More Strokes, Especially Ischemic, With LAAO vs Medical Therapy: Meta-analysis

For patients with atrial fibrillation (AF), left atrial appendage occlusion (LAAO) is associated with a higher rate of stroke, driven by a difference in ischemic stroke, compared with oral anticoagulation (OAC), according to a new Bayesian meta-analysis.

The analysis showed, however, that the association was not statistically strong and the “estimated absolute excess is small,” leading the authors to question whether this risk should influence decision-making when counterbalanced by a 45% lower incidence of hemorrhagic stroke.

With results of both the CHAMPION-AF and CLOSURE-AF trials putting a damper on enthusiasm for LAAO—despite the overall noninferior efficacy of the procedure compared with oral anticoagulation in the former study—the field is craving more context for where it might be most helpful. Other recent meta-analyses published over the summer, including one with a random-effects model of the same six trials included here and another of seven trials, have shown conflicting findings regarding stroke and bleeding.

Lead author Amar Rai, MBBS (University of Pennsylvania, Philadelphia, and Imperial College London, England), said that as a clinician working with patients who have high bleeding risks in general, he wanted to look not only at stroke risks, but also at bleeding and overall procedural risks with LAAO versus oral anticoagulation.

“[For] the patients that I work with, it’s mostly shared decision-making, and I want them to be informed, and clinicians around the world [to] also be informed, about [risks and benefits],” he told TCTMD.

Andrew Goldsweig, MD (University of Massachusetts-Baystate, Springfield), who was not involved in the study, said the key contribution from the new analysis, published earlier this month in JSCAI, is its Bayesian hierarchical methodology used to calculate the probabilities of true differences in outcomes.

“While the present analysis is congruent to other recent meta-analyses in identifying slightly higher ischemic stroke and lower nonprocedural bleeding rates, Rai et al suggest the novel findings that all-cause stroke may be higher with LAAO (70% chance) but that major bleeding may be lower with LAAO (86% chance),” he commented to TCTMD in an email.

With known heterogeneity in the trials—medication options included warfarin and direct oral anticoagulants (DOACs), device options included Watchman 2.5 and Watchman FLX (Boston Scientific) as well as Amplatzer Amulet (Abbott), and operator experience varied—it’s hard to directly compare findings, according to Goldsweig. “I was very pleasantly surprised to find that probabilities favor less hemorrhagic stroke and less major bleeding with LAAO,” he said. “This meta-analysis finding is hypothesis-generating and strongly supports continued development of LAAO as an alternative to anticoagulation that may be superior both in terms of stroke and in terms of bleeding.”

Six Trials

The meta-analysis included 7,004 patients with nonvalvular AF eligible for long-term oral anticoagulation who were enrolled in one of six trials: PROTECT AF, PREVAIL, PRAGUE-17, OPTION, CLOSURE-AF, and CHAMPION-AF. Patients were randomized to either LAAO (n = 3,681) or oral anticoagulation (n = 3,323), and follow-up ranged from a median of 1.0 to 3.8 years.

Using a Bayesian hierarchical analysis, the researchers found the primary outcome of ischemic stroke occurred more often with LAAO than medical treatment with 94% confidence (incidence rate ratio (IRR) 1.34; 95% credible interval [CrI] 0.91-1.97). The effect was smaller with any stroke (IRR 1.09; 95% CrI 0.76-1.56), primarily due to a lower rate of hemorrhagic stroke in the LAAO group (IRR 0.55; 95% CrI 0.23-1.18), with these observations made with 70% and 94.5% confidence, respectively.

Among trials using DOACs, researchers noted an attenuation in the drop in hemorrhagic stroke (IRR 0.80; 95% CrI 0.33-1.75).

“The reason we did this is mainly [to see if this could] be an effective alternative to long-term anticoagulation in patients,” Rai said. “And when you go back to that, it is useful, but we don’t know what population group yet.” Any procedure increases ischemic risk acutely, he continued, “but from the results we have, it shows that long-term [LAAO] could be potentially more beneficial than DOACs, but there’s no clear answer at this very moment in time.”

More data on specific subgroups are needed to see if LAAO would be especially helpful for certain populations—potentially in older adults or those with end-stage renal disease, Rai said. The ongoing CATALYST trial, as well as other registry studies, should be able to provide some answers in this regard, he added.

Rai said that most patients with whom he discusses LAAO are at high bleeding risk. “We have the discussion that, ‘Yes, anticoagulation will be maybe not as beneficial for you—hence, we would think about a LAAO device,’” he said. “From there—this is where this paper really comes in strong—‘Yes, initially as we do this procedure, there is an increased risk of stroke and we can see that. But in the long term, your bleeding risk will be minimized in comparison to anticoagulation and also your stroke risk to a degree.’”

More Data, Clinical Guidance Needed

With long-term data still forthcoming, Rai said he supports many of these patients moving forward with LAAO, although “it depends on the individual. There’s no clear answer that one is better than the other. It’s just about which risks you’re willing to take and which benefits you are more likely to feel. Because again, from this data, we haven’t got a clear answer yet, but we’re sure both are equally beneficial in patients.”

Taking a step back, this meta-analysis suggests that “LAAO may be truly equivalent or even superior to OAC for both stroke prevention and bleeding prevention,” Goldsweig said. The findings presented here “strongly support” the argument made in the 2025 practice guideline for LAAO from the Society for Cardiovascular Angiography and Interventions (SCAI) and the Heart Rhythm Society (HRS) that “LAAO and OAC offer equivalent overall stroke/bleeding protection.”

Even so, the options are “not interchangeable for most patients,” he said. “There are clinical and anatomic factors as well as patient values and preferences that may make either LAAO or OAC more favorable for an individual patient.”

In addition to more subgroup research, Goldsweig also advocated for the creation of a “decision score or risk-stratification algorithm . . . to help guide this decision in clinical practice.”