Authors :
Presenting Author: James Poe, B.S.2025 – Yale University
Hal Blumenfeld, MD, PhD – Professor, Yale University, Yale University; Arthur Cukiert, MD – Sao Paulo Epilepsy Clinic; Adithya Sivaraju, MD – Associate Professor, Department of Neurology, Yale University
Rationale:
Patients with Lennox-Gastaut syndrome (LGS) have a high epileptic burden, both ictal and interictal epileptiform discharges (IED). Comorbid behavioral and cognitive issues limit the value of clinically reported seizures. Given the disease burden and refractoriness to anti-seizure medications, deep brain stimulation (DBS) of the centromedian thalamic nucleus (CM) is an option in patients with LGS. Finding an early and reliable biomarker could predict efficacy in clinical trials and improve patient selection.
Methods:
Patients with LGS and CM-DBS were recruited from the Sao Paolo epilepsy clinic. One hour of scalp EEG was recorded at an outpatient visit approximately one month following DBS surgery, and all patients had at least one-year follow-up. Stimulation was titrated during the one-hour EEG recording. Stimulation for all subjects was bipolar, continuous, and bilateral at 130 Hz and 300 μs pulse width. Before stimulation, 10 min of baseline EEG was recorded, after which the stimulator was turned on at 1V for five minutes, followed by a five minute period of no stimulation. This continued in increments of 1V until either 5V of stimulation or the subject experienced side effects of paresthesias. IEDs were identified by automated detection using Persyst (v14c) and then screened by an expert human reader. Spikes were counted in ten second windows. Spike count data was normalized and charted as a heat map to show IED burden and response to stimulation. Spearman rank correlation assessed the relationship between seizure reduction rate and maximum percentage change in IED burden.
Results:
IED burden was charted as a heatmap to display the percentage change in IED burden vs. stimulation for each patient (figure 1). Except for one patient, the rest (n=10) had a reduction in IED burden ranging from 31% to 100%. Eight patients were responders (responder defined as 50% reduction in seizure burden), and two non-responders. The degree in reduction of IED burden during the acute titration phase did not correlate with seizure reduction rate at one year (p=0.35). Spearman rank correlation was 0.3 suggesting a weak correlation between the two variables (Figure 2).
Conclusions:
While there was a reduction in IED burden in virtually every patient, the degree of reduction did not correlate with the mean reduction of seizure frequency at one year. One patient who did not have any notable change in IED burden with stimulation was a non-responder. Our study has several limitations: small sample size, skewed outcome data due to high responder rate, and limited duration of EEG sampling. Our study strengths include quantification of IED burden along a continuum instead of binarizing the data and a prospective study design to evaluate EEG changes with increasing stimulation currents. Whether or not a certain percentage acute reduction of IED burden can predict chronic response, and thus serve as a positive biomarker is unclear from our data set and needs to be studied in larger sample sizes.
Funding: None.