Abstracts

Rationale: Patients with epilepsy often have cognitive impairments in different domains. Contributing factors can include the underlying cause of the epilepsy, such a brain injury, the effects of repeated seizures, medication, or comorbid psychiatric disorders. Conventional neuropsychometric assessments are time and labour intensive, limiting their broad use in routine epilepsy practice. There are few reports of the neuropsychological profile of large populations of epilepsy patients assessed using neuropsychological screening tools. Methods: The Neuropsychiatry Unit Cognitive (NUCOG), a brief bed-side tool assessing attention, visuoconstructional, memory, executive and language functioning, has been routinely administered to patients admitted to the Royal Melbourne Hospital in-patient Video-EEG Monitoring (VEM) unit since 2007.1 Here the data collected from February 2005 to February 2017 was retrospectively assessed. Patients were classified with respect to seizure etiology (epileptic seizures or non-epileptic seizures (NES)), epilepsy syndrome (focal or generalized epilepsy; temporal lobe, extra-temporal lobe epilepsy or uncertain classification), seizure frequency, anti-epileptic drugs (AEDs) and psychiatric comorbidities (anxiety and/or depression or no psychiatric condition). Repeated measures ANOVA was performed with NUCOG scores as output variable and the different group classifications as predictors, by controlling for age.  Results: A total of 810 patients aged 12-77 years were included in the analysis. Patients with NES performed significantly better than patients with epilepsy (P = 0.002) in total NUCOG scores and in the subscales memory, language, visuoconstructional and executive functioning specifically. Bilateral temporal lobe seizure localization was associated with poorer performance compared to both right-sided and left-sided temporal lobe epilepsy (P = 0.003) in total NUCOG score and in the attention, memory, language and visuoconstructional domains. Statistically significant differences were not observed between patients with focal and generalized epilepsies, nor between patients with temporal, extra-temporal and unknown classification, nor between patients with anxiety and/or depression and those without psychiatric conditions. Patients who had 1-3 seizures a month had lower scores than patients who had 4-11 seizures a year (P = 0.038) for total score and in attention and language domains. Patients treated with valproate, phenytoin, and topiramate scored lower than patients on other AEDs (P < 0.001). Conclusions: The study shows that the NUCOG is a practical bedside tool for characterizing cognitive deficits in groups of patients with different seizure etiologies. Differences in seizure etiology, temporal lobe laterality, as well as seizure frequency and medication are factors that affect cognitive performance as measured with this tool.   Funding: No funding to declare