NSAID-Induced Gastrointestinal Damage

GASTROENTEROLOGY BOARD REVIEW MANUAL STATEMENT OF EDITORIAL PURPOSE The Hospital Physician Gastroenterology Board Review Manual is a study guide for fellows and practicing physicians preparing for board examinations in gastroenterology. Each quarterly manual reviews a topic essential to the current practice of gastroenterology. PUBLISHING STAFF PRESIDENT, GROUP PUBLISHER Bruce M. White EDITORIAL DIRECTOR Debra Dreger SENIOR EDITOR Becky Krumm ASSISTANT EDITOR Rita E. Gould EXECUTIVE VICE PRESIDENT Barbara T. White EXECUTIVE DIRECTOR OF OPERATIONS Jean M. Gaul PRODUCTION DIRECTOR Suzanne S. Banish PRODUCTION ASSISTANT Kathryn K. Johnson ADVERTISING/PROJECT MANAGER Patricia Payne Castle SALES & MARKETING MANAGER Deborah D. Chavis NOTE FROM THE PUBLISHER: This publication has been developed without involvement of or review by the American Board of Internal Medicine. Endorsed by the Association for Hospital Medical Education NSAID-Induced Gastrointestinal Damage Series Editor: Robert M. Craig, MD Professor of Medicine, Division of Gastroenterology, Department of Medicine, The Feinberg School of Medicine, Northwestern University, Chicago, IL Contributor: James M. Scheiman, MD Professor of Medicine, Division of Gastroenterology, University of Michigan Medical Center, Ann Arbor, MI Table of Contents Introduction............................... 2 Assessing Risk.............................. 2 Pathogenesis............................... 4 Management............................... 5 Preventive Strategies......................... 6 NSAID-Induced Small Bowel and Colonic Disease.................................. 10 References............................... 10 Cover Illustration by Scott Holladay Copyright 2004, Turner White Communications, Inc., 125 Strafford Avenue, Suite 220, Wayne, PA 19087-3391, www.turner-white.com. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, mechanical, electronic, photocopying, recording, or otherwise, without the prior written permission of Turner White Communications, Inc. The editors are solely responsible for selecting content. Although the editors take great care to ensure accuracy, Turner White Communications, Inc., will not be liable for any errors of omission or inaccuracies in this publication. Opinions expressed are those of the authors and do not necessarily reflect those of Turner White Communications, Inc. Gastroenterology Volume 10, Part 2 1

GASTROENTEROLOGY BOARD REVIEW MANUAL NSAID-Induced Gastrointestinal Damage James M. Scheiman, MD INTRODUCTION CASE PRESENTATION It s 5:30 AM, and you ve just left the intensive care unit following successful endoscopic therapy to control a bleeding gastric ulcer in a 55-year-old man. He was generally healthy, but he was taking one of those new, safer, NSAIDs for knee osteoarthritis. He also had been taking an enteric-coated aspirin at night, because all his friends take one, too. He takes no other medications. At the time of the endoscopy, biopsies were obtained for Helicobacter pylori status. EPIDEMIOLOGY OF NSAID-INDUCED ULCERS Each year, more than 40 billion aspirin tablets are consumed and more than 110 million prescriptions for nonsteroidal anti-inflammatory drugs (NSAIDs) are filled in the United States. At least one third of the prescriptions are for cyclooxygenase (COX)-2 specific inhibitors (coxibs) at a cost of $5 billion dollars/yearly. 1 These drugs can have significant adverse gastrointestinal (GI) effects ranging from symptoms such as nausea and dyspepsia (ie, persistent pain or discomfort in the upper abdomen) to serious ulcer complications such as bleeding and perforation. Among chronic NSAID users (defined as daily use for more than 1 year), the risk of developing a symptomatic ulcer, gastrointestinal bleed, or perforation is 1% to 4%. 2 Although there is a low probability that any individual NSAID user will experience a drug-related complication, the huge number of chronic users translates NSAID toxicity into a major health care problem. It is estimated that more than 100,000 hospitalizations and up to 10,000 to 20,000 deaths each year in the United States can be attributed to NSAID complications. 2 There is no consensus among clinicians on how best to weigh the potential clinical benefits of various antiinflammatory agents against the possibility of adverse events associated with their use. Choosing between a generic traditional NSAID and the safer but more costly medications in the COX-2 specific inhibitor class requires a framework for informed decision making. Considerations include the patient s risk for toxicity, the need for concurrent aspirin therapy, and whether the patient is already taking a gastroprotective agent such as a proton pump inhibitor (PPI). Patient risk factors should drive clinical decision making to target these medications in cost-effective manner. ASSESSING RISK NSAID users have an approximately 3-fold elevated risk of an ulcer complication (ie, bleeding, perforation, or a clinical event requiring hospitalization or causing death) compared to nonusers. Although 1% to 4% of chronic NSAID users develop an adverse event, individual risk varies considerably across the population. When chronic NSAID therapy is necessary, it is essential to assess a patient s individual risk for complications. Recognized risk factors (in order of their relative importance) are listed in Table 1. RECENT OUTCOME STUDIES Until the recent CLASS 3 and VIGOR 4 outcome studies were completed evaluating the safety of coxibs (celecoxib and rofecoxib, respectively) compared to traditional NSAIDs, the only prospective data regarding the risk of serious complications due to NSAIDs was from the MUCOSA 5 trial, which compared the outcome of rheumatoid arthritis (RA) patients taking non-aspirin NSAIDs plus either misoprostol or placebo. 0.95% of patients on non-aspirin NSAIDs plus placebo developed serious GI complications within 6 months as compared to 0.57% of patients on NSAIDs plus misoprostol. These findings are consistent with the frequently quoted 2% to 4% risk included in the prescribing information for prescription NSAIDs. The CLASS and VIGOR studies confirmed these rates approximately 2% annually for complicated ulcers, and 4% annually for symptomatic ulcers. 3,4 RISK FACTORS Patients with a history of ulcer complications and those taking concomitant anticoagulant therapy have the highest risk of developing NSAID-associated serious GI complications. Moderate risk factors include advanced age, 2 Hospital Physician Board Review Manual

concomitant corticosteroid use, underlying major organ impairment, the use of high-dose or multiple NSAIDs, and arthritis-related disability. Gender 6 and a history of GI symptoms 7 do not appear to predict increased risk. The past occurrence of an NSAID-associated serious GI complication is unequivocally associated with a markedly increased likelihood of another complication with recurrent NSAID use. Unsurprisingly, the concomitant use of traditional NSAIDs and anticoagulants markedly exacerbates the risk of GI bleeding and results in a 2.2-fold increased risk of hospitalization. 8 Although corticosteroids alone do not increase the risk of peptic ulcer disease, their use with NSAIDs leads to nearly a 2-fold increase in risk of complications, and an even greater increase in the risk of death. Given the increasing use of aspirin to prevent cardiovascular disease and colorectal cancer, attention to the toxicity of low-dose aspirin is particularly relevant. When aspirin is added to an ongoing NSAID treatment regimen, the risk of toxicity rises substantially. Doses of aspirin as low as 10 mg daily deplete prostaglandins and can cause acute GI toxicity. 9 Aspirin doses as low as 75 mg daily have been associated with a significantly increased risk of bleeding gastric and duodenal ulcers in epidemiologic studies from both the United States and the United Kingdom. 10 The risk of upper GI bleeding with low-dose aspirin in these studies was increased 3-fold, and enteric coating and buffering provided no protection from bleeding. These findings support the notion that the costs and benefits of prophylactic aspirin use should be carefully weighed in every patient. Physicians should warn NSAID users that aspirin use increases the risk of ulcers and should encourage these patients to consult a health care provider if they are considering taking low-dose aspirin for cardiovascular health. 11 For patients who are concerned about their cardiovascular risk, such as the one presented here, a risk calculator designed for patient use is available at the National Cholesterol Education Program Web site (http://hin. nhlbi.nih.gov/atpiii/calculator.asp). Increasing age is an independent predictor of experiencing an NSAID-associated GI complication. 12 Clinically significant major organ impairment was identified as an independent risk factor in the MUCOSA trial. Patients with cardiovascular disease have been shown to be at a nearly 2-fold increased risk for GI complications due to NSAID therapy, independent of aspirin use. Symptoms, or the lack thereof, are not good predictors of NSAID complication risk. In one study, 58% of patients admitted with an NSAID complication had no antecedent dyspeptic symptoms compared with the presence of symptoms in 75% of those with non-nsaid complicated ulcers. 7 Table 1. Risk Factors for Aspirin- and NSAID- Associated Ulcer Complications* Personal history of complicated ulcer disease Concurrent use of more than 1 NSAID (including aspirin) Use of high doses of NSAIDs Concurrent use of an anticoagulant Personal history of uncomplicated peptic ulcer disease Age greater than 70 years Concurrent use of steroids NSAID = nonsteroidal anti-inflammatory drug. *Listed in order of relative importance. Meta-analyses have assessed the effect of different types and dosages of NSAIDs on serious GI complications, using ibuprofen as the reference medication. NSAIDs with increasing potency and duration of COX-1 inhibition were associated with increasing risks of serious GI complications. The relative risk was greater than 2-fold higher with high- versus low-dose NSAID therapy. 13 The data indicated a trend for ibuprofen being less likely to cause serious complications than naproxen or indomethacin, likely due to the lower dosages of ibuprofen frequently used by patients. 13 These results likely explain ibuprofen s apparent safety when used appropriately at over-the-counter dosages. Several studies have evaluated agents that bypass gastric absorption (eg, salsalate, nabumetone) or agents that are less potent COX-1 inhibitors (eg, etodolac, nabumetone, meloxicam). 14 These trials demonstrated a significant reduction in endoscopic gastric ulcers and erosions with salsalate, etodolac, and nabumetone. Meloxicam and nabumetone also have been associated with a low rate of symptomatic ulcers in analyses of their respective clinical trial programs. 14 However, the results of these studies should be interpreted cautiously for several reasons, including variability in the relationship of NSAID-associated endoscopic damage to clinical events. Most importantly, these agents have not been subjected to large-scale outcome trials designed to examine the incidence of serious GI complications (eg, bleeding, perforation, hospitalization, or death). The COX-2 specific inhibitors were developed with the goal of reducing ulcers and complications by sparing inhibition of COX-1, which mediates prostaglandin levels. At usual clinical doses, these agents do not inhibit COX-1. As predicted, these drugs, even at very high doses, do not significantly reduce upper GI prostaglandin levels. 15 Clinical trials have demonstrated equivalent anti-inflammatory efficacy of COX-2 specific inhibitors to commonly used Gastroenterology Volume 10, Part 2 3

NSAIDs in arthritis patients. 16 When studied with endoscopy, both celecoxib and rofecoxib produce rates of injury nearly equivalent to placebo. 17 Rofecoxib also has been shown to be equivalent to placebo in fecal blood loss studies in humans. The reduction in serious GI complications observed in the CLASS and VIGOR clinical outcome trials is discussed later in this paper. PATHOGENESIS UPPER GASTROINTESTINAL DEFENSE MECHANISMS In general, the key causes of upper GI injury to consider are NSAIDs and H. pylori infection. The upper GI mucosa has in place a remarkable barrier that, under normal conditions, withstands the hostile acidic environment of the gastric lumen. Indeed, the fundamental question with regard to understanding the development of NSAID-related ulcers is not so much why they occur, but why they do not occur more frequently. The complex elements that defend the gastroduodenal mucosa from damage are largely dependent on endogenous prostaglandins synthesized in the upper gastrointestinal mucosa. COX is the rate-limiting catalytic enzyme in prostaglandin production. At least 2 isoforms of COX have been identified. COX-1 is constitutively expressed in the GI tract, and COX-1 derived prostaglandins play an important role in the maintenance of normal gastric and duodenal physiology. COX-2 is an inducible form that is upregulated in areas of injury, including H. pylori gastritis. 18 However, COX-2 also is regulated in response to physiologic stimuli in other tissues, including the kidney, brain, and reproductive tract. The nonselective NSAIDs nonspecifically inhibit COX isoforms, leading to both beneficial (anti-inflammatory) and toxic (GI bleeding) outcomes. The stomach and duodenum are covered by a mucusbicarbonate barrier that is regulated by prostaglandins and provides the primary defense against the acidic gastric lumen. The surface epithelium provides the second line of gastroduodenal defense. Regeneration, the process by which larger epithelial defects (eg, ulcers) heal, requires cellular proliferation, which is at least partly dependent upon prostaglandins and growth factors. 19 Although very little COX-2 is present in the intact stomach, prostaglandins derived from COX-2 induced in the damaged stomach play an important role in ulcer healing, in part related to angiogenesis stimulated by growth factors. 20 Another key factor preventing mucosal injury is maintenance of microvascular blood flow, also regulated by COX-1 derived prostaglandins. NSAID MECHANISMS OF INJURY Although the mechanisms by which NSAIDs cause ulcers remain incompletely understood, they are known to involve both topical injury and systemic effects mediated by depletion of endogenous prostaglandins. Analogous to H. pylori associated ulcer disease (ie, a high exposure risk coupled with a low absolute risk), the biologic basis of the factors that place certain individuals at an increased risk for NSAID-related ulcers remains unknown. Local Effects Aspirin and most NSAIDs undergo ion trapping within the proximal GI mucosa, causing direct cellular injury. NSAIDs also directly attenuate the hydrophobic, or non-wettable, properties of the mucous barrier independent of prostaglandin-mediated actions. 21 Although topical effects can largely be prevented by administering enteric-coated NSAID formulations or prodrugs, the failure of these approaches to reduce the incidence of symptomatic NSAID-induced ulcers demonstrates that topical injury is not the critical determinant of NSAID-induced injury. For example, parenteral administration of an NSAID such as ketorolac may lead to ulcer complications. Certain NSAIDs, including indomethacin, piroxicam, oxaprozin, and ketorolac, also undergo an extensive enterohepatic recirculation, resulting in repeated exposure to the GI mucosa and increased toxicity. 19 Systemic Effects The clinically important adverse effects of NSAIDs ulcers with an increased risk of complications appear to be due largely to systemic actions. Nonselective COX inhibition with a resultant decrease in the endogenous prostaglandins that are critical to mucosal defense is thought to be the most important mechanism of action. In addition, platelet COX-1 is inhibited irreversibly by aspirin and for as long as 18 hours by other NSAIDs. The impaired platelet function may potentiate bleeding from both the upper and lower GI tracts. COX-1 inhibition leads not only to quantitative but also to qualitative decreases in mucous barrier function. Because prostaglandin deficiency impairs regenerative responses, erosions created by direct topical injury are exposed to acid with the mucosa in a vulnerable state. Ulcerations tend to occur in areas of decreased blood flow; NSAIDs induce microvascular ischemia, in part, by promoting neutrophil adherence in the microcirculation, in addition to reduced prostaglandin production. Nitric oxide (NO) has a role in the maintenance of epithelial integrity that is related to its ability to maintain mucosal blood flow. In animal models, inhibition of NO 4 Hospital Physician Board Review Manual

synthesis exacerbates NSAID injury and thus NO donors reduce NSAID toxicity. 22 Nitric oxide releasing aspirin and NSAID derivatives, currently not available in clinical practice, cause reduced damage despite marked inhibition of prostaglandin levels. This strategy may be of greatest value when aspirin therapy is needed because it would allow the utilization of a non-ulcerogenic antiplatelet agent; clinical trials supporting the animal data are anxiously awaited. The pathophysiology that formed the mechanistic basis for the COX-2 hypothesis is that gastroduodenal mucosal prostaglandin deficiency initiates ulceration, and virtually all mucosal prostaglandins are derived from COX-1. Recent animal data, however, challenge this notion and indicate that for gastric ulceration to occur, both COX-1 and COX-2 must be inhibited. 23 In that study, selective inhibition of COX-1 alone or COX-2 alone did not cause gastric damage, whereas inhibition of both COX isoforms produced considerable gastric ulceration. Thus, the explanation for reduced GI toxicity for COX- 2 specific inhibitors may be their lack of dual COX inhibition rather than simply their COX-1 sparing effects. In this theoretical framework, it may be anticipated that the addition of low-dose aspirin (primarily a COX-1 inhibitor at a low dose) to a COX-2 inhibitor creates an ulcerogenic dual COX inhibitor, as in the case patient. Finally, acid plays an important secondary role in NSAID-induced ulceration. The role of acid is supported by recent data demonstrating the efficacy of PPIs in the treatment and prevention of NSAID damage. 24 These data suggest that topical injury is the first step in NSAID ulceration, with secondary acid injury occurring in the setting of prostaglandin depletion and impaired repair mechanisms, all leading to the development of clinically important ulceration (Figure 1). ROLE OF HELICOBACTER PYLORI INFECTION In the absence of NSAID use, H. pylori is accepted as the cause of most ulcers; its role in the pathogenesis of NSAID-associated ulcers, however, remains controversial. The pathophysiologic mechanisms of NSAID-associated and H. pylori associated ulcers are distinct: NSAID ulcers occur without gastritis (the injury associated with NSAIDs occurs with little to no microscopic inflammation), whereas H. pylori ulcers occur in the setting of diffuse inflammation. The degree of inflammation is likely related to the virulence of the H. pylori strain and host factors. Although NSAIDs decrease prostaglandin synthesis and H. pylori increases the synthesis of prostaglandins, there is little evidence that this increase provides protection from ulceration. Most importantly, it is clear from epidemiologic studies that H. pylori infection is not a required cofactor for NSAID-associated ulcers because these ulcers may occur in the absence of H. pylori. In patients who use NSAIDs chronically, the prevalence of H. pylori infection appears to be similar to those with or without ulcers. 25 Whether eradication of H. pylori protects against NSAID-associated ulcers is another area of controversy. Although a study from Hong Kong found that NSAIDnaive patients who had successful H. pylori eradication had fewer ulcers, follow-up studies in non NSAID-naive patients failed to confirm these results, supporting the independence of these two ulcerogens. 26,27 These same investigators demonstrated that eradication of H. pylori alone was insufficient to prevent recurrence of a bleeding NSAID-associated ulcer in patients with a history of NSAIDassociated ulcer bleeding. 28 In another arm of that study, eradication appeared as effective as omeprazole maintenance therapy for individuals taking low-dose aspirin for cardiovascular protection. 28 A study of similar design from a different group of Hong Kong investigators supported the value of additional prophylaxis with a PPI. However, some of the ulcer recurrences were associated with reinfection, a common problem in endemic areas. 29 In summary, because H. pylori and NSAIDs appear to produce ulcers by different mechanisms, and in the absence of sound evidence suggesting a therapeutic advantage, testing for H. pylori does not appear to be costeffective for all patients starting on NSAID therapy. 30 Patients with a history of peptic ulcer disease, such as the case patient, should be tested for H. pylori and treated with antibiotics if the test is positive to reduce recurrence of H. pylori associated ulcers. Eradication of the infection does not sufficiently reduce NSAID ulcer risk to preclude the need for additional therapies for reducing risk. MANAGEMENT ALTERNATIVES FOR PAIN RELIEF The traditional approach to a chronic NSAID user who bleeds from an ulcer is to stop the NSAID until one would expect the ulcer to be healed (typically 8 weeks). Because patients taking NSAIDs for chronic pain may suffer functional losses from increased pain when the NSAIDs are stopped, an alternative pain reliever may be desirable. The typical choice for pain relief in a patient off NSAID therapy is high doses of acetaminophen (not to exceed 4000 mg daily, in divided doses), assuming that heavy alcohol use or impaired liver or renal function does not make this therapy hazardous. Small doses of a short-acting narcotic medication (eg, propoxyphene, codeine, hydrocodone) may be used temporarily for more severe pain. The Gastroenterology Volume 10, Part 2 5

NSAID Epithelial injury Direct effects Microvascular injury Increased adhesion molecule expression Neutrophil adherence Stasis Microvascular ischemia Prostaglandinmediated effects Mucin Surface active phospholipids Bicarbonate secretion Mucosal proliferation Figure 1. Pathogenesis of NSAID-induced ulcers. NSAID = nonsteroidal antiinflammatory drug. (Adapted from Scheiman JM. NSAIDs, gastrointestinal injury, and cytoprotection. Gastroenterol Clin North Am 1996;25:279, with permission from Elsevier.) Free radical formation Chronic ulcer Erosion non-narcotic analgesic tramadol also may be tried. Other approaches to pain management, such as topical medications, intra-articular corticosteroids or hyaluronic acid injection, can be used concurrently with oral analgesics. ULCER HEALING Several anti-ulcer therapies may be used for healing NSAID-induced ulcers. When NSAIDs are continued, H 2 receptor antagonists (H 2 RAs) have impaired effectiveness for healing ulcers compared to when the NSAID is discontinued. In a study of ulcer patients treated with ranitidine 150 mg twice daily, 63% of gastric ulcer patients who continued NSAID use were healed at 8 weeks compared with 95% of those who discontinued NSAIDs. Duodenal ulcers healed in 84% of patients continuing NSAIDs and in 100% of those who discontinued NSAIDs. 31 PPIs are superior to either H 2 RAs or misoprostol for healing NSAID ulcers in the setting of continued NSAID use. In a study of patients randomized to omeprazole 20 mg daily, omeprazole 40 mg daily, or ranitidine 150 mg twice daily, the rates of healing at 8 weeks were higher in those treated with omeprazole compared to those treated with ranitidine. The higher PPI dose was not superior to the lower dose. 32 Similar data exist for other PPIs. 33 Omeprazole 20 or 40 mg once daily for 8 weeks also has been demonstrated to be significantly more effective than misoprostol for gastric ulcer healing. 24 When patients can discontinue NSAIDs, all forms of anti-ulcer therapy are effective. If the patient has a large or a complicated ulcer, many clinicians use a PPI once or twice daily, based upon evidence that larger ulcers heal faster with more potent acid suppression. In a patient with an uncomplicated NSAID-related ulcer who is able to discontinue the NSAID, any anti-ulcer therapy will be effective. It is the standard of care that all patients with peptic ulcer disease, whether taking NSAIDs or not, undergo testing for and treatment of H. pylori infection. Literature suggests that because healing occurs more quickly with PPIs when the organism is present due to improved acid suppression, it is reasonable to defer eradication until after a course of therapy to heal the ulcer. 27 This may improve compliance as it can be done electively in the outpatient setting after the bleeding hospitalization. PREVENTIVE STRATEGIES At present, data from rigorously designed clinical and economic trials are not available to make accurate 6 Hospital Physician Board Review Manual

head-to-head comparisons between different strategies to prevent ulcers and complications. PROSTAGLANDIN REPLACEMENT Prostaglandin depletion is central to the development of NSAID-induced ulcers, and replacement therapy with a synthetic prostaglandin reduces NSAID toxicity. Placebo-controlled studies have demonstrated the efficacy of misoprostol for the prevention of endoscopic ulcers in NSAID-using arthritis patients. 34 Misoprostol is the only agent currently approved by the US Food and Drug Administration (FDA) for the prevention of NSAIDassociated serious GI complications (eg, bleeding, perforation, obstruction). This approval is based upon the results of the MUCOSA trial, 5 in which a 40% relative risk reduction was observed. Misoprostol appears to be costeffective in high-risk patients only, as the number needed to treat is 264 for 6 months to prevent a single definite upper GI complication. 35 In certain subgroups with an increased risk of serious upper GI complications, the use of misoprostol would be more cost-effective and associated with a lower number needed to treat. 36 Those who benefited the most from misoprostol were patients with a previous history of GI bleed (50% risk reduction), history of previous peptic ulcer disease (52% risk reduction), significant cardiovascular disease (38% risk reduction), significant functional disability (87% risk reduction), and those who required concomitant antacid use (48% risk reduction). 36 Despite this agent being the only FDA-approved regimen for prevention of NSAID ulcers and their complications, it rarely is used due to adverse effects of diarrhea and cramps, a consequence of the prostaglandin analogue on small bowel motility and secretion. ACID SUPPRESSION Proton Pump Inhibitors Omeprazole proved effective in primary prophylaxis of NSAID-induced ulcers compared to placebo in a study of 169 patients requiring continuous NSAID therapy. 37 After 6 months, 78% of the omeprazole group remained in remission compared with 53% in the placebo group (P = 0.004). There were 3 gastric ulcers and no duodenal ulcers in the omeprazole group compared with 8 and 3, respectively, in the placebo group. In the ASTRONAUT study, 32 patients requiring continuous NSAID therapy were randomized following ulcer healing to ranitidine 150 mg twice daily or omeprazole 20 mg daily. Gastric ulcers recurred in 5.2% of the omeprazole-treated patients versus 16.3% of the ranitidine-treated group (P < 0.001). For duodenal ulcers, there was a 0.5% versus a 4.2% (P = 0.02) rate of recurrence in the two groups, respectively. In the OMNIUM study, 24 omeprazole 20 mg once daily was compared to misoprostol 200 µg twice daily and placebo over 6 months in a cohort of patients with recently healed ulcers. In this study, 32% of patients taking placebo developed a gastric ulcer relapse compared to 10% in the misoprostol group and 13% in the omeprazole group. Duodenal ulcers relapsed in 12% of those given placebos, 10% of those given misoprostol, and 3% of those given omeprazole. Omeprazole also was superior in the maintenance of overall remission, largely because of its ability to improve NSAID-associated dyspepsia and overall quality of life. In a randomized, double-blind, placebo-controlled, prospective study of 458 long-term NSAID users with endoscopically documented, H. pylori negative, recently healed gastric ulcers, Graham et al 38 found that patients in the lansoprazole 15- or 30-mg groups remained free from gastric ulcer significantly longer than those who received placebo. Furthermore, whereas the misoprostol-treated group remained free of gastric ulcer longer than the lansoprazole groups at either the 15-mg or 30-mg dose, 31% in the misoprostol group prematurely discontinued use of the study medication due to reported adverse drugrelated events, compared with only 10% and 7%, respectively, in the lansoprazole groups. Recently completed studies with esomeprazole 20 mg and 40 mg have demonstrated the efficacy of this agent to reduce NSAID dyspepsia and associated reductions in quality of life. In studies of H. pylori negative patients taking NSAIDs who were at increased risk of developing ulcers (age > 60 years or recent gastric ulcer or duodenal ulcer), there were greater than 50% reductions in ulcer recurrence. Strikingly, among coxib users taking concomitant esomeprazole therapy, very few ulcers were seen, whereas coxib users taking concomitant placebo experienced similar rates of ulcer recurrence as those taking traditional NSAIDs. 39 This finding emphasizes the value of using PPI cotherapy even when taking coxibs. H 2 Receptor Antagonists The level of acid suppression provided by traditional doses of H 2 RAs does not prevent most NSAID ulcers; only ulcer formation in the duodenum is significantly reduced. In a single study, however, when H 2 RAs were given at double the traditional dose (twice daily instead of once daily), they were effective at reducing both gastric and duodenal ulcers. 40 There are no studies comparing high doses of H 2 blockers to misoprostol or PPIs for the prevention of NSAID ulcers. Practically speaking, given the compliance concerns with twice-daily dosing, PPI therapy is generally preferable to H 2 RAs in this clinical setting. Gastroenterology Volume 10, Part 2 7

Table 2. Incidence of GI Events in the VIGOR Study Rate (per 100 patient-yr) Rofecoxib Naproxen Relative Risk Relative Risk Event (n = 4047) (n = 4029) (95% CI) Reduction, % P Value Clinical upper GI Event 2.1 4.5 0.46 (0.33 0.64) 54 < 0.001 Complicated upper GI Event 0.6 1.4 0.43 (0.24 0.78) 57 0.005 Any GI bleeding 1.2 3.0 0.38 (0.25 0.57) 62 < 0.001 Data from Bombardier C, Laine L, Reicin A, et al. Comparison of upper gastrointestinal toxicity of rofecoxib and naproxen in patients with rheumatoid arthritis. VIGOR Study Group. N Engl J Med 2000;343:1520 8. CI = confidence interval; GI = gastrointestinal. COX-2 SPECIFIC ANTI-INFLAMMATORY DRUGS Coxibs versus Nonselective NSAIDs Deeks and colleagues 41 recently reviewed the results of 4 trials that compared celecoxib with other NSAIDs on the incidence of ulcers detected by endoscopy in patients not receiving aspirin and those receiving aspirin at doses up to 325 mg/day. While the reduction in ulcer risk was significant with celecoxib versus other NSAIDs among those individuals not receiving aspirin (relative risk = 0.27; P = 0.001), no significant difference in ulcer risk was observed between those individuals receiving aspirin and celecoxib and those receiving aspirin with another NSAID (relative risk = 0.49; P = 0.54). Two 6-month, placebo-controlled, endoscopy studies have been performed with rofecoxib 25 mg and 50 mg compared to ibuprofen 800 mg 3 times daily. In the 2 studies, a total of 1517 patients were randomized. In a study of 742 patients, 42 at 12 weeks, 7.3% of patients on placebo developed an endoscopic ulcer compared to 4.7% on 25 mg of rofecoxib, 8.1% on 50 mg of rofecoxib, and 28.5% in the ibuprofen group. In a second study of identical design, 17 similar safety with rofecoxib was observed. When the two studies were combined, the incidence of endoscopic ulcers in patients taking rofecoxib met predefined criteria for equivalence to placebo. 17 However, similar to the combination of aspirin with celecoxib, 41 a recent endoscopic study with rofecoxib found that the benefits of the coxib were lost with the addition of aspirin. 43 When evaluated by endoscopy over a 12-week study period, there was no difference in the cumulative incidence of endoscopic gastroduodenal ulcers in patients taking low-dose (81 mg) enteric-coated aspirin with rofecoxib 25 mg as compared to those taking ibuprofen 2400 mg daily. Because the relationship between endoscopic injury and serious GI complication rates among NSAID users remains controversial, manufacturers of both leading coxibs performed long-term safety trials. In the VIGOR study, the GI safety of rofecoxib 50 mg daily was compared to naproxen 500 mg twice daily in more than 7000 rheumatoid arthritis patients worldwide. 4 The primary and secondary endpoints of this study were the occurrence of symptomatic and complicated upper GI events, respectively. The results of the trial are shown in Table 2. The coxib was associated with a 50% to 60% reduction in ulcers, bleeding, or both. In the CLASS study, 8000 patients worldwide with osteoarthritis (90%) or rheumatoid arthritis (10%) were randomized to celecoxib 400 mg twice daily, Ibuprofen 800 mg 3 times daily, or diclofenac 75 mg twice daily for up to 12 months. 3 In this trial, which included patients taking low-dose (325 mg or less) aspirin for cardiovascular prophylaxis, the primary endpoint was the development of complicated upper GI events, with symptomatic ulcer development as the secondary endpoint. The primary endpoint of ulcer complications did not reach significance in the CLASS trial (Figure 2); this was largely attributable to the impact of low-dose aspirin (Figure 3). When the entire 12 months of the trial became available, the primary endpoint did not reach significance even when those taking aspirin were excluded, likely reflecting numerous methodologic problems with the design of the trial, including the high drop-out rate that was observed in the study. 44 Both endoscopic and clinical outcome studies have demonstrated that for those patients receiving coxibs who have a need for an antiplatelet agent and also must take low-dose aspirin, the coxib protection from ulceration is markedly impaired. Coxibs versus an NSAID with a PPI In a trial from Hong Kong, Chan et al 45 randomized patients with a history of ulcer bleeding to either a PPI (omeprazole 20 mg daily) plus a traditional NSAID (diclofenac 75 mg twice daily) or celecoxib 200 mg twice daily for 6 months. All patients had H. pylori eradicated or were uninfected. In this large study, no significant 8 Hospital Physician Board Review Manual

Annualized incidence (%) 4 3 2 1 0 Celecoxib Conventional NSAIDs 0.76 P = 0.09 1.45 Complications 2.08 P = 0.02 3.54 Complications and symptomatic ulcers Figure 2. Incidence of symptomatic upper gastrointestinal ulcers and ulcer complications in the CLASS trial at 6 months (all patients). (Adapted with permission from Silverstein FE, Faich G, Goldstein JL, et al. Gastrointestinal toxicity with celecoxib vs nonsteroidal antiinflammatory drugs for osteoarthritis and rheumatoid arthritis: the CLASS study. A randomized controlled trial. Celecoxib Long-term Arthritis Safety Study. JAMA 2000;284:1251. Copyright 2000, American Medical Association. All rights reserved.) difference was found in the number of patients presenting with an ulcer bleed: 4.9% on the coxib and 6.4% on diclofenac with the PPI. Although the study was not powered to demonstrate equivalence, the results suggest both that the strategies are similar in efficacy and that neither approach is sufficient in this high-risk group of patients. Most experts recommend that these highest-risk patients receive cotherapy with a PPI (or misoprostol) in addition to a coxib to provide multiple risk-reducing strategies. Formulating a Cost-Effective Preventive Strategy Key components of the overall cost of disease management with NSAIDs include physician visits, medication expenditure, and endoscopies related to the development of GI side effects such as dyspepsia. Coxibs appear to have fewer nuisance GI side effects compared to traditional NSAIDs, although such side effects are more frequent than with placebo. Rofecoxib was noted to have significantly fewer GI symptoms and need for GI medication cotherapy in clinical trials of up to 6 months duration. 46 In general, the incidence of these adverse events and rates of anti-inflammatory drug discontinuation with rofecoxib were intermediate between those of traditional NSAIDs and those of placebo. In a trial specifically designed to address the question of GI tolerability, the rates of cumulative discontinuation due to GI adverse events were statistically significantly lower in the rofecoxib group than in the naproxen group (5.9% versus 8.1%), as were rates of cumulative use of medication to treat GI symptoms Incidence (events/100 patient-yr) 5 4 3 2 1 P = 0.618 Celecoxib NSAIDs P = 0.02 0 Aspirin No aspirin Figure 3. Ulcers and ulcer complications in the CLASS trial at 12 months, according to aspirin use. P value by log rank test. (Adapted from Lefkowith JB. Safety profile of celecoxib. CLASS: Celecoxib Long-Term Arthritis Safety Study. Slide no. 110. Available at www.fda.gov/ohrms/dockets/ac/01/slides/3677s1_01_ sponsor.pdf. Accessed 26 Mar 2004.) (9.1% versus 11.20%). 47 Although the superior GI tolerability was statistically significant, however, it was modest in absolute terms, limiting the economic value of the coxib because of its higher cost. Because coxibs are as effective as traditional NSAIDs but safer, they represent a logical alternative on the basis of both clinical and economic considerations if the traditional NSAID is being given with another medication, such as an H 2 RA or PPI, that can be discontinued if the patient is switched to a coxib. Patients with a history of ulcer complications and concomitant anticoagulant therapy have the highest risk of developing NSAIDassociated serious GI complications. These patients would be expected to derive the greatest overall benefit from the reduced GI risk associated with coxibs. Moderate risk factors include advanced age, corticosteroid use, chronic major organ impairment (particularly cardiovascular disease), and the use of high-dose or multiple NSAIDs. Coxibs in these patients who are at aboveaverage risk are likely to be cost effective, particularly if more than 1 of these risk factors is present. 48 Additional economic benefits will be achieved if additional medications are not required, or if adequate symptom control occurs with the addition of a low-cost generic antisecretory drug to the coxib therapy. Management of the patient requiring low-dose aspirin is more complex. Testing and treatment of H. pylori may prove beneficial in these patients. 28 In high-risk patients, concomitant aspirin therapy may mandate antisecretory cotherapy to prevent bleeding, regardless of whether a coxib is being used. An algorithm for preventing GI complications in patients who require aspirin therapy (with or without anti-inflammatory drugs) is shown in Figure 4. Gastroenterology Volume 10, Part 2 9

Patient requiring low-dose aspirin Anti-inflammatory drug not required Anti-inflammatory drug required Prior history of ulcer or GI bleeding Prior history of ulcer or GI bleeding Yes No No Yes Eradicate H. pylori if present Monitor for signs and symptoms of bleeding No H. pylori testing Eradicate H. pylori if present GI prophylactic medication added based upon ulcer risk independent of coxib or traditional NSAID choice Figure 4. Algorithm for preventing gastrointestinal (GI) complications in patients requiring aspirin therapy. NSAID = nonsteroidal antiinflammatory drug. NSAID-INDUCED SMALL BOWEL AND COLONIC DISEASE Clinically important NSAID-induced GI damage beyond the duodenum has been recognized. Autopsy studies have confirmed a marked increase in nonspecific small bowel ulceration in NSAID users (8.4%) compared to control subjects (0.6%). 49 Inflammation and increased blood and protein loss from the small bowel occur in some chronic NSAID users and is known as NSAID enteropathy. Rare reports suggest that intestinal ulcers can become chronic and lead to strictures. Recent studies utilizing wireless capsule endoscopy have confirmed a high incidence of endoscopic damage in NSAID users. These studies have shown a surprising background prevalence of lesions, which is increased following administration of a traditional NSAID. 50,51 Coxibs were shown to cause far fewer lesions, and cotherapy with PPI did not, as was expected, reduce this damage. The clinical relevance of these lesions remain uncertain; in the analysis of the VIGOR trial, however, bleeding events attributed to the lower GI tract represented approximately one third of the events, with the risk reduction for rofecoxib compared to naproxen similar to that seen in the upper tract. 52 NSAIDs also may have adverse effects on the colon, including nonspecific ulcerations in the cecum and transverse and sigmoid colon. NSAIDs may induce a colitis whose appearance may be confused with inflammatory bowel disease (IBD); however, the biopsy changes are nonspecific and are inconsistent with IBD. Rectal administration of NSAIDs can cause proctitis. In case-control studies, preexisting lesions (eg, diverticula) have been reported to bleed or perforate with increased frequency in NSAID users. NSAIDs also have been implicated in the relapse of quiescent IBD. The pathogenesis of intestinal injury is thought to be due to direct effects of NSAIDs to alter cellular metabolism in enterocyte mitochondria with subsequent loss of cellular integrity. In this hypothesis, cyclooxygenase inhibition plays a secondary role by impairing prostaglandindependent healing mechanisms. Increased permeability due to a damaged barrier then allows luminal aggressive agents (eg, bacteria, bile) to further perpetuate injury. Drugs such as indomethacin or piroxicam, which undergo enterohepatic recirculation, appear to cause the most injury by virtue of their repeated exposure to the mucosa. It has been observed that metronidazole and sulfasalazine, but not misoprostol, can decrease the inflammation and blood loss due to NSAID enteropathy without affecting intestinal permeability. Clinically significant intestinal injury due to NSAIDs is rare, and there is little information to guide patient management. Coxibs would be anticipated to cause less injury in most patients, 53 except those with pre-existing intestinal damage (eg, those with IBD). Although there have been reports of the successful use of coxibs in this patient population, these agents should be used with caution owing to concern of exacerbating the existing disease. REFERENCES 1. Laine L. Approaches to nonsteroidal anti-inflammatory drug use in the high-risk patient. Gastroenterology 10 Hospital Physician Board Review Manual

2001;120:594 606. 2. Wolfe MM, Lichtenstein DR, Singh G. Gastrointestinal toxicity of nonsteroidal antiinflammatory drugs [published erratum appears in N Engl J Med 1999;341:548]. N Engl J Med 1999;340:1888 99. 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. Celecoxib Long-term Arthritis Safety Study. JAMA 2000;284:1247 55. 4. Bombardier C, Laine L, Reicin A, et al. Comparison of upper gastrointestinal toxicity of rofecoxib and naproxen in patients with rheumatoid arthritis. VIGOR Study Group. N Engl J Med 2000;343:1520 8. 5. Silverstein FE, Graham DY, Senior JR, et al. Misoprostol reduces serious gastrointestinal complications in patients with rheumatoid arthritis receiving nonsteroidal antiinflammatory drugs. A randomized, double-blind, placebo-controlled trial. Ann Intern Med 1995;123:241 9. 6. Gabriel SE, Jaakkimainen L, Bombardier C. Risk for serious gastrointestinal complications related to use of nonsteroidal anti-inflammatory drugs. A meta-analysis. Ann Intern Med 1991;115:787 96. 7. Armstrong CP, Blower AL. Non-steroidal anti-inflammatory drugs and life threatening complications of peptic ulceration. Gut 1987;28:527 32. 8. Shorr RI, Ray WA, Daugherty JR, Griffin MR. Concurrent use of nonsteroidal anti-inflammatory drugs and oral anticoagulants places elderly persons at high risk for hemorrhagic peptic ulcer disease. Arch Intern Med 1993; 153:1665 70. 9. Cryer B, Feldman M. Effects of very low dose daily, longterm aspirin therapy on gastric, duodenal, and rectal prostaglandin levels and on mucosal injury in healthy humans. Gastroenterology 1999;117:17 25. 10. U.S. Preventative Services Task Force. Aspirin for the primary prevention of cardiovascular events: recommendation and rationale. Ann Intern Med 2002;136:157 60. 11. Hayden M, Pignone M, Phillips C, Mulrow C. Aspirin for the primary prevention of cardiovascular events: a summary of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med 2002;136:161 72. 12. Griffin MR, Piper JM, Daugherty JR, et al. Nonsteroidal anti-inflammatory drug use and increased risk for peptic ulcer disease in elderly persons. Ann Intern Med 1991; 114:257 63. 13. Henry D, Lim LL, Garcia Rodriguez LA, et al. Variability in risk of gastrointestinal complications with individual non-steroidal anti-inflammatory drugs: results of a collaborative meta-analysis. BMJ 1996;312:1563 66. 14. Delvalle J, Chey WD, Scheiman JM. Acid peptic disorders. In: Textbook of gastroenterology, 4th ed. Yamada T, editor. Philadelphia: Lippincott; 2003:1321 76. 15. Scheiman JM, Greenson JK, Lee J, Cryer B. Effect of cyclooxygenase-2 inhibition on human Helicobacter pylori gastritis: mechanisms underlying gastrointestinal safety and implications for cancer chemoprevention. Aliment Pharmacol Ther 2003;17:1535 44. 16. Feldman M, McMahon AT. Do cyclooxygenase-2 inhibitors provide benefits similar to those of traditional nonsteroidal anti-inflammatory drugs with less gastrointestinal toxicity [published erratum Ann Intern Med 2000;132:1011]? Ann Intern Med 2000;132:134 43. 17. Hawkey C, Laine L, Simon T, et al. Comparison of the effect of rofecoxib (a cyclooxygenase 2 inhibitor), ibuprofen, and placebo on the gastroduodenal mucosa of patients with osteoarthritis: a randomized, double-blind, placebo-controlled trial. The Rofecoxib Osteoarthritis Endoscopy Multinational Study Group. Arthritis Rheum 2000;43:370 7. 18. McCarthy CJ, Crofford LJ, Greenson J, Scheiman JM. Cyclooxygenase-2 expression in gastric antral mucosa before and after eradication of Helicobacter pylori infection. Am J Gastroenterol 1999;94:1218 23. 19. Scheiman JM. NSAIDs, gastrointestinal injury, and cytoprotection. Gastroenterol Clin North Am 1996;25:279 98. 20. Jones MK, Wang H, Peskar BM, et al. Inhibition of angiogenesis by nonsteroidal anti-inflammatory drugs: insight into mechanism and implications for cancer growth and ulcer healing. Nat Med 1999;5:1418 23. 21. Lichtenberger LM, Wang ZM, Romero JJ, et al. Nonsteroidal anti-inflammatory drugs (NSAIDs) associate with zwitterionic phospholipids: insight into the mechanism and reversal of NSAID-induced gastrointestinal injury. Nat Med 1995;1:154 8. 22. Wallace JL, Reuter B, Cicala C, et al. Novel nonsteroidal anti-inflammatory drug derivatives with markedly reduced ulcerogenic properties in the rat. Gastroenterology 1994;107:173 9. 23. Wallace JL, McKnight W, Reuter BK, Vergnolle N. NSAIDinduced gastric damage in rats: requirement for inhibition of both cyclooxygenase 1 and 2. Gastroenterology 2000; 119:706 14. 24. Hawkey CJ, Karrasch JA, Szczepanski L, et al. Omeprazole compared with misoprostol for ulcers associated with nonsteroidal antiinflammatory drugs. Omeprazole versus Misoprostol for NSAID-induced Ulcer Management (OMNIUM) Study Group. N Engl J Med 1998; 338:727 34. 25. Chan FK. COX-2 inhibition, H. pylori infection and the risk of gastrointestinal complications. Curr Pharm Des 2003;9:2213 9. 26. Chan FK, Sung JJ, Chung SC, et al. Randomized trial of eradication of Helicobacter pylori before non-steroidal antiinflammatory drug therapy to prevent peptic ulcers. Lancet 1997;350:975 9. 27. Hawkey CJ, Tulassay Z, Szczepanski L, et al. Randomized controlled trial of Helicobacter pylori eradication in patients on non-steroidal anti-inflammatory drugs: HELP NSAIDs study. Helicobacter Eradication for Lesion Prevention [published erratum appears in Lancet 1998;352: 1634]. Lancet 1998;352:1016 21. 28. Chan FK, Chung SC, Suen BY, et al. Preventing recurrent upper gastrointestinal bleeding in patients with Gastroenterology Volume 10, Part 2 11

Helicobacter pylori infection who are taking low-dose aspirin or naproxen. N Engl J Med 2001;344:967 73. 29. Lai KC, Lam SK, Chu KM, et al. Lansoprazole for the prevention of recurrences of ulcer complications from longterm low-dose aspirin use. N Engl J Med 2002;346:2033 8. 30. Scheiman JM, Bandekar RR, Chernew ME, Fendrick AM. Helicobacter pylori screening for individuals requiring chronic NSAID therapy: a decision analysis. Aliment Pharmacol Ther 2001;15:63 71. 31. Lancaster-Smith MJ, Jaderberg ME, Jackson DA. Ranitidine in the treatment of non-steroidal anti-inflammatory drug associated gastric and duodenal ulcers. Gut 1991;32: 252 5. 32. Yeomans ND, Tulassay Z, Juhasz L, et al. A comparison of omeprazole with ranitidine for ulcers associated with nonsteroidal antiinflammatory drugs. Acid Suppression Trial: Ranitidine versus Omeprazole for NSAID-associated Ulcer Treatment (ASTRONAUT) Study Group. N Engl J Med 1998;338:719 26. 33. Agrawal NM, Campbell DR, Safdi MA, et al. Superiority of lansoprazole vs ranitidine in healing nonsteroidal antiinflammatory drug-associated gastric ulcers: results of a double-blind, randomized, multicenter study. NSAID- Associated Gastric Ulcer Study Group. Arch Intern Med 2000;160:1455 61. 34. Graham DY, White RH, Moreland LW, et al. Duodenal and gastric ulcer prevention with misoprostol in arthritis patients taking NSAIDs. Misoprostol Study Group. Ann Intern Med 1993;119:257 62. 35. Scheiman JM. Meta analysis: Misoprostol reduced NSAID-induced gastrointestinal injury. ACP J Club 1997; 126:36. [Comment on Koch M, Dezi A, Ferrario F, Capurso I. Prevention of nonsteroidal anti-inflammatory drug-induced gastrointestinal mucosal injury. A metaanalysis of randomized controlled clinical trials. Arch Intern Med 1996;156:2321.] 36. Simon LS, Hatoum HT, Bittman RM, et al. Risk factors for serious nonsteroidal-induced gastrointestinal complications: regression analysis of the MUCOSA Trial. Fam Med 1996;28:204 10. 37. Cullen D, Bardhan KD, Eisner M, et al. Primary gastroduodenal prophylaxis with omeprazole for non-steroidal anti-inflammatory drug users. Aliment Pharmacol Ther 1998;12:135 40. 38. Graham DY, Agrawal NM, Campbell DR, et al. Ulcer prevention in long-term users of nonsteroidal anti-inflammatory drugs: results of a double-blind, randomized, multicenter, active- and placebo-controlled study of misoprostol vs lansoprazole. NSAID-Associated Gastric Ulcer Prevention Study Group. Arch Intern Med 2002;162:169 75. 39. Scheiman JM, Yeomans N, Hawkey CJ, et al. Esomeprazole reduces gastric and duodenal ulcer development among high risk NSAID users. Am J Gastroenterol 2003;98:S52. 40. Taha AS, Hudson N, Hawkey CJ, et al. Famotidine for the prevention of gastric and duodenal ulcers caused by nonsteroidal anti-inflammatory drugs. N Engl J Med 1996; 334:1435 9. 41. Deeks JJ, Smith LA, Bradley MD. Efficacy, tolerability, and upper gastrointestinal safety of celecoxib for treatment of osteoarthritis and rheumatoid arthritis: systematic review of randomised controlled trials. BMJ 2002;325:619 26. 42. Laine L, Harper S, Simon T, et al. A randomized trial comparing the effect of rofecoxib, a cyclooxygenase 2-specific inhibitor, with that of ibuprofen on the gastroduodenal mucosa of patients with osteoarthritis. Rofecoxib Osteoarthritis Endoscopy Study Group. Gastroenterology 1999;117:776 83. 43. Vioxx [package insert]. White House (NJ): Merck & Co.; 2003. Available at www.vioxx.com/vioxx/shared/ documents/english/pi.pdf. Accessed 26 Mar 2004. 44. Scheiman JM. Outcomes studies of the gastrointestinal safety of cyclooxygenase-2 inhibitors. Cleve Clin J Med 2002;69 Suppl 1:SI40 6. 45. Chan FK, Hung LC, Suen BY, et al. Celecoxib versus diclofenac and omeprazole in reducing the risk of recurrent ulcer bleeding in patients with arthritis. N Engl J Med 2002;347:2104 10. 46. Watson DJ, Harper SE, Zhou PL, et al. Gastrointestinal tolerability of the selective cyclooxygenase-2 (COX-2) inhibitor rofecoxib compared with nonselective COX-1 and COX-2 inhibitors in osteoarthritis. Arch Intern Med 2000;160: 2998 3003. 47. Lisse JR, Perlman M, Johansson G, et al. Gastrointestinal tolerability and effectiveness of rofecoxib versus naproxen in the treatment of osteoarthritis: a randomized, controlled trial. ADVANTAGE Study Group. Ann Intern Med 2003;139:539 46. 48. Fendrick AM, Bandekar RR, Chernew ME, Scheiman JM. Role of initial NSAID choice and patient risk factors in the prevention of NSAID gastropathy: a decision analysis. Arthritis Rheum 2002:47:36 43. 49. Allison MC, Howatson AG, Torrance CJ, et al. Gastrointestinal damage associated with the use of nonsteroidal antiinflammatory drugs. N Engl J Med 1992;327:749 54. 50. Goldstein J, Eisen G, Lewis B, Gralnek I, et al. Abnormal small bowel findings are common in healthy subjects screened for a multi-center, double blind, randomized, placebo controlled trial using capsule endoscopy. Gastroenterology 2003;124:A37. 51. Goldstein J, Eisen G, Lewis B, Gralnek I, et al. Celecoxib is associated with fewer small bowel lesions than naproxen + omeprazole in healthy subjects as determined by capsule endoscopy. Am J Gastroenterol 2002:98:S297. 52. Laine L, Connors LG, Reicen A, et al, Serious lower gastrointestinal clinical events with nonselective NSAID or coxib use. Gastroenterology 2003;124:288 92. 53. Sigthorsson G, Crane R, Simon T, et al. COX-2 inhibition with rofecoxib does not increase intestinal permeability in healthy subjects: a double blind crossover study comparing rofecoxib with placebo and indomethacin. Gut 2000;47:527 32. Copyright 2004 by Turner White Communications Inc., Wayne, PA. All rights reserved. 12 Hospital Physician Board Review Manual