Review and update of the concept of embolic stroke of undetermined source.

Ischaemic strokes have traditionally been classified according to the TOAST criteria, in which strokes with unclear aetiology are classified as cryptogenic strokes. However, the definition of cryptogenic stroke did not meet the operational criteria necessary to define patient populations for randomized treatment trials. To address this problem, the concept of embolic stroke of undetermined source (ESUS) was developed and published in 2014. A hypothesis that underpinned this concept was that most strokes in patients with ESUS are caused by embolic events, perhaps many cardioembolic, and that anticoagulation would prevent secondary ischaemic events. On this basis, two large randomized trials were conducted to compare the non-vitamin K antagonist oral anticoagulants (NOACs) dabigatran and rivaroxaban with aspirin. Neither NOAC was superior to aspirin in these trials, although subgroups of patients with ESUS seemed to benefit specifically from anticoagulation or antiplatelet therapy. The neutral results of the trials of anticoagulation and insights into ESUS from research conducted since the concept was introduced warrant reassessment of the ESUS construct as a research concept and a treatment target. In this Review, we discuss the evidence produced since the concept of ESUS was introduced, and propose updates to the criteria and diagnostic algorithm in light of the latest knowledge.