The pathogenic form of the cyclooxygenase (COX) enzyme, COX-2, is also constitutively present in the spinal cord and has been implicated in chronic pain states in rat and man. A number of COX-2 inhibitors, including celecoxib and rofecoxib, are already used in man for the treatment of inflammatory pain. Preclinically, the dual-acting COX-2 inhibitor, GW406381X [2-(4-ethoxyphenyl)-3-[4-(methylsulfonyl)phenyl]-pyrazolo[1,5-b]pyridazine, where X denotes the free base], is as effective as rofecoxib and celecoxib in the rat established Freund's Complete Adjuvant model with an ED(50) of 1.5 mg/kg p.o. compared with 1.0 mg/kg p.o. for rofecoxib and 6.6 mg/kg p.o. for celecoxib. However, in contrast to celecoxib (5 mg/kg p.o. b.i.d.) and rofecoxib (5 mg/kg p.o. b.i.d.), which were without significant effect, GW406381X (5 mg/kg p.o. b.i.d.) fully reversed mechanical allodynia in the chronic constriction injury model and reversed thermal hyperalgesia in the mouse partial ligation model, both models of neuropathic pain. GW406381X, was also effective in a rat model of capsaicin-induced central sensitization, when given intrathecally (ED(50) = 0.07 mug) and after chronic but not acute oral dosing. Celecoxib and rofecoxib had no effect in this model. Several hypotheses have been proposed to try to explain these differences in efficacy, including central nervous system penetration, enzyme kinetics, and potency. The novel finding of effectiveness of GW406381X in these models of neuropathic pain/central sensitization, in addition to activity in inflammatory pain models and together with its central efficacy, suggests dual activity of GW406381X compared with celecoxib and rofecoxib, which may translate into greater efficacy in a broader spectrum of pain states in the clinic.