Premium
Caesarean delivery and the risk of offspring overweight and obesity over the life course: a systematic review and bias‐adjusted meta‐analysis
Author(s) -
Sutharsan R.,
Mannan M.,
Doi S. A.,
Mamun A. A.
Publication year - 2015
Publication title -
clinical obesity
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.64
H-Index - 12
eISSN - 1758-8111
pISSN - 1758-8103
DOI - 10.1111/cob.12114
Subject(s) - overweight , medicine , meta analysis , confounding , obesity , offspring , publication bias , relative risk , confidence interval , systematic review , demography , childhood obesity , pregnancy , medline , sociology , biology , political science , law , genetics
Summary A causal role of C aesarean delivery ( CD ) on developing overweight and obesity in the life course of offspring has been postulated. However, the true strength of this association is not clear and the potential for confounding has not been adequately addressed. A systematic review and meta‐analysis were conducted to evaluate the strength of this association, this time using a bias‐adjusted model in addition to conventional methods. Our search yielded 32 estimates from 14 publications ( n = 261 000) for meta‐analysis. The pooled analysis of seven estimates ( n = 194 463) demonstrated a trend only towards a risk increase ( RR = 1.15; 95% CI :0.94, 1.40) in overweight and obesity combined (ow+ob) due to CD in early childhood (0–5 years) and a similar trend was observed for mid‐childhood and adolescence (5–18 years). In adulthood, a moderate increase in risk for ow+ob due to CD was observed ( n = 30 200) ( RR = 1.28; 95% CI 1.02, 1.34). Results for obesity and overweight separately were stronger for obesity and demonstrated a decreasing effect across the three life stages. Conventional methods of analysis suggested less uncertainty than we report and publication bias assessment was strongly suggestive of a bias in favour of positive studies. The current analysis therefore suggests that the small effects seen with CD in this and previous meta‐analyses are probably a cumulative consequence of several biases we have outlined, including confounding effect and publication bias.