Uh-oh, here's a troubling paper: Effects of selective serotonin reuptake inhibitors on motor neuron survival.

According to Anderson et al,

Motor neurons were challenged with fluoxetine and paroxetine at clinically relevant doses ... In fluoxetine-treated motor neurons there was ~52% cell death while in paroxetine-treated cells there was 14% cell survival.... Both SSRIs decreased cell survival in a dose-dependent manner. This study is provocative enough to call for further in vivo studies.

Standard doses of fluoxetine (Prozac) and paroxetine (Paxil), two of the biggest-selling SSRI antidepressants, both had a devastating effect on the survival of mammalian nerve cells growing in a dish. Within just 24 hours, the antidepressants had caused over half the cells to expire, 85% of them in the case of Paxil. Take a look:

Oh dear. Should we be worried for the safety of the nerves of everyone taking SSRIs? I'm not, for two reasons.

Reason #1 is that one of the authors on this paper was a Dr Amy Bishop, aka the University of Alabama Shooter. Three months ago, Bishop shot dead 3 colleagues and injured 3 others during a departmental meeting at the University of Alabama in Huntsville.

A blog called Shepherd and Black Sheep did some nice investigating into Bishop's work following the shootings. They pointed out that of the five authors on the SSRI paper, one was Amy Bishop, and the other four were her husband and three children. They're all listed as working for a "Cherokee Lab Systems", but
Hmm. But, hang on, the fact that this research was done by a spree killer and her family operating from their garage, doesn't mean it's wrong. Maybe the research was why she did it: Was Dr. Bishop Mind-Controlled To Shoot Her Victims As Retaliation For Her Critical Research Paper On Prozac In 2009? Hey, I'm just asking questions.

Fortunately there's a second reason to doubt that SSRIs cause this kind of motor neuron toxicity in people: SSRIs don't cause this kind of motor neuron toxicity in people. If they did, everyone would collapse and/or die after popping a single Prozac.

That millions of people have been taking SSRIs for years and, whatever else may have happened to them, they are still standing, is proof they don't. The fact that a low dose of alcohol (0.08% i.e. just "legally drunk") also killed a large proportion of nerve cells in this study is more evidence that these results can't be extrapolated to humans. Antidepressants are in fact commonly used to help treat the pain resulting from neuropathic nerve damage.

It's always possible to explain away such differences. Maybe SSRIs can cause motor neurotoxicity in humans, but only at higher doses, or only after very long periods, or only in some people. Anything's possible. But given that the effects actually seen in this study - dramatic, rapid, and present at normal doses - never happen in humans, such possibilities are no more likely, in the light of these results, than they were before.

This is the serious lesson of this paper: if you've got data showing that something causes some effect in animal models or cell cultures, but it doesn't do that in real life, the simplest explanation is that there's something wrong with your model, or your data. Unfortunately, Anderson et al is by no means the only example of putting the experimental cart before the clinical horse...

Lily B Anderson, Phaedra B Anderson, Thea B Anderson, Amy Bishop, & James Anderson (2009). Effects of selective serotonin reuptake inhibitors on motor neuron survival International Journal of General Medicine, 2, 109-115