By The first year of life is critical for the gut microbiome. A large study published August 12, 2026 in Nature followed 714 mother-infant pairs to understand how multiple maternal and early-life factors work together to shape this process. The establishment and evolution of the infant gut microbiome have a vital role in early immune development and later health outcomes.
- Cohort: 714 mother-infant pairs in the Netherlands, with 1,587 maternal fecal samples and 2,939 infant fecal samples
- Maternal age: Average 32 years at delivery, 45.1% first-time mothers
- Infants: 92.7% born full-term, 84.8% delivered vaginally, 22.5% born at home
- Feeding: 83.8% breastfed at birth, 49.7% at 3 months, 18.4% at 12 months
- Additional data: Vaginal microbiome at birth (n=82), breast milk microbiome at 1 month (n=90), 24 HMOs measured in 586 milk samples
Maternal and Infant Microbiomes Develop Very Differently
The researchers found that maternal alpha diversity remained stable during pregnancy and after delivery, with only subtle compositional shifts (overall R2 time = 0.77%). Infant alpha diversity, by contrast, significantly increased with age (R2 time = 7.78%).
At 2 weeks old, the infant gut was largely dominated by Bifidobacterium species, Escherichia coli, and skin and oral bacteria such as Staphylococcus epidermidis and Streptococcus salivarius. After 2 weeks, skin and oral bacteria decreased significantly. By 12 months, there were marked increases in adult bacteria capable of fermenting carbohydrates, such as Ruminococcus gnavus and Faecalibacterium prausnitzii.
Clustering at 2 weeks did not accurately predict later infant communities, suggesting infant gut microbiome maturation follows a non-deterministic trajectory.
Delivery Mode Is the Strongest Early Driver
Beyond time, delivery mode emerged as the strongest biological variable associated with infant gut microbiome composition. The analysis divided the infant gut into three time-dependent factors collectively explaining 19.4% of variation.
- Infants born through C-section showed a depletion of Bacteroides until month 3, whereas differences persisted until 12 months
- Vaginal birth was associated with Bacteroides uniformis, Phocaeicola vulgatus, and Parabacteroides distasonis – previously found representative of vaginal births
- Among vaginally delivered infants, hospital delivery was linked to nominally significant reductions in B. uniformis (P=0.002) and Bacteroides xylanisolvens (P=0.006) compared to home birth
- Prolonged labour, maternal use of painkillers and anaesthetics, and longer duration of ruptured membranes were also linked to reduced Bacteroides abundance
The researchers observed that reduced Bacteroides colonization likely reflects multiple characteristics of prolonged or complicated labour rather than a single delivery factor.
Feeding Mode Shapes Function, Not Just Species
Feeding mode was associated with Factor 3, which showed an inflection point at 9 months when effects diminished after introduction of solid foods. Overall, bacterial profiles of infants formula-fed at birth closely resembled those who remained formula-fed until 12 months or were never breastfed.
Vaginally delivered and formula-fed infants had higher alpha diversity compared with C-section-delivered and breastfed infants. The study found 125 significant associations between predictors and gut-brain modules. Acetate synthesis I pathway was significantly enriched in breastfed infants, while quinolinic acid degradation was decreased in formula-fed infants.
In 277 exclusively breastfed infants, the team assessed 24 HMOs. The study found no FDR-adjusted significant associations between measured HMOs and infant gut microbial diversity, species or pathways. The authors note this may reflect limited power, high within-feed variability, and limited variation in early exposure since 83.8% were breastfed at birth.
After solid foods, after adjusting for feeding mode and caloric intake, only dairy, sweet drinks and legumes remained associated. Notably, yogurt and quark intake was positively associated with Streptococcus thermophilus (FDR = 8.14 × 10-12).
Functional Capacity Matures Rapidly
The researchers annotated carbohydrate-active enzymes (CAZymes). Maternal CAZyme profile remained stable; infant profile was highly dynamic. Bacteroides, Parabacteroides and Phocaeicola were main contributors to infant CAZyme variation.
- Vaginally delivered infants were enriched for CAZymes involved in non-starch polysaccharide metabolism (FDR=0.002), while C-section infants enriched for starch-metabolizing CAZymes (FDR=0.008)
- Breastfeeding enriched for CAZymes related to resistant oligosaccharides, mucin and human milk glycans
- Early timepoints (≤3 months) enriched for resistant oligosaccharides, starch, peptidoglycan; later timepoints (≥6 months) enriched for glycoprotein, pectin and arabinogalactan metabolism
- Glycoside hydrolase family 89 (GH89) was associated with infant weight, length and head circumference
Other Factors That Mattered
The team tested 221 predictors. After correction for delivery and feeding mode, 193 species associations remained, mostly linked to parity, stool consistency, frequency and diet. Only single associations were found for infant sleeping and no significant associations for crying time.
Maternal alpha diversity was significantly associated with delivery mode and place, gestational age, educational level, pre-pregnancy BMI and smoking, and infant eczema. Maternal diversity was lower in women who delivered in hospitals than at home (FDR=0.006), with differences already evident at week 28 of pregnancy. The authors suggest this likely reflects underlying health and lifestyle, since home birth in the Netherlands is typically limited to low-risk pregnancies.
What This Means and What It Doesn’t
What the study found: In this Dutch cohort, temporal dynamics, delivery mode and feeding mode were the primary drivers of infant gut microbiome development, with minor effects from place of delivery, parity, birth weight and maternal health.
What remains uncertain: This is an observational study and cannot prove causation. Maternal communities before birth did not accurately predict clustering at 2 weeks, nor did early clustering predict later communities. The study may have been underpowered for some effects. The authors did not test interventions to restore the microbiome.
The findings highlight how maternal influences extend beyond direct transmission and suggest careful language is needed when discussing implications for infant health.
Conclusion: The infant gut microbiome is shaped by a combination of time, how a baby is born, and how a baby is fed. This large, deeply phenotyped study provides an integrated view of those influences and a platform for future research into how early microbial development relates to health.
Dr. Samuel Kemp is a Consultant Physician at the Royal Brompton Hospital with specialist expertise in respiratory medicine, including lung cancer, COPD, and advanced interventional bronchoscopy.