Disclosure: John M. Mandrola, MD, has disclosed no relevant financial relationships
In nearly a decade of previewing major cardiology meetings, I have never chosen to highlight a single trial. The exception this time is because I believe that electrophysiology is on the brink of a large wave of overtreatment.
Among the late-breaking trials at the upcoming American Heart Association Scientific Sessions is the OPTION trial. It will compare atrial fibrillation (AF) ablation done concomitantly with percutaneous left atrial appendage closure (LAAC) vs ablation plus continuation of direct-acting oral anticoagulants (DOACs). The results of this trial could herald the beginning of a dubious new practice pattern where the two procedures are combined.
The Centers for Medicare & Medicaid Services in the United States have recently created a billing code for concomitant LAAC and AF ablation. I've already seen social media pictures of doctors who performed the two procedures together.
Both procedures are popular and well reimbursed. Technology has reduced the time and effort to do AF ablation. While we do many more AF ablation procedures than in the past, its success rates remain modest.
LAAC is also well compensated, but the claim that LAAC can reduce stroke by plugging the opening in the LAA has far weaker evidence than AF ablation. Seminal trials of LAAC vs warfarin either failed to pass FDA muster (PROTECT) or did not meet noninferiority for one of its co-primary endpoints (PREVAIL).
Most LAAC is now done in patients unlike those in the seminal trials. Hundreds of thousands of LAAC procedures have been done worldwide, and yet, because so few patients have been randomly assigned in trials, we have no convincing evidence of net benefit from this procedure.
At the AHA meeting, the OPTION trial probably will deliver positive results. Enthusiastic marketing will follow. A colleague recently showed me his email invite to the OPTION "celebration meeting" at AHA.
Given the popularity of AF ablation, adding concomitant LAAC could lead to a massive increase in this procedure -- without convincing evidence of net benefit or exclusion of net harm.
Proponents of LAAC cite the ability to interrupt or stop oral anticoagulation after the ablation as the main advantage of adding LAAC. Indeed, many patients would embrace the idea of not taking oral anticoagulants after their ablation.
Five flaws in the OPTION study design nearly guarantee a "positive" result and probably will limit its use in clinical decision-making. The trial uses a noninferiority design to test the primary efficacy endpoint -- designating LAAC as the active arm and continued DOAC as the control arm.
The first flaw comes in the trialists' choice of stroke, systemic embolism, and all-cause death as the primary endpoint. Since no study -- in either ablation or appendage occlusion -- has found statistically significant differences in all-cause death, the most likely result in OPTION is that death rates will be similar in both arms. Consider that the LAAOS III trial of concomitant surgical appendage closure (vs no closure) found significant reductions in stroke with surgery, but this was not enough to reduce death rates.
Because no difference in outcomes is a "positive" result in a noninferiority trial, adding a relatively common endpoint, which is not affected by either treatment, makes noninferiority easier to reach. In OPTION, all-cause death will "dilute" any differences in the true efficacy question -- stroke and systemic embolism.
The second flaw in OPTION is the low numbers of patients. OPTION seeks to show noninferiority for thrombotic events and superiority of bleeding rates with LAAC vs DOACs. This is similar to the DOAC-vs-warfarin trials. But the drug trials enrolled more than 10,000 patients each. OPTION will enroll only 1600 patients; it will be underpowered.
The third flaw of the trial is the estimate of event rates and choice of noninferiority margin. Trialists estimate an event rate of 10% in the ablation/DOAC arm and a fixed noninferiority margin of 5%. A meta-analysis of AF ablation trials in patients with heart failure, who would presumably be at higher thrombotic risk than those in OPTION, reported stroke rates of about 4%. Given advances in ablation technology and its ability to reduce AF burden, stroke and embolism rates in OPTION probably will be lower than those observed in trials of AF ablation in patients with heart failure.
The likely scenario, therefore, is that primary outcome event rates will be lower than 10%. Stroke and systemic embolism alone will surely be lower than 10%. The noninferiority margin of 5% was set with the assumption of a 10% rate of events. If the observed primary event rate is lower than that, say 5%, then a 5% margin allows for the active arm to be declared noninferior even if those patients have twice as many events. OPTION could have allowed for lower-than-expected event rates by also testing noninferiority with a risk ratio (relative difference). The rationale and design paper does not mention risk ratio testing.
The fourth flaw is that the primary noninferiority endpoint will be tested in the intention-to-treat (ITT) population. While ITT is the appropriate choice in superiority trials, in noninferiority trials, the better test is the as-treated population. By including patients who were randomly assigned to LAAC but did not have the procedure, noninferiority is easier to reach. The authors write that they will test both ITT and as-treated, but they designated that their primary analysis will be done in the ITT group.
The fifth flaw involves the choice of safety endpoints. The trialists' choice of "nonprocedural" bleeding as a primary safety endpoint is not appropriate. Nonprocedural endpoints are reasonable only as exploratory secondary endpoints. LAAC is an invasive procedure; one of its primary safety issues is procedural bleeding due to perforation, pericardial effusion, and tamponade. A patient cannot avoid procedural bleeding.
How best to deal with stroke prevention and reduction of bleeding after ablation in the left atrium is a relevant clinical issue. After AF ablation, patients often ask whether they need to take oral anticoagulants; we have no evidence to guide our recommendation.
The challenges are that postablation thrombotic and bleeding events occur infrequently, and AF is a chronic condition that can recur years after ablation. OPTION will provide some evidence, but its serious design flaws will limit its use in decision-making. These flaws, however, probably will not dampen the marketing or spin after the trial.
I write about the flaws in trial design not to attack the sponsors or authors. Instead, I hope to improve clinical translation of this trial.
The current popularity of LAAC far exceeds its evidentiary basis. By identifying the many limitations of this trial before the results are presented, I aim to help my colleagues see through the enthusiasm of a positive trial.
For if they do not, many people will have foreign bodies placed in their hearts -- with extremely weak supportive evidence and the possibility of long-term harm.