Nonconvulsive status epilepticus in neurocritical care: A critical reappraisal of outcome prediction scores.

ABSTRACT

OBJECTIVE: Nonconvulsive status epilepticus (NCSE) is a frequent condition in the neurocritical care unit (NCCU) patient population, with high morbidity and mortality. We aimed to assess the validity of available outcome prediction scores for prognostication in an NCCU patient population in relation to their admission reason (NCSE vs. non-NCSE related). METHODS: All 196 consecutive patients diagnosed with NCSE during the NCCU stay between January 2010 and December 2020 were included. Demographics, Simplified Acute Physiology Score II (SAPS II), NCSE characteristics, and in-hospital and 3-month outcome were extracted from the electronic charts. Status Epilepticus Severity Score (STESS), Epidemiology-Based Mortality Score in Status Epilepticus (EMSE), and encephalitis, NCSE, diazepam resistance, imaging features, and tracheal intubation score (END-IT) were evaluated as previously described. Univariable and multivariable analysis and comparison of sensitivity/specificity/positive and negative predictive values/accuracy were performed. RESULTS: A total of 30.1% died during the hospital stay, and 63.5% of survivors did not achieve favorable outcome at 3 months after onset of NCSE. Patients admitted primarily due to NCSE had longer NCSE duration and were more likely to be intubated at diagnosis. The receiver operating characteristic (ROC) for SAPS II, EMSE, and STESS when predicting mortality was between .683 and .762. The ROC for SAPS II, EMSE, STESS, and END-IT when predicting 3-month outcome was between .649 and .710. The accuracy in predicting mortality/outcome was low, when considering both proposed cutoffs and optimized cutoffs (estimated using the Youden Index) as well as when adjusting for admission reason. SIGNIFICANCE: The scores EMSE, STESS, and END-IT perform poorly when predicting outcome of patients with NCSE in an NCCU environment. They should be interpreted cautiously and only in conjunction with other clinical data in this particular patient group.