Biological agents directed against tumor necrosis factor (TNF) represent therapeutic options for patients with ankylosing spondylitis with high disease activity despite use of non-steroidal anti-inflammatory drugs. To evaluate the efficacy and safety of the anti-TNF agents infliximab, etanercept, adalimumab, golimumab, and certolizumab for the treatment of ankylosing spondylitis, we performed a systematic review of randomized clinical trials on adult patients with ankylosing spondylitis using articles culled from the EMBASE, MEDLINE, Cochrane Controlled Trials Register and LILACS databases (September/2012), manual literature search, and the gray literature. Study selections and data collection were performed by two independent reviewers, with disagreements solved by a third reviewer. The following outcomes were evaluated: ASAS 20 response, disease activity, physical function, vertebral mobility, adverse events, and withdraws. The meta-analysis was performed using the Review Manager(®) 5.1 software by applying the random effects model. Eighteen studies were included in this review. No study of certolizumab was included. Patients treated with anti-TNF agents were more likely to display an ASAS 20 response after 12/14 weeks (RR 2.21; 95 % CI 1.91; 2.56) and 24 weeks (RR 2.68; 95 % CI 2.06; 3.48) compared with controls, which was also true for several other efficacy outcomes. Meta-analysis of safety outcomes and withdraws did not indicate statistically significant differences between treatment and control groups after 12 or 30 weeks. Adalimumab, infliximab, etanercept, and golimumab can effectively reduce the signs and symptoms of the axial component of ankylosing spondylitis. Safety outcomes deserve further study, especially with respect to long-term follow-ups.