Abstract Background and aims Electroencephalography (EEG) is a sensitive marker at the bedside of the tissue at risk. Previous quantitative EEG (qEEG) studies often averaged bilateral/frontal activity, focused on thrombectomy, and were recorded at 24 h. We tested whether a lesion-directed alpha-to-delta ratio (lesion-ADR) within 72 h after reperfusion predicts malignant cerebral edema (MCE). Methods In a prospective single-center cohort of anterior-circulation ischemic stroke treated with intravenous alteplase, endovascular thrombectomy, or both, a 7-minute EEG was obtained ≤72 h after reperfusion. Lesion-ADR was the mean ADR on the ipsilesional electrodes; the reperfusion-to-EEG interval (ΔT) was recorded. The primary outcome was MCE. Secondary outcomes were a favorable function at discharge and 90 days. Results Forty-nine patients were included. For MCE, the combined lesion-ADR+ΔT model achieved an AUC of 0.888 (95% CI 0.765–0.960); p* 0.18 produced 86% sensitivity and 88% specificity. For discharge independence (mRS ≤2; n=9), the ADR-only model showed an AUC of 0.899 (cross-validated AUC 0.85); p* 0.77 gave 88.9% sensitivity and 82.5% specificity and remained independently associated (aOR 18.63; p=0.016). For 90-day independence (n = 13), the ADR-only AUC was 0.759 (95% CI 0.615–0.869); p* 0.912 gave 46.2% sensitivity and 97.2% specificity; including ADR increased the cross-validated AUC to 0.912. Conclusions A 7-minute ADR directed at the lesion enables rapid prognostication after reperfusion: p* 0.18 (ADR+ΔT) indicates the risk of MCE for intensive monitoring, while p* 0.77 (ADR-only) identifies patients who are likely to leave the hospital with minimal disability. Conflict of interest Jin-An Huang: nothing to disclose.
Huang et al. (Fri,) studied this question.