Accuracy of seizure reporting in patients with communication deficits

Inaccuracy in seizure reporting has been well-described in patients with epilepsy, with self-reporting at a rate of less than 50% compared to objective seizures on ambulatory EEG (aEEG). Although much of the inaccuracy has been attributed to the lack of self-awareness that occurs because of some seizures, the extent to which developmental delay, language impairment, or neurodegenerative disease contributes is not known. In these patients, clinical evaluation and management is determined by the seizure frequency as reported by family members or caregivers, rather than by the patients themselves. As such, reporting frequency may be markedly unpredictable in this subset of patients.

Clinical, radiographic, and electrophysiologic data were reviewed retrospectively for patients older than age two years with prior aEEG of >47 h, objective EEG evidence of epilepsy, and some degree of communication deficit between January 2013 and February 2020. Caregiver reported seizure frequency was determined on review from clinical documentation for comparison to objective seizure frequency on aEEG.

Fifty-seven patients, 28 of whom were below age 18, were identified. Eighteen patients had isolated speech impairment, 31 had cognitive impairment, and eight had a general encephalopathy. All except six of the patients were reporting seizures at the time of the ambulatory study. Eighteen patients (32 %) had seizures, and 39 patients (68 %) did not have seizures on EEG. The number of seizures captured ranged from one seizure to 312 seizures. Excluding two outliers who had >100 seizures, the average number of seizures in all patients was 1.05 (±2.50). The average number of seizures among those who had seizures on EEG was 3.62 (±3.55). Seventy-eight percent of the patients who had seizures experienced more seizures than reported in clinic, while 22 % had fewer seizures. Eighty-nine percent of the patients who had seizures and 59 % who had no seizures were ultimately deemed to have inaccurate reporting. Twenty-six of the 57 patients in the study (45 %) had events during the aEEG without EEG change. Nineteen of these 26 patients (73 %) were ultimately deemed to have inaccurate reporting.

Our findings support prior studies that have indicated that aEEG findings do not correlate well with reported seizure frequency, and this appears to hold true for patients with communication deficits. Additionally, even when caregivers are identifying seizures appropriately during the aEEG, this does not necessarily indicate that the frequency reported during clinic visits is accurate. Further prospective studies with improved means of self-reporting may provide more insight on the accuracy of aEEG in comparison with clinically appreciated seizures in patients with epilepsy.