Cryptogenic stroke: clinical consideration of a heterogeneous ischemic subtype.

Background: Cryptogenic stroke can be subdivided into 3 distinct categories: stroke of no determined cause (CyNC), stroke due to multiple etiologies (Cy >1), and stroke etiology unclear due to incomplete evaluation. Although these subdivisions may be very different from one another with respect to baseline features and outcomes, they are often reported as a composite group in clinical trials.

Methods: Patients treated at our academic institution between July 2008 and June 2013 for acute ischemic stroke were retrospectively assessed in our prospective registry. CyNC and Cy >1 patients were compared to other Trial of Org 10172 in Acute Stroke Treatment (TOAST) stroke subtypes and to each other using univariate analyses and multivariate logistic regression. The primary outcome of interest was good functional outcome, defined as a discharge modified Rankin Scale score of 0-2.

Results: Of the 1311 included patients, 260 (19.8%) experienced a CyNC and 49 (3.7%) experienced a Cy >1. Cy >1 classification was associated with history of atrial fibrillation (odds ratio [OR], 3.17; 95% confidence interval [CI], 1.16-6.12; P = .001). In comparison to other TOAST classifications, CyNC strokes were more likely to have good functional outcome (OR, 1.97; 95% CI, 1.38-2.82; P < .001) after adjusting for baseline National Institutes of Health Stroke Scale, admission glucose, age, and intravenous tissue plasminogen activator (IV tPA).

Conclusions: Even after adjusting for higher IV tPA treatment rates, ischemic stroke patients with no identified cause had better outcomes than other TOAST groups. Conversely, patients coded as cryptogenic with more than 1 likely cause represent a different patient subpopulation. These data argue against the consolidation of cryptogenic stroke subcategories in future investigations.