After cardiac arrest (CA), epileptiform EEG, occurring in about 1/3 of patients, often but not invariably heralds poor prognosis. We tested the hypothesis that a combination of specific EEG features identifies patients who may regain consciousness despite early epileptiform patterns.
We retrospectively analyzed a registry of comatose patients post-CA (2 Swiss centers), including those with epileptiform EEG. Background and epileptiform features in EEGs 12-36 hours or 36-72 hours from CA were scored according to the American Clinical Neurophysiology Society nomenclature. Best Cerebral Performance Category (CPC) score within 3 months (CPC 1-3 vs 4-5) was the primary outcome. Significant EEG variables were combined in a score assessed with receiver operating characteristic curves, and independently validated in a US cohort; its correlation with serum neuron-specific enolase (NSE) was also tested.
Of 488 patients, 107 (21.9%) had epileptiform EEG <72 hours; 18 (17%) reached CPC 1-3. EEG 12-36 hours background continuity ≥50%, absence of epileptiform abnormalities (p < 0.00001 each), 12-36 and 36-72 hours reactivity (p < 0.0001 each), 36-72 hours normal background amplitude (p = 0.0004), and stimulus-induced discharges (p = 0.0001) correlated with favorable outcome. The combined 6-point score cutoff ≥2 was 100% sensitive (95% confidence interval [CI], 78%-100%) and 70% specific (95% CI, 59%-80%) for CPC 1-3 (area under the curve [AUC], 0.98; 95% CI, 0.94-1.00). Increasing score correlated with NSE (ρ = -0.46, p = 0.0001). In the validation cohort (41 patients), the score was 100% sensitive (95% CI, 60%-100%) and 88% specific (95% CI, 73%-97%) for CPC 1-3 (AUC, 0.96; 95% CI, 0.91-1.00).
Prognostic value of early epileptiform EEG after CA can be estimated combining timing, continuity, reactivity, and amplitude features in a score that correlates with neuronal damage.