Cohort studies have demonstrated that high consumption of fast-food at first assessment is associated with higher BMIs later in childhood or adult life.12 24 If this is Tipifarnib the case, despite an apparently lower BMI with more frequent fast-food consumption in the adolescent group, the high proportion of adolescents eating fast-food frequently is a matter of concern, as fast-food

consumption has been directly linked with insulin resistance, hypertension and other health sequelae.31 Strengths and limitations The major strengths of this study were its size and multicentre structure, with 199 135 adolescents from 36 countries and 72 900 children from 17 countries surveyed. Many of the centres were from middle-income and low-income countries from which previous data on the association between fast-food consumption and BMI had not been reported. The main limitation to this study is the cross-sectional design which allows identification

of associations, but not of temporal sequence or causality. The assessment of participants’ heights and weights and their fast-food consumption was primarily undertaken by a questionnaire which raises the possibility of misclassification error, particularly with respect to the parent-reported weights of their children and self-reported weights of the adolescents, thereby influencing the calculated BMI. It is also possible that parents and adolescents may have misreported fast-food consumption. A further consideration is the interpretation of the question about the consumption of burgers/fast-foods. While a ‘burger’ is almost universally understood, a ‘fast-food’ can be interpreted in a number of ways, including foodstuffs from global restaurant

chains, smaller non-franchised food stores, street vendors and even frozen meals heated and served at home. Without knowing the exact kind of fast-food consumed, we can make no assessment of the nutritional content or energy density of the food in question. Additionally, while the questionnaire asked about the frequency of burgers/fast-food consumption, we have no indication of portion sizes, Batimastat or if it was accompanied by other items such as sugar-sweetened beverages. Finally, centres that objectively measured heights and weights received no standardised instructions for doing this. We have controlled for GNI, centre, measurement type and sex in our analysis; however, we have no data available on individual socioeconomic status or parental BMI which could potentially affect young peoples’ BMI, and nor did we control for physical activity/inactivity of the participants. Conclusions This cross-sectional study has found that one-quarter of children and half of the adolescents consume fast-food frequently or very frequently.