ObjectiveThe existence of catch-up lung function growth and its predictors is uncertain. We aimed to identify lung function trajectories and their predictors in a population-based birth cohort.MethodsWe applied group-based trajectory modelling to z-scores of forced expiratory volume in 1 second (zFEV₁) and z-scores of forced vital capacity (zFVC) from 1151 children assessed at around 4, 7, 9, 10, 11, 14 and 18 years. Multinomial logistic regression models were used to test whether potential prenatal and postnatal predictors were associated with lung function trajectories.ResultsWe identified four lung function trajectories: a low (19% and 19% of the sample for zFEV₁and zFVC, respectively), normal (62% and 63%), and high trajectory (16% and 13%) running in parallel, and a catch-up trajectory (2% and 5%) with catch-up occurring between 4 and 10 years. Fewer child allergic diseases and higher body mass index z-score (zBMI) at 4 years were associated with the high and normal compared with the low trajectories, both for zFEV₁and zFVC. Increased children’s physical activity during early childhood and higher zBMI at 4 years were associated with the catch-up compared with the low zFEV₁trajectory (relative risk ratios: 1.59 per physical activity category (1.03–2.46) and 1.47 per zBMI (0.97–2.23), respectively). No predictors were identified for zFVC catch-up growth.ConclusionWe found three parallel-running and one catch-up zFEV₁and zFVC trajectories, and identified physical activity and body mass at 4 years as predictors of zFEV₁but not zFVC catch-up growth.