AAOS Bulletin - June, 2005

COX-2 inhibitors and bone: 2005 update

Prolonged use appears to interfere with bone healing

By Stuart B. Goodman, MD, PhD, and Barbara D. Boyan, PhD

In the October 2002 AAOS Bulletin, Thomas Einhorn, MD, wrote a timely article titled: “Use of COX-2 inhibitors in patients with fractures. Is there a trade off between pain relief and healing?”1 Dr. Einhorn discussed the results of several animal studies that suggested that continuous usage of COX-2 inhibitors may interfere with fracture healing. However, he also emphasized the paucity of clinical data in humans on this subject.

Dr. Einhorn suggested that if COX-2 inhibitors were to be used in exchange for narcotic analgesics, their administration should probably not exceed 10 to 14 days in patients without any other comorbid conditions that might impair fracture healing. If any adverse effects on fracture healing were realized during this short 10-to-14-day period, it was theorized that these effects would reverse after the drug was discontinued.

Since Dr. Einhorn’s article, there have been several more studies investigating the effects of COX-2 inhibitors on bone. Furthermore, there has been great controversy on the potential systemic adverse effects of COX-2 inhibitors on the cardiovascular system. Indeed, two drugs, Vioxx (rofecoxib) and Bextra (valdecoxib), have been withdrawn from the U.S. market. The remaining COX-2 inhibitor on the market, Celebrex (celecoxib), has also undergone scrutiny for potential cardiovascular and other side effects. The Food and Drug Administration is requiring revised warnings on labeling for both COX-2 and nonselective nonsteroidal anti-inflammatory drugs (NSAIDs), highlighting the increased risk for cardiovascular events and gastrointestinal bleeding associated with their use.

This article will not delve into the details of the COX-2 cardiovascular controversy. However, it is worthwhile to review new data on the effects of COX-2 inhibitors and NSAIDs on bone, because these drugs are commonly prescribed as part of a multimodal pain management approach in clinical situations such as the treatment of fractures and nonunions, the use of bone ingrowth prostheses and spinal fusion.

Mechanism of action

Cyclooxygenase (COX) is an enzyme that controls the conversion of arachidonic acid, a phospholipid, to prostaglandins. Prostaglandins are ubiquitous molecules that have numerous effects on all cells in the body, including regulating bone formation by osteoblasts. There are two isoforms of the COX enzyme: COX-1 and COX-2.

COX-1 is a “housekeeping” enzyme that regulates homeostatic activities in virtually all cells. It is particularly important in controlling platelet function, gastric acid secretion and renal blood flow. The COX-2 enzyme is rapidly induced when the body is under systemic or local stress such as during shock, injury, infection or inflammation. Non-specific NSAIDs inhibit both COX-1 and COX-2 activity, whereas COX-2 inhibitors are more selective in their enzyme inhibition.

Effect on fracture healing

During the last 30 years, numerous studies in animals have shown that nonspecific NSAIDs such as indomethacin, ibuprofen, aspirin and others delay fracture healing, bone ingrowth or spinal fusion.2-9 In some studies, a dose-response effect was reported, with higher doses eliciting more profound inhibitory effects.

More recently, studies on selective COX-2 inhibitors have suggested that bone formation is inhibited by these drugs. A group from Rochester University showed delayed fracture healing in homozygous knockout mice that were deficient for the COX-2 gene, COX-2 (-/-), but not the COX-1 gene, COX-1 (-/-).10 In bone marrow cultures from the COX-2-deficient mice, bone nodule formation was reduced, but could be normalized by adding prostaglandins to the culture.

Several research groups have shown that fracture healing in rats was delayed when the animals were administered celecoxib, rofecoxib, paracoxib or etodolac at various doses for periods of 3 to 8 weeks, although the effect on spinal fusion is more controversial.11-14 On the contrary, Brown and co-workers found significant adverse effects on the histological and mechanical properties of healing femoral fractures in rats given indomethacin (a non-specific NSAID) but not celecoxib (a selective COX-2 inhibitor) over a 12-week period, compared to controls with no drug.15 However, more fibrous tissue was present in the celecoxib group at four and eight weeks compared to controls. The treatment and control groups were similar at 12 weeks. A similar negative effect of celecoxib use for eight weeks in a spine fusion model in rabbits was reported but the data are controversial.16

A time-response effect of COX-2 inhibitors on fracture healing has been found using the rat femoral fracture model. Fracture healing assessed at eight weeks was delayed when celecoxib was given for 5, 10 and 15 days post-fracture.12 Etodolac delayed fracture healing at three weeks when given continuously or during the first week only post-fracture.13 COX-2 mRNA levels were shown to peak during the first 14 days of fracture healing in rats but returned to normal levels by day 21, whereas COX-1 mRNA were constant over the 21-day period.14 Taken together, these studies suggest that the first two weeks of fracture healing in the rat appear to be most susceptible to the effects of COX-2 inhibition.

These observations are supported by in vitro studies examining the effects of COX-1 and COX-2 inhibition on the response of osteoblasts to titanium substrates. Inhibition of both enzymes with indomethacin blocked the stimulatory effects of the surface on osteoblast function, and this was dependent on the time at which the cells were exposed to the drug.17 The results indicated that early exposures were more detrimental than later exposures, suggesting that NSAIDs might be more detrimental in vivo during the first weeks of fracture healing.

These in vitro studies also showed that both COX-1 and COX-2 are important for the normal growth and differentiation of osteoblasts in culture.18 Specific inhibition of COX-1 or COX-2 blocks some, but not all, of the responses of osteoblasts to titanium substrates, suggesting that both enzymes may be required for bone ingrowth in vivo.

Effects on bone ingrowth

Are the effects of COX-2 inhibitors on bone ingrowth the same as for fractures? Our research group showed that both naproxen sodium (a nonspecific NSAID) and rofecoxib (a selective COX-2 inhibitor) given continuously for four weeks significantly decreased bone ingrowth in a harvest chamber model implanted in the rabbit tibia.19

In a follow-up study, rofecoxib given continuously for six weeks suppressed bone ingrowth; however, when the drug was given during the initial or final two weeks only of a six-week treatment, bone ingrowth was only slightly (but not significantly) decreased compared to controls.20 This suggests reversibility of the adverse biological effects on bone when the drug is stopped.

Clinical applications

The effects of NSAIDs on bone formation have several clinical uses. In humans, short-term administration of NSAIDs after hip replacement has been used to help prevent heterotopic ossification, without any apparent long-term deleterious effects on clinical outcomes.21-24

In spinal fusion, no adverse effects on union rate at one year have been reported after short-term use of celecoxib for five days postoperatively as part of a comprehensive pain management protocol.25 However, one clinical study identified concomitant NSAID use to be a major factor contributing to femoral nonunion after intramedullary nailing.26 Another study showed that when indomethacin was given for prophylaxis of heterotopic ossification in acetabular fractures, there was a concomitant increase in the nonunion rate of long bone fractures.27

Current situation

So where are we with regards to NSAIDs, COX-2 inhibitors and bone healing in 2005? The timing of administration of NSAIDs and COX-2 inhibitors appears to be an important determinant of the effects of these drugs on bone. Continuous usage of NSAIDs and COX-2 inhibitors in the early stages of bone healing appears to delay bone formation in animal models and probably in humans. Putting aside the potential systemic effects of these drugs on the cardiovascular system and other organs, NSAIDs and COX-2 inhibitors should still be used very judiciously and only for short time periods in patients with conditions involving bone healing such as fractures, nonunions, bone ingrowth prostheses and spinal fusion.

Dr. Einhorn’s conclusions still appear to be valid. Prolonged use of these drugs probably interferes with key biological processes of osteoblasts. The adverse effects of short-term use of NSAIDs and COX-2 inhibitors are probably reversible when the drug is discontinued. However, the greater the exposure to NSAIDs (including COX-2 inhibitors) during the first six weeks of fracture repair, bone ingrowth, spine fusion and related biological processes, the greater the negative effects. If suitable alternatives to NSAIDs are available, it may still be prudent to use the alternative and avoid the use of NSAIDs entirely in these conditions.

Stuart B. Goodman, MD, PhD, is a professor, department of orthopaedic surgery at Stanford University Medical Center, and a member of the AAOS Biological Implants Committee. Barbara D. Boyan, PhD, is a consultant to the Biological Implants Committee and is the Price Gilbert, Jr. Chair in Tissue Engineering in the department of biomedical engineering at Georgia Institute of Technology.


1. Einhorn T: Use of COX-2 inhibitors in patients with fractures. Is there a trade off between pain relief and healing? AAOS Bulletin 2002;50(5):43-44

2. Bo J, Sudmann E, Marton PF: Effect of indomethacin on fracture healing in rats. Acta Orthop Scand 1976;47(6):588-599.

3. Sudmann E, Dregelid E, Besseses A, Morland J: Inhibition of fracture healing by indomethacin in rats. Eur J Clin Invest 1979;9(5):333-339.

4. Altman RD, Latta LL, Keer R, Renfree K, Hornicek FJ, Banovac K: Effect of nonsteroidal anti-inflammatory drugs on fracture healing: a laboratory study in rats. J Orthop Trauma 1995;9(5):392-400.

5. Allen HL, Wase A, Bear WT: Indomethacin and aspirin: effect of nonsteroidal anti-inflammatory agents on the rate of fracture repair in the rat. Acta Orthop Scand 1980;51(4):595-600.

6. Dimer JR, Ante WA, Zhang YP, Glassman SD: The effects of non-steroidal anti-inflammatory drugs on posterior spinal fusions in the rat. Spine 1996;21(16):1870-1876.

7. Keller J, Trancik TM, Young FA, St. Mary E: Effects of indomethacin on bone ingrowth. J Orthop Res 1989;7(1):28-34.

8. Trancik T, Mills W, Vinson N: The effect of indomethacin, aspirin, and ibuprofen on bone ingrowth into a porous-coated implant. Clin Orthop 1989;249:113-121.

9. Cook SD, Barrack RL, Dalton JE, Thomas KA, Brown TD: Effects of indomethacin on biologic fixation of porous-coated titanium implants. J Arthroplasty 1995;10(3):351-358.

10. Zhang X, Schwarz EM, Young DA, Puzas JE, Rosier RN, O’Keefe RJ: Cyclooxygenase-2 regulates mesenchymal cell differentiation into the osteoblast lineage and is critically involved in bone repair. J Clin Invest 2002;109(11):1405-1415.

11. Simon AM, Manigrasso MB, O’Connor JP. Cyclo-oxygenase 2 function is essential for bone fracture healing. J Bone Mineral Research 2002;17(6):963-976.

12. Simon AM, O’Connor JP: Acute inhibition of cyclooxygenase-2 impairs fracture healing. Transactions of the 49th Annual Orthopaedic Research Society: 2003;29:314.

13. Endo K, Sairyo K, Komatsubara S, et al: Cyclooxygenase-2 inhibitor inhibits the fracture healing. J Physiol Anthropol Appl Human Sci 2002;21(5):235-238.

14. Gerstenfeld LC, Thiede M, Seibert K, et al: Differential inhibition of fracture healing by non-selective and cyclooxygenase-2 selective non-steroidal anti-inflammatory drugs. J Orthop Res 2003;21(4):670-675.

15. Brown KM, Saunders MM, Kirsch T, Donahue HJ, Reid JS: Effect of COX-2-specific inhibition on fracture-healing in the rat femur. J Bone Joint Surg Am 2004;86-A(1):116-123.

16. Long J, Lewis S, Kuklo T, Zhu Y, Riew D: The effect of cyclooxygenase-2 inhibitors on spinal fusion. J Bone Jt Surg 2002;84-A:1763-1768.

17. Sisk MA, Lohmann CH, Cochran DL, et al: Inhibition of cyclooxygenase by indomethacin modulates osteoblast response to titanium surface roughness in a time-dependent manner. Clin Oral Implants Res 2001;12(1):52-61.

18. Boyan BD, Lohmann CH, Sisk M, et al: Both cyclooxygenase-1 and cyclooxygenase-2 mediate osteoblast response to titanium surface roughness. J Biomed Mater Res 2001;55(3):350-359.

19. Goodman SB, Ma T, Trindade M, et al: COX-2 selective NSAID decreases bone ingrowth in vivo. J Orthop Res 2002;20(16):1164-1169.

20. Goodman SB, Ma T, Mitsunaga L, Miyanishi K, Genovese MC, Smith RL: Temporal effects of a COX-2-selective NSAID on bone ingrowth. J Biomed Mater Res A 2005;72A(3):279-287.

21. Kjaersgaard-Andersen P, Nafei A, Teichert G, et al: Indomethacin for prevention of heterotopic ossification. A randomized controlled study in 41 hip arthroplasties. Acta Orthop Scand 1993;64(6):639-642.

22. Neal B, Rodgers A, Dunn L, Fransen M: Non-steroidal anti-inflammatory drugs for preventing heterotopic bone formation after hip arthroplasty. Cochrane Database Syst Rev 2000;3:CD001160.

23. Trnka H, Zenz P, Zembsch A, Easley M, Ritschl P, Salzer M: Stable bony integration with and without short-term indomethacin prophylaxis: A 5-year follow-up. Arch Orthop Trauma Surg 1999;119:456-460.

24. Wurnig C, Schwameis E, Bitzan P, Kainberger F: Six-year results of a cementless stem with prophylaxis against heterotopic bone. Clin Orthop 1999;361:150-158.

25. Reuben SS, Ekman EF. The effect of cyclooxygenase-2 inhibition on analgesia and spinal fusion. J Bone Joint Surg 2005;87-A(3):536-542.

26. Giannoudis PV, MacDonald DA, Matthews SJ, Smith RM, Furlong AJ, De Boer P. Nonunion of the femoral diaphysis. The influence of reaming and non-steroidal anti-inflammatory drugs. J Bone Joint Surg 2000; 82-B(5):655-658.

27. Burd TA, Hughes MS, Anglen JO: Heterotopic ossification prophylaxis with indomethacin increases the risk of long-bone nonunion. J Bone Joint Surg Br. 2003;85(5):700-705.

Close Archives | Previous Page