Background Understanding participation in physical activity from a life course perspective may aid the development of more effective interventions. Studies exploring the developmental origins of health have linked a heavier birth weight and earlier motor development with increased muscle strength, endurance and fitness during adulthood, and motor coordination in youth with physical capability decades later. This study examined the early life origins of adult physical activity.

Methods Data used were from the MRC National Survey of Health and Development, a nationally representative sample of men and women followed up since birth in March 1946. Birth weight was extracted from birth records and grouped into five categories. Ages at sitting, standing and walking unaided were reported by the mother when the child was aged 2 years. Motor speed and coordination were assessed by a school physician at age 15 through finger and foot-tapping tests. Leisure-time physical activity (LTPA) was self-reported at ages 36, 43, 53 and 60–64 years and classified as inactive (no LTPA reported), moderately active (<5 times/month) or most active (≥5 times/month) at each age. The odds of participation in LTPA (vs. inactivity) by each different early life factor were examined using multinomial logistic regression. Models were adjusted for sex, birth order, father’s occupation and mother’s education during childhood and adult health status.

Results 3,309 participants had at least one available measure of LTPA. Compared with those of low birth weight (<2.51 kg), all other birth weight groups had greater odds of participating in LTPA. For example, after adjustment for all covariates, those weighing 3.01–3.50 kg at birth were twice as likely (OR = 2.00; 95% CI: 1.26–3.19) to be most active as opposed to inactive at age 36 when compared with those of low birth weight and the equivalent OR for age 53 was 1.90 (95% CI: 1.17–3.09). Higher scores on finger and foot-tapping tests were also associated with increased odds of more frequent participation in LTPA across adulthood (e.g. adjusted OR of being moderately active vs. inactive at age 60–64 per 10-unit increase in finger-tapping score = 1.04; 95% CI: 1.02–1.21). There was less evidence of association between the age at attainment of motor milestones and LTPA and any associations found were fully attenuated after adjustment for childhood socioeconomic circumstances.

Conclusion A heavier birth weight and better adolescent motor coordination were associated with more frequent participation in LTPA across adulthood. These early life factors could be considered when designing physical activity interventions.