Evaluating diagnostic yield of ambulatory EEG for interictal epileptiform discharges and seizures

Ambulatory EEG (aEEG) can be high yield, accessible, and cost-effective for prolonged outpatient monitoring, but it is unclear how patient and clinical characteristics can be used to select patients for whom aEEG may be diagnostic. To this end, this study aims to identify factors associated with diagnostic yield for interictal epileptiform discharges (IEDs) and seizures.

We evaluated all ambulatory EEGs from a Comprehensive Epilepsy Center over a 6-month period performed on adults. Using logistic regression and recursive feature elimination (RFE), we evaluated how patient and clinical characteristics were associated with observing IEDs or seizures.

AEEG was performed for 179 patients (average age 45, 58% female). The median aEEG duration was 48 h (range 18.5 to 86 h). Diagnostic yield was 35% for IEDs and 10% for seizures. IED observation was more likely when a prior aEEG was abnormal (Odds Ratio [OR] 3.37, p = 0.039) or with more concurrent antiseizure medications (OR 1.45/medication, p = 0.007). Seizure observation was more likely on 72-hour studies (OR 26, p = 0.003) predominantly within the first 48 h, with higher seizure frequency (OR 1.71 per log seizure/month, p = 0.034), and with IEDs observed in prior epilepsy monitoring unit admission (OR 22, p = 0.004).

This study identified factors associated with diagnostic yield of aEEG for observing IEDs and seizures. At least 48-hours of aEEG was highest yield for seizures, and longer studies may be appropriate for people with lower seizure frequency. IEDs were less dependent on study length and were more associated with pre-aEEG certainty of epilepsy diagnosis.