Stroke, Bleeding Risks Key for Assessing Potential LAAO Benefit

The potential for left atrial appendage occlusion (LAAO) to provide a clinical benefit relative to lifelong oral anticoagulation in atrial fibrillation (AF) varies based on a patient’s risks of stroke and bleeding, as well as the type of anticoagulant, a modeling study shows.

LAAO is most likely to be a good option when a patient’s bleeding risk is high and stroke risk is low, with less-certain benefits compared with oral anticoagulation when bleeding risk comes down and stroke risk rises, lead authors Derek Chew, MD (Libin Cardiovascular Institute, University of Calgary, Canada), and Ke Zhou, MD, PhD (Duke-National University of Singapore Medical School, Singapore), report.

That pattern is similar irrespective of whether the blood-thinning agent is warfarin or a direct oral anticoagulant (DOAC). But because of the lower risk of bleeding with the latter agents, the subset of patients with AF in whom LAAO has a predicted clinical advantage over anticoagulation is smaller when the comparator is a DOAC rather than warfarin.

The findings were published online this week in the Annals of Internal Medicine.

“It does make sense that in cases of very high risk of stroke, protection from an agent like a direct-acting oral anticoagulant likely would be superior to focal mechanical occlusion of the left atrial appendage,” senior author Jonathan Piccini, MD (Duke University Medical Center, Durham, NC), told TCTMD. “And it’s also, I think, intuitive that the higher someone’s bleeding risk [is], the more they would benefit from a bleeding-avoidance strategy.”

For patients in the gray zone, where there is no obvious treatment choice based on the balance of bleeding and stroke risks, shared decision-making—and consideration of patient preferences—is key, Piccini said: “It’s always important, but it’s even more important in those situations.”

Uncertainty About Use of LAAO

Although LAAO has been shown to be noninferior to oral anticoagulation when it comes to overall clinical benefits, with lower risks of intracranial hemorrhage and major bleeding, there are some concerns about the procedural risks and possibly higher rates of ischemic stroke or systemic embolism versus medical therapy. In addition, without definitive trials comparing LAAO with DOACs, how extensively the procedure should performed—and in which patients—continues to be debated.

To explore which types of patients might be expected to benefit the most from LAAO after taking into account their bleeding and stroke risks, Chew, Zhou, and colleagues performed a decision analysis with a Markov model that simulated a clinical trial comparing LAAO with warfarin or a DOAC in 70-year-old patients with AF and no prior stroke.

Model inputs regarding clinical effectiveness came primarily from a patient-level meta-analysis pooling 5-year data from the PROTECT AF and PREVAIL trials, which compared the first-generation Watchman device (Boston Scientific) with warfarin. That analysis showed that LAAO reduced risks of major bleeding and intracranial hemorrhage (RRs 0.48 and 0.20, respectively) and was associated with a nonsignificantly higher risk of ischemic stroke (RR 1.40; 95% CI 0.76-2.59).

Baseline risks of stroke and bleeding were calculated using the CHA₂DS₂-VASc and HAS-BLED scores, respectively, and the interplay between the two determined the expected benefit of LAAO versus oral anticoagulation.

“We can actually see that trade-off between bleeding and stroke risk, where at very high bleed risk, that’s where left atrial appendage occlusion starts to shine and where it becomes more preferred. [But] as patients have increasing stroke risk, the uncertainty of benefit becomes more of a concern with left atrial appendage occlusion and oral anticoagulation is preferred,” Chew told TCTMD.

For a patient with a HAS-BLED score of 5, for example, LAAO had an advantage over warfarin—in terms of quality-adjusted life-years (QALYs)—in more than 80% of model simulations when the CHA2DS2-VASc score was between 2 and 5. At lower bleeding risk—HAS-BLED 0 to 1—a likely benefit of LAAO was limited to patients with a CHA₂DS₂-VASc score of 2. At low bleeding risk and very high stroke risk, warfarin was preferred over LAAO.

Appendage closure had an advantage over oral anticoagulation in a smaller group of patients when the comparator was a DOAC. A benefit in terms of QALYs was estimated only for patients with a CHA₂DS₂-VASc score of 2 and a HAS-BLED score of 2 to 5 and for those with a CHA₂DS₂-VASc score of 3 and a HAS-BLED score of 5.

The findings were similar when clinical benefit was assessed using life-years, but fewer combinations of bleeding and stroke risks pointed toward an advantage for LAAO when the outcome was net clinical benefit, which takes into account risks of ischemic stroke and intracranial hemorrhage. In fact, in terms of net clinical benefit, there were no scenarios in which LAAO was preferred over DOACs with at least 80% certainty.

The investigators performed multiple sensitivity analyses and found largely consistent results when using a different bleeding risk score (ORBIT) and other sources of LAAO clinical effectiveness data, including the PRAGUE-17 trial comparing LAAO to DOACs.

Identifying the Best Candidates for Device Therapy

The researchers say their results could be used to help physicians discuss the risks and benefits of LAAO when considering treatment options. “For patients with low-to-moderate CHA2DS2-VASc scores and high HAS-BLED scores, LAAO may be particularly beneficial,” they write. They note that this could apply to many patients, citing a secondary analysis of the ARISTOTLE trial showing that 12% of participants had a CHA₂DS₂-VASc score of 1 or 2 and a HAS-BLED score of at least 3.

Commenting for TCTMD, Faisal Merchant, MD (Emory University, Atlanta, GA), acknowledged the challenge of balancing risks of stroke and bleeding when considering LAAO for an individual patient, particularly because the two risks are closely related—if a patient has a high stroke risk they often also have a high bleeding risk.

This analysis “is starting to help put some meat on the bones of parsing out those sorts of categories of patients,” he said, adding, however, that there are limitations stemming from the model inputs—which came mostly from the large RCTs of Watchman versus warfarin. Both LAAO technology and oral anticoagulation have advanced since then. Moreover, he said, it’s difficult to take modeling data like these and translate them into real-world practice.

“At the end of the day, the reason we’re struggling in this field is we still don’t have good data on Watchman versus DOAC,” Merchant said. There are ongoing trials of LAAO versus DOACs, including CHAMPION-AF, OPTION, and CATALYST, but those include patients with average bleeding risks, he said, which contrasts with the types of patients mostly being considered for LAAO in everyday practice (higher bleeding risks and multiple comorbidities). So additional trials are needed in those higher-risk patients.

Nonetheless, Merchant said the results of the current analysis resonate with his clinical intuition about which patients might benefit the most from LAAO. The fact that there were no risk groups in which LAAO was favored over a DOAC when considering net clinical benefit “would suggest that . . . for the vast majority of patients, DOAC is still preferable until we have data to suggest otherwise.”

That calculus changes when a patient has a bleed while taking a DOAC, “but I do think that this probably reinforces that DOACs should be the preferred anticoagulation strategy for most patients, particularly those who are at higher stroke risk,” Merchant said. “I think that LAAO should really be a sort of niche, limited-use device for right now,” used in patients who have a high bleeding risk or who have already had a bleed and who have a relatively low stroke risk.

For the majority of patients being considered for LAAO, who tend to be older with multiple comorbidities and to have high bleeding and stroke risks, “it’s much more difficult to know, and I think it’s hard to make any real inferences from a modeling paper like this,” Merchant said.

Chew and Piccini both cautioned against putting too much weight in their comparisons between LAAO and DOACs because those models were based on data from a single trial (PRAGUE-17). That said, Chew stated, it makes sense that any benefits of LAAO over oral anticoagulation would be lessened when comparing the devices with agents carrying lower risks for bleeding.

Piccini said that based on the totality of the data from clinical trials and this analysis, “I would agree for the vast majority of patients, direct-acting oral anticoagulants would be preferred.” He added, however, that the answer could be different a few years in the future after more trial data are available. “Given that devices are evolving and we have better devices with better outcomes, I think given the results of this model, in the future . . . it may very well be that there may be true equipoise [between LAAO and oral anticoagulation] in persons who are at lower risk for bleeding and lower risk for stroke.”

For now, Piccini said, his “gut reaction” is that “left atrial appendage occlusion is preferred in patients who cannot take oral anticoagulation, and there probably is a sizeable group of patients who should consider left atrial appendage occlusion to minimize their bleeding risk. But I don’t think the majority of stroke prevention should be left atrial appendage occlusion.”

In an accompanying editorial, Bharat Kantharia, MD (Icahn School of Medicine at Mount Sinai, New York, NY), urges caution when interpreting the results and highlights some limitations of the analysis, including the lack of long-term data from LAAO trials and the use of assumptions for the model that might not be appropriate.

In addition, he says, the CHA₂DS₂-VASc and HAS-BLED scores are not valid in all patients, such as those with reduced ejection fraction, aortic atheroma, carotid disease, or cancer.

Though the study “has substantial scientific merit,” Kantharia said, “clinicians caring for patients with AF need further research to guide decisions when faced with the dilemma of anticoagulation versus LAAO.”