Abstracts

Rationale: In the treatment of infantile spasms (IS), successful response requires elimination of both spasms and hypsarrhythmia (HYPS). Confirmation of response is routinely accomplished with video-electroencephalography (VEEG) because (1) HYPS may persist despite resolution of spasms, and (2) subtle spasms may escape parental observation. However, the use of VEEG is subject to considerable variability; whereas some practitioners rely on < 1h recordings, others mandate 24-h VEEG. The ideal duration of VEEG to verify response is unknown.  Using a large cohort in which treatment responses were confirmed with overnight VEEG, we set out to determine whether shorter duration studies might adequately confirm responses to therapy.  Methods: Patients with VEEG-confirmed IS were retrospectively identified. For each patient, we identified the first post-treatment VEEG, tabulated clinical impression of response prior to VEEG, and determined latency to first spasm during the study, if present. As we were chiefly interested in quantifying false-negatives (patients with persistent spasms on the overnight VEEG, but no spasms on shorter VEEGs), we determined the sensitivity of shorter recordings, using overnight EEG as the gold standard. Sensitivity was defined as the proportion of patients with spasms observed in the first 1, 2, 4, 8, 12, or 18 hours of the recording, among all patients with spasms on the entire overnight VEEG. Results: We identified 119 children with VEEG-confirmed IS and overnight VEEG following treatment. Persistent spasms and/or HYPS were observed in 65 (55%) cases. Of note, we frequently encountered cases in which the clinical impression of response was contradicted by VEEG: Among 44 cases with clinically suspected response, we found 7 (16%) children with persistent HYPS or subtle spasms, and among 35 cases with lack of clinical response, we found 4 (11%) with resolution of both spasms and HYPS on overnight VEEG. Among 40 cases with uncertain clinical response, 27 (68%) exhibited spasms and/or HYPS. Although HYPS was described as intermittent in some cases, a specific time for first appearance of HYPS patterns was never documented. In contrast, the exact time of first spasm was identified in 30 cases. Considering only those patients with persistent spasms and no HYPS (n=27) on follow-up VEEG, the first spasm occurred after a median of 176 min (range 3.0 min – 23.1 h). Using just the first 1, 2, 4, 8, 12, or 18 hours of the recording to identify persistent spasms, sensitivity was 19%, 37%, 67%, 74%, 78% and 96%, respectively, and did not vary according to clinical impression of response prior to EEG. Conclusions: Given a high rate of clinical-VEEG discordance, this study supports the view that VEEG should be implemented to assess response to therapy for infantile spasms. If HYPS is not observed at the beginning of a recording, longer—and especially overnight—VEEG duration confers incremental benefit. Further study in a larger cohort, with blinded review of actual VEEG data is warranted to confirm these findings. Funding: None