Early life and adult multiple risk behaviours in the 1958 British Birth Cohort: a prospective study of associations with midlife central adiposity

BackgroundCauses of obesity are multifactorial, including modifiable risk behaviours from early and later life such as physical inactivity and diet. Such risk behaviours are known to co-occur, but effects of co-occurrence on central obesity are not well understood. In the 1958 British Birth Cohort, we aimed to establish the extent of co-occurrence of early life and adult risk behaviours and associations with waist circumference.MethodsWe used data from the ongoing 1958 cohort study of all births across Britain during 1 week in March, 1958. From prospectively collected information we assessed two risk behaviours from early life (parental body-mass index ≥25 kg/m², maternal smoking in pregnancy) and four in adulthood at 33 years and 42 years (smoking, physical inactivity activity less than once a week], low dietary fibre wholemeal bread consumed occasionally or never], heavy alcohol consumption ≥22 units a week for women, ≥36 for men]). The prevalence of combinations of risk behaviours and their associations with waist circumference at age 45 years were assessed (n=9299) using covariate (eg, education level) adjusted linear regression. Missing values were imputed with multiple imputation chained equations; maximum sample analyses in univariable models gave similar results. South East MREC (ref 01/1/44) gave ethics approval and participants gave written informed consent.FindingsIn fully adjusted models, adult dietary fibre, smoking, and alcohol consumption were not associated with waist circumference. Adults whose mothers smoked in pregnancy had a 1·68 cm larger waist than those whose mothers had not smoked in pregnancy (95% CI 1·13–2·24). Trends were observed for number of overweight or obese parents, compared with having no overweight or obese parents (larger waist of 2·49 cm 1·92–3·06] for one parent, 5·56 4·66–6·45] for two parents) and ages at which inactive, compared with being inactive at neither age (larger waist of 1·36 cm 0·82–1·89] for one age, 2·57 1·86–3·29] for two ages). The most common combination of risk behaviours was having one overweight or obese parent and inactivity at one age (9·91%, n=921), which was associated with a 3·15 cm (2·09–4·21) larger waist than individuals with no risk behaviours. Individuals with two overweight or obese parents, maternal smoking in pregnancy, and inactivity at two ages had a 11·46 cm (7·36–15·57) larger waist than individuals with none of these risk factors.InterpretationSelf-reported risk behaviours and possible reverse causation are limitations, but the large, nationwide sample and lifetime risk behaviours are study strengths. Modifiable risk behaviours for large waist circumference are prevalent in the general population and co-occur across the lifetime.FundingPublic Health Research Consortium.