Neurology. 2025 Dec 23;105(12):e214426. doi: 10.1212/WNL.0000000000214426. Epub 2025 Nov 24.
ABSTRACT
BACKGROUND AND OBJECTIVES: In comatose patients after cardiac arrest (CA), only "highly malignant" or "benign" EEG patterns are recognized as robust markers of poor and good outcomes, respectively, whereas many patients with "malignant" EEG remain in prognostic uncertainty. We aimed to assess the prognostic value of combined malignant EEG criteria for poor neurologic outcome in this clinical setting and to evaluate their effect on reduction of prognostic uncertainty.
METHODS: We conducted a bicentric cohort study using institutional registries (CHUV, Switzerland; Cochin Hospital, France), including comatose adults who underwent EEG >24 hours after CA. EEGs were categorized according to Westhall et al.: highly malignant, malignant, or benign. We assessed the prognostic performance of isolated and combined malignant EEG criteria (recorded on a single EEG after targeted temperature management) and their ability to reduce prognostic uncertainty in patients with an "indeterminate prognosis" according to current European recommendations. Poor neurologic outcome was defined as a Cerebral Performance Category 3-5 at 3 months.
RESULTS: Between 2017 and 2024, 1,076 patients were included (CHUV: 655; Cochin: 421). The mean age was 61.7 (SD ± 15.2) years, and 291 (27.2%) were female; 633 (59.5%) died while 317 (29.5%) had a good outcome. EEG was highly malignant in 23.6%, malignant in 39.1%, and benign in 36.8%. One malignant EEG feature predicted poor outcome (specificity 81.4% [95% CI 77.10-85.67]), sensitivity 24.2% ([95% CI 21.19-27.29]). Two, 3, or 4 malignant EEG criteria were highly specific (98.1% ([95% CI 96.61-99.61], 100% [95% CI 100.00-100.00], and 100% [95% CI 100.00-100.00], respectively), with limited sensitivity (16.9% [95% CI 14.20-19.53], 4.6% [95% CI 3.12-6.10], and 0.4% [95% CI 0.00-0.84]). Among the 725 lacking 2 criteria for poor outcome according to recommendations, 386 had a benign EEG, leaving 339 in the indeterminate outcome group. Among these, multiple malignant EEG features reclassified 106 patients into the "likely poor outcome" category, reducing prognostic uncertainty by 31.3%.
DISCUSSION: The combination of multiple malignant EEG criteria seems to be as specific as a highly malignant EEG in predicting poor neurologic outcome. Using these criteria significantly reduces prognostic uncertainty.
PMID:41284957 | DOI:10.1212/WNL.0000000000214426