Richard Whitlock MD, PhD of McMaster University in Hamilton, Ontario presented the outcomes of the LAAOS III trial as a late breaking clinical trial (LBCT) at the 2021 virtual ACC scientific sessions. The findings were simultaneously published in the New England Journal of Medicine.

Surgical left atrial appendage (LAA) closure is intended to reduce thromboembolic events. The indications for this procedure and the methods by which it should be performed have remained controversial for decades. While commonly employed, there has been a lack of strong scientific data to support this practice.

After pioneering work by Dr. John Madden in New York more than 70 years ago, a number of small uncontrolled studies have shown high surgical LAA closure success rates in patients with atrial fibrillation (AF) undergoing cardiothoracic surgery. Several of these studies have demonstrated reductions in both systemic thromboembolic and stroke rates. These positive outcomes have been buttressed by retrospective database analyses, which suggest that this procedure has the potential to reduce post-operative stroke rates in AF patients undergoing cardiac surgery. However, several manuscripts indicate that incomplete surgical closure might actually be harmful by augmenting systemic thromboembolic rates.

Given the lack of prospective randomized trials to inform clinicians, published society guidelines have labelled surgical LAA occlusion with a class IIb level of recommendation. In addition, the European guidelines recommend that anti-coagulation be continued post-operatively in these patients whose risk profile places them at a moderately high risk for subsequent thromboembolic events.

The results from LAAOS III shed much light on this issue and provide evidence to help inform clinician decision-making. In this study 4811 patients with a history of AF undergoing cardiothoracic surgery were randomized to either surgical LAA occlusion, performed by several modalities, or no LAA occlusion. Interestingly, 66.6% of the patients underwent valve surgery. The mean CHA2DS2-VASc score was moderately high at 4.2. Concomitant surgical AF ablation was performed in 32.7% of the patients. Surprisingly only 48.4% of patients were treated with either a DOAC or VKA preoperatively.

The cumulative incidence of subsequent stroke or systemic thromboembolism was 4.8% in the ‘no LAA occlusion’ group vs. 7.0% in the ’LAA occlusion’ group with a hazard ratio (HR) of 0.67 (95% CI: 0.53-0.83), p= 0.001. These results were driven by a larger reduction in ischemic stroke (HR: 0.66 [0.57-0.84]) than systemic embolism (HR: 0.86 [0.29-2.55]). There were no differences by intention to treat (ITT), per protocol (PP) or on treatment (OT) assessments. By study design patients were asked to continue anticoagulant therapy and compliance with this recommendation was 76.8%.  The major adverse clinical event (MACE) rate was very low and not different between the two groups.

The robust results observed in LAAOS III shed much light on this issue and provide strong evidence that patients with a history of AF and a CHA2DS2-VASc score > 2 undergoing cardiac surgery have a lower systemic thromboembolic risk when they undergo concomitant LAA closure. Based upon these findings it is my opinion that surgical LAA occlusion should be performed routinely as a standard of care at the time of cardiothoracic surgery in this patient population.

Remaining unanswered questions

  • Should the guidelines be updated based upon this large, well-designed trial or should an update be postponed until further confirmatory studies, which could be generated in a somewhat less rigorous manner (e.g. databases or registry outcomes), become available?
  • Should AF patients with a CHA2DS2-VASc score < 2 undergoing cardiothoracic surgery also have LAA occlusion?
  • What should be done in patients with a relatively high CHA2DS2-VASc score but no history of AF since AF is relatively common in this patient group?
  • Are there subgroups of AF patients in whom we can eventually stop anticoagulant therapy after surgical LAA occlusion or should we continue anticoagulation in all of these patients pending further investigations?
  • Is it important to perform an anatomical analysis (e.g. CT scan, TEE, etc.) to assess for complete closure to ensure that the thrombo-embolic risk is minimal or is this overkill?
  • What is the role for percutaneous LAA closure in these patients?


While this very important study advances our scientific knowledge on the role of surgical LAA closure, it leaves many questions unanswered. As clinical trialists we must build upon this foundation; determine the most important residual questions and design future clinical investigations that will address these issues and ultimately improve the quality of cardiovascular care.

Please let us know what you think of this trial and what questions we should prioritize?


Leave a Reply