The 2010 update of ASAS/EULAR recommendations for managing ankylosing spondylitis (AS) specify that continuous non-steroidal anti-inflammatory drug (NSAID) treatment should be preferred in patients with persistently active, symptomatic disease. Here, our objective was to assess whether continuous NSAID therapy improves disease control and influences radiographic progression compared to on-demand therapy. We also assessed the safety profiles of both regimens. We performed a review by searching the PubMed and Embase databases using two MeSH term combinations to compare continuous and on-demand NSAID therapy in terms of disease control, radiographic progression, and safety. The only study evaluating the impact of continuous NSAID therapy on disease control showed no significant difference with on-demand therapy. In four studies, continuous treatment was associated with slower radiographic progression, as assessed in three studies using the modified Stoke Ankylosing Spondylitis Spinal Score (m-SASSS). Three studies compared the safety of continuous and on-demand celecoxib, two in osteoarthritis and one in AS, and found no significant differences regarding the usual side effects of Cox-2 inhibitors. Several studies showed slower radiographic progression with continuous NSAID therapy in AS. No studies demonstrated superiority of continuous NSAID therapy regarding symptom control. Continuous NSAID therapy (at least with Cox-2 inhibitors) does not modify safety compared to on-demand therapy.