Abstract Introduction The Maintenance of Wakefulness Test (MWT) is widely used in clinical trials to assess excessive daytime sleepiness (EDS). Mean sleep onset latency (SOL), assessed by MWT, the primary endpoint in these studies, measures the ability to stay awake; it does not provide insight into the quality of wakefulness. We identified microsleeps, short (3-15 sec) periods of sleep prior to sleep onset during MWT, and artificial intelligence (AI)-derived sleepiness scores from electroencephalography (EEG) recordings as biomarkers for EDS in participants with narcolepsy type 1 (NT1), then compared these pre-post treatment with oveporexton, an oral orexin receptor 2-selective agonist. We report results from two randomized, double-blind, phase 3 studies (The First Light FL: NCT06470828; The Radiant Light RL: NCT06505031). Methods Participants (16-70 years) with NT1 were randomized to oveporexton 1mg/1mg (FL only), oveporexton 2mg/2mg, or placebo, at least 3 hours apart (first dose 8AM) for 12 weeks. MWT sessions consisted of 40-minute wake trials at 10AM, 12PM, 2PM, and 4PM to assess sleepiness fluctuations. Exploratory microsleeps data were scored manually during MWTs conducted at baseline and end of treatment. EEG recordings were processed to generate AI-derived sleep-stage probabilities, with all non-wake stage probabilities summed into sleepiness scores. Zero-inflated negative binomial mixed-effects models (with an offset SOL term), mixed-effects survival analysis, and linear mixed-effect model were used to analyze microsleep rates, time-to-first microsleep, and AI-derived sleepiness scores, respectively. Placebo-corrected change from baseline within treatments were computed. Results MWT data were collected from 158 FL participants (placebo: n=37, 1mg/1mg: n=57, 2mg/2mg: n=64) and 105 RL participants (placebo: n=35, 2mg/2mg: n=70). In all oveporexton-treated groups across both studies, microsleep rates decreased significantly relative to baseline (from 3 microsleeps per 10 min at baseline to 1 per 10 min; each p 0.05). Additionally, oveporexton-treated groups experienced prolonged time-to-first microsleep versus placebo (each p 0.005). Average microsleep duration (excluding participants with no microsleep) showed no significant change in any group from baseline. Conclusion Oveporexton significantly reduced microsleep rates and delayed onset of the first microsleep in MWTs, demonstrating its potential as a digital biomarker and benefit over SOL in participants with NT1. Support (if any) Funded by Takeda Development Center Americas, Inc.
Gong et al. (Fri,) studied this question.