Initially, 1,492 patients with ankylosing spondylitis (AS) were assessed by a new disease activity index. This addressed the degree of pain, severity and overall disease activity (scale 3-26). As expected, there was a normal distribution with a median of 12. Sixty-five patients, selected from the extremes of the disease activity scale, (< 5, > 22), were prospectively followed over a 2-year period: 30 with high activity mean score 24.1 (SD 1.2) and 35 with low mean score 3.4 (SD 0.5). At followup, subjects were assessed with this new index and a validated functional index. Although at followup the high activity mean score was significantly reduced to 21.1 (p < 0.001) and the low activity mean score was significantly increased to 6.1 (p = 0.002), the majority of the patients had remained in their original quartiles i.e., 63 and 77%, respectively. Disease status at followup was independent of disease duration; i.e., high activity group mean 27.1 (SD 5.3) and remission group 26.9 (SD 13.77) years, respectively (NS). When the 2 disease indices were compared, an excellent correlation existed: r = 0.788, p < 0.001. Our data suggest (1) < 1% of patients with AS who present to a rheumatologist enter longterm remission ("burn out"). (2) Some 20% of patients in remission will develop active disease 2 years later. (3) The prognosis over 2 years for those with active disease is poor.