Non-convulsive status epilepticus (NCSE) is an electro-clinical syndrome, with a wide spectrum of underlying aetiologies. The electroencephalogram (EEG) is a key investigative tool for NCSE, but diagnosis is often challenging. We conducted a retrospective audit from May 2017–February 2019 in the Clinical Neurophysiology Department at St. George’s University Hospitals NHS Foundation Trust, looking at all EEG-referrals for suspected NCSE. We ascertained whether referrals were appropriate based on semiology according to International League Against Epilepsy (ILAE) recommendations for NCSE¹; and whether reporting of EEGs adhered to Salzburg Consensus Criteria for NCSE². For EEGs reported as NCSE, we ascertained if patients were managed according to guidelines^3-4; whether follow-up EEGs were performed; and whether treatment of NCSE affected overall outcome.

A total of 183 referrals were made during this period. 18 were paediatric patients with reduced GCS (3 were neonates in SCBU, 6 in paediatric ICU). 70 were adult intensive care patients with GCS≤8 (15 had out-of-hospital cardiac arrest, 40 had other structural brain injury including cerebrovascular-event or traumatic brain injury, and 15 had low GCS without clear structural abnormality). 86 were adult non-ITU inpatients with low GCS. 9 referrals did not adhere to clinical criteria and were for acute confusional state or syncopal episodes.

Of the 183 referrals, 52 patients (28%) had EEG features which fulfilled or partially fulfilled the Salzburg criteria – 41 were intensive care patients. Of these, 27 were reported as definitive NCSE, while 25 were reported as possible NCSE that could not be excluded although another pathology was more likely (encephalopathy, structural brain abnormality with epileptic tendency, or encephalitis were suggested as more likely causes for these cases).

Out of the 52 cases, 36 patients were given a trial of intravenous benzodiazepine or loading dose of intravenous anti-epileptic drug. 24 patients had follow-up EEGs, and 22 showed resolution of epileptiform discharges. A total of 25 patients later passed away from their underlying illness, including all those whose EEG showed resolution of NCSE but remained in severe cerebral dysfunction.

We concluded that EEG reporting for NCSE is challenging due to variability in aetiologies as well as interpretation of the guidelines⁵. There was also variability in management of the patients. A large majority presented with coma in intensive care, whose overall clinical picture and prognosis appeared independent of resolution of EEG discharges. We plan to disseminate our findings to referring departments, instigate more stringent EEG-reporting criteria, and perform a re-audit.