THE EFFECT OF COX–2 INHIBITORS ON SPINAL FUSION

K. Daniel Riew, MD,
Stephen Lewis, MD,
John Long, DVM,
Timothy Kuklo, MD
St. Louis, MO, USA

INTRODUCTION:
Most spine surgeons discourage the use of NSAIDs following spine fusion because of their inhibitory effect on bone healing. To our knowledge, there is no data on the effects of the new COX–2 inhibitors on bone healing. We undertook this study to determine the effects of these more selective NSAIDs on spinal fusion.

METHODS:
Fifty New Zealand White rabbits underwent intertransverse fusions at the L5–6 level using autogenous iliac crest bone. The rabbits were randomly divided into 3 groups. Group I received celecoxib 50 mg orally, group II received indomethacin 50 mg, and the control group III received oral saline. The rabbits received the treatment daily for 8 weeks at which point they were killed and the lumbar spines harvested. The specimens were palpated for motion and radiographed. On gross inspection, fusion was defined as complete absence of movement. Radiographically, the fusion masses were graded as fully remodeled, moderately remodeled, minimally remodeled or resorbed. The fusion mass was considered fused radiographically if there was contiguous bone across the transverse processes. The readings were done by 2 observers blinded to the rabbits' treatment group.

RESULTS:
There were 17 rabbits in each of the control and celecoxib groups, and 16 in the indomethacin group because of an early death. Radiographically, there was no statistical difference between the control (82%, 14/17) and celecoxib (88%, 15/17) groups for fusion rate or remodeling of the fusion mass; the difference with indomethacin (50%, 8/16) was significant to p<0.05. The fusion masses were fully remodeled in 62% of control and 68% of celecoxib rabbits, vs 41% in the indomethacin group. By gross inspection and palpation, the rate of fusion was 65% (11/17) in the control group. The fusion rate was 47% (8/17) in the celecoxib treated rabbits. The difference was not statistically significant. On the other hand, only 25% (4/16) of the rabbits treated with indomethacin were fused by palpation (p< 0.05). The radiographic and gross interpretation of fusion corresponded in only 26 of 50 rabbits. There were 19 cases that were interpreted as fused radiographically that had some motion on gross inspection. Five cases were fused on palpation but were interpreted as pseudarthrosis on radiographs. In the cases with resorption or minimal remodeling of the fusion mass, there were no discrepancies between the radiographic and gross palpation interpretations.

CONCLUSIONS AND SIGNIFICANCE:
These results suggest that COX–2 inhibitors do not significantly inhibit the fusion rate in the rabbit model. The radiographic appearance of the fusion masses in the celecoxib group closely resembled the control group. There was no statistically significant difference between the fusion rate in the celecoxib group and the control group. To our knowledge, this is the first investigation of the COX–2 inhibitors on bone healing. Our results suggest that COX–2 inhibitors would be the preferred choice if NSAIDs are deemed necessary following spinal fusion.