Background The global childhood obesity epidemic poses great challenges in understanding how growth trajectories are determined from early life and across generations. It is increasingly clear that factors causing intra-uterine growth retardation and macrosomia may place children on adverse growth trajectories that lead to overweight and obesity. The Lifeways study was established to study the influence of familial and lifestyle factors including diet on long-term outcomes for children.

Methods Of 1082 mother-child pairs recruited antenatally in 2001–3, after exclusion of twins, data were available for 981 children with at least one body mass index (BMI) measurement available at one of 3 waves of follow-up to age of 9 years old. Infants were categorised into low (<2.5 kg, N = 48), normal (2.5–4 kg, N = 761) and high (≥4 kg, N = 172) birth weight categories. We examined a variety of maternal and child characteristics at univariate level in relation to the trajectory groups. A linear mixed model was fitted with wave and birth weight category as factors with a fixed intercept and allowing correlated residuals over time with an unstructured covariance matrix (selected by BIC) and with an interaction term to allow for different trajectories over time, both adjusted and unadjusted for gestational age.

Results There was a positive association for maternal age (p < 0.001), pre-pregnancy BMI (p < 0.001), and an inverse association for maternal smoking status (p < 0.001) with these trajectory groups. Maternal fat intake (% Energy Intake) was higher and carbohydrate intake (% Energy Intake) was lower during pregnancy in the low birth-weight group, p = 0.004 and p = 0.008 respectively. There was a statistically significant interaction (p < 0.001), indicating a different pattern of change in trajectory over time in the three groups, whether adjusted for gestational age or not. The low birth weight group showed a rapid catch-up response by age 5, before overtaking the normal weight group by age 9. The high birth weight group continued to track higher than the normal weight group over time.

Conclusion These novel prospective cohort data confirm the adverse influence of birth-weight outcomes on subsequent growth trajectories in this cohort of children, in turn associated with maternal characteristics during gestation and pregnancy.