Authors :
Presenting Author: Siddharth Jain, MBBS, MD – University of Louisville
Jeetendra Sah, MD – University of Louisville
Sean Woods, MD – University of Louisville
Feride Un Candan, MD – University of Louisville
Darren Farber, DO – University of Louisville
Cemal Karakas, MD – University of Louisville
Rationale:
The spike-wave index (SWI) is a crucial quantitative marker in pediatric epilepsy, particularly in assessing sleep-potentiated epileptiform discharges. However, the quantity of electroencephalography (EEG) time required to compute a reliable SWI has not been formally established. The primary objective of this study was to ascertain the minimum EEG review time necessary to estimate SWI within a margin of ±5% of its final value.
Methods:
Prolonged EEGs from 24 pediatric epilepsy patients were analyzed, with spike burden recorded every 10 seconds. Each EEG was divided into three segments, and two independent reviewers (DF and FC) scored each segment. SWI was calculated cumulatively across the recording. Stabilization time for each reviewer was defined as the earliest point at which the cumulative SWI remained within ±5% of the final value. For each patient, mean stabilization time was calculated from both reviewers, and standard deviation was used to represent inter-rater variability. Additional analyses included rolling coefficient of variation (CV), bootstrap-derived confidence intervals, and Bland-Altman agreement between early and final SWI estimates. Summary statistics across patients were used to determine threshold times for reliable SWI estimation.
Results:
The median stabilization time across patients was 455 seconds (7.5 minutes), with a 90th percentile of 900 seconds (15 minutes). The shortest observed stabilization time was 30 seconds, and the longest was 990 seconds. Inter-rater agreement was strong, with an average standard deviation of 6.6 seconds between reviewers. Figure 1 illustrates mean stabilization times per patient with error bars representing reviewer variability. Figure 2 demonstrates the distribution of stabilization times across all patients.
Conclusions:
Most patients obtain a reliable SWI estimate within 7–8 minutes of EEG review. However, to account for inter-patient variability, a conservative cutoff of 15 minutes may be necessary. These findings provide an evidence-based framework for optimizing the duration of EEG review required to assess SWI accurately and efficiently in clinical practice.
Funding: None