Page 33 - Read Online

Horwell et al. Microbiome Res Rep 2025;4:1 https://dx.doi.org/10.20517/mrr.2024.32 Page 11 of 20

Addressing the first point, definitively showing that exposure to environment A leads directly to
microbiomic outcome B, might appear pedantic given the voluminous observational data. However,
without knowing specifically how species rarely found in the mostly aerobic outdoor environment colonise
the infant GIT, such as Akkermansia Mucinophilia, devising future interventions to aid the developing
microbiome is going to be problematic and imprecise. This can be seen contemporarily in novel vaginal
seeding techniques used after CS, which have so far failed to convincingly replicate the microbiome of
children born via VD. The entero-mammary pathway of translocating bacteria from the maternal colon to
breast milk appears to be a fruitful avenue for future research. Furthermore, while the widespread use of 16S
genomic profiling has transformed our understanding and ability to measure the microbiome in a cost-
effective way, it does not provide the entire picture. For example, 16S RNA is not sufficient for fungal or
viral profiling, and both methods fail to give clarity on strain-specific population changes. As an example, it
has been estimated that 75% of our microbiome, at the bacterial strain level, has yet to be characterised,
colloquially known as dark matter [190,191] . Furthermore, metagenomics is not specific to the metabolome of

the microbiome, which provides the mainstay of bioactivity. The fine details of this are beyond the scope of
[192]
this paper, but a recent review covers the topic well . New technologies are showing promise to give
researchers the tools to track specific strains that might produce clear evidence of transfer from one
environment to the GIT, and the variation in metabolic output from different strains (e.g., WISH-tagging);
however, economic viability has limited widespread use, and they have not yet been employed in this
specific field [192,193] . The use of genetically tagged bacteria has been used for several decades as a cost-effective
way of tracking microbiome dynamics in adults across different disciplines [194,195] . To date, and to the best of
our knowledge, there has only been one study using this technique on the infant microbiome . More
[25]
studies using these techniques are needed to clarify with precision modes of microbial transmission.
Addressing the second point, observational studies that demonstrate differences in health outcomes all
suffer from well-established biases and the inability to control for unknown confounders. As already
highlighted, these studies can, therefore, only infer causality . Due to the nature of the subject matter (the
[196]
paediatric population), it is not surprising that randomised trials have not been performed. Ethical approval
for a randomised trial into CS vs. VD would be near impossible to justify, let alone finding volunteers for
recruitment. Animal models have been, therefore, crucial for deepening our understanding of how
differences in the microbiome affect metabolic and immunogenic health. Despite the critiques of
observational data listed above, pragmatism and the highly consistent and reproducible evidence of
associated long-term health benefits for the infant should, and do, guide public health organisations as well
as professional bodies to advocate for VD over CS, and for exclusive breastfeeding for at least six months. It
should be noted that there are further benefits to breastfeeding that have not been discussed, such as the
nutrition provided by breast milk in nations that have food insecurity or unsanitary water conditions ; the
[197]
protective effects of breastfeeding on the mother’s risk of breast cancer, etc. , but these are beyond the
[198]
scope of this paper.

The critical window of opportunity
It is clear that the infant microbiome is influenced by numerous variables in a highly orchestrated and
dynamic ecosystem that follows a similar pattern across different global populations. CS is associated with
an altered microbiome compared to VD, although the mechanism of this difference is not fully established.
The largest single effector of the microbiome is breastfeeding, with very clear evidence that once this ceases,
the microbiome of the infant quickly starts to change to that of an adult. During this time, environmental
stimuli also become more apparent (e.g., diet, living in a rural environment, going to day care, etc.) and the
microbiomic differences from variables seen earlier in life, such as mode of birth or breast feeding, start to
converge. At around the age of 3-5 years, the microbiome resembles that of an adult, and populations
stabilise.