ISC Feature: Telemetry shows promise for cryptogenic stroke detection
“What we don’t know is the cause of these cryptogenic strokes,” the study's lead author Daniel J. Miller, MD, told Cardiovascular Business. “However, paroxysmal atrial fibrillation is a potential cause of cryptogenic stroke and TIA [transient ischemic attack]."
Paroxysmal AF is often missed during initial workups because patients may no longer be experiencing an AF episode during the visit. “If you are not looking for paroxysmal atrial fibrillation, you won’t find it,” said Miller, a senior staff neurologist at Henry Ford Hospital in Detroit.
Two previous studies have shown mobile cardiac outpatient telemetry (MCOT) monitoring to detect high rates of paroxysmal AF in cryptogenic stroke patients. For example, in 2008, Tayal et al found a 23 percent rate of AF in cryptogenic stroke patients who were monitored with over 21 days of MCOT (Neurology 2008:71;1696-1701). Miller et al set out to confirm previous results and evaluate the optimal duration of cardiac monitoring for this patient population.
Miller's retrospective study included 156 patients who had a cryptogenic stroke or TIA within six months. MCOT records were received for the 18 months between June 2009 and January 2011. At study enrollment, 97 percent of patients were not administered anticoagulants and 51 percent were not on any type of rate control medication.
Of the 156 patients, 50 percent were female, 49 percent were African American and 35 percent had a history of prior stroke or TIA. Seventy-nine percent of the index events were stroke.
“At what point do you stop getting benefit from continued remote monitoring?” Miller asked. Since Miller and colleagues’ goal was to find the optimal time point for monitoring, the authors looked at several different time periods: 48 hours, seven days, 14 days and 21 days. “We found statistically significant increases in rate detections between these time points,” Miller said.
Twenty-seven patients had AF on monitoring; however, those monitored for 21 days had the highest rate of AF detection—19.5 percent. The authors only reported a 3.8 percent rate of detection after 48 hours of monitoring, 9 percent after seven days of monitoring and 14.1 percent after 14 days of monitoring.
“Our study concludes that a minimum of 21 days of monitoring should be done in these patients to adequately assess for paroxysmal AF,” Miller said.
Interestingly, the researchers also found that five risk factors increased AF detection: female gender, a dilated left atrial heart chamber, a decreased left ventricular ejection fraction, higher stroke severity and premature atrial complexes. “While these risk factors were all found to be statistically significant on a Cox multivariate analysis, having a premature heart rate on the EKG [premature atrial complexes] carried the strongest development for AF,” Miller noted. In fact, this single risk factor carried a 13.7-fold increased risk of AF detection.
While these developments could help depict the causes of cryptogenic stroke, Miller said that the duration of AF episodes must be better defined. During the study, Miller et al found that two-thirds of AF episodes were less than 30 seconds long. “These episodes were short and this is a very important point,” Miller offered. “When you have shorter episodes of AF, it is not yet clear whether these will carry a higher risk of recurrent stroke.”
Previous evidence has shown that patients with shorter episodes of AF go on to have longer episodes, which could in the long run, turn out to be persistent AF. “While the single episodes of AF in our study are probably too short to cause stroke themselves, they may be markers for other AF episodes,” Miller concluded. "[P]rompt treatment of this disorder with appropriate thinning of the blood, otherwise known as anticoagulation, should decrease that risk by approximately 40 percent over treatment with aspirin alone."
These shorter paroxysmal AF episodes of less than 30 seconds must further be defined, as should monitoring these patients beyond 21 days, Miller summed.