Has Psychiatry Tamed the “Ketamine Tiger?” Considerations on its Use for Depression and Anxiety published in Progress in Neuro-Psychopharmacology and Biological Psychiatry (Jan. 2015)


Ketamine has been available for approximately 50 years as an anesthetic agent. It is known to have potent effects on the central nervous system glutamatergic system, in particular blockade of n-methyl-d-aspartate (NMDA) receptors. Based upon pre-clinical evidence of involvement of the glutamatergic system in mood disorders, studies have been undertaken to test the antidepressant properties of ketamine. Several well-controlled studies, along with open-label case series, have established that ketamine can have rapid antidepressant effects. Additionally, data exist showing benefits of ketamine in post-traumatic stress disorder as well as obsessive compulsive disorder. However, improvements in these conditions tends to be short-lived with single infusions of ketamine. Of concern, ketamine has been associated with neurotoxicity in pre-clinical rodent models and is well-known to cause psychotomimetic effects and addiction in humans. While ketamine has proven safe for use in sub-anesthetic doses administered once or a few times, the safety profile of prolonged use has not been established. Aspects of safety, possible mechanisms of action, and future directions of ketamine research are discussed in addition to the clinical literature on its use in psychiatric conditions.

1.0. Introduction

Ketamine [2-(2 chlorophenyl) -2 — (methylamino) cyclohexanone] was introduced as a better-tolerated anesthetic alternative to phencyclidine (PCP), which was associated with a high incidence of prolonged emergence psychotic reactions (Domino, 2010). First administered to humans in the early 1960’s, ketamine was found to induce anesthesia reliably with minimal respiratory or circulatory depression, characteristics that were highly desired for certain clinical settings in anesthesiology practice. It was noted early on that unusual psychological reactions occurred with ketamine, notably, feelings of being disconnected with one’s environment, leading to its being named a “dissociative” anesthetic (Domino, 2010). The name “ketamine” is a portmanteau of “ketone” and “amine,” reflecting two of the moieties in its molecular structure.

Ketamine is a compound with a fascinating duality about it. It has been described as neuroprotective yet also neurotoxic (Olney et al, 1989; Olney et al, 1991). It has been studied as a model for induction of psychotic symptoms of schizophrenia (Olney et al, 1999) yet, as discussed in detail below, it is an antidepressant. It is implicated in addiction but yet has also been used to treat addictions. It is indispensable for anesthesia, yet some people are so traumatized by their experiences given this drug they never want to take it again (Johnstone, 1973). It is incumbent upon the field of psychiatry to balance risks with expected benefits as it embarks to investigate longer-term antidepressant and anti-anxiety effects of this drug.

The discovery that ketamine blocks n-methyl-d-aspartate (NMDA) receptors has fueled a new generation of research on the mechanisms of psychiatric illness, and new data collections on its use for depression and anxiety disorders are reported with dizzying frequency. Ketamine clinics are popping up in multiple American cities. Indeed, this drug is currently very much a hot topic in modern clinical and research psychiatry. However, in this author’s view, there are aspects of this flurry of popularity that require caution, both in terms of the safety of this drug as well as the confidence with which its presumed NMDA-related mechanism of action can be deduced. There are lessons to be learned by a close examination of the currently available data on the use of ketamine for depression, obsessive compulsive disorder, and post-traumatic stress disorder that cast ambiguity on the “NMDA hypothesis” of its mechanism of action. In this paper, the author reviews the data on the use of ketamine in psychiatric conditions. Also discussed are perspectives on the safety issues pertinent to ketamine use as well as considerations on mechanisms of action alternative to the inherent biologic actions of the drug. Finally, some recommendations for further research are provided. Of note, to review the use of ketamine in psychiatric populations, an internet-based literaure search was undertaken using search terms such as “ketamine” and “major depression,” “obsessive compulsive disorder,” and “post-traumatic stress disorder.” All relevant papers and the bibliographies of these were reviewed and are discussed below.

2.0. Ketamine for Depression

Berman et al (2000) first postulated, based upon data of glutamatergic dysfunction in animal models of depression and outcome with NMDA receptor blocking drugs in such models, that ketamine may have antidepressant effects in humans. Thus far, there have been seven randomized, controlled single-infusion trials of ketamine versus another treatment, details of which are presented in Table 1.0. Five of these studies involved intravenous ketamine versus saline, two in patients with bipolar depression (Diazgranados et al, 2010a; Zarate et al, 2012) and the rest in unipolar depressives (Berman et al, 2000; Zarate et al, 2006a; Sos et al, 2013). In all these studies, saline infusions were associated with essentially no antidepressant response at 24 hours post-infusion while ketamine was associated with strong responding which generally abated by a few days to a week or two following the infusions. Dissociative effects, as measured by a dissociative states scale (Bremner et al, 1998) and psychotomimetic effects as measured by the Brief Psychiatric Rating Scale (Overall and Gorham, 1988), were common with ketamine in these studies during the infusions but disappeared within 20–30 minutes post-infusion. Generally, the depressed patients in these studies were chronically ill, highly medication refractory and were not acutely ill or psychotic. Rise in psychotomimetic effects during ketamine infusions did weakly correlate with degree of reductions in depression scores at day 3 following the infusions in the Sos et al (2013) study.

In a study providing an alternative to the inconvenience of intravenous ketamine administration, Lapidus et al (2014) randomized 18 depressed patients to 50 mg intranasal ketamine versus intranasal saline, again in a crossover design, and found 24 hour response rates of 44% with ketamine and no response with saline. There was a higher rise in dissociative scores in ketamine responders versus non-responders. By day 7 following the dosing, ketamine responses had dissipated.

Responding to criticisms that ketamine versus saline studies are not truly blind given the dramatic side effects to ketamine and none with saline, Murrough et al (2013a) used the short-acting benzodiazepine midazolam as “active placebo.” Midazolam was chosen because it is available intravenously and has pharmacokinetics similar to ketamine. As can be seen in Table 1.0, response rates were greater with ketamine but also quite high in the midazolam-treated group, a surprising finding considering the highly chronic, medication-refractory nature of the patients.

In addition to these randomized, controlled comparisons, there are several case studies (reviewed in detail in aan het Rot et al, 2012)and open-label single-infusion ketamine studies documenting rapid antidepressant response, generally with return to baseline severity within days to two weeks or so of the infusion (Phelps et al, 2009; Diazgranados et al, 2010b; Mathew et al, 2010; Salvadore et al, 2010; Valentine et al, 2011; Ibrahim et al, 2011, 2012; Thakurta et al, 2012; Duncan et al, 2013). Additionally, Chilukuri et al (2014) randomized depressed patients to ketamine 0.5 mg/kg intravenously over 40 minutes versus 0.5 mg/kg or 0.25 mg/kg intramuscularly and found similar reductions in depression ratings at 2 hours and 3 days post-dosing. However, there was no non-ketamine control group in that study.

As this single-infusion literature suggests impressive acute antidepressant responses to ketamine, a logical next question is whether multiple infusions in a series could result in greater response rates. Thus, this would be analogous to a course of electroconvulsive therapy (ECT) in which case a patient receives a series of ECT treatments, typically twice or thrice weekly, until maximal symptomatic improvement occurs. There are currently three data sets shedding light on multiple-infusion ketamine therapy for depression. Murrough et al (2013b) and aan het Rot et al (2010), reporting two phases of a single study, administered six thrice weekly ketamine infusions (ie, two weeks of infusions) to depressed patients, each infusion being 0.5 mg/kg over 40 minutes. Analysis of the graphically-displayed depressive severity scores in each cohort reveals that the most dramatic reduction in scores occurred after the first infusion, and scores stayed relatively constant after that point. Thus, there did not seem to be much further improvement beyond the first infusion.

In another study of serial-infusion ketamine, Rasmussen et al (2013) administered up to 4 twice-weekly infusions of ketamine, 0.5 mg/kg over 100 minutes, to depressed patients. In this study, patients were treated until either pre-defined remission occurred or four infusions without a remission. Five of the 10 patients met criteria for remission: one after one infusion, three after two infusions, and one after four infusions. Thus, there was a signal that serial infusions may enhance efficacy rates. Lara et al (2013) used very low dose ketamine sublingually (10 mg doses) serially at intervals of every 2 to 7 days and found that 20 of 26 patients seemed to achieve remission or response. However, there was no systematic assessment of depression severity and no standardized depression rating scale scores were used, thus rendering this an impressionistic data collection. It is noteworthy, though, that such low doses seemed effective in a large proportion of their depressed patients.

Ghasemi et al (2013) randomized 18 hospitalized depressed patients deemed in need of ECT to either ECT as usual for 3 treatments or to 3 serial infusions of ketamine 0.5 mg/kg over 40 minutes over the same time period (ie, one week). Depression ratings (the raters were blind, but obviously not the patients) revealed slightly lower scores in the ketamine-treated group during the week of treatment and one week later. Inspection of the graphically presented serial depressive severity scores reveals that virtually all of the reduction with ketamine was associated with the first infusion — scores remained the same thereafter. Thus far, then, two of three ketamine studies have provided essentially no signal of enhanced acute response rates with serial infusions whereas one study has.

A further very important clinical issue regarding ketamine for depression is post-response (or remission) relapse rates. If ketamine cannot induce sustainable improvement, then it is not worth much as a clinical treatment. As pointed out earlier in the single-infusion studies, return of depression scores to baseline levels within one to two weeks post-infusion was the norm. Eight out of 13 initial ketamine responders relapsed over one month in one study (Mathew et al, 2010) while 27% of ketamine responders did not relapse over one month in another trial (Ibrahim et al, 2012). In the Rasmussen et al (2013) serial infusion study, two of the five remitters sustained the improvement over a month of follow up. In the Murrough et al (2013b) serial infusion study, the chances of remaining relapse-free for 83 days of post-treatment follow up was 25%. Thus, with these two small serial infusion studies, there does seem to be a signal that this more aggressive treatment approach may be associated with more sustained improvement. To date, there are no “maintenance ketamine” studies testing the hypothesis that serial infusions at spaced intervals, say weekly (analogous to maintenance ECT), can prevent relapse. In the Lara et al (2013) report with very low dose sublingual ketamine, it is alluded that some patients received ketamine in an ongoing manner, however, no outcome data are presented relevant to the efficacy of this strategy.

Several data sets have explicitly pointed out acute reductions in suicidality in depressed patients treated with ketamine (Price et al, 2014; Larkin and Beautrais, 2011; Zarate et al, 2012; Diazgranados et al, 2010b; Rasmussen et al, 2013). However, it is not clear whether ketamine has a specific anit-suicidal effect or if suicidal symptom reduction occurs only in tandem with other depressive symptoms.

Ketamine is generally available as a racemic mixture of s- and r-ketamine. There are some studies of differential effects of the two stereoisomers in pre-clinical models. Zeilhofer et al (1992) showed that s-ketamine was approximately twice as potent as r-ketamine in blockade of NMDA receptors. Another group has shown that the two enantiomers have differential effects on dopamine and serotonin efflux in the rat nucleus accumbens and basal ganglia which might indicate greater psychotomimetic effects of the s- enantiomer (Hancock and Stamford, 1999; Tso et al, 2004). There are some human data available. Paul et al (2009) found in two depressed patients that infusions of s-ketamine did not cause the dissociative side effects of racemic ketamine. Hashimoto (2014) has argued that r-ketamine may be better tolerated. Vollenweider et al (1997), in a study in healthy volunteers given infusions of either stereoisomer, found that s-ketamine did cause more psychotomimetic effects while r-ketamine induced more relaxation, thus arguing in favor of r-ketamine as a possibly better tolerated drug. Clearly, further research is needed regarding differential therapeutic and side effects of racemic versus stereoisomeric forms of ketamine.

6.0. Future Directions

There are several avenues and recommendations for further study of clinical applications of ketamine in psychiatry. First, there probably should be no more ketamine-versus-saline studies. It is too easy for patients to distinguish between inert saline and the classic effects of even sub-anesthetic ketamine infusions; thus, any so-called double-blind comparison between ketamine and saline is in reality a two-group open-label study. Regarding the subject of using an “active placebo” as comparator, the efficacy differences between ketamine and midazolam have not been as dramatic as those between ketamine and saline; this either reflects that ketamine has a combination of expectation-related as well as inherent neurobiologically-based mechanisms or perhaps all expectation-related efficacy, as it is quite unlikely that midazolam’s efficacy is anything but expectation-related. Midazolam, at the time of its infusion, is likely to be associated with strong anti-anxiety effects that “set the patient up” cognitively to expect longer-lasting improvement, thus accounting for its remarkable apparent efficacy in the two studies of depression and PTSD (Murrough et al, 2013a; Feder et al, 2014). Perhaps a better placebo comparator than midazolam is dexmedetomidine, a selective alpha-2 adrenergic agonist used as an adjunct in anesthetic procedures and for sedation in mechanically ventilated intensive care patients (Anger, 2013). It is not a “psychotropic” drug, so no immediate anxiolytic or euphoriant properties would likely occur that might beset a series of expectations of ultimate efficacy as might occur with midazolam. Further, the pharmacokinetics and pharmacodynamics of dexmedetomidine probably more closely mimic those of ketamine than does midazolam.

Further studies should attempt to enhance the acute remission/response rates with ketamine. Manipulation of such variables as total ketamine dose, infusion rate, and frequency of treatments might shed more light on whether acute efficacy may be enhanced. The use of intranasal or intramuscular preparations obviously would be more convenient, and further study of those is indicated. The differential efficacy/tolerability of s- versus r- versus racemic ketamine should be further studied. Prevention of post-ketamine relapse rates is probably the predominant issue in the clinical use of ketamine in psychiatry at this time. It has become clear from the acute phase study follow ups that relapse rates are high and occur quickly. The obvious next step is the use of maintenance ketamine infusions, perhaps weekly or thereabouts much like maintenance ECT treatments. Investigators of this technique will need to be vigilant for signs of delayed onset, persisting psychosis or precipiation of ketamine cravings and addiction.

It is probably also worth recommending that in future publications of ketamine in psychiatry, ketamine should not be referred to simply as an “NMDA blocker,” as has been the case in some of the currently existing publications (Diazgranados et al, 2010b; Zarate et al, 2006a). The use of this phrase presupposes that the mechanism of ketamine is known and seems to rule out other biologic and psychologic mechanisms. Indeed, there are multiple other compounds with known NMDA receptor antagonistic properties which are not effective antidepressants, such as riluzole (Mathew et al, 2010), memantine (Zarate et al, 2006b), dextromethorphan, amantadine, and ethanol. Clearly, more is at play in this field than mere “NMDA blockade.”

The study of ketamine affords an opportunity to distinguish between a “drug high” and what is in some sense a “true” antidepressant effect. Ketamine, like other abused drugs (alcohol, benzodiazepines, opiates, stimulants) can cause an immediate euphoriant effect — is this a “false drug high” or merely a short-lived “true” antidepressant effect? The current enthusiasm for ketamine must be tempered by its known addictive potential and probable tolerance effects.

Another avenue worth exploring with ketamine is to elaborate on its full psychopathologic symptom profile. In other words, which symptoms of depression or anxiety disorders does ketamine improve? So far, the types of patients enrolled in ketamine studies have been chronic, treatment refractory, non-psychotic patients. Does ketamine have efficacy in fulminantly melancholically or psychotically depressed patients? If not, then this is an opportunity to explore the pathophysiologic differences between these two types of depressions. Furthermore, it is apparent that induction of mania does not occur with ketamine (Niciu et al, 2013a) — might this drug actually be effective for mania?

In the anesthesiology field, it was noted long ago that agitated emergence reactions with ketamine could be muted with benzodiazepine pre-treatment (Domino, 2010). Interestingly, Olney et al (1991) found in their laboratory studies that NMDA-blockade-induced vacuolar lesions could be prevented by pre-treatment with GABA-ergic or anti-muscarinic drugs. One wonders whether such use clinically in depression and anxiety studies with ketamine might improve the tolerability of this drug.

In summary, ketamine has generated enormous interest both as a potential treatment that might be useful clinically in its own right and as a neuropharmacologic probe into glutamatergic mechanisms in psychiatric disorders. While short-term studies do show promise, more data are needed on long-term ketamine use before this drug can be recommended for routine use. Indeed, several prominent psychiatrists have recommended against promiscuous use of ketamine before more data are published (aan het Rot et al, 2012; Rush, 2013; Kellner, 2014; Schatzberg, 2014). The clinician would be wise to heed these recommendations. It is premature to declare that psychiatry has “tamed the ketamine tiger.”

Conflict of interest statement: the author reports no conflicts of interest in the preparation of this manuscript.

Funding: there was no funding for the preparation of this manuscript.

A single golf clap? Or a long standing ovation?

By clapping more or less, you can signal to us which stories really stand out.