MR-imaging pattern is not a predictor of occult atrial fibrillation in patients with cryptogenic stroke

Abstract

Background: To date, insertable cardiac monitors (ICMs) are the most effective method for the detection of occult atrial fibrillation (AF) in cryptogenic stroke. The overall detection rate after 12 months, however, is low and ranges between 12.4 and 33.3%, even if clinical predictors are considered. Ischemic stroke patients due to cardiogenic embolism present with particular lesion patterns. In patients with cryptogenic stroke, MR-imaging pattern may be a valuable predictor for AF. Methods: This is an MRI-based, retrospective, observational, comparative, single-center study of 104 patients who underwent ICM implantation after cryptogenic stroke. The findings were compared to a reference group with related stroke etiology, i.e., 166 patients with embolic stroke due to AF detected for the first time by long-term ECG. Lesion patterns were evaluated with regard to affected territories, distribution (cortical, lacunar, scattered), lesion volume, and lesion size (diameter of the lesion size > 20 mm). Results: The MR-imaging analysis of acute ischemic lesions yielded no association between AF and lesion size or volume, arterial vessel distribution, or the number of affected territories. There was no significant difference between the cohorts regarding ischemic patterns (cortical lesions, scattered lesions, and lacunar infarcts). An important clinical inference of our findings is that 10% (2 of 20) of cases in the ICM group in whom AF was detected had a lacunar infarct pattern. Similar results were shown in cases of ischemic stroke patients with AF detected for the first time by long-term ECG, with 10.9% (16 of 147) of them showing lacunar infarcts. The analysis of chronic MRI lesions revealed no differences between the groups in the rate of chronic lesions, arterial vessel distribution, or the number of affected territories. Left atrial size (LA size) and the presence of atrial runs in long-term ECG were independently associated with AF. Conclusions: In this MRI-based analysis of patients with cryptogenic stroke who had received ICM implantation, the detection rate of AF in patients with ICM was not related to the imaging pattern. In addition, the lacunar infarct pattern should not be an exclusion criterion for ICM insertion in patients with cryptogenic stroke. ICM insertion in patients with cryptogenic stroke should not be evaluated solely on the basis of reference to infarct patterns.