Abstracts

Rationale: The accuracy of patient seizure frequency tracking is crucial to understanding seizure trends and treatment response. The Human Epilepsy Project 2 (HEP2) is a prospective, observational, multicenter US study with the goal of identifying treatment response in participants with Focal Treatment Resistant Epilepsy (FTRE). Here, we evaluated how much seizure frequency aligned when estimated with a global retrospective estimate, retrospective daily diary entries, or prospective daily dairy entries.

Methods: All participants had FTRE, failed at least 4 ASMs, and were receiving at least 1 ASM. At enrollment, subjects provided pre-enrollment seizure counts for at least the two preceding months, and a global estimate of baseline frequency. Prospective seizure tracking involved daily entries via electronic seizure diary, monthly check-ins with coordinators, and medical record review. We analyzed subjects who provided at least 2 months of retrospective daily seizure counts, 3 months of continuous post-enrollment daily seizure counts without corrections from medical record review and had no medication changes in the times that were analyzed. Patients who corrected their diaries during check-ins or record review were excluded to ensure methods of tracking were comparable. Estimated baseline, retrospective, and prospective seizure frequencies were analyzed using rank correlation (Spearman’s Rho) and percent differences.

Results: 49/146 of subjects had no corrections based on medical record review in the post-enrollment window, and no medication changes in the pre- and post-enrollment windows. 35/49 provided at least 2 months of pre-enrollment seizure counts, and all 49 provided estimated baseline frequencies. The correspondence between the estimated baseline frequency and pre-enrollment daily seizure counting was 91%. The correspondence between pre-enrollment and post-enrollment daily seizure counting was 63%. The correspondence between estimated baseline frequency and post-enrollment daily seizure counting was 66%. When compared to pre-enrollment daily seizure counts, 14/35 (40%) had at least 25% lower post-enrollment daily seizure counts, 11/35 (31.4%) had more than 25% higher counts, and 10/35 (29%) were within 25%. When compared to the global estimate of baseline seizure frequency, 8/49 (16%) had at least 25% lower counts by 25%, 24/49 (49%) had more than 25% higher counts, 17/49 (35%) were within 25%.

Conclusions: Estimated baseline and pre-enrollment tracking demonstrated a strong correlation, implying patients’ estimated baseline frequencies closely followed their retrospective tracking. The natural variability of seizure counts was highlighted by a 63-66% correlation between pre- and post-enrollment seizure counts. There was no systematic increase or decrease in seizure frequency when tracking occurred retrospectively as compared to prospectively. Future directions involve analyzing how often monthly check-ins and record-review caused revisions in seizure counts, and in which direction.

Funding: The HEP2 study was funded by UCB, Neurelis, & SK Life Sciences.