Two Opposite Diet Trials Found the Same Gut Microbe Signal

Two opposite diet moves, one shared microbiome result

You can push diet quality in opposite directions and still get the same gut-microbe signal. In 2025, one randomized trial asked adults with abdominal obesity to add a whole food every day: avocado. Another asked adults with obesity to cut ultra-processed foods while restricting calories. Both trials reported higher levels of Faecalibacterium, a genus that includes Faecalibacterium prausnitzii, a bacterium often linked to butyrate production and gut health.

That convergence suggests food quality may shape the gut ecosystem in obesity independent of the usual calorie math. But the second half of the story matters just as much: neither trial showed that these microbial shifts improved insulin sensitivity, inflammation, or gut barrier function. In the avocado trial, the investigators looked for functional metagenomic differences in microbial metabolic pathways and found none. The microbiome changed. Its clinical meaning is still up for trial.

What the avocado trial actually found

In a 2025 ancillary study of the HAT randomized controlled trial by Yang et al., researchers analyzed 230 adults with abdominal obesity from a single site within a larger 1,008-participant multicenter trial. Participants were randomized to daily avocado consumption or to continue their habitual diet without avocado for 26 weeks. The microbiome substudy used shotgun metagenomics, which can identify bacteria at species level and also estimate microbial gene content.

The main taxonomic result was consistent over time. Alpha diversity increased by week 4 in the avocado group and stayed higher through week 26 after false-discovery correction. By week 26, the avocado group also showed enrichment of F. prausnitzii and Bacterium AF16_15, and the overall microbial community separated significantly from the control group on Bray-Curtis beta-diversity analysis.

That sounds like a clean win for the "feed better bugs" story. Then the paper hits the brakes. The same study's functional metagenomics analysis found no statistically significant between-group difference in metabolic pathway gene content. In other words, the trial showed a taxonomic shift without a detectable functional genomic shift. That does not prove the microbial change was meaningless, but it blocks the easy leap from "more F. prausnitzii" to "more butyrate" to "better metabolic health."

The HAT investigators also reported a post-hoc signal that participants with lower baseline Healthy Eating Index scores, defined as 52.7 or less, appeared to get a larger microbial response. That is interesting and biologically plausible. It is also exploratory, unpreregistered, and exactly the kind of subgroup result that can disappear on replication.

What the ultra-processed-food trial adds

A 2025 secondary analysis by Santos de Oliveira et al. tested a different idea. Instead of adding a whole food, the trial compared two calorie-restricted diets in adults with obesity: one that specifically targeted ultra-processed foods and a generic energy-restriction control. The microbiome analysis included 34 completers, with 18 in the ultra-processed-food restriction arm and 16 in the generic restriction arm, using 16S rRNA sequencing.

The ultra-processed-food restriction group showed significantly greater increases in Ruminococcaceae, with a p-value for interaction of 0.03, and in Faecalibacterium genus, with p=0.02, than the generic calorie-restriction group. Both groups reduced body fat and waist circumference, with p<0.01 over time. That internal comparison is the key feature of the study. Since the calorie deficit and body-composition improvements showed up in both groups, the differential Faecalibacterium increase is the best evidence here that food quality itself affected the microbiota.

Still, this was a small completer analysis, and the paper reported an awkward detail: alpha diversity increased at phylum and family levels but slightly decreased at genus level. That mismatch weakens any simple ecological summary. It may reflect technical limits of 16S sequencing, sampling noise, or a genuinely mixed biological response. With 34 completers, modest p-values warrant caution, not confetti.

Why the overlap is interesting

Taken together, the two trials point in the same direction from opposite dietary moves. Add a fiber-rich whole food. Remove a chunk of ultra-processed intake. In both cases, Faecalibacterium goes up in adults with obesity. That makes dietary quality look like a real ecological variable rather than a side effect of weight loss.

That interpretation fits broader evidence. A 2024 systematic review on the dietary inflammatory index and the gut microbiome reported consistent directional associations between better-quality diets and taxa that include Faecalibacterium. A 2025 systematic review of gut microbiota in obesity also tied microbiome patterns to dietary substrate, while stressing that the field is heterogeneous and not ready to treat any single taxon as a clinical biomarker.

The likely mechanism is not avocado magic. It is substrate. A 2022 randomized trial of oats versus rice in mildly hypercholesterolemic adults found that beta-glucan-rich oats increased short-chain-fatty-acid-producing taxa and raised fecal short-chain fatty acids. That supports a simpler explanation: when you feed the microbiome more fermentable fiber and less industrially convenient junk, some butyrate-associated microbes do what they tend to do.

Why this still falls short of clinical proof

The weak point is translation. Neither obesity trial measured the outcomes people actually care about from a health perspective: HOMA-IR, fasting insulin, C-reactive protein, interleukin-6, serum zonulin, lipopolysaccharide, or any other direct test of whether the microbial change improved metabolic or barrier function.

The nearest supporting clue comes from outside the obesity population. The 2021 MaPLE randomized trial found that a polyphenol-rich dietary pattern lowered serum zonulin, a marker of intestinal permeability, in older adults. That shows diet-quality interventions can reach a functional endpoint. It does not show that the same thing happened here, in adults with obesity, through Faecalibacterium specifically.

The strongest reason to stay skeptical comes from the avocado trial itself. Because it used shotgun metagenomics, it could do more than count taxa. It could ask whether the microbial gene-content profile shifted in a way that suggested altered metabolic capacity. It did not. If the taxonomic signal were already translating into a large functional effect, this was the place most likely to catch it.

The current evidence supports a real microbiome signal from better diet quality in obesity. It does not yet support a proven metabolic benefit from that signal.

Other explanations and competing levers

Diet is not the only way to nudge the same microbes. A 2025 systematic review and meta-analysis of exercise interventions in people with obesity and type 2 diabetes found that exercise increased alpha diversity and produced similar taxonomic shifts, including Faecalibacterium enrichment, without changing diet quality. Neither primary diet trial controlled for or measured physical activity, so some part of the observed microbial change could reflect exercise differences between groups.

There is also a useful negative result on supplements. A 2022 randomized trial in women with overweight or obesity and polycystic ovary syndrome found that an energy-restricted low-glycemic-index diet changed fecal short-chain fatty acids and improved lipid measures, but adding Lactobacillus rhamnosus did not provide extra microbiota or short-chain-fatty-acid benefit. That makes the whole "just take a probiotic and call it a day" approach look less like precision science and more like wishful shopping.

The limitations are not decorative

The avocado microbiome study drew its 230 participants from a single site of a much larger parent trial, which limits generalizability. The ultra-processed-food study had only 34 completers, which raises the odds that modest between-group findings reflect noise, imbalance, or both. The two studies also used different sequencing methods. Shotgun metagenomics in HAT could identify F. prausnitzii at species level and test pathway gene content. The ultra-processed-food study used 16S sequencing, which resolved Faecalibacterium only at genus level and could not test functional gene content in the same way. That means the overlap between studies is real but not perfectly matched.

Duration is another open question. Both interventions ran for 26 weeks or less. We do not know whether these microbial shifts persist after the intervention stops or vanish as soon as the avocados or ultra-processed-food restrictions do.

Bottom line

The cleanest reading of the evidence is modest and useful. In adults with obesity, improving dietary quality appears to shift the gut microbiota toward higher Faecalibacterium abundance, whether you get there by adding a whole food or by cutting ultra-processed foods. That signal seems at least partly independent of calorie restriction alone.

What the evidence does not show is that this microbial shift improves health in a measurable clinical sense. The most detailed trial found no significant difference in microbial metabolic pathways, and neither primary trial measured insulin sensitivity, systemic inflammation, or gut permeability. For now, "better diet quality changes the microbiome" is a defensible statement. "Better diet quality fixes your gut" is marketing.