To test the predictive ability of remission in terms of long-term disability in patients with recent-onset inflammatory polyarthritis (IP). Methods Consecutive patients with early IP, recruited between 1990 and 1994 (first cohort) and 2000 and 2004 (second cohort), were included in this study. Remission was defined as the absence of clinically detectable joint inflammation on a 51–joint count. In additional analyses, less stringent definitions of remission were used based on the 40– and 28–joint counts. Remission was assessed at 1, 2, and 3 years after inclusion. A 5-year Health Assessment Questionnaire score =1 (moderate disability) was chosen as the primary outcome measure. Results A total of 841 and 498 patients from the first and second cohorts, respectively, completed 5 years of followup. In the first cohort, patients with at least 1 episode of remission had lower odds of 5-year disability (odds ratio [OR] 0.26, 95% confidence interval [95% CI] 0.17–0.41). The number of times in remission correlated with the odds of disability, with a mean decrease in the probability of disability of ~64% for each additional time point in remission (OR 0.38, 95% CI 0.28–0.52). The time until first remission was not associated with functional disability. Remission according to less stringent criteria showed a weaker protection against future disability. Similar results were found in the second cohort. Conclusion Patients with IP achieving a state of sustained remission early are less likely to show long-term deterioration of function compared with patients who do not achieve remission. The most persistent remission under the most stringent definition of remission has the lowest probability of long-term disability.