[en] Objective: To evaluate the long-term safety and efficacy of etanercept in patients with rheumatoid arthritis. Methods: 549 patients entered this 5-year, open-label extension study and received etanercept 25 mg twice weekly. All patients showed inadequate responses to disease-modifying antirheumatic drugs before entry into the double-blind studies. Safety assessments were carried out at regular intervals. Primary efficacy end points were the numbers of painful and swollen joints; secondary variables included American College of Rheumatology (ACR) response rate, Disease Activity Score and acute-phase reactants. Efficacy was analysed using the last-observation-carried-forward approach. Results: Of the 549 patients enrolled in the open-label trial, 467 (85%), 414 (75%) and 371 (68%) completed 1, 2 and 3 years, respectively; 363 (66%) remained in the study at the time of this analysis. A total exposure of 1498 patient-years, including the double-blind study, was accrued. In the open-label trial, withdrawals for efficacy-related and safety-related reasons were 11% and 13%, respectively. Frequent adverse events included upper respiratory infections, flu syndrome, rash and injection-site reactions. Rates of serious infections and malignancies remained unchanged over the course of the study; there were no reports of patients with central demyelinating disease or serious blood dyscrasias. After 3 years, ACR20, ACR50 and ACR70 response rates were 78%, 51% and 27%, respectively. The Disease Activity Score score was reduced to 3.0 at 3 months and 2.6 at 3 years from 5.1. A sustained improvement was found in Health Assessment Questionnaire scores throughout the 3-year time period. Conclusion: After 3 years of treatment, etanercept showed sustained efficacy and a favourable safety profile.