BackgroundAnxiety has been suggested to increase among young individuals, but previous studies on longitudinal trends are inconclusive. The aim of this study was to analyze longitudinally, the changes over time of prevalence of self-reported anxiety in the Swedish population between 1980/1981 and 2004/2005, in different birth cohorts and age groups.MethodsA random sample of non-institutionalized persons aged 16–71 years was interviewed every eighth year. Self-reported anxiety was assessed using the question” Do you suffer from nervousness, uneasiness, or anxiety?” (no; yes, mild; yes, severe). Mixed models with random intercepts were used to estimate changes in rates of anxiety (mild or severe) within different age groups and birth cohorts and in males and females separately. In addition to three time-related variables – year of interview, age at the time of the interview, and year of birth –the following explanatory variables were included: education, urbanization, marital status, smoking, leisure time physical activity and body mass index.ResultsOverall prevalence of self-reported anxiety increased from 8.0 to 12.4% in males and from 17.8% to 23.6% in females, during the 25-year follow-up period. The increasing trend was found in all age groups except in the oldest age groups, and the highest increase was found in young adults 16–23 years, with more than a three-fold increase in females, and a 2.5-fold increase in males, after adjustments for covariates.ConclusionsBetween 1980/81 and 2004/05, there was an increasing prevalence of self-reported anxiety in all age groups except in the oldest, which indicates increased suffering for a large part of the population, and probably an increased burden on the health care system. Clinical efforts should focus particularly on young females (16–23 years), where the increase was particularly large; almost one third experienced anxiety at the end of the 25-year follow-up.