SURPRISE: AF Common in Cryptogenic Stroke

Daniel M. Keller, PhD

December 11, 2012

Brasilia, Brazil — In a Danish study of patients with cryptogenic stroke and no cardiac symptoms, long-term cardiac monitoring post-stroke revealed that almost one fifth of patients had atrial fibrillation (AF).

The results, from the Stroke Prior to Diagnosis of Atrial Fibrillation Using Long-term Observation with Implantable Cardiac Monitoring Apparatus Reveal (SURPRISE) study, were presented at the 8th World Stroke Congress (WSC).

First-year resident Louisa Christensen, MD, from the Department of Neurology at Bispebjerg University Hospital of Copenhagen, Denmark, pointed out that cardioembolic and large- and small-vessel disease factors underlie most strokes, "but we are still left with 25-30% of stroke that is of unknown cause," even in the presence of known risk factors.

Paroxysmal AF can become persistent and then permanent if not caught and treated early. Electrocardiography and even Holter monitoring after hospital admission for stroke can miss short bursts of AF. "By considering longer term monitoring we would be able to look at a patient over time and maybe catch early episodes of atrial fibrillation before a recurrence," Dr. Christensen said.

Significant Burden of Undetected AF

The aim of the SURPRISE study was to use long-term monitoring of patients with apparent cryptogenic transient ischemic attacks or minor ischemic strokes to establish the frequency and burden of paroxysmal and persistent AF.

To do this, the researchers successfully implanted MRI-compatible cardiac monitoring devices (CareLink Monitor, Medtronic) into 84 patients. Patients held a small pick-up unit over the device, which then sent data to a secure computer server over mobile Internet or phone line connections. Physicians could then review the data on the CareLink clinician Web site.

The mean time between stroke onset and device implantation was 78.2 ± 74.9 days. A senior cardiologist adjudicated events. If AF was confirmed, the patient was prescribed anticoagulant therapy. Of the 84 patients, about three quarters were in sinus rhythm, 14 (16.7%) had paroxysmal AF, and 1 (1.2%) had permanent AF.

These 15 patients were in AF for a median of 0.3 hour per day of monitoring, with a minimum of 0.1 hour and a maximum of 17 hours. The mean time between device implantation and the first recorded AF event was 50.9 ± 48.1 days, and the mean time between stroke onset and the first recorded AF event was 96.6 ± 43.5 days.

Dr. Christensen concluded that AF is common in patients with cryptogenic stroke and no cardiac history or symptoms; in this case, AF was documented among 17.9% of the cohort with long-term monitoring. The results she presented were based on data obtained through October 2012, and she said further data analysis will be published by the end of the year.

Cryptogenic Strokes

Session moderator Ralph Sacco, MD, MS, professor and chairman of the Department of Neurology at the University of Miami Miller School of Medicine in Florida and past president of the American Heart Association, commented to Medscape Medical News that a large proportion of stroke patients whose strokes are labeled "cryptogenic" or who have infarcts of undetermined cause may have actually had embolic strokes, "but [we] haven't been able to determine the source of the embolism. A few studies have now shown that asymptomatic, silent, or undiagnosed atrial fibrillation or even subclinical atrial fibrillation may be occurring more frequently in these cryptogenic stroke cases."

He said the SURPRISE study is intriguing in that it found that about 25% of patients may have had various cardiac arrhythmias undetected by usual Holter monitoring or electrocardiography in these cases of cryptogenic infarcts.

Dr. Sacco said documenting AF in these cases is clinically relevant because it changes patient management. "That puts them into a cardioembolic category, and that usually means you should be using oral anticoagulants," he advised.

He mentioned other monitoring devices may exist, such as skin patches, which can provide long-term monitoring without the need for implantation. But he said it remains to be seen whether they, too, "can still improve the detection of subclinical or silent or undiagnosed or asymptomatic atrial fibrillation."

Dr. Sacco noted that he and many other experts believe that paroxysmal AF is just as risky as permanent AF, "and most of the guidelines in terms of oral anticoagulant use have now combined paroxysmal Afib with permanent or persistent Afib."

Dr. Christensen has received grants from Medtronic for conference attendance and for research support, from Boehringer Ingelheim for conference attendance, and from The Arvid Nilssons Foundation for research support. Dr. Sacco was not involved in the study and has disclosed no relevant financial relationships.

8th World Stroke Congress (WSC). Abstract 665. Presented October 12, 2012.