Repeat-dose ketamine augmentation for treatment-resistant depression with chronic suicidal ideation: A randomized, double blind, placebo controlled trial

Major depressive disorder (MDD) is highly prevalent and a substantial minority of patients remain symptomatic despite multiple adequate treatments; approximately 30% of patients meet criteria for treatment-resistant depression (TRD). Chronic suicidal ideation (SI) is common among people with TRD and these individuals are more likely to engage in suicidal behaviour than treatment responders. Most clinical trials have excluded patients with high suicide risk, leaving a gap in evidence for interventions that might benefit this high-risk outpatient population. Ketamine, an NMDA–receptor antagonist originally developed as an anaesthetic, has attracted interest because of its rapid antidepressant and putative antisuicidal effects that appear mechanistically distinct from monoaminergic antidepressants, but evidence on repeated-dose efficacy in TRD patients with chronic, current SI is limited and mixed. Ionescu and colleagues designed a double-blind, randomized, placebo-controlled trial to test whether repeated, non-escalating intravenous ketamine (0.5 mg/kg over 45 minutes) would produce superior short-term and longer-term reductions in depressive symptoms and suicidal ideation compared with saline in medicated outpatients with severe TRD and documented chronic SI (≥ 3 months). The primary aim was to assess antidepressant efficacy using the Hamilton Depression Rating Scale (HDRS); a secondary aim was to assess antisuicidal effects using the Columbia-Suicide Severity Rating Scale (C-SSRS). The investigators hypothesised that ketamine would show rapid and sustained superiority over placebo.

This study was a double-blind, placebo-controlled, randomised trial conducted at Massachusetts General Hospital between January 2013 and November 2015. After an initial screening visit, participants were seen twice more during a pre-infusion phase to confirm ongoing eligibility and symptom stability; the third pre-infusion visit served as baseline. Eligible outpatients were randomised by a computer algorithm to receive six 45-minute intravenous infusions of ketamine (0.5 mg/kg) or saline placebo, administered twice weekly over three weeks. All clinicians, patients, and raters were blinded to allocation; a senior anaesthesiologist held the randomisation list. During infusions a physician remained present, vital signs were monitored every five minutes, and adverse events could prompt immediate discontinuation. Inclusion and exclusion criteria were detailed and intended to target severely ill, chronically suicidal outpatients. Key inclusion criteria included: age 18–65; current primary diagnosis of MDD by SCID; HDRS-28 score ≥ 20; history of ≥ 3 failed antidepressant trials in the current episode; SI for ≥ 3 months with C-SSRS SI score ≥ 1 and HDRS suicide item ≥ 2 at screening or pre-infusion visits; and a stable antidepressant regimen for ≥ 4 weeks prior to infusions. Major exclusions included pregnancy, unstable medical illness, bipolar or psychotic disorders, substance-use disorder within the past year, positive urine toxicology, ketamine abuse history, or SI requiring immediate hospitalisation. Concomitant antidepressant medication was maintained for ethical reasons; some medications with interaction risk were exclusionary. Primary and secondary outcome measures were prespecified. The primary outcome was HDRS total score (clinician-rated, 28 items); response was defined as ≥ 50% improvement. Suicidal ideation was assessed with the C-SSRS (presence/absence and an intensity rating 0–25). Dissociative effects were measured with the Clinician Administered Dissociative States Scale (CADSS) at multiple time points during infusions. Depression and SI scales were administered approximately 240 minutes (4 hours) after each infusion. After the infusion phase participants entered a naturalistic follow-up with visits every other week for three months; medication adjustments were allowed and recorded. Analyses used an intent-to-treat (ITT) approach including all randomised patients, modelled with mixed-effects models with repeated measures (MMRM) to test group-by-time interactions for HDRS and C-SSRS outcomes, controlling for baseline scores (the third pre-infusion visit). Demographic comparisons employed chi-square tests and t-tests. Statistical significance was set at p < 0.05 (two-sided) and analyses were conducted in STATA SE v12. The extracted text does not report a formal sample size calculation or target power, though the investigators anticipated large effect sizes based on prior studies.

Thirty-seven outpatients were screened, 26 met criteria and were randomised (13 to ketamine, 13 to placebo). Demographic and clinical characteristics did not differ significantly between groups on reported variables (all p > 0.05). During the infusion phase, 9/13 (69%) ketamine and 9/13 placebo participants completed all six infusions; overall 24/26 randomised patients completed at least one set of post-infusion ratings. Dropouts occurred for a variety of reasons including side effects (two ketamine patients withdrew before the second infusion and one stopped an infusion due to hallucinatory side effects), belief of receiving placebo (three placebo participants withdrew after the first infusion), and one inpatient hospitalisation in the placebo arm. Depression outcomes during the infusion phase showed a significant overall decrease in HDRS total scores across time for the whole sample (p < 0.01), but there was no significant group-by-time interaction indicating superiority of ketamine over placebo (p = 0.47). Mean HDRS scores in both groups remained above 20 at the end of the infusion series, consistent with at least moderate depression. After the final infusion, antidepressant response rates were 3/12 (25%) in the ketamine group and 4/12 (33%) in the placebo group; remission rates were 2/12 (17%) and 1/12 (8%), respectively. Differences in proportions of responders or remitters between groups were not statistically significant. Suicidal ideation measured pre-infusion (C-SSRS) decreased significantly across the first three pre-infusion visits for the whole sample (F(23.806) = 14.679, p < 0.001), suggesting an improvement during the pre-randomisation phase. There was no significant group-by-time interaction for C-SSRS SI score (p = 0.32) or SI intensity rating (p = 0.23) during the infusion phase. After the final infusion, absence of SI (C-SSRS = 0) occurred in 5/12 (42%) ketamine patients and 3/12 (25%) placebo patients, a non-significant difference. Dissociative symptoms were greater in the ketamine group: mean CADSS total scores during infusion visits were significantly higher with ketamine than placebo (F(396) = 34.514, p < 0.001), with the largest difference at 30 minutes post infusion start (p < 0.01). In the three-month naturalistic follow-up phase, two participants in each group were lost to follow-up and 14/26 (54%) completed all visits (7 per group). At three months after the final infusion, among remaining participants 2/9 (22%) ketamine patients met criteria for both antidepressant response and remission, compared with 3/11 (27%) response and 2/11 (18%) remission in the placebo group; group differences were not significant. Of the five ketamine-treated patients who achieved absence of SI at the last infusion, four maintained absence at the beginning of follow-up but only one of the nine remaining ketamine patients (11%) continued to meet absence of SI criteria at the end of follow-up. The extracted text indicates that all three placebo patients who had absence of SI at the last infusion “continued” but the sentence is incomplete in the extraction and it is not clearly reported whether they maintained absence of SI through the end of follow-up. Overall, the trial found no statistically significant advantage of repeated 0.5 mg/kg ketamine infusions over saline for depression or suicidal ideation in this sample.

Ionescu and colleagues interpret their negative findings—no superiority of repeated 0.5 mg/kg ketamine over placebo for antidepressant or antisuicidal outcomes—as potentially attributable to several factors. They emphasise the high degree of chronicity and treatment resistance in the sample: many participants had failed more than five adequate antidepressant trials in the current episode, nearly 50% had previously not responded to electroconvulsive therapy (ECT), and the mean length of the current depressive episode was 115 months (about 9.6 years). Such severe, chronic TRD may be less responsive to the commonly used ketamine dose of 0.5 mg/kg. The investigators contrast these results with their prior open-label study of escalating doses (0.5 mg/kg for three infusions then 0.75 mg/kg for three), in which significant antidepressant and antisuicidal effects emerged only after dose escalation to 0.75 mg/kg (p < 0.01; Cohen’s d = 1.01). They therefore propose that higher or escalating doses may be required for similarly ill outpatient populations. Other possible explanations raised by the authors include limited statistical power due to the small sample, and an imbalance in sex distribution between groups that, while not statistically substantiated as an effect modifier, could have influenced outcomes. Acknowledged limitations are several. Recruitment and retention difficulties produced a small sample and attrition (only 14/26 completed the entire protocol), which may have left the study underpowered. Maintaining participants on stable outpatient medication regimens throughout the infusion phase precludes ruling out interactions between ketamine and concomitant treatments. By design, the study excluded patients requiring immediate hospitalisation, so the sample reflects outpatients with chronic rather than acute SI and may not generalise to emergency or inpatient populations; this may have produced a floor effect for change in SI. Timing of assessments was another constraint: symptom ratings were obtained at 4 hours post-infusion only, with no same-day baseline or daily assessments in the week immediately after the final infusion, limiting the ability to capture acute or short-term durability of effects. The authors note strengths as well: targeting a severely ill TRD outpatient group with chronic SI, using a double-blind placebo-controlled design, and including a three-month follow-up period. They also highlight that the pre-infusion lag and multiple baseline visits revealed a significant decline in C-SSRS SI scores before infusions began, suggesting the act of study participation or baseline assessment may produce treatment-independent improvements in SI even when depressive symptoms do not change. Finally, they recommend continued investigation of ketamine’s antidepressant and antisuicidal potential, including dose-finding studies, examination of patient subtypes (for example, treatment-resistant anxious depression or SI with insomnia), and further work to identify who is most likely to benefit from ketamine-based interventions.