International Stroke Conference (ISC)
February 12-14, 2014

Paroxysmal atrial fibrillation (PAF) is an elusive cause of stroke, but because anticoagulation is an effective intervention to prevent recurrent stroke in patients with AF, it is important to diagnose it. While most stroke patients undergo telemetry monitoring for several days during the acute stroke admission, PAF is not always detected, and long-term outpatient cardiac monitoring has been proposed as an important element in the evaluation of stroke etiology. Several studies presented at this year’s ISC looked at the issue, and while they shed some light on the issue, many questions still remain.



Two studies that used mobile cardiac outpatient telemetry (MOCT) for 25-28 days in patients with cryptogenic stroke found an incidence of PAF in 13%. Investigators in both studies looked for imaging and echocardiographic fetures that could identify patients who were most likely to benefit from MOCT. Among 132 patients (mean age 72 years), Omar Kass-Hout and his group from Emory University found that cortical lesions on MRI and higher mean left atrial diameter, lower tissue Doppler velocity, and higher left atrium volume index (LAVI) and mean LAVI/late diastolic Doppler velocity on echocardiography were associated with a diagnosis of PAF. Scott Kasner and colleagues at the University of Pennsylvania retrospectively analyses data from 227 patients (mean age 62 years). They did not find any echocardiographic factors associated with incident PAF (they only looked at 4 factors: LA size, ejection fraction, aortic arch atheroma and PFO) but noted that age >60 years and a prior infarct, particularly cortical or cerebellar, were associated with PAF on MOCT. It is possible that the Emory group had the echocardiographic associations because the patients were older or because the investigators looked at many more factors (repeat testing).

While Kass-Hout and Kasner studied patients with cryptogenic stroke, it is possible that MCOT should not be limited to patients with stroke of undetermined etiology. Some patients may have more than one potential cause for the stroke (vascular diseases tend to go together), and some question whether all patients should be evaluated with MOCT, independent of alternative identified etiologies. Dr. Ghazala Basir and her colleagues from the University of Alberta favor this approach because they found, in a study of 96 patients who had MOCT for 3 weeks, that 32% of those with cryptogenic stroke and 36% of those with other potential etiologies had PAF.

While looking for PAF using MOCT is an attractive strategy, several questions remain unanswered. Are the implications (in terms of stroke risk) of finding PAF on MOCT similar to those of chronic AF, as suggested by a study presented by Seiji Miura and colleagues from Kyushu Medical Center in Japan? Are all episodes of PAF equally embologenic, independent of duration? Rolf Watcher and colleagues from the University of Mainz in Germany compared the stroke severity among patients with different duration of PAF and found that patients with paroxysmal AF of at least 30 seconds during 7-day-Holter have strokes similar to patients with persistent AF. It is unclear whether runs of PAF shorter than 30 seconds confer an increased stroke risk.

While these studies tell us about the prevalence of PAF in patients with stroke, we do not know the baseline prevalence of PAF among patients without stroke, or even without vascular risk factors, and it is thus difficult to place the rates in these studies in context. We also do not know whether anticoagulants have a similar risk-benefit profile in patients with chronic AF and with PAF. DO we need a clinical trial to evaluate whether anticoagulation is warranted in these patients?

– José G. Merino