Abstracts

Rationale: Functional seizures (FS) are disabling paroxysmal neurological symptoms that, most commonly, are associated with acute and chronic biopsychosocial stressors and are otherwise known as psychogenic nonepileptic seizures (PNES). Often, FS are initially diagnosed as epileptic seizures (ES), which are disabling paroxysmal neurological symptoms caused by hyperexcitable and hypersynchronous electrical activity. Due to the excess mortality in ES from Sudden Unexpected Death in Epilepsy (SUDEP) and other causes, there is a perception that accurately identifying FS is “good news.” However, there is emerging evidence for similar excess mortality in FS. In this study, we used electronic health records from a large healthcare system to directly compare excess mortality rates in FS and ES, while controlling for other factors that may contribute to mortality rates.

Methods: Based on electronic health records, we identified patients over 18 years old who had a clinical encounter at any location of the University of Pittsburgh Medical Center with an International Classification of Disease Codes Version 10 diagnosis of ES (G40), FS (F44.5), or mixed ES and FS. Patients were considered to possibly have mixed ES and FS if there was at least one code for ES (G40) that occurred after their initial code of FS (F44.5). For each group, we evaluated the time from initial diagnosis until last follow up or death using Cox-Proportional Hazards analysis controlling for confounding factors as well as a standardized mortality rate (SMR).

Results: We identified 29,655 patients with ES only (53.9% female; age mean 54.5 interquartile range [IQR] 37-70), 831 with FS only (80.0% female, age mean 41.5, IQR 29-52), and 658 with mixed (74.9% female; age mean 46 IQR 34-57). The median duration of follow up was 4.0 years (IQR 1.7-7.2). In 140,384 patient-years of observation, there were 4,012 deaths in ES, 27 deaths in FS, and 22 deaths in mixed. The raw mortality rate was higher in ES (29.6 per 1,000 patient-years) compared to FS (10.4 per 1,000 patient-years) and mixed (9.8 per 1,000 patient-years). The SMRs for each condition were similar (ES: 2.38 [95% Confidence Interval (C): 2.31-2.46], FS: 2.37 [95% CI 1.56-3.48], mixed: 1.77 [95% CI 1.11-2.72]). The Cox-Proportional Hazards analysis further demonstrated a lack of a significant difference in mortality associated with FS as compared to ES (Hazard Ratio 0.73, 95% CI 0.50-1.06, p=0.10), after accounting for confounding factors.

Conclusions: This work further demonstrates that the excess mortality associated with FS was not less than ES, when accounting for underlying differences in these populations. In comparison to ES, the lower raw mortality rate in FS was associated with differences in non-seizure factors. Similar excess mortality in FS as compared to ES emphasizes that patients with FS also warrant treatment and other interventions aimed to reduce or prevent death. Future work is needed to identify, evaluate, and implement interventions to address this excess mortality.

Funding: This work was supported by the American Epilepsy Society, the Epilepsy Foundation, the American Academy of Neurology, and the American Brain Foundation.