Background Evidence suggests that childhood socioeconomic position (CSEP) is associated with adult physical functioning (PF). However, literature is limited on key life-course intermediaries that explain this relationship. In the 1958 British birth cohort, we aimed to examine (1) the association between CSEP and PF in midlife, and (2) potential adult intermediaries (i.e. social, psychological, behavioural and biological pathways) that could explain the CSEP-PF relationship.

Methods CSEP recorded at birth (or at 7y) was represented by 4 categories of father’s occupational class (professional/managerial, skilled non-manual, skilled manual and partly skilled/unskilled). PF was assessed at 50y using the validated PF subscale of the SF-36 questionnaire. Participants with a score at or below the gender-specific 10th-centile were classified as having poor PF. Prospectively collected data represented each of the four potential intermediary pathways: social (42y socioeconomic position, 33y education), psychological (42y mental health), behavioural (42y smoking, drinking and physical activity) and biological (45y measured BMI, waist-hip ratio, blood pressure, lung function, HbA1c and blood lipids). Gender specific associations were examined using logistic regression, in STATA v15. Adjustments were made for prospectively recorded early-life covariates (birthweight, maternal smoking, childhood disability and presence of childhood illnesses).

Results In unadjusted analysis (n=4056 men, n=4375 women) odds ratios (ORs) for poor PF increased with decreasing CSEP. This trend was maintained after covariate adjustment e.g. in women, compared to the professional/managerial class, adjusted-OR for skilled non-manual was 1.31[0.77,2.24]; skilled manual was 1.86[1.28,2.69] and partly skilled/unskilled was 2.23[1.49,3.35]. After accounting for the social pathway, associations were markedly attenuated, particularly in women; e.g. compared to the professional/managerial class OR for skilled non-manual was 0.87[0.46,1.68]; skilled manual was 1.24[0.80,1.93] and partly skilled/unskilled was 1.30[0.79,2.14]. Behavioural factors attenuated associations to a lesser extent. Biological factors completely attenuated associations in women, whereas associations remained in men. Associations were little affected by psychological factors.

Conclusion While there was a trend between lower CSEP and higher risk of poor midlife PF in men and women, intermediary pathways differed by gender. Better adult education and employment opportunities may help alleviate the burden of low CSEP for all, however improving biomedical profiles may have greater benefits for women. Study strengths include the large nationally representative cohort followed from birth and prospective measures of CSEP, PF and covariates. However, the current examination based on a complete case analysis, disregards missing data, which will be addressed in future work.