Electrocardiography

Importance: Existing pharmacological treatments for insomnia have significant limitations.

Objective: To assess the effective dose range, safety, and tolerability of the novel selective orexin-2 receptor antagonist seltorexant in insomnia disorder.

Design, setting, and participants: This randomized, double-blind, active- and placebo-controlled, dose-finding, polysomnography study was conducted from November 2017 to April 2019 at 55 sites in 6 countries and analyzed in August 2019. The timeline for submission of this data for publication was impacted by internal strategic decision-making. Adults (aged 18-64 years) and older adults (aged 65-85 years) with insomnia (Insomnia Severity Index score ≥15) and no psychiatric comorbidity were included.

Interventions: Participants were randomized 1:1:1:1:1 to receive nightly oral-seltorexant (5 mg, 10 mg, or 20 mg), placebo, or zolpidem (5-10 mg) for 14 days.

Main outcomes and measures: Primary and key secondary outcomes included the dose-response relationship of night 1 latency to persistent sleep (LPS) and wake after sleep onset over the first 6 hours (WASO-6). Other secondary outcomes included night 13 LPS and WASO-6. Due to asymmetrical distributions of LPS and WASO-6 at baseline, log transformation was applied and results were expressed as back-transformed least-squares mean (LSM) ratios for comparisons between groups.

Results: Overall, 364 participants (mean [SD] age, 57.8 [12.4] years; 246 [67.6%] female) received seltorexant, 5 mg (n = 71), 10 mg (n = 74), or 20 mg (n = 71); placebo (n = 75); or zolpidem (n = 73). The night 1 dose-response relationship for LPS was significant (with trend test t statistics ≥3.99 and adjusted P values <.001 for all 4 prespecified models), with greater improvements in seltorexant, 10 mg and 20 mg, vs placebo (10 mg: LSM ratio, 0.64; 90% CI, 0.51-0.81; 20 mg: LSM ratio, 0.51; 90% CI, 0.41-0.64) and in seltorexant, 20 mg, vs zolpidem (LSM ratio, 0.71; 90% CI, 0.57-0.88). The night 1 dose-response relationship for WASO-6 was also significant, with trend test t statistics ≥3.99 and adjusted P values <.001 for all 4 prespecified models (seltorexant, 10 mg: LSM ratio, 0.68; 90% CI, 0.55-0.85; seltorexant, 20 mg: LSM ratio, 0.60; 90% CI, 0.48-0.74). Night 1 LPS and WASO-6 improvements were maintained on night 13 for seltorexant, 10 mg and 20 mg, but diminished for zolpidem. On night 13, compared with zolpidem, seltorexant, 10 mg and 20 mg, improved LPS by 30% and 28%, respectively, and seltorexant, 20 mg, improved WASO-6 by 31%. Treatment-emergent adverse events (TEAEs) were lower across the combined seltorexant doses (73/216 [33.8%]) relative to placebo (37/75 [49.3%]) and zolpidem (31/73 [42.5%]). Two participants experienced serious TEAEs during the double-blind phase (1 in the seltorexant, 20 mg, group and 1 in the zolpidem group). Three participants in the seltorexant, 5 mg, and 1 in the seltorexant, 20 mg, group experienced asymptomatic electrocardiogram-related TEAEs leading to discontinuation.

Conclusions and relevance: Among participants with insomnia in this study, seltorexant, 10 mg and 20 mg, improved sleep initiation and maintenance throughout 14 days of treatment. Seltorexant was generally well tolerated.

Trial registration: ClinicalTrials.gov Identifier: NCT03375203.