Birthweight, lifetime body-mass Index, and physical function in mid-adulthood: evidence from a nationwide birth cohort study

Abstract

BackgroundPhysical functioning (PF) from mid-adulthood is important for maintaining independence and is linked to better health outcomes. Studies have linked obesity to poor PF; however, they have typically been cross-sectional, limited to two timepoints, or relied on retrospectively recalled height and weight. We aimed to establish associations between physical functioning at age 50 years and birthweight and body-mass index (BMI) across the life-course; BMI gains at specific life stages; and duration of obesity. MethodsThe 1958 birth cohort includes all individuals born in 1 week, in March, 1958, across Britain. BMI was calculated (4173 males, 4501 females) using height and weight measured at school (ages 7, 11, 16 years) or in participants' homes (33, 45) or were self-reported (23, 50). Primary outcome (required to be included in this study) was PF at 50 years, assessed via postal questionnaire with the validated PF subscale of the Short Form 36 Health Survey; the lowest sex-specific tenth percentile were defined as having poor PF. Missing data were imputed via multiple imputation. Associations were examined with logistic regression, and adjusted for social class, education, and health behaviours. Ethics approval and informed consent was obtained from participants at various ages. For the 50 year survey, ethics approval was provided by the London Multi-centre Research Ethics Committee (ref 08/H0718/29). FindingsBirthweight was not associated with PF. BMI at all ages from 11 years was associated with poor PF (eg, in males, adjusted odds ratio [OR] for poor PF per SD increase in BMI was 1·13 (95% CI 1·02–1.25, p=0·018) at 11 years and 1·34 (1·22–1·47, p<0·0001) at 50 years. BMI gains from adolescence were related to poor PF (eg, for females, adjusted OR per SD increase in BMI was 1·28 [1·13–1·46, p<0·0001] at 16–23 years and 1·36 [1·11–1·65, p<0·0001] at 45–50 years. Longer duration of obesity was associated with poor PF. For example, in males adjusted OR was 2·32 (1·26–4·29, p=0·007) for childhood obesity onset and 1·50 (1·16–1·96, p=0·002) for mid-adulthood obesity onset (vs never obese, ptrend<0·001). InterpretationStudy strengths include the large nationally representative cohort followed from birth and prospective measures of BMI and PF, though PF was self-reported. BMI and BMI gains from adolescence were associated with mid-adult PF. A particularly novel finding relates to duration of obesity, with earlier obesity onset associated with increased risk of poor PF, highlighting the importance of maintaining a healthy BMI from early life to mitigate the risk of poor PF in mid-adulthood. FundingDepartment of Health Policy Research Programme through the Public Health Research Consortium.