The famous notation that “a camel is a horse defined by a committee” underscores attempts to achieve concurrence on rheumatologic issues, wherein the presence of 2 physicians may result in 3 opinions. One wonders if it might be more advantageous to present the variant perspectives, rather than attempt to present the impression of a unified assessment. Given the substantial medicolegal implications of any pronouncements such as those offered in the American College of Rheumatology (ACR) White Paper, published in a recent issue of Arthritis Care & Research (1), their foundations must be unequivocal.
Concern is expressed as to one such White Paper (1) unreferenced statement: “If a patient is taking aspirin for cardioprotective benefit, then selective and nonselective NSAIDs [nonsteroidal antiinflammatory drugs] should be avoided. This combination is associated with an elevated risk of GI [gastrointestinal] bleeding.” While it might seem intuitive that combining aspirin with an NSAID would increase GI bleeding risk, that actually has not been found with the selective agent celecoxib (2). This leaves the question of cardiovascular risk, and it must be mentioned that even in the White Paper (1) it is noted that “there are insufficient data to ascertain if low-dose aspirin eliminates the [presumed] cardiovascular risk associated with inhibition of COX-2 [cyclooxygenase 2].” Therefore, avoiding selective NSAIDs with concurrent aspirin might even be counterproductive. It certainly is not contraindicated.
There is confusion in the citation of the American Heart Association Position Statement on NSAIDs (3) in the White Paper (1). The title of the American Heart Association article is somewhat misleading, calling it a “scientific statement,” when actually it is only an opinion. However, the comments in the position paper on cardiovascular risk are relatively balanced, and importantly (3) and specifically caution “against relying on meta-analyses that involve an incomplete set of trials….”
The White Paper statement (1) and that of Strand (4) do not seem to adequately address the issue of benefit/risk ratios and the rationale for use of NSAIDs. This is especially true, given failure to emphasize their own statement (1) that “There is insufficient evidence from placebo-controlled trials to demonstrate increased cardiovascular risk with nonselective NSAIDs.”
Rather than compare the frequency of clinically significant non-lethal events, Strand (4) attempted to determine the relative risk of death from nonselective NSAIDs versus the cardiovascular risk. However, Strand (4) compared cardiac mortality in a “high-risk indigent” MediCal population with GI-related deaths among individuals ingesting nonselective NSAIDs in very different populations. Close analysis revealed that the mortality with such bleeding was actually 6–16% (5). Valid comparison of the rate of cardiac and GI deaths among individuals consistently consuming NSAIDs is needed to document whether cardiac concerns outweigh GI-sparing therapeutic interventions. Until that is available, one must rely on the incident data related to hospitalizations, and that clearly demonstrates the overwhelming importance of the gastroprotective approach (2).
One approach to the issue of GI toxicity is to use a gastroprotective agent. One such agent, misoprostol, has been documented as reducing the frequency of major GI events related to nonspecific NSAIDs by 50%, if taken 3–4 times per day (note that 1 or 2 times per day did not work) and if NSAID ingestion is appropriately timed (20 minutes after a misoprostol dose) (6). This cumbersome regimen was overcome by combining an NSAID (diclofenac) with misoprostol, in the form of Arthrotec (G. D. Searle & Co., New York, NY). Arthrotec coordinates the timing of release of the diclofenac to that of misoprostol, insuring the desired timing and equivalent GI safety to that observed with 3–4 daily doses of misoprostol (6, 7). Therefore, Arthrotec can be prescribed as 1 or 2 tablets per day. The alternative is 4–6 tablets per day, with anticipated adherence issues, both as to number ingested and timing. Indeed, misoprostol is now the more expensive portion of the tablet. Prescribing the components rather than Arthrotec actually inflates cost 2–3-fold. That insurance company formulary-mandated approach seems fiscally irresponsible.
Therefore, the critical question is whether there is clinically (not just statistically) significant cardiac risk associated with celecoxib use. Contrasting the Vioxx (Merck, Rahway, NJ) experience, early celecoxib studies revealed no such risk (8). Subsequent studies documented cardiac events attributed to celecoxib that were associated with doses typically not used in the care of rheumatology patients (800 mg/day) (9, 10). More clinically pertinent may be the study by Solomon et al (11), which suggested increased risk at the 400 mg dose, but only if taken as a divided dose of 200 mg twice a day.
In the best of all worlds, there would be no need for NSAIDs. In the worst of all worlds, they would become unavailable, whether because of agent withdrawal or because of White/Position Paper–induced medicolegal compromises. Then would narcotics and steroids be used? Celecoxib, at single daily doses up to 400 mg per day, is still clinically important.
That being said, there is a very important observation buried in the text of the White Paper (1): “studies have now confirmed that drugs specifically inhibiting COX-2 are not cross-reactive with aspirin in patients with aspirin-exacerbated respiratory disease and suggest that celecoxib might be safe to use in these patients.” Perhaps the most important contribution of the White Paper is promulgating the apparent safety of celecoxib in patients with aspirin-intolerance syndrome.