Objective Adult obesity is associated with a variety of acute and chronic illnesses. Although high birthweight is known to predict obesity in middle age, the relationship between low birthweight and obesity is less clear. This study examines the association between birthweight and obesity in middle-aged women, and investigates whether the association is modified by other factors.

Methods The Million Women Study is a large population-based prospective cohort study of middle-aged UK women. This analysis is based on 372 542 women who reported their birthweight, current body size, and other information in a follow-up survey administered approximately 3 years after recruitment (mean age 58 years at follow-up). Logistic regression was used to estimate relative risk for being obese in adulthood (body mass index>30 kg/m²) by birthweight, both unadjusted and adjusted for reported adult height, parental heights, and a range of social and lifestyle factors, including socio-economic status, parental smoking at birth, being breast fed, reproductive history and health behaviours.

Results There was a U-shaped relationship between birthweight and adult obesity. Compared to women of intermediate birthweight (3.0–3.5 kg), the relative risk of being obese was 1.26 (95% CI 1.23 to 1.29) for women with low birthweight (<2.5 kg), and 1.33 (1.30 to 1.37) for women with high birthweight (>4.0 kg). After adjustment for height, the relative risk of being obese associated with low birthweight was attenuated (from 1.26 to 1.15), while the relative risk associated with high birthweight was increased (from 1.33 to 1.44). Other health and lifestyle characteristics did not substantially change the obesity-related relative risks at different birthweight.

Conclusion Birthweight has a U-shaped relationship with obesity in middle age. This relationship tends towards being J-shaped after adjustment for adult height.