Antisuicidal and antidepressant effects of ketamine and esketamine in patients with baseline suicidality: A systematic review

Suicide is a major global public-health problem, with around 800,000 deaths annually and many more attempts. Mood disorders account for a large proportion of suicides, and active suicidal ideation (SI) is a recognised predictor of suicide. Despite this, there are few evidence-based pharmacotherapies shown to reduce SI rapidly; clozapine and lithium have some evidence, but other readily applicable agents are lacking. Ketamine, an NMDA receptor antagonist, and its S-enantiomer esketamine have shown rapid antidepressant effects at subanaesthetic doses, and prior meta-analyses and trials have reported reductions in SI following single doses or short courses. However, most clinical trials have excluded people with moderate-to-high baseline suicidality for ethical and safety reasons, leaving uncertainty about efficacy and safety in this high-risk subgroup. Siegel and colleagues set out to review the literature specifically addressing whether intranasal esketamine or intravenous ketamine reduce SI in patients who had SI at baseline. The review aimed to evaluate both antisuicidal and antidepressant effects in this population and to assess the safety of including patients with active SI in clinical trials. By focusing on studies that required SI at enrolment, the authors intended to clarify efficacy and tolerability in a subgroup that is under-represented in standard RCTs and for whom rapid interventions are particularly clinically important.

The investigators conducted a systematic review following PRISMA guidance. Electronic searches of EMBASE, PsycINFO and PubMed were performed from database inception to July 2020, with an additional search on 21 September 2020 that incorporated suicidality terms (suicid* OR "suicidal ideation" OR "suicidal behaviour"). A manual citation search of included articles was also undertaken. Eligibility criteria encompassed original interventional and observational studies (clinical trials, retrospective studies, open-label trials, and case reports/series) in which ketamine or esketamine was used to treat depressive symptoms and where having SI at baseline was an explicit inclusion criterion. Titles and abstracts were screened for relevance and full texts were reviewed. Data were extracted using a structured form capturing author/year, study type, sample size, demographics and diagnoses, suicidality criteria, drug formulation and dosing, concurrent medications, outcome measures, suicidality and depression outcomes, and serious adverse events related to suicide or depression. Quality appraisal of randomised controlled trials (RCTs) followed domains recommended in the Cochrane Handbook: randomisation process, deviations from intended interventions, missing outcome data, outcome measurement, and selection of reported results. Domains were rated as high or low risk of bias depending on adequacy of reported protocols. Retrospective and case studies were included for comprehensiveness but were recognised to have lower methodological rigour by virtue of non-randomisation and lack of blinding.

Searches identified 762 records; after removing 129 duplicates and screening titles/abstracts, 202 full texts were assessed. From an initial set of 46 papers relating to special populations, and after additional searches focused on suicidality, 15 studies met inclusion criteria for this review. Study types and pooled sample sizes reported in the Discussion comprised three double-blind, placebo-controlled RCTs of intranasal esketamine (total n receiving esketamine = 263), five RCTs of intravenous (IV) ketamine (pooled n = 88), four retrospective open-label ketamine trials (pooled n = 126), and three case reports (pooled n = 3), yielding 480 individuals treated with esketamine or IV ketamine across the included studies. Esketamine RCTs: Three double-blind, placebo-controlled trials enrolled patients with major depressive disorder and active SI. Dosing in several trials used 84 mg intranasal esketamine administered twice weekly for up to four weeks, with standard-of-care oral antidepressants initiated or optimised at randomisation and initial dosing often administered in emergency or inpatient settings. Primary endpoints varied (MADRS change at 4 h or 24 h). Across trials, depressive symptoms measured by MADRS improved significantly more at 4 h and/or 24 h post-dose in esketamine versus placebo (one trial reported a between-group MADRS reduction at 24 h with p = 0.006). By later follow-up (day 25 to day 90 depending on study) group differences in MADRS were no longer significant. Measures of suicidality (for example the CGI-SS-r and clinician-rated assessments) showed rapid improvement in both treatment and placebo groups, but between-group differences were generally non-significant at most time points; an exception was the single MADRS suicidal-thought item at 4 h in one trial. Treatment-emergent serious suicide-related events occurred in several trials during double-blind and follow-up phases (attempts and hospitalisations), but investigators commonly deemed these events unrelated to esketamine. IV ketamine RCTs: Five double-blind RCTs used midazolam as an active comparator. The largest trial (N = 80, MDD with SSI ≥4) administered ketamine 0.5 mg/kg over 40 min and found a 24-h SSI advantage favouring ketamine: an average 4.96-point lower SSI score versus midazolam (p ≤ 0.001), with responder rates of 55% versus 30% at day one and benefits evident as early as 230 min post-infusion. Other RCTs showed mixed results: a bipolar depression trial (N = 16) had large numerical SSI reductions with ketamine but did not reach statistical significance (p = 0.074); a heterogeneous mood/anxiety disorders trial (N = 24) found a significant effect on BSSI at 48 h (p = 0.047) but not at the 24-h primary endpoint; a repeat-dose trial (N = 26, chronic SI) using twice-weekly infusions for three weeks reported no group differences on the C-SSRS, although some participants achieved absence of SI after treatment. One RCT suggested that patients with anxious depression showed greater antisuicidal and antidepressant responses to ketamine compared with non-anxious patients, indicating a possible moderator. Serious adverse events related to suicidality or depression were reported in four of the five RCTs, including suicide attempts, inpatient admissions and, in two cases, suicides occurring months after infusion; causal attribution to ketamine was generally uncertain or judged unrelated. Retrospective and open-label studies: Seven retrospective/open-label reports (including case series and single cases) generally administered 0.5 mg/kg IV ketamine, sometimes increasing to 0.75 mg/kg for later infusions. These studies frequently documented rapid reductions in SI and depressive symptoms within hours to days, with varying response and remission rates. For example, an open-label cohort (N = 86) reported antisuicidal response rates of 57% at 24 h and 70.6% at two-week follow-up. However, durability was limited for many patients: several reports noted relapse within weeks after completing the infusion series. Case reports described rapid but transient benefit in refractory cases, and one case maintained improvement with adjunctive memantine. Across studies, common non-serious adverse events were those typically seen with ketamine/esketamine (headache, nausea, transient dissociation). When pooled, the proportion of patients experiencing treatment-emergent serious events related to suicidality or depression ranged up to 3% in treatment groups and up to 2% in control groups during double-blind phases, according to the authors' tabulation.

Siegel and colleagues concluded that evidence differs between intranasal esketamine and IV ketamine when trials specifically enrol patients with SI at baseline. The three esketamine RCTs consistently showed greater rapid antidepressant effects at 24 h post-dose compared with placebo, but did not demonstrate superior antisuicidal effects versus placebo on clinician-rated suicidality scales, except for isolated findings on a single MADRS suicidal-thought item at very early time points. This pattern led the authors to suggest that, in these trials, antidepressant and antisuicidal effects may be dissociable. Results for IV ketamine were more mixed. Several RCTs and open-label studies reported rapid reductions in SI within hours to 24 h, and the largest RCT found a statistically and clinically meaningful reduction in SSI at 24 h. Nevertheless, not all trials found significant between-group differences, and retrospective studies pointed to frequently short-lived benefits after cessation of treatment. The authors highlighted a substantial placebo or non-specific treatment effect in inpatient settings—admission and the therapeutic milieu themselves appear to improve suicidality and may have contributed to high placebo response rates, complicating detection of drug effects in this population. On safety and feasibility, most patients with baseline SI were able to participate in trials and complete procedures, with adverse events largely consistent with known drug profiles. Treatment-emergent serious events related to suicidality were observed across some trials, but the rates were low during double-blind phases and investigators often judged events unrelated to study drug. The authors emphasised limitations of the evidence base: small sample sizes, scarcity of RCTs that include people with active SI, and lack of long-term follow-up to determine durability of antisuicidal and antidepressant effects. They recommended future research with longer follow-up, exploration of adjunctive strategies to sustain benefit (memantine was noted as an example from a case report), and consideration of concurrent psychotherapeutic interventions. Finally, the review supports the continued inclusion of patients with SI in clinical research, while calling for careful ethical and safety procedures to enable rigorous study in this high-risk group.