The study, published in the journal Metabolomics, followed 72 infants with an elevated hereditary risk of allergic disease through their entire first year. All of the children were exclusively breastfed for at least their first 16 weeks, a design choice that allowed the researchers to isolate the effects of later feeding changes without the confounding influence of early formula exposure. Fecal samples were collected at three time points, and parents kept detailed records of allergy manifestations, delivery mode, and the timing of complementary feeding. The work was part of the larger TEMPO trial, a randomized, double-blind, controlled, multi-country study registered with ClinicalTrials.gov.

What sets this investigation apart from much of the existing microbiome literature is its methodological reach. Rather than relying solely on DNA sequencing, which reveals which microbes are present but says little about what they are actually doing, the team deployed a two-pronged chemical and cellular arsenal. Fecal samples were analyzed using a combination of gas chromatography with flame ionization detection and liquid chromatography-mass spectrometry, allowing the researchers to profile hundreds of metabolites produced by both the infants themselves and their resident microbes — aromatic amino acid derivatives, B vitamins, bile acids, short-chain and long-chain fatty acids, and lactic acid among them. In parallel, the researchers used fluorescence in situ hybridization, or FISH, to directly count bacterial cells belonging to two keystone groups: Bifidobacterium species, the darling microbes of the healthy breastfed infant gut, and the Eubacterium rectale/Clostridium coccoides group, a collection of Lachnospiraceae species that dominates the adult microbiome.

The decision to enumerate bacteria by microscopy rather than relative sequencing abundance proved consequential. Sequencing-based surveys can distort temporal comparisons because they express each taxon as a fraction of a shifting whole, making it difficult to know whether a species truly declined or simply yielded numerical ground to a flourishing neighbor. FISH, by contrast, provides absolute cell counts, giving the researchers a firm quantitative footing as they tracked microbial dynamics across infancy.

The results paint a picture of a gut metabolome in constant, age-driven flux. The strongest signal in the entire dataset was time itself: as infants matured from three months to twelve months, their fecal chemistry shifted dramatically, particularly in the metabolism of aromatic amino acids, the composition of the bile acid pool, the availability of B vitamins, and the balance between short-chain and long-chain fatty acids. This chemical maturation mirrors the known ecological succession of the infant gut, in which a community dominated by Bifidobacterium species gradually gives way to a more diverse, adult-like assemblage enriched in obligate anaerobes. Each successive wave of colonizers brings its own enzymatic toolkit, and the metabolome records that turnover in exquisite detail.

But age was not the only force at work. The introduction of complementary foods — the baby’s first tastes of solid food — and the cessation of breastfeeding both left statistically significant fingerprints on the fecal metabolome, confirming that feeding transitions act as genuine perturbations to gut chemistry, not merely as background context. Human breast milk delivers not only a balanced package of macronutrients and bioactive compounds but also human milk oligosaccharides, complex sugars that the infant cannot digest but that certain bifidobacteria, such as Bifidobacterium breve, B. bifidum, and B. longum, devour with specialized enzymatic machinery. When breastfeeding ends and solid foods arrive, the fuel supply for these microbes changes fundamentally, and the downstream metabolites change with it.

Perhaps the most striking finding concerned birth itself. Infants delivered by Cesarean section carried a fecal metabolome measurably different from that of their vaginally delivered peers, and the differences persisted until six months of age. This aligns with a growing body of evidence linking C-section delivery to early-life dysbiosis and to elevated risks of atopic dermatitis and other immune-mediated conditions later in childhood. During vaginal delivery, an infant is exposed to the maternal vaginal and fecal microbiota, an inoculation that C-section infants largely miss. The new metabolomic data show that this missing exposure is not simply a matter of which bacteria are present — it propagates into the chemical output of the entire gut ecosystem, altering the very molecules with which the developing immune system interacts.

The allergy connection emerged from the comparison between infants who developed allergic manifestations during the first year and those who did not. Before the age of 16 weeks — the window covered by the first sampling time point — the future allergy group harbored significantly lower numbers of Bifidobacterium species and, simultaneously, significantly higher levels of polyunsaturated fatty acids in their feces. Bifidobacteria have long been regarded as crucial architects of immune education in infancy; they metabolize human milk oligosaccharides into short-chain fatty acids and other compounds that help calm inflammatory signaling and promote the development of regulatory T cells, the immune system’s own peacekeepers. A deficit of these microbes during the earliest months, when immune tolerance is being established, could plausibly tilt the balance toward the allergic phenotype.

The elevated polyunsaturated fatty acid signal is particularly intriguing because it hints at altered lipid handling in the pre-allergic gut. Long-chain fatty acids are signaling molecules as well as nutrients: they bind to cell-surface receptors on immune cells and epithelial cells, influencing barrier function and inflammatory tone. Whether the higher fecal levels in future allergy cases reflect differences in the infants’ own metabolism, differences in dietary exposure through breast milk, or differences in microbial lipid transformation is a question the study raises but cannot yet fully resolve. The researchers’ metabolomic approach, which captured both host-derived and microbe-derived metabolites, was designed precisely to make such distinctions addressable in future work.

The clinical context of the findings is the so-called “atopic march” — the well-documented tendency of allergic disease to progress in a temporal sequence, beginning with atopic dermatitis and food allergy in infancy and often advancing to allergic asthma and allergic rhinitis in childhood. If the seeds of that march are sown in the first months of life, as this study suggests, then the infant gut metabolome becomes not just a scientific curiosity but a potential early-warning system. Distinct metabolic signatures measurable at three months of age could, in principle, help identify infants most likely to benefit from preventive interventions, whether probiotic supplementation with targeted Bifidobacterium strains, tailored dietary guidance for breastfeeding mothers, or modified timing of complementary feeding.

The study’s design carries important caveats. All 72 infants were exclusively breastfed for at least 16 weeks, which enhances internal validity but raises questions about how the findings generalize to formula-fed infants, whose gut microbiomes are known to differ substantially. The sample size, while respectable for such an intensive longitudinal metabolomic study, is modest, and the participants were all infants at elevated allergy risk, drawn from a parent trial that enrolled 855 healthy term infants across multiple countries. The researchers also applied rigorous statistical hygiene, normalizing metabolite signals to fecal dry weight, excluding compounds detected below five times the blank signal, subjecting metabolites with missing data to formal group-bias testing, and validating their analytical workflows with quality-control software.

Even with those caveats, the work marks a meaningful shift in how scientists approach the infant microbiome. For years, the field has been dominated by compositional studies — catalogs of who lives in the gut. The new findings demonstrate the value of asking what the residents are producing, and how that chemical production is sculpted by the circumstances of birth, feeding, and development. The gut microbiome influences host physiology largely through its metabolic output, and it is that output, more than the microbial roster itself, that interfaces with the immune system. By capturing the metabolome longitudinally, at three time points across the first year, and pairing it with absolute bacterial counts, the researchers have produced something closer to a functional movie of infant gut development than the static snapshots that sequencing alone can offer.

For parents, the takeaways are appropriately measured. Breastfeeding remains the recommended infant nutrition for allergy prevention, and the study adds metabolomic texture to that recommendation by showing how profoundly feeding transitions reshape gut chemistry. The persistence of delivery-mode effects to six months reinforces what pediatric microbiome researchers have long argued: that Cesarean birth leaves a measurable ecological legacy in the infant gut, one that may intersect with immune development. And the finding that allergy-prone infants diverge chemically before 16 weeks of age offers hope that the window for prevention opens far earlier than symptoms do — a window that future studies, armed with metabolomics and absolute microbial quantification, are now well positioned to explore.

Savova, M. V., Zhu, P., Kindt, A., the TEMPO study team, Wopereis, H., Belzer, C., Harms, A. C., & Hankemeier, T. (2026). Fecal metabolome alterations in infants at risk of developing allergies during the first year of life. Metabolomics, 22(1), Article 112. https://doi.org/10.1007/s11306-026-02478-6