We conducted two separate double-blind withdrawal studies in parallel; one involved 20 patients taking piroxicam 20 mg daily, and the other involved 20 patients taking naproxen 750 mg daily. After 48 hour withdrawal of active medication, the naproxen group showed significant deterioration in eight of the 14 parameters measured and the piroxicam group in one of the 14 parameters measured. The deterioration in the naproxen group was significantly greater than that in the piroxicam group in five of the 14 parameters measured. These findings are consistent with the long half-life of piroxicam and suggest that non-compliance leading to omission of piroxicam for two days is of less clinical importance than omission of naproxen for two days.