We investigated the effects of cyclooxigenase-2 (cox-2) on fracture healing. After closed non-displaced fractures were created at the middle of both femoral shafts in 12-week-old Wister rats, a cox-2 specific inhibitor, etodolac (20 mg/day; intra-peritoneal) was administered every day for three weeks (E group). Bone union and callus formation were evaluated by weekly radiographs. Three weeks after surgery, the mechanical strength of the fractured femur was evaluated by a three-point-bending test. These results were compared with those of a vehicle control group (V group). The fracture healing score on radiographs in the E group three weeks after the surgery was 3.3 +/- 0.9, and in the V group it was 5.8 +/- 1.5, indicating that fracture healing was significantly poorer in the E than the V group (p < 0.05). From the three point bending test, the ultimate strength and stiffness of etodolac-treated fractured femurs were shown to be significantly lower than those in vehicle control group (p < 0.05). Mechanically, femurs of etodolac treated rats were weaker than those of control rats. Thus, it was concluded that etodolac, a cox-2 specific inhibitor, inhibited fracture healing.