Objective: To investigate the rate and stability of impairments in children born preterm by assessing (1) early and school-age outcome in four developmental domains and (2) individual changes in outcome at both timepoints. Design: Prospective, longitudinal cohort study in children born in 2006–2007, <32 weeks' gestation. Follow-up at 2 and 10 years of age included standardized neurological, motor, cognitive and behavioral assessments. Children were categorized as having no, mild or moderate-severe impairment in these four domains. A composite impairment score was composed and the number of domains with impairments counted. For each child, individual outcomes at both timepoints were compared. Results: Follow-up at both time-points was available in 71/113(63%) children. At group level, there were no significant changes in the severity of impairments per domain. However, at individual level, there were less children with a mild abnormal composite score at 10 years of age (44 vs. 20%; p = 0.006), and more with a moderate-severe abnormal composite score (12 vs. 35%; p = 0.001). Especially children with normal/mild outcome at 2 years were likely to shift to other outcome categories over time. Conclusions: Children with early severe impairment are likely experiencing impairments later on, but early normal/mild abnormal outcomes should be interpreted with care, considering the large individual shifts over time. Long-term follow-up in all children born very preterm should therefore be continued to at least school-age.