Background Ultra-processed foods (UPFs) are increasingly consumed in Western diets and have been associated with obesity, cardiometabolic disease, and colorectal cancer. However, the gut microbial and metabolic pathways underlying these associations remain incompletely understood. We examined associations of habitual UPF intake with the gut microbiome and fecal metabolome in a large U.S. cohort of women.

Methods In the Microbiome Among Nurses (Micro-N) study, we performed shotgun metagenomic sequencing and high-throughput liquid chromatography-tandem mass spectrometry (LC-MS) metabolomic profiling of fecal samples collected in 2019 from participants in the Nurses’ Health Study II. Long-term diet was assessed using validated food-frequency questionnaires administered every four years and cumulatively averaged from 1991 to 2019. UPFs were classified using the NOVA system (total UPFs and major subgroups). Associations with Shannon diversity, Bray-Curtis community composition (PERMANOVA), species-level abundance, microbial functional pathways, and fecal metabolites were evaluated using multivariable models adjusting for demographic, lifestyle, dietary, medication, and stool-related factors, with correction for multiple testing.

Results Among 2,288 participants with metagenomic data (mean ± SD age 64.7 ± 4.4 years; mean UPF intake 6.3 ± 2.0 servings/day), UPF intake was not associated with alpha diversity (β = -0.007, p = 0.13) and explained only 0.4% of between-person variation in overall microbiome composition (p < 0.001). Species-level analyses identified targeted associations with modest effect sizes. Higher UPF intake was associated with increased abundance of Akkermansia muciniphila, Parabacteroides distasonis, and an unclassified species bin (GGB9468 SGB14861), and lower abundance of several butyrate-producing taxa, including Faecalibacterium-related species, Lachnospiraceae members, and Fusicatenibacter. Associations differed across UPF subgroups: dairy and grain products were positively associated with Streptococcus thermophilus and other lactic acid bacteria, whereas beverages, mixed dishes, and meat-based UPFs were more strongly associated with depletion of beneficial anaerobes. Among 1,861 participants with fecal metabolomics, UPF intake was associated with broader metabolic perturbations than observed for the microbiome. Higher UPF intake correlated with coordinated alterations in lipid, amino acid, and bile acid pathways, including enrichment of sphingolipid- and glycerolipid-related metabolites and lower levels of compounds related to whole-food intakes. Signals indicative of food additive exposures were limited.

Conclusion: Habitual UPF intake was associated with selective species-level microbial shifts rather than broad restructuring of the gut microbiome, alongside more pronounced alterations in the fecal metabolome. These findings suggest that metabolic dysregulation, together with depletion of butyrate-producing taxa, may represent key pathways linking UPFs to chronic gastrointestinal and metabolic disease risk.