Ketamine and Esketamine in Acute Suicidal Presentations: Evidence, Limits, and a Conceptual Recovery-Enabling Downstaging Framework

Abstract

Objective

To synthesize evidence for intravenous racemic ketamine and regulated intranasal esketamine in adults with acute suicidal presentations; distinguish the outcomes these interventions have and have not been shown to change; and place any rapid response within a complete emergency suicide-care pathway.

Methods

A narrative review of PubMed/MEDLINE-indexed literature was updated through July 16, 2026. Searches and study selection were performed by Wiesław Jerzy Cubała and Anita Słupska. Priority was given to randomized trials, systematic reviews and meta-analyses, pivotal ASPIRE publications, current US and European product information, and studies of emergency-department safety planning, lethal-means counseling, and post-discharge care. No formal risk-of-bias assessment or pooled reanalysis was undertaken.

Results

Intravenous racemic ketamine has randomized and meta-analytic evidence for rapid, short-term reduction of suicidal ideation in selected adults, with important variation by population, comparator, measure, and follow-up. The evidence does not establish prevention of suicide attempts or deaths. In ASPIRE I and II, regulated intranasal esketamine added to hospitalization, a newly initiated or optimized oral antidepressant, and comprehensive standard care accelerated improvement in depressive symptoms; clinician-rated suicidality improved markedly in both groups and did not consistently separate from placebo plus standard care. Across both evidence streams, a post-dose score cannot substitute for dynamic risk formulation, level-of-care decisions, safety planning, lethal-means restriction, or a confirmed transition of care.

Conclusions

Intravenous racemic ketamine may produce a rapid, short-term reduction in suicidal ideation in selected adults, whereas regulated intranasal esketamine has demonstrated accelerated improvement in depressive symptoms in hospitalized patients with major depressive disorder and acute suicidal ideation with intent, without a consistent treatment-specific effect on suicidality. Neither intervention has been shown to prevent suicide attempts or deaths. We propose a conceptual, non-validated recovery-enabling downstaging framework to describe a time-limited movement from high acuity toward a state in which collaboration and definitive treatment become more feasible. The framework is intended to complement, not replace, dynamic risk formulation, safety planning, and disposition assessment.

Introduction

An acute suicidal presentation is a dynamic clinical state that requires assessment of current risk, recent changes, and factors affecting immediate safety, rather than being reducible to a diagnosis, rating scale, or expected medication response. Although ketamine and esketamine are clinically attractive because their effects may emerge within hours rather than weeks, this rapid onset can encourage overinterpretation, with improvements in depression or suicidal ideation being mistaken for reduceAI Declaration suicide risk or readiness for discharge. Intravenous racemic ketamine has been studied off label, with a relatively mature evidence base for short-term reductions in suicidal ideation. By contrast, intranasal esketamine is a regulated, product-specific intervention evaluated in the ASPIRE trials for hospitalized adults with major depressive disorder and active suicidal ideation with intent, where it was tested against placebo only as an adjunct to intensive standard care, rather than as a stand-alone rescue treatment.^1,2

Acute suicidal presentations are clinically heterogeneous. They may occur in major depressive or bipolar episodes, psychotic or mixed states, substance intoxication or withdrawal, personality-related crises, acute reactions to loss or violence, severe pain, and other syndromic contexts. Trial populations, by contrast, are usually highly selected and may exclude many features encountered in emergency psychiatry. Findings from a study conducted in one diagnostic and care context should therefore not be generalized automatically to all patients presenting with suicidal thoughts or behavior.

This review focuses on intravenous racemic ketamine and regulated intranasal esketamine because both are considered in clinical discussions of rapid intervention during acute depressive presentations with suicidality, while their evidence bases and defensible claims differ materially. Inclusion of esketamine should not be interpreted as classifying it as an established anti-suicidal treatment. The review examines the limits of the available evidence, methodological challenges in interpreting treatment effects, and the possible role of rapid symptom improvement in supporting short-term stabilization and engagement with comprehensive suicide-focused care. The proposed recovery-enabling downstaging framework is presented as a conceptual clinical and documentation model rather than a validated risk-assessment, disposition, or treatment algorithm.

Methods

This narrative review searched PubMed/MEDLINE from inception to July 16, 2026. Search terms combined ketamine, esketamine, depression or mood disorders with suicidality, suicidal ideation, suicidal behavior, suicide attempt, psychiatric emergency, hospitalization, discharge, safety planning, lethal means, and emergency care. Reference lists of pivotal trials, recent reviews, and suicide-care studies were also screened. The literature search and study selection were performed by Wiesław Jerzy Cubała and Anita Słupska.

For the pharmacological evidence stream, eligible evidence focused on adults receiving intravenous racemic ketamine or regulated intranasal esketamine in studies that enrolled acute suicidal presentations or reported suicide-specific outcomes. Priority was given to randomized comparator-controlled trials, individual-participant or aggregate meta-analyses, the pivotal ASPIRE program and its pooled or secondary analyses, and current FDA and EMA product information. Studies of chronic suicidal ideation were retained when they were informative about boundary conditions and generalizability. Evidence from oral, subcutaneous, compounded intranasal, or unsupervised racemic ketamine was not used to support efficacy claims for the acute indications reviewed here.

A separate evidence stream addressed the emergency suicide-care pathway, including dynamic risk formulation, hospitalization and discharge, safety planning, lethal-means counseling, and post-discharge transition. Evidence was synthesized narratively by preserving the original study population, comparator, concomitant care, follow-up, and outcome measure. No formal risk-of-bias assessment, certainty grading, PRISMA-style study-flow analysis, or pooled reanalysis was undertaken; the aim was a clinically bounded narrative synthesis rather than a systematic review.

The Outcome Problem: Symptoms, Ideation, Attempts, Mortality, and Recovery

The word suicidality is often used as though it referred to one continuous variable.

In practice, it can denote thoughts, intent, planning, preparatory acts, attempts, nonsuicidal self-injury, or death by suicide. These phenomena overlap, but they are not interchangeable endpoints. A trial can demonstrate a change in one while remaining uninformative about another.

Depressive-symptom scales are especially easy to overread. The Montgomery-Asberg Depression Rating Scale (MADRS), for example, contains a suicide item, but its total score is dominated by a broader set of depressive symptoms. A statistically significant total-score difference may be clinically relevant while providing no independent evidence that suicidal ideation, behavior, or mortality changed. Conversely, a suicide-specific measure can improve even when the full depressive syndrome remains severe. The correct interpretation depends on the instrument, prespecified endpoint, time point, and between-group contrast, not on the emotional weight of the clinical setting.

Suicidal-ideation outcomes usually capture the intensity, frequency, or persistence of thoughts at the time of assessment. Ideation may fluctuate within hours, may be affected by sedation or relief from acute distress, and may be disclosed differently as trust, perceived consequences, or cognitive state changes. A lower ideation score is therefore clinically meaningful but not self-authenticating evidence that the environment is safe or that an attempt has been prevented.

Suicide attempts and deaths are more difficult to study than suicidal symptoms because they are relatively infrequent, clinically heterogeneous, and influenced by multiple post-treatment factors. Consequently, most ketamine and esketamine trials lack sufficient size and follow-up to estimate these outcomes reliably, meaning that low event rates or the absence of observed deaths should not be interpreted as evidence of reduced suicidal behavior or mortality.³ Clinical improvement should also be distinguished from functional recovery and disposition. Reduced mental pain does not necessarily restore decision-making capacity or the ability to maintain safety, whereas modest symptom improvement may meaningfully enhance engagement, cognitive flexibility, and participation in safety planning. These outcomes should therefore be assessed and documented directly, rather than inferred from changes in depression severity (Table 1).

Table 1.

Outcome Domains that Should Not be Conflated

Outcome domain	Typical evidence source	What a positive result supports	What it does not establish
Depressive-symptom burden	MADRS, HDRS, self-report depression scales	Average change in the depressive syndrome at the measured time point	A specific effect on suicidal ideation, attempts, death, function, or disposition
Suicidal ideation	SSI, BSI, CGI-SS-r, C-SSRS ideation items, suicide items within broader scales	Change in reported or clinician-rated thoughts, intent, or severity as operationalized	Prevention of an attempt or death; durable safety; resolution of access to means or environmental risk
Preparatory or suicidal behavior	Structured behavior scales, clinical records, collateral information	Observed change in defined behaviors during follow-up	Mortality benefit or generalization to behaviors not captured by the measure
Suicide attempts	Adjudicated events, records, prospective follow-up	Difference in attempt incidence if the study is adequately powered and follow-up is complete	Reduction in death by suicide; mechanism; safety in an individual patient
Death by suicide	Mortality linkage or very large prospective studies	A mortality association or treatment effect when design and power are adequate	Equivalent effects on ideation, function, or quality of life
Function, capacity, and disposition	Clinical assessment, functional measures, engagement, care transitions	Whether the person can collaborate, make decisions, and enter a feasible next step	That medication caused the change or that a lower level of care is automatically safe

Notes: Measures are listed as examples, not as interchangeable instruments.

Abbreviations: BSI, Beck Scale for Suicide Ideation; C-SSRS, Columbia-Suicide Severity Rating Scale; CGI-SS-r, Clinical Global Impression-Severity of Suicidality-revised; HDRS, Hamilton Depression Rating Scale; MADRS, Montgomery-Asberg Depression Rating Scale; SSI, Scale for Suicidal Ideation.

Intravenous Racemic Ketamine and Suicidal Ideation

Evidence for intravenous racemic ketamine in acute suicidality is derived from a limited number of heterogeneous randomized trials rather than a single definitive emergency study. Across these studies, the most common regimen is 0.5 mg/kg administered over approximately 40 minutes as an adjunct to standard clinical care.

Murrough et al⁴ randomized 24 adults with mood or anxiety-spectrum disorders and clinically significant suicidal ideation to ketamine or midazolam. The primary Beck Scale for Suicide Ideation (BSSI) outcome did not differ significantly at 24 hours, although a difference emerged at 48 hours, and a suicide item embedded in the MADRS favored ketamine at 24 hours. Grunebaum et al⁵ provided a stronger suicide-specific signal. Eighty adults with major depression and clinically significant suicidal ideation received a single adjunctive ketamine or midazolam infusion. At 24 hours, the reduction in BSSI score was 4.96 points greater with ketamine (95% CI [2.33, 7.59]). A reduction of at least 50% occurred in 55% of the ketamine group and 30% of the midazolam group (OR = 2.85, 95% CI [1.14, 7.15]; NNT = 4). Abbar et al⁶ enrolled 156 hospitalized adults with severe suicidal ideation and randomized them to two ketamine infusions 24 hours apart or placebo. Full remission of suicidal ideation at day 3 occurred in 63.0% of participants receiving ketamine and 31.6% receiving placebo (OR = 3.7, 95% CI [1.9, 7.3]).

Two additional randomized studies illustrate both the promise and uncertainty of the acute signal. Domany et al⁷ randomized 18 depressed adults presenting to an emergency department with acute suicidal ideation requiring hospitalization to intravenous ketamine 0.2 mg/kg or saline. At 90 minutes, remission of suicidal ideation was observed in 88% of the ketamine group and 33% of the placebo group. By contrast, Barzkar et al⁸ randomized 36 inpatients with severe acute suicidal ideation to ketamine 0.5 mg/kg or midazolam 0.02 mg/kg. Suicidal ideation decreased in both groups, but between-group differences on the BSSI and S-VAS were not statistically significant at 12 or 24 hours. These small trials reinforce the importance of population, comparator, baseline imbalance, measure, and time point when interpreting an average treatment effect.

Comparator choice and functional unblinding are important methodological limitations in ketamine trials. Ketamine’s acute dissociative and perceptual effects can reveal treatment allocation when saline is used, whereas midazolam may better mask some nonspecific acute effects but is itself pharmacologically active. An integrative analysis found larger ketamine-control effect sizes in saline-controlled than in midazolam-controlled studies, largely because participants receiving midazolam improved more than those receiving saline.⁹ Accordingly, neither saline nor midazolam constitutes a perfect comparator, and expectancy, sedation, anxiolysis, and unblinding may influence estimated treatment effects.

A negative result in a different clinical state is equally informative. Ionescu et al¹⁰ randomized 26 outpatients with treatment-resistant depression and chronic suicidal ideation to six ketamine or saline infusions over 3 weeks. Ketamine did not show significant superiority for depression or suicidal ideation. Chronic suicidal ideation in a medically stable outpatient population is clinically distinct from an acute suicidal crisis requiring immediate assessment and containment.

An individual-participant data meta-analysis of 10 comparison studies included 167 participants with baseline suicidal ideation and found that a single ketamine dose reduced suicidal thoughts rapidly, with evidence of benefit extending across several days.¹¹ Later systematic reviews have generally confirmed a short-term group-level signal while emphasizing heterogeneity in diagnosis, baseline severity, comparator, suicide measure, and concomitant treatment.^3,12

The 2026 meta-analysis by Shim et al¹³ synthesized 26 randomized trials involving 1166 participants with a major depressive episode. Single and repeated infusions were associated with rapid average reductions in depressive and suicidal symptoms.

The broader evidence base supports a reproducible average short-term symptom signal across heterogeneous trials; it does not establish comparable effects in every setting or patient group. Moreover, limited long-term follow-up means that the clinical significance of any average treatment effect remains dependent on the individual’s residual suicide risk, functioning, environment, and access to lethal means (Table 2).

Table 2.

Pivotal Randomized Evidence for Intravenous Racemic Ketamine and Suicidal Ideation

Study	Population and co-care	Comparison	Suicide-specific result	Bounded interpretation
Murrough et al (2015)4	24 adults with mood/anxiety-spectrum disorders and clinically significant ideation	Single IV ketamine vs midazolam	Primary BSI difference not significant at 24 h; other measures/time points favored ketamine	Early mixed evidence; outcome and time-point sensitivity are material
Grunebaum et al (2018)5	80 adults with major depression and clinically significant ideation; adjunctive treatment	Single IV ketamine vs midazolam	SSI reduction 4.96 points greater at 24 h; response 55% vs 30% (NNT = 4)	Short-term reduction of thoughts in selected adults; no attempt or mortality inference
Abbar et al (2022)6	156 hospitalized adults with severe ideation across diagnostic groups	Two IV ketamine infusions 24 h apart vs placebo	Day-3 full remission 63.0% vs 31.6% (OR = 3.7)	Supports an acute inpatient ideation signal; subgroup and durability claims require caution
Domany et al (2020)7	18 depressed adults presenting to an emergency department with acute suicidal ideation requiring hospitalization	Single IV ketamine 0.2 mg/kg vs saline	Remission of suicidal ideation at 90 min: 88% vs 33%	Supports a rapid ED signal, but the very small proof-of-concept sample limits precision and generalizability
Barzkar et al (2025)8	36 inpatients with severe acute suicidal ideation	Single IV ketamine 0.5 mg/kg vs midazolam 0.02 mg/kg	BSSI and S-VAS decreased in both groups; no significant between-group difference at 12 or 24 h	Highlights comparator sensitivity and uncertainty in small heterogeneous acute samples
Ionescu et al (2019)10	26 outpatients with treatment-resistant depression and chronic current ideation	Six IV ketamine infusions over 3 weeks vs saline	No significant superiority for depression or ideation	Shows that chronic outpatient ideation should not be collapsed into the acute-crisis evidence base

Notes: Trial regimens are evidence summaries, not universal prescribing recommendations. Reference numbers correspond to the manuscript reference list.

Abbreviations: BSI, Beck Scale for Suicide Ideation; BSSI, Beck Scale for Suicide Ideation; IV, intravenous; NNT, number needed to treat; OR, odds ratio; SSI, Scale for Suicidal Ideation; S-VAS, Suicide-Visual Analog Scale.

The durability of treatment effects should be interpreted in layers: rapid reductions in suicidal ideation may occur within hours to days, sustained improvement may reflect both pharmacological and non-pharmacological influences, and effects on subsequent suicide attempts or deaths require different evidence altogether. Consequently, persistence of symptom improvement should not be interpreted as persistence of protection against suicidal behavior.

For acute psychiatric use, intravenous racemic ketamine should be understood as a supervised, off-label intervention with a narrowly defined target. The service should specify who confirms diagnosis and medical eligibility, who administers and monitors the infusion, how dissociation, sedation, blood-pressure elevation, respiratory compromise, agitation, and emerging mania or psychosis are managed, and when the suicide-risk formulation is repeated. A protocol that ends at the infusion chair is not an emergency suicide-care protocol.^14,15

The evidence also does not justify transferring an intravenous effect to oral, subcutaneous, or compounded intranasal racemic ketamine. Those routes differ in formulation, exposure, supervision, and evidence. In an acute suicidal presentation, convenience is not a sufficient reason to replace a studied, monitored intervention with an unvalidated take-home product.

Regulated Intranasal Esketamine and the ASPIRE Program

Regulated intranasal esketamine is a distinct therapeutic intervention rather than simply a nasal formulation of ketamine. It is a device-based product with defined dosing, supervised administration, monitoring requirements, and jurisdiction-specific indications. In the United States, the acute indication concerns depressive symptoms in adults with major depressive disorder and acute suicidal ideation or behavior, used together with an oral antidepressant. The current label describes 84 mg twice weekly for 4 weeks, with reduction to 56 mg for tolerability, and states that use beyond 4 weeks for this indication has not been systematically evaluated.¹⁶

The US label also preserves two decisive limitations: effectiveness in preventing suicide or in reducing suicidal ideation or behavior has not been demonstrated, and treatment does not preclude hospitalization when clinically warranted. These statements are integral to the interpretation of the evidence, defining the appropriate boundaries of the acute treatment claim. The product must be administered in a certified healthcare setting under the applicable risk-management program, with monitoring for sedation, dissociation, respiratory depression, blood-pressure effects, and clinical stability before the person leaves the healthcare setting.

European wording differs. The EMA product information describes esketamine, coadministered with oral antidepressant therapy, as acute short-term treatment for rapid reduction of depressive symptoms that constitute a psychiatric emergency in adults with a moderate-to-severe major depressive episode. The acute indication has not been studied in adults aged 65 years or older.¹⁷ Neither formulation should be considered applicable to all psychiatric emergencies, nor should evidence for either be generalized to compounded intranasal racemic ketamine.

ASPIRE I and ASPIRE II enrolled adults with major depressive disorder, active suicidal ideation with intent, substantial depressive symptoms, and a clinical need for acute psychiatric hospitalization. Participants received esketamine 84 mg or placebo nasal spray twice weekly for 4 weeks. Crucially, both groups also received comprehensive standard care, including hospitalization and initiation or optimization of an oral antidepressant. The trial comparison was therefore esketamine plus intensive care vs placebo nasal spray plus the same intensive care - not medication vs waiting.

In ASPIRE I, 226 participants were randomized. The least-squares mean difference in MADRS change at 24 hours favored esketamine by 3.8 points (95% CI [−6.56, −1.09], p = 0.006).¹ In ASPIRE II, 230 participants were randomized and 227 analyzed; the corresponding difference was 3.9 points (95% CI [−6.60, −1.11], p = 0.006).² Both studies therefore demonstrated faster average improvement in depressive symptoms.

The suicide-specific result was different. Clinician-rated severity of suicidality, measured with the Clinical Global Impression-Severity of Suicidality-revised (CGI-SS-r), improved substantially in both groups, and the between-group difference at 24 hours was not statistically significant in either pivotal trial. Pooled analyses reached a similar conclusion: esketamine accelerated improvement in depressive symptoms, whereas suicide-specific outcomes appeared to reflect the overall effects of comprehensive acute care rather than a consistent advantage over placebo.¹⁸

Additional ASPIRE analyses add clinical consistency without changing the endpoint grading. Esketamine was associated with a shorter time to depressive remission in pooled analyses, and patient-reported depressive burden and quality of life improved.^19,20 These findings may matter to patients and services. They still do not demonstrate fewer attempts, lower mortality, or a medication-driven reduction in the need for hospitalization.

This endpoint distinction is also consistent with the 2025 systematic review and meta-analysis by Fountoulakis et al²¹ which found antidepressant efficacy for intranasal esketamine but did not demonstrate a significant pooled effect on suicidality.

The appropriate interpretation is therefore not that esketamine is an established anti-suicidal treatment, but that it may accelerate improvement in depressive symptoms within the specific acute-care context in which it was studied.

The most defensible conclusion is that regulated intranasal esketamine can accelerate improvement in depressive symptoms among carefully selected, hospitalized patients with major depressive episode (MDD) and active suicidal ideation with intent who are receiving comprehensive standard care. This distinction is important because both treatment groups improved rapidly, while suicide-specific outcomes did not consistently favor esketamine over placebo. The ASPIRE findings also have limited external validity beyond this selected population and treatment pathway: they do not establish efficacy across the broader heterogeneous population of psychiatric emergency patients, including acute suicidal presentations outside major depressive disorder. Although earlier relief of depressive symptoms may facilitate communication, engagement, and participation in treatment, decisions regarding safety, hospitalization, and subsequent care must remain grounded in comprehensive clinical assessment rather than symptom rating scales or medication response alone (Table 3).

Table 3.

What the ASPIRE Program Supports - and What it Does Not

Evidence source	Studied context	Observed finding	Clinically defensible conclusion
ASPIRE I (Fu et al, 2020)1	226 hospitalized adults with MDD, active ideation with intent; esketamine or placebo plus comprehensive standard care	24-h MADRS difference −3.8 points; CGI-SS-r did not significantly separate	Faster reduction of depressive symptoms in the studied high-risk population
ASPIRE II (Ionescu et al, 2021)2	230 randomized (227 analyzed); parallel design and co-care	24-h MADRS difference −3.9 points; CGI-SS-r did not significantly separate	Replication of the rapid depressive-symptom effect, not a suicide-prevention finding
Pooled and secondary analyses (Canuso et al, 2021; Fu et al, 2023; Jamieson et al, 2023)18–20	Pooled Phase 3 data, remission timing, and patient-reported outcomes	Earlier depressive improvement and favorable patient-reported change in some domains	Useful for depressive burden and experience; does not establish fewer attempts or deaths
Fountoulakis et al (2025)21	Systematic review and meta-analysis of intranasal esketamine studies in adult depression, including suicidality outcomes	Antidepressant efficacy was observed; pooled suicidality outcomes did not show a significant treatment effect	Supports separating antidepressant efficacy from demonstrated anti-suicidal efficacy; esketamine should not be characterized as an established anti-suicidal treatment
Current product information (FDA; EMA)16,17	Supervised, regulated product within jurisdiction-specific indications	Explicit monitoring and label limitations	Hospitalization, risk formulation, and suicide-specific care remain necessary when indicated

Notes: Negative numbers indicate greater MADRS reduction with esketamine. Reference numbers correspond to the manuscript reference list.

Abbreviations: CGI-SS-r, Clinical Global Impression-Severity of Suicidality-revised; MADRS, Montgomery-Asberg Depression Rating Scale; MDD, major depressive disorder.

Ketamine or Esketamine within the Full Emergency Suicide-Care Pathway

The first priorities are environmental protection and treatment of time-sensitive medical problems. Recent overdose, injury, intoxication, withdrawal, delirium, hypoxia, metabolic disturbance, severe agitation, or reduced consciousness can alter both risk and the reliability of the interview. These conditions should be addressed before a ketamine-class antidepressant intervention is interpreted. A person who cannot yet participate in assessment needs observation and treatment, not a premature conclusion that suicidal intent has resolved.

The assessment should identify the syndrome driving the crisis and the factors likely to change near-term risk. At minimum, it should cover current and recent ideation, intent, plan, access to means, preparatory acts, prior attempts, ambivalence, agitation, insomnia, hopelessness, psychosis, mania or mixed activation, substance use, pain, recent losses or violence, capacity, supports, and the safety of the proposed setting. Rating scales can support structured assessment but cannot replace comprehensive clinical formulation or collateral information when available.^22,23

The need for observation or hospitalization should be considered before the rapid intervention is chosen. Hospitalization may be voluntary or involuntary under local law; the review cannot define universal legal criteria. Clinically, continued containment is more likely to be necessary when there is sustained intent, recent or escalating behavior, psychosis or mania, intoxication or withdrawal, inability to collaborate, an unsafe or violent environment, inability to restrict lethal means, impaired capacity, serious medical risk, or no viable follow-up.

Only after the target and level of care are clear should route-specific treatment be considered. Intravenous racemic ketamine may be considered off label within a protocol that matches the evidence and monitoring capacity. Regulated esketamine should follow current product information and the jurisdiction-specific acute indication. In both cases, the target should be written in advance: for example, a short-term reduction in severe depressive burden, a reduction in suicidal ideation within a monitored inpatient pathway, or improved capacity to engage in definitive treatment.

The meaningful post-treatment assessment begins after dissociation, sedation, perceptual change, and impaired coordination have resolved sufficiently for a valid interview. Depression, suicidal ideation, intent, behavior, activation, psychosis, cognition, function, and adverse effects should be assessed separately. A person may feel calmer because distress has fallen, because they are sedated, or because they are relieved to be in a protected setting. The clinical task is to distinguish these possibilities without treating any one of them as proof of future safety.

Any therapeutic window should be used immediately. The person and team can identify warning signs, rehearse coping steps, involve trusted supports with consent, reduce access to lethal means, clarify medication and transport arrangements, and confirm the next contact. These interventions are not reserved for discharge; they can improve inpatient and observation care as well (Figure 1).

Figure 1.

Ketamine or esketamine inside the emergency suicide-care pathway.

Note; The sequence is conceptual. Local emergency procedures, legal standards, and product requirements take precedence. Medication may be considered only after immediate stabilization and does not determine disposition.

Hospitalization and Discharge Decisions

Hospitalization should be viewed neither as definitive treatment nor as evidence of failure of rapid pharmacological intervention. It is a level-of-care intervention used when the person cannot be managed safely or effectively in a less restrictive setting. Its potential benefits include protection from immediate means, medical and diagnostic observation, sleep and substance stabilization, intensive treatment, collateral gathering, and organization of follow-up. Its burdens include loss of autonomy, disruption, stigma, cost, and, for some people, distress or trauma. The decision should therefore be individualized, proportionate, and revisited as the clinical state changes.

A rapid reduction in ideation may inform that reassessment, but it cannot settle it. Persistent intent, recent lethal behavior, psychosis, mania, severe agitation, intoxication, impaired capacity, inability to collaborate, or an unsafe environment may outweigh symptom improvement. Conversely, a person may remain depressed or acknowledge residual thoughts while demonstrating stable behavior, credible collaboration, restricted access to means, reliable support, and a feasible intensive outpatient pathway.

For regulated esketamine, the product information requires observation until the patient is clinically stable to leave the healthcare setting after the treatment session. That judgment concerns acute medication effects. It is not equivalent to discharge from the emergency department, observation unit, or psychiatric hospital. The same distinction applies after intravenous ketamine. A person can meet physiologic and cognitive criteria to leave the administration area while still requiring inpatient psychiatric care.

Before any lower-intensity disposition, the team should confirm that acute psychoactive effects have resolved; vital signs and sensorium are acceptable; no new mania, psychosis, severe agitation, or worsening intent has emerged; a collaborative safety plan is usable; lethal-means steps are concrete and feasible; transport and supervision are appropriate; medications and substance-related risks are addressed; and the next clinical contact is confirmed. The rationale should explain residual uncertainty rather than hide it behind a score threshold (Table 4).

Table 4.

Proposed Non-Validated Framework for Risk-Based Hospitalization and Discharge Considerations After a Rapid Intervention

Domain	Findings that generally support continued observation or admission	Conditions required before a lower-intensity disposition can be considered
Intent and behavior	Persistent or fluctuating intent; recent attempt or preparatory behavior; inability to explain a credible change	Repeat formulation after acute effects; behavior and intent sufficiently stable for the proposed setting; residual risk explicitly documented
Mental state and capacity	Psychosis, mania/mixed activation, delirium, severe agitation, cognitive impairment, inability to collaborate	Orientation and decision-making restored; no emergent activation or psychosis; person can understand and use the plan
Medical and substance state	Overdose/injury, intoxication, withdrawal, unstable vital signs, prolonged sedation or dissociation	Medical issues treated; acute drug effects resolved; safe transport and activity restrictions understood
Means and environment	Immediate access to highly lethal means; unsafe home; violence, exploitation, or homelessness without a viable alternative	Specific means-restriction steps completed or reliably delegated; destination assessed as workable and safe enough
Support and continuity	No reachable support when needed; refusal or inability to engage; no follow-up; major treatment barriers unresolved	Trusted support involved when appropriate; named receiving clinician/service; confirmed early contact and return pathway
Medication response	Nonresponse, partial response, or response that is inconsistent across measures and behavior	Response may inform but never determine disposition; all other domains independently support the lower level of care

Note: This is a proposed, non-validated clinical documentation framework, not a validated score, a legal standard, or a list of absolute criteria. Local law and policy govern involuntary care and discharge. It should not be used as a stand-alone suicide-risk assessment or disposition instrument.

Discharge does not mark the end of risk. Meta-analyses have found very high suicide rates after psychiatric discharge, particularly in the first week and first month, although estimates are heterogeneous and subject to publication and selection bias.^24,25 A follow-up assessment during the first week can identify risk that was not apparent or has changed since the discharge interview.²⁶ These data do not prove that every discharge is unsafe; they do show why an early, active transition is part of the disposition decision rather than an optional extra.

Discharge should not create a gap in care. Follow-up should be timely and clearly coordinated, with explicit responsibility for ongoing treatment, communication, and practical arrangements, while ensuring that the receiving team is fully informed about the recent suicidal presentation, interventions, clinical response, current risk formulation, safety plan, and outstanding concerns.

Safety Planning and Lethal-Means Restriction

A collaborative safety plan translates an acute suicidal crisis into a structured, personalized set of coping strategies and support actions that can be used if suicidal thoughts recur or intensify. It typically identifies personal warning signs, internal coping strategies, people and places that can provide distraction, trusted contacts who can help, professional and crisis resources, and steps to make the environment safer. The plan should use the person’s own language, be accessible outside the clinic, and be rehearsed rather than merely signed.

In a large cohort comparison, the Safety Planning Intervention combined with structured follow-up was associated with fewer suicidal behaviors and better treatment engagement after emergency-department care.²⁷ ED-SAFE similarly found that an emergency-department intervention incorporating brief in-visit components and follow-up contacts reduced subsequent suicidal behavior over 52 weeks.²⁸ ED-SAFE 2 extended the question to implementation, showing that a department-level process-improvement package could increase reliable delivery of suicide-prevention care and improve outcomes.²⁹

Suicidal crises are often time limited, while some methods are far more lethal and less forgiving than others. Reducing access during the high-risk period can create time and distance between impulse and action. The evidence for means restriction is strongest at the population and systems level; direct randomized evidence for an individual counseling encounter is necessarily limited. Even so, lethal-means counseling is a rational and recommended component of acute care, and emergency-department studies show that it is not delivered consistently.^30,31

Effective lethal-means counseling is individualized, focusing on the person’s access to means, the immediacy of that access, available support, and realistic temporary strategies for reducing access during periods of elevated risk. The plan should identify who will implement each step, by when, and how feasibility or completion will be confirmed; if a necessary restriction cannot be implemented, that limitation should inform the level-of-care decision.

The conversation also needs tact. It should be framed as a temporary safety measure during a period when judgment and impulse control may be under strain, not as a moral judgment or confiscation threat. Involving family or other supports requires attention to consent, safety, coercion, and the possibility that the support person is part of the risk environment. When the proposed restriction cannot be implemented, that failure changes the level-of-care decision.

A reduction in mental pain or cognitive constriction may make safety planning more authentic. The person may be better able to imagine alternatives, contact someone they had avoided, discuss access to means, or tolerate a conversation about hospitalization. This is one of the most plausible clinical values of a rapid intervention. Documentation should record the plan itself, barriers, means addressed, support persons involved, the person’s understanding, and the follow-up arrangement.

Recovery-Enabling Downstaging as a Time-Limited Operational Framework

Wilkowska and Cubala³² proposed downstaging to describe movement from a more severe or treatment-resistant depressive state toward a less severe and potentially more treatable one. The present recovery-enabling downstaging framework is an adaptation and extension of that earlier concept to acute suicidal presentations. It integrates, rather than replaces, established dynamic suicide-risk formulation, safety planning, and disposition assessment. Its proposed novelty lies in explicitly prespecifying a bounded acute intervention window, multidomain targets, stopping rules, and a mandatory transition to definitive care; it is not a new predictive risk scale, a validated staging instrument, or a stand-alone disposition algorithm. Recovery-enabling downstaging is defined here as a prespecified, time-limited reduction in acute burden that makes a safer and more constructive next step feasible. The intervention may include intravenous racemic ketamine or regulated esketamine, but medication is not the whole model. Hospitalization or observation, diagnostic treatment, sleep restoration, substance stabilization, safety planning, means restriction, family or support involvement, and continuity work are part of the same acute window.

Downstaging begins with an explicit high-acuity state and an explicit time horizon. The entry formulation may include severe depressive burden, active suicidal intent, mental pain, agitation, cognitive constriction, inability to engage, or need for intensive containment. The target should be narrower than remission. Examples include enough relief to participate in an ECT decision, enough cognitive flexibility to complete a credible safety plan, or enough behavioral regulation to move from continuous observation to a structured inpatient program.

The intervention window should be fixed in advance by the evidence and service context. For a single intravenous infusion, the key decision points may be several hours, 24 hours, and the next transition review. For the acute esketamine indication, the product-defined 4-week course provides an outer boundary, with earlier review whenever risk, response, or tolerability changes. A bounded model prevents an acute bridge from quietly becoming indefinite maintenance without a separate evidence review and consent process.

Downstaging should be defined by meeting across multiple clinical domains rather than by improvement in depressive symptoms alone. Assessment should demonstrate improvement across the clinical domains that contributed to the initial high-acuity presentation, including suicidal thoughts and behaviors, cognitive flexibility and decision-making capacity, arousal and impulse control, engagement with treatment, functional status, environmental safety, access to lethal means, reliability of social support, and the feasibility of the next stage of care. Although improvement in every domain is not required, the overall pattern should support the proposed clinical transition (Table 5).

Table 5.

Proposed Non-Validated Operational Framework for Recovery-Enabling Downstaging

Domain	High-acuity entry state	Short-term downstaging signal	Mandatory transition action
Suicidal crisis	Intent, planning, recent behavior, severe mental pain, inability to maintain safety	Lower acute intensity plus credible collaboration; no new behavior or concealed escalation	Repeat risk formulation; choose level of care; complete safety plan and means intervention
Cognition and capacity	Cognitive constriction, indecision, inability to understand options	Improved attention, flexibility, and participation after acute drug effects resolve	Shared decision-making; document capacity and residual limitations
Arousal and behavior	Agitation, panic, insomnia, impulsivity, dyscontrol	Improved regulation without sedation, mania, psychosis, or intoxication-related impairment	Treat sleep/activation; retain appropriate observation and diagnostic care
Engagement and alliance	Withdrawal, refusal, hopeless disengagement, inability to identify help	Willingness and ability to collaborate with clinicians or trusted supports	Warm handoff; confirm who is responsible for the next contact
Environment and means	Immediate access to lethal means; unsafe or unstable setting	Specific restrictions implemented; a viable setting and support plan exist	Verify feasibility; escalate level of care if the plan cannot be enacted
Function and agency	Collapse of self-care, communication, or goal-directed behavior	Early return of self-care, communication, or purposeful action	Begin rehabilitation and diagnosis-specific continuation treatment
Care feasibility	Continuous containment, repeated crisis use, or an unsustainable plan	A safer, deliverable lower-intensity pathway is available	Transition at the end of the acute window; measure treatment burden and continuity

Notes: This is a proposed, non-validated conceptual and documentation framework, not a validated outcome scale, predictive risk model, or disposition instrument. It adapts the downstaging concept described by Wilkowska and Cubala.³² A downstage is judged after acute medication effects have resolved and never overrides risk-based disposition.

The assessment should occur after acute psychoactive effects have resolved. Sedation can look like calm; dissociation can temporarily distance a person from intent; and enthusiasm immediately after a striking experience can look like restored agency. A recovery-enabling state is one in which the person is sufficiently stable to understand the available options, participate in planning concrete next steps, and engage with care that is appropriate and realistically accessible.

The model fails when there is no clinically meaningful, convergent benefit; when acute effects prevent valid assessment; when mania, psychosis, severe hypertension, respiratory compromise, prolonged dissociation, cognitive deterioration, misuse, or other harm emerges; or when the proposed lower-intensity pathway remains unsafe or unavailable. In those circumstances, treatment should stop or the plan should be revised. Repeating exposure without a measurable target is not downstaging; it is therapeutic drift.

A favorable response should also define the endpoint of acute intervention. Once immediate stabilization has been achieved and ongoing care is established, continuation of ketamine or esketamine should be considered a separate longitudinal treatment decision, requiring its own clinical indication, evidence base, informed consent, monitoring, and assessment of benefits, burdens, and discontinuation criteria (Figure 2).

Figure 2.

Operational downstaging as a bounded bridge to definitive treatment.

Note: The red line marks the end of the prespecified acute intervention window. Improvement must lead to a concrete transition; it is not evidence that the long-term natural history of a mood disorder has changed. This is a conceptual, non-validated framework and should not be interpreted as a clinical decision rule, predictive risk instrument, or validated staging tool.

Current evidence does not show that a short ketamine or esketamine course changes the long-term natural history of major depression, prevents recurrence, restores a defined biological reserve, or protects against future suicide. Terms such as disease-modifying, transformative, and neoadjuvant should therefore be used cautiously and only as explicitly identified hypotheses or conceptual frameworks, rather than established properties of ketamine or esketamine treatment.

A Practice Synthesis for Acute Services

The service should be able to answer five questions before treatment begins. What exact problem is expected to change? Why does the evidence for this route and product fit this patient and setting? What outcomes will be measured separately? What level of care is required regardless of response? What definitive treatment or transition will follow? When these questions do not have concrete answers, the intervention is not ready to be offered.

Baseline documentation should include diagnosis and differential diagnosis; current and recent suicidal thoughts and behaviors; intent, plan, and means; medical and substance state; psychosis, mania, and mixed features; capacity; supports; functional baseline; prior treatment; and the patient’s own priority outcome. Consent should distinguish approved from off-label use, explain the route-specific evidence, and state plainly that symptom or ideation improvement does not prove prevention of an attempt or death.

Administration should be supervised and protocolized. Monitoring should match the product, route, medical risk, and acute effects. The environment should support calm observation without turning dissociation into a performance or therapeutic goal. Staff should record what the person experiences, how long it lasts, what treatment is required, and whether adverse effects compromise the later interview. New mania, psychosis, severe agitation, medical instability, or worsening suicidal behavior takes priority over completing the planned course.

The review should separate depression, suicidal ideation, suicidal behavior, global clinical state, function, capacity, activation, adverse effects, and treatment burden. Risk formulation should be updated at each reassessment rather than carried forward unchanged. Continuation of ketamine or esketamine should depend on clinically meaningful benefit within a clearly defined treatment pathway and reassessment interval, and treatment should be discontinued if benefit is insufficient, adverse effects outweigh gains, or the intervention no longer represents the most appropriate use of clinical resources. Clinical handover should communicate the treatment administered, clinical response, adverse effects, residual symptoms, current suicide risk formulation, safety and lethal-means plans, support arrangements, and any unresolved medical or legal issues, while explicitly identifying remaining uncertainties. Patients should also leave with a collaboratively developed plan they understand and can actively follow, rather than a summary of the care they received.

Research Priorities

Future studies should preserve the distinction between depressive symptoms, suicidal ideation, attempts, mortality, function, and disposition from protocol to publication. Suicide-specific trials need prespecified behavioral outcomes, adequate follow-up, and linkage to high-quality event or mortality data when feasible. Because single trials will rarely be powered for death by suicide, coordinated registries and pragmatic data linkage may be more informative than repeatedly underpowered efficacy samples.

Future comparative research should prioritize clinically relevant care pathways rather than simple comparisons between ketamine formulations. Questions include whether intravenous racemic ketamine and regulated esketamine differ in the speed or durability of suicide-specific outcomes when embedded in equivalent care; which patients experience improvement in ideation independent of overall depression; how response affects collaboration and function; and whether any strategy improves the quality of safety planning, means restriction, continuity, or engagement without increasing premature discharge.

The proposed downstaging framework also requires prospective validation. Feasibility studies should test whether teams can define entry states, targets, time limits, and stopping rules reliably; whether multidomain improvement predicts a safer and more effective transition; and whether the model reduces treatment drift. Validation should not assume that ketamine is the active ingredient in every successful downstage.

The framework is intended to evaluate a pathway, not to certify a drug class.

Limitations

Several limitations should be acknowledged. Trial populations are narrower than many real-world emergency cohorts, outcomes and follow-up vary across studies, suicide attempts and deaths remain underpowered, and concomitant care is often intensive.

In ketamine trials, functional unblinding is incompletely evaluated and comparator choice is consequential: saline can make treatment allocation apparent, whereas midazolam may reduce masking differences but can itself influence anxiety, sedation, and nonspecific response. The ASPIRE program enrolled a selected population with major depressive disorder, active suicidal ideation with intent, and a need for hospitalization, so its findings should not be generalized to all acute suicidal presentations. Consequently, findings from intravenous ketamine and regulated intranasal esketamine cannot be extrapolated to compounded, oral, subcutaneous, or unsupervised formulations. In addition, the proposed hospitalization, discharge, and recovery-enabling downstaging framework is principle based rather than empirically validated and should be regarded as a conceptual clinical and service-design model rather than a biological staging system, validated clinical decision tool, or evidence of disease modification.

Conclusions

Intravenous racemic ketamine and regulated intranasal esketamine have distinct evidence bases in patients presenting with acute suicidality. Evidence supports short-term reductions in suicidal ideation with intravenous racemic ketamine in selected adults, whereas intranasal esketamine has demonstrated accelerated improvement in depressive symptoms in the specific context of the ASPIRE trials without a consistent treatment-specific reduction in suicidality. Neither intervention has been shown to prevent suicide attempts or deaths. Clinical management must therefore remain centered on comprehensive suicide care, including immediate stabilization, dynamic risk assessment, appropriate hospitalization, collaborative safety planning, lethal-means restriction, and continuity of care. Within this context, the proposed recovery-enabling downstaging framework is a conceptual, non-validated model for defining the objectives and limits of rapid intervention. Its purpose is to help determine whether a bounded therapeutic window enables a safer transition to definitive ongoing care, not to imply disease modification or durable protection against future suicidal behavior.

Funding Statement

This research study is part of the project 01-10026/0010396/01/221/221/0/2026 funded by the Medical University of Gdańsk, Poland.

AI Declaration

The authors acknowledge the use of ChatGPT, powered by GPT-5.5 and GPT-5.6 (OpenAI), during the preparation and revision of this manuscript. The tools were used solely for language and stylistic editing, including improving clarity, readability, and phrasing, and for assistance in organizing responses to reviewer comments. They were not used to generate or independently interpret study data. All manuscript content was critically reviewed and verified by the authors, who confirm the originality and accuracy of the final text, including the accuracy and appropriateness of all references. The authors take full responsibility for the integrity of the entire manuscript and its final published content. The authors also confirm that they reviewed the applicable terms of use for ChatGPT and consider its use suitable for publication in this context.

Data Sharing Statement

No datasets were generated or analyzed for this narrative review.

Ethics Approval and Informed Consent

Not applicable. This narrative review did not involve human participants, human data, or animals.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

WJC: Grants: Acadia, Alkermes, Allergan, Angelini, Auspex Pharmaceuticals, Beckley Psytech, BMS, Celon, Cephalon, Compass Pathways, Cortexyme, Ferrier, Forest Laboratories, GedeonRichter, GH Research, GWPharmaceuticals, Helus Pharma, HMNC Brain Health, IntraCellular Therapies, Janssen, KCR, Lilly, Lundbeck, MindMed (Definium), Minerva, MSD, NIH, Neumora, Novartis, Orion, Otsuka, Recognify Life Sciences, Sanofi, Seaport, Servier; Honoraria: Adamed, Angelini, AstraZeneca, BMS, Celon, GH Research, GSK, Janssen, KRKA, Lekam, Lundbeck, Minerva, NeuroCog, Novartis, Orion, Pfizer, Polfa Tarchomin, Sanofi, Servier, Zentiva;Advisory boards: Angelini, Celon (ended 2021), Douglas Pharmaceuticals, GH Research, Janssen, MSD, Novartis, Polpharma, Sanofi, Tasman Therapeutics; Consulting fees from Douglas Pharmaceuticals, outside the submitted work.

JSz: Subinvestigator in clinical trials for Janssen, Novartis, GH Research, Celon, Compass Pathways, KCR, Lilly, MSD, Minerva, Neumora, Otsuka.

AW: Subinvestigator in clinical trials for Janssen, Novartis, GH Research, Celon, Compass Pathways, KCR, Lilly, MSD, Minerva, Neumora, Otsuka.

JSł: Subinvestigator in clinical trials for Janssen, Novartis, GH Research, Celon, Compass Pathways, KCR, Lilly, MSD, Minerva, Neumora, Otsuka.

The authors report no other conflicts of interest in this work.