Short-term cognitive effects of repeated-dose esketamine in adolescents with major depressive disorder and suicidal ideation: a randomized controlled trial

Ketamine and its S-enantiomer esketamine, both N-methyl-d-aspartate (NMDA) receptor antagonists, have demonstrated rapid antidepressant and antisuicidal effects in adults. Lan and colleagues note that, despite accumulating adult data and a small number of open-label and case-series reports in adolescents, there is limited systematic evidence about the cognitive effects of therapeutic ketamine/esketamine in young people. Concerns persist because chronic recreational ketamine use has been associated with deficits in executive function and memory, and animal studies suggest vulnerable neurodevelopmental windows during adolescence. This paper reports a secondary analysis of a randomised, double-blind, placebo-controlled trial in adolescents with major depressive disorder (MDD) and current suicidal ideation, aiming to evaluate short-term cognitive outcomes after three subanesthetic esketamine infusions. The investigators hypothesised that three low-dose infusions (0.25 mg/kg) would not produce cognitive decline relative to baseline, and they additionally explored whether baseline cognitive performance predicted antidepressant or antisuicidal response.

The analysis used data from a single-centre, randomised (1:1), double-blind, placebo-controlled trial conducted at the Affiliated Brain Hospital of Guangzhou Medical University between December 2020 and April 2022. This report is a secondary analysis focusing on cognitive outcomes; the parent trial assessed safety and efficacy of three esketamine infusions versus midazolam in adolescents with MDD and suicidal ideation. Eligible participants were inpatients aged 13–18 years with DSM-5 major depressive episode without psychotic features, a HAMD-17 score ≥ 17 at screening, and current suicidal ideation of at least 3 months’ duration (clinician-rated C-SSRS suicidal ideation score ≥ 1 and a self-report Beck SSI item 4 or 5 score ≥ 2). Exclusion criteria included substance/alcohol addiction, primary psychotic or bipolar disorder, neurodevelopmental disorders, significant medical illness, recent active suicide attempt, and nonpsychiatric neurological disorders. All participants were physically healthy on examination and laboratory testing. Randomisation was computer-generated and allocation prepared by a designated psychiatrist who did not take part in assessments. Participants, raters and treating clinicians were blinded. Study drug (esketamine hydrochloride 0.25 mg/kg or midazolam 0.02 mg/kg) was administered in 50 ml normal saline by intravenous infusion over 40 minutes on Days 1, 3 and 5. Clinical monitoring included vital signs and physician supervision during infusions; infusions were to be halted for intolerable emergent events. Standard oral medication was continued during the infusion phase; structured psychotherapy, TMS and ECT were not permitted during that period. Cognitive function was measured with four dimensions from the Chinese MATRICS Consensus Cognitive Battery (MCCB) at baseline, Day 6 (one day after the final infusion) and Day 12 (one week after the final infusion). Selected domains were processing speed (BACS, Trail Making A, Category Fluency), working memory (Spatial Span subtest of WMS-III), verbal learning (Hopkins Verbal Learning Test–Revised) and visual learning (Brief Visuospatial Memory Test–Revised). Scores were standardised to T-scores (mean 50, SD 10). Clinical measures included MADRS for depression and the first five items of the Beck SSI for suicidal ideation, assessed at the same timepoints. Different test versions were used across visits for the learning measures to reduce practice effects. Statistical planning used an effect-size estimate of 0.85 for suicidal ideation change to derive a sample of 22 per group for 80% power; the trial aimed for at least 25 per group allowing dropout. Analyses followed an intention-to-treat principle. Between-group baseline comparisons used t-tests and chi-squared tests. Primary longitudinal analyses employed linear mixed-effects repeated-measures models including fixed effects for time (baseline, Days 6 and 12), drug (esketamine versus midazolam) and time-by-drug interaction, with covariates for oral antidepressant class (SSRI versus non-SSRI), augmentation therapy (yes/no) and the relevant baseline score. Bonferroni-corrected post hoc tests and Cohen’s d effect sizes were reported for within- and between-group changes. A generalized estimation equation model tested whether baseline MCCB scores predicted antidepressant response (≥ 50% MADRS improvement), remission (MADRS ≤ 12) or being free of suicidal ideation (SSI-5 = 0), with study drug and medication covariates; similar models were run within each treatment subgroup.

From 155 screened adolescents, 54 were enrolled and received at least one infusion; five participants did not complete all three infusions (three midazolam, two esketamine). The final analysis sample comprised 51 participants who had baseline and Day 6 cognitive and clinical data (26 in the esketamine group, 25 in the midazolam group). Forty-five participants (88.2%) completed the Day 12 assessment. Groups were balanced at baseline on age, education, sex, BMI, illness duration, medication types, baseline MADRS, SSI-5 and MCCB domain scores. Clinical outcomes replicated the parent-trial findings: three infusions of esketamine were associated with larger reductions in MADRS and SSI-5 scores compared with midazolam. At Day 12, responder, remission and being free of suicidal ideation rates were 38.5%, 26.9% and 30.8% in the esketamine group versus 20.0%, 8.0% and 16.0% in the midazolam group. The MADRS showed significant time, drug and time-by-drug interaction effects (MADRS time-by-drug F = 3.345, P = 0.039; time F = 64.685, P < 0.001; drug F = 4.409, P = 0.041). SSI-5 had significant time and drug main effects but no significant interaction. Cognitive outcomes: processing speed demonstrated a significant time main effect (F = 12.803, P < 0.001), drug main effect (F = 6.607, P = 0.013) and time-by-drug interaction (F = 3.315, P = 0.041) in the linear mixed model. The esketamine group showed significant increases in processing speed from baseline to Day 6 and to Day 12, with Cohen’s d effect sizes of 0.566 and 1.032 respectively, indicating improvement. The midazolam group did not show changes in processing speed. Working memory and verbal learning exhibited significant time main effects (all P < 0.05) but no significant drug or interaction effects; the authors report that working memory improved in the esketamine group from baseline to Day 12. Visual learning did not show a drug-related effect in the reported results. Finally, the generalized estimation equation analyses found no significant associations between baseline cognitive performance and subsequent antidepressant response, remission or freedom from suicidal ideation (all P > 0.05), a pattern that held within each treatment subgroup.

Lan and colleagues interpret their findings as evidence that three subanesthetic esketamine infusions did not impair short-term cognitive function in adolescents with MDD and suicidal ideation over a 12-day follow-up; rather, they observed an improvement in processing speed among esketamine-treated participants. The authors relate these results to adult literature in which single and repeated subanesthetic ketamine/esketamine regimens generally do not produce sustained cognitive deficits and in some cases are associated with modest improvements in memory, processing speed and working memory. The discussion acknowledges contrasting data from recreational and chronic high-dose ketamine users, where persistent cognitive deficits have been observed, and from animal studies in which repeated low-dose NMDA antagonists during adolescence induced neurobiological changes in prefrontal cortex and hippocampus that might impair cognition. The investigators highlight developmental considerations: adolescents may metabolise drugs faster and have higher NMDA receptor expression during adolescence, so identical adult doses could exert different effects in younger patients. They therefore advise that dose-ranging and longer-term studies are needed to balance antidepressant benefit and neurodevelopmental safety. Key limitations noted by the authors include that this analysis was a secondary endpoint from a trial not specifically powered for cognitive outcomes and that the sample size was relatively small. Cognitive assessment was limited to baseline, Day 6 and Day 12, so acute (within hours) and longer-term effects were not captured. Only four MCCB domains were assessed, omitting sustained attention and social cognition, and practice effects remain a concern despite use of alternate versions for the learning tests. Concomitant oral antidepressant medication during the infusion phase could confound cognitive changes, and suicidal ideation was measured with the first five SSI items rather than the full instrument. The authors conclude that larger, longer, and dose-ranging studies with more comprehensive cognitive batteries are required to confirm these short-term findings and to evaluate long-term neurodevelopmental safety.

The study concluded that three low-dose (0.25 mg/kg) esketamine infusions did not impair short-term cognitive function in adolescents with MDD and current suicidal ideation over a 12-day follow-up; instead, an improvement in processing speed was observed. The authors recommend further research to clarify long-term cognitive effects and optimal dosing in this population.