Esketamine (Spravato): Independent Evidence on TRD, Side Effects & Risks

Table of contents

• Evidence summary
• What esketamine is
• How it works
• What it is prescribed for
• What works and what does not
• Benefits by claim
• Risks and all side effects
• All interactions
• Who should avoid esketamine
• Dosage and how to take it
• Follow the money: who makes it and who funded the evidence
• Related research
• Frequently asked questions
• Sources and funding notes

Evidence summary

Independent evidence and credibility scorecard

Independence tier: 1 = no money from the product's seller; 4 = seller-funded. Credibility: A (highest) to D.

What esketamine is

Esketamine is the S-enantiomer (the "S" mirror-image form) of racemic ketamine, a long-established anaesthetic. The FDA label describes it as "a non-competitive N-methyl D-aspartate (NMDA) receptor antagonist" (FDA label). It is sold as Spravato, a single-use nasal spray device that delivers 28 mg of esketamine in two sprays. A 56 mg dose uses 2 devices and an 84 mg dose uses 3.

• Originator and owner: Janssen Pharmaceuticals, Inc. holds US application NDA 211243, which the FDA approved on 5 March 2019 under priority review (Drugs@FDA). Janssen is part of Johnson & Johnson, headquartered in New Brunswick, New Jersey, USA (J&J annual report). The EU marketing authorisation holder is Janssen-Cilag International NV of Beerse, Belgium. EU authorisation was granted on 18 December 2019 (EMA EPAR).
• Development path: the FDA granted Breakthrough Therapy Designation in November 2013, based on an early intravenous esketamine study (FDA briefing). Later US approvals added depressive symptoms with acute suicidal ideation (supplement approved 31 July 2020) and use as monotherapy for treatment-resistant depression (supplement approved January 2025) (Drugs@FDA).
• Generic status: no US generic is approved. Drugs@FDA lists only Janssen's NDA. J&J reports that it has sued Sandoz, Hikma and Alkem, which filed generic applications. It settled with Hikma in January 2026 (J&J annual report).
• Legal status: in the US it is a Schedule III controlled substance, supplied only through a restricted REMS programme. In the EU and UK it is prescription-only, and the SmPC says the decision to prescribe "should be determined by a psychiatrist" (EU SmPC). We did not verify its UK controlled-drug schedule.
• Ketamine itself: racemic ketamine is FDA-approved for general anaesthesia (VA/DoD 2022). It "has not been approved for depression in the USA or in any other country" but is used off-label (Gastaldon et al. 2019).

How it works

Esketamine blocks the NMDA receptor, a glutamate receptor in the brain. Its major metabolite, noresketamine, acts at the same receptor with lower affinity. The FDA label is blunt: "The mechanism by which esketamine exerts its antidepressant effect is unknown" (FDA label §12.1). Scotland's SMC summary describes the working theory. By acting on NMDA receptors, esketamine "changes the production of glutamate, which supports the creation and stability of new connections in the brain" (SMC patient summary). This is a hypothesis, not an established fact.

Pharmacokinetics (FDA label):

• Nasal bioavailability is about 48%, and peak blood levels come 20 to 40 minutes after the last spray.
• The terminal half-life is 7 to 12 hours.
• It is metabolised mainly by the liver enzymes CYP2B6 and CYP3A4.
• The label reports no plasma accumulation with twice-weekly dosing.

Independent reviewers have questioned how to read the drug's fast effect. It is not clear whether the rapid change reflects improvement in the underlying depression or a temporary drug-induced mental state that lowers rating-scale scores (Gastaldon et al. 2019). The FDA raised a related issue in its review of SUSTAIN-1. Its own modelling found that higher dissociation scores were associated with a longer time to relapse (HR 0.63 per unit increase in square-root CADSS score). The reviewers said this "introduces the possibility of alternative interpretations" of the relapse result (FDA briefing).

What it is prescribed for

US (FDA label, revised 09/2026):

• Treatment-resistant depression (TRD) in adults, either "as monotherapy or in conjunction with an oral antidepressant".
• Depressive symptoms in adults with major depressive disorder with acute suicidal ideation or behaviour, together with an oral antidepressant.
• It is not approved for children or as an anaesthetic (FDA label).

EU and UK (SmPC):

• Combined with an SSRI or SNRI for adults with treatment-resistant major depressive disorder "who have not responded to at least two different treatments with antidepressants in the current moderate to severe depressive episode".
• As acute short-term treatment of a depressive episode that "according to clinical judgement constitute[s] a psychiatric emergency" (EU SmPC).

Where guidelines and health technology assessment (HTA) bodies place it:

In practice, NICE's committee heard that esketamine "has a higher treatment burden than oral antidepressants". Patients must attend twice a week at first, for about 2 hours or more each visit, and cannot drive afterwards. Clinical experts therefore expected it to be used "later in the treatment pathway" than in the trials, after 1 or 2 augmentation therapies (NICE TA854 committee discussion).

What works and what does not

Benefits by claim

1. Short-term treatment of treatment-resistant depression (add-on)

The FDA's 2019 briefing set out the Phase 3 programme in full (FDA briefing document). All results below are placebo-adjusted changes on the MADRS (0–60 scale) at day 28, with esketamine or placebo spray added to a newly started oral antidepressant:

• TRANSFORM-2 (flexible dose, adults under 65): the difference was −4.0 (SE 1.7), one-sided p = 0.010. This was the only positive short-term trial.
• TRANSFORM-1 (fixed dose, adults under 65): the 84 mg dose (−3.2; 95% CI −6.9 to +0.5) "did not separate from placebo". Because of the pre-specified testing order, the 56 mg result (−4.1) could only be called "nominally" significant. The study "was not positive based on the prespecified analysis plan".
• TRANSFORM-3 (adults 65 and over): the difference was −3.6 (95% CI −7.2 to 0.07), p = 0.029 one-sided, which is not statistically significant. The FDA also flagged "data integrity concerns".

The EU SmPC uses a stricter analysis that treats dropouts as returning to their baseline score. Under that analysis, the TRANSFORM-2 difference is −3.5 and the TRANSFORM-1 84 mg difference is −1.2 (EU SmPC). The EMA called the 3.5-point difference "clinically relevant" and concluded that the three studies taken together "convincingly showed" superiority over placebo (EMA EPAR overview).

Independent re-analyses reach more cautious conclusions:

• Gastaldon et al. pooled the three short-term trials and found a mean difference of about −4.08 MADRS points. They noted that the trial authors had set 6.5 points as the difference needed for clinical significance (Gastaldon 2019).
• The 2025 individual-patient-data re-analysis of five initiation trials (1,505 patients across 7 RCTs overall) found −2.94 (95% CI −5.39 to −0.48), with GRADE moderate certainty. The authors concluded that the benefit was "statistically significant but small" and that its "clinical relevance... is unclear, particularly given the risks of adverse events" (Naudet et al. 2025).

Placebo response matters here. In TRANSFORM-2, the placebo-plus-new-antidepressant group improved by 15.8 points, against 19.8 on esketamine (FDA label Table 10). Most of the improvement a patient experiences is therefore shared with the comparison group. Gastaldon et al. attribute that shared improvement mainly to "the high intensity of care received by patients in these trials". Blinding is also imperfect for a drug that causes dissociation in most users, which can make rating scales less reliable.

2. Speed of onset

Cochrane found that esketamine "likely results in a large increase in participants achieving remission at 24 hours compared with placebo" (OR 2.74, 95% CI 1.71 to 4.40; 5 studies, 894 people; moderate certainty). Scores fell more modestly (SMD −0.31, 95% CI −0.45 to −0.17). Response rates also rose (OR 2.11), but that estimate was low certainty (Dean et al., Cochrane 2021). The FDA noted that in TRANSFORM-2 the only formally testable key secondary endpoint, sustained response starting at day 2, was not statistically significant (8% vs 5%; one-sided p = 0.161). Across the Phase 3 trials, the only secondary endpoint consistently in the nominally significant range was the self-rated PHQ-9 (FDA briefing).

3. Relapse prevention (SUSTAIN-1)

In SUSTAIN-1, 297 people who had improved on esketamine were randomised either to continue it or to switch to placebo spray. Everyone stayed on their oral antidepressant. Among stable remitters, relapse occurred in 26.7% on esketamine vs 45.3% on placebo (HR 0.49; 95% CI 0.29 to 0.84; NNT 6). Among stable responders, it was 16/62 vs 34/59 (HR 0.30; NNT 4) (Daly et al., JAMA Psychiatry 2019). The FDA reviewers raised three problems (FDA briefing):

• One site in Poland had a "100% rate of placebo arm relapses (16 out of 16 subjects on placebo versus 2 out of 9 on esketamine)".
• Most relapses happened within 2 to 4 weeks, faster than in typical maintenance trials. The reviewers said this raised "concern that it could reflect functional unblinding".
• Counting a randomised withdrawal study as one of the two pivotal trials was a first. The Division of Psychiatry Products "has not previously considered a randomized withdrawal trial" this way, though it said doing so was "not unreasonable".

The independent re-analysis confirmed a reduced relapse risk overall (HR 0.38, 95% CI 0.26 to 0.57). However, the remission result "was not statistically significant" when all Polish centres were excluded (Naudet 2025).

4. Monotherapy

The trial behind the 2025 US monotherapy approval randomised adults 1:1:2 to 56 mg, 84 mg or placebo after an antidepressant-free lead-in. At day 28, the differences were −5.1 (95% CI −7.91 to −2.33) for 56 mg and −6.8 (95% CI −9.48 to −4.07) for 84 mg (Janik et al., JAMA Psychiatry 2025). Independent reviewers noted two concerns (Naudet 2025):

• People who improved during the lead-in were excluded from the primary analysis, which could produce an "enriched sample".
• About 75% of the esketamine group and 50% of the placebo group "correctly guessed their group allocation with certainty".

5. Versus an active comparator (ESCAPE-TRD)

This Janssen-funded, open-label, rater-blinded trial involved 676 patients. Remission at week 8 was 27.1% on esketamine vs 17.6% on extended-release quetiapine (P = 0.003). The proportion both in remission at week 8 and relapse-free at week 32 was 21.7% vs 14.1% (Reif et al., NEJM 2023). More people stopped quetiapine than esketamine: 40.3% vs 23.2% overall over 32 weeks (EU SmPC). Because the primary endpoint counts anyone who discontinued as a non-remitter, some of this difference reflects tolerability rather than antidepressant effect. The independent re-analysis notes the MADRS difference at 32 weeks was only −2.2 points (95% CI −3.6 to −0.8) (Naudet 2025).

6. Esketamine vs racemic (generic) ketamine

No large randomised head-to-head trial has been published. The two meta-analyses below compare the drugs indirectly, across different trials, doses and routes, so the comparisons are weaker than a direct trial:

• A 2021 meta-analysis of 24 trials (1,877 participants) reported that racemic ketamine showed greater response (RR 3.01 vs 1.38) and remission (RR 3.70 vs 1.47) than esketamine, with fewer dropouts. It concluded that "intravenous ketamine appears to be more efficacious than intranasal esketamine" (Bahji et al., J Affect Disord 2021).
• A 2023 meta-analysis of 49 RCTs (3,299 participants) found that during repeated dosing, effects stayed significant for racemic ketamine but not for esketamine (Esket-High −0.22, 95% CI −0.54 to 0.10). Effect sizes were "numerically greater for racemic ketamine" (Nikolin et al., eClinicalMedicine 2023).

Separately, a publicly funded trial of 403 people referred for ECT found that IV ketamine was noninferior to ECT for non-psychotic treatment-resistant depression. Response was 55.4% with ketamine vs 41.2% with ECT, and ketamine caused less short-term memory loss (Anand et al., NEJM 2023). That result applies to IV ketamine, not esketamine.

Risks and all side effects

FDA boxed warning, in summary: the boxed warning covers five risks: sedation, dissociation, respiratory depression, abuse and misuse, and suicidal thoughts and behaviours. Patients "must be monitored for at least 2 hours at each treatment session", including pulse oximetry. Because of these risks, the drug "is only available through a restricted program" called the SPRAVATO REMS (FDA label).

Deaths in the development programme. The FDA reported "six deaths in the esketamine for treatment-resistant depression development program as of January 8, 2019, all in esketamine-treated subjects". Three were suicides, occurring 4, 12 and 20 days after the last dose. The others were a motorcycle accident, a sudden death and a myocardial infarction. The reviewers judged that "it is difficult to consider these deaths as drug-related", given the small numbers and the severity of the patients' illness (FDA briefing). Independent critics argue these safety signals "were minimised" (Horowitz & Moncrieff 2021). The 2025 individual-patient-data re-analysis found no increase in serious adverse events (IRR 1.35, 95% CI 0.54 to 3.40). It did find more sedation (RR 3.70) and dissociation (RR 2.36) (Naudet 2025).

Post-marketing signals. An analysis of the FDA Adverse Event Reporting System (FAERS) covered the first year after approval (962 cases). It found signals for dissociation, sedation, suicidal ideation (ROR 24.03) and completed suicide (ROR 5.75). When esketamine was compared with venlafaxine, the signals for suicidal ideation remained, but those for suicide attempt and completed suicide did not (Gastaldon et al., Psychother Psychosom 2021). Spontaneous-report analyses can detect signals, but they cannot prove causation or measure frequency.

Withdrawal. The EU SmPC says stopping "does not require tapering off; based on data from clinical trials the risk of withdrawal symptoms is low". The FDA label says "no withdrawal symptoms" were captured for up to 4 weeks after stopping. It still advises monitoring for dependence, because withdrawal has been reported with frequent, high-dose ketamine misuse (EU SmPC; FDA label §9.3). Any oral antidepressant taken alongside it has its own stopping rules.

All interactions

Who should avoid esketamine

• Contraindicated (US): aneurysmal vascular disease (including thoracic and abdominal aorta, intracranial and peripheral arterial vessels) or arteriovenous malformation; a history of intracerebral haemorrhage; and hypersensitivity to esketamine, ketamine or the excipients (FDA label §4).
• Additional EU contraindication: a "recent (within 6 weeks) cardiovascular event, including myocardial infarction". The SmPC also says people with unstable cardiovascular or respiratory conditions should be treated only where resuscitation equipment and trained staff are available (EU SmPC §4.2–4.3).
• Substance-use history: careful consideration is needed because of abuse risk, including a history of alcohol use disorder (FDA label §9.2).
• Psychosis: assess carefully and start only "if the benefit outweighs the risk" (FDA label §5.2).
• Pregnancy: "not recommended during pregnancy". Animal data show neurotoxicity with NMDA blockers during brain development. If pregnancy occurs, esketamine should be stopped. Women of childbearing potential should consider pregnancy planning and prevention (FDA label §8.1, §8.3). The EU SmPC says it is not recommended in women of childbearing potential who are not using contraception.
• Breastfeeding: esketamine is present in human milk, and the US label says breastfeeding "is not recommended" during treatment (FDA label §8.2).
• Older adults: the elderly trial did not reach significance. Blood levels are higher, and larger blood-pressure rises were seen (17% had a systolic rise of 40 mmHg or more). The EU label starts adults 65 and over on 28 mg and notes "limited efficacy" over 75 (EU SmPC).
• Liver: people with moderate impairment may need longer monitoring. Use in severe impairment (Child-Pugh C) is "not recommended" (FDA label §8.6).
• Kidneys: no dose adjustment for mild to severe renal impairment. Patients on dialysis were not studied (EU SmPC).
• Children and adolescents: not approved; safety and effectiveness have not been established.

Dosage and how to take it

Your prescriber and treatment centre set the dose. The ranges below are the official licensed schedules, summarised for information only. The patient sprays the dose themselves under the direct observation of a healthcare professional, in a certified setting, and never takes it home.

Sources: FDA label §2; EU SmPC §4.2.

• On the day: no food for at least 2 hours and no drinks for at least 30 minutes beforehand. Blood pressure is measured before the dose and at about 40 minutes. Each 28 mg device is used with a 5-minute rest between devices.
• Observation: at least 2 hours under the US label, including pulse oximetry. Under the EU label, until the patient is "considered clinically stable". Arrange someone to take you home, and do not drive until the next day after a restful sleep.
• How long before it works: the trials measured effects from 24 hours onward. The label says benefit "should be evaluated at the end of the 4-week induction phase" to decide whether to continue.
• Missed sessions: continue the current schedule if symptoms have not worsened. If they have worsened during maintenance, the prescriber may return to the previous, more frequent schedule.
• Stopping: the EU label says tapering is not required. Decisions about stopping esketamine, and especially about stopping the accompanying oral antidepressant, belong with the prescriber.

Follow the money: who makes it and who funded the evidence

Who makes and sells it

• Developer and sole US supplier: Janssen Pharmaceuticals, Inc., part of Johnson & Johnson (New Brunswick, New Jersey, USA). The EU authorisation is held by Janssen-Cilag International NV (Beerse, Belgium) (Drugs@FDA; EU SmPC).
• Revenue: J&J reports worldwide Spravato sales as follows (J&J 2025 annual report):
  • 2023: US$689 million.
  • 2024: US$1,077 million.
  • 2025: US$1,696 million, up 57.4%. Of this, US$1,485 million came from the US and US$210 million from outside it.
  The company attributes the growth to "continued increased physician and patient demand" and says more than 200,000 patients have been treated worldwide. That patient figure is the company's own claim and has not been independently verified.
• UK list price: £163 per 28 mg device, so £326 for a 56 mg dose and £489 for an 84 mg dose, excluding VAT. NICE noted that the company had a confidential commercial arrangement (NICE TA854 §2.3). In Scotland, acceptance took "into account a confidential discount" (SMC).
• Generic competition: none approved in the US. J&J is litigating patents against generic applicants (J&J annual report). Racemic ketamine, by contrast, has long been generic, and no company holds exclusive rights to sell it for depression.

Who funded the evidence

• Every pivotal and label-supporting trial was sponsored by Janssen. This includes TRANSFORM-1, 2 and 3, SUSTAIN-1, ASPIRE I and II, the monotherapy trial and ESCAPE-TRD (ClinicalTrials.gov; NEJM 2023).
  • In SUSTAIN-1, most named authors were "employees of Janssen Research & Development LLC and hold company equity" (Daly 2019).
  • In the monotherapy trial, several J&J-employed authors hold "a patent for esketamine for treatment of depression, the rights of which were assigned to Johnson & Johnson" (Janik 2025).
• Data sharing: J&J shares trial data through the Yale Open Data Access (YODA) Project. This allowed the independent French re-analysis. The re-analysis could not resolve the FDA's concerns about the Polish site "due to data privacy restrictions" (Naudet 2025).
• Independent reviews:
  • Cochrane was funded by the UK MRC and NIHR.
  • The re-analysis was funded by the French Ministry of Health.
  • The ketamine-versus-ECT trial was funded by PCORI.
  • Gastaldon et al. declared no conflicts.
  The ketamine-comparison meta-analyses include an author who is a co-inventor on ketamine-for-depression patents, with rights assigned to the US government. That is a declared interest that could favour ketamine.

Documented integrity and process events

• Regulatory precedent. The FDA counted a randomised withdrawal study as one of the two adequate and well-controlled trials, which the Division of Psychiatry Products had not done before. The drug was approved on the strength of "only one positive short-term trial" (FDA briefing). The joint advisory committees voted 14–2, with one abstention, in favour (Psychiatric News, 2019).
• Data concerns noted by the FDA. In TRANSFORM-3, the FDA cited "discrepancies between the locked datasets and reported protocol violations". In SUSTAIN-1, it said one Polish site drove the result (FDA briefing).
• Non-submissions to HTA bodies. For the suicide-risk / psychiatric-emergency indication, Janssen did not submit evidence to NICE (TA899, terminated) or to Scotland's SMC (SMC2539, not recommended) (NICE TA899; SMC2539).
• Critical voices. Erick Turner is a former FDA medical officer who did not take part in the 2019 meeting. He published "seven concerns about efficacy and FDA approval" (Turner, Lancet Psychiatry 2019). He declares no financial interest in esketamine, ketamine or competing treatments (Cochrane 2021 declarations).

None of this proves the drug does not work. The randomised data do show a real average effect. What the record shows is that the size and durability of that effect rest mostly on sponsor-run trials, which independent regulators and reviewers have read more cautiously than the company.

Related research

• Stress, anxiety and depression prevention guide
• Stress, anxiety and depression supplements: the evidence
• St John's wort
• Saffron for depression
• Sleep prevention guide
• Sibling medicine reviews: sertraline, escitalopram, venlafaxine, bupropion, mirtazapine, vortioxetine, zuranolone

Frequently asked questions

Sources and funding notes

• SPRAVATO (esketamine) US prescribing information, DailyMed, revised 09/2026: regulator-approved label; the text is written by the manufacturer (Janssen) and approved by the FDA.
• Drugs@FDA, NDA 211243 submission history: US government database.
• FDA Briefing Document, PDAC/DSaRM Advisory Committee, 12 February 2019: written by FDA reviewers (US government; drug review is partly funded by industry user fees); copy hosted by NATAP.
• Psychiatric News Alert, 13 February 2019: news service of the American Psychiatric Association (US); used only for the vote count.
• EMA, Spravato EPAR overview: EU regulator, mostly fee-funded.
• EU Summary of Product Characteristics, Spravato: regulator-approved; the marketing authorisation holder is Janssen-Cilag International NV (Belgium).
• NICE TA854, Esketamine nasal spray for treatment-resistant depression (2022), with section 2 and section 3: UK public body; companies pay appraisal fees.
• NICE TA899, terminated appraisal (2023): UK public body.
• SMC2258, SMC patient summary and SMC2539: Scottish NHS body; SMC2258 took account of a confidential company discount.
• VA/DoD Clinical Practice Guideline for the Management of MDD (2022): US federal government.
• Naudet et al., BMC Med 2025, IPD meta-analysis: funded by the French Ministry of Health (PHRC); authors declare no manufacturer interests; France / Romania / Italy / USA.
• Dean et al., Cochrane Database Syst Rev 2021: UK MRC and NIHR grants; one author declared Janssen advisory work and runs ketamine clinics; UK.
• Gastaldon et al., Epidemiol Psychiatr Sci 2019: University of Verona (WHO Collaborating Centre); conflicts of interest "None"; Italy.
• Horowitz & Moncrieff, Br J Psychiatry 2021: UCL, UK; NIHR grant listed on Europe PMC; the authors are known critics of psychiatric drugs (see their declarations in the 2025 WHO-database letter).
• Turner, Lancet Psychiatry 2019, "seven concerns": former FDA medical officer (Oregon, USA); no financial interest declared (per Cochrane 2021). Only the title and metadata were available to us.
• Gastaldon et al., Psychother Psychosom 2021, FAERS analysis: academic (Italy / USA); funding statement not available in the abstract.
• Bahji, Vazquez & Zarate, J Affect Disord 2021: NIMH intramural funding (USA / Canada); the senior author is a co-inventor on ketamine-related patents assigned to the US government.
• Nikolin et al., eClinicalMedicine 2023: funding "None"; authors declare Janssen advisory-board work (Loo), ketamine patents (Zarate) and various industry ties (Vazquez); Australia / Canada / USA.
• Anand et al., NEJM 2023 (ELEKT-D): funded by the Patient-Centered Outcomes Research Institute; USA.
• Popova et al., Am J Psychiatry 2019 (TRANSFORM-2): Janssen-sponsored (NCT02418585).
• Daly et al., JAMA Psychiatry 2019 (SUSTAIN-1): Janssen-sponsored; most authors were Janssen employees holding equity.
• Janik et al., JAMA Psychiatry 2025 (monotherapy): Janssen-sponsored; J&J-employed authors hold an esketamine patent assigned to J&J.
• Reif et al., NEJM 2023 (ESCAPE-TRD): funded by Janssen EMEA.
• ClinicalTrials.gov records for NCT02417064, NCT02418585, NCT02422186, NCT02493868, NCT04599855, NCT03039192, NCT03097133 and NCT04338321: all list Janssen as lead sponsor.
• Johnson & Johnson 2025 Annual Report (SEC filing): company source; sales figures are audited, but its descriptions of the product are promotional.

Last reviewed: 26 September 2026. Educational information from Pure City Research, not medical advice. Medicine decisions belong with a qualified prescriber.