Left atrial appendage closure is a proven method for prevention against stroke in patients with non-valvular atrial fibrillation. This method is proven for having clinical utility as well as safety.

The need for coverage in this area is immense. Atrial fibrillation affects 1%–2% of the entire adult population.1 Of all ischaemic strokes, 15% are associated with atrial fibrillation. Both the CHADS₂ and the CHA₂DS₂-VASc risk scores have been used for determining the level of anticoagulation in these patients. Multiple anticoagulants have been used for prevention of ischaemic stroke, including warfarin, dabigatran, rivaroxaban, apixaban and edoxaban. The main purpose of the ‘new’ oral anticoagulants is to provide anticoagulants without the additional need for monitoring.

The Amplatzer Cardiac Plug and the second-generation Amulet device have a large history of registry data since the device was first implanted in 2008. The paper by Lopez-Minguez et al2 reports results from 167 patients who were contraindicated for oral anticoagulation. Procedural success was high at 94.6% of patients, making the procedure very safe. Stroke was at 2.4% per year, which consisted of a 75% reduction compared with expected risk. For these patients, anticoagulation is not an option. This is in line with other large databases, including the ones mentioned in their table 4. The Canadian registry, which had a higher baseline stroke risk and a similar follow-up period, also showed a similar reduction as well as the ASA Plavix Feasibility Study With Watchman Left Atrial Appendage Closure Technology (ASAP) study. Of all the studies listed, the Iberian study had the highest annual risk of stroke. It is important here to mention the results from the international multicentre European Amplatzer Cardiac Plug (ACP) registry,3 which followed 1047 patients for a mean of 13 months with 98.2% complete follow-up. This showed a procedural success of 97.3% and periprocedural major adverse event rate of 4.97%. Annual rates of stroke, systemic thromboembolism, and major bleeding were 0.9%, 2.3% and 2.1%, respectively.

Randomised trials exist for the Watchman device (Boston Scientific, Natick, Massachusetts, USA). The PROTECT-AF trial4 was a randomised trial of over 707 patients from 59 centres which showed statistical superiority for the combined efficacy endpoint, as well as improvements in all-cause mortality and cardiovascular mortality. The confirmatory PREVAIL trial showed improvements in safety, and there was found to be a learning curve with the device, which required trans-septal access and placement of a device into the left atrial appendage (LAA). There are also improvements in quality of life and cost-effectiveness. This trial used warfarin for postprocedural management; however, a small study in patients who could not tolerate oral anticoagulation showed high safety and efficacy using aspirin and clopidogrel in its place.5

The patients referred to in this paper, that is, those who cannot tolerate oral anticoagulation, are at potentially higher risk for bleeding with dual antiplatelet therapy, as mentioned in the discussion. The authors do well in listing specific procedural bleeding indications in over 80% of patients, as well including patients with a mean HAS-BLED score of 3. It is important to note that most patients in this series started with dual antiplatelet therapy (87.3%), but then went on to mostly single antiplatelet therapy (87.8%) at 12 months. Even with this therapy, there were still bleeding events (as seen in the paper's table 5); one can only imagine that with oral anticoagulation, or persistent dual antiplatelet therapy, there would be many more.

Procedural safety has been raised as an important issue with left atrial appendage closure. Here procedural complications occurred in 5.4% of patients. The vast majority were vascular complications. Of note, none of the complications required surgery—all were managed percutaneously. This is important to note, as many readers may see pericardial tamponade as a necessarily life-threatening event. These effusions are most often seen periprocedure and can be responded to by the implanting physician; if caught quickly, the blood can be autotransfused immediately using the venous sheath. It is important to focus on the vascular access, trans-septal puncture, sheath de-airing, and gentle manoeuvring in the left atrial appendage space. Physician experience and learning curve have been shown to be important in LAA closure; by doing so, the rate of device-related safety events can be as low as 3.7%, as seen in one continued access protocol study.6

One of the myths of oral anticoagulation and medical therapy is that patients are compliant and take these medications regularly. This is far from the truth. Patients are rarely compliant with these medications. In the trials mentioned, time in target therapeutic range for warfarin ranged from 55% to 64%. In the community non-trial setting, one study showed that only 1/3 of all eligible patients were taking Coumadin.7 Discontinuation of dabigatran was 16% at 1 year and 21% at 2 years, discontinuation rate of rivaroxaban and apixaban was over 20% in the trials. Even in a very rigorously controlled and monitored trial setting, within only 1 year, one out of five patients were not protected from stroke due to atrial fibrillation. In a community setting, the numbers are probably even higher, and may continue to grow with time.

Another significant issue with oral anticoagulation is the issue of bleeding. In the three clinical trials of newer anticoagulants, the rate of major bleeding was 3%–4% with warfarin, and 2%–4% with the newer anticoagulants. For patients with contraindications to oral anticoagulation or previous bleeding, these numbers are much higher. When these patients bleed on new oral anticoagulants, the long half-lives can lead to issues with management. Reversal of these agents is an area of active discovery, but this has not been fully resolved. Interrupted use due to procedures, drug–drug interactions, supratherapeutic dosing and low time in therapeutic range are all issues that lead to lower efficacy for these agents. Additionally, the medications are costlier alternatives to warfarin, with only apixaban showing a mortality benefit compared over warfarin.8

For patients with atrial fibrillation, left atrial appendage closure therapy is a valuable option with clinical efficacy and safety. Patients with the inability to tolerate oral anticoagulation can definitely benefit as no other options exist. As for those who can tolerate oral anticoagulation, a randomised clinical study has shown superiority in the primary clinical endpoint. This is similar to transcatheter aortic valve replacement, which is a viable option for patients who are inoperable as well as those who are high-risk candidates.

In our practice, the ability to have combined decision making is supremely important. For patients who cannot tolerate oral anticoagulation, this is the first option, and we recommend this strongly. For patients who can tolerate oral anticoagulation, we discuss the benefits and risks of both strategies. A large group of well-informed patients are interested in a one-time device therapy than continue being on potentially life-long medications. Even the ‘minor’ bleeding, or as it is otherwise known, ‘nuisance’ bleeding, is often a major issue for our patients. The nuisance is probably for the physician who continues to listen to these symptoms without being able to recommend discontinuation because he or she is committed to this strategy. The ability to avoid anticoagulation is, for these patients, an attractive if not preferential one.

In short, there is enough safety and clinical efficacy data to make left atrial appendage closure a viable option for stroke prevention in patients with atrial fibrillation (box 1).