Published: June 2002
ADR update

Can Patients Stomach COX-2 Inhibitors?

Prescriber Update 23(2): 18–19
June 2002

Dr Ruth Savage, Medical Assessor,
Centre for Adverse Reactions Monitoring,Dunedin

Adverse reaction reports and epidemiological studies suggest that cyclo-oxygenase-2 (COX-2) inhibitors cause gastroduodenal ulceration and subsequent complications.  The risk, however, is halved compared with conventional non-steroidal anti-inflammatory agents (NSAIAs).  Case reports suggest that COX-2 inhibitors, like NSAIAs, may exacerbate inflammatory bowel disease and cause intestinal strictures.  As with NSAIAs, COX-2 inhibitors should be withdrawn in patients with significant gastrointestinal symptoms, pending investigation.

New Zealand reports for COX-2 inhibitors include a high proportion of gastrointestinal adverse effects

Gastrointestinal adverse effects account for about 30% of the COX-2 inhibitor reactions reported to the Centre for Adverse Reactions Monitoring (CARM).  There are 65 reports of adverse gastrointestinal effects attributed to celecoxib, 17 of which are serious.  These include melaena, gastroduodenal ulcer, intestinal perforation and three deaths.  All except two patients had at least one risk factor, other than use of a COX-2 inhibitor, for gastroduodenal ulceration.  Fourteen patients were aged 75 years or older.  Gastrointestinal reactions have also been reported with rofecoxib.*

Gastrointestinal toxicity reduced by COX-2 inhibitors but not abolished

The COX-2 inhibitors exert their therapeutic effect by inhibiting the production of prostaglandins involved in inflammation.  At therapeutic doses they have little or no inhibitory effect on cyclo-oxygenase-1, which is expressed in many tissues and is necessary for the production of prostaglandins that protect the mucosa of the upper gastrointestinal tract.  It is therefore expected that COX-2 inhibitors will have reduced gastroduodenal toxicity compared with conventional NSAIAs.

Studies of clinical importance are those that assess the risk of symptomatic ulcers, and bleeding and perforated ulcers.  A meta-analysis1  of randomised trials of rofecoxib and a prospective study2  of rofecoxib and naproxen in 8076 patients (the VIGOR trial) each showed a 50% reduction in risk of peptic ulcer complications and/or upper gastrointestinal bleeding compared with their NSAIA comparators.  A large prospective study3  of celecoxib compared with diclofenac and ibuprofen (the CLASS trial) showed a similar reduction in risk for ulcer complications although this was not significant.  The reduction in risk was significant when patients with symptomatic ulcers were included in the analysis.

These studies1-3 indicate that COX-2 inhibitors do carry an increased risk of gastroduodenal ulcer complications but this is less than with conventional NSAIAs.  Since a number of case-control studies4 have indicated an approximately 4-fold increase in risk of gastroduodenal ulcer complications with conventional NSAIAs compared with no NSAIA use, it can be assumed from the above studies1-3 that the increase in risk with the COX-2 inhibitors is approximately 2-fold.  However, some of this benefit may be lost if low-dose aspirin is taken concurrently.  The CLASS study3 showed that the risk with celecoxib was greater when low-dose aspirin users were included in the study.

The adverse reaction reports to CARM suggest serious upper gastrointestinal reactions occur predominantly in elderly patients and those with other risk factors.  It is likely that this observation in part reflects preferential prescribing of COX-2 inhibitors to at-risk patients.  However, a recent report5 from the United Kingdom indicates that most cases of serious gastroduodenal disease attributed to conventional NSAIAs also occur in patients with other risk factors.  It is not known what degree of risk reduction is achieved when switching high-risk patients from NSAIAs to COX-2 inhibitors.  The studies1-3 described above did include elderly patients and those with a history of ulcer, so some reduction in risk in these patient groups is possible.

Using lower doses of NSAIAs may also reduce harm

When deciding whether to use a COX-2 inhibitor, also consider whether more cautious prescribing of a conventional NSAIA would instead be appropriate.  In the above studies,1-3 the maximum doses of conventional NSAIAs were used as comparators.  It has been shown that the risk of gastroduodenal ulcer complications increases with the dose of NSAIA.6   For patients without significant joint inflammation, low doses and intermittent use may be sufficient.  NSAIAs, even at low doses, are contraindicated in patients with active or previous gastroduodenal ulceration.

Avoid COX-2 inhibitors in patients with active gastroduodenal disease

COX-2 inhibitors should not be prescribed to patients with active gastroduodenal disease.  Where there is a clear need for an anti-inflammatory agent, they may be prescribed cautiously for patients with a history of gastroduodenal ulceration.  Evidence that there is a small risk of COX-2 inhibitors causing gastroduodenal ulceration suggests co-prescribing of a gastroprotective agent should be considered in patients who have previously had serious gastrointestinal reactions to conventional NSAIAs.

Adverse effects on the large intestine have been reported

There are two reports in the literature of exacerbation of inflammatory bowel disease with celecoxib.7  Also a 46-year-old woman developed diaphragm-like strictures of the colon while taking celecoxib.8  Conventional NSAIAs are also known, on rare occasions, to cause ulceration, haemorrhage and diaphragm-like strictures of the distal small intestine and large intestine.9

Use lowest possible dose and discontinue if symptoms of gastrointestinal toxicity occur

In order to reduce gastrointestinal toxicity, both COX-2 inhibitors and conventional NSAIAs should be prescribed at the minimum effective dose.  Unless there is ongoing inflammatory joint disease, they should preferably be used short-term or intermittently.  Where strongly indicated, COX-2 inhibitors may be prescribed to patients with a past history of gastroduodenal disease, and gastroprotection should be considered.  To avoid progression of adverse effects in the upper and lower gastrointestinal tract, patients should stop taking NSAIAs and COX-2 inhibitors, pending investigation, if pain, bleeding, signs of obstruction or altered bowel habit occur. 

* At the time of this data analysis, the use of celecoxib in New Zealand was double that of rofecoxib.

Competing interests (author): none

Correspondence to Dr Ruth Savage, CARM, PO Box 913, Dunedin.

  1. Langman MJ, Jensen DM, Watson DJ, et al. Adverse upper gastrointestinal effects of rofecoxib compared with NSAIDs. JAMA 1999;282:1929-1933.
  2. Bombardier C, Laine L, Reicin A, et al. Comparison of upper gastrointestinal toxicity of rofecoxib and naproxen in patients with rheumatoid arthritis (the VIGOR study). N Engl J Med 2000;343:1520-1528.
  3. Silverstein FE, Faich G, Goldstein JL, et al. Gastrointestinal toxicity with celecoxib vs nonsteroidal anti-inflammatory drugs for osteoarthritis and rheumatoid arthritis. The CLASS study: a randomized controlled trial. JAMA 2000;284:1247-1255.
  4. Savage RL, Moller PW, Ballantyne CL, et al. Variation in the risk of peptic ulcer complications with nonsteroidal anti-inflammatory drug therapy. Arthritis and Rheumatism 1993;36:84-90.
  5. Medicines Control Agency (MCA)/Committee on Safety of Medicines (CSM). Non-steroidal anti-inflammatory drugs (NSAIDs) and gastrointestinal (GI) safety. Current Problems in Pharmacovigilance April 2002;28:5.
  6. Langman MJ, Weil J, Wainwright P, et al. Risk of bleeding peptic ulcer associated with individual non-steroidal anti-inflammatory drugs. Lancet 1994; 343:1075-1078.
  7. Bonner GF. Exacerbation of inflammatory bowel disease associated with use of celecoxib [letter]. Am J Gastroenterology 2001;96:1306-1308.
  8. Eid RY, Quirk DM. Diaphragm-like strictures of the colon induced by the selective COX-2 inhibitor celecoxib: a case report and review of the literature. Am J Gastroenterology 2001;96(Suppl):S194.
  9. Savage RL. NSAIAs can cause lower GIT damage. Prescriber Update June 2001;(21):25-27.