre-registry data collection [12]. Despite a 96 % closure rate (137 of 143 patients), serious complications were observed in ten patients (three ischemic strokes, two device embolisms, five pericardial effusions). Urena and colleagues reported a 98.1 % successful implantation rate in 52 patients who did not tolerate or desire long-term anticoagulation. There was one stroke (1.9 %), one transient ischaemic attack (1.9 %) and one major bleeding event (1.9 %) [16]. A comparison of safety and efficacy of first- versus second-generation Amplatzer occluders showed similarly high success rates. However, Amulet (second generation) was not superior regarding the combined endpoint death, stroke, cardiac tamponade and bail out by surgery [3]. A clinical trial comparing the ACP with optimal medical treatment is presently on hold.