Ketamine: Magic Antidepressant Or Illusion? Revisited

By Neuroskeptic | December 18, 2012 11:54 am

There’s a lot of interest in the idea that ketamine provides unparalleled rapid, powerful antidepressant effects, even in people who haven’t responded to conventional antidepressants.

Earlier this year, I asked:

Ketamine – Magic Antidepressant, or Expensive Illusion?
There have now been several studies finding dramatic antidepressant effects of ketamine, the “club drug” aka “horse-tranquilizer”. Great news? If you believe it. But hold your, er, horses… there’s a problem.

My concern was that although depressed patients certainly do report feeling better after an injection of ketamine, compared to people given placebo, that doesn’t prove that the drug is actually an antidepressant.

Rather, patients might be experiencing an enhanced, ‘active’ placebo effect, because ketamine causes subjectively powerful hallucinogenic experiences. So the placebo-controlled trials weren’t really blinded.

To settle the question, I suggested a three-way trial comparing it both to an inert placebo, and to some other hallucinogen; if ketamine has a specific antidepressant effect, it should produce more improvement than the comparison drug.

This has never been done.

Given this background, a new trial from NIMH‘s Ketamine King, Carlos Zarate, makes interesting reading: A Randomized Trial of a Low-Trapping Nonselective N-Methyl-D-Aspartate Channel Blocker in Major Depression.

Zarate et al tried a novel drug, AZD6765 in depressed people. AZD6765 works much like ketamine in that it blocks brain NMDA receptors. But it is a less powerful trapping blocker than ketamine, meaning that AZD6765 causes less dramatic effects on the target receptors, in some respects.

In practice, this makes AZD6765, much less hallucinogenic than ketamine.

So it’s interesting that, compared to placebo, the new drug only produced small benefits. On the MADRS depression symptom scale, patients felt a little better on AZD6765, but the boost only lasted a few hours.

The effect was far smaller than in an earlier ketamine trial as my crudely-mashed-up graph shows (although note that the patient populations were somewhat different, one bipolar and one unipolar depression, although their baseline severity was the same.)

While on ketamine people experienced significant subjective effects, with AZD6765 they didn’t, and couldn’t tell whether they got drug or placebo. Is that why they got a smaller benefit?

This is what we’d see if NMDA blockers do have a modest antidepressant effect but the dramatic improvements seen on ketamine are largely active placebo phenomena. Then again, it’s also consistent with ketamine being a powerful antidepressant and AZD6765 just being less effective because it’s a milder blocker of NMDA – effectively, a low dose of ketamine.

To tell the difference, we need… an active placebo controlled trial, like I’ve been banging on about for ages. But I wasn’t the first one to suggest it – that was none other than Carlos Zarate et al in 2006.

ResearchBlogging.orgZarate CA Jr, Mathews D, Ibrahim L, Chaves JF, Marquardt C, Ukoh I, Jolkovsky L, Brutsche NE, Smith MA, and Luckenbaugh DA (2012). A Randomized Trial of a Low-Trapping Nonselective N-Methyl-D-Aspartate Channel Blocker in Major Depression. Biological psychiatry PMID: 23206319

  • Bernard Carroll

    Given that mood episodes likely represent abnormal circuit activity, then it isn’t surprising that almost anything that acts on the neurochemical components of circuits will have some effect, at least temporarily, that looks like an antidepressant effect. So, in the prehistory of antidepressants we saw opiates and amphetamines used; before that we saw alcohol self-administered, to be replaced for a while by benzodiazepines; we know that anticholinergic drugs have some effect. These run the gamut from opioid to dopaminergic to GABAergic to cholinergic actions. Later, norepinephrine and serotonin were added in. So, the current interest in glutamatergic actions was predictable… but, placebo effect aside, that doesn’t mean that glutamate is the key to mood disorder circuit pathology.

  • Kelly Wilson

    Why would we assume that mood episodes likely represent abnormal neural circuitry? I am a scientist and I do not see compelling evidence that this is so. Nor does it need to be so in order for drugs that tinker with neurochemistry to alter mood. The two are orthogonal. At best abnormal neural circuitry is a widely believed hypothesis.

  • Curtis R.

    Two questions for you:

    The 2006 study you address shows that the antidepressant effect “continued to remain significant for 1 week”. If it were simply a placebo effect, shouldn't the symptom reduction have gone away shortly after the hallucinogenic side effects went away?

    Also, a similar RCT using memantine – another NMDAR antagonist with hallucinogenic side effects – found no antidepressant effect. If ketamine's efficacy was due to placebo effects stemming from hallucinogenic experiences, then why wasn't memantine also effective as an antidepressant?

  • Bernard Carroll

    Kelly, I suggested abnormal neural circuit activity, but not abnormal neural circuitry in the anatomical sense. Given that mood episodes are, well, episodic, my suggestion seems reasonable, but I am open to your alternative ideas.

  • DS

    Opioids are dissociative relievers of pain – I still feel the pain but don't care. Maybe ketamine is a dissociative antidepressant – I still feel the depression but don't care?

  • Anonymous

    It seems to me that something like major depression, which probably does definitely have some neural wiring issues that are causing unhealthy thought patterns, would be very hard (read impossible) to treat chemically. As some use the analogy, it's at best a splint or cast on the problem while therapy is the real treatment.

    However, these kinds of drugs, and especially MDMA, should be tried on traumatic instances of depression like post-partum, which seem very likely entirely neural transmitter related.

  • Zigs

    It is important to note that only 30% of patients responded to the new drug vs. 70% who respond to ketamine. This alone could account for the less dramatic difference in MADRS scores.

  • John

    Perhaps it is time we reconsidered what we mean by the placebo effect. Pleasure has physiological consequences. Eg. Both THC and MDMA reduce inflammation, as do opioids, and inflammation is related to depression.

  • neurosulcing

    Just out of curiosity, an active placebo would, presumably, have to have similar hallucinogenic effects as ketamine. Which drugs would be candidate drugs for this?

  • CM

    Maybe, as I tried to point out on the last post, it's an issue that has received little attention because no one sensible actually cares a lot about this. The clinicians don't care, because the patient is better. The health economists don't care, because it's fairly cheap and effective. The patients don't care, because initially they were too depressed to care and, though they now aren't, it doesn't matter to them how they got there.

    What does active placebo effect even mean in this context? That a compound affects your level of depression through some (unspecified) legitimate action, rather than through your mind? I don't know what the “legitimate” action is meant to look like, but if it doesn't involve changes in mental states, then it isn't really affecting depression. You can spend your life trying to work out what the full structural equation model for ketamine's antidepressant effect looks like, and you can probably label the nodes of that model with either psychological constructs or neural substrates and it will be equally valid. What it won't be is clinically relevant, unless it leads to better treatment (either through improved protocols or discovery of a better drug).

    Earlier, you proposed testing against a benzo. Let's face it… many severely depressed patients are already on a benzo, many more have already experienced benzos. Benzos might reduce the energy one has to obsess over suicide, but they don't have a strong antidepressant effect. If you find another substance with a similar effect to ketamine, that also reduces depression, you may not have found an active placebo effect, you will have found another potentially useful substance.

    If you want to find out if the effect is partially mediated by unblinding, then rather than finding the compound that looks and smells and feels like ketamine, perhaps ask the patients if they think they are on placebo or treatment and compare the success rate of guessing across the groups? It's pretty obvious and standard practice. Then, if there is evidence of unblinding, you could compare the magnitude of the treatment effect in the incorrect and correct guessers (there would be several problems with this analysis, and I suspect you would find correct guessers do have higher effects, but you could do it).

    Nothing you have raised calls into question the potential clinical benefit of ketamine.

  • Neuroskeptic

    neurosulcing: “Just out of curiosity, an active placebo would, presumably, have to have similar hallucinogenic effects as ketamine. Which drugs would be candidate drugs for this?”

    It would be tricky to find a drug similar to ketamine, it's quite unusual, the closest is probably PCP which would not be a good idea.

    So long as the patients had never taken ketamine before, the active placebo wouldn't need to be like ketamine, it would just need to cause some effects that make the same degree of 'impression' that ketamine does.

    i.e. you should tell people “You're going to get one of two drugs and they will both cause mind-altering effects… and one of them is also believed to be an antidepressant” but don't tell them which effects go with the antidepressant.

    If you do that, you could use almost anything. A high dose benzo, an opiate, LSD, amphetamine…

  • Neuroskeptic

    CM: I can see your point but I think it's very short-sighted.

    Even granting that whether ketamine is an active placebo isn't relevant to clinicians or patients, it's of great importance to researchers and Pharma.

    For example, if we knew that ketamine was just an active placebo, there'd be no point in researching AZD6765 because we'd have known, from the fact that it's not psychoactive, that it wouldn't work nearly as well.

    And there are surely many other similar drugs Pharma has in the pipeline.

    All that money could be wasted.

    Likewise there's lots of animal work going on trying to find out why ketamine is an antidepressant. If ketamine really is one, that's good research.

    But if it's an active placebo, it's all for nought: lots of money wasted, animals killed, for no reason.

    And that's just directly – indirectly, the whole 'glutamate hypothesis' – if it's wrong because ketamine's an active placebo – could be diverting attention away from the real causes of depression which would end up hurting patients.

  • CM

    Which point is short-sighted? The point about it being relatively cheap to just ask the patients if they think they got the drug?

    But this is the real question: what would it mean for ketamine to “really be an anti-depressant”? It sounds like you want it to affect mood without affecting mind in a way of which the patient is conscious. This is a very odd standard and is bordering on dualist: it must affect the physiological underpinnings without causing a consciously detectable shift in mental states? An alternative to just asking the patient would be to combine a dissociative substance with an amnesic substance to test this, but, again, why would you?

    The reason I object in such strong terms is that you have previously referred to ketamine as an expensive illusion. It is not expensive and it has an effect. Should basic science try to explain the effect? Absolutely. Would it be interesting if it was mediated by a placebo/belief type effect? Definitely! Although, perhaps even more interesting, such a belief might be made possible by particular dissociative substances. Depressed people are not reknowned for changing themselves through self-belief. It certainly doesn't seem to be improved by most or all psychoactive substances, as self-medication with illicit drugs is reasonably common and typically worsens rather than improves mental illness. As for the “glutamate hypothesis”… well… you can work on that if you like.

    Please articulate what a “real” anti-depressant should do, and why it can't affect the mind in a way that causes the patient to realise they have had treatment.

  • Neuroskeptic

    A real antidepressant should make you less depressed, even if you don't expect it to do so.

    The active placebo concern with ketamine is that you're essentially telling people “If you get ketamine, you will become less depressed, if you get placebo, you won't” – and then informing them that you've given them ketamine (because of the unmistakable psychoactive effects).

    Now if you just took random depressed people who'd never heard of ketamine and slipped them some ketamine in their coffee, and they still got less depressed, that would go a long way to getting around that problem – however, it's obviously not ethical…

    So the next best approach is an active placebo controlled study.

    You're right that asking people which group they think they're in would be a step forward. However the problem with that is that, if ketamine is a real antidepressant, people might guess their group based on the fact that they got less depressed. So finding that people on ketamine who guessed ketamine got less depressed wouldn't really answer the question.

  • DS


    Why are you so convinced that an antidepressant should make you less depressed? Is it not possible that the action of an effective medical treatment of depression could be through dissociation?

  • CM

    I think you're starting to get it:

    “You're right that asking people which group they think they're in would be a step forward. However the problem with that is that, if ketamine is a real antidepressant, people might guess their group based on the fact that they got less depressed. So finding that people on ketamine who guessed ketamine got less depressed wouldn't really answer the question.”

    So, following that logic, any effective antidepressant is “effectively” (sorry) unblinded. Yet, in a real and important sense, they don't know they've got it, because they might have got the placebo. They are unblinded only by the action of the drug. Furthermore, other drugs with consciously detectable effects (caffeine, benzos, anti-psychotics) do not appear to have the same anti-depressant properties. So, perhaps detecting the action of the drug is necessary to its effect. That would be interesting, and it would be interesting to know what properties of the experience elicited the antidepressant effect. But it wouldn't make it an expensive illusion, which is what I am objecting to so strenuously.

    Now, you also say that ketamine has “unmistakeable psychoactive effects”, but I bet you'd find that people in the placebo group also experience some of these and some of those people think they have got ketamine. That is how placebo groups work. They get (typically lesser versions of) the effects and side-effects that are expected from the treatment. So, especially for people with no prior experience of ketamine, who knows if the effects are “unmistakeable” or not?

    If ketamine was on-patent it would already be over-prescribed and we'd have experts talking it up in every city.

  • Neuroskeptic

    CM: Well, we don't know if the effects are unmistakable or not, another reason your plan of testing unblinding is a good one.

    However just from personal experience, you'll probably know if you've taken ketamine…it wouldn't be such a popular illicit drug if you didn't.

    And rates of psychoactive effects are always higher than placebo in these ketamine trials (unlike with AZD6765).

    Now on the point about all effective antidepressants leading to unblinding because they work: yes, true, but in that case the unblinding can't explain the antidepressant effects.

    And that's true of all drugs. If you're dying of sepsis and someone injects you with a mystery drug and you're cured, you'll know it's an antibiotic.

    That's unavoidable. But quite different from my concern about ketamine.

  • Neuroskeptic

    P.S. “Furthermore, other drugs with consciously detectable effects (caffeine, benzos, anti-psychotics) do not appear to have the same anti-depressant properties.”

    That's exactly what I'd like to establish. It's true that benzos are not regarded as clinically effective antidepressants. But we don't know what would happen if you injected someone with a benzo having first established a strong expectation that it would be an antidepressant by recruiting them into a very involved trial at the NIMH involving being admitted as an inpatient.

    Now I don't believe that such expectation effects would actually be able to cure someone of severe depression, over a timescale of weeks and months, but I've suffered depression and I know that even in the midst of an episode you can kid yourself into thinking you're better, for a few days, especially if intoxicants are involved.

    That's what I worry about. Maybe it's a senseless worry but that's why all I'm calling for is a trial to find out.

  • Zigs

    @Neuroskeptic: What are your thoughts about GLYX-13? Initial reports suggest rapid ketamine-like antidepressant effects without any side effects.

  • Neuroskeptic

    Oh, I hadn't seen that one, thanks. Looks promising, but we'll have to wait until it's published to be sure.

  • Anonymous

    I don't know if this will be of any use or not but, I've been in a major depressive episode for 2 years. I take Lithium, Mirtazapine and Quetiapine to no avail…I take dizapam Prn and i dont like taking it because well, it works! benzos work, they make you feel floaty and relaxed.. as would alcohol or heroin.. so i use it with caution.. and in a severe depressed state lets face it, we would take anything to not feel even remotely as bad as we do..

    As you can imagine I was desperate so I took part in a Ketamine trial in England, a few months ago.. here is the trial;

    I was given 0.5mg per Kilo, Infused over 40 mins, 2wice a week for a month..

    Just as the infusion finished i was starting to feel better, not well better, but i was in “ketamine land” i was disassociated form my body.. things looked odd which was scary and noises where loud and bizzare.. so one part of me wanted it to end, but i wasn't in reality so i wanted to stay there because well.. it wasnt real life..Real life is horrific.

    You come down from the trippy land rather fast once the infusion stops and within 10 mins of that i was depressed again..

    I sadly had no response to the ketamine.. and i asked the team about the trial.. They found apparently that most people do not respond.. 30% have a “favourable” response.. They said anyone that has benefited will be given ketamine off label, but it has to be given at the hospital and they still don't know the long term safety of that..

    They applied for more funding but did not receive it.. i think because well there just hasn't been this “amazing” response in the majority of us major depressives given ketamine.. the handfull that it worked for, brilliant yes it is rapid and magical and marvellous.. but thats not the reality for the majority.

    The trial has no ended and a paper on the results is being done so it will be interesting to see figures and percentages etc, when that comes out..

    so now i must place my faith back in trying new meds.. and consider ECT..

    sorry if it doesn't make sense, im very ill, my cognition is poor and to top that off i'm Dyslexic.

  • MoSho

    A friend of mine participated in a ketamine trial here in Oslo, Norway done by the norwegian public health institute. I don't think they're published the results yet, but my friend and the staff that had conducted the study were very clear that the healthy controls given ketamine experiences extreme paranoia after administration of the drug. Just thought I'd put it out there, I am awaiting the results eagerly.

  • Neuroskeptic

    Yes the psychoactive effects of ketamine are powerful. A friend of mine was a heavy user for a while & he described taking a moderate dose on a public bus once, the experiences he had were straight out of a textbook definition… of schizophrenia.

  • Jack Juni

    What about recent paper in Science showing dramatic effect of ketamine on synapse creation and density. Although not done in humans, this would seem to provide evidence that ketamine does, in fact, change the circuitry

  • Anonymous

    I've never been what I'd call a heavy user of ketamine, though I've definitely taken ketamine on the bus before, like your friend. Buddy who used to slang the stuff gets on the bus, where a friend of mine and I are sitting at the back, with those kinds of seats that face each other. He immediately busts out a vial and asks if I want a bump, to which I say “sure, a small one”. Pops it on my hand, I snarf it up, and I immediately say to these people who just got on the seats across from me, who obviously noticed that I was doing drugs, “are you ketamine? you gotta be ketamine” to which they laughed pretty hard at.

    Anyways, I didn't take very much ketamine – Maybe enough to make it feel like I had a beer or two, and on the comedown, what i'd liken to (but not claim to be, though I think I heard somewhere that this was a thing) a delightful dopamine kick on the tail end the experience. I don't like doing a lot at a time when it comes around, and I don't do it often because I choose to still know what 'treat' means while everyone else in this society seems to have chosen to forget. I also don't feel that I suffer from any depressive/anxious symptoms. I have demonstrated to myself that I can handle my business on a dose of ketamine in public, with nobody being the wiser, as long as I keep the dose a lot lower than what I see most other people electing themselves to take (Goddamn cocaine generation and their ridiculous gangsta rails). Reading the above comment from the really depressed guy who did that trial in britain, and the comment on the Norwegian study where healthy controls end up paranoid, it makes me wonder what kind of equivalent dosages people are getting out of these trials where the stuff is being poked right into their systems in comparison to 'a little bump'. I was under the impression that the initial studies where this was being done were barely perceptible dosages? I'd like to see more studies where people are given less than .5mg/kg, not enough to get googly-eyed, and see how they turn out. I get the sense that this is a dose-dependent treatment.



No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.


See More

@Neuro_Skeptic on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar