I am on the autism spectrum and I take a 21-strain probiotic, around 60 billion CFU, twice a day. When I stop for one or two weeks, I become permanently bloated. In terms of volume, it feels like I have eaten a big meal after eating almost nothing. And a gas smell like having eaten something completely off.
It is not just a comfort detail, it eats at your physical self-confidence.
This article separates what I observe on myself from what the literature actually holds. A testimony is individual and a study is general. Both are useful, they simply do not weigh the same.
What the gut sends to the brain
The gut and the brain are not two organs chatting now and then. They are wired together permanently by the vagus nerve, the immune system, stress hormones and the molecules bacteria make. That is what the reference review on the gut-brain axis describes, and most of the traffic runs from the gut upward, not the other way.
The figure that makes this concrete comes from a Belgian population study. In 1,054 people, validated in 1,070 more, two butyrate-producing genera, Faecalibacterium and Coprococcus, were consistently associated with higher reported quality of life. Coprococcus and Dialister were also depleted in people with depression, and the association held after correcting for antidepressant use, which is the first trap in this kind of analysis.
Autism and the gut: the solid part, and the part that derails
The solid part is clear. A meta-analysis of fifteen controlled studies found roughly a 4.4-fold higher risk of gastrointestinal problems in autistic children, led by constipation, diarrhoea and abdominal pain. This is not the impression of a few autistic people, it is one of the most solid observations in the article.
The part that derails is the meaning we can give it. The idea of a failing microbiome producing autism circulates, and the best-powered study says something else. In 247 children with stool metagenomics, the authors found almost no direct association between diagnosis and microbiome. Exactly one bacterium stood out, Romboutsia timonensis.
That same study was, however, well powered to detect other links, and it described a coherent chain: restricted interests go with a less varied diet, a less varied diet with a less diverse microbiome, and that with looser stools. The authors explicitly cautioned against the idea of a microbiome driving autism.
This reading minimises nothing, it moves the lever. If the depleted microbiome mostly comes from the plate, then it is worked on from the plate, which is both less spectacular and far more reachable than a treatment. Even though in my case a varied plate or not did not change much about the 'outcome'.
Then there is the heavy lead, microbiota transfer. The two-year follow-up of Microbiota Transfer Therapy reported sustained digestive improvements and autism-related symptoms still declining after treatment ended. The problem fits in one line: 18 participants, open label, no placebo, and the authors themselves call for a controlled trial. In a field where parental placebo effects are large, that stays a lead, not a treatment.
The stress that goes through the gut
For years, before a morning exam, my gut emptied completely. Not discomfort: emptying. Afternoons were fine, mornings were not. Since my transit stabilised, only a faint trace is left, and the symptom changed shape: now it is nausea when stress climbs really high.
My trick against it is not scientific: a menthol lozenge until it passes. A strong mint taste gives attention something else to do. The molecule does have a real literature next door, for a different use: peppermint oil beats placebo on irritable bowel symptoms, with a real effect but very low quality of evidence and more side effects, mainly reflux.
The interesting part lies elsewhere. I do not feel stress until it is very high: mentally, nothing happens. I see the effects before I see the cause. A gut emptying tells me about my state when my own perception tells me nothing. For someone who reads internal signals poorly, the body becomes the sensor, and that is exactly what objective tracking is supposed to make readable.
Then comes the loop. Once the symptom starts, telling yourself not to think about it is thinking about it, and thinking about it makes it worse. The gut-brain axis runs both ways: that circuit is what needs cutting, not the event that started it.
21 strains, 60 billion: what it is actually worth
The international consensus on the term probiotic defines a probiotic at three levels: genus, species and strain. Two strains of the same species can behave differently, the way two breeds of the same dog do not have the same aptitudes.
Meta-analyses that sort at that level show it plainly. Across 42 trials and 3,856 participants, some strains clearly reduce abdominal pain while others show no measurable effect on the same outcome. The meta-analysis published in Gastroenterology reaches the same place: overall efficacy on irritable bowel symptoms, but no way to recommend a specific product given how species, strains and endpoints differ between trials.
| What the label says | What it is worth |
|---|---|
| « 60 billion CFU » | A count of live bacteria, not evidence of effect |
| « 21 strains » | Useful only if each is named down to its strain code |
| « Lactobacillus plantarum 299v » | Verifiable: the strain maps to published trials |
| « Lactic ferments » | Maps to nothing |
On my own intake, honesty requires two caveats. I cannot know which of the 21 strains is working, and the observation is not blinded, so expectation counts (one day I should really get my microbiome analysed). But the stop and restart alternation has repeated several times, in the same direction, with a sign I did not choose and would rather not have. That is a repeated observation, below a controlled trial and above an impression.
What the plate does better than the capsule
A capsule brings a few known strains. Fermented food brings a whole living community, plus the material it feeds on. Neither replaces the other.
Kefir is the best example, provided it is real kefir and not a flavoured milk drink: the difference is in the yeasts, and I wrote a whole article about it. Yoghurt, for its part, has something rare here: EFSA accepted a cause and effect relationship between live yoghurt cultures and improved lactose digestion, on thirteen concordant human studies out of fourteen. It is one of the very few validated claims in a field where nearly all the others were rejected.
It is also my practical way around lactose intolerance. Milk does not pass, yoghurt does, and the EFSA opinion explains why: live cultures digest part of the lactose for me. The glycine I take at night, which is supposed to act only on sleep, also slightly improves my digestion.
What this article does not say
- That probiotics treat autism. No solid data supports it, and the best-powered study points the other way.
- That my 21 strains are a good product. I cannot verify it, which is exactly my complaint about labels.
- That persistent digestive trouble is a microbiome matter. Chronic diarrhoea, blood, weight loss or anaemia are reasons to see a doctor, not to buy a supplement.
- That butyrate in a capsule reproduces what bacteria producing it on site do. Nobody has shown that.
- That my experience counts as proof. It counts as a starting point, and that is how I present it.
The common thread is simple enough. A gut that misbehaves costs energy, attention and confidence every single day, and that cost appears in no blood panel. It is the kind of variable nobody measures, and it decides the quality of a whole day.