We concluded a randomized, controlled trial to compare once-daily 20 mg piroxicam versus once-daily 20 mg tenoxicam in ankylosing spondylitis. We recorded patients' dosing histories with electronic monitors for an average of 225 days (range 55-379) in 34 recipients of piroxicam and 31 recipients of tenoxicam. Dosing histories with the two agents were similar and are combined. Patients took 81% of prescribed doses; 78% once daily (as prescribed) and 3% as two or more daily doses. On 19% of all monitored days, there was no record of a dose being taken; 68% were single no-dose days, the rest (32%) being 2 to >10 consecutive no-dose days. In 3% of monitored days, extra doses were evidently taken, 88% as twice daily and 12% as three or more doses. Only 22% of all patients (14/65) strictly complied with the regimen: one dose daily every day. The remainder alternated between no-dose days and extra-dose days. We found no correlation between patient compliance and improvement in reported pain or morning stiffness.