A few weeks ago my colleagues and I published a paper in the BMJ Open on mode of delivery at birth and the metabolic syndrome in adulthood. Like many research papers, a bigger research story lies behind the 3500 words of the classic ‘scientific’ paper.

Initially we were interested in the works that were showing links between: a) mode of delivery and obesity; b) microbiota in the gut and obesity; c) mode of delivery and microbiota in the gut. These three associations have been linked into a biologically plausible hypothesis explaining the exponential rise in obesity and metabolic diseases in countries with elevated rates of caesarean births.

The research story here is that vaginal births expose babies to the vaginal flora, which initiates the bacterial colonization of their gut. This, by consequence favours the development of their immune/ inflammatory systems for their given environment. Babies who are born by caesarean section and therefore who are not exposed to the vaginal flora are likely to have a delayed/ different development of their immune/ inflammatory systems which has been linked with higher rates of obesity.

The hypothesis goes like this: Being born by caesarean section is a risk for later metabolic disorders.

So we thought to test this in a historical birth cohort study (the UK 1958 birth cohort) where we know about mode of delivery and other birth related characteristics and we have biological measures in mid-life on the metabolic syndrome. Surely, if the biological hypothesis of the microbiota colonization is true now, it was surely true then. But of course things are inevitably more complex…

In the UK in 1958 only 2.6% of births happened by caesarean section, versus around 30% nowadays. So the historical versus contemporary contexts became a huge factor to think through. If anything, the context of births and how births happened in 1958 is likely to be so different compared to today, that it is worthy of a detailed qualitative study (which someone somewhere has probably done a lengthy PhD on).

Both planned and unplanned caesareans were recorded by the midwives in the 1958 birth cohort. It seems that ‘at risk’ women who had had previous pregnancy complications were typical of the planned caesarean group. Caesareans were so rare that the medical staff may have been nervous or worried about carrying them out unless absolutely necessary. Therefore being in the rare situation of having an emergency caesarean could have been a very stressful event for the woman, the staff and the baby.

To summarize: Births happened differently back then; the few women who ended up having caesareans were probably a highly specific subgroup of the population; the environment was different too – so the strains of bacteria we were being exposed to were probably different. Suffice to say that caesareans did not happen under the same circumstances as today, in our age of the “too-posh-to-push” birth fashion.

But when we ran our analyses we observed that the individuals who were born by emergency caesareans were more likely to have the metabolic syndrome in later life. This meant that the ‘microbiota’ hypothesis couldn’t apply, since if it did no difference should be observed between types of caesarean. Here, people born via planned caesareans had the same disease risk as those with vaginal births.

So what may have been different between those born by emergency versus planned caesareans in 1958? In the end, we don’t really know enough about the context but we can speculate:

-Mothers having an emergency caesarean may have been a ‘special’ group of women. If they had all been more disadvantaged or had other socioeconomic traits in common, we would hope to eliminate that explanation by including mother’s education, overcrowding and other socioeconomic measures into the model.

-Mothers who ended up with emergency caesareans had usually had long labours and had broken their water more than 12 hours before the birth, so their exposures during labour and the birth may be different. We would expect to control for some of this in our model by adding the birth variables.

-Maybe mothers who had emergency caesareans were more likely to have had gestational diabetes which would be a risk for later metabolic disorders in the baby. We didn’t have information on gestational diabetes at the time, but did control for birthweight and gestational age which would account for the bigger babies born of diabetic mothers (macrosomia).

-It seems that the overarching difference between the two types of caesarean is the context in which they happen. It may seem overly simple, but when your caesarean is planned, it can be prepared and better ‘controlled’. When it happens as an emergency in a world where caesareans are rare – it is likely to be fraught, frightening and after a prolonged labour which is onerous for the baby and mother.

So, at the end of our research story, we had neither challenged nor proved the ‘microbiota hypothesis’. Rather, by plunging a modern day hypothesis into the past, we had thought about the whole relationship between mode of delivery and later health a bit differently. All we can do is ask whether extreme stress has a lasting impact on our health, from the earliest moments. But it is unlikely we will ever know for sure.

For me, using historical cohorts is a challenge that is hugely enriching. They usually tell us that research stories are never as straightforward as they seem.


Béatrice Bouhanick, Virginie Ehlinger, Cyrille Delpierre, Bernard Chamontin, Thierry Lang, Michelle Kelly-Irving, 2014, Mode of delivery at birth and the metabolic syndrome in midlife: the role of the birth environment in a prospective birth cohort study BMJ Open 2014;4:e005031