Patients with active ankylosing spondylitis of at least 6 months' duration were stabilised on diclofenac 50mg, given 2, 3 or 4 times daily in a 4-week open-label run-in. Patients were then randomised to receive a diclofenac 50mg/misoprostol 200 micrograms fixed combination tablet (n = 331) or diclofenac 50mg (n = 339) for 8 weeks at the same dosage frequency as in the open-label phase. For the intent-to-treat cohorts differences between treatment groups in the primary measures of efficacy were statistically significant. This between-group difference was thought to be attributable to the high incidence of 'unknown' outcomes, particularly in the diclofenac/misoprostol recipients. No significant differences were observed in the modified intention-to-treat or evaluable cohorts, which excluded patients with an 'unknown' outcome. No significant between group differences were observed in the change from baseline in the ankylosing spondylitis assessments. A difference was apparent between groups in the incidence of adverse events, particularly abdominal pain and diarrhoea (18.1 and 17.2%, respectively, in diclofenac/misoprostol recipients versus 12.4% each in diclofenac recipients). However, it was thought that the 4-week open-label phase of the study, during which patients received only diclofenac, may have selected for diclofenac-tolerant patients. Thus, in the treatment of ankylosing spondylitis, the diclofenac/misoprostol fixed tablet was as effective as diclofenac. Furthermore, diclofenac/misoprostol was well tolerated.