Penumbra volume predicts unfavorable outcome in patients with acute minor stroke or transient ischemic attack

Background: A subgroup of patients with acute minor stroke (AMS) or transient ischemic attack (TIA) become disabled due to disease progression (DP) or recurrent stroke within 3 months. The aim of this article is to identify the risk factors for DP in AMS/TIA patients. In the literature, no studies focused on computed tomography perfusion (CTP) in AMS/TIA patients at the acute stage. Methods: This retrospective study included patients with AMS or TIA (onset of symptoms <= 4.5 hours, baseline National Institutes of Health Stroke Scale [NIHSS] score of 0-4). DP was defined as a deterioration of NIHSS score of >= 2 points during hospitalization or modified Ranking Scale >= 2 at 3-month follow-up. Clinical data and imaging results were retrieved and measured for statistical analysis. Results: From 2011 to 2017, total 135 patients were eligible for further analysis: 28 patients (20.7%, DP group) and 107 patients (79.3%, non-DP group). The DP group had significantly higher larger penumbra volumes (p = 0.028). In univariate model of the logistic regression, patients with the following risk factors tended to have unfavorable outcome: female gender, higher HbA1c, chronic kidney disease stage >= 3b, intracranial atherosclerosis, and penumbra volume were associated unfavorable outcome, but larger deadcore volume was not. In further multivariate analysis, only penumbra volume >5 cm(3) (p = 0.049, odds ratio [OR] = 3.21, 95% CI: 1.00-10.27) had the statistical significance. The cut-point value of the penumbra volume for unfavorable outcome in AMS/TIA patients was 4.73 cm(3). Conclusion: One fifth of the AMS/TIA patients had unfavorable outcome at 90 days. In CTP performed within 4.5 hours after the onset of AMS/TIA, the penumbra volume (>5 cm(3)) was a significant risk factor for DP, and the cut-point value was 4.73 cm(3). Further studies could be designed to involve this subgroup of patients for more aggressive treatment.