Oral esketamine in patients with treatment-resistant depression: a double-blind, randomized, placebo-controlled trial with open-label extension

Major depressive disorder (MDD) is a leading cause of global disease burden and about one third of patients do not achieve adequate response to standard pharmacological and psychological treatments, constituting treatment-resistant depression (TRD). Intravenous and intranasal forms of (es)ketamine have shown rapid but often short-lived antidepressant effects; repeated dosing can prolong benefit but intravenous administration and supervised intranasal dosing have limitations of feasibility, accessibility, and cost. Oral administration is widely used for other indications and could offer greater scalability and acceptability, but evidence for oral (es)ketamine in TRD is limited and largely derived from small, uncontrolled studies or studies at high risk of bias. Smith-Apeldoorn and colleagues therefore conducted a multicentre, double-blind, randomized, placebo-controlled, parallel-group trial to test the antidepressant efficacy, safety, and tolerability of fixed low-dose oral esketamine in patients with severe TRD, with an optional open-label extension offering individually titrated, higher-dose oral esketamine. The primary outcome was change in depressive symptom severity (HDRS17) over six weeks; secondary aims included systematic assessment of adverse events and functional outcomes. The open-label extension was included to evaluate flexible higher-dose oral esketamine in the same participant population.

The trial was approved by a medical ethics committee and registered. It was performed at three psychiatric centres in the Netherlands between February 2017 and February 2021 and comprised three phases: a six-week double-blind RCT of fixed low-dose oral esketamine versus matched placebo (both added to participants' ongoing antidepressant medication); a four-week post-treatment wash-out; and an optional six-week open-label phase of individually titrated higher-dose oral esketamine (twice weekly) for RCT completers who opted in. Eligible participants were aged 18–80 years, met DSM-5 criteria for a current major depressive episode (MINI interview), scored >18 on the HDRS17, were on a stable therapeutic antidepressant dose, and met the study definition of TRD (insufficient lifetime response to three or more different antidepressant classes given for at least four weeks at adequate dose). Key exclusions included lifetime psychotic disorder, bipolar disorder, moderate-to-severe substance-use disorder, recent use of non-prescribed psychoactives or high-dose benzodiazepines, and active suicidal intent. Recruitment involved psychiatric departments and public advertisement. Randomisation was 1:1 by computer-generated blocks of eight; placebo capsules were matched in appearance and blisters were identically labelled. Participants, clinicians, and study personnel remained blind until study completion; blinding success was assessed with a post-treatment guess form. During the RCT, intervention participants took esketamine capsules three times daily (08:00, 14:00, 20:00) for 42 days, with titration from 30 mg/day to 90 mg/day over the first three days and tapering over the final three days (maximum daily dose 90 mg divided across three doses). Control participants received matched placebo capsules. All participants were hospitalised for the first five days for supervised dosing and monitoring, then self-administered capsules at home for the remaining 37 days, with compliance checks at visits. Primary outcome was change in total HDRS17 score from baseline to six weeks; secondary efficacy outcomes included response (≥50% HDRS17 reduction), partial response (25–49% reduction), IDS-SR, CGI severity and improvement scores, and EQ-5D-5L health-related quality-of-life index. Safety and tolerability were systematically monitored: general adverse events via SAFTEE, dissociation via the Dissociation Tension Scale (DSS), psychotic experiences via the Questionnaire for Psychotic Experiences (QPE), and sleep disturbances via the Iowa Sleep Disturbance Inventory (ISDI). Blood pressure was assessed daily for the first five days and at subsequent visits, with clinically relevant increases predefined (SBP ≥30 mmHg, DBP ≥15 mmHg). Liver enzymes, weight, and other laboratory measures were assessed at specified visits; predefined thresholds for clinically relevant liver test abnormalities were applied. Statistical analysis used IBM SPSS v28. Baseline comparisons employed t-tests, Fisher's Exact Tests, or Mann–Whitney U tests as appropriate. The primary longitudinal analysis used a linear mixed model (LMM) with restricted maximum likelihood estimation (REML), including a random intercept, fixed effects for time (modelled categorically as dummy variables) and the treatment-by-time interaction, thereby adjusting for baseline HDRS17 implicitly via the intercept. Random slopes were not fitted. Two-tailed tests with significance at p<0.05 were specified; sensitivity analyses are reported in supplementary materials.

Screening and retention: Of 360 patients screened, 247 were excluded (187 not meeting inclusion criteria, 22 declined, 38 other reasons). One hundred and thirteen participants were randomised; two in the placebo arm were excluded before treatment start, leaving 111 participants who received study drug. Most participants completed the six-week RCT (87.4%) and the four-week wash-out (84.7%). Seventy-two participants entered the open-label programme; HDRS17 data for the open-label phase were available for 59 participants. Baseline clinical profile: Participants had severe, chronic illness: mean current episode duration 52.5 months, mean number of failed antidepressant treatments 6.3, 77.5% had insufficient response to psychotherapy, 40.9% had insufficient response to electroconvulsive therapy, and 64.9% had at least one comorbid axis I disorder. Treatment groups were comparable on demographic and clinical characteristics. Blinding: At end of the RCT, 95.5% of participants completed a guess form. In the esketamine arm 55.6% correctly guessed they had esketamine; in the placebo arm 57.7% correctly guessed placebo. The proportion choosing “I don’t know” was similar across arms. Primary efficacy (RCT): Mean HDRS17 decreased from 23.2 (SD 3.48) to 21.4 (SD 5.79) in the fixed low-dose esketamine arm and from 23.2 (SD 3.49) to 20.4 (SD 6.47) in the placebo arm. The LMM found no significant treatment-by-time interaction overall (F=0.65, df=4, 210, p=0.626) and no significant interaction for individual time points; sensitivity analyses were consistent. Response rates at six weeks were 4.0% in the esketamine arm versus 10.4% in the placebo arm (χ2=1.52, p=0.264). Partial response rates were 18.0% versus 12.5% (χ2=0.57, p=0.577). Secondary measures (IDS-SR, CGI, EQ-5D-5L) showed no statistically significant differences between arms. Safety and tolerability (RCT and wash-out): Two serious adverse events of suicide occurred, one in each arm (one in week 1 in the placebo arm, one in week 4 in the esketamine arm); neither had active suicidal intent at baseline per HDRS17 item 3. Six participants discontinued due to adverse events (four in the esketamine arm, two in placebo). Three esketamine participants requested dose reduction because of adverse effects. Of 55 potential SAFTEE items, “dizziness or faintness” and “dizziness when standing up” were more often moderately to severely increased in the esketamine arm. No significant differences were found in dissociation (DSS), psychotic experiences (QPE), or ISDI-assessed sleep disturbances except for a transient increase in sleep hallucinations after one week in the esketamine arm; this difference was not present at six weeks. Blood pressure, weight, and liver enzyme findings are reported in the supplement; the main text states no major safety or tolerability problems attributable to esketamine beyond the noted items. Open-label programme: Seventy-two participants entered open-label individually titrated treatment with oral esketamine twice weekly at doses from 0.5 to 3.0 mg/kg. The distribution of highest administered doses among participants was: below 2.0 mg/kg for 22 participants (30.6%), 2.0 mg/kg for 16 participants (22.2%), between 2.0 and 3.0 mg/kg for nine participants (12.5%), and 3.0 mg/kg for 24 participants (33.3%); dosing data were missing for one participant (1.4%). For the 59 participants with HDRS17 data, mean score decreased from 21.0 (SD 5.09) at open-label baseline to 15.1 (SD 7.27) at end-of-treatment (mean difference -6.0, 95% CI -7.71 to -4.29, p<0.001). In the intention-to-treat sample of 72, 18 participants (25.0%) met response criteria and nine (12.5%) met partial-response criteria. Six participants dropped out of the open-label phase before end-of-treatment (four due to adverse events, two due to non-response), and data were missing for seven participants.

Smith-Apeldoorn and colleagues conclude that fixed low-dose oral esketamine (30 mg three times daily for six weeks, maximum 90 mg/day divided TID) did not reduce depressive symptoms more than placebo in this sample of patients with severe TRD. The null RCT result contrasts with some prior reports of antidepressant effects with low-dose or fixed-dose oral and sublingual ketamine regimens; the authors note that earlier positive findings often came from small or uncontrolled studies or from samples that were less treatment resistant than the present cohort. Several explanations for the negative RCT outcome are offered. First, participants were severely treatment resistant (long episode duration, multiple failed medication trials, many with inadequate response to ECT), which reduces probability of response to subsequent interventions. Second, the dosing regimen divided daily dose into three administrations to limit side-effects, which likely produced lower peak esketamine blood levels than less frequent, higher-dose schedules used in other studies; oral bioavailability varies between individuals and a single low fixed dose may not be optimal for all patients. Third, unresolved questions remain about potential differences in efficacy between ketamine enantiomers and whether racemic ketamine might perform differently from esketamine, though preliminary clinical data do not clearly favour one enantiomer over the other. The open-label, flexible higher-dose programme produced a clinically meaningful average HDRS17 reduction (mean ≈6 points) and a 25% response rate in the intention-to-treat sample, which the authors interpret as supportive of potential antidepressant effects at individually titrated higher oral doses. However, they caution that the open-label phase lacked blinding or a control, so expectancy and other non-specific effects may have contributed to benefit, although participants had already experienced a failed blinded low-dose trial and extensive prior treatment failures which might limit expectancy effects. Regarding safety, the study found no major safety or tolerability signal attributable to oral esketamine beyond higher rates of dizziness-related complaints and a transient increase in sleep hallucinations; serious adverse events included one suicide in each arm. Strengths noted by the authors include the relatively large randomised sample, systematic outcome and safety assessments, and preservation of blinding for many participants. Key limitations highlighted are the open-label nature and lack of control in the flexible-dose phase and the possibility that the fixed low-dose regimen produced insufficient plasma exposure in some participants. The authors recommend further controlled studies of flexible-dose oral esketamine to better define efficacy, optimal dosing, safety, and tolerability.