PHOENIX — Anticoagulation worked at least as well as antiplatelet medication for stroke prevention after cervical artery dissection, an observational study suggested.

Subsequent ischemic stroke was numerically but not significantly less common with anticoagulation than with antiplatelet use in this population both at 30 days (adjusted HR 0.71, 95% CI 0.45-1.12) and 180 days (adjusted HR 0.80, 95% CI 0.28-2.24), according to Shadi Yaghi, MD, of Brown Medical School in Providence, Rhode Island.

For patients with occlusive dissection, the advantage of anticoagulation was significant (aHR 0.40, 95% CI 0.18-0.88, P=0.009 for interaction), Yaghi reported at the American Stroke Association International Stroke Conference. The STOP-CAD study findings also were published in Stroke.

Anticoagulation didn’t carry higher risk of major hemorrhage in the first 30 days (aHR 1.39, 95% CI 0.35-5.45, P=0.637), but that risk became substantial by 180 days (aHR 5.56, 95% CI 1.53-20.13, P=0.009).

“Our study does not rule out a benefit of anticoagulation in reducing ischemic stroke risk, particularly in patients with occlusive dissection,” the researchers concluded. “If anticoagulation is chosen, it seems reasonable to switch to antiplatelet therapy before 180 days to lower the risk of major bleeding.”

Late-breaking science session moderator Lauren Sansing, MD, of Yale School of Medicine in New Haven, Connecticut, agreed with those conclusions.

“What it looks like is that probably early anticoagulation will be better, but that after 30 days we should be switching, probably not keeping those patients on anticoagulation,” she told MedPage Today.

Cervical artery dissection accounts for only about 2% of ischemic strokes but up to 25% of those in young adults ages 50 and younger. Some 3% result in subsequent stroke, so guidelines have recommended antithrombotic therapy for 3 to 6 months afterward without specifying anticoagulation or antiplatelet use.

With relatively small numbers of events and a low overall risk of stroke on either of those medications is low, “it’s hard to then show that one treatment is beneficial over the other,” Sansing noted. “But it’s incredibly important because, again, these are young people, so another stroke is a devastating event.”

Practice has been in equipoise with clinicians relying on intuition about what might be better for any given patient, she added.

Two prior trials, CADISS and TREAT-CAD, compared vitamin K antagonist anticoagulation against antiplatelet therapy after cervical artery dissection but didn’t definitively answer the question. One suggested anticoagulation was superior but was underpowered; the other failed to show non-inferiority of aspirin.

With its 4,023 patients, STOP-CAD “dwarfs any kind of studies that were done before, so that’s the big strength. And it’s a well done study for a retrospective study,” noted session co-moderator Tudor G. Jovin, MD, of the Cooper Neurological Institute in Camden, New Jersey.

However, “with every retrospective study, there’s potential for bias,” he added. “It doesn’t, in my opinion, give us any definitive answers, but it gets us close to that answer.”

Nevertheless, in the absence of a clear winner, Sansing said, “It’s helpful to know that there’s no wrong way. We haven’t been doing it wrong.”

Yaghi’s group called for large prospective studies to validate their findings.

The study included 63 sites in 16 countries that enrolled their adult patients presenting for acute care of cervical artery dissection that was confirmed by imaging. Incidental or iatrogenic chronic dissection cases, dissecting pseudoaneurysm, and those with major head or neck trauma within the prior 4 weeks were excluded.

In total, 3,636 patients (11.1%) received only therapeutic anticoagulation, whether parental anticoagulation, heparin or low-molecular-weight heparin, oral vitamin K antagonist, or a direct oral anticoagulant. Another 2,453 (67.5%) got exclusively antiplatelets, whether a single or dual antiplatelet regimen. The rest of the cohort crossed over between the two antithrombotic types or received both concurrently.

In subgroup analyses, results were similar between the direct oral anticoagulants and vitamin K antagonists.

Limitations included the retrospective observational design with lack of central and blinded outcome adjudication and likely confounding by indication, although mitigated by inverse probability weighting and propensity matching. Endpoint events were only considered if they happened while on treatment.

Also, Yaghi noted at the session, most sites were large academic institutions, “thus limiting generalizability to small community hospitals.”

Please enable JavaScript to view the

comments powered by Disqus.