HSS Physicians Review Literature on the Safety of COX-2 Inhibitors

Joseph M. Lane, MD

Attending Orthopaedic Surgeon, Hospital for Special Surgery
Professor of Orthopaedic Surgery, Weill Cornell Medical College

Joseph A. Markenson, MD

Attending Physician, Hospital for Special Surgery
Professor of Clinical Medicine, Weill Cornell Medical College

Michael K. Urban, MD, PhD

Associate Attending Anesthesiologist, Hospital for Special Surgery

Following the recall of Vioxx and concern over the safety of Celebrex and other COX-2 inhibitors and NSAIDs, many patients and physicians have been left trying to sort out the facts. In response, Hospital for Special Surgery convened a physician panel on December 17, 2004 to review the literature on three controversial areas in the use of COX-2 inhibitors-acute pain management, cardiac problems, and bone healing and fusion.

Cox-2 Inhibitors and the Treatment of Acute Postoperative Pain: Dr. Urban

The goals of post-operative pain control are to treat pain effectively; ameliorate the post-operative stress response (which can include hypertension, myocardial ischemia, water retention, splinting, anxiety/depression, and muscle spasm/atrophy); facilitate postoperative rehabilitation; minimize side-effects; improve outcomes; and decrease hospital stays. A multi-modal approach to pain control is most effective at achieving these objectives since it reduces the dose of each analgesic, improves antinociception due to synergistic/additive effects, and may reduce the severity of side-effects. When possible, HSS physicians use regional anesthetic techniques and a combination of local anesthetics, narcotics, and anti-inflammatory agents to treat post-operative pain. Attempts are made to minimize the administration of narcotics, since they cause sleepiness and nausea that may slow post-operative recovery.

Non-steroidal anti-inflammatory drugs (NSAIDs) work by inhibiting cyclooxygenase, an enzyme that controls inflammation. The use of post-operative NSAIDs is advantageous because they reduce pain, particularly movement-evoked pain; decrease swelling and inflammation; improve musculoskeletal function; reduce opioid dosage; and accelerate postoperative recovery. However, as with all medications, the use of NSAIDs has potential disadvantages, including gastric damage, and ulceration; bleeding caused by loss of platelet aggregation; renal ischemia; and delayed healing and bone fusion.

In the 1990s, scientists discovered that there were actually two cyclooxygenase enzymes: COX-1, a "constitutive" enzyme that protects the gastric mucosa and hemostasis, and COX-2, an "inducible" enzyme involved in the inflammatory response[1]. Biochemists then developed a new generation of NSAIDs that block only the COX-2 enzyme, thereby reducing the harmful GI side-effects of non-specific NSAIDs. Following the recent recall of rofecoxib (Vioxx), the two COX-2 inhibitors available on the market today are celecoxib (Celebrex) and valdecoxib (Bextra). Meloxicam (Mobic) has both COX-1 and COX-2 properties, and hence is more likely to cause GI side-effects than pure COX-2 medications.

The advantages of using COX-2 inhibitors for pain relief have been well demonstrated in the literature. Randomized controlled studies have shown COX-2 inhibitors to be superior to placebos and to non-specific NSAIDs in the treatment of perioperative pain. In addition, a study of total hip replacement patients at HSS showed that patients who received celecoxib required less narcotics and had less post-operative pain, enhancing their rehabilitation efforts and reducing lengths of stay[2]. Finally, two major randomized studies, VIGOR and CLASS, investigated the gastrointestinal safety of rofecoxib and celecoxib and found that they were effective in reducing the GI side-effects associated with traditional NSAIDs.

The Current Controversies on Cardiovascular Implications of Chronic Cox-2 Inhibitor Usage: Dr. Markenson

COX-2 inhibitors block the production of prostacyclin but do not inhibit thromboxane. Prostacyclin inhibits platelet aggregation while thromboxane promotes platelet aggregation. There has been concern, therefore, that COX-2 inhibitors may lead to increased thrombotic activity and therefore increased cardiovascular events[3]. Garrett Fitzgerald first raised this prothrombotic hypothesis in 1999 and, though it has never been proven, it spurred Merck to begin to examine the cardiovascular impact of COX-2 inhibitor usage.

Merck's examination of data from several large trials yielded conflicting results regarding the effect of COX-2 inhibitors versus traditional NSAIDs on cardiovascular events. The CLASS trial showed no significant difference in cardiovascular event rates in patients taking celecoxib and those taking NSAIDs. In contrast, the VIGOR trial, which compared rofecoxib with naproxen (a non-specific NSAID), found that cardiovascular events were more frequent in patients treated with rofecoxib[4]. However, this apparent increased risk could actually be the result of a cardioprotective effect from naproxen rather than a detrimental effect from rofecoxib. In fact, Merck's retrospective examination of data from studies involving thousands of elderly and Alzheimer's patients found no further evidence of an increased risk of cardiovascular events with chronic COX-2 inhibitor use[5]. Nevertheless, in April 2003 rofecoxib received a revised label that included precautions about cardiovascular risk cited in the VIGOR trial[6].

Since the VIGOR trial, other studies have continued to add to the data that perhaps there is a cardiovascular problem with COX-2 usage. One study found that patients taking rofecoxib had a higher risk of acute myocardial infarction than those taking celecoxib or those not taking any type of NSAID[7]. Another study showed an increase in the risk of cardiovascular events with rofecoxib doses greater than 25 mg per day[8].

The APPROVe study finally convinced Merck to withdraw rofecoxib from the market when, after 18 months, around twice as many subjects receiving a daily 25 mg dose of rofecoxib experienced cardiovascular events than those taking placebo[9].

In summarizing the current state of knowledge about the cardiovascular implications of COX-2 usage, an FDA representative addressing the ACR conference commented that there is no definitive evidence of class effect; agents differ in terms of their selectivity; dose response may be an important factor; and the cardiovascular risk of conventional NSAIDs is unknown[10]. She also spoke about the evaluation challenges, including the difficulty in conducting long-term placebo controlled trials in pain syndromes, the fact that high-risk groups may be on concomitant aspirin, and the lack of statistical power in many studies[11]. Looking ahead, the FDA expects increasing certainty with regards to celecoxib, further evaluation of valdecoxib, and careful scrutiny of other existing NSAIDs[12]. The FDA plans to convene a public advisory committee to discuss the cardiovascular safety issue[13].

Subsequent to this panel, Pfizer reported that the National Cancer Institute suspended the use of Celebrex in two of its long-term cancer prevention trials because in one of those trials, the Adenoma Prevention with Celecoxib Trial, patients who were treated with celecoxib (400-800mg) for an average of 33 months were found to have an increased risk of experiencing heart problems (2.5 for 400mg celecoxib and 3.4 for 800mg celecoxib) compared to those taking placebo. However, the comparable simultaneously conducted cancer prevention study, Prevention of Spontaneous Adenomatous Polyps, did not find any increased risk of heart problems in patients taking celecoxib.

These findings were further complicated by results reported from the Alzheimers's Disease Anti-Inflammatory Prevention Trial in which 2,400+ patients over the age of 70 who were considered to be at risk of developing Alzheimer's disease because of family history were randomly assigned to take either naproxen (220mg twice daily of Aleve), celecoxib (200mg twice daily of Celebrex), or placebo. In this trial, the patients taking naproxen were found to have an apparent increase in cardiovascular and cerebrovascular events compared with those taking celecoxib or placebo.

While this data is concerning, it is by no means clean and definitive. Despite being marketed for many years, NSAID's have yet to be studied for cardiovascular safety in large, multi-year placebo controlled trials. In addition, this data must also be viewed in relation to the large body of evidence, encompassing more than 30,000 patients over ten years, in which an increased risk of cardiovascular events with the use of Celebrex has not been demonstrated.

In conclusion, each physician should weigh the risks and benefits of using COX-2 inhibitors and conventional NSAID's in each individual patient. Long-term use of these drugs before further analysis of their risks may be justified when it is necessary to improve quality of life and maintain activities of daily living. There is no real evidence of increased risk with short-term use of the available agents (<6 weeks). We intend to continue reviewing all the data and look forward to hearing the results of the FDA panel that will meet in February 2005.

COX-2: Bone Healing/Spinal Fusion: Dr. Lane

The effect of COX-2 inhibitors on fracture healing has been studied in comparative models. A group at the University of Rochester showed that COX-2 knockout mice had profoundly compromised bone healing and smaller bone nodules and did not respond to growth factors that would stimulate bone healing in normal animals[14]. In addition, several studies have shown that celecoxib, rofecoxib, and indomethacin (a traditional NSAID) either delay or inhibit fracture healing in rats[15],[16]. Another study found that COX-2 inhibitors decreased the strength of fracture healing at 21 days[17].

The effect of NSAIDs and COX-2 inhibitors on spinal fusion has also been studied. A controlled rabbit study showed a decrease in fusion rates with indomethacin and no statistical difference with celecoxib[18]. A human retrospective study showed that the non-union rate was related to the dosage of Toradol (a traditional NSAID) and that the patients who took Toradol had five-times the rate of non-union than those that did not[19].

In fact, studies have shown that COX-2 inhibitors work very effectively in preventing heterotopic ossification in comparative models and that celecoxib shows the same efficacy as indomethacin in the prevention of heterotopic ossification in total hip replacement patients[20],[21]. Finally, studies have also shown that traditional NSAIDs and COX-2 inhibitors reduce the penetration into the porous ingrowth[22],[23],[24].

Limitations of these studies include uncertainty about how dosing in comparative models compares to humans; a lack of statistical significance because of a small number of study subjects; concerns about the validity of translating findings in young, vigorous animals to an elderly and frail patient population; and the lack of randomized controlled human studies.

Despite these limitations, we can conclude that the use of NSAIDs/COX-2 inhibitors effect fracture healing and spine fusion. Therefore, judicious use of NSAIDs/COX-2 inhibitors is recommended. Specifically, administration of NSAIDs/COX-2s should be delayed for three to four weeks in fracture healing and should never be used in the case of spinal fusion. In addition, physical therapy for patients with porous ingrowth should be delayed, especially in elderly patients, and the time line of patients receiving NSAIDs to inhibit heterotopic ossification should be adjusted to take into account the effect of the drug on bone healing.

Summary prepared by Reesa Kaufman

[1] Bakhle et al. Med Inflamm. 1996;5:305-323Vane et al. Inflamm Res. 1995;44:1-10

[2] Urban et.al. 2004|

[3] Catella-Lawson et al. J Pharmacol Exp Ther. 1999;289:735-741; McAdam et al. Proc Natl Acad Sci U S A. 1999;96:272-277

[4] Committee Briefing Document. NDA 21-042, s007. VIOXX Gastrointestinal Safety. February 8, 2001. Available at: http://www.fda.gov/ohrms/dockets/ac/01/briefing FDA Advisory /367762_03_med.pdf. Date accessed: March 4, 2002.


[6] VIOXX® (rofecoxib tablets and oral suspension) [prescribing information]. Whitehouse Station, NJ: Merck & Co., Inc.; April 2002; Scott-Levin PDDA, YTD November 2002.

[7] Solomon et al. Arthritis Rheum. 2003;48:S697. Abstract 1823.

[8] Ray,WA Lancet 2002:360:1071-73

[9] Bresalier R, ACR 2004, Monday Oct 18

[10] Woodcock J, ACR 2004, Monday Oct 18

[11] Woodcock J, ACR 2004, Monday Oct 18

[12] Woodcock J, ACR 2004, Monday Oct 18

[13] Woodcock J, ACR 2004, Monday Oct 18

[14] Zhang et al., Cyclooxygenase-2 regulates mesenchymal cell differentiation into the osteoblast lineage and is critically involved in bone repair. The Journal of Clinical Investigation, 109(11) 1405-1415.

[15] Simon AM, et al. Cyclo-Oxygenase 2 Function is Essential for Bone Fracture Healing. Journal of Bone and Mineral Research, 17(6): 963-976.

[16] Brown KM, et al. Effect of COX-2-specific inhibition on fracture-healing in the rat femur. JBJS-A, 86A(1): 116-123.

[17] Gerstenfeld LC, et al. Differential Inhibition of fracture healing by non-selective and cyclooxygenase-2 selective non-steroidal anti-inflammatory drugs. Journal of Orthopaedic Research, 21(2003): 670-675.

[18] Long J, et al. The Effect of Cyclooxygenase-2 Inhibitors on Spinal Fusion. JBJS-Am, 2002. 84A(10): 1763-1768

[19] Glassman SD, et al. The effect of postoperative Nonsteroidal Anti-inflammatory Drug Administration on Spinal Fusion. Spine, 1998. 23(7): 834-838.

[20] Rapuano/Mineralized Tissues Lab

[21] Romanò CL, et al. Celecoxib Versus Indomethacin in the Prevention of Heterotopic Ossification After Total Hip Arthroplasty. The Journal of Arthroplasty, 2004. 19(1): 14-18.

[22] Trancik T, et al. The effect of indomethacin, aspirin, and ibuprofen on bone ingrowth into a porous coated implant. Clin Orthop, 1989. 249: 113-121.

[23] Keller JC, et al. Effects of Indomethacin on bone ingrowth. J Orthop Res, 1989. 7: 28-34.

[24] Cook SD, et al. Effects of indomethacin on biologic fixation of porous coated titanium implants. J Arthroplasty, 1995. 10: 351-358

^ Back to Top
Request an Appointment