BACKGROUND: The assumption that lifestyles formed early in life track into adulthood has been used to justify the targeting of health promotion programmes towards children and adolescents. The aim of the current study was to use data from the Northern Ireland Young Hearts Project to ascertain the extent of tracking, between adolescence and young adulthood, of physical activity, aerobic fitness, selected anthropometric variables, and diet. METHODS: Males (n 245) and females (n 231) were assessed at age 15 y, and again in young adulthood [mean (SD) age 22 (1.6) y]. At both timepoints, height, weight and skinfold thicknesses were measured, and physical activity and diet were assessed by questionnaire and diet history method respectively. At 15y, fitness was assessed using the 20 metre shuttle run, while at young adulthood, the PWC170 cycle ergometer test was used. For each measurement made at 15y, subjects were ranked into 'low' (L1; lowest 25%), 'medium' (M1; middle 50%) or 'high' (H1; highest 25%) categories. At young adulthood, similar categories (L2, M2, H2) were created. The extent of tracking of each variable over time was calculated using 3 x 3 matrices constructed using these two sets of categories, and summarised using kappa (kappa) statistics. RESULTS: Tracking of diet and fitness was poor (kappa </= 0.20) in both sexes, indicating substantial drift of subjects between the low, medium and high categories over time. The tracking of physical activity in males was fair (kappa 0.202), but was poor in females (kappa 0.021). In contrast, anthropometric variables such as weight, body mass index and sum of skinfolds tracked more strongly in females (kappa 0.540, kappa 0.307, kappa 0.357 respectively) than in males (kappa 0.337, kappa 0.199, kappa 0.216 respectively). CONCLUSIONS: The poor tracking of fitness and diet in both sexes, and physical activity in females, suggests that these aspects of adolescent lifestyle are unlikely to be predictive of behaviours in young adulthood. In contrast, the fair to moderate tracking of anthropometric variables, particularly in females, suggests that attempts to reduce the ever increasing incidence of overweight and obesity in adults, should probably begin in earlier life.