Background. The aim was to identify risk indicators from preadolescence (age period 10-12) that significantly predict unfavorable deviations from normal anxiety development throughout adolescence (age period 10-17 years). Methods. Anxiety symptoms were assessed in a community sample of 2,220 boys and girls at three time-points across a 5-year interval. Risk indicators were measured at baseline and include indicators from the child, family, and peer domain. Associations with anxiety were measured with multilevel growth curve analyses. Results. A stable difference in anxiety over adolescence was found between high and low levels of a range of child factors (frustration, effortful control), family factors (emotional warmth received from parents, lifetime parental internalizing problems), and peer factor (victims of bullying) (P<.001). In contrast, the difference in anxiety between high and low levels of factors, such as self-competence, unfavorable parenting styles, and bully victims, decreased over adolescence (P<.001). For other family factors, associations were weaker (.05<P<.001). Associations with parental education and family composition were not significant. Adjustment for concurrent depressive symptoms attenuated the associations, but those that were significant at P<.001 remained to be so. Specificity for anxiety subtypes (generalized anxiety, separation anxiety, social phobia, panic, and obsessive-compulsive symptoms) was reported for each association. Conclusions. Several child, family, and peer factors measured in preadolescence were risk indicators of high levels of anxiety symptoms throughout adolescence. Some factors (such as rejective parenting) were vulnerability indicators for anxiety in early adolescence only, whereas other factors (such as peer victimization) were indicators of long-term elevated anxiety levels.