NEW ORLEANS – Continuous cardiac outpatient telemetry monitoring for at least 21 days may be necessary to accurately identify atrial fibrillation as the cause of cryptogenic stroke, according to the results of a single-center, retrospective study.
Identifying these patients is important because many could be treated with anticoagulants that would reduce their stroke risk, said Dr. Daniel J. Miller, a neurologist at Henry Ford Hospital, Detroit.
About a third of all strokes are of unknown etiology. Paroxysmal atrial fibrillation (AF) is a potential cause of cryptogenic stroke and transient ischemic attacks (TIAs), and carries a high risk of future strokes, Dr. Miller said during a press briefing at the International Stroke Conference.
Physicians have not been sure about the best way to monitor these patients or how long to monitor them.
He and his colleagues followed-up on two small studies: one conducted by Dr. Ashis H. Tayal and colleagues (Neurology 2008;71:1696-1701) and another by Dr. Archit Bhatt and associates (Stroke Res. Treat. 2011 [doi:10.4061/2011/172074]). The Tayal study showed a 23% AF detection rate in cryptogenic stroke patients monitored for longer than 21 days. The Bhatt study found a similar AF rate in cryptogenic stroke.
Dr. Miller and his coinvestigators reviewed the records of 156 Henry Ford Hospital patients who’d had a cryptogenic stroke or TIA and had undergone monitoring with the Cardionet Mobile Cardiac Outpatient Telemetry (MCOT) device, which is commonly used at Henry Ford
The patients had a mean age of 68 years and half were women. The vast majority (97%) were not taking an anticoagulant. Hypertension was common, present in 87% of the patients. They had a fairly low mean score of 3 on the National Institutes of Health Stroke Scale (NIHSS), indicating that the stroke had not been severe.
After monitoring, 27 (17.3%) of the 156 patients had paroxysmal AF. The AF events lasted less than 30 seconds in two-thirds of the patients and longer than 30 seconds in 26%. The remaining 8% had persistently occurring AF. This finding was not surprising and was consistent with the Tayal report, Dr. Miller said in an interview. The monitors are very sensitive, in particular for those events longer than 30 seconds, but they might miss some episodes of less than 30 seconds, he said.
The review showed that the rate of AF detection rose with increased duration of monitoring. The detection rate on a Kaplan-Meier curve was estimated to be 4% at 48 hours, 9% at 7 days, 14% at 14 days, and 20% by 21 days. "Our study does show that in order to capture all these events you should continue to monitor for at least 21 days," Dr. Miller said.
A multivariate Cox regression analysis showed that for all patients, female gender (P = .002), premature atrial complexes (PAC) on electrocardiogram (P = .001), a 1-cm increase in the diameter of the left atrium (P = .033), and a 10% decline in left ventricular ejection fraction (P = .008) all were associated with increased risk of AF. For stroke patients, female gender, PAC, and increasing stroke severity on the NIHSS were all associated with increased risk.
PAC, a premature beat from the atrium, has been shown to be associated with AF. It was an especially strong predictor of AF in this study, Dr. Miller said.
He urged further study to determine optimal monitoring beyond 21 days for patients with cryptogenic stroke or TIA.