Correspondence
Gastrointestinal Toxicity of Nonsteroidal Antiinflammatory Drugs
N Engl J Med 1999; 341:1397-1399October 28, 1999
- Article
To the Editor:
In their article, Wolfe et al. (June 17 issue)1 stated that the number of deaths associated with the use of nonsteroidal antiinflammatory drugs (NSAIDs) in the United States every year is almost identical to that associated with AIDS. This statement, misinterpreted by the press, seriously overstates the risks associated with the use of NSAIDs.
Wolfe et al. extrapolated the estimated number of deaths associated with NSAIDs from death rates in the Arthritis, Rheumatism, and Aging Medical Information System (ARAMIS). Although data from ARAMIS are valuable for examining relative risks, one cannot extrapolate incidence rates from ARAMIS to other populations, such as users of lower doses of NSAIDs. The estimated number of U.S. patients with arthritis who are taking NSAIDs — 13 million — has not been verified and is not traceable to any particular source. This figure2 appears to include 5 million patients with rheumatoid arthritis or “probable rheumatoid arthritis,” a number substantially higher than the American College of Rheumatology's estimate of 2 million.3
The authors appear to assume that all patients with rheumatoid arthritis take NSAIDs daily in the doses needed to reduce inflammation. However, even in the ARAMIS studies cited, only 69 percent of person-time was accumulated while the patients were taking NSAIDs. Estimates from the Third National Health and Nutrition Examination Survey show that only 24 percent of people who state they have been told by a doctor that they have rheumatoid arthritis have taken NSAIDs in the previous month. The authors' assumptions about death rates for patients with probable rheumatoid arthritis and osteoarthritis are not supported by any published data; their estimate of the yearly number of NSAID-related deaths from gastrointestinal causes (16,500) is more than twice the previous estimates from the same ARAMIS data set.4,5 In Figure 1 of their article, Wolfe et al. attribute all 16,500 deaths to NSAIDs. However, in the text, they refer variously to “NSAID-related” and “NSAID-associated” effects. In previous reports involving the same data, 16,500 was stated as the number of associated, not attributable, deaths.2
Finally, deaths from AIDS is a dramatic but unstable and misleading comparison, because mortality from AIDS has dropped sharply. The cited number of deaths from AIDS may be inaccurate; the Centers for Disease Control and Prevention reported 21,437 deaths from AIDS in 1997 (substantially more than the 16,500 NSAID-related deaths mentioned by the authors).6
The article's title is misleading, because it does not refer to the well-known relation between dose and toxicity. Consequently, press reports referred to the risks associated with NSAIDs for all users, generating great anxiety among the millions of people who safely take lower doses of over-the-counter NSAIDs to relieve their symptoms or to prevent cardiovascular events.
6 ReferencesRobert W. Morgan, M.D., S.M.Hyg.
Exponent Health Group, Menlo Park, CA 940251
Wolfe MM ,Lichtenstein DR ,Singh G . Gastrointestinal toxicity of nonsteroidal antiinflammatory drugs. N Engl J Med 1999;340:1888-1899
Full Text | Web of Science | Medline2
Singh G ,Triadafilopoulos G . Epidemiology of NSAID induced gastrointestinal complications. J Rheumatol 1999;26:Suppl 56:18-24
Web of Science3
American College of Rheumatology. (See: http://www.rheumatology. org/patients/factsheet./ra.html.)
4
Fries J . Toward an understanding of NSAID-related adverse events: the contribution of longitudinal data. Scand J Rheumatol Suppl 1996;102:3-8
CrossRef | Medline5
Fries JF . Assessing and understanding patient risk. Scand J Rheumatol Suppl 1992;92:21-24
CrossRef | Medline6
Centers for Disease Control. HIV/AIDS surveillance report. Year-end ed. Vol. 10. No. 2. 1998:38. (See: http://www.cdc.gov/nchstp/hiv_aids/stats/hasr1002.pdf.)
To the Editor:
Wolfe and colleagues have presented an accurate and balanced review of the substantial numbers of severe gastrointestinal complications, including death, associated with the use of NSAIDs. Unfortunately, they failed to distinguish clearly between occasional use of over-the-counter drugs and prescription doses of these medications, and their historical development of the subject tended to implicate aspirin more than other NSAIDs. As a consequence, the heavy response from the media, typical for a Journal review article presenting some startling numbers, included misinterpretations that may have alarmed consumers unnecessarily.
At least two thirds of serious side effects from NSAIDs occur with prescription doses, despite the far larger numbers of exposures to over-the-counter drugs. At typical over-the-counter levels of use of NSAIDs, taken occasionally (e.g., an average of 6 to 10 tablets monthly), the risk of serious complications with these agents has not yet been established as measurably different from base line. Regular daily use — for the prevention of coronary events, for example — carries some risk, but the risk is relatively small. My colleagues and I have contributed to the recognition of the silent epidemic of NSAID-related gastropathy, and we continue to make efforts to decrease the rates of complication.1-3 At the same time, a data- and risk-based perspective must be maintained. With occasional use in a healthy person, NSAIDs pose little risk of serious problems. In the abuser of analgesics, even over-the-counter drugs may pose substantial risk.
Aspirin, when used for cardioprotection, can save far more lives by reducing the number of coronary events than are lost because of gastrointestinal consequences of its use. It would be most unfortunate if consumers were unnecessarily frightened into discontinuing cardioprotective medication. In addition, there is likely to be a role for this class of drugs in reducing the incidence of colon cancer and possibly Alzheimer's disease. The quality-of-life benefits of reduced headache, reduced menstrual or musculoskeletal symptoms, and the reduced risk of heart attack or stroke are the driving reasons behind the widespread (and generally responsible) use of these agents. Patients need to be given accurate information about the risks and benefits associated with the use of NSAIDs so that they can make informed decisions about their use, and we must attempt to correct the media when they fail to present a full perspective.
(ARAMIS, of which I am principal investigator, is funded primarily by the National Institutes of Health but also contracts for studies with multiple pharmaceutical companies that market NSAIDs and over-the-counter analgesics.)
3 ReferencesJames F. Fries, M.D.
Stanford University School of Medicine, Palo Alto, CA 94304-57551
Fries JF ,Williams CA ,Bloch DA ,Michel BA . Nonsteroidal anti-inflammatory drug-associated gastropathy: incidence and risk factor models. Am J Med 1991;91:213-222
CrossRef | Web of Science | Medline2
Fries JF ,Ramey DR ,Singh G ,Morfeld D ,Bloch DA ,Raynauld JP . A reevaluation of aspirin therapy in rheumatoid arthritis. Arch Intern Med 1993;153:2465-2471
CrossRef | Web of Science | Medline3
Fries JF . Quality-of-life considerations with respect to arthritis and nonsteroidal anti-inflammatory drugs. Am J Med 1998;104:14S-20S
CrossRef | Medline
To the Editor:
Wolfe et al. present a thorough review of the gastrointestinal toxicity of NSAIDs. Although they correctly point out that proton-pump inhibitors are superior to histamine H2–receptor antagonists and mucosal protective agents for the management of NSAID-related dyspepsia and NSAID-related gastroduodenal ulcers, they fail to comment on the effect of proton-pump inhibitors on the bioavailability of NSAIDs.
Aspirin and several other NSAIDs are weak acids that cross the mucosa in their lipid state. The suppression of acid production reduces the lipophilic nature of these compounds. This, in turn, reduces their absorption and bioavailability. Omeprazole has been shown to reduce the absorption of aspirin and to decrease its bioavailability significantly.1 This evidence suggests that patients require larger doses of NSAIDs for an analgesic effect to be exerted if they are also taking a proton-pump inhibitor. The safest approach with such patients is for them to use acetaminophen or a nonacetylated salicylate for analgesia, rather than to use an NSAID concurrently with a proton-pump inhibitor.
1 ReferencesGavin Harewood, M.D.
Mayo Clinic, Rochester, MN 559011
Anand BS ,Sanduja SK ,Lichtenberger LM . Effect of omeprazole on the bioavailability of aspirin: a randomized controlled study on healthy volunteers. Gastroenterology 1999;116:A371-A371 abstract.
Web of Science
Author/Editor Response
The authors reply:
To the Editor: Dr. Morgan suggests that the exposure to NSAIDs in the United States is considerably less than the 13 million patient-years estimated in ARAMIS studies. This figure may, in fact, underestimate true exposure. In 1995, the National Arthritis Data Workgroup estimated that there were nearly 40 million people with arthritis in the United States and an additional 32 million with frequent low back pain.1 Even if only 25 percent of these people took NSAIDs, the total exposure would be even greater than the figures used in making our estimates. In addition, the ARAMIS estimates did not include the use of NSAIDs for conditions other than arthritis.
With regard to Dr. Morgan's concern about the data base used in our review, in contrast to early studies in which ARAMIS extrapolated incidence rates to estimate mortality, more recent studies, including those cited in our review, have used actual data. The case fatality rate for rheumatoid arthritis was calculated from available data, and we assumed a similar case fatality rate in patients with osteoarthritis, even though these patients are likely to be older and have coexisting conditions. The numbers of deaths attributable to NSAIDs were not reported, because those numbers were not available for the other conditions shown for comparison. The number of AIDS-related deaths in 1977 was 16,516,2 a figure very close to the preliminary data used to estimate mortality in our review. The number of deaths that Dr. Morgan cited includes all deaths in persons with AIDS, even those not related to the infection.
We agree completely with Dr. Fries that the risk is considerably less with over-the-counter doses of NSAIDs than with prescription doses.3 We also concur with both Dr. Fries and Dr. Morgan that the failure of the media to distinguish between over-the-counter and prescription doses of NSAIDs may have caused unwarranted anxiety among the users of these valuable agents. Clearly, as with all interventions, benefit should outweigh risks.
Nevertheless, despite the relatively low risk, the magnitude of the problem associated with over-the-counter NSAIDs (including aspirin) is important, a result of the millions of people who consume these drugs. According to the registry of the American College of Gastroenterology, 84 percent of all cases of NSAID-related serious gastrointestinal hemorrhage were associated with the use of over-the-counter NSAIDs, and the vast majority of cases were associated with the use of aspirin.4 Approximately two thirds of all gastrointestinal hemorrhages in the United Kingdom every year are related to the use of over-the-counter aspirin, and 25 percent are due to prophylactic use.5 Moreover, even occasional use of these drugs carries risk.5
We also concur with Dr. Harewood's recommendations that acetaminophen or a nonacetylated salicylate probably represents the safest approach to analgesia, provided that hepatotoxic doses of acetaminophen are avoided. The preliminary report he cited indicates that omeprazole may reduce the bioavailability of aspirin. However, NSAIDs with different dissociation constants were not examined, and the study did not determine whether the desired therapeutic effects of aspirin were prevented. Finally, in the studies cited in our report, the analgesic effects of various NSAIDs used in combination with omeprazole were not diminished.
5 ReferencesM. Michael Wolfe, M.D.
David R. Lichtenstein, M.D.
Boston University Medical Center, Boston, MA 02118Gurkirpal Singh, M.D.
Stanford University Hospital, Stanford, CA 943051
Lawrence RC ,Helmick CG ,Arnett FC , et al. Estimates of the prevalence of arthritis and selected musculoskeletal disorders in the United States. Arthritis Rheum 1998;41:778-799
CrossRef | Web of Science | Medline2
Hoyert DL ,Kochanek KD ,Murphy SL . Deaths: final data for 1997. Natl Vital Stat Rep 1999;47:54-543
Singh G ,Ramey D ,Balise R ,Triadafilopoulos G . Serious gastrointestinal complications of over-the-counter NSAIDs. Digestion 1998;59:Suppl 3:4-4 abstract.4
Peura DA ,Lanza FL ,Gostout CJ ,Foutch PG . The American College of Gastroenterology Bleeding Registry: preliminary findings. Am J Gastroenterol 1997;92:924-928
Web of Science | Medline5
Weil J ,Colin-Jones D ,Langman M , et al. Prophylactic aspirin and risk of peptic ulcer bleeding. BMJ 1995;310:827-830
CrossRef | Web of Science | Medline
- Citing Articles (5)
Citing Articles
1
Kristopher Silver, Ludovic Leloup, Lisa C. Freeman, Alan Wells, James D. Lillich. (2010) Non-steroidal anti-inflammatory drugs inhibit calpain activity and membrane localization of calpain 2 protease. The International Journal of Biochemistry & Cell Biology 42:12, 2030-2036
CrossRef2
Ingvar Bjarnason, Ken Takeuchi. (2009) Intestinal permeability in the pathogenesis of NSAID-induced enteropathy. Journal of Gastroenterology 44:S19, 23-29
CrossRef3
Ling-dong Quan, Geoffrey M Thiele, Jun Tian, Dong Wang. (2008) The development of novel therapies for rheumatoid arthritis. Expert Opinion on Therapeutic Patents 18:7, 723-738
CrossRef4
Shan Yuan, Victoria Reyes, Mary P. Bronner. (2003) Pseudomembranous Collagenous Colitis. The American Journal of Surgical Pathology 27:10, 1375-1379
CrossRef5
(2000) Current Awareness. Pharmacoepidemiology and Drug Safety 9:2, 165-180
CrossRef
