Objective: In this retrospective study, we aimed to evaluate the sensitivity and negative predictive value of long-term EEG (L-EEG) in patients being assessed for epilepsy, who had already undergone non-specific standard EEG(s) (S-EEG). Secondary endpoints of this study were: (1) the correlation of non-specific changes on EEG with epileptiform patterns on L-EEG; and (2) the correlation of clinical parameters such as subjective frequency of seizures or epileptogenic lesions on cerebral imaging with epileptiform changes on L-EEG.

Methods: We retrospectively analysed clinical and electrophysiological data of 75 patients, assessed for epilepsy at the University Hospital Zurich, who had undergone an L-EEG for at least 48 hours, between 2010 and 2015. All patients had already undergone S-EEG(s) before L-EEG, which showed no epileptic changes. Furthermore, the association with clinical parameters, such as frequency of presumptive seizures, abnormalities on standard-EEG, AED intake and cerebral imaging with the final diagnosis, was analysed.

Results: Out of 75 patients, 14 (19%) patients were finally diagnosed with epilepsy. In eight of these patients, L-EEGs showed typical ictal/interictal patterns, with a sensitivity of 57% and negative predictive value of 91%. Neither the subjective frequency of seizures nor potentially epileptogenic lesions on cerebral imaging were associated with a positive epilepsy diagnosis.

Significance: In this preselected cohort of patients, who had already undergone a non-diagnostic S-EEG, the sensitivity of L-EEG remained considerable. Nonetheless, our study also revealed a significant false-negative rate. Based on the high negative predictive value in this study, L-EEG appears to be most useful at excluding epilepsy. Nevertheless, thorough evaluation of seizure history and clinical findings remain crucial for a reliable diagnosis.