Gut Microbes Track With Autism Symptoms, but Inflammation Tells a Different Story

For more than a decade, one of the most tantalizing ideas in autism research has been that the gut holds a clue to the brain. Children with autism spectrum disorder (ASD) experience gastrointestinal problems at strikingly high rates, and many parents report that digestive troubles and behavioral flare-ups seem to travel together. A new cross-sectional study published in Pediatric Research by a team at UTHealth Houston and Baylor College of Medicine puts that idea to one of its most rigorous tests yet, and the results are as revealing for what the researchers did not find as for what they did.

The study, led by Vaishali Harne and J. Marc Rhoads, examined baseline data collected between June 2016 and August 2023 from children aged 4 to 16 with confirmed ASD diagnoses, alongside stool samples from their siblings and typically developing children. Seventy-two children with ASD provided blood and stool samples for biomarker analysis, while 135 children contributed stool for microbiome sequencing. Every child in the ASD group met DSM-5 criteria, verified with gold-standard instruments including the Autism Diagnostic Interview-Revised and the Autism Diagnostic Observation Schedule, Second Edition, or its telehealth adaptation developed during the COVID-19 pandemic.

The central finding is a statistically significant correlation between gastrointestinal symptom severity and autism symptom severity. The researchers quantified digestive complaints using a modified GI Severity Index that incorporated Bristol stool scoring, and they measured autistic traits with the Social Responsiveness Scale, Second Edition. The two measures moved together: children with more disturbed social awareness, particularly on the social awareness and social motivation subscales, tended to have more severe GI symptoms. The correlation between total GI severity and total SRS-2 scores reached statistical significance, with correlation coefficients around 0.29 to 0.30, and the social awareness subscale showed an even stronger association. Children with the highest GI burden scored, on average, more than seven points higher on the social awareness T-score than those with milder symptoms, after adjusting for sex, age, ethnicity, and verbal ability.

But here is where the story takes an unexpected turn. The dominant hypotheses for gut trouble in autism have long centered on inflammation and a so-called leaky gut, in which the intestinal barrier becomes permeable and allows microbial products to seep into circulation. To test this, the team measured an unusually comprehensive panel of biomarkers: plasma zonulin and occludin, two markers of intestinal tight junction integrity; plasma S100A9, an alarmin protein central to innate immune activation; fecal calprotectin, the gold-standard marker of gut inflammation; and plasma S100B, a glial protein used to gauge neuroinflammation and blood-brain barrier disruption. The verdict was unambiguous. None of these markers were elevated above published normal values, none correlated with GI severity, and none differed meaningfully between children with high and low symptom scores. Fecal calprotectin was low in both groups, and the two children who did show elevated values had no GI symptoms at all. S100B was undetectable in so many participants that it was excluded from further analysis.

This null result matters because it cuts against a substantial body of conflicting literature. Some earlier studies reported elevated zonulin in children with ASD, others found reduced levels, and still others found no difference at all. Similar contradictions surround occludin, with one recent study reporting increases and another reporting decreases. The UTHealth team argues that blood markers of gut permeability in autism have been inconclusive across the field, and their own data, in which zonulin values actually ran low, offer no support for the leaky gut hypothesis as an explanation for digestive symptoms. Likewise, the suspicion that autistic children harbor a smoldering intestinal inflammation akin to inflammatory bowel disease finds no confirmation here, consistent with a recent systematic review that found no significant differences in fecal calprotectin or lactoferrin in ASD.

If inflammation is not the culprit, what is? The microbiome data point toward a different kind of explanation rooted in microbial composition rather than immune activation. Using 16S rRNA sequencing of the V4 region on an Illumina platform, with the DADA2 algorithm and QIIME2 pipeline for processing and the SILVA database for taxonomic assignment, the researchers found that children with ASD had reduced fecal microbial richness compared with both their siblings and typically developing children. Beta diversity analysis revealed distinct clustering of the ASD samples, separating them from the more diffuse microbial profiles of the comparison groups. At the genus level, children with ASD showed increased relative abundance of Fusicatenibacter, Blautia, Sarcina, and Eubacterium species compared with typically developing peers.

The correlations between specific microbes and specific symptoms were the study’s most granular contribution. Children with ASD who had significant GI symptoms carried higher levels of Bacteroides and Ruminococcus, and these organisms tracked positively with flatulence, foul-smelling stools, nighttime awakening, and overall GI severity. Seven bacterial groups showed strong correlations with global symptom burden: the Ruminococcus torques group, Coprococcus, Faecalibacterium, the Eubacterium hallii group, Anaerostipes, Lachnospiraceae, and Sarcina. Sarcina, a genus previously associated with delayed gastric emptying but never before linked to autism, correlated positively with flatulence, abdominal pain, diarrhea, and disrupted sleep. Butyrivibrio correlated strongly with diarrhea, while Anaerotruncus and Candidatus Soleaferrea showed negative correlations with symptom scores. Linear discriminant analysis identified Blautia and Faecalibacterium as the two taxa that best discriminated the ASD group, both with effect sizes above the conventional threshold.

Several of these organisms carry intriguing biological baggage. Ruminococcus degrades the mucus layer that protects the intestinal lining, and Ruminococcus torques is known to expand in pre-clinical Crohn’s disease. Higher abundance of Ruminococcus in infancy has been linked to poorer social functioning at age three, and mono-colonized mice carrying Ruminococcus gnavus produce elevated tryptamine, a microbial metabolite with documented effects on gut-brain signaling. Faecalibacterium, by contrast, is generally considered a beneficial commensal that produces butyrate, a short-chain fatty acid with anti-inflammatory properties, yet it has appeared at elevated levels in other autism cohorts, possibly as a compensatory response to restricted diets. Blautia presents its own paradox: a live biotherapeutic product of Blautia stercoris has been shown to attenuate social deficits in mouse models of autism, raising the question of whether its presence in symptomatic children is harmful or a failed attempt at self-correction.

The authors are careful about causality, and their limitations section is candid. The study is cross-sectional, so it cannot determine whether dysbiosis drives symptoms, symptoms reshape the microbiome, or both are downstream of a third factor such as food selectivity, which affects the majority of autistic children and can dramatically alter microbial substrates. The researchers did not draw blood from healthy controls for permeability markers, owing to ethical concerns about phlebotomy in healthy children, and the analysis relied on univariable comparisons that leave room for residual confounding. Notably, however, recent population-based studies have found microbial differences in children who later develop autism as early as one year of age, before dietary selectivity typically emerges, suggesting that at least some of the dysbiosis may precede both diet and symptoms.

The practical implications are twofold. Clinically, the findings suggest that GI complaints in autistic children, which affect somewhere between 31 and 84 percent of this population, are real and track with behavioral severity, but they are unlikely to reflect occult inflammation that would warrant anti-inflammatory treatment. Therapeutically, the results temper expectations for microbiome-targeted interventions. Probiotic trials have shifted microbial profiles without consistently improving symptoms, and an open-label fecal microbiota transfer study showed promising gains in both GI and behavioral measures but lacked a placebo control. What this study offers is a sharper map: a reproducible signature of reduced diversity and specific symptom-correlated taxa, and a clear signal that the long-sought inflammatory smoking gun is absent, at least in children without underlying gastrointestinal disease. The gut-brain conversation in autism, it appears, will continue, but it will have to proceed without the leaky gut narrative it so often leaned on.

Subject of Research: The relationship between gut microbiota, gastrointestinal symptoms, behavior, inflammation, and intestinal permeability in children with autism spectrum disorder

Article Title: Gut microbiota are linked to gastrointestinal symptoms and behavioral changes but not inflammation in autism spectrum disorder: a cross-sectional analysis

Article References: Gut microbiota are linked to gastrointestinal symptoms and behavioral changes but not inflammation in autism spectrum disorder: a cross-sectional analysis. (n.d.). https://doi.org/10.1038/s41390-026-05548-6

Image Credits: AI Generated

DOI: 10.1038/s41390-026-05548-6

Keywords: autism spectrum disorder, gut microbiota, gastrointestinal symptoms, dysbiosis, fecal calprotectin, zonulin, intestinal permeability, 16S rRNA sequencing, Social Responsiveness Scale, Ruminococcus, Faecalibacterium, gut-brain axis