Ketamine for suicidality: an umbrella review

The investigators conducted a comprehensive search across databases and registries and supplemented this with expert contact and web searches; PROSPERO and the Research Registry were searched on 29 October 2021 and the umbrella review completed on 29 November 2021. Duplicates were removed using EndNote X9 and manual checking. All authors independently screened citations and assessed full texts in parallel, resolving disagreements by three‑person discussion. Inclusion criteria required publications to be systematic reviews or meta‑analyses published as full papers in peer‑reviewed journals, to include the terms ketamine and suicid* in title/abstract/keywords, and to investigate the efficacy of ketamine for suicidal ideation and/or behaviours in humans. Reviews were excluded if they were preclinical only, not full papers (for example abstracts), lacked clear methodology, mixed preclinical and clinical evidence without separation, or included fewer than two human core studies addressing ketamine and suicidality. There was no language restriction; corresponding authors were contacted when necessary. Data extraction was performed independently by all authors using Review Manager 5.4, and discrepancies were resolved by consensus. Extracted items included first author, year, country, number of eligible core studies on suicidality, patient numbers and follow‑up (when reported), how methodological quality of core studies was assessed, intervention details, meta‑analytic results (if present), evaluation of outcomes, authors’ final inference on effectiveness, adverse events and funding sources. The primary outcome for the umbrella review was the included reviews’ final inference about ketamine effectiveness for suicidal ideation or behaviours; secondary outcomes included adverse events and comparative effects of esketamine. Quality appraisal of included systematic reviews used AMSTAR‑2 (a 16‑item tool with seven major items). Reviews were classified as high, moderate, low or critically low quality according to AMSTAR‑2 guidance: no or only minor flaws = high; multiple minor but no major flaws = moderate; one major flaw = low; more than one major flaw = critically low.

The search retrieved 274 citations; after duplicate removal 131 records were screened and 27 systematic reviews met the inclusion criteria. Reasons for exclusion at screening and full‑text stages included editorials, non‑clinical studies, protocols, conference abstracts, non‑systematic reviews and studies that did not address the intervention/outcome of interest. The included reviews are described in summary tables (as reported in the extracted text). On AMSTAR‑2 quality assessment, 17 reviews were judged critically low, 2 low, 2 moderate and 6 high quality. Item 10 (reporting funding sources of included studies) scored poorly across reviews, whereas items 8 and 16 (reporting details of included studies and funding sources of the review itself) were more consistently satisfied. Overlap between reviews was noted; for example, one review (Rosenblat et al.) included only case reports and no clinical trials, and another (Lima et al.) included only meta‑analyses. Narrative synthesis was organised around whether suicidal ideation (SI) was treated as a primary or secondary outcome. Fifteen reviews had SI as a primary outcome; 12 of these focused on adult major depressive disorder (MDD) and/or bipolar depression (BD) populations, with or without treatment‑resistant depression (TRD). All reviews emphasised short‑term and acute effects because long‑term data were scarce. Of the reviews with SI as primary outcome, five were rated high or moderate quality and the rest were low or critically low. Across reviews that examined racemic IV ketamine, multiple reports described rapid reductions in suicidal ideation within hours, with some effects persisting up to 72 hours and, in pooled analyses reported by some authors, up to 2 weeks. Several randomised controlled trials (RCTs) showed early improvements (within the first few hours); however, only a minority of RCTs demonstrated statistically significant superiority over comparators at specific time points, and placebo/active‑comparator response rates were sometimes high—possibly because hospital admission itself may be an effective intervention. Subgroup analyses in some reviews suggested IV racemic ketamine produced more pronounced acute antidepressant effects than intranasal (IN) esketamine, but differences specifically for suicidality were inconsistent. Evidence for esketamine was more limited and mixed. Of the reviews that considered esketamine (nine reviews in total, per the Discussion), about half reported significant antisuicidal findings, but several of those positive conclusions relied on case reports or on the same single esketamine core trial. Other reviews and RCTs reported marginal or no significant superiority of esketamine over placebo/control for suicidal ideation despite effects on depressive symptoms. When SI was a secondary outcome (12 reviews), the pattern was similar: acute reductions in SI after ketamine were commonly reported, with effects often emerging by 4 hours and sometimes lasting days, but longer‑term benefits were not consistently documented. Evidence in adolescents and older adults was limited to small numbers of RCTs and case reports; some case reports described sustained benefit after multiple infusions, while the single RCTs produced mixed results. One review (Dean et al.) found no significant pooled effect for ketamine or esketamine on SI at any time point, based on the trials it included. Pooling only higher‑quality reviews (eight medium/high quality) left a cautiously positive picture: four of five reviews that treated SI as a primary outcome suggested IV ketamine has potential for treating SI, while one reported mixed results. Subgroup and meta‑regression analyses reported in higher‑quality reviews generally did not find significant differences in ketamine efficacy by age, gender, TRD status, unipolar versus bipolar diagnosis, or monotherapy versus adjunctive therapy. Adverse events reported across reviews were usually transient and included increased pulse rate and blood pressure, dissociation, agitation/confusion, blurred vision, nausea, vomiting, dizziness, drowsiness, vertigo, and dysgeusia. Most side effects were described as mild and short‑lived. The extracted text did not provide pooled incidence rates or long‑term safety data.

Shamabadi and colleagues interpret the body of evidence as indicating preliminary support for the short‑term, rapid antisuicidal effects of ketamine—particularly IV racemic ketamine—but emphasise that the evidence base is limited by heterogeneity and frequent methodological shortcomings in existing reviews. The majority of included reviews reach positive conclusions about acute efficacy, but the authors note that long‑term effectiveness and safety remain largely unknown. Esketamine evidence is described as inconclusive: among nine reviews that addressed esketamine, only about half reported significant antisuicidal effects, and several of those positive findings were based on case reports or a single shared core trial. The discussion highlights conceptual issues raised in the literature: some studies suggest ketamine’s antisuicidal effects may be partly independent of its antidepressant effects, because reductions in suicidal ideation have been observed in the absence of concurrent antidepressant response in some cases. However, the authors also acknowledge many other known contributors to suicide risk (hopelessness, comorbid mental disorders, impulsivity, substance misuse, personality traits) that were not consistently accounted for in core studies, limiting causal inference. Safety and feasibility considerations are reviewed: low‑dose ketamine appears to be effective, tolerable and reasonably safe in the short term, with rapid onset of action lasting days to about a week in many reports. Concerns discussed include psychotomimetic effects, potential for misuse or dependence, neurotoxicity, cognitive effects and cardiovascular adverse events, all of which complicate long‑term administration. Given these unknowns and the current evidence quality, the authors suggest that other established options may be preferable for long‑term management while ketamine may be considered for acute risk reduction where rapid effect is needed. Key limitations of the umbrella review itself are recognised: substantial heterogeneity across included systematic reviews (populations, diagnostic mixes, outcome definitions and measurement methods), inclusion of reviews that treated suicidality as a secondary outcome, and the generally low methodological quality of many reviews. The authors recommend future primary studies and systematic reviews adhere to methodological standards (for example AMSTAR‑2 guidance) and call for more RCTs and longer‑term safety and effectiveness data to clarify ketamine’s role in managing suicidal ideation and behaviours.

Most included systematic reviews reported that ketamine has rapid, short‑term efficacy for reducing suicidal ideation in acutely suicidal patients, with IV racemic ketamine most commonly implicated. Of the 27 reviews included in this umbrella review, only a small number reported mixed or negative results, although the extracted text is inconsistent about whether that number is three or four. Evidence concerning esketamine is inconclusive: nearly half of reviews reporting on esketamine described no clear antisuicidal benefit. Reported adverse effects were generally transient (for example increased pulse and blood pressure, dissociation, nausea and dizziness), and long‑term efficacy and safety remain uncertain. The authors conclude that, because of heterogeneity and frequent methodological limitations in the evidence base, further high‑quality research is required to establish longer‑term benefits, comparative effectiveness and safety.