Objective: To review the literature on the association between birthweight and body mass index (BMI) and obesity in later life.
Methods: Included in the review were papers appearing in Medline since 1966 and identified using the search terms obesity, body fat, waist, body constitution, birthweight and birth weight. Further papers were identified by examining bibliographies.
Results: There is good evidence that there is an association between birthweight and subsequent BMI and overweight in young adults and children, which is linear and positive in some studies and J- or U-shaped in others. The evidence is less strong for middle-aged subjects. Studies that have assessed lean body mass (LBM) and fat body mass have tended to find that birthweight is positively associated with LBM and negatively associated with relative adiposity. This suggests that the association between birthweight and BMI/overweight does not necessarily reflect increased adiposity at higher birthweights. On controlling for current body mass there is fairly consistent evidence of a negative association between birthweight and a central pattern of fat distribution as measured by central:peripheral skinfold ratios. It has been suggested that the prenatal period is a 'critical' period for the development of adiposity, but it is unclear how far associations between birthweight and subsequent body habitus are genetic in origin and how far they result from intrauterine 'programming'. Two lines of evidence would suggest that the association is predominantly genetic. Studies of monozygotic twins have found environmentally determined differences in birthweight to be unrelated to subsequent BMI, and the association between birthweight and BMI is substantially reduced on controlling for parental BMI. However, some evidence of an influence of intrauterine environment on later obesity comes from studies of subjects who were exposed in utero to the effects of diabetes, famine conditions or smoking.
Conclusions: The reasons for the positive association between birthweight and BMI remain unclear. More studies including accurate measurement of body composition are needed to assess how far this relation is accounted for by changes in fat mass or by changes in lean mass. Studies with accurate measures of parental BMI would also be useful in assessing the importance of this confounder.