Value of clinical features to differentiate refractory epilepsy from mimics: a prospective longitudinal cohort study

Misdiagnosis of refractory epilepsy (rE) is common and such patients experience a long diagnostic delay. Our aim was to identify key clinical/laboratory factors in order to obtain an alternative diagnosis in patients referred for rE.

Between January 2010 and December 2015, 125 consecutive patients with a diagnosis of rE were prospectively enrolled. All patients underwent a comprehensive neurological, neuropsychiatric and cardiological evaluation, and had an observation time of at least 1 year after the study entry.

Diagnosis of rE was confirmed in 104/125 (83.2%) patients (55 women, mean age 38.8 ± 14.3 years). Thirteen/125 patients (10.4%, seven women, mean age 50.8 ± 20.9) were diagnosed with syncope, which was cardiac/cardio inhibitory in 9/13 (69%). The remaining 8/125 patients (6.4%, six women, mean age 41.2 ± 14.6 years) were diagnosed with psychogenic non-epileptic seizures. Age at onset had a high accuracy in differentiating patients with syncope from others, with the best cut-off age at 35 years and above. Abnormal brain magnetic resonance imaging (MRI) had a significant yield of about 70% in rE. A diagnostic model including age at onset and brain MRI was highly accurate in differentiating patients with syncope from others. In patients with cardiac/cardio inhibitory syncope, the point score of historical features was ≥1 and falsely favoured the diagnosis of epileptic seizures.

This prospective cohort study identifies rE mimics who are at high risk of morbidity and mortality. rE starting in adulthood should raise a high suspicion of cardiac syncope. Brain MRI is accurate in differentiating rE from other conditions.