Efficacy of ketamine in the rapid treatment of major depressive disorder: a meta-analysis of randomized, double-blind, placebo-controlled studies

Major depressive disorder (MDD) is a common and disabling psychiatric condition for which many patients do not achieve remission with existing antidepressants; about 33% do not respond to available treatments. Ketamine, an N-methyl-d-aspartate (NMDA) receptor antagonist used as an anaesthetic, has attracted attention because several clinical studies and earlier reviews suggested it can produce a rapid antidepressant effect, including in treatment-resistant depression (TRD). A prior meta-analysis up to December 2013 included only four trials, and more randomized controlled trials (RCTs) have since been published. Han and colleagues therefore conducted an updated meta-analysis to synthesise randomized, double-blind, placebo-controlled trials comparing intravenous ketamine with placebo in adult patients with MDD, aiming to assess ketamine's rapid antidepressant efficacy. The review focused on response and remission at three early time points (24 hours, 72 hours, and day 7) to capture the acute effects of ketamine infusion.

The investigators searched multiple international databases (PubMed, Cochrane Central Register of Controlled Trials, Web of Science, Embase) plus two Chinese databases (CBM-disc and CNKI) and relevant websites for studies published up to April 2016. Search terms were "depress*" and "ketamine"; no language restriction was applied. Reference lists, conference summaries and the WHO International Clinical Trials Registry Platform were also checked. Eligibility criteria required randomized, double-blind, placebo-controlled trials that compared ketamine with placebo in adults aged >18 years diagnosed with MDD, with mood assessed using HDRS, MADRS, or CGI. Exclusions were studies without a control group, studies of narrow or secondary depression diagnoses (for example postpartum or vascular depression), case reports and reviews, and duplicate reports. Two reviewers independently screened studies and extracted data; disagreements were resolved by consensus or a third reviewer. When data were not directly retrievable, the authors attempted to contact primary investigators. Primary and secondary outcomes were response and remission rates. Response was defined as at least a 50% reduction from baseline in HDRS or MADRS, or significant improvement on CGI; HDRS was prioritised if multiple scales were reported. Efficacy was assessed at 24 hours, 72 hours and day 7 to capture rapid effects. For handling dropouts, a worst-case scenario approach was used (assuming all dropouts were non-responders). Statistical synthesis used pooled odds ratios (ORs) with 95% confidence intervals (CIs) and the Mantel–Haenszel random effects model, reflecting the expectation of between-study variability. Heterogeneity was evaluated with the chi-square test (P≤0.10 threshold) and the I2 statistic (where higher values indicate greater heterogeneity). Because fewer than 10 studies limit the interpretability of funnel plots, the authors used Egger's test to assess publication bias. Analyses were performed with RevMan5 software and two-sided tests with P<0.05 considered statistically significant.

The search identified 169 records; after screening and exclusions, nine randomized, double-blind, placebo-controlled trials met inclusion criteria and were included in the meta-analysis, comprising a total of 368 adult patients with MDD. Study quality indicators reported were generally favourable: all nine trials conducted allocation concealment and matched baseline characteristics, and outcome assessment was blinded. Eight studies used an intravenous ketamine dose of 0.5 mg/kg and one used 0.54 mg/kg. Six trials were conducted in the United States, one in the Czech Republic and two in the People's Republic of China. Two trials used an active placebo (electroconvulsive therapy and midazolam) while the others used inactive placebo. At 24 hours (data from eight RCTs for response), 84 of 161 patients (52.2%) receiving ketamine and 11 of 140 (7.8%) receiving placebo met response criteria. The pooled OR for response at 24 hours was 10.09 (95% CI 4.96–20.52, z=6.38, P<0.00001), indicating markedly higher odds of response with ketamine; statistical heterogeneity was negligible (I2=0%, P=0.57). For remission at 24 hours (six RCTs), 26 of 126 patients (20.6%) receiving ketamine and 3 of 105 (2.8%) receiving placebo met remission criteria; pooled OR was 5.25 (95% CI 1.82–15.17, z=3.06, P=0.002), with I2=0%. At 72 hours (nine RCTs for response), 94 of 196 patients (47.9%) receiving ketamine and 23 of 172 (13.4%) receiving placebo were responders. The pooled OR was 7.42 (95% CI 3.97–13.88, z=6.27, P<0.00001) with I2=0% (P=0.90). For remission at 72 hours (six RCTs), 30 of 126 (23.8%) in the ketamine group and 5 of 105 (4.7%) in the placebo group met remission criteria; pooled OR was 4.04 (95% CI 1.66–9.85, z=3.07, P=0.002), I2=0%. At day 7 (nine RCTs for response), 78 of 196 patients (39.8%) receiving ketamine and 23 of 172 (13.4%) receiving placebo were responders. The pooled OR was 5.66 (95% CI 2.92–10.97, z=5.13, P<0.00001) with I2=0%. For remission at day 7 the extracted text reports 33 of 126 patients (26.2%) in the ketamine group; the count for the placebo group in the extracted Results section is not clearly reported. Across the reported outcome analyses heterogeneity was very low and Egger's test did not indicate significant publication bias. A subgroup analysis separated four crossover studies from five parallel-design studies. Effect sizes were broadly similar between subgroups: crossover studies showed ORs for response of 16.21, 7.19 and 5.13 at 24 hours, 72 hours and day 7 respectively, while the parallel studies showed ORs of 9.36, 7.59 and 5.96 at the same time points. The pooled results for remission by subgroup likewise followed similar patterns. Regarding safety, the authors reported that ketamine infusions were generally well tolerated and no serious adverse effects were reported in the ketamine groups within the included trials, though detailed adverse-event reporting was limited in the extracted text.

Pooling nine randomized, double-blind, placebo-controlled trials, Han and colleagues concluded that a single intravenous ketamine infusion produced rapid antidepressant effects in adults with MDD: higher response and remission rates were apparent at 24 hours and persisted through 72 hours and day 7. Reported pooled effect sizes (ORs) for response were 10.09, 7.42 and 5.66 at 24 hours, 72 hours and day 7 respectively; for remission the corresponding ORs reported were 5.25, 4.04 and 4.60. The authors noted consistency across studies, with negligible statistical heterogeneity. The investigators highlighted that over half of the pooled sample had some degree of treatment-resistant depression, suggesting ketamine may be useful in the acute management of TRD. They also discussed potential mechanisms cited in the literature, including induction of neuroplastic processes, increased synaptic signalling proteins and spine synapses in prefrontal cortex, modulation of cortical connectivity relevant to depressive symptoms, and enhancement of synaptogenesis. Mention was made of a ketamine metabolite identified in preclinical work that might produce rapid antidepressant-like effects via an NMDA-independent mechanism and potentially lack some of ketamine's side effects. Key limitations acknowledged by the authors included the relatively small total sample size, the short treatment window of 1–7 days across included trials which precluded conclusions about mid- and long-term efficacy, and inability to assess whether repeated infusions produce tolerance. Psychotomimetic effects and longer-term neurocognitive safety were incompletely assessed in the available trials; while some reports described transient memory impairment immediately after administration, other studies found no evidence of cognitive impairment. The authors also noted that eight of nine trials used the same dose (0.5 mg/kg) so optimal dosing was not established here. They recommended larger, longer-term clinical trials to address mid- and long-term efficacy, optimal dosing and safety/tolerance concerns.