Coffee, vegetables and exercise share a surprising connection in your gut

What connects your morning coffee, fermented foods, exercise habits, and time spent sitting? A large international microbiome investigation examines how everyday choices correspond with gut bacteria, inflammation, and chronic disease, uncovering unexpected differences between familiar foods and lifestyle behaviors.

Want to read later? Download your clean, ad-free, print-friendly PDF copy by clicking here.

Background

The World Health Organization (WHO) estimates that more than two-thirds of worldwide mortality is attributable to chronic diseases. Lifestyle factors such as diet, alcohol and tobacco use, physical inactivity, and prolonged sedentary behavior contribute to chronic disease risk. The gut microbiome has recently been associated with lifestyle-related diseases and health conditions. Most studies have examined lifestyle factors and disease risk separately, making it difficult to understand the microbiome’s mediating role. Large population-based studies that integrate clinical outcomes with microbial data are needed to clarify these relationships and identify shared and disease-specific microbial patterns.

About the study

The study evaluated fecal shotgun metagenomic data from 6,803 adult participants in the Japanese 4D (Disease, Drug, Diet, and Daily Life) research project, aged 16–95 years. The analysis examined 28 lifestyle-related factors, including eating habits, alcohol use, tobacco smoking, physical activity, and sedentary behavior. Outcomes included general health status and 15 chronic diseases, including inflammatory bowel disease (IBD), irritable bowel syndrome (IBS), type 2 diabetes (T2D), cardiovascular conditions, and respiratory disorders. Most disease diagnoses were physician-confirmed, but IBS was assessed using self-reported Rome IV questionnaires. The 2,119 participants classified as generally healthy had no comorbidities, previous chronic disease or surgery, medication use, or specified chronic viral infections.

The researchers characterized microbial composition using bacterial species profiles, while microbial functional potential was inferred from genes annotated using Kyoto Encyclopedia of Genes and Genomes (KEGG) orthologs and modules. Fecal short-chain fatty acids (SCFAs) were measured in 1,713 participants, and plasma inflammatory cytokines and chemokines were assessed in 1,740. Lifestyle exposures were self-reported, with dietary questions covering the preceding month. Prolonged sitting was defined as 9.5 hours or more per day. Mediation analyses evaluated bacterial species and functional microbial features as potential intermediaries between lifestyle exposures and clinical outcomes. Models were adjusted for age, sex, body mass index (BMI), and medication use. Researchers also tested reverse mediation to reduce the likelihood of reverse causation, but this approach could not establish causality.

The longitudinal investigation examined changes in lifestyle-associated microbiome patterns in 402 participants using 804 samples collected over a median period of two years. Researchers developed Microbiome-based Assessment of Disease and hEalth (MADE) scores from bacterial species profiles using statistical prediction models. They evaluated the scores in the Japanese cohort and tested them across 13 independent external cohorts comprising 3,095 individuals.

Study results

The analysis identified two opposing groups of microbial features statistically linking lifestyle factors with general health and chronic diseases. The researchers identified numerous significant mediation relationships involving bacterial species and microbial functional genes. Plant-based diets, fermented foods, coffee consumption, and physical activity predominantly demonstrated health-oriented mediation patterns. By comparison, refined carbohydrate consumption, smoking, and prolonged sitting were primarily associated with disease-oriented microbial pathways. Digestive, metabolic, psychiatric, and immune-related conditions exhibited relatively large numbers of microbial mediators.

Health-oriented mediators were positively correlated with fecal SCFAs, especially butyrate, and negatively correlated with inflammatory markers, such as tumor necrosis factor (TNF-α), interleukin-8 (IL-8), and interferon gamma (IFN-γ). Disease-oriented mediators generally showed weaker associations with SCFAs and stronger positive associations with inflammatory cytokines. Oral-derived bacteria represented 26.3% of disease-oriented mediators, compared with none of the health-oriented mediators. Pathobionts, microorganisms with the potential to cause disease under certain conditions, accounted for 14.3% of disease-oriented mediators and were absent from the health-oriented group.

Dietary analyses revealed specific microbial functions associated with different food groups. Vegetables, fruits, and mushrooms were linked to predicted microbial pathways for breaking down plant-derived polysaccharides, including pectin, hemicellulose, and cellulose. Several butyrate-producing bacteria, including Faecalibacterium prausnitzii, Anaerostipes hadrus, and Fusicatenibacter saccharivorans, were identified as health-oriented mediators. Coffee consumption was associated with similar health-oriented microbial patterns involving butyrate-producing bacteria. Lawsonibacter asaccharolyticus emerged as a potential microbial mediator linking coffee intake with a lower likelihood of IBD.

Dairy consumption showed more complex associations. Milk and yogurt intake was associated with Bifidobacterium species and genes involved in lactose metabolism, but individual species showed different mediation patterns across diseases. For IBD, T2D, and depression, disease-oriented mediation was more prominent than health-oriented mediation in associations with dairy consumption. The researchers also identified microbial genes involved in downstream lactate and acetate production as disease-oriented mediators.

The consumption of refined carbohydrates, including rice, bread, noodles and sodas, was primarily associated with disease-oriented microbial features, including different Streptococcus species found in saliva, Lactobacillus species, and Klebsiella pneumoniae, Escherichia coli and Morganella morganii. The microbial functions associated with the disease-related features of the gut microbiota included starch and sugar utilization, lactate production, biofilm formation, and Gram-negative cell envelope biosynthesis.

Alcohol consumption demonstrated bidirectional microbial mediation involving host-derived glycan utilization. Disease-oriented mediation predominated across several conditions, including IBS, dyslipidemia, hypertension, hyperuricemia, and depression. Physical activity was associated with butyrate-producing bacteria, whereas sedentary behavior was linked to disease-associated species, including E. coli and Ruminococcus gnavus. Longitudinal analyses supported the consistency of several lifestyle-associated microbial signatures but did not establish whether these changes influenced later disease outcomes.

The MADE scores distinguished participants meeting the study's strict general-health criteria from other participants, while disease-specific scores differentiated individuals with and without particular chronic conditions in the Japanese cohort. External validation across 13 independent cohorts demonstrated consistent disease-discriminating performance in 11 cohorts. Performance differed by disease, with limited discrimination for IBS. Combining MADE scores with age, sex, BMI, and lifestyle factors improved discrimination for nine of the 16 outcomes, but clinical usefulness remains unproven. The scores also captured distinct microbial patterns associated with individual diseases.

Conclusion

The findings linked plant-based foods, coffee, and physical activity with health-oriented microbial profiles, while refined carbohydrates and sedentary behavior were linked to disease-oriented profiles. The MADE scores suggested that microbial profiles may help distinguish health from disease across populations, with performance varying by condition. Experimental studies and clinical lifestyle intervention trials are needed to establish causality, clarify underlying mechanisms, and determine their clinical applicability. Other limitations include self-reported dietary exposures, possible unmeasured confounding, and uncertainty about how well the findings apply beyond Japan. The cost, accessibility, and clinical value of microbiome testing also require further study.

Source:

Journal reference:

• Nagata, N., Wei, R., Kaneko, G., Nishijima, S., Aoki, R., Kojima, Y., Tamura, Y., Sugiyama, M., Hamabata, T., Nakagawa, T., Ueki, K., Ohsugi, M., Nakanishi, Y., Ito, A., Asami, S., Suzuki, R., Kuhn, M., Bork, P., Hattori, M., Ohno, H., & Uemura, N. (2026). Opposing microbiome-mediated axes link modifiable lifestyles to health and chronic diseases. Nature Medicine. DOI: 10.1038/s41591-026-04744-5, https://www.nature.com/articles/s41591-026-04744-5