select search filters
roundups & rapid reactions
before the headlines
Fiona fox's blog

expert reaction to study on caesareans and neurodevelopmental disorders

Research, published in JAMA Network Open, reports a link between cesareans and increased risk of neurodevelopmental disorders.


Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“This study has careful and thorough statistical analysis in it, but unfortunately its findings haven’t clarified matters much. There are several problems. The patterns of association, if any, between caesarean delivery and developmental and psychiatric disorders turn out to vary a great deal between different studies. The researchers made great efforts to find explanations for this variability, but much of it remains unexplained, and that means that the ‘average’ measures of risk that they found are difficult to interpret. The research puts together data from observational studies, where it’s always difficult to know what is causing what, and this research is no exception to that general rule. Overall, the disorders that were studied were generally uncommon, so that even if increases in risk are really caused by caesarean delivery, the absolute increases in risk are small. There may be other good reasons, in terms of the health of the offspring, to avoid unnecessary caesarean deliveries, and some of those reasons are reviewed in the introduction to the paper. But, in my view, the new findings in this research shouldn’t be a huge cause for concern, at least until more has been learned about the details of what might be going on.

“In relation to all but one of the eight different groups of neurodevelopmental and psychiatric disorders covered in this study, the researchers found fairly high or very high levels of heterogeneity between the results that they combined. (The exception was for affective psychosis and depression, where heterogeneity was low, but in this case there was no overall statistically significant evidence of an association with caesarean delivery anyway.) That means that, for seven of the eight disorders, it’s not reasonable to assume that the underlying strength of association is the same across all the studies that were pooled, with any differences stemming from random statistical variation. That’s not really surprising, since the studies were done in different populations in different parts of the world, with different caesarean rates, sometimes different diagnostic criteria for the disorders, different study designs, and so on. All this variability doesn’t mean that the results can’t be pooled, but it does make a major difference to how the pooled results can be interpreted. The researchers used random effects models for the pooling (meta-analyses), which means that they aren’t assuming the strength of association is the same everywhere, and instead are estimating a kind of average effect. So how that might apply to a specific population of mothers in a specific country under specific circumstances is hard to say. The researchers used appropriate statistical methods to examine the extent to which this variability might be related to various features of the individual studies, such as the research design, the quality of the study, how rich the country is where the study was performed, and the proportion of caesarean deliveries, and in some of these factors do account for some of the variation, but much of the variability still remains. So it’s really difficult to say what any of the results might mean within a particular population of mothers.

“Because all the studies that were pooled in this research were observational, they are subject to the issue that one can never be certain what is causing what in an observational study. The problem is that there are very likely to be other differences between the caesarean births and the vaginal births, apart from the fact that some are caesarean, in terms of factors to do with the mother, the child, or something else. Such differences are called confounders. One specific problem in a study like this is what’s called confounding by indication. Caesareans aren’t performed just at random – there’s usually a reason (an indication) in terms of the condition and health of the mother and/or baby. So it’s possible that the underlying reason for having a caesarean is the cause of any association between caesareans and the health outcomes, and not the actual process of the caesarean itself. The problem with this pooling of studies is that they will all have been subject to possible confounding, of various sorts, and although many of them would have made statistical adjustments to allow for some known confounders, this would have been done in different ways using different potential confounders in different studies. That’s another possible reason for the heterogeneity, but more importantly it’s a reason why we can’t be certain that the caesarean births are the cause of any increased risk.

“How big is the increased risk anyway? The measures given in the research paper are relative (odds ratios) and don’t tell us how big the increased risk is in absolute terms. One can make an estimate of the absolute risks from the data from cohort studies included here, together with the pooled odds ratios. For the two groups of disorders where there’s the most evidence in this review, Autism Spectrum Disorder (ASD) and Attention Deficit/Hyperactivity Disorder (ADHD), that works out as follows. For ASD, the studies indicate that somewhere round 7 babies in 1000 had an ASD diagnosis if they had a vaginal delivery, and that that rate might instead be 9 or 10 per 1000 caesarean deliveries. ADHD was a bit more common, with somewhere round 20 or 25 diagnoses per thousand vaginal deliveries, which would instead be around 25 or 30 per thousand in caesarean deliveries. There’s a lot of uncertainty about these figures, but generally we’re talking about a fairly small risk difference in disorders that are not very likely anyway. And, of course, these are averages of some very heterogeneous studies, and we don’t know to what extent the risk differences are actually caused by the birth method.”


Dr Pat O’Brien, Consultant Obstetrician and Spokesperson for the Royal College of Obstetricians and Gynaecologists:

“This systematic review and meta-analysis shows an association between caesarean birth and autism and ADHD, but a number of underlying factors which may have led to the development of these conditions were not accounted for. Therefore, the findings of this paper do not show that caesarean birth leads to autism and ADHD. As the authors of the paper suggest, more research is needed to determine the underlying cause for this association, and if and how exactly this may be related to caesarean birth.

“Women who have a caesarean birth should be reassured that it is a safe procedure. In many cases, a caesarean birth can be a lifesaving intervention, as well as the right choice for mother and baby. The vast majority of caesareans in the UK are performed for medical reasons, and should a woman request a caesarean, this should be fully supported in line with guidance from The National Institute for Health and Care Excellence.”

Supporting documentation:  

NICE guidance

RCOG supports maternal request for caesareans

RCOG information leaflet for women on choosing to have a caesarean


Prof Andrew Shennan, Professor of Obstetrics at Kings College London, said:

“The need for caesarean is often caused by problems that could influence brain function, such as a poorly functioning placenta. It is highly unlikely the caesarean delivery itself is causal in these mental health conditions, from our current understanding of brain physiology and the effects of caesarean.

“Women should not be alarmed by the need for a caesarean which is often performed to reduce risk to their baby.”


Mr Richard Kennedy, Medical Director, Birmingham & Solihull Local Maternity System, said:

“Women should be reassured that caesarean section carried out in the UK is a very safe operation. Performed for appropriate clinical reasons it can be life saving for the mother and baby. Any association with paediatric outcomes is more likely to relate to underlying maternal conditions rather than the caesarean itself.”


Dr James Findon, Lecturer in Psychology at the Institute of Psychiatry, Psychology and Neuroscience, King’s College London, said:

“This study brings together the evidence of an association between Caesarean delivery and neurodevelopmental disorders. It is important to note the results do not suggest that caesarean section causes neurodevelopmental disorders. Indeed, there is good evidence from sibling studies that there is no causal link between caesarean section and autism.

“It is possible that the association stems from a genetic or environmental factor common to both neurodevelopmental disorders and the need for caesarean delivery. Parents should be reassured that caesareans are a largely safe procedure when medically indicated.”


‘Association of Cesarean Delivery With Risk of Neurodevelopmental and Psychiatric Disorders in the Offspring’ by Tianyang Zhang et al. was published in JAMA Network Open at 4pm UK time on Wednesday 28 August. 

DOI: 10.1001/jamanetworkopen.2019.10236


Declared interests

Prof Kevin McConway: “Prof McConway is a member of the SMC Advisory Committee, but his quote above is in his capacity as a professional statistician.”

Prof Shennan: No conflicts.

Dr Findon: I disclose no conflict of interest.

No others received.

in this section

filter RoundUps by year

search by tag