Methods appropriate for studying the relationship of breast-feeding to obesity.

A vast literature examines the association of breast-feeding with body composition and risk of overweight and obesity in childhood and adulthood. Several recent systematic reviews, including one by the World Health Organization, concluded that there is a small protective effect of breast-feeding against overweight in later life. Nearly all studies covered by these reviews used observational study designs that limit causal inferences. Methodological strengths, weaknesses, and main results of the epidemiologic studies that have drawn conclusions about the relation between infant feeding and overweight in childhood and adulthood are briefly reviewed to provide a methodological perspective for the subsequent presentations in this symposium. The focus is on the role of recall and selection bias, appropriate representation of exposures, inadequate control for confounding, and the utility of alternative study designs that may circumvent some of the problems.