To the Editor:—
We read with great interest the excellent meta-analysis by Marret et al. 1and the accompanying editorial by Professor Kehlet2on the effects of combined opioid and nonsteroidal antiinflammatory drugs (NSAIDs) use to relieve postsurgical pain. Marret et al. conclude that NSAIDs (cyclooxygenase 2 [COX-2] selective and nonselective), in the aggregate, provide approximately 30% reduction in morphine consumption, with associated reductions in postoperative nausea and vomiting and sedation, but not pruritus, urinary retention, and respiratory depression. Although the efficacy data support the use of multimodal analgesia involving opioids and COX-2 selective NSAIDs, we believe the safety data on the short-term perioperative use of COX-2 selective inhibitors are less than clear for the following reasons.
First, the US New Drug Application for rofecoxib was based almost exclusively on studies in patients with chronic pain. Studies to directly support the postsurgical pain indication consisted of 741 patients, of whom 85% received one dose for dental pain and 15% received five doses for orthopedic pain.*Marret et al. note that the cardiovascular risk of rofecoxib was associated with “long-term use.” However, in the absence of robust studies in high-risk surgical patients and based on data on parecoxib, one cannot preclude similar safety risks with short-term postsurgical use of rofecoxib.
Second, although parecoxib (the injectable prodrug of the now withdrawn valdecoxib) is an effective analgesic, there remain serious unanswered questions about the safety of short-term use in the postsurgical setting. An important early adverse postsurgical safety signal came from a coronary artery bypass graft study included in the original US New Drug Application.†In the parecoxib and valdecoxib group, 19.0% had serious adverse events, versus 9.9% in the placebo group. Citing deficiencies in the data, including a numerically higher incidence of myocardial infarctions (1.9% vs. 0.7%) and cerebrovascular events (2.6% vs. 0.7%) and deaths (4 vs. 0), parecoxib received a nonapprovable letter from the Food and Drug Administration in 2001. The Food and Drug Administration concluded that “the adverse event profile of parecoxib was generally worse than that of placebo in this trial. Although not statistically significantly different, the number of deaths, myocardial infarctions, cerebrovascular accidents, pulmonary embolisms, along with renal and pulmonary complications were also numerically more frequent for parecoxib during this IV dosing period than placebo. In fact, during the entire study period, the incidence of these clinically relevant adverse events associated with parecoxib/valdecoxib was statistically significantly different than placebo. Similarly, during the entire study period, more patients in the parecoxib/valdecoxib versus the placebo group withdrew from the study due to an adverse event.”†
Third, follow-up studies conducted with parecoxib have raised additional safety issues. In one trial,3cardiovascular events (including myocardial infarction, cardiac arrest, stroke, and pulmonary embolism) occurred significantly more frequently in the parecoxib and valdecoxib group than with placebo (2.0% vs. 0.5%; P = 0.03). In another trial,4there were significantly more sternal wound infections with parecoxib than with placebo (3.2% vs. 0%; P = 0.03).
Fourth, 3 yr ago, the European Medicines Evaluation Agency issued a public statement on parecoxib regarding the risk of serious hypersensitivity and skin reactions, including Stevens-Johnson syndrome, toxic epidermal necrolysis, erythema multiforme, and exfoliative dermatitis, as well as anaphylaxis and angioedema.5The European Medicines Evaluation Agency has since contraindicated the use of parecoxib in patients with ischemic heart disease and stroke. Excluding individuals with silent ischemia, this translates to approximately 20 million at-risk patients in the United States. Immediately before its withdrawal in the United States, the Food and Drug Administration required a similarly worded box warning for valdecoxib. Notably, the warning indicated (1) a higher risk of serious skin reactions within the first 2 weeks, (2) a greater propensity for such reactions with valdecoxib than other COX-2 inhibitors, and (3) a recommendation to discontinue valdecoxib at the first appearance of skin rash. Since self-limiting pruritus and hypersensitivity are common in the postsurgical setting, causality assessment in patients with early signs of serious and unrelated skin reactions may prove difficult.
Fifth, there are indications that the increased COX-2 levels observed under pathologic conditions in endothelial cells and atherosclerotic lesions provides atheroprotection and modulates vascular remodeling through its principal metabolite, prostacyclin.6In addition, in patients with acute pain, there is considerable spinal up-regulation of COX-1, which may provide a further rationale for the use of drugs that inhibit both isoforms of COX in acute pain.
Finally, survey data continue to support the view that postsurgical pain is undertreated. Although the efficacy of COX-2 selective NSAIDs such as parecoxib is comparable to nonselective NSAIDs, further safety data are required to support their short-term use in the perioperative setting.
‡TheraQuest Biosciences, Blue Bell, Pennsylvania. nbabul@theraquestinc.com