OBJECTIVES: To evaluate the cost-effectiveness of golimumab (Gol) treatment in severe, active Ankylosing Spondylitis (AS).   METHODS: A Markov model was constructed based upon the literature, to model the progression of a cohort of AS patients treated with Gol and its comparators over a 20 year time frame. The comparators considered were adalimumab (Ada) and etanercept (Eta) and all were compared to standard care which was comprised of a combination of nonsteroidal anti-inflammatory drugs (NSAIDs), disease modifying antirheumatic drugs (DMARDs), Cox-2 inhibitors and physiotherapy. Long-term efficacy was based on regressions estimated from the Gol phase III trial (GO RAISE) and the literature. Short-term comparative efficacy was derived from a mixed treatment comparison. The outcome measure was quality-adjusted life-years (QALYs). Utilities were estimated through use of an algorithm translating BASFI and BASDAI progression to EQ-5D. Costs were based on the literature (long-term) and expert opinion (short-term). Uncertainty was explored through deterministic and probabilistic sensitivity analysis (PSA).   RESULTS: Compared to conventional therapy, the incremental cost-effectiveness ratios (ICERs) of the biologic TNF-α inhibitors Gol, Ada and Eta were £30,043, £30,187 and £30,810 respectively, with the TNF-α inhibitors having similar QALYs and costs. In the sensitivity analysis, time horizon and baseline BASFI and BASDAI scores had the biggest impact on the results. Gol was seen to provide the greatest net monetary benefit (NMB) of all the TNF-α inhibitors at all willingness to pay (WTP) thresholds up to £30,000 per QALY.   CONCLUSIONS: Gol is a highly effective and well-tolerated therapy for the treatment of patients with severe, active AS and represents a treatment option with similar cost-effectiveness versus conventional care as the TNF-α inhibitors currently approved by the National Institute for Health and Clinical Excellence (NICE) in the United Kingdom.