In a proof-of-concept randomised sequential trial in treatment‑resistant depression, responders to six ketamine infusions were randomised to 14 weeks of CBT versus treatment‑as‑usual and CBT showed a significant group-by-time advantage on self-report depressive symptoms (QIDS; moderate-to-large Cohen d ≈ 0.71), while the primary clinician-rated outcome (MADRS) showed a non-significant moderate effect (d = 0.65). A subset analysis also found improved accuracy on an emotional N‑back task among ketamine responders.
Papers cited by this study that are also in Blossom
Introduction
Ketamine has emerged as a rapid-acting antidepressant. While ongoing treatment can prevent relapse, concerns exist regarding long-term exposure.
Objective
We conducted a randomized trial to examine the feasibility and efficacy of cognitive behavioral therapy (CBT) following intravenous ketamine in treatment-resistant depression (TRD).
Methods
Subjects with TRD were recruited and treated with 6 intravenous infusions of ketamine over 3 weeks. Subjects who experienced a clinical response (≥50% improvement in depression severity) were then randomized to receiving CBT or treatment as usual (TAU) for an additional 14 weeks, using a sequential treatment model.
Results
Of the 42 patients who signed consent, 28 patients achieved a response and were randomized to CBT or TAU. When measured using the Montgomery-Asberg Depression Rating Scale (primary outcome measure), the effect size at the end of the study was moderate (Cohen d = 0.65; 95% CI –0.55 to 1.82), though the group-by-time interaction effect was not significant. There was a significant group-by-time interaction as measured by the Quick Inventory of Depressive Symptomatology (F = 4.58; p = 0.033), favoring a greater sustained improvement in the CBT group. This corresponded to a moderate-to-large effect size of the Cohen d = 0.71 (95% CI –0.30 to 1.70) at the end of the study (14 weeks following the last ketamine infusion). In a subset of patients (N = 20) who underwent cognitive testing using the emotional N-back assessments before and after ketamine, ketamine responders showed improvement in the accuracy of emotional N-back (t[8] = 2.33; p < 0.05) whereas nonresponders did not (t[10] <1; p ns).
Conclusions
This proof-of-concept study provides preliminary data indicating that CBT may sustain the antidepressant effects of ketamine in TRD. Further study and optimization of this treatment approach in well-powered clinical trials is recommended.
Wilkinson and colleagues frame the study against the backdrop of ketamine's emergence as a rapid-acting antidepressant and the practical and safety concerns surrounding long-term ketamine/esketamine maintenance. The introduction notes that maintenance regimens can be time-consuming, costly, and disruptive, and that although one-year safety data exist for esketamine, concerns about abuse liability and other adverse outcomes remain. The authors highlight evidence that ketamine enhances synaptic neuroplasticity and may rapidly improve cognitive flexibility, suggesting a time-limited window in which psychotherapeutic interventions could be more effective. Building on this rationale, the investigators tested whether cognitive behavioural therapy (CBT) delivered after a course of intravenous ketamine could sustain antidepressant benefit in treatment-resistant depression (TRD). Using a sequential treatment model, they conducted a proof-of-concept randomised trial to evaluate feasibility and efficacy, hypothesising that CBT would improve longer-term outcomes following ketamine by leveraging ketamine-induced increases in neuroplasticity and cognitive control.
Participants were adults aged 18–65 with severe major depressive disorder recruited from patients seeking ketamine at the Yale Interventional Psychiatric Service between February 2017 and August 2019. Inclusion required resistance to two or more adequate antidepressant trials (assessed via the MGH Antidepressant Treatment History Questionnaire) and a 17-item Hamilton Depression Rating Scale score of 21 or greater. Individuals with active substance use (other than nicotine) or comorbid schizophrenia/schizoaffective disorder were excluded. Forty-two patients consented; 41 entered the ketamine phase. Phase 1 comprised six open-label intravenous ketamine infusions (0.5 mg/kg over 40 minutes) delivered over three weeks. Patients who achieved a clinical response (≥50% improvement in depression severity at the end of phase 1) were randomised 1:1 to receive either CBT or treatment as usual (TAU) for a 14-week phase 2. Randomisation used computer-generated codes in randomly varying blocks and was implemented by a biostatistician not otherwise involved in the trial. No additional ketamine infusions were given during phase 2. The CBT intervention, based on the Beck model, included psychoeducation, cognitive restructuring, and behavioural activation. Two therapists experienced and certified in CBT delivered twice-weekly sessions for the first two weeks of phase 2, then weekly sessions thereafter; sessions included agenda setting and homework (activity charts, thought records). TAU comprised weekly or every-other-week visits with a study physician focused on medication management and adverse-event monitoring, with oral antidepressant adjustments permitted as clinically indicated. Depression severity was assessed weekly using the clinician-rated Montgomery–Åsberg Depression Rating Scale (MADRS; primary outcome) and the self-report Quick Inventory of Depressive Symptomatology, 16-item (QIDS-SR-16). MADRS interviews were conducted by telephone by raters blind to treatment allocation. A protocol amendment added neuropsychological testing (emotional 2-back and emotional Stroop tasks) administered before and after the ketamine course to assess cognitive control and flexibility in a subset of participants. For analysis, scores were normalised to baseline (week 1) and repeated-measures analysis of variance was used to test time-by-treatment interactions. Baseline comparisons used two-sample t tests and Pearson χ2 tests. Analyses were run in STATA 15.1. The trial was registered as NCT03027362.
Murrough, J. W., Perez, A. M., Pillemer, S. et al. · Biological Psychiatry (2012)
Zanos, P., Moaddel, P. J., Morris, P. J. et al. · Nature (2016)
Girgenti, M. J., Ghosal, S., Lopresto, D. et al. · Neurobiology of Disease (2017)
Wilkinson, S. T., Katz, R. B., Toprak, M. et al. · Journal of Clinical Psychiatry (2018)
Zarate, C. A., Brutsche, N. E., Ibrahim, L. et al. · Biological Psychiatry (2012)
Papers in Blossom that reference this study
Moore, T. M., Walker, K., Tung, E. et al. · Journal of Affective Disorders (2025)
Mills, N. T., Nikolin, S., Glozier, N. et al. · British Journal of Psychiatry (2025)
Earleywine, M., Falabella, G. S., Oliva, A. B. et al. · Journal of Psychoactive Drugs (2024)
Aepfelbacher, J., Panny, B., Price, R. · Biological Psychiatry (2024)
Zavaliangos-Petropulu, A., Mcclintock, S. M., Khalil, J. et al. · Journal of Affective Disorders (2023)
Chmielewska, Z., Jakuszkowiak-Wojten, K., Wiglusz, M. S. et al. · Brain Sciences (2023)
Price, R. B., Spotts, C., Panny, B. et al. · American Journal of Psychiatry (2022)
Cavarra, M., Falzone, A., Ramaekers, J. G. et al. · Frontiers in Psychology (2022)
Grabski, M., McAndrew, A., Lawn, W. et al. · American Journal of Psychiatry (2022)
Of 42 consented participants, 41 received at least one ketamine infusion and 28 met response criteria (≥50% improvement) at the end of phase 1 and were randomised into phase 2. Among those 28 randomised patients, 53.6% were female and 92.9% were white; mean years of education was 15.1 (SD = 1.7). The sample was chronically and severely ill: 71% reported at least one prior psychiatric hospitalisation, 21% had previously received electroconvulsive therapy, half had a melancholic subtype, and the median duration of the current depressive episode was 156 weeks. On the primary outcome (MADRS), the between-group effect size at the end of the study was moderate (Cohen's d = 0.65; 95% CI -0.55 to 1.82), but the time-by-treatment interaction was not statistically significant. By contrast, self-reported depressive symptoms on the QIDS showed a significant time-by-treatment interaction favouring CBT (F = 4.58; p = 0.033). The corresponding QIDS end-of-study effect size was moderate-to-large (Cohen's d = 0.71; 95% CI -0.30 to 1.70) at 14 weeks after the last ketamine infusion. Neurocognitive testing was completed by 20 participants (9 ketamine responders, 11 nonresponders). Responders improved in accuracy on the emotional 2-back task (t[8] = 2.33; p < 0.05), whereas nonresponders showed no significant change (t[10] < 1, not significant). Responders also showed greater improvement in reaction time on the valence emotional Stroop task compared with nonresponders, but this difference did not reach statistical significance. The protocol was generally tolerated without treatment-emergent adverse events leading to discontinuation. Five serious adverse events occurred (3 in the CBT group, 2 in TAU), all consisting of worsening depression or suicidal ideation requiring hospitalisation; one aborted suicide attempt occurred in the CBT group and one suicidal gesture in the TAU group. Investigators judged these events related to relapse/worsening of illness rather than the study interventions. The most common adverse events attributed to ketamine were headache (24.4%, 10/41) and nausea (12.2%, 5/41); other single-occurrence events were reported. No adverse events were attributed to CBT.
Wilkinson and colleagues interpret the findings as preliminary evidence that post-ketamine CBT may help sustain antidepressant benefit in TRD when delivered in a sequential model. Although the clinician-rated MADRS did not show a significant time-by-treatment interaction, the authors note a moderate between-group effect size in favour of CBT at study end. Self-reported symptoms (QIDS) did demonstrate a statistically significant time-by-treatment interaction, suggesting greater sustained improvement with CBT. Two mechanistic pathways are proposed to explain how CBT might extend ketamine's effects. First, ketamine may rapidly improve cognitive control and reduce dysfunctional information processing, with CBT then consolidating and maintaining those cognitive gains. Second, ketamine may induce a transient state of enhanced neuroplasticity in regions such as the medial prefrontal cortex and hippocampus, increasing sensitivity to psychotherapy-driven learning and allowing CBT to stabilise synaptic changes and reduce relapse risk. The authors cite preclinical and early clinical evidence that ketamine can rapidly modify cognitive flexibility, dendritic architecture, and synaptic function, and that these changes could plausibly interact with psychotherapeutic learning. The paper emphasises several important limitations. The ketamine phase was open-label and there was no placebo lead-in. Sample size was small, which may account for discrepancies between clinician-rated and self-report outcomes. Participants could not be blinded to CBT versus TAU, and MADRS interviews were performed by telephone to preserve rater blinding, potentially limiting the assessments' richness. Medication regimens were not tightly controlled given the chronic and severe nature of the sample. Safety concerns are acknowledged: serious adverse events reflected relapses of depression and suicidal ideation, and the authors note the possibility that abrupt discontinuation of ketamine might produce a rapid return of symptoms akin to withdrawal in some cases; they suggest future trials might test tapered discontinuation strategies. Concluding, the investigators recommend larger, well-powered randomised trials to confirm whether CBT plus ketamine improves long-term outcomes compared with ketamine alone, to test the proposed mechanisms, and to optimise timing and delivery of CBT relative to ketamine administration.
In their final remarks, the authors state that this pilot randomised trial provides preliminary support for the idea that CBT administered after a course of intravenous ketamine may sustain antidepressant effects in treatment-resistant depression. They propose that ketamine and CBT could act synergistically—ketamine enhancing metaplasticity and CBT stabilising therapeutic gains—or that ketamine may rapidly correct maladaptive cognitive patterns that CBT then maintains. The authors recommend further investigation of this sequential approach in adequately powered clinical trials.