Aug 26 2013

Probiotics for Mental Health?

I frequently criticize science journalism for falling into one or more common traps – false balance, hyperbole, misleading framing, failure to put one study into the proper context, inappropriate extrapolation, and others. Here is one article discussing the relationship between gut bacteria and mental health that can be used as a textbook example of how not to write a science news story.

The article begins, as is unfortunately often the case, with an emotional anecdote. Journalists are taught to find a human angle to draw the reader in, and I get this. But this style is better suited to fluff pieces than serious science journalism.

There are multiple problems with this style as applied to a science topic. The first is, of course, that the story is anecdotal. We cannot know what the implications of this story are. It is likely highly selected – chosen out of many possible examples to be the most dramatic and emotionally appealing example of whatever story the journalist wants to tell.

In this case we are told the story of Mary who has severe refractory obsessive compulsive disorder (OCD). All the emotional hooks are there – her parents are desperate, all previous treatments have failed, and the doctor they are seeing (the focus of the piece) is their last hope.

The doctor, James Greenblatt, is presented as a maverick, almost a guru, who sees something no other doctor has seen. His ideas, specifically that gut bacteria are a major contributor to mental illness and can be treated with probiotics, are treated as prophetic, ahead of their time, and that they are in the process of being validated, which at this point is almost just a formality.

His treatment consists of drugs, therapy, and probiotics, but of course the probiotics get the credit for any improvement, which is described by the journalist as “miraculous,” thus completing the standard medical anecdote formula. In case you missed it, here is that formula:

– Sick patient no one can figure out or successfully treat
– Desperate family/parents
– Maverick doctor with unusual treatment ideas
– Miraculous response to treatment
– Reporting of preliminary and cherry picked research as if the conclusions are inevitable or obvious
– Token skepticism optional.
– Finish with an emotional flourish that immediately negates the token skepticism.

This is fluff journalism inappropriately applied to science topics typically by people who are not trained science journalists.

Probiotics for mental health

What about the topic itself – does the bacteria that colonizes our GI system have a dramatic effect on brain function, including mental illness? At present this is an intriguing idea with some promise and preliminary evidence, but that’s all. It is very unlikely that intestinal flora have as much impact as the article suggests, and there are numerous other factors that are well established, such as genetics and environmental factors.

The research cited in the article is mostly animal research, which essentially shows that mice with certain GI flora have higher levels of stress hormone. That is a plausible simple mechanism – increased levels of cortisol raise the overall level of stress, which could exacerbate things like OCD or anxiety. Exacerbating a pre-existing condition is not the same as causing that condition, however. Even this relatively simple effect is not established in humans, who, it turns out, are not mice. (Animal research is legitimate, but caution is always recommended in extrapolating to humans, especially in psychology studies.)

Some of the studies also do not use probiotics but fecal transplants to reconstitute the gut biome of the mice being studied. This cannot be extrapolated to probiotic use (I’ll discuss this more below).

The article author, Carrie Arnold, mentioned one human study, although the link she provided was to the wrong study. I did manage to track down the study to which she referred – a study in women exposed to fermented milk with four probiotics, milk without probiotics, and no intervention (number of subjects in each group 12,11,13 respectively). They found that emotional response to faces as measured by fMRI was decreased in the probiotic group.

The usual caveats apply – this was one study, the sample size was very small, and fMRI scans (functional magnetic resonance imaging) is a tricky technology that requires great care and controls, and lots of averaging to get a signal out of the copious noise. No reliable conclusions can be drawn from one small fMRI study.

To summarize – the state of the research is very preliminary, mostly involves animals, the one human study is small and not much can be concluded from it, and often the interventions are more extreme (fecal transplants) than simply taking a probiotic product. Also the outcomes are difficult to interpret and we may just be seeing basic physiological effects on the brain.

For example, I already mentioned stress hormone activity. Gut bacteria might also plausibly affect glucose metabolism, or  aspects of the endocrine system such as pituitary function. What this means is that gut bacteria likely do have a primary effect on some basic physiological functions of the body, which in turn can secondarily affect brain function in terms of mood,  anxiety level, and other basic emotions.

The evidence does not point to a direct effect of gut bacteria on brain wiring or function. It can be explained more simply as just the brain responding to the physiological environment of the body (nothing new or surprising). Some of the results may be as simple as animals being in a bad mood when their tummy aches, but I’ll grant that it’s likely there is more to it than that.

Even if we grant that the constitution of bacteria in the GI system has an indirect effect on stress, anxiety, and emotions it still remains to be seen if current interventions have a significant effect on the GI flora. The evidence so far warrants caution. Mark Crislip wrote an excellent review of this topic on science-based medicine. In short, the species of bacteria that make your GI system their home exist in a stable ecosystem. It is very difficult to alter that ecosystem by adding one or a few probiotic species. The new species simply get edged out by the established species., and the existing equilibrium is maintained.

Therefore, if we do discover a relationship between GI flora and mental health, we currently don’t have the knowledge or technology to do much about it.

All of this means that existing probiotic products and interventions are very  unlikely to have a significant effect on the treatment of any mental health issue. In any case, there are no clinical trials demonstrating the efficacy of probiotics in mental disorders. Especially given the preliminary and implausible nature of the basic science, such definitive clinical efficacy trials would be needed before any health benefit claims for specific probiotic products would be justified.

In contrast, Carrie Arnold characterized the situation with this paragraph – easily the worst in her piece:

Although plenty of questions remain, the benefits of using probiotics to treat human behavior are becoming increasingly obvious. Yogurts like Dannon’s Activia have been marketed with much success as a panacea for all of our intestinal ills. Other probiotic supplements have claimed to support immune health. Probiotics’ potential to treat human behavior is increasingly apparent, but will manufacturers one day toss an anxiety-fighting blend into their probiotic brews?

She is essentially using marketing claims to justify the scientific conclusion that the benefits of probiotics are “increasingly obvious.” This, of course, is absurd and flies in the face of recent history. Antioxidant claims, for example, vastly exceeded the evidence and now it seems that the hype of antioxidants was entirely misguided. Such marketing claims (especially in the unregulated market of supplements) tend to be premature, over-hyped, and ultimately wrong.


When I write about probiotics I always feel compelled to point out that I have nothing against the concept. I actually think that we are at the early stages of discovering that the bacteria that are symbiotic with humans play a large role in our health. They are therefore potentially a useful area of research to explore this relationship, and figure out ways to improve health by altering the bacteria that live on and inside us.

Like any single approach to health or new technology, however, early hype tends to vastly exceed the reality. What is likely to be just one more complex and limited tool in our box is talked about as if it will be a panacea. While scientists are still trying to figure out the “ifs,” the sensationalist media is presenting future scientific knowledge as if it is inevitable and already known. The point of research is not to demonstrate that something is true but to ask if it is true, and until that research is done, we truly do not know. Further, most new ideas tend not to be true, or to be far more limited than initial enthusiasm would suggest.

This is now a common pattern. In the decade or two when a new scientific idea in medicine is being worked out, and may or may not pan out as a good idea and successful treatment, the hype surrounding this new idea is exploited to sell snake oil. Stem cells are an exciting idea, but current stem cell clinics are fraudulent. Antioxidants were an interesting idea for an intervention, but as our knowledge deepened it became apparent that they were more of a double-edged sword than elixir of life. Viral vector gene therapy will likely be an important treatment option, but the technology has proved trickier than we imagined, to the tune of decades of further research.

Probiotics are still in this early phase as well. Existing treatments and products, according to the research currently available, range from worthless to perhaps having a minor benefit in certain specific situations. This is a useful area of research, but it remains to be seen what the ultimate potential of probiotic therapy will be, what level of interventions will be required, and how long this will all take.

12 responses so far