Real-time detection of neonatal seizures improves with on demand EEG interpretation

Abstract

ObjectiveUnder- and overtreatment of seizures may result in adverse outcomes; thus, early, reliable seizure identification is crucial. Continuous conventional ElectroEncephaloGram (cEEG) is the gold standard, but amplitude integrated EEG (aEEG) is most widely-used in the Neonatal Intensive Care Unit (NICU). We aimed to determine whether a novel pathway combining cEEG and aEEG for seizure detection would improve real-time seizure identification. MethodsA single-center, prospective quality improvement project cohort. Patients at-risk of seizures were monitored by cEEG and aEEG concurrently, with the option for the neonatologist to contact a neurologist for real-time cEEG interpretation. The primary outcome was correct identification of seizures by the new combined pathway compared to aEEG alone. ResultsSeizure detection using aEEG had a sensitivity of 46.2%; specificity of 93.3%; PPV of 66.7%; and NPV of 85.7%. Utilizing the new on-demand, real-time cEEG interpretation by a neurologist, correct seizure identification increased by 27% (95%CI: 10–56%). Over-diagnosis of seizures was avoided in 33.3% (95% CI: 12.1–64.6%) and misuse of anti-seizure medication was prevented. ConclusionsCombining aEEG with on-demand cEEG interpretation improved accurate seizure detection in a high-risk NICU population, with the potential to avoid over-treatment. SignificanceWe describe a novel combined EEG monitoring pathway to improve seizure detection, and prevent unnecessary treatment.