C-Section Babies Stay Short a Key Gut Microbe for a Full Year, Study Finds

If you have spent any time researching your own gut health, you have probably met the claim that it all starts at birth. Until now that claim leaned on relatively small studies and a lot of inference. A new paper in Nature followed 714 mother and infant pairs from pregnancy through the first birthday and put hard numbers on which early-life factors actually shape a baby's gut bacteria, how long those effects last, and which ones track with a health outcome parents research constantly: eczema.

Key takeaways

• Where a baby's bacteria come from: the mother's gut, not her vaginal or breast milk microbiome, was the major reservoir of shared bacterial strains.
• What shaped the infant gut most: delivery mode. Cesarean-delivered infants were depleted in Bacteroides from the early weeks, and the gap was still visible at 12 months.
• What the mother's gut predicted: infant eczema, with an AUC of 0.74. Lower maternal microbial diversity was associated with higher infant eczema risk.
• The caveat that matters: the eczema result is associational, comes from a single Dutch cohort, and has not yet been replicated elsewhere.

Maternal Gut Bacteria Are the Infant's Main Strain Reservoir

The infant gut is seeded mainly from the mother's own gut rather than from her vaginal or breast milk microbiome. That is the central result of a longitudinal study led by Trishla Sinha, Siobhan Brushett and Alexandra Zhernakova at the University of Groningen and UMC Groningen, published in Nature. Working in the Dutch Lifelines NEXT cohort, the team tracked individual bacterial strains (genetically distinct versions of a single species, precise enough to tell one mother's Bifidobacterium from another's) as they moved from mother to baby. The maternal gut was the major strain-sharing reservoir, while transmission from the vaginal and milk microbiomes happened only occasionally.

Sharing was also strongly front-loaded. Roughly 65% of an infant's strains were shared with the mother at 2 weeks, falling to about 25% by 12 months as the baby picked up bacteria from food, household and environment. Some inherited strains dug in hard: Bifidobacterium adolescentis, when it was shared, carried an odds ratio of 17 for persisting in the infant gut. Bifidobacterium longum strain sharing was associated with higher abundance early (beta = 0.39) and lower abundance later (beta = -0.42), a reminder that inherited does not mean permanent.

Cesarean Birth Leaves Bacteroides Depleted Through 12 Months

Delivery mode was the strongest single driver of infant gut composition in this cohort, and the clearest signal was a depletion of Bacteroides in cesarean-delivered infants through month 3 that was still detectable at 12 months. Bacteroides is one of the workhorse genera of the mature human gut, and its early arrival is normally a marker of an on-schedule microbiome. This is not a new or isolated observation: a 2019 Nature study of nearly 600 UK newborns also found disrupted transmission of maternal Bacteroides strains after cesarean birth.

Read that as a description of bacteria, not a verdict on anyone's birth. Cesareans are performed for reasons that comfortably outrank gut colonization, and with about 32.4% of US deliveries by cesarean, this pattern describes a very large and largely healthy group of children. The study measured who was living in the gut, not disease outcomes. It is also worth noting what mattered less: among the 585 vaginal deliveries, the 155 that took place at home (26.5%) showed only moderate independent effects. How a baby was born counted for far more than where.

Breastfeeding Shifts Infant Gut Chemistry Toward Acetate

Feeding mode changed what the infant microbiome does, not just who lives in it. Breastfed infants' gut communities were enriched for acetate synthesis pathways (acetate is a short-chain fatty acid that gut bacteria make from milk sugars and that the gut lining burns as fuel), while formula-fed infants showed decreased quinolinic acid degradation. That functional split echoes the TEDDY cohort's finding that breast milk was the most significant factor shaping the early gut microbiome, and it is a useful corrective to species-list thinking: two babies can host similar bacteria and still be running different chemistry.

4,526 Stool Samples Across 714 Dutch Families

The credibility here comes from density rather than from a novel method. The team collected 4,526 fecal samples from 714 mother and infant pairs spanning pregnancy to 12 months postpartum, integrated them with 474 clinical variables, and added ultra-deep sequencing of breast milk and vaginal microbiomes so that strain transmission could be traced to a specific maternal body site rather than assumed.

That density produced a clean picture of both sides. The maternal gut turned out to be remarkably steady: only subtle changes across pregnancy and the postpartum year, with alpha diversity (roughly, how many different species are present and how evenly they are spread) holding stable, and diet, infections and pre-pregnancy smoking accounting for most of what movement there was. The infant gut did the opposite. Alpha diversity climbed significantly with age, the 2-week community was dominated by Bifidobacterium, E. coli and skin and oral bacteria, and by 12 months adult-type organisms such as Ruminococcus gnavus had increased substantially.

Maternal Gut Diversity Predicted Infant Eczema at AUC 0.74

A model built from the mother's gut microbiome predicted which infants went on to develop eczema with an AUC of 0.74 (area under the curve, where 0.5 is a coin flip and 1.0 is a perfect call). Specifically, lower maternal alpha diversity was associated with higher infant eczema risk (p = 0.004), and one species, Dialister invisus, stood out as protective.

For families already managing atopic disease this is the most interesting line in the paper and also the one to hold most loosely. Eczema is the most common chronic inflammatory skin disease and affects 10 to 30 percent of children in the United States, so a maternal-side predictor would genuinely matter. But an AUC of 0.74 describes an association across a population, not a test that tells one mother anything reliable about her own baby, and nothing in this design shows that changing a mother's microbiome would change her child's skin.

What Parents and Patients Can Actually Do With This

For now the honest answer is to understand this work, not to act on it. No supplement, probiotic, seeding procedure or maternal diet was tested here as a way to change any of these outcomes, so there is no validated intervention to take away. What the study does support is a sharper question for a clinician: whether a family history of eczema is worth flagging to a pediatrician early rather than after the first flare.

It is equally worth knowing what this study does not say. It does not say that cesarean birth or formula feeding harmed these children, because it did not measure that. It measured which bacteria arrived, when, and from where.

One Dutch Cohort, Nine Milk Samples, and What Comes Next

The limitations are specific, and the authors state them. The analysis of human milk oligosaccharides (HMOs, the complex sugars in breast milk that feed Bifidobacterium) rested on just 9 samples, including 3 from non-secretor mothers, with high within-feed variability and no data on total breast milk intake, so that arm is preliminary at best. The study may also be underpowered to detect deterministic effects of early microbiome clustering on later composition, and associations with the more detailed questionnaire variables were limited.

The larger caveat is structural. This is one Dutch cohort, and diet, obstetric practice and home-birth rates in the Netherlands are not universal. The eczema prediction and the strain-persistence results in particular need replication in an independent population before anyone builds a screening tool on them. That replication, plus trials that actually manipulate the maternal or infant microbiome and then measure health outcomes, is what would turn this map into something clinically usable.

Common Questions

Does a cesarean birth permanently damage my baby's microbiome? This study found Bacteroides depletion still present at 12 months, so the effect is long-lasting within the first year, but depleted relative to vaginally born infants is not the same as damaged, and the study did not track health outcomes caused by it.

If I breastfeed after a cesarean, does that undo the difference? The study did not test that combination directly. It found that delivery mode and feeding mode acted on different aspects of the microbiome, composition versus chemistry, so they are not simply interchangeable levers.

Can I get my gut tested to predict my baby's eczema risk? No. The AUC of 0.74 came from a population-level model in a single research cohort, not from a validated clinical test, and it has not been replicated elsewhere.

Did giving birth at home make a difference? Only a moderate one. Among the 585 vaginal deliveries in the cohort, 155 were home births, and home birth showed only moderate independent effects next to the much stronger influence of delivery mode itself.

References

Trishla Sinha, Siobhan Brushett, Asier Fernández-Pato, et al. (2026). Maternal influences on infant gut microbiome and health. Nature. https://doi.org/10.1038/s41586-026-10922-9