Skip to content

Delivery Mode Affects Stability of Early Infant Gut Microbiota

Key findings

  • This study involved sequencing of maternal microbiome samples and samples from 40 infants delivered vaginally, 19 delivered by scheduled, pre-labor Cesarean section (CS) and 16 delivered by emergent CS following labor
  • Unexpectedly, 94% of infants delivered by pre- or post-CS had detectable Bacteroides levels in their first week of life, which usually disappeared by week two
  • Mother-to-child bacterial transmission occurred mostly in vaginally delivered infants, but the maternal source was rectal rather than vaginal
  • Delivery mode had a larger influence on infant Bacteroides colonization and persistence than maternal antibiotics or breastfeeding

Delivery mode is one of the most significant influences on the composition of the gut microbiome during the first year of life. Infants born by Cesarean section (CS) have significantly lower diversity and richness of gut microbial features than vaginally delivered infants do, including less diversity of species and lower colonization with Bacteroides species, which assist in food breakdown.

The conventional explanation for the differences holds that the infant gut is "seeded" with microbes from the mother during labor, particularly from the vaginal microbiome. However, in Cell Reports MedicineCaroline M. Mitchell, MD, MPH, a researcher with the Vincent Center for Reproductive Biology in the Department of Obstetrics & Gynecology at Massachusetts General Hospital, and colleagues show that although maternal seeding can occur, there is not a straightforward transfer of the maternal vaginal microbiota.

Study Methods

An important feature of the Mass General study was that instead of simply comparing infants born by vaginal delivery (n=40) or scheduled, pre-labor CS (n=19), it included 16 infants delivered by emergent CS following labor. Like infants delivered vaginally, the latter group was exposed to the maternal vaginal microbiome.

The researchers sequenced the microbiome using three groups of samples:

  • Maternal rectal and vaginal samples collected <24 hours before delivery
  • Daily stool samples collected from the newborns while they were at the hospital (days 0–4 of life, 2–4 samples per infant)
  • One stool sample collected at home at a median age of 17 days

A Surprising Finding

In contradiction to earlier reports, the researchers observed detectable levels of Bacteroides species during the first week of life in 94% of infants born by pre- or post-labor CS. During week two, as expected, those infants were much less likely to be colonized by Bacteroides than infants delivered vaginally.

Source of Bacteroides in CS-Born Infants

Vaginal exposure couldn't be the source of Bacteroides in infants delivered by CS, the researchers determined, because only one maternal vaginal sample had a detectable level of that genus.

In fact, among 85 species found in any vaginal sample, colonization of the infant was detected in only five dyads (two vaginal deliveries, three post-labor CS). Thus, transmission of vaginal species to infants was infrequent—and not exclusive to vaginal deliveries.

In contrast, when the team sequenced rectal samples, they found:

  • Week 1: 28% of vaginally delivered infants shared at least one species with their mothers vs. 11% of CS-born infants
  • Week 2: 39% vs. 18%
  • Most of the shared species were Bacteroides

Why Did Bacteroides Disappear?

The researchers found evidence that loss of Bacteroides by week 2 in CS-delivered infants may be due to competition within the infant's gut:

  • Streptococcus and Haemophilus were more abundant in CS-born infants in week one samples, before the disappearance of Bacteroides, than in week two samples
  • Across all samples and all time points, the relative abundances of Bacteroides and either Streptococcus or Haemophilus were inversely correlated

Foundations for Future Research

This study challenges the dogma that the difference between vaginally and CS-delivered infants in early colonization of Bacteroides is due to lack of vaginal exposure. It seems the major factor underlying the association between birth mode and infant microbial community is not entering the birth canal but rather exiting through it.

This study also suggests CS influences the colonization efficiency of Bacteroides compared with other bacteria. Thus, establishment of the initial infant gut microbiota seems to require more than topical exposure to maternal vaginal secretions.

Learn more about the Vincent Center for Reproductive Biology

Refer a patient to the Department of Obstetrics & Gynecology


Vulvar and vaginal discomfort is common associated withmenopause, but new research reveals that vulvovaginal symptoms are common both before and after menopause. Overall 40% of women surveyed reported moderate to severe vulvovaginal symptoms within the past month.


Even in the absence of maternal over- or undernutrition, the choice of diet for pregnant mice significantly affected fetal brain development and the neonatal and adult behavior of offspring.