Practical approaches to minimizing gastrointestinal and cardiovascular safety concerns with COX-2 inhibitors and NSAIDs
© BioMed Central Ltd 2005
Published: 15 September 2005
Skip to main content
© BioMed Central Ltd 2005
Published: 15 September 2005
Nonsteroidal anti-inflammatory drugs (NSAIDs) are highly effective in treating the pain and inflammation associated with osteoarthritis and rheumatoid arthritis, but it is well recognized that these agents are associated with substantial gastrointestinal toxicity. Treatment guidelines suggest that patients with one or more risk factors for NSAID associated ulcers should be prescribed preventive treatment. However, well over 80% of such patients may not receive an appropriate therapeutic intervention. Multiple strategies are available to reduce the risk for NSAID associated gastrointestinal complications. First, risk may be reduced by using non-NSAID analgesics. Second, use of the minimum effective dose of the NSAID may reduce risk. Third, co-therapy with a proton pump inhibitor or misoprostol may be desirable in at-risk patients. Use of cyclo-oxygenase-2 inhibitors may also reduce the risk for gastrointestinal events, although this benefit is eliminated in patients who receive aspirin, and cyclo-oxygenase-2 inhibitors may increase cardiovascular adverse events. The optimal management of NSAID related gastrointestinal complications must be based on the individual patient's risk factors for gastrointestinal and cardiovascular disease, as well as on the efficacy and tolerability of both the NSAID and accompanying gastroprotective agent.
Nonsteroidal anti-inflammatory drugs (NSAIDs) are among the most prescribed classes of medications worldwide, with over 111 million prescriptions written between September 1999 and August 2000 . In addition, more than 30 billion over-the-counter (OTC) NSAID tablets are purchased annually. NSAID use is common in all age groups, with most frequent use among the elderly, of whom nearly 70% take NSAIDs at least weekly.
The NSAIDs are highly effective in treating the pain and inflammation associated with osteoarthritis (OA) and rheumatoid arthritis (RA), but it is well recognized that these agents are associated with an increased risk for upper gastrointestinal toxicity, ranging from dyspepsia to gastroduodenal ulcers and bleeding. Although only a small proportion of patients who use these agents develops gastrointestinal complications, the widespread use of these agents magnifies the frequency into a large absolute number of clinical gastrointestinal events. Individually and together, NSAID related adverse events have an important impact on medical outcomes, patient quality of life, and health care costs.
In view of the recent controversies surrounding the cardiovascular effects of COX-2 selective agents, the number of patients who receive traditional NSAIDs is likely to increase substantially. Consequently, the number at risk for NSAID related gastrointestinal complications is also expected to increase. Accurate identification of those who are at high risk for NSAID related gastrointestinal toxicity is therefore essential. (These issues are taken up in greater detail elsewhere in this supplement .)
A number of factors have been identified that increase the risk for NSAID associated upper gastrointestinal complications, including ulcers . Use of multiple NSAIDs (including OTC NSAIDs and aspirin) and high dosages of medication increase risk. Interestingly, the greatest relative risk for gastrointestinal complications exists during the first month of treatment. Other important risk factors include prior ulcer complications, advanced age, and concomitant corticosteroid or anticoagulant use. The severity of RA may appear to increase risk independently for adverse gastrointestinal events. In contrast, dyspepsia and other upper gastrointestinal symptoms do not predict the development of upper gastrointestinal events .
The role of Helicobacter pylori infection in NSAID associated gastrointestinal disease remains somewhat controversial. However, a recent meta-analysis  indicated that both H pylori infection and NSAID use are independent risk factors for gastrointestinal complications. It is clear that eradication of the infection, although indicated in all patients with a history of ulcer disease, is insufficient to reduce the NSAID associated risk for ulcer and its associated complications.
Recommendations for the treatment of NSAID related dyspepsia and mucosal injury
Empirical treatment with H2 receptor antagonist or PPI; individualize therapy
Helicobacter pylori infection
Treatment to eradicate infection in patients with a history of peptic ulcer; PPI therapy
Active gastroduodenal ulcer
Treatment with H2 receptor antagonist or a PPI
Treatment with a PPI
Treatment with a PPI or misoprostol or a COX-2 (preferential) or selective NSAID
Finally, the use of COX-2 inhibitors may reduce the risk for gastrointestinal events. However, this benefit is eliminated in patients who receive aspirin, and COX-2 inhibitors may also be associated with increased risk for adverse cardiovascular events.
The efficacy of misoprostol, a synthetic prostaglandin E1 analog, as prophylaxis for NSAID related gastric ulcers has been demonstrated in several trials. In one study  638 patients who were taking a traditional NSAID but did not have a gastric or duodenal ulcer at screening were randomly assigned to 12 weeks of treatment with misoprostol 200 μg or placebo four times daily. Endoscopy was performed at baseline and after 4, 8, and 12 weeks. In the intent-to-treat population, 1.9% of those receiving misoprostol and 7.7% of those in the placebo arm developed a gastric ulcer.
In a second study  8843 ulcer free patients receiving continuous therapy with NSAIDs were randomly assigned to therapy with misoprostol 200 μg/day or placebo four times daily. The primary outcome measure was the development of serious upper gastrointestinal complications, as detected by clinical symptoms or findings. Over the course of the 6-month study, the incidence of serious upper gastrointestinal complications was 40% lower among patients who received misoprostol than among those who received placebo. More patients in the misoprostol group (42%) than in the placebo group (36%) withdrew from therapy, primarily because of gastrointestinal side effects (e.g. diarrhea, abdominal pain, and flatulence).
Although these data indicate that misoprostol is effective in the prevention of ulcers, poor compliance because of the relatively high rate of adverse events and the need for frequent dosing are important concerns. One study  found that lower doses of misoprostol are better tolerated. However, the drug had to be taken at least twice daily to provide effective prophylaxis against NSAID related ulcers as quantified by endoscopy. Outcomes data are not available for such lower dosages.
Misoprostol may also be effective in the treatment of patients with established NSAID associated ulcers, but comparative studies suggest that proton pump inhibitors (PPIs) are substantially more effective in these patients .
The potential for poor compliance with misoprostol may, in part, be reduced by fixed dose combinations of misoprostol with an NSAID. Arthrotec® (Pfizer, Inc., New York, USA), a combination agent consisting of diclofenac 50 mg or 75 mg plus misoprostol 200 μg, was evaluated in a clinical study that enrolled 572 patients with symptomatic OA of the knee or hip and a history of gastric or duodenal ulcers or 10 or more erosions . In this study, endoscopically diagnosed ulcers were significantly less frequent in patients receiving the combination agent than in those receiving diclofenac alone. Outcomes data are not available for this combination.
Sucralfate, a basic aluminum salt of sucrose octasulfate, forms an ulcer adherent complex at duodenal ulcer sites, protecting the ulcer and promoting healing. It may also inhibit pepsin activity in gastric fluid. Sucralfate has been shown to be effective in the treatment of NSAID associated duodenal ulcers, particularly when the NSAID is stopped, but it is not effective in the treatment or prevention of NSAID related gastric ulcers . Its use is not recommended because of the availability of far superior therapeutic options.
H2 receptor antagonists modulate gastric pH through thecompetitive inhibition of the action of histamine at H2 receptor sites on the gastric parietal cell. The efficacy of the H2 antagonist famotidine at high dose (double the usual dose) for preventing ulcers in patients receiving long-term therapy with NSAIDs was examined in a double blind, parallel group, randomized study conducted in patients with RA or OA . A total of 285 patients were randomly assigned to therapy with famotidine (20 mg or 40 mg twice daily) or placebo, and followed for 24 weeks for the incidence of gastric or duodenal ulceration. Gastric ulcers were observed in 13% and 8% of patients receiving the 20 mg and 40 mg dosages of famotidine, respectively, and in 20% of patients in the placebo group (P = 0.03 for famotidine 40 mg versus placebo). The percentages of patients with gastroduodenal ulcers were significantly lower in the famotidine 20 mg group (4%) and the 40 mg group (2%) compared with the placebo group (13%).
Although this agent has been shown to be effective in preventing ulcers in patients taking NSAIDs in this single study, H2 receptor antagonists are not recommended for routine treatment of asymptomatic patients for a variety of reasons, including their potential to mask dyspeptic symptoms associated with mucosal injury . Furthermore, data suggest that the H2 receptor antagonists are less effective in healing gastroduodenal ulcers than are PPIs, whether or not NSAIDs are continued, and are inferior in preventing ulcer recurrence [13, 14].
Unlike the H2 receptor antagonists, PPIs act by binding irreversibly to resident proton pumps (H+/K+-ATPase), thus inhibiting the final common pathway for acid secretion. PPIs are administered as prodrugs that are activated in the acidic environment of the parietal cell secretory canaliculus . Once converted to their active form, PPIs bind to cysteine residues in the proton pump and inhibit acid secretion into the canalicular lumen .
Several large scale clinical trials have compared the efficacies of PPIs and H2 receptor antagonists in patients with established ulcers. In one trial  541 patients with an ulcer or more than 10 erosions were randomly assigned to receive omeprazole 20 mg/day or 40 mg/day, or ranitidine 150 mg twice daily. Healing of gastric lesions was observed in significantly more patients who received omeprazole (80%) than in those who received ranitidine (61%; P < 0.001). Ulcer healing was substantially more rapid with the PPI than with the H2 receptor antagonist.
In another trial  353 patients with confirmed gastric ulcers were randomly assigned to therapy with lansoprazole 15 mg/day or 30 mg/day, or ranitidine 150 mg twice daily. After 8 weeks of therapy, ulcer healing rates in the lansoprazole group ranged from 69% to 73%, as compared with 53% for ranitidine.
The efficacies of PPIs and misoprostol have also been compared in several studies conducted in patients with established ulcers and erosions. In a double blind study conducted by Hawkey and colleagues  935 patients who had developed ulcers or more than 10 erosions in the stomach or duodenum, or both, were randomly assigned to 4–8 weeks of therapy with omeprazole (20 mg/day or 40 mg/day) or 200 μg misoprostol four times daily. Patients in whom treatment was successful were re-randomized to treatment with 20 mg/day omeprazole, 200 μg misoprostol twice daily, or placebo for an additional 6-month maintenance phase. Patients were assessed clinically and endoscopically during both phases. Treatment success, defined as the absence of ulcers or the presence of fewer than five erosions at each site and no more than mild dyspepsia, was observed in 76% and 75% of patients receiving omeprazole 20 mg/day and 40 mg/day, respectively, and in 71% of those given misoprostol. The rate of gastric ulcer healing was significantly higher in patients receiving the 20 mg dosage of omeprazole (87%) than in those receiving misoprostol (73%); the rate of duodenal ulcer healing was higher in the omeprazole 20 mg and 40 mg groups (93% and 89%, respectively) than in the misoprostol group (77%). The percentage of patients with moderate or severe dyspepsia declined to a significantly greater extent in the omeprazole 40 mg group than in the misoprostol group. During the maintenance phase, significantly fewer patients in the omeprazole 20 mg/day group than in the misoprostol group suffered a relapse (27% of patients in the placebo group).
A prospective, double blind, active and placebo controlled study conducted by Graham and colleagues  compared the efficacy of lansoprazole and misoprostol in the prevention of ulcer recurrence among patients with a history of endoscopically detectable gastric ulcers who were free from H pylori infection at baseline. A total of 537 patients were randomized to 12 weeks of therapy with 30 mg/day lansoprazole, 15 mg/day lansoprazole, misoprostol 200 μg four times daily, or placebo. The primary end-point was ulcer status as determined by endoscopy. Evaluable patients in the lansoprazole groups remained free from gastric ulcer significantly longer than those who received placebo but for a shorter time than those who received misoprostol. After 12 weeks of therapy, 93% of patients in the misoprostol group were free from gastric ulcers, compared with 80% and 82% in the lansoprazole 15 mg and 30 mg groups. Significantly more patients in the misoprostol groups reported treatment related adverse events and withdrew from the study. When withdrawal from the study was considered as a treatment failure, lansoprazole and misoprostol were clinically equivalent.
The efficacy of PPIs in the prevention of recurrences of ulcer complications from long-term, low-dose aspirin use was examined in a trial conducted in 123 patients in Hong Kong . All patients initially received antibiotic treatment for H pylori plus 30 mg lansoprazole twice daily, followed by treatment with famotidine 20 mg twice daily for 5 weeks. Patients with healed ulcers and in whom H pylori infection had been eradicated were randomized to 12 months of therapy with aspirin 100 mg/day plus lansoprazole 30 mg/day or to aspirin plus placebo. Recurrence of ulcer complications was observed in 1.6% of patients in the lansoprazole group (one episode of gastrointestinal bleed) and in 14.8% of those in the placebo group (nine gastrointestinal bleeds). After controlling for confounding covariates, patients who received placebo were at a 9.6-fold increased risk for recurrent ulcers.
Similarly, a study conducted by Chan and colleagues  demonstrated that PPIs are superior in preventing recurrent ulcer complications in patients receiving aspirin or other NSAIDs who were also infected with H pylori. In this study, 400 patients with a history of upper gastrointestinal bleeding were treated with omeprazole 20 mg/day until ulcer healing, and then assigned to treatment with omeprazole 20 mg/day or to 1 week of H pylori eradication therapy. Patients previously receiving aspirin were treated with aspirin 80 mg/day; those receiving NSAIDs took naproxen 500 mg twice daily. The risk for recurrent bleeding was 1.9% in aspirin patients who received eradication therapy and 0.9% in those who received omeprazole; among patients receiving other NSAIDs the corresponding risks were 18.8% with eradication therapy and 4.4% with omeprazole.
The reduction in incidence of gastrointestinal complications in patients receiving COX-2 inhibitors is reviewed elsewhere in this supplement . Given recent concerns regarding the cardiovascular safety of this class of medications, and additional data showing that the incidence of gastrointestinal complications is not reduced in patients who receive concomitant aspirin, the risks and benefits of COX-2 inhibitors must be carefully balanced when considering therapy with these agents. In light of these concerns, in July 2005 the US Food and Drug Administration recommended labeling changes for OTC and prescription NSAIDs, including COX-2 selective NSAIDs, that include a boxed warning highlighting the potential not only for increased risk for cardiovascular events but also for life-threatening gastrointestinal bleeding associated with their use .
It should be noted, however, that additional long-term outcomes studies will be necessary to quantify better the benefits of PPI co-therapy for chronic NSAID users compared with the selective COX-2 inhibitors with regard to the risk for bleeding throughout the gastrointestinal tract, because PPI co-therapy affects only injury in the acid vulnerable proximal gastrointestinal tract.
The clinicians' guide to NSAID therapy in the post-rofecoxib/valdecoxib age
No/low NSAID gastrointestinal risk
NSAID gastrointestinal risk
No cardiovascular risk (no aspirin)
Traditional NSAID + PPI
Coxib + PPI1 (if necessary)
Consider non-NSAID therapy
Cardiovascular risk (consider aspirin)
Traditional NSAID2 plus PPI if gastrointestinal risk warrants gastroprotection
A gastroprotective agent must be added if a traditional NSAID2 is prescribed
Consider non-NSAID therapy
Consider non-NSAID therapy
For patients without cardiovascular risk factors who require aspirin prophylaxis and who are at low risk for gastrointestinal events, monotherapy with a traditional NSAID remains the initial approach to anti-inflammatory therapy. Data suggest that strategies consisting of COX-2 inhibitors or NSAIDs plus a PPI are equally effective in reducing the risk for ulcer bleeding. Therefore, patients who do not require aspirin and who are at relatively high risk for NSAID related gastrointestinal events may be prescribed a COX-2 inhibitor or an NSAID with PPI co-therapy. In treating those with cardiovascular risk that requires aspirin prophylaxis, COX-2 inhibitors should generally be avoided; these patients should receive a traditional NSAID plus a PPI if the absolute risk of a gastrointestinal event warrants gastroprotection.
The management of NSAID related ulcer disease must be individualized according to the patient's risk factors for gastrointestinal complications and cardiovascular disease, as well as the efficacy and tolerability of both the NSAID and gastroprotective co-therapy.
As discussed elsewhere in this supplement , the concomitant use of aspirin, particularly in an older population at generally higher cardiovascular risk, is an important factor to consider. Although it is clear that COX-2 inhibitors reduce the risk for gastrointestinal events, use of concomitant aspirin negates the gastrointestinal benefits of COX-2 inhibitors. Moreover, use of certain NSAIDs such as ibuprofen may interfere with the antiplatelet effect of aspirin. In general, a strategy that carefully balances treatment benefits, potential for gastrointestinal complications, and cardiovascular risk is warranted in all patients who receive NSAIDs.
Finally, given the widespread use of OTC NSAIDs, careful attention should be paid to educating patients about the gastrointestinal and cardiovascular risks associated with NSAIDs and the potential for undesirable interactions among these medications.
nonsteroidal anti-inflammatory drug
proton pump inhibitor