Abstracts

Rationale: Nonconvulsive seizures (NCSz) are defined as any seizures without prominent motor activity diagnosed with electroencephalography (EEG). Previous studies suggest that clinical signs such as ocular movement abnormalities, facial/periorbital twitching, or presence of remote risk factors for seizures (such as history of stroke, tumor, neurosurgery, dementia, or meningitis) are sensitive for presence of NCSz on EEG. We sought to explore the clinical utility of these signs for identifying NCSz.

Methods: This is a case-control study obtained from a retrospective review of 432 patients that were admitted at the University of North Carolina (UNC) Medical Center and UNC Rex Hospital. Patients over 18 years of age that underwent continuous video EEG (cEEG) were included. We excluded patients admitted for neurological diagnoses. A total of 27 patients with seizures on EEG were found (Sz-EEG). We matched 27 controls to Sz-EEG patients by demographics and EEG duration ± 12 hours (NoSz-EEG). We recorded elements of patient’s history such as neuroimaging findings, medical diagnoses, remote risk factors for seizures, seizure-like activity on presentation including clinical findings previously shown to be sensitive for presence of NCSz, elements of neurologic exam, and treatment with anti-seizure medications (ASMs). Descriptive statistical analysis, including Chi squared and Fisher’s Exact tests, was used.

Results: The majority of patients were admitted to the medical ICU (n=39, 72.2%). There were no differences between groups in elements of patients’ histories. Clinical activity on presentation or on neurological exam, including ocular movement abnormalities, facial/periorbital twitching, extremity movement abnormality, and episodic behavior change were not different between groups. Positive and negative likelihood ratios for these clinical signs also generally fell in the 1 to 2 or 0.5 to 1 range respectively, indicating these signs are not clinically useful markers for presence of seizure on EEG. Prior to EEG monitoring, half of Sz-EEG and NoSz-EEG patients were started on ASMs (48.1% and 59.2%, respectively). Almost all (92.6%) of Sz-EEG patients were on ASMs after EEG monitoring. Interestingly, all but one patient (n=15, 93%) in the NoSz-EEG group were continued on ASMs despite negative EEG results for seizures. There was higher in-hospital mortality in the Sz-EEG group vs NoSz-EEG group (p=0.002). See attached tables.

Conclusions: Contrary to previously reported, our matched case-control study showed that incidence of ocular movement abnormalities, facial/periorbital twitching, and presence of remote risk factors for seizures are not clinically useful markers for presence of NCSz. This suggests that in the hospitalized majority critically ill population, these clinical signs cannot reliably stratify risk for nonconvulsive seizures. Patients with NCSz had significantly higher mortality, and 93% of the patients without NCSz remained on ASMs. Further study of influence of EEG findings on treatment and outcomes of patients with NCSz will better inform care of this population.

Funding: This work was assisted by the North Carolina Translational and Clinical Sciences Institute supported by NCATS NIH Grant Award #UL1TR002489.