Authors :
Presenting Author: Fatemeh Mohammad Alizadeh Chafjiri, MD – Boston Children's Hospital
Emily Peter, BA – Research Assistant, Neurology, Boston Children's Hospital; Lillian Voke, BS – Research Assistant, Neurology, Boston Children's Hospital; Sarah Cantley, BA – Research Assistant, Boston Children's Hospital; Michele Jackson, BA – Neurology – Boston Children's Hospital; Solveig Vieluf, PHD – Research Fellow, Neurology, Boston Children's Hospital; Tobias Loddenkemper, MD – Professor of Neurology, Boston Children's Hospital
Rationale:
Wearable electrodermal activity (EDA) and heart rate (HR) recordings show changes during seizures.
1 Anti-seizure medications (ASMs) may affect autonomic nervous system (ANS) function in patients with epilepsy (PWE). Studies suggest that heart rate variability (HRV) was lower during periods of high ASM levels as compared to lower levels.
2 Here, we assess acute ASM effects on EDA and HRV from wearables in PWE before and after ASM intake compared to PWE without ASMs.
Methods:
We enrolled patients, ages one month to 21 years, admitted for long-term video-EEG monitoring between Feb 2015 and 2021 who wore at least one wearable sensor (Empatica E4) on wrists and/or ankles to record EDA and photoplethysmography, from which HR is derived. We compared PWE on 1 ASM (case group) at home and during enrollment to patients with or without epilepsy on no ASMs (control group). We excluded patients with intravenous or continuous infusion ASM, seizures within 2 hours of ASM intake, short recordings, or recordings where the sensor lost skin contact. We included one recording per patient. For the case group, we evaluated a baseline for one hour before individualized ASM intake and an hour around ASM-specific peak time (30 min before and after peak). For controls, we used mean case group intake and peak times for comparison. From baseline and peak wearable recordings, we analyzed mean EDA and HR, and HRV (root mean square of successive differences, RMSSD). We calculated a two cohort (case, control) by two time windows (baseline, peak) repeated measure analysis of variance.
Results:
We included 19 patients
(52.6% female, median age: 9.62 years
) on eight ASMs in the case group and nine in the control group (66.7% female, median age: 3.06 years, Table 1). Mean EDA was lower in the case group than in controls (F (1, 26) = 4.30, p = 0.048, η² = 0.142).
HR showed a trend for an interaction between the factors group and time window (F (1, 26) = 4.02, p = 0.056, η² = 0.134). The group and time window interaction was significant for RMSSD (F (1, 26) = 4.47, p = 0.044, η² = 0.147)
. Bonferroni post-hoc comparisons show RMSSD lower during ASM mean peak compared to baseline for the case group (p = 0.039), while RMSSD does not differ for controls (p = 0.293) (Figure 1).Conclusions:
Evaluation of ASM acute effects on wearable signals showed lower RMSSD during mean peak time compared to the period before intake, presumably the time point of the lowest ASM effects. Our results may provide a future biomarker to monitor ASMs non-invasively and suggest that tracking ASM intake times may be relevant for developing a seizure monitoring system, as ASM effects on the ANS may interrelate with seizure likelihood.
1 Vieluf S. Twenty-four-hour patterns in electrodermal activity recordings of patients with and without epileptic seizures.
Epilepsia. 2021;62(4).
2 Halimeh M. Wearable device assessments of antiseizure medication effects on diurnal patterns of electrodermal activity, heart rate, and heart rate variability.
Epilepsy & Behavior. 2022;129:108635.
Funding:
Epilepsy Research Fund. SV was funded by the AES Early Career Grant.