The aim of this study was to assess long-term golimumab therapy in rheumatoid arthritis (RA) patients who discontinued previous tumor necrosis factor-α (TNF)-inhibitor(s). Patients enrolled into this multicenter, randomized, double-blind, placebo-controlled study of active RA (≥4 tender, ≥4 swollen joints) received placebo (Group 1) or golimumab 50 mg (Group 2) or 100 mg (Group 3) injections every 4 weeks. Patients in Groups 1 and 2 with inadequate response at week 16 escaped to golimumab 50 and 100 mg, respectively. At week 24, Group 1 patients crossed-over to golimumab 50 mg, Group 2 continued golimumab 50/100 mg per escape status, and Group 3 maintained dosing. During the long-term-extension (LTE), golimumab 50 mg could be increased to 100 mg, and 100 mg could be decreased to 50 mg. Data through 5 years are reported for all patients (safety) and patients using methotrexate (efficacy, intention-to-treat (ITT) analysis with last-observation-carried-forward for missing data and non-responder imputation for unsatisfactory efficacy discontinuations). In total, 459 of 461 randomized patients received the study agent, 304 of whom were methotrexate-treated and included in efficacy analyses. Through week 256, the proportions of methotrexate-treated patients achieving American-College-of-Rheumatology (ACR) responses were 37.6% to 47.0% for ACR20, 21.4% to 35.0% for ACR50, and 7.8% to 17.0% for ACR70 response across randomized groups. Golimumab safety through week 268 was generally consistent with that at week 24 and week 160 and other anti-TNF agents. In some patients with active RA discontinuing previous TNF-antagonist therapy, golimumab safety and efficacy, assessed conservatively with ITT analyses, was confirmed through 5 years. Clinicaltrials.gov NCT00299546 . Registered 03 March 2006.