Effect of Esketamine on Depressive Symptoms in Adolescents with Major Depressive Disorder at Imminent Suicide Risk: A Randomized Psychoactive-Controlled Study

Major depressive disorder (MDD) is highly prevalent in adolescents and is the psychiatric condition most commonly associated with suicide. Earlier research and clinical practice highlight that conventional oral antidepressants require several weeks to reach full effect and that adolescents with acute suicidal behaviour or intent are frequently excluded from antidepressant trials, leaving a gap in evidence-based, rapid pharmacological interventions for this high-risk group. Hospitalisation, psychotherapy and initiation or optimisation of oral antidepressants constitute the current standard-of-care (SOC), but each has limitations in speed of onset and durability of protection from suicide in the weeks after discharge. Kosik-Gonzalez and colleagues set out to evaluate whether intranasal esketamine, an N-methyl-D-aspartate (NMDA) receptor antagonist, administered in addition to comprehensive SOC, can produce a rapid reduction in depressive symptoms in adolescents with MDD who are assessed to be at imminent risk for suicide. The trial compared three fixed doses of esketamine (28, 56, 84 mg) with an active psychoactive control (oral midazolam) given twice weekly for 4 weeks, with the primary endpoint focused on change in depressive symptoms 24 hours after the first dose. The study therefore addresses the question of rapid-onset antidepressant effect in a population typically excluded from registrational trials and conducted in the context of usual acute-care practices for suicidal adolescents.

This was a randomized, double-blind, double-dummy, psychoactive placebo-controlled, multicentre Phase IIb study conducted from January 2018 to March 2023 at 37 sites across seven countries. Eligible participants were adolescents aged 12 to <18 years who met DSM-5 criteria for MDD without psychosis, had current suicidal thinking with intent as assessed by MINI-KID items, were judged to require acute psychiatric hospitalisation for imminent suicide risk, and had moderate-to-severe depressive symptoms (CDRS-R total score ≥58 at baseline). Key psychiatric comorbidities (for example bipolar disorder, moderate-to-severe substance use disorder, psychotic disorders) and certain other conditions were exclusionary; a full list was reported in supplemental material. Candidates were screened after presentation to emergency departments or inpatient psychiatric units. Participants were randomised 1:1:1:2 to fixed-dose intranasal esketamine 28 mg, 56 mg, 84 mg, or to oral midazolam (0.125 mg/kg), administered twice weekly for 4 weeks. A double-dummy design ensured masking: esketamine groups received oral placebo and the midazolam group received intranasal placebo. Randomisation used randomly permuted blocks stratified by centre; an unblinded pharmacist prepared study drug to preserve the blind. All participants received comprehensive SOC, including initial hospitalisation (minimum recommended 5 days unless clinically adjusted), initiation or optimisation of one of three oral antidepressants (fluoxetine, escitalopram or sertraline) by day 7 at the latest, and evidence-based psychotherapy through at least day 81; type and frequency of therapy were not mandated. The primary efficacy endpoint was change from baseline in the Children's Depression Rating Scale–Revised (CDRS-R) total score to 24 hours post-first dose, analysed using ANCOVA with treatment and analysis centre as factors and baseline CDRS-R as covariate. A prespecified pooled sequential multiple-testing procedure controlled type I error: first pooled 56 mg and 84 mg esketamine versus midazolam (two-sided α=0.05), then individual 56 mg and 84 mg versus midazolam only if the pooled comparison was significant. Dose–response was examined using multiple trend tests, a sigmoid Emax model, and MCP-Mod. Secondary assessments included MADRS, suicidality measures from the Suicide Ideation and Behaviour Assessment Tool (SIBAT) including CGI-SS-R and CGI-SR-I, and clinician- and patient-reported frequency of suicidal thinking. Safety assessments comprised adverse events, CADSS for dissociation, MOAA/S for sedation, BPRS+ for psychosis-like symptoms, YMRS for mania, vital signs (including blood pressure), a computerized cognitive battery, PWC-20 for withdrawal, and timeline follow-back for ketamine/PCP use. Efficacy analyses used a full efficacy dataset (participants with baseline and ≥1 post-dose CDRS-R); missing data were imputed using last observation carried forward (LOCF). The study was sized to detect an effect size of 0.65 for the pooled 56 mg/84 mg comparison, targeting approximately 29 participants per esketamine dose and 58 to midazolam; 147 adolescents were ultimately randomised.

A total of 147 adolescents were randomised (28 mg n=29; 56 mg n=31; 84 mg n=24; midazolam n=63). Most participants were moderately-to-severely depressed and about 95% were rated moderately-to-extremely suicidal at enrolment; 54% reported a suicide attempt within the month prior to enrolment. Completion rates were high: 139 (94.6%) completed the double-blind phase and 111 (80.4%) completed 6-month follow-up. The primary efficacy analysis found that pooled esketamine 56 mg and 84 mg produced a greater reduction in CDRS-R total score at 24 hours post-first dose versus midazolam: least-squares (LS) mean difference -5.8 (95% CI -11.19, -0.35), p=0.037, with an effect size of 0.36. Follow-up testing on the individual 84 mg and 56 mg doses did not reach statistical significance: LS mean differences versus midazolam were -5.7 (95% CI -12.91, 1.55), p=0.123 for 84 mg, and -5.9 (95% CI -12.25, 0.53), p=0.072 for 56 mg. A significant dose–response for change in CDRS-R was observed at 24 hours (multiple trend test one-sided p=0.030). Improvements on MADRS were directionally consistent with CDRS-R but the reported between-group differences were smaller and their confidence intervals included zero. By 24 hours post-first dose, response rates (≥50% improvement in CDRS-R adjusted score) in esketamine groups ranged from 53.6% to 61.3%, with remission rates (CDRS-R ≤28) of 16.1% to 21.7%; absolute differences versus midazolam for response were 4.4%, 12.1%, and 7.3% for the 28, 56, and 84 mg groups respectively. At day 25 (end of double-blind phase, 4 hours post-final dose), the majority across groups met response criteria: esketamine 28 mg 96.4%, 56 mg 71.0%, 84 mg 73.9%, and midazolam 77.8%; remission rates at day 25 were higher than at 24 hours across groups (for example, 60.7% in the 28 mg group and 41.3% in the midazolam group). CDRS-R and MADRS improvements were generally sustained through the post-treatment follow-up phase. Measures of suicidality improved in all four treatment arms by 24 hours post-first dose as assessed by CGI-SS-R; LS mean changes from baseline were roughly -1.6, -1.6, -1.3 for the 28, 56 and 84 mg esketamine groups, respectively, and approximately -1 for midazolam, and the between-group CIs included zero, indicating no statistically significant separation on suicidality indices. Safety findings showed common treatment-emergent adverse events (AEs) with esketamine (incidence ≥20% across doses) including dizziness, nausea, dissociation, headache, dysgeusia, somnolence, vomiting and hypoesthesia; intentional self-injury was reported at similar rates in esketamine-treated versus midazolam-treated participants (20.5% vs 19.0%) during the double-blind phase. Most common AEs occurred on dosing days and resolved the same day; the majority were mild or moderate. Serious adverse events (SAEs) during the double-blind phase occurred in 12 participants across groups (esketamine 28 mg: 4 [13.8%]; 56 mg: 7 [22.6%]; 84 mg: 1 [4.3%]; midazolam: 9 [14.3%]); investigators judged none to be related to study drug. Suicide attempts during the double-blind phase were reported in 7 (8.4%) esketamine-treated and 5 (7.9%) midazolam-treated participants. Over the 6-month follow-up, SAEs were reported for 31.6% of those randomised to esketamine and 32.2% of those randomised to midazolam; suicide attempts in the 6-month period occurred in 15.2% and 15.3% of participants, respectively. One participant randomised to midazolam died by suicide on day 193, more than five months after last study drug dose. Dissociative symptoms increased at the 40-minute post-dose assessment in the esketamine groups (dose-related), typically resolving by about 1.5 hours post-dose and attenuating with repeated dosing. Transient increases in blood pressure were observed with esketamine (mean maximum systolic increases from pre-dose approximately 9.4, 10.7 and 14.8 mmHg for the 28, 56 and 84 mg groups, respectively, versus approximately 5.0 mmHg for midazolam). Cognitive testing revealed no systematic adverse effects on attention, executive function, working memory, or learning/memory.

Kosik-Gonzalez and colleagues interpret the trial as evidence that intranasal esketamine administered alongside comprehensive SOC produced a rapid reduction in depressive symptoms in adolescents at imminent suicide risk, with a statistically significant benefit for the pooled 56 mg and 84 mg doses at 24 hours after the first dose. The authors note that the observed between-group difference on the primary endpoint (LS mean difference -5.8) compares favourably in magnitude to effect sizes reported for standard oral antidepressants at study endpoints, while acknowledging differences in assessment timing and study populations. The investigators point out a significant dose–response relationship and consistency of rapid antidepressant effects with prior phase 2/3 esketamine studies in adults. Nevertheless, they caution that individual 56 mg and 84 mg dose comparisons did not reach significance, likely reflecting limited sample sizes, since the study was powered for the pooled primary comparison only. Improvements in suicidality measures occurred across all groups, but between-group differences were not statistically significant; the authors attribute this to several factors including non-specific therapeutic effects of comprehensive care (hospitalisation, psychosocial support, initiation/optimization of oral antidepressant and psychotherapy) and the use of midazolam as an active psychoactive comparator, which may have dampened the detectable signal on some measures. In terms of safety, the study detected no new or unexpected adverse events in adolescents compared with adult trials; common esketamine-associated effects such as dissociation, dizziness and somnolence occurred more frequently than in adults but were generally transient and resolved on dosing days. The potential for functional unblinding due to characteristic side effects is highlighted as a methodological limitation, and the authors discuss that midazolam—while chosen to provide psychoactive effects—remains an imperfect control. Additional limitations acknowledged by the investigators include the relatively small sample size (powered only for the primary endpoint), the modified use of CDRS-R for short recall periods (which may require further psychometric evaluation), and the inherent difficulty of isolating drug effects in an acutely suicidal inpatient population where non-specific elements of study participation and SOC can have substantial therapeutic impact. The authors conclude that esketamine plus comprehensive SOC produced rapid reductions in depressive symptoms in this high-risk adolescent sample and that the safety profile aligned with prior adult data, while underscoring the need for cautious interpretation given the study’s design constraints and the unresolved questions about effects on suicidality measures and longer-term outcomes.