Prescribing Ketamine and Esketamine

Why Ketamine Still Changes the Conversation

For seventy years, every antidepressant available worked the same way: nudge monoamines (serotonin, norepinephrine, dopamine) and wait four to six weeks to find out whether it helped. Ketamine is the first antidepressant to break that pattern, both in mechanism (glutamatergic, not monoaminergic) and, more strikingly, in tempo. A single dose can lift a severe depression within hours and can cut suicidal ideation on day one, before ECT or TMS have even begun to work.

The speed isn't marketing. In head-to-head terms, ketamine's antidepressant effect size runs 0.6 to 1.0, roughly two to three times that of a standard SSRI (~0.3). Around half of treatment-resistant patients show a meaningful response (≥50% symptom reduction) at 24 hours, versus under 10% on placebo. In patients who have, by definition, already failed the conventional playbook, that number is remarkable.

The catch, and the reason this guide is long, is that ketamine's benefit is fast but not durable. A single dose fades over days to weeks. Holding the response means repeated dosing, which raises questions the field has not fully answered: long-term safety, tolerance, abuse liability, and a post-discontinuation suicide signal that is still unsettled. Add the REMS bureaucracy that governs esketamine (Spravato) and the two-hour in-office monitoring every dose requires, and you have a treatment that asks more logistical commitment than almost anything else in outpatient psychiatry.

The terminology trips people up constantly:

• Ketamine is the racemic mixture: 50% S-ketamine (esketamine) plus 50% R-ketamine (arketamine). Used off-label, most commonly IV (0.5 mg/kg) or intranasal/compounded. Schedule III. Cheap. Not FDA-approved for depression.
• Esketamine (Spravato) is the purified S-enantiomer, delivered intranasally, FDA-approved for TRD and for depression with acute suicidality. It is governed by a REMS program, must be given in a certified center, and costs considerably more.

Counterintuitively, the FDA-approved, patented, expensive product is probably not the more effective one (more on that below). It is, however, the one insurers are most likely to cover and the one with the regulatory scaffolding built around it.

Part 1: Indications

FDA-Approved Uses (Esketamine / Spravato)

• Treatment-resistant depression (TRD) in adults, as an adjunct to an oral antidepressant. Approved March 2019.
• Depressive symptoms in adults with major depressive disorder and acute suicidal ideation or behavior (MDSI), again alongside an oral antidepressant. Approved 2020.

Both approvals are for esketamine added to an antidepressant; the label never covers it as standalone monotherapy.

The Evidence-Based Clinical Uses

Treatment-resistant depression. The evidence is strongest here, and TRD is where ketamine belongs in the algorithm. The working definition of TRD across most trials and clinical use is failure of at least two adequate antidepressant trials (adequate meaning roughly six weeks at a therapeutic dose). In the fixed-dose registration trial (TRANSFORM-1), response at four weeks was 53.1% for esketamine 84 mg plus an antidepressant versus 38.9% for placebo plus an antidepressant, with remission 38.8% versus 30.6%. IV ketamine's numbers are frequently higher (40 to 70% response by 24 hours), and indirect meta-analytic comparisons favor IV racemic ketamine over intranasal esketamine.

Acute suicidal ideation. Here ketamine does something nothing else in the pharmacopeia does: it lifts depressive symptoms within hours. In hospitalized patients with active suicidal ideation and intent, esketamine 84 mg improved depressive symptoms at 4 and 24 hours, faster than ECT or TMS can act. Suicidality itself improved just as fast on standard care alone, and the label states that effectiveness in preventing suicide or in reducing suicidal ideation has not been demonstrated. That makes it a real option for the acutely suicidal patient, potentially as an alternative to prolonged hospitalization when paired with close follow-up (intensive outpatient, partial hospital).

Bipolar depression: emerging, use with caution. Some trials show benefit, but the data are thinner and the risk of manic or mixed activation is real. If used, screen carefully for mixed and hypomanic features, monitor mood closely, and ensure adequate mood-stabilizer coverage.

Off-label and pilot-level indications. PTSD (where dissociation may cut both ways: benefit versus traumatic recall), OCD, and generalized and social anxiety all have pilot data only. Reserve these for specialists.

Who Is the Ideal Candidate?

• Unipolar treatment-resistant depression (≥2 adequate failed trials): the strongest evidence.
• Acute suicidality where rapid action is critical and close follow-up is available.
• Severe depression where ECT is unavailable or declined, or that failed to respond to TMS (roughly half of TMS non-responders respond to ketamine).
• A patient with the temperament and life logistics to tolerate a dissociative experience and navigate a high-touch, supervised treatment.

Predictors of a Better Response

The literature is noisy here, but several signals recur:

• Family history of alcohol dependence (a fairly consistent early-response predictor)
• Prior suicide attempt (paradoxically, a positive prognostic sign)
• Comorbid anxiety
• History of childhood trauma has been reported as positive in some ketamine data. It is counterintuitive and not something to lean on clinically.

Who Is a Poor Candidate?

• Active substance use disorder (beyond nicotine) or high addiction risk: iatrogenic misuse risk.
• Active psychosis or psychotic-spectrum symptoms: ketamine is dissociative and PCP-like; it may worsen psychosis, and safety data are insufficient.
• Cardiovascular fragility: recent MI, uncontrolled hypertension, aneurysm or recent cerebral hemorrhage, significant coronary disease.
• Patients who cannot tolerate dissociation or the supervised model, or who lack the follow-through the treatment demands.

Part 2: Before You Start: Workup and Candidacy

The pre-start workup for ketamine is less about lab panels (this is not lithium) and more about cardiovascular screening, psychiatric and substance-use screening, and expectation-setting.

Baseline Assessment

Routine labs (CBC, CMP, LFTs) are not required for standard psychiatric-dose courses, because hepatic and bladder toxicity come from chronic high-dose abuse, not supervised therapeutic dosing. Consider LFTs and urologic screening only if planning prolonged high-frequency maintenance.

Two Things to Do Before the First Dose

Consider tapering benzodiazepines first. GABAergic drugs (benzodiazepines, Z-drugs) may blunt ketamine's glutamatergic mechanism and reduce efficacy. Where feasible, taper the benzodiazepine before starting, ideally with the last dose about 2 nights before the first ketamine dose. One clinic found the benzodiazepine taper could run concurrently with ketamine initiation without worsening depression, anxiety, or sleep; ketamine is powerful enough to cover the taper, so this is a good deprescribing opportunity.

Set concrete behavioral goals. Ketamine works best embedded in a behavioral plan. Set two or three SMART goals ("out of bed and showered by 9 a.m. daily") so the rapid mood lift turns into changed behavior instead of evaporating.

Part 3: How to Start and Dose

There are three practical delivery routes. Match the route to coverage, access, and setting.

Oral Ketamine (Off-Label)

Bioavailability is low and erratic; dosing is unstandardized (ranges from roughly 1 mg/kg up to 50 to 300 mg). It carries the highest diversion risk of any route because it is home-dispensed. Not preferred for psychiatric use.

Route Selection at a Glance

Esketamine: the Second Indication and Special Populations

• Strengths: Spravato is supplied as 28 mg nasal-spray devices, dispensed as 56 mg (two devices) or 84 mg (three devices) dose kits; C-III, REMS-restricted.
• Second FDA indication (MDD with acute suicidal ideation or behavior): 84 mg twice weekly for 4 weeks (reduce to 56 mg if 84 mg is not tolerated), as adjunct to an oral antidepressant. This schedule is separate from, and more intensive than, the TRD induction above.
• Geriatric (≥65): the geriatric trial started esketamine at 28 mg twice weekly and titrated to 56 or 84 mg by tolerability; the label itself mandates no reduction.
• Hepatic impairment: moderate (Child-Pugh B) is usable, but expect longer monitoring; severe (Child-Pugh C) is not recommended (not studied). Renal impairment needs no dose adjustment.
• Contraindications (esketamine): aneurysmal vascular disease (thoracic/abdominal aortic, intracranial, or peripheral arterial), arteriovenous malformation, or a history of intracerebral hemorrhage (all because of the transient blood-pressure rise); and hypersensitivity to ketamine/esketamine.
• Blood pressure: esketamine causes a transient, dose-related BP increase, so measure BP before each dose and ~40 minutes after; do not dose if baseline BP is unsafe, and monitor longer if it rises.
• Pregnancy/lactation: avoid, given potential fetal harm and excretion into milk; not for use in pregnancy.
• Boxed warnings (esketamine): sedation, dissociation, respiratory depression, abuse and misuse, and suicidal thoughts and behaviors. These are the reasons for the REMS, in-office administration, and ≥2-hour post-dose observation with no driving that day.

Part 4: Monitoring

The Non-Negotiable Peri-Dose Monitoring

Every dose, every route:

• Blood pressure and heart rate: baseline before dosing, then during and after. Expect the systolic rise to peak around 40 minutes and settle within a few hours. Apply predetermined stopping criteria (Emory: SBP rise >45 mmHg, DBP rise >30 mmHg, or symptomatic severe hypertension: headache, chest pain, dyspnea, blurred vision).
• Sedation and dissociation: observe until resolved, typically 1–2 hours. For esketamine this ≥2-hour observation is REMS-mandated.
• No driving on treatment days. Arrange transport in advance.
• Suicidality: formal re-assessment after each dose and during maintenance.

Timing Landmarks

• BP peak: ~40 minutes post-dose.
• Dissociation/sedation resolution: 1–2 hours.
• Maximum antidepressant effect: ~24 hours post-dose (well after the drug itself has cleared: half-life is only 2–4 hours). Don't judge efficacy in the dosing room; judge it the next day.

Longitudinal Monitoring

Part 5: Side Effects and How to Manage Them

The common side effects are acute, transient, expected, and self-limited. They resolve within 1 to 2 hours, and you manage them with preparation and observation rather than intervention. The serious long-term concerns come almost entirely from chronic high-dose abuse, not supervised therapeutic dosing, and keeping that distinction straight prevents most of the unwarranted fear around this drug.

Dissociation / Altered Perception (30–80%; >80% Report Something Strange)

This is the hallmark effect. Patients describe feeling "spacey," "floating," "woozy," "numb," "out of body." It tracks the drug's acute window and resolves in 1 to 2 hours.

Management:

• Prepare in advance: normalize it, frame it as temporary and expected, coach a "let go and be curious" stance rather than fighting it.
• Set the environment: dim room, blindfold, instrumental music (avoid lyrics with interpretable content); minimal interaction during the acute phase.
• Do not reflexively suppress it. Dissociative intensity may correlate with antidepressant response, so blunting it, especially with benzodiazepines, may reduce efficacy.
• A small fraction (1–5%) find the experience frightening enough to drop out; supportive presence and preparation minimize this.

Hypertension (the Most Reliable Physiologic Effect)

Mean placebo-adjusted systolic rise ~7–10 mmHg and diastolic ~4–6 mmHg, peaking ~40 minutes and lasting about 4 hours. In the registration trials, systolic BP reached 180 mmHg or higher in about 3% of esketamine-treated patients and diastolic BP reached 110 mmHg or higher in about 4%.

Management:

• Baseline BP before each dose; monitor through the observation window.
• Predetermined stopping criteria (Emory thresholds, above): rarely triggered in practice.
• No routine antihypertensive is needed in most cases; the rise is self-limited.
• Extra caution in patients with cardiovascular risk; screen out the truly high-risk up front.

Nausea, Dizziness, Bitter Aftertaste (~5–10%)

These are usually mild and transient and tend to resolve as the dissociation abates.

Management: antiemetic PRN; the bitter aftertaste (intranasal) is brief and needs little beyond reassurance; often habituates over repeated doses.

Sedation (~10–20%)

Time-limited to the 1 to 2 hour window. Managed simply by the observation period and the no-driving rule; some patients prefer evening dosing.

The "Long-Term" Concerns: Keep Them in Proportion

These effects cause the most anxiety and the most confusion, because the frightening data come almost entirely from chronic high-dose recreational abuse rather than supervised therapeutic use.

• Urinary/bladder toxicity (ketamine cystitis). About 25% of heavy recreational users report urinary symptoms; the mechanism is bladder cellular death and fibrosis. In supervised psychiatric use, it is extraordinarily rare: only a handful of documented cases across decades, and animal data show inflammation only at doses far above the clinical range. Ask about dysuria/hematuria on long maintenance; otherwise the risk is negligible.
• Hepatotoxicity. Seen with massive-dose, prolonged (40–50 hour) infusions and chronic abuse, not with standard psychiatric dosing. LiverTox, the NIH database of drug-induced liver injury, reports no association between esketamine and serum enzyme elevations or clinically apparent liver injury.
• Cognitive impairment. Acute distortion lasts a few hours; there is no documented permanent impairment at clinical doses. Neuropsych testing at 6 weeks post-infusion shows no residual deficits. Cognitive problems appear only in heavy chronic abusers (grams per day, multiple days per week for years), and even those largely reverse with abstinence.
• Addiction/dependence. Case reports at therapeutic doses are rare, and most involve atypical circumstances. There is no chemical dependence and no withdrawal syndrome; the risk is psychological, and selection and supervision manage it (see Parts 2 and 4).

Part 6: Overdose, Safety, and the Discontinuation Signal

There Is No Classic Overdose Toxidrome

Ketamine has a wide therapeutic index. Anesthetic doses (1–4.5 mg/kg IV) dwarf the psychiatric dose (0.5 mg/kg) and are used routinely in operating rooms; recreational users tolerate doses many times higher. It has far more margin than lithium, and its acute danger is not "overdose" in the classic sense.

The Boxed Warnings

Racemic ketamine carries no boxed warning of its own; as a Schedule III controlled substance, it instead requires careful record-keeping and DEA compliance.

The Post-Discontinuation Suicide Signal: Take This Seriously

This is the one safety issue that calls for caution rather than reassurance. In the esketamine registration data, there were 6 deaths, including 3 completed suicides occurring 4, 12, and 20 days after the last esketamine dose (following the open-label active-treatment phase). The FDA attributed these to underlying illness severity, and they did not reach statistical significance. But two features are unsettling: the frequency looked unusual compared with ECT/TMS trials in similar populations, and some patients had no documented suicidality immediately before.

The leading hypotheses are a withdrawal-like phenomenon, possibly opioid-mediated (recall that naltrexone reduces ketamine's antidepressant effect, implicating the opioid system), or glutamatergic rebound. A trial is underway examining whether buprenorphine can prevent these post-discontinuation events and extend ketamine's benefit.

Part 7: Drug Interactions

Ketamine is metabolized largely by CYP3A4, CYP2C9, and CYP2B6. In practice, standard psychiatric medications rarely cause clinically meaningful pharmacokinetic interactions, and there are no absolute contraindications with common psychotropics, since esketamine was studied as an add-on to ongoing antidepressants. The interactions that matter are pharmacodynamic.

Benzodiazepines and Other GABA Agonists: the Efficacy Interaction

• Mechanism: GABA agonists (benzodiazepines, Z-drugs, barbiturates) may blunt ketamine's glutamatergic action.
• Consequence: reduced acute dissociative effect and possibly reduced antidepressant efficacy.
• Management: taper the benzodiazepine before (or concurrently with) ketamine initiation; last dose ideally about 2 nights before the first ketamine dose. The taper protects efficacy and is also a deprescribing opportunity.

Naltrexone: the Efficacy-Abolishing Interaction

• Mechanism: ketamine's antidepressant effect depends partly on the opioid system.
• Consequence: in a small double-blind crossover trial stopped at interim analysis, 7 of 12 patients responded under placebo and none under naltrexone, while dissociation was unaffected. A larger crossover trial in 26 patients found naltrexone attenuated rather than abolished the acute antidepressant effect.
• Management: avoid co-administering naltrexone with ketamine given for depression. If a patient needs opioid-antagonist therapy, expect a reduced antidepressant response.

Sympathomimetics and Blood-Pressure-Raising Agents

Ketamine itself raises BP via sympathomimetic-like effects. Stimulants, other sympathomimetics, and even the additive load with the patient's own baseline hypertension can compound the transient rise. Coordinate monitoring in patients on antihypertensives, and be cautious stacking BP-raising agents.

CNS Depressants

Because sedation is an expected effect, layering additional CNS depressants (opioids, sedative-hypnotics, alcohol) adds to sedation and respiratory-depression risk, especially in obstructive sleep apnea.

Agents That Failed as Maintenance

Lithium, riluzole, lamotrigine, and d-cycloserine have all been tried as continuation agents to prolong ketamine's benefit, and all have failed. (Lithium may also theoretically dampen ketamine's NMDA-related mechanism; anecdotal.) Only repeated dosing and, to some degree, structured psychotherapy sustain the response.

Part 8: Special Populations

Pregnancy

No controlled human data. Ketamine is teratogenic in high-dose animal studies, though obstetric anesthetic use suggests relative safety at clinical doses. Avoid unless depression or suicidality is severe and the risk-benefit clearly favors treatment; consider deferring to after delivery.

Lactation

Ketamine enters breast milk; infant risk is unclear and data are insufficient. Likely avoid; discuss risk-benefit explicitly if considered.

Elderly (≥65)

The flexible-dose geriatric esketamine trial fell just short of statistical significance (p = 0.06); the efficacy signal is present but weaker than in younger adults. Older patients also carry higher baseline cardiovascular risk, so monitor BP closely and consider slower titration. The label sets no mandatory geriatric dose reduction. The 28 mg twice-weekly start in Part 3 is what the geriatric trial used rather than a label requirement.

Renal Impairment

Ketamine is cleared hepatically, not renally, so no specific renal dose adjustment is defined. Likely safe; monitor for accumulation only with repeated dosing in severe renal failure.

Hepatic Impairment

Ketamine is metabolized by hepatic CYP enzymes, so significant cirrhosis likely warrants caution and possible dose reduction. No toxicity reports in the psychiatric population at standard doses.

Children and Adolescents

Off-label; not FDA-approved for pediatric depression. Data are limited to case reports. Reserve for severe, refractory cases under specialist oversight.

Part 9: Discontinuation and Maintenance

Maintenance is the hardest part of ketamine treatment. The antidepressant effect is fast but not durable, and there is no reliable way to make it last except to keep dosing.

The Relapse Problem

After a completed acute course, relapse rates run 55 to 89% within 4 weeks without maintenance. Median time-to-relapse across studies ranges from about one week to several months. A single course is an acute intervention, not a cure, and its gains have to be actively held.

There Is No Chemical Withdrawal, But There Is a Discontinuation Signal

Ketamine produces no chemical dependence and no withdrawal syndrome in the classic sense. But as covered in Part 6, there is an unresolved post-discontinuation suicide signal (3 suicides 4 to 20 days after last dose in the trials). You won't see tremor or autonomic withdrawal, but discontinuation is not pharmacologically inert, and it demands vigilance.

What Works for Maintenance

• The only proven maintenance strategy is repeated dosing at intervals matched to the drug's duration. The standard esketamine taper widens gradually: twice weekly × 4 weeks → weekly × 4 weeks → every 1 to 2 weeks thereafter, individualized to the longest interval that holds remission.
• Failed maintenance add-ons: lithium, riluzole, lamotrigine, and d-cycloserine were all tried, and all failed to extend benefit.
• Psychotherapy helps. CBT is the one psychotherapy with controlled evidence for extending ketamine's benefit; the Montreal ketamine-assisted psychotherapy model (SMART goals, preparation, integration with ACT) is a promising framework. The clinical maxim: "Ketamine opens the door; therapy walks through it; the patient does the work."

A Reasonable Discontinuation Approach

By analogy to a standard depressive episode, a defensible practice is a roughly 6-month continuation after achieving response, followed by a carefully monitored discontinuation trial to see whether improvement is sustained. If relapse occurs, reinitiate and attempt periodic discontinuation trials over time. Because the suicide signal appeared in the post-discontinuation window, taper spacing gradually, keep the standard antidepressant and therapy in place, and monitor closely through the transition.

Part 10: Ketamine vs the Alternatives

Ketamine's value is clearest when you set it beside the other tools for severe/resistant depression.

In head-to-head comparisons with ECT over about 3 weeks, both produced similar 40 to 50% improvement; ECT's benefits may last longer, and satisfaction may be higher with ketamine. Remission rates for ketamine and TMS are broadly comparable (~20 to 30% each; ECT exceeds both).

Racemic Ketamine vs Esketamine

Esketamine (the S-enantiomer) has stronger NMDA antagonism, but NMDA antagonism is likely a secondary antidepressant mechanism rather than the primary one, so the greater potency at that receptor doesn't translate into greater clinical benefit. A 24-trial meta-analysis favored IV racemic ketamine (roughly twice as likely to achieve response/remission), and animal data suggest the R-enantiomer (arketamine, in development) may be the more potent, longer-lasting, less-rewarding, less-neurotoxic component. Per dose, compounded intranasal racemic ketamine is far cheaper than esketamine, while IV racemic ketamine costs about the same. Esketamine is the FDA-approved, widely used product because racemic ketamine is generic and unpatentable (Janssen pursued esketamine specifically because the pure enantiomer could be patented) and because insurance covers esketamine, not racemic ketamine. Much of the prescribing follows coverage rather than clinical superiority.

Mechanism

Ketamine's mechanism differs from that of every monoamine antidepressant, and it explains both the speed and the maintenance problem.

• NMDA receptor antagonism is the classic mechanism (ketamine blocks the receptor within the channel), but the field increasingly views this as secondary rather than the primary driver of the antidepressant effect. (Other NMDA antagonists like memantine lack ketamine's antidepressant punch, which supports that view.)
• The likely primary mechanism is downstream: a glutamate surge driving preferential AMPA receptor activation, which triggers mTOR signaling, BDNF release, and rapid synaptogenesis, the strengthening of synaptic connections thought to underlie the fast antidepressant effect. The active metabolite hydroxynorketamine may drive much of this AMPA-mediated effect independent of NMDA blockade.
• An opioid contribution is likely: naltrexone reduces the antidepressant effect (while leaving dissociation intact), implicating the endogenous opioid system, possibly relevant to the post-discontinuation signal.
• Pharmacokinetics: elimination half-life is only 2 to 4 hours, yet the peak antidepressant effect is at ~24 hours and a single dose can last days to weeks. The therapeutic effect comes from a downstream neuroplastic cascade rather than the drug sitting on a receptor. That is why the benefit outlasts the drug, and why it eventually fades once the plasticity window closes.
• Enantiomers: racemic ketamine equals 50% S-ketamine (esketamine) plus 50% R-ketamine (arketamine). Esketamine has higher NMDA affinity; arketamine may (in animals) be the more potent and better-tolerated antidepressant. Racemic-ketamine findings cannot simply be extrapolated to esketamine alone.

The Bedside Cheat Sheet