May 23, 2012

caesarean and risk of obesity

Title, Date of Publication & Journal

‘Delivery by caesarean section and risk of obesity in preschool age children: a prospective cohort study’ by Susanna Huh et al. published in Archives of Disease in Childhood on Wednesday 23 May 2012.

Claim supported by evidence?

The paper provides reasonable evidence of a link between caesarean section and obesity in children at 3 years of age.

Summary

• This is a well designed large prospective cohort study which provides reasonable evidence of a 2 fold increase in the odds of obesity (BMI >95th percentile) at age three for children born by caesarean section compared to vaginal delivery. 
• In the population from Boston in this study, this was an absolute increase from 7.5% obesity for vaginal deliveries to 15.7% for caesarean. As recognised by the authors, because the study looked at relatively affluent educated women in Boston, these rates may not be applicable to the general population. Similarly they are not likely to be applicable to the UK population.
• The results were properly adjusted to take into account maternal pre-pregnancy weight and child birth weight.
• The study adjusted for a sensible list of potential confounders and several plausible biological explanations are provided along with references of this supporting evidence. Breastfeeding, timing of introduction to solids, energy intake, TV watching etc. were all investigated by the authors and rejected as confounders.

Study Conclusions

• After consideration of likely confounders, it was found that children born by caesarean were more likely to become obese at 3 years of age but a causal biological mechanism has not been established.
• The authors’ conclusions are well constructed and the studies limitations are addressed.
• The study did not find a similar link between caesarean children being overweight (as opposed to obese, i.e. BMI 85th-<95th percentile) but this is not discussed in detail in the paper.

Strengths/Limitations

A large well-designed prospective cohort study in which a reasonable and sensible approach has been taken to potential confounders.

Authors have provided a couple of plausible explanations for the link along with supporting references.

As the authors acknowledge, there is always a risk of the results being caused by an unknown confounding factor.

Maternal glucose intolerance was almost statistically significantly associated with whether the mother had a caesarean or not but it was not included in the multivariate model. However, it is unlikely that this alone would change the effect size to a significant extent.

Glossary

None

Any specific expertise relevant to studied paper (beyond statistical)?

None

‘Before the headlines’ is a service provided to the SMC by volunteer statisticians: members of the Royal Statistical Society (RSS), Statisticians in the Pharmaceutical Industry (PSI) and experienced statisticians in academia and research.  A list of contributors, including affiliations, is available here.