Abstracts

Rationale: The gold standard in the diagnosis of psychogenic nonepileptic seizures (PNES) is video EEG. One challenge that patients have with the diagnosis is that they perceive PNES to be a diagnosis of exclusion, which it is not. Scales administered by neuropsychologists such as the Personality Assessment Inventory (PAI) have been validated as supportive data in the diagnosis of PNES; specifically, if scales of somatization are elevated. Females have a higher reported incidence of PNES and may differ in presentation of symptoms from males. What has yet to be established is how to apply the PAI subscales separately for men and women, in order to improve the accuracy of diagnosing PNES. Methods: Patients admitted to the Epilepsy Monitoring Unit (EMU) at Mayo Clinic Hospital in Arizona are routinely administered personality testing, such as the PAI, as part of comprehensive neuropsychological evaluation. Patients admitted between January 2008 and March 2014 for spell classification were considered for the analysis. The dataset included demographics (including gender), PAI scores (including subscale scores), as well as final diagnosis based on results of video EEG (epilepsy only, PNES only, Epilepsy + PNES, other physiological diagnosis, or indeterminate). Only patients with PNES only or epilepsy only were included. 6 were excluded as they completed only the short form of the PAI, which does not provide subscale data. Patients with an invalid PAI protocol due to random responding (n=16) were excluded. In the case of multiple admissions, only the first diagnostic admission was retained for analysis. The T-scores from the somatization scale (SOM) and the conversion somatization subscale (SOM-C), separated into females and males, were collected. Analysis was performed to determine the optimal cutoff score to differentiate PNES from epilepsy in each group. Sensitivities and specificities were calculated for each cutoff score. Results: A total of 646 patients admitted for spell classification successfully completed the PAI during their EMU stay during the defined time period. After excluding those patients as defined above, a total of 387 patients were included in the analysis. 181 (females= 125, males=56) patients with the diagnosis of PNES, and 206 (females=108, males=98) with the diagnosis of epilepsy. The table summarizes the data. To support the diagnosis of PNES, the optimal cutoff on SOM was a T score of 70 for females and 67 for males. For SOM-C the optimal cutoff T-scores were 69 for females and 66 for males. At these cutoff scores, sensitivities and specificities were similar for males and females, more similar for the SOM-C subscale. Conclusions: The optimal cutoff score for males was lower for both the SOM and SOM-C scales on the PAI to support the diagnosis of PNES. This suggests that there is a lower threshold to support a PNES diagnosis of males from the PAI, than for women. This exemplifies that men and women express PNES symptoms differently on personality testing, as well as a need to take into account gender when interpreting the PAI.