OBJECTIVE: To investigate disease course, outcome, and predictors of outcome in an unselected population-based cohort of individuals diagnosed with juvenile chronic arthritis (JCA) followed for 17 years. METHODS: The cohort consisted of 132 incidence JCA cases identified 1984-1986 according to EULAR criteria. At 5-year followup, 129 individuals underwent joint assessment, laboratory measurements, radiographic examination, and medication and functional assessment. At 17-year followup, 86 were examined with joint assessment, laboratory measurements, medication assessment, Health Assessment Questionnaire (HAQ), Keitel functional test (KFT), and Medical Outcomes Study Short Form-36 (SF-36). RESULTS: At 17-year followup, 40% were in remission, 44% changed subgroups, median HAQ score was 0.0 (range 0.0-1.5), and median KFT was 100 (range 54-100). SF-36 scores were significantly lower compared to a reference group. Thirty-nine percent of those in remission at 5-year followup were not in remission at 17-year followup. In multivariate analyses of variables from the 17-year followup: remission was predicted by remission at 5-year followup (OR 4.8); HAQ > 0 by rheumatoid factor (RF)-positivity at 5-year followup (OR 3.6); KFT < 100 by nonremission (OR 11.3); and RF-positivity (OR 5.6) at 5-year followup; and the SF-36 physical component summary score above average of the reference group by remission at 5-year followup (OR 5.8). CONCLUSION: This longterm study of 86 individuals with JCA showed large variability of disease courses and of impaired health-related quality of life. Sixty percent were not in remission at 17-year followup. Longterm outcome was best predicted by and associated with characteristics at 5-year followup rather than those at onset.