Four Gut Bacteria Track With Muscle Mass in 2,476 Adults, But Not Bone

If you have been researching the gut microbiome because of thinning bones, unexplained weakness, or a parent who fell, a new paper in the journal Bone deserves a careful read, including the part that came back empty. Researchers pooled stool samples and full-body scans from 2,476 middle-aged and older adults in two of the longest-running population studies in the world. Four gut bacterial genera tracked consistently with muscle mass. Bone density tracked with nothing at all.

Key takeaways

• Four bacterial genera were associated with appendicular lean mass across both cohorts after strict multiple-testing correction: less Oscillibacter, Anaerotruncus and Eisenbergiella, and more Agathobacter, went with more muscle.
• No gut bacteria were associated with bone density or bone microarchitecture, despite the study's "musculoskeletal" framing. That null result is real, and the authors blame statistical power.
• Three extra genera and a microbial biotin-production pathway showed up only in women, hinting at sex-specific biology the study was too small to confirm.
• This is cross-sectional and observational. Nobody has shown that changing these bacteria changes muscle.

Four Gut Genera Tracked With Appendicular Lean Mass

Four bacterial genera were consistently associated with appendicular lean mass (ALM, the combined lean tissue of both arms and both legs, and the standard scan-based proxy for muscle mass) in a two-cohort study published in the journal Bone by Ruolin Li, Carolina Medina-Gomez and colleagues at Erasmus MC in Rotterdam, working with Douglas P. Kiel's group at the Marcus Institute for Aging Research at Hebrew SeniorLife. Lower abundance of Oscillibacter (β = −0.51, 95% CI −0.74 to −0.29), Anaerotruncus (β = −0.41) and Eisenbergiella (β = −0.39) went with higher lean mass, while higher abundance of Agathobacter (β = 0.40) did the same. All four cleared a deliberately harsh significance threshold of P < 1.98 × 10⁻⁴, set to account for testing 84 genera against three different body measurements.

The team then ran a sanity check against actual muscle function. In the Rotterdam cohort, where handgrip strength was measured, more Eisenbergiella went with weaker grip (β = −0.37, P = 0.01), and among Rotterdam women more Anaerotruncus went with weaker grip (β = −0.41, P = 0.02). Mass and strength pointed the same way, which is reassuring, though grip was not measured in Framingham at all.

Bone Mineral Density and Trabecular Bone Score Showed No Signal

Not a single bacterial genus was associated with femoral neck bone mineral density or with trabecular bone score (TBS, a texture measure of spine bone microarchitecture) once multiple testing was accounted for. Overall microbiome diversity behaved the same way: alpha diversity, a measure of how many different bacteria live in one person's gut, was not associated with any musculoskeletal trait, and beta diversity, which compares whole communities between people, was linked to lean mass but never to bone.

The authors' own explanation is statistical power, not biology. Their earlier two-cohort analysis of gut bacteria and bone microarchitecture in Framingham and the MrOS cohort estimated that roughly 6,000 people would be needed to detect microbiome-bone associations at the effect sizes typically observed. At 2,476, this study was not close. Read the headline accordingly: if you are researching osteoporosis specifically, this paper does not support a gut-bone link. It just fails to rule one out.

Sarcopenia, Not Osteoporosis, Is the Condition Actually in Play

The muscle finding matters most to people worried about sarcopenia, the progressive loss of muscle mass and strength that drives frailty, falls, fractures and loss of independence in older adults. ALM is not an obscure research metric: it is the mass criterion in the European consensus definition of sarcopenia, used clinically alongside grip strength. With the World Health Organization projecting that 1 in 6 people worldwide will be 60 or older by 2030, any modifiable input into muscle maintenance is worth chasing.

If you live with a condition that reshapes your gut community, such as inflammatory bowel disease, coeliac disease, chronic antibiotic exposure or long-term proton pump inhibitor use, this is the kind of finding you have probably been waiting for. Note, though, that the researchers statistically adjusted for PPI and metformin use precisely because those drugs alter the microbiome, and excluded anyone who had taken antibiotics in the previous 30 days.

Two Independent Cohorts Ran One Shared Analysis Plan

The credibility here comes from replication in populations that had almost nothing in common except the protocol. The Rotterdam Study contributed 1,249 Dutch adults (mean age 62.7) and the Framingham Heart Study contributed 1,227 Americans (mean age 55.2), each with 16S rRNA gene sequencing of stool, DXA scans, and a common pre-specified analysis plan circulated across both sites. Models were adjusted for age, sex, body composition, type 2 diabetes, smoking, diet quality and medication use, plus cohort-specific technical factors such as DNA isolation and sequencing batch in Rotterdam. Only the 84 genera and 87 predicted metabolic pathways present in both cohorts were carried into the joint analysis.

Sex-stratified models produced a second layer of findings. In women only, Anaerotruncus, Hungatella and Clostridiales bacterium DTU089 were each associated with lower lean mass, and a predicted biotin biosynthesis II pathway was associated with higher lean mass (β = 0.44 in women versus 0.05 in men, difference P = 0.007). For the three genera, however, the formal test for a sex difference was not statistically significant, which the authors attribute to low power. Treat the female-specific results as hypotheses, not conclusions.

What Gut-Muscle Research Does and Does Not Justify Doing

Nothing in this study justifies buying a probiotic. These are genus-level associations measured at a single point in time, not one of these bacteria has been shown to build muscle in a human trial, and the direction of causality is unresolved. Age-related muscle loss could be reshaping the gut rather than the reverse, since frailty changes diet, activity and gut transit.

What the evidence does support is unglamorous: adequate protein intake and regular resistance training, which the National Institute on Aging identifies as a reliable defence against age-related muscle loss. If you already track your gut health, a reasonable thing to raise with a doctor is whether your muscle mass and grip strength are being measured at all, since sarcopenia is routinely missed. The bacteria are a research lead. The strength training is the intervention.

Cross-Sectional Design and Genus-Level Resolution Limit the Conclusions

This is association evidence with several specific weaknesses. 16S sequencing resolves bacteria only to genus, and Oscillibacter alone contains more than thirty species that may work in opposite directions. The two cohorts used different sequencing regions (V3–V4 versus V4), different stool preservation methods (ambient-temperature mail versus immediate ethanol) and different reference database versions, all of which likely pushed effect estimates toward zero. Both cohorts are predominantly of European ancestry, so generalisability elsewhere is unknown. Laxative use, which alters gut transit and microbial composition, was not recorded.

Some results also cut against expectation. Anaerotruncus is a butyrate producer, generally filed under "beneficial," yet more of it went with less muscle. And on funding transparency: the authors declare no competing interests, but the Framingham trabecular bone score data acquisition was supported by a grant from Dairy Management Inc., an industry checkoff organisation. It is disclosed, and it sits on the bone measurements that produced no findings, but it is worth naming.

Shotgun Metagenomics and Larger Cohorts Come Next

The authors call for shotgun metagenomic sequencing to reach species and strain resolution, multi-cohort collaborations large enough to test the bone question properly, and animal transplantation or human intervention studies to establish whether the relationship runs from microbe to muscle. Comparable work such as the HUNT study, which linked three bacterial species to higher lean mass in 5,196 Norwegians, shows the design is achievable. Expect years, not months.

Common Questions

Does this mean gut bacteria cause muscle loss? No. The study measured bacteria and muscle mass at the same moment in time, so it cannot separate cause from consequence. Reverse causation, where declining muscle changes diet and gut transit, remains entirely plausible.

Should I take a probiotic containing these bacteria? No supplement has been tested for this purpose, and three of the four genera were associated with lower muscle mass, meaning "more bacteria" is not the goal. Neither Agathobacter nor Oscillibacter features in standard commercial probiotics, which are built around Lactobacillus and Bifidobacterium strains.

Why did bone show nothing when animal studies suggest a gut-bone link? The most likely reason is sample size. The authors estimate roughly 6,000 participants are needed to detect microbiome-bone associations, and they had 2,476. Bone remodelling is also slower than muscle turnover and may need longitudinal follow-up to show up at all.

Is Oscillibacter bad for you? Not obviously. The same genus was recently identified in Framingham participants as a major cholesterol-metabolising group linked to lower blood cholesterol, which is cardiovascularly favourable. A single bacterium can be helpful for one organ system and neutral or unhelpful for another.

References

1. Ruolin Li, Paul C. Okoro, M. Carola Zillikens, et al. (2026). The association of gut microbiome composition with musculoskeletal features in middle-aged and older adults: A two-cohort joint study. Bone. https://doi.org/10.1016/j.bone.2026.118051
2. Paul C. Okoro, Eric S. Orwoll, Curtis Huttenhower, et al. (2023). A two-cohort study on the association between the gut microbiota and bone density, microarchitecture, and strength. Frontiers in Endocrinology. https://doi.org/10.3389/fendo.2023.1237727
3. Alfonso J. Cruz-Jentoft, Gülistan Bahat, Jürgen Bauer, et al. (2019). Sarcopenia: revised European consensus on definition and diagnosis. Age and Ageing. https://doi.org/10.1093/ageing/afy169
4. How can strength training build healthier bodies as we age? National Institute on Aging, NIH. https://www.nia.nih.gov/news/how-can-strength-training-build-healthier-bodies-we-age
5. Louise Grahnemo, Maria Nethander, Eivind Coward, et al. (2023). Identification of three bacterial species associated with increased appendicular lean mass: the HUNT study. Nature Communications. https://doi.org/10.1038/s41467-023-37978-9
6. Chenhao Li, Martin Stražar, Ahmed M. T. Mohamed, et al. (2024). Gut microbiome and metabolome profiling in Framingham heart study reveals cholesterol-metabolizing bacteria. Cell. https://doi.org/10.1016/j.cell.2024.03.014
7. Ageing and health. World Health Organization fact sheet. https://www.who.int/news-room/fact-sheets/detail/ageing-and-health