Background Periods early in life, between birth and adolescence, could be especially important for the establishment of health behaviours. Early adversity may influence the level and types of physical activity in adulthood.

Objectives To investigate the associations between early adversity and adult physical activity, and the role of depression and education as potential mediators of these associations. Structural equation models were used to 1) identify five latent factors of early adversity: socio-economic conditions (father's unskilled job, lack of home amenities and overcrowding), parental health (parental poor heath, maternal neuroticism); family structure during childhood (death of parents, divorce), chronic illness (hospitalisation for >1 month); and social isolation; 2) model the effect of depression/anxiety (at ages 15 and 36 years) and education level (at age 26 year) on the relationships between early adversity and physical activity.

Design Prospective cohort study.

Setting England, Scotland and Wales.

Participants Approximately 3300 men and women from the MRC National Survey of Health and Development, followed up since birth in March 1946.

Main outcome measures Four physical activity types measured at age 36 years—cycling/walking, heavy gardening, sports/recreational activities, and activity during the working day—with three levels of intensity (most active, less active, inactive).

Results There was no evidence of direct paths between early adversity and adult physical activity (p>0.05). However, there was a strong indirect effect of socio-economic adversity on activity at work (0.24) and on sports/recreational activity (−0.10) via education. The significant indirect effects of adolescent emotional problem on activity at work day (−0.01) and on sports/recreational activity (−0.01) via adult depression/anxiety were of smaller magnitude. There were inverse associations between adult depression/anxiety with sports/recreational activity (−0.06) and activity at work (−0.07). Lower educational level was associated with a higher level of activity at work (0.31) and a lower level of sports/recreational activity (−0.12).

Conclusions We found that the effect of childhood socio-economic adversity on adult physical activity was entirely mediated by educational attainment. The association between adolescent depression/anxiety and activity was entirely mediated by current mental health problems. These results highlight the importance of education in reducing the adverse effect of childhood socio-economic conditions on adult physical activity. Furthermore, addressing current mental health status should be seen as a priority for policies aimed at   physical activity in adulthood.