Background Mixture modelling is a useful approach to identify subgroups in a population who share similar trajectories. We aimed to identify distinct body mass index (BMI) trajectories between 10 and 42 years and investigate how known early-life risk factors are related to trajectories.

Methods Sample: 9187 participants in the 1970 British Birth Cohort Study, with BMI observations between 10 and 42 years and data on birth weight, parental BMI, socioeconomic status, breast feeding and puberty. Latent growth mixture modelling in Mplus was used to model age-related BMI trajectories and test associations of risk factors with trajectory membership.

Results A three latent class model was most credible: (1) normative: 92%: started normal weight but gradually increased BMI to become overweight in adulthood; (2) childhood onset persistent obesity (COP): 4%: persistently high BMI from childhood; (3) adolescent and young adulthood onset obesity (AYAO): 4%: normal weight in childhood but had a steep ascending trajectory. Higher maternal and paternal BMI and early puberty increased the probability of being in either the COP or the AYAO classes compared with the normative class.

Conclusion Most individuals gradually increased BMI and became overweight in mid-adulthood. Only 8% demonstrated more severe BMI trajectories. Further research is needed to understand the underlying body composition changes and health risks in the COP and AYAO classes.