• radiographic change (mean) of the spine after 2 years in the per-protocol population. Radiographs will be collected and centrally digitized. Scoring will be done by 2 readers who were blinded to treatment and sequence of the films [ Time Frame: 2 years ] [ Designated as safety issue: No ]
  

• the proportions of patients with any progression (change in the mSASSS ≥ 1) and change in the mSASSS > smallest detectable change (SDC), i.e. change in mSASSS which is greater than the measurement error. [ Time Frame: 2 years ] [ Designated as safety issue: No ]
  
• ITT analysis of radiographic change. [ Time Frame: 2 years ] [ Designated as safety issue: No ]
  
• Change in VAS back pain, BASDAI, BASFI, BASMI, CRP. [ Time Frame: 2 years ] [ Designated as safety issue: No ]
  
• event rates of serious and non-serious adverse events will be documented and compared between the two groups. [ Time Frame: 2 years ] [ Designated as safety issue: Yes ]
  

Ankylosing spondylitis (AS) is a common chronic inflammatory rheumatic disease with a prevalence of about 0.5%. First symptoms normally occur in young adulthood. Early in its course, AS is dominated by chronic pain, fatigue and morning stiffness, later on by ankylosis and loss of function. Nonsteroidal anti-inflammatory drugs (NSAID) and tumor necrosis factor (TNF) alpha blocking agents are the only drugs with proven efficacy for signs and symptoms. It is not clear, however, whether these drugs are also capable of retarding or stopping structural damage, i.e. prevention of bony ankylosis. Earlier investigations indicated that NSAIDs have, in addition to their anti-inflammatory, also an anti-osteoproliferative effect. In this study we will investigate whether treatment with 150 mg diclofenac, a non-selective NSAID, on a daily basis (continuous treatment) over 2 years is capable to slow down the development of bony ankylosis as compared to treatment with 75-150mg diclofenac as needed according to clinical symptoms (on-demand treatment). In this national multi-centre randomized trial 360 patients with symptomatic AS and indication for NSAID therapy will be enrolled in about 40 centres. The primary outcome parameter is the proportion of patients with radiographic progression in the spine after 2 years in each treatment arm. If continuous NSAID treatment results in less radiographic progression as compared to on-demand treatment, a true disease modifying effect of NSAID has to be assumed which will most likely change the place of NSAID treatment in AS.