Abstract
INTRODUCTION: Obesity in adulthood is associated with reduced physical functioning (PF) at older ages. However, mechanisms underpinning this association are not well understood. We investigated whether and the extent to which C-reactive protein (CRP) mediates the association between early-adult obesity and mid-life PF. METHODS: We used data from 8495 participants in the 1958 British birth cohort study. Body mass index (BMI), CRP and PF were measured at 33, 45 and 50y, respectively. Poor PF was defined as the lowest (sex-specific) 10% on the Short-form 36 Physical Functioning subscale. We accounted for prospectively measured confounders in early-life (e.g., social class at birth) and in mid-adulthood (e.g., 42y comorbidities). We decomposed the total effect of early-adult obesity on mid-life PF into direct and indirect (via CRP) effects, by employing a mediation analysis based on parametric g-computation. RESULTS: The estimated total effect of obesity at 33y on poor PF at 50y, expressed as an odds ratio (OR), was 2.41 (95% CI: 1.89, 3.08). The direct effect of obesity on poor PF (i.e., not operating via CRP), was 1.97 (95% CI: 1.51, 2.56), with an indirect effect of 1.23 (95% CI: 1.10, 1.37). As such, the proportion of the total effect which was mediated by the effect of obesity on CRP at 45y, was 23.27% (95% CI: 8.64%, 37.90%). CONCLUSION: Obesity in early-adulthood was associated with over twice the odds of poor PF in mid-life, with approximately 23% of the obesity effect operating via a downstream effect on CRP. As current younger generations are likely to spend greater proportions of their life course in older age and with obesity, both of which are associated with poor PF, there is an urgent need to identify mechanisms, and thus potential modifiable intermediaries, linking obesity to poor PF.