麻豆传媒

Among patients with ischemic stroke attributed to large artery atherosclerosis or small vessel occlusion, atrial fibrillation (Afib) detection was increased more than seven-fold with an implantable cardiac monitor (ICM) compared with standard care, according to the STROKE AF trial.

After 1 year, the Reveal LINQ device picked up previously unidentified Afib in 12.1% of patients, while usual care identified just 1.8% (HR 7.41, 95% CI 2.60-21.28, P<0.001), reported Lee H. Schwamm, MD, of Massachusetts General Hospital in Boston, at the American Stroke Association's virtual International Stroke Conference.

The findings of non-cardioembolic stroke of presumed known origin tracked almost exactly with those seen previously in the , which established loop recorder use in cryptogenic strokes, said Schwamm.

"Forget about the first stroke," he told 麻豆传媒. "If you find fibrillation, you should treat it. ... The reality is it sort of doesn't matter what provoked it, other than direct manipulation of the heart during heart surgery."

The real question, he said, is "should we start monitoring anybody with a stroke or a sufficiently elevated CHADS VASC score to try to detect undetected atrial fibrillation? Because it will change how we try to prevent their stroke."

That makes this an important study, commented Louise McCullough, MD, PhD, of the University of Texas Health Science Center at Houston.

"Every time we send somebody with a stroke, either with large vessel atherosclerosis or small vessel disease, out on antiplatelets, we feel like we're protecting them," she said. "But if they have occult Afib, we're not."

Such patients who initially appear to be in sinus rhythm typically haven't been considered for extended rhythm monitoring, she noted.

"Even if you prevent a few cardioembolic strokes, that's a huge cost savings for public health," she said. "We need to catch these people; we need to have a very low bar and a high index of suspicion. This study shows it needs to be even higher than we had previously thought."

Notably, median time to the first Afib episode in with the device was 99 days, and 78% would have been missed with 30-day monitoring, suggesting that a patch or Holter monitor wouldn't catch many of these cases.

The number needed to monitor to detect one case of Afib was just eight in the ICM group, compared with 56 in the standard-of-care group.

The open-label trial included 496 patients with ischemic stroke attributed to small vessel occlusion or cervical or intracranial artery atherosclerosis treated at 33 U.S. centers. The population was an elevated-risk group: patients who weren't at least 60 years old could be enrolled as young as 50 only if they had a documented medical history of at least one additional stroke risk factor (congestive heart failure, hypertension, diabetes, second stroke more than 90 days prior, or vascular disease).

Participants were randomized within 10 days of the index stroke to the continuous monitoring arm with the Reveal LINQ ICM or to whatever the site-specific standard of care was for arrhythmia detection.

Afib had to last for at least 30 seconds (although the ICM only picked up episodes of at least 2 minutes' duration) without detectable P waves and be adjudicated by a clinical events committee.

For 55% of the patients with Afib, the duration of the longest detected episode was an hour or more, which "would be considered by many physicians appropriate to treat with anticoagulation for secondary prevention," Schwamm told the virtual session attendees.

Indeed, 18 of the 27 Afib patients (66.7%) in the ICM group were started on an oral anticoagulant compared with three of the four patients identified with usual care (75%).

While the trial was underpowered for such clinical endpoints, there were numerically fewer recurrent strokes in the ICM group at 12 months (15 vs 23 patients in the usual care group).

"Our long-term results (36 months) and other ongoing studies may shed light on the clinical significance of ICM-detected [Afib] for stroke recurrence," Schwamm noted in the presentation.

Comment