Cryptogenic stroke is common, frustrating for both the patient and providers, and can be a tough nut to crack. Recent studies – namely CRYSTAL AF and EMBRACE – confirmed the long-standing hypothesis that some proportion of cryptogenic stroke was actually caused by occult atrial fibrillation, unmasked by implantable loop recorders. These data were paradigm shifting in my practice, compelling prolonged outpatient telemetry in patients with the unofficial but surely familiar designation of cryptogenic-suspect-embolic stroke. If and how these data affected practice patterns broadly speaking remains to be seen and published.



Some early data in this regard come from some German colleagues who surveyed stroke directors around the country to survey their practice as regards cryptogenic stroke. This survey was devoid of patient-level data, opting for estimations of practice patterns, but still provides valuable insight into stroke workup practice patterns.

The main findings from this survey of 179 stroke directors (~72% survey response rate) were the differences in duration and means of continuous cardiac monitoring as well as the great difference in duration of monitoring of patients with transitory acute neurovascular episodes (e.g., “TIA”) versus those with demonstrated ischemia. In their survey, two thirds of patients with ischemic stroke underwent >48h of cardiac telemetry whereas only one in five patients with TIA had that duration of monitoring. Furthermore, in contradistinction with the aforementioned encouraging data on loop recorders, utilization was very low in German stroke units (~1 in 10 patients) and very few had a defined mechanism for ECG or telemonitoring follow up even though most providers (~90%) identified that kind of follow up as important.

These data are interesting and potentially representative of North American practice: we all acknowledge that prolonged cardiac telemetry is likely to help us chip away at a fair proportion of cryptogenic strokes, but exactly how? Is a month long enough? Is an implantable recorder safe and effective enough to justify its use in this context? Will all of this be obviated by a trial of empiric NOACs after cryptogenic-suspect-embolic strokes? Wait…aren’t there other non-occult-atrial-fibrillation mechanisms – some yet undiscovered – that cause a cryptogenic stroke?!

Clear as mud, but this is a start.