Do Antidepressants Make Some People Worse?

Antidepressants may help depression in some people but make it worse for others, according to a new paper.

This is a tough one so bear with me.

Gueorguieva, Mallinckrodt and Krystal re-analysed the data from a number of trials of duloxetine (Cymbalta) vs placebo. Most of the trials also had another antidepressant (an SSRI) as well. And the SSRIs and duloxetine seemed to be indistinguishable so from now on I'll just call it antidepressants vs. placebo as the authors did.

People on placebo got, on average, moderately better over 8 weeks.

People on antidepressants fell into two classes. The largest class got, on average, a lot better. But about 25% did poorly, staying just as depressed as before. This "nonresponder" group did much worse than the placebo group - again on average. Here you can see the mean "trajectories" of depression symptoms (HAMD scores) in the three groups:

This raises the scary possibility that while antidepressants are helping some people, they're harming others. But hang on. It's complicated.

First off, maybe this is all a statistical illusion. When the authors say that the people on drug fell into two classes, what they mean is that when you try to model the data according to a certain mathematical model, assuming either 1, 2, 3 or 4 underlying classes, the 2 class solution was the best fit. While for placebo a 1 class solution was best.

We considered linear, quadratic, and cubic trends over time, with between 1 and 4 trajectory classes. We also considered piecewise models with a change point at 2 weeks, linear change before week 2, and quadratic change after week 2. The selection of the best model was based on the Schwartz-Bayesian information criterion and on the Lo-Mendell-Rubin (LMR) likelihood ratio test...

That's nice... but they don't present the raw data. They don't tell us whether, looking at the individual trajectories of people on antidepressants, you'd actually see two classes. What I want is a graph of how likely people are to get better by a certain amount. If Gueorguieva et al are right, I want it to look like this i.e. bimodal -

We're not shown this graph. I'll eat my hat if it does look like that, frankly, because if it did people would have noticed the bimodality in antidepressant trials ages ago.

True, statistical models can tell us things that aren't obvious by inspection, so even if this isn't what the data look like, they might still be right. It could be that the two "peaks" are so broad, and there's so much random noise, that they blur into one.

However, it's also true that you can fit an infinite number of models to any set of data and at some point you have to step back and say - am I making this more complicated than it needs to be?

It could be that a 2-class model is better than a 1-class model for the people on antidepressants, but only because they're both crap, and really, every patient has a different, unpredictable trajectory which is poorly captured by such models.

Let's assume however that this is true. What would it mean?

Firstly, the fact that one class of people on antidepressants does worse than people on placebo doesn't mean that antidepressants are harming them. The authors miss this point, when they say

there are 2 trajectories for patients treated with antidepressants and 1 trajectory for patients treated with placebo [so] some patients would seem to be more effectively treated with placebo than with a serotonergic antidepressant.

But that's fallacious. It treats a purely statistical entity as representing individual people. Suppose that what antidepressants do is to take people who, on placebo, would have improved a bit, and make them improve a bit more than they otherwise would have. You'd then end up with more people doing well, but also fewer people doing moderately because they'd have been "moved up" out of the middle ground.

That "nudging people off the fence" could lead to a bimodal distribution and two distinct classes. But in this case the people doing badly would have done badly either way. The drug didn't make them do badly, it just made doing-badly into a class. On the other hand it's consistent with antidepressants doing real harm. We can't tell.

We do know that other randomized controlled trials show very convincingly that in a small minority of people, mostly but not exclusively young people, antidepressants do worsen suicidal thoughts and behaviours. So it's plausible. But we just don't know yet.

What worries me is that this paper is the latest in a series of attempts  to use, well, creative statistical approaches to antidepressant trial data. This one is nowhere near as dodgy as the Cherrypicker's Manifesto I discussed last year, but it cites that paper and others by the same group. The first sentence of the Abstract of this paper makes the intention clear:

The high percentage of failed clinical trials in depression may be due to high placebo response rates and the failure of standard statistical approaches to capture heterogeneity in treatment response.

In other words, the reason clinical trials of new antidepressants often fail to show a benefit over placebo is not because the drugs are crap but because the statistics aren't subtle enough. And you can see where this is going: if only we could use statistical models to find the people who do benefit from antidepressants, and compare them to placebo, there'd be no problem...

Gueorguieva R, Mallinckrodt C, and Krystal JH (2011). Trajectories of depression severity in clinical trials of duloxetine: insights into antidepressant and placebo responses. Archives of General Psychiatry, 68 (12), 1227-37 PMID: 22147842

A Psychedelic Tale of Two Neurotransmitters

An unexpected interaction between neurotransmitter systems may explain psychosis and hallucinations, according to a fascinating new paper.

Serotonin (5HT) and glutamate are two neurotransmitters. Up until now, it was thought that they acted independently. A given neuron might have receptors for both serotonin and glutamate, but they didn't interact: serotonin would never affect the glutamate receptors, and vice versa.

The new research overturns that view. Authors Miguel Fribourg and colleagues of Mount Sinai School of Medicine show, in a series of elegant experiments in mice, that different receptors can cluster together, forming a complex. The two receptors, serotonin's 5HT2A and glutamate's mGluR2, can talk to each other.

However, this doesn't seem to happen under normal conditions. Serotonin and glutamate don't seem to trigger the receptor interaction, or at least not very much. Only certain drugs can do it. And this is where it gets really interesting.

Psychedelic drugs, like LSD, have long been thought of as 5HT2A agonists, binding to the receptor and activating it. It turns out that this was only half right. They also inhibit mGluR2 transmission via the receptor complex. Serotonin itself is a 5HT2A agonist, but it doesn't do that. So psychedelics seem to be a kind of (for want of a better word) "superagonist".

It also works in reverse. The antipsychotic drugs clozapine and risperidone are known as 5HT2A antagonists. But Fribourg et al show that they also activate the mGluR2 receptor.

And the cross-talk can go in the other direction. Certain molecules that act on mGluR2 can either inhibit or promote 5HT2A. Unlike psychedelics and antipsychotics, these mGluR2 drugs have not been tested in humans yet. But these data predict that they will have psychedelic-like or antipsychotic-like effects, depending which way they work.

The interaction turns out to be all about G proteins, which are part of the chain of transmitter substances that convey signals within the cell, in response to neurotransmitters outside it. Here's a chart showing the effects of various drugs on the balance between different G proteins: the LSD-like psychedelic DOI has the opposite effect from the antipsychotics clozapine and risperidone.

This paper builds on a previous one from the same team showing that psychedelic 5HT2A "agonists" (like LSD and DOI) have different effects on G proteins from other, non-psychedelic agonists. That was interesting in itself but by adding glutamate to the picture, this new paper is really ground-breaking.

This goes a long way to explaining one of the mysteries of serotonin which is this:  if 5HT2A agonists like LSD are psychedelic, why aren't antidepressants the same? Almost all antidepressants work by increasing extracellular 5HT levels. That ought to mean that they activate 5HT2A receptors (indirectly). This explains why not - 5HT alone doesn't promote the crucial 5HT2A-mGluR2 interaction.

Taken together, these interesting results show clearly that 5HT2A and mGluR2 are hooking up and doing something exciting. Certainly in terms of how hallucinogens work.

I'm less convinced that this can directly explain antipsychotic effects though. The problem is that while newer "atypical" antipsychotics act on 5HT2A, the older antipsychotics don't, and atypicals are at best only slightly more effective on average.

What we don't yet know is whether this kind of complex receptor interactions can happen with other receptors. I'd have thought it unlikely that these two receptors were the only ones that could ever do it. The synapse looks like it's more complex than we could have imagined.

Fribourg M, et al. (2011). Decoding the Signaling of a GPCR Heteromeric Complex Reveals a Unifying Mechanism of Action of Antipsychotic Drugs. Cell, 147 (5), 1011-23 PMID: 22118459

Another Antidepressant Bites The Dust

Yet another up-and-coming antidepressant has flopped.

A paper just out reveals that the snappily-named GSK372475 doesn't work and has lots of side effects. It's a report of two clinicals trials in which Glaxo's contender was pitched against placebo and against older antidepressants in the treatment of depression.

GSK372475 failed to improve depression any better than placebo, even though the trials were large (393 and 504 patients respectively) and twice as long as most antidepressant trials (10 weeks whereas 4 or 6 is more usual)which ought to have given it plenty of room to shine.

The comparison drugs, the widely used venlafaxine and paroxetine, did work. A bit.

One of the trials even used the Bech "Melancholia Subscale" as an outcome measure, which Neuroskeptic readers may remember as I've praised it before. Venlafaxine worked on that, GSK's new pill didn't. If anything, the new drug was worse than placebo, in that patients improved slower.

In terms of side effects it caused dry mouth, insomnia, and nausea serious enough to make many people quit the study early. But even worse, it raised heart rate by almost 10 beats per minute on average, which is really never a good sign.

So, overall, it was an utter flop. In one sense this is not surprising. New "antidepressants" that don't work in trials have been all too common recently. Just last week we learned about the failure of "Serdaxin" in a Phase II trial. Actually Serdaxin isn't a new drug but an old antibiotic called clavulanic acid that a company was trying to rebrand as a mood lifter.

However the failure of GSK372475 is a bit of a mystery. The drug is a potent triple reuptake inhibitor (TRI) which acts on the neurotransmitters serotonin, noradrenaline and dopamine. By contrast, venlafaxine is a double reuptake inhibitor which doesn't hit dopamine, and paroxetine only targets serotonin. I've written about other TRIs before.

Now it seems surprising that venlafaxine worked, but a TRI didn't, in the same trial. That would imply that blocking the reuptake of dopamine makes you more depressed, enough to cancel out the other actions which are shared with venlafaxine. Which is not what I'd have predicted.

There are other differences between the drugs though. Venlafaxine has a very short half-life - it's broken down in the body in a matter of hours. But GSK372475 has a halflife of 8-10 days. Could this be the problem?

Learned S, Graff O, Roychowdhury S, Moate R, Krishnan KR, Archer G, Modell JG, Alexander R, Zamuner S, Evoniuk G, & Ratti E (2011). Efficacy, safety, and tolerability of a triple reuptake inhibitor GSK372475 in the treatment of patients with major depressive disorder: two randomized, placebo- and active-controlled clinical trials. Journal of psychopharmacology (Oxford, England) PMID: 22048884

The Teen Happiness Gene?

Whether you were happy with life as a teenager could be down to a certain gene, says a new study.

In a large study of American adolescents, the AddHealth project, teens who carried the long form of the 5HTTLPR locus were more likely to say they were satisfied or very satisified with their lives (at age 18 to 26). People with two long variants were the most cheerful, with short/long carriers in the middle and short/short being the least so.

The effect was significant controlling for ethnicity (p=0.013), however looking at the data shows that this effect was largely driven by the unhappy teens who reported being "Dissatisfied" or "Neither" on the 5 point scale of life satisfaction - but there were only a small number of these, because the great majority said they were "Satisfied" or "Very Satisfied". Still, there you go.

Incidentally, Neuroskeptic readers may remember AddHealth because of its role in the "black women are ugly" race row from earlier this year.

This study is the latest in a long, long line of attempts to correlate 5HTTLPR with happiness, depression, stress and so on. A few months ago I discussed the history of this busy little gene and covered a meta-analysis of no fewer than 54 papers which claimed that there was indeed a link, with the short allele increasing the risk of depression in response to stressful events.

However many studies failed to find one, and worryingly the three largest studies were all negative which is a classic tell-tale sign of publication bias - maybe people were only bothering to publish smaller studies if they did find a link and hence were "exciting findings". This is quite possible because so many researchers collect DNA as part of psychology studies these days. When the 5HTTLPR story got big (about 5 years ago) I know a lot of people decided to jump on the bandwagon by looking at it in the context of their old data.

Personally I have no idea whether 5HTTLPR is associated with anything. I used to think it probably did, but now I'm just confusion. There have been so many studies and so much inconsistency that it's very hard to know. What worries me is that I'm not sure whether we'll ever get a consensus. We've already had a gigantic study (over 80,000 people) showing no link and many meta-analyses coming to different conclusions.

What will it take to settle the issue? An even bigger study? Would 200,000 people do it? A million? I don't know.

De Neve JE (2011). Functional polymorphism (5-HTTLPR) in the serotonin transporter gene is associated with subjective well-being: evidence from a US nationally representative sample. Journal of human genetics, 56 (6), 456-9 PMID: 21562513

Life With Low Serotonin, Revisited

Last year I covered the case of a young man born with a genetic disorder which caused him to suffer low levels of the monoamine neurotransmitters - serotonin, dopamine, and noradrenaline.

These are the chemicals that are widely thought to be deficient in depression, and they're the target of antidepressant drugs (especially serotonin).

If low monoamines cause depression, you'd expect someone with low monoamines to be depressed, at least on the simplest view. But the case from last year had no reported mood problems, although he did show appetite, sleep and concentration problems that were cured by serotonin replacement therapy.

Now a new case report has just appeared that tells a different story. Gabriella Horvath and colleagues from British Columbia describe two sisters. Both had a normal birth and childhood, but at the ages of 11 and 15 respectively, began to suffer severe migraines and other symptoms. Sister 1:

started having hemiplegic migraine at age 11 years, initially occurring every 3–8 weeks, lasting 4–48 hours, presenting with right or left-sided numbness and paralysis, no visual disturbances, but slurred speech, associated with vomiting, headache, and confusion, followed by weakness lasting up to 7 days, and then complete recovery. The frequency of her migraine increased slowly with age up to twice a month...

Between 12 and 20 years she had developed progressive spastic paraparesis; sensory loss in stocking distribution... urinary and bowel incontinence; bladder instability... irritable bowel syndrome; sleep problems; depressed mood; and anxiety. She needed to use a wheelchair for most of the time by the age of 17.

Sister 2 had a rather different course:

The older sister originally presented at the age of 15 years with a history of hemiplegic migraine and seizures and myoclonic jerks. EEG showed generalized spike-and-wave activity, and polyspikes with photoconvulsive [light-induced seizures] response, in keeping with juvenile myoclonic epilepsy. Her seizures were brief and infrequent and not associated with the migraine episodes...

She subsequently developed progressive weakness, frequent falls, depression, and mild bladder instability...

Various blood and genetic tests failed to get to the bottom of it. MRI scans showed abnormalities in the spinal cord and parts of the brainstem in both cases, but why?

Spinal tap studies in Sister 1 revealed very low levels of 5HIAA, which is a by-product of brain serotonin (5HT). This suggested low 5HT levels. So doctors started her on 5HTP to try to boost it.

They report that 5HTP treatment caused "improvement" in all symptoms, including the migraines, slurred speech, depression, and movement, but not immediately. She gradually went from being in a wheelchair to being able to walk around the house on crutches, although she used a wheelchair outside. However, after 3 years of treatment, at age 20, she suddenly fell into a coma lasting 2 months. She is now recovering.

Sister 2 also had low 5HIAA, and was given 5HTP. She also reported symptomatic improvement.

Blood tests reported very low platelet serotonin levels. 5HTP treatment increased this but they were still below normal. Platelet 5HT reuptake rate was also low, suggesting a problem with the 5HT reuptake transporter protein 5HTT.

But the 5HTT gene (famously known as "The Happiness Gene" although that's questionable) seemed entirely normal in these patients. The authors say however that the symptoms are, in some ways, reminiscent of mice who lack the 5HT reuptake protein (5HTT knockout mice), who also show low serotonin. Also, if it were genetic, that wouldn't explain why there were no problems at all during childhood.

So this case is a mystery. The low serotonin has no known cause, and it might just be a side effect of a deeper underlying problem, but serotonin has long been linked to migraines so it might account for some of the symptoms. The fact that 5HTP helped supports this, though it wasn't a controlled trial so we can't know for sure.

As for the depression and anxiety, improved by 5HTP, this could have been a result of low serotonin, but it could also have been a psychological reaction to the severe medical problems. It's impossible to know.

Horvath GA, Selby K, Poskitt K, Hyland K, Waters PJ, Coulter-Mackie M, & Stockler-Ipsiroglu SG (2011). Hemiplegic migraine, seizures, progressive spastic paraparesis, mood disorder, and coma in siblings with low systemic serotonin. Cephalalgia : an international journal of headache PMID: 22013141