Unknown etiology group of late onset epilepsy more likely to show epileptiform EEG abnormalities than cerebrovascular disease group

Objective

To determine whether the cerebrovascular disease (CVD) group of late-onset epilepsy shows lower detection rates of interictal and ictal epileptiform abnormalities (EAs) using routine EEG than other etiology groups.

Methods

We retrospectively reviewed consecutive patients who had first seizure at age 65 years or older. The detection of EAs was investigated as well as background factors including age at EEG, timing of EEG after the last seizure, recording time of EEG, EEG-vigilance, and antiseizure medications. Multivariable logistic regression analysis was used to compare the detection rate of EAs between etiologies.

Results

EAs were detected in 53 (27 %) of 196 patients with late-onset epilepsy: 17 (20.7 %) of 82 patients with CVD, 9 (29.0 %) of 31 with traumatic brain injury, 4 (23.5 %) of 17 with other structural etiologies, 1 (20.0 %) of 5 with metabolic etiology, 9 (28.1 %) of 32 with dementia, and 13 (44.8 %) of 29 with unknown etiology. The odds ratios for EA detection, using the CVD group as the reference, were 1.60 for traumatic brain injury, 1.32 for other structural etiologies, 0.98 for metabolic etiology, 1.57 for dementia, and 4.10 for unknown etiology. The unknown etiology group showed significantly higher detection rate than the CVD group (p = 0.010).

Conclusion

The higher detection rate in the unknown etiology group may indicate the presence of latent etiologies with high epileptic activity, while the lower detection rate in the CVD group may reflect electrophysiologically inactive lesions.