The effect of cyclooxygenase-2 inhibitors on spinal fusion

Spine surgeons discourage the use of nonsteroidal anti-inflammatory drugs following spine arthrodesis because of their inhibitory effect on bone-healing. To our knowledge, there are no data on the effects of the new cyclooxygenase-2 inhibitors on bone-healing. We undertook this study to determine the effects of these more selective nonsteroidal anti-inflammatory drugs on spinal fusion in a rabbit model.

Seventy-two New Zealand White rabbits underwent a posterolateral intertransverse process arthrodesis with use of autologous iliac crest bone. Sixty-six rabbits survived the surgical procedure and the perioperative period and had an uneventful postoperative course. These rabbits were randomly divided into three groups. One group received 10 mg/kg of celecoxib orally, the second group received 10 mg/kg of indomethacin orally, and the third group (the control group) received 1 cm (3) of saline solution orally. The rabbits received the treatment daily for eight weeks, after which they were killed and the lumbar spine was harvested. The specimens were palpated for motion, radiographed, and prepared for histological analysis. The quality of the fusion was graded at each level by assigning a histological score of 0 to 7.

Gross inspection and palpation revealed that 64% (fourteen) of the twenty-two control spines and 45% (ten) of the twenty-two spines in the rabbits treated with celecoxib were fused. With the numbers available, this difference was not significant (p = 0.224). Of the twenty-two spines in the indomethacin-treated rabbits, 18% (four) were fused, and this percentage was significantly different from the control value (p = 0.002). On radiographic assessment, the spine segment was deemed to be fused in 82% (eighteen) of the twenty-two controls, 86% (nineteen) of the twenty-two rabbits treated with celecoxib, and 41% (nine) of the twenty-two indomethacin-treated animals. Only the difference between the indomethacin-treated and control groups was significant (p = 0.004). The histological scores averaged 5.2, 4.8, and 3.5 for the control, celecoxib, and indomethacin groups, respectively. There was a significant difference between the control and indomethacin groups (p = 0.002) but not between the celecoxib and control groups (p = 0.161).

These results suggest that celecoxib does not significantly inhibit the rate of spinal fusion in rabbits. They also suggest that the inhibitory effects of nonsteroidal anti-inflammatory drugs on bone-healing are likely mediated by inhibition of cyclooxygenase-1 and that celecoxib is the better choice if treatment with nonsteroidal anti-inflammatory drugs is deemed necessary following spinal arthrodesis.