Abstracts

Rationale: Psychogenic non-epileptic seizures (PNES) are diagnosed in ~25% of patients admitted to Epilepsy Monitoring Units (EMU) but there is relatively little information on outcome following diagnosis. Previous studies have focused on seizure frequency and antiepileptic drug utilization. However, somatoform disorders often have broad health care impact, not captured by seizure measures alone. We examined long-term health care utilization outcomes in Veterans diagnosed with PNES vs. epileptic seizures (ES). Methods: A structured chart review included all patients meeting research criteria for PNES or ES at the Portland, Oregon VAMC Epilepsy Monitoring Unit (2005-2010), and with at least 3 years of continuous records before and 3 years after the diagnostic admission. We reviewed 3 sources; (1) the electronic medical record; (2) linked medical records from other VAMCs; (3) scanned medical records from non-VA providers. For each patient we abstracted (a) all Emergency Department (ED) visits and the primary reason for each visit; (b) the number and length of hospitalizations and the primary reason for each hospitalization; (c) the number of outpatient (OP) visits to primary care, neurology, or psychiatry. Results were compiled separately for the first year post-diagnosis (months 1-12), and months 13-36, and compared to the same intervals pre-diagnosis. We also abstracted demographic, seizure, and psychiatric characteristics. Results: Twenty PNES patients and 25 ES patients met inclusion criteria. All cause ED visits and hospital admissions (events) were more common in PNES vs. ES patients during the 3 years following diagnosis (median 3.5 vs. 1 event; p<0.05). Events decreased during the first year following PNES diagnosis (p<0.05 as compared to the 1^st year pre-diagnosis) but then increased with no overall change for the 3 years post-diagnosis as compared to the 3 years pre-diagnosis. Pain complaints (rather than seizures) were the most common presenting symptom during 3 years post-diagnosis and were more common than in ES patients (p<0.01). Medically unexplained symptoms (chest pain, abdominal pain) were also more common in the PNES group (p<0.05). Seizure presentations decreased from pre-diagnosis levels (ns). Total hospital days were greater in PNES vs. ES patients both before and after diagnosis (p=0.02). Forty percent of all hospital days in PNES patients were for psychiatric care. Among psychiatric factors a diagnosis of personality disorder had the strongest effect on all cause events during 3 years of follow-up (p<0.01). Conclusions: Following EMU diagnosis of PNES and conventional treatment there is a transient reduction in ED visits and hospitalizations followed by a return to pre-admission levels. Pain complaints are the most common presenting symptoms and are often unexplained, whereas seizure presentations decrease mildly. These results emphasize the heterogeneous medical complaints in PNES patients and the need for comprehensive long-term treatment strategies.