Use of ketamine and esketamine for depression: an overview of systematic reviews with meta-analyses

Depression is a prevalent, disabling psychiatric disorder characterised by persistent low mood, anhedonia, cognitive and somatic symptoms, and sometimes suicidal ideation. Standard antidepressant treatments and electroconvulsive therapy can be effective but typically require days to weeks to achieve response, and many patients experience suboptimal outcomes. Early studies identified modulation of the glutamatergic system—specifically NMDA receptor antagonism—as a potential rapid-acting mechanism; subsequent work has shown that subanaesthetic doses of ketamine (a racemic mixture) and esketamine (S-ketamine) can produce rapid antidepressant effects and may also involve opioid-system activation. Intranasal esketamine has regulatory approval in some jurisdictions, but off‑label and clinical use of these agents has expanded despite remaining uncertainties over longer-term efficacy and safety. This overview sought to summarise the available evidence from systematic reviews with pairwise meta-analyses concerning the efficacy and safety of ketamine and esketamine for depression in adults. In addition to aggregating reported effect estimates, the investigators examined the methodological quality of included reviews to inform the reliability of the synthesized evidence and identify gaps for future research.

A protocol for the overview was prepared in advance (available in an appendix). The investigators conducted a comprehensive search of Medline (via PubMed), LILACS, the Cochrane Library, and the Centre for Reviews and Dissemination (CRD) up to 29 November 2020, using MeSH terms and keywords for depression, ketamine/esketamine, and systematic reviews with meta-analysis. Reference lists of included articles were also screened. Detailed search strategies are reported in an appendix. Eligibility required that publications be systematic reviews with pairwise meta-analyses of randomised controlled trials in adults with major depressive disorder or bipolar depression, published in English, Spanish, or Portuguese; they had to evaluate ketamine and/or esketamine (any route, dose, frequency, or as monotherapy/adjunct) versus placebo or other drugs and report efficacy or safety outcomes. Exclusions covered narrative reviews, systematic reviews without meta-analysis, network meta-analyses, reviews combining ECT with ketamine/esketamine as the intervention, reviews with different target populations or study designs, and reviews without available full text. Study selection proceeded in three stages (duplicate removal, title/abstract screening, and full-text review) with two investigators performing selection independently and a third resolving disagreements. Data extraction captured review characteristics and effect estimates; reported meta-analytic measures (mean difference, standardised mean difference, relative risk, odds ratio) and 95% confidence intervals were extracted as presented by the review authors. Methodological quality of included reviews was appraised using AMSTAR-2. The authors synthesised characteristics and methodological assessments descriptively in structured tables and assessed overlap of primary RCTs across reviews.

The searches identified 118 records; after screening and full-text review, 11 systematic reviews with pairwise meta-analyses met inclusion criteria. Included reviews were published between 2015 and 2020 and collectively covered primary RCTs in major depressive disorder and bipolar depression. Most reviews examined intravenous ketamine as the intervention; esketamine was typically evaluated via the intranasal route. Common outcome measures included MADRS and HDRS/HAM-D; some reviews also used BPRS, CADSS, and PHQ-9. Six reviews assessed outcomes such as suicidality, acceptability, disability, and adverse events. Clinical response: Multiple reviews reported that ketamine produced a significantly greater clinical response than control at very early time points (40 min, 80 min, 2 h, 4 h) and at 24 h in seven reviews. Significant effects for ketamine were also reported at 48 h, 72 h (in several reviews), and at 1 week in four reviews, but evidence was inconsistent for certain subgroups: very‑low‑dose ketamine often showed no benefit, oral ketamine did not show significant effects, and some analyses in bipolar depression failed to show benefit at later time points. Esketamine demonstrated statistically superior clinical response versus placebo across time points from 2 h up to 4 weeks (evidence described as high quality in the reviews that assessed it). Clinical remission: Ketamine showed significant superiority over control at acute time points (80 min, 2 h, 4 h) and in many analyses at 24–72 h; however, findings were inconsistent at and beyond 1 week, and oral ketamine did not show significant remission benefits. Esketamine again was reported as statistically superior to placebo for remission across measured time frames between 2 h and 4 weeks (high-quality evidence in the relevant reviews). Depression rating scales: Several reviews found ketamine improved HDRS/HAM-D and MADRS scores at 24 h and 1 week, with some benefit persisting up to about 3 weeks in pooled analyses; results were heterogeneous and limited for very‑low doses and bipolar depression. Esketamine was reported to produce significant improvements on MADRS and PHQ-9 (high-quality evidence reported by the reviews that assessed these outcomes). BPRS and CADSS (dissociation-related measures) were also reported to improve in some analyses with ketamine. Suicidality, acceptability, disability, and adverse events: Suicidal ideation was reported to be reduced versus control up to 72 h in some reviews (evidence rated low quality). Acceptability (dropout rates) findings were mixed: three reviews found no difference in overall dropout or dropout due to lack of efficacy, whereas one review found higher dropout due to adverse events with esketamine (high-quality evidence). Only two systematic reviews with meta-analyses reported adverse-event meta-analyses: oral ketamine did not show a significant increase in adverse events versus control, but intranasal esketamine was associated with significantly more adverse events and more dropouts due to adverse events, including dissociation, dizziness, hypoesthesia, vertigo and other events (high-quality evidence cited). The authors noted that many RCTs did not present usable adverse-event data, limiting pooled safety assessment and leaving acute (up to 2 weeks) and long-term safety incompletely characterised. Methodological quality and overlap: Using AMSTAR-2, three reviews were rated "low quality" and eight as "critically low quality." Common deficiencies included failure to report funding sources for included primary studies, lack of protocol registration or an a priori design in several reviews, and generally not assessing the potential impact of risk of bias in primary RCTs on pooled estimates. Two pairs of ketamine reviews included overlapping RCTs but sometimes reported differing sample sizes or event counts for the same outcomes; no overlap was identified across esketamine reviews. Across the included literature, the authors counted approximately 20 RCTs evaluating ketamine and four RCTs evaluating esketamine.

De Mendonça Lima and colleagues conclude that both ketamine and esketamine produce rapid antidepressant effects, often apparent within hours and lasting days to a few weeks. Esketamine had consistent evidence of superiority to placebo on response, remission and several scales out to 4 weeks in the reviews that assessed it, while ketamine showed substantial short-term benefits for response and remission up to about 72 h and, in many analyses, up to 1 week. The overview notes that effect sizes reported for ketamine and esketamine in these systematic reviews tended to be larger than pooled effects seen for standard antidepressants at around 8 weeks in an external network meta-analysis cited by the authors, highlighting the rapid onset of action as a potentially important clinical advantage. At the same time, important uncertainties remain. The reviewers emphasise that evidence on longer-term efficacy and safety is lacking: most RCT data synthesised by the included reviews cover up to 2 weeks for ketamine and up to 4 weeks for esketamine. Safety data were incomplete across RCTs, with only limited meta-analytic pooling for adverse events; where pooled data exist, intranasal esketamine was associated with more adverse events and greater discontinuation due to adverse events. The authors also highlight concerns about selective reporting and under‑assessment of long-term harms in primary studies. Methodological limitations of the existing review literature were underscored: most included systematic reviews were rated low or critically low on AMSTAR-2, frequently failing to report funding sources of primary studies, register protocols, or assess the impact of study-level risk of bias on results. The overview therefore recommends more primary RCTs with longer follow-up, direct head-to-head comparisons between ketamine and esketamine and against standard antidepressants, exploration of different doses and routes, and better reporting of adverse events. Future systematic reviews should adhere to established methodological standards—pre-registered protocols, assessment and reporting of funding sources, consideration of risk of bias in interpretation, and adequate investigation of publication bias—so that more reliable syntheses can inform clinical and policy decisions. The authors acknowledge limitations of their overview: restricting inclusion to systematic reviews with meta-analysis may have omitted relevant narrative or qualitative syntheses; reviews combining ECT with ketamine/esketamine were excluded, possibly omitting pertinent data; grey literature and ongoing or unpublished reviews were not sought; and the overview relied on review authors' own assessments of evidence quality where available. They also did not calculate prediction intervals to quantify between-study variation.

The overview found that ketamine and esketamine show significant short-term superiority to control in many measures of clinical response, remission, depression rating scales and suicidal ideation, with effects often evident within hours and persisting for days to a few weeks. However, the evidence base is limited to acute treatment windows (generally up to 2 weeks for ketamine and 4 weeks for esketamine), adverse-event data are incompletely reported, and most systematic reviews were of low or critically low methodological quality, which reduces confidence in the pooled estimates. The authors call for more high-quality primary trials with longer follow-up, head-to-head comparisons and better safety reporting, and for future systematic reviews to follow rigorous methodological standards so that clinical and policy decisions about ketamine and esketamine for depression can rest on more reliable evidence.