Treatment effect and safety of seltorexant as monotherapy for patients with major depressive disorder: a randomized, placebo-controlled clinical trial

Abstract

The antidepressant efficacy and safety of seltorexant monotherapy in major depressive disorder (MDD) was investigated in a placebo-controlled, placebo lead-in, randomized, double-blind, phase 1b study. Participants were randomized to receive seltorexant (20 mg or 40 mg) or placebo. The treatment effect was assessed by changes in the Hamilton Rating Scale for Depression-17 item (HDRS₁₇) from treatment-period baseline to week 5 in lead-in placebo non-responders ("enriched" intent-to-treat analysis set). As a secondary outcome, the effect of seltorexant on HDRS₁₇ was assessed in patients with and without subjective insomnia. Seltorexant's effects on polysomnography, serum cortisol, and cortisol waking response were also measured. In total, 128 participants were enrolled, including 86 in the enriched sample (lead-in placebo non-responders). The mean changes from baseline (SD) in HDRS₁₇ score at week 5 differed significantly across arms: -7.0 (5.04) for seltorexant 20 mg, -5.5 (4.34) for seltorexant 40 mg, and -4.4 (3.67) for placebo (p = 0.0456), which was attributable to the difference between the 20 mg and placebo arms (p = 0.0049). Improvement in depression severity at week 5 for seltorexant 20 mg was greater in patients with higher baseline insomnia severity (nominal p = 0.0059). The treatment benefit in the 20 mg arm remained significant when HDRS scores were adjusted by removing the sleep items (nominal p = 0.0289). The mean HDRS₁₇ change versus placebo was numerically larger in the 20 mg than the 40 mg arm, consistent with data from a previous study in which seltorexant was administered adjunctively to conventional antidepressants. In secondary analyses, the waking cortisol response decreased in the 20 mg arm but not the 40 mg or placebo arms, and while total sleep increased more in the 40 mg arm, this arm also showed reduced REM onset latency and increased stage N1 sleep, which were not evident in the 20 mg arm. These biomarker data suggest mechanistic hypotheses that may account for the apparent curvilinear dose-response relationship of seltorexant. Trial Registration: ClinicalTrials.gov, NCT03374475.

Fig. 1. Study Design.

This phase 1 study consisted of a screening phase (up to 3 weeks), a double-blind treatment phase (8 weeks), and a follow-up phase (1 week). The double-blind treatment phase consisted of 3 periods: a blinded placebo lead-in period, a treatment period (5 weeks), and a blinded placebo withdrawal period. The study duration for each patient was ~13 weeks. MDD, major depressive disorder; MADRS, Montgomery–Åsberg Depression Rating Scale; HDRS₁₇, Hamilton Depression Rating Scale-17; green dotted line, placebo responders based on reduction from lead-in baseline in HDRS₁₇; purple dotted line, placebo non-responders based on reduction from lead-in HDRS₁₇; gold solid line, contributes to primary analysis.

Fig. 2. Evaluation of HDRS₁₇ in the eITT analysis set.

A Changes in HDRS₁₇ Sleep Adjusted Total Score and HAM-D₆ Subscale Score; eITT Analysis Set. B HDRS₁₇ Total Score: Least Squares Mean Changes from Baseline During the Treatment Period; eITT Analysis Set. Sleep disorder as documented in the sleep diary by subjective Sleep Onset Latency >30 min and total sleep time of <6 h at least 3 nights over 7 recorded days. eITT, “enriched” intent-to-treat; HAM-D₆, 6-item subscale from HDRS₁₇; HDRS₁₇, Hamilton Rating Scale for Depression-17 item; ISI Insomnia Severity Index, JNJ, seltorexant, RRS Rumination Response Scale, SE standard error.

Fig. 3. Cortisol waking response and Nadir levels.

A Mean Cortisol Waking Response (±SE); eITT Analysis Set. B Cortisol Waking Response (±SE); Least Squares Mean Changes (in nmol/L) and Comparison Versus Placebo; eITT Analysis Set. C Mean Cortisol Nadir Levels (±SE) Over Time (nmol/L); eITT Analysis Set. D Cortisol Nadir Levels: Mean Changes (in nmol/L) From Time-Matched Lead-In Baseline and Comparison Versus Placebo at Week 3–4; eITT Analysis Set.

Fig. 4. Mean changes in N1 sleep duration and REM latency (in Minutes): eITT analysis set.

N1 sleep duration (minutes) and REM onset latency (minutes) duration change from baseline at week 3–4: Mean values (±90% CI).