ANNALS O F CLINICAL AND LABORATORY SCIENCE, Vol. 28, No. 2 Copyright © 1998, Institute for Clinical Science, Inc.

Adverse Effects of Nonsteroidal Anti-inflammatory Drugs on the Gastrointestinal System*

EGIL FOSSLIEN, M.D.

Department of Pathology, University o f Illinois at Chicago, College of Medicine, Chicago, IL 60612

ABSTRACT

Two enzymes, cyclo-oxygenase (COX) and 5-lipoxygenase, act upon arachidonic acids to produce prostaglandins and leukotrienes. Inhibition of COX-2 by non­ steroidal anti-inflammatory drugs (NSAIDs) lowers synthesis of proinflammatory prostaglandins and produces analgesia. COX-2 is highly inducible by endotoxin, IL-1, hypoxia, epidermal growth factor (EGF), benzo[a]pyrene, and transforming growth factor beta 1(TGF-01). COX-1 is constitutively expressed. Conventional NSAIDs also inhibit the synthesis of cytoprotective prostaglandins by COX-1 in the gastro­ intestinal tract. Surplus arachidonic acids accumulate and enhance the generation of leukotrienes via the lipoxygenase pathway inducing neutrophil adhesion to endothe­ lium and vasoconstriction. The NSAIDs harboring a carboxyl group also inhibit oxidative phosphorylation (OXPHOS) lowering adenosine-triphosphate (ATP) generation leading to loss of mucosal cell tight junctions and increased mucosal permeability. Administration of NSAIDs that do not interfere with OXPHOS, and concomitant use of prostaglandin analogues to restore cytoprotection reduces complications of NSAID use. However, no NSAID that lacks potential for serious gastrointestinal toxicity is currently avail­ able. Selective inhibitors of COX-2 and 5-lipoxygenase are newer, promising drugs. Surprisingly, COX-2 null mice are able to mount an inflammatory response, suffer­ ing however, from kidney dysfunction and a shortened life span. Results of clini­ cal studies on the long-term use of NSAID drugs such as selective inhibitors are still pending.

Introduction phlogistic agents.2,3,4 Worldwide, physicians write over 100 million prescriptions for Nonsteroidal anti-inflammatory drugs NSAIDs annually. Additional use arises from (NSAIDs) are effective analgesic1 and anti­ patient self-medication with one or more of a large number of non-prescription drugs con­ taining NSAIDs.5 In the United States over- * Send reprint requests to: Egil Fosslien, M.D., Profes­ the-counter drug use of N S AIDs is 7 times the sor of Pathology, Department of Pathology (M/C847), College of Medicine, University of Illinois at Chicago, Chi­ prescription use. Over 2 billion dollars are cago, IL 60612. spent for NSAIDs annually. The gastrointesti­

6 7 0091-7370/98/0300-0067 $03.75 © Institute for Clinical Science, Inc. 68 FOSSLIEN nal tract remains the most acceptable and reactions and with serious upper gastrointesti­ popular route of administration of NSAIDs. It nal disorders.21 Spontaneous reports of offers the advantage of convenience and adverse effects from diclefenac, nabumetone, results in good absorption of the therapeu­ naproxen, and piroxicam show higher associa­ tic compounds.6 tion of gastrointestinal events with nabume­ Uses include the short-term treatment of tone and piroxicam.22 It has been estimated dental7 or perisurgical pain8 and long term that there is one hospital admission per 2,823 treatment of painful musculoskeletal disorders NSAID prescriptions.23 Endoscopic data and such as rheumatoid arthritis. Aspirin is benefi­ review of clinical histories of NSAID use by cial in the treatment of coronary heart disease, patients within one week of admission showed and prophylactic aspirin use significantly that a history of NSAID use was approximately reduces the risk of myocardial infarction.9,10 equal in patients with upper and lower intes­ Newer studies suggest that certain NSAIDs tinal bleeding and significantly greater than may find further uses. Sulindac, for instance, controls.24 The risk of diverticular bleeding can induce regression in rectal adenomatous was higher than that of duodenal ulcer bleed­ polyps.11 Aspirin and other NSAIDs may ing. The study included 461 patients with prevent or inhibit colonic cancer.12,13 Treat­ upper and 105 with lower GI bleeding and ment with NSAIDs exerts a stronger protec­ 1,895 controls. tive influence than steroids in Alzhei­ One half of the patients taking NSAIDs mer's disease.14,15 regularly have gastric erosion and 10 percent However, between 20 and 30 percent of to 30 percent have gastric ulcers.25 NSAID patients under NSAIDs therapy develop diges­ exposure is more common in gastric than in tive symptoms.16 Adverse effects of NSAIDs prepyloric, pyloric and duodenal perforation involve all segments of the alimentary tract. and accounts for 20 to 35 percent of ulcer For example, patients with esophageal reflux bleedings.26 About 70 percent of a study group taking NSAIDs have an increased risk of of 76 consecutive patients with gastrointestinal esophageal stenosis.16 The more common, (GI) perforations had used aspirin or other serious adverse effects of NSAIDs involve the NSAIDs.27 Twenty percent were lower GI and stomach, duodenum, and the lower gastroin­ 80 percent upper GI perforations.27 Low dose testinal tract. The NSAID-induced gastroin­ aspirin therapy may precipitate myocardial testinal effects range from nausea, vomiting, infarction or unstable angina by inducing dyspepsia, diarrhea, constipation, mucosal irri­ upper gastrointestinal bleeding, especially in tation, erosions to peptic ulcerations and mas­ patients with ischemic heart disease.28 Patients sive, fatal gastrointestinal hemorrhage. Gas- suffering fatal bleeding from gastric or duode­ tropathies consist of erythematous, erosive or nal ulcerations may experience no pain at all ulcerative lesions located mainly antral or pre­ because of the analgesic effect of NSAIDs. pyloric.17,18 The NSAID-induced gastropathy Results of a study on the exposure to may be asymptomatic despite a broad spec­ NSAIDs in 272 patients with bleeding or per­ trum of abnormal endoscopic findings, but in forated peptic ulcer documented by endos­ the majority of cases there are clinical symp­ copy are shown in figure 1. The NSAIDs are toms.19 Most importantly, NSAID exposure is widely used for the treatment of chronic strongly associated with both upper and lower arthropathies, such as rheumatoid arthritis. A gastrointestinal perforations.20 Life threat­ study of 2,400 consecutive patients with rheu­ ening hemorrhage may be the first clini­ matoid arthritis followed prospectively for an cal presentation. average of three and a half years indicated that NSAID gastropathy in patients with rheuma­ Epidemiology and Nosology toid arthritis accounts for at least 20,000 hos­ The NSAID use is associated with hepatic, pitalizations and 2,600 deaths annually in the renal, hematological, and hypersensitivity United States.29 ADVERSE EFFECTS OF NSAIDS 6 9

F igure 1. The NSAID ingestion in patients with bleeding or perforated pep­ tic ulcers. Histogram of data from a retrospective study of 272 patients?10 About one half of gastric ulcer patients and one third of perforated gastric ulcer patients had a history of recent NSAID exposure. In the much larger group of patients with duo­ denal ulcers, NSAID expo­ sure was somewhat higher than in the gastric ulcer group, however, it was much lower as a percentage of the total. Number of patients

Overall, there are more than 20,000 Major risk factors in patients with rheuma­ NSAID-induced fatalities a year in the United toid arthritis include age over 60 years, con­ States.30 In England, NSAID use leads to the comitant therapy with corticosteroids, and premature death of over 3,000 patients annu­ longer duration or larger dose of NSAID treat­ ally.31 Of 65,000 emergency upper gastrointes­ ment.15 Another study shows a relative risk of tinal admissions per annum in the UK, 12,000 gastroduodenal ulceration 4 to 5 times higher are attributable to NSAID use resulting in in the NSAID user.35 Higher doses of NSAIDs 2230 deaths.32 Another 330 deaths attributable and concomitant use of corticosteroids further to NSAID exposure occur in the community. increases the risk. A seven-fold higher risk of In the Nordic countries, 20 to 50 percent of gastric bleeding in elderly patients than in patients using NSAIDs have an ulcer at some younger patients is associated with failure of time, and 1 to 2 percent on continuous therapy normal mucosal adaptation.36 Evidence from a are hospitalized for ulcer complications result­ population-based retrospective case-control ing in 450 deaths annually.33 study of 1,377 cases of upper gastrointestinal A one-year prospective study of the natural bleeding and perforation (UGIB) and 10,000 history of dyspepsia in 545 adult primary care control subjects reveal that age is the most patients indicated that exposure to NSAIDs important predictor of UGIB 3 The risk varies increased the risk for gastrointestinal bleeding widely with the individual drug. Administra­ by a factor of 7.34 The NSAID-associated seri­ tion of azapropazone and piroxicam carries the ous gastrointestinal events in 2,747 patients highest risk of UGIB. Ibuprofen, naproxen, with rheumatoid arthritis and 1,091 patients diclofenac, ketoprofen, and indomethacin has with osteoarthritis3 resulted in an annual hos­ relative risks similar to that for overall NSAID pitalization incidence of 1.58 percent during use. The risk is higher in the female than the NSAID treatment in rheumatoid arthritis male patient.38 patients. The risk of gastrointestinal related The NSAIDs induce or exacerbate damage death was 0.19 percent per year with NSAIDs. of the distal gastrointestinal tracts.39 Seventy The hazard ratio of patients taking NSAIDs to percent of patients receiving long term those not taking NSAIDs was 5:2. NSAID therapy have evidence of inflamma­ 7 0 FOSSLIEN tion of the small intestine.40 Patients may suf­ acid is converted to prostaglandins.48 Two fer from asymptomatic ileal dysfunction, cyclo-oxygenase isoenzymes, referred to as increased mucosal permeability,41 protein and COX-1 and COX-2, are inhibited by conven­ blood loss.42 Occasionally, diaphragm-like tional NSAIDs (figure 2).49 The COX isoforms small intestinal strictures necessitating surgery share structural and enzymatic similarities. may develop. The prevalence of NSAID- Human COX-1 and COX-2 genes consist of induced damage to the large intestine is 599 and 604 amino acids, respectively, but unknown.43 It is associated with diarrhea, exhibit only about 60 percent homology.50 colonic bleeding, anemia, ulcerations and Both are expressed to a similar extent in the strictures, perforations, and death.43,44 Diar­ human stomach. However, in the mouse stom­ rhea occurs in 3 to 5 percent of patients on ach, COX-1 is more abundant. In the rat stom­ NSAIDs. The NSAIDs may trigger exacerba­ ach, COX-1 and COX-2 are found in surface tion of inflammatoiy diseases and can compli­ mucous cells and mucous neck cells, respec­ cate diverticular disease of the large bowel. tively. The COX isoenzymes exhibit cellular Rare cases of segmental ischemic colitis in two compartmentalization: COX-2 is located both patients with no possible etiologic factors other in the nuclear envelope and the endoplasmic than NSAID use have been reported.45 Use of reticulum, COX-1 only in the latter. slow release NSAIDs shifts adverse effects The COX genes exhibit striking differences from the upper to the distal gastrointesti­ in their regulation due to differences in pro­ nal tract.39 moters and transcripts. The constitutively expressed COX-1 gene shows poor inducibility Pathophysiology and is classified as a ‘housekeeping gene.’ However, gastric COX-1 is induced by endo­ The NSAIDs exert their analgesic, antiphlo­ toxin administration in the rat.51 It provides gistic, and antipyretic effects through periph­ prostaglandins for mucosal cytoprotective eral and central inhibition of prostaglandin functions in the normal mucosa. The COX-2 (PG) synthesis and through a variety of other (prostaglandin H synthase-2, PGHS-2, EC peripheral and central mechanisms (table 1.14.99.1) is a multifunctional enzyme.52 Its I).46,47 Aspirin was the first NSAID shown to gene is highly inducible by agents such as pro- inhibit the COX pathway in which arachidonic inflammatory endotoxins that enhance the

TABLE I

Group Agent Function 1 Flurbiprofen COOH group lowers OXPHOS yield 2 NO-flurbiprofen COOH group removed; OXPHOS intact 3 Sodium salicylate No COX inhibition; inhibits PMN adhesion via adenosine release 4 Indomethacin COX inhibitor; increases PMN p-2-integrin and PMN adhesion 5 Flosulide COX-2 selective inhibitor 6 Tepoxalin COX and 5-LO inhibitor; NF kappa B inhibitor 7 Zileuton Selective 5-LO inhibitor; inhibits PMN adhesion 8 Misoprostol PG analogue; restores cytoprotection OXPHOS - Oxidative phosphorylation. COX - Cyclo-oxygenase. 5 - LO - 5 - lipoxygenase. NF * Nuclear factor. ADVERSE EFFECTS OF NSAIDS 7 1

Antacid * ATP - - IOXPHOS FlosM e | Permeability [ C00H Endotoxin NSAIDs Inflammation r -P G sf Arachidonic acids ► Q PGs I Tepoxalin QPain Zileuton Sodium PMN__ salicylate ^ (g )-(3 2 -in te g rin # SMC Leukocyte adhesion ^ ------Leukotrienes i t Vasoconstriction

F igure 2. Beneficial (open arrows) and adverse (black arrows) effects induced by non-steroidal anti-inflammatory drugs (NSAIDs). Numbers in brackets refer to agent group in table I. Lower production of proinflammatory prosta­ glandins (PG) is achieved by NSAID-induced inhibition of cyclo-oxygenase-2 (COX-2) [2,4(NO-flurbiprofen, Indo- methacin)] reducing pain and inflammation (open arrows). However, conventional NSAIDs also inhibit COX-1 [2,4{NO- flurbiprofen, Indomethacinj] causing loss of cytoprotective functions in the intestinal mucosa. The NSAIDs such as Flurbiprofen, [1] harbor a carboxyl moiety that inhibits oxidative phosphorylation (OXPHOS) reducing local adenosine- triphosphate (ATP) generation and increasing mucosal permeability. Misoprostol [8], a PG analogue, partially restores cytoprotection. The COX inhibition provides surplus arachidonic acids increasing 5-LO conversion to leukotrienes that contract vascular smooth muscle cells (SMC). Microcirculatory blood flow is further impaired by leukocyte (PMN) adhesion to vascular endothelium. Tepoxalin [6] inhibits both COX and 5-LO. Use of newer drugs that selectively inhibit COX-2 (Flosulide) [5] or 5-LO (Zileuton) [7] result in fewer NSAID-induced adverse effects. For further detail see text.

COX-2 transcript levels increasing the synthe­ ratio of inhibition of the two COX isoforms sis of proinflammatory prostaglandins at sites varies for different NSAIDs (figure 3). Strong of inflammation. inhibition of COX-1 by NSAIDs such as piroxi- Prostaglandins derived from arachidonic cam reduces mucosal prostaglandin cytopro­ acid via COX-2, in particular PGE2, initiate tective functions, induces mucosal vascular inflammation and pain.53 Inhibition of COX-2 injury, and leads to mucosal erosions and to activity, partially by conventional NSAIDs and ulcers. There may be local damage due to local more completely by selective COX-2 inhibitors effects where the NSAID tablet sits adjacent such as flosulide is therefore a major mecha­ to the mucosa. But gastric ulcerations occur nism for the analgesic and anti-inflammatory even when NSAIDs are administered intra­ action of NSAIDs.46 Treatment with monoclo­ muscularly or intravenously. nal anti-PGE2 antibody fully reverses hyperal­ Ligation of the bile duct reduces NSAID gesia in experimental models.53 Interference damage in animal models indicating that with G-protein-mediated signal transduction enterohepatic recirculation is important in the by NSAIDs forms a PG independent pathway pathogenesis of NSAID-induced pathologies. for analgesic action.46 Diclofenac but not nitrofenac undergoes Some NSAIDs also inhibit the lipoxygenase extensive enterohepatic circulation when pathway.46 However, conventional NSAIDs administered to rats, exposing parts of the simultaneously inhibit COX-1 lowering syn­ intestine repeatedly to the drug.’4 Diclofenac thesis of cytoprotective prostaglandins. The but not nitrofenac induces frank intestinal 7 2 FOSSLIEN SC5766 SC58125 Flosulide Nabumetone F igure 3. Inhibition of Naproxen the two COX isoforms varies for different NSAIDs. Histo­ Diclofenac gram of data111 ranking of NSAIDs by relative selectiv­ Indomethacin ity of COX-1 and COX-2 inhibition. Piroxicam inhibits Sulindac COX-1 the most, flosulide is Aspirin a strong COX-2 inhibitor. Piroxicam

0.00001 0.001 0.1 10 1000 Log COX-2/COX-1 ulcerations in such animals despite similar comitant inhibition of prostaglandin synthesis alterations in gastrointestinal permeability. further reduces the amount and quality of gas­ Foods may have adverse effects. Refeeding trointestinal mucosal protection.58 after NSAID administration enhances macro­ All of 10 NSAIDs tested uncoupled rat liver scopic damage in animals. mitochondrial respiration in vitro56 and Microbe involvement in the pathogenesis of decreased the mitochondrial membrane lesions is suggested by experiments showing potential.59 Acetylsalicylic acid, diclofenac that pretreatment with antimicrobial drugs sodium, piroxicam, and mefenamic acid both reduces damage in animals. Neutrophil adhe­ uncouple and inhibit OXPHOS in rat renal sion to endothelial cells is damaging. Mucosal cortex mitochondria in vitro.59 Dipyrone only damage is limited when neutropenia is uncouples and paracetamol only inhibits induced or when antibodies to adhesion mole­ OXPHOS. Evidence of mitochondrial damage cules in experimental models limit neutrophil caused by NSAIDs in animal models include adhesion. The pathogenesis of NSAID- activation of mitochondrial marker enzymes induced gastric ulcers is also associated with after oral NSAIDs, damage to mitochondria regionally disturbed gastric microcirculation shown on electron micrographs after oral and with the presence of acid.55 NSAIDs, and reduced ATP generation.60 The NSAIDs harboring a carboxyl moiety Flurbiprofen, ibuprofen, and ketoprofen sig­ cause interfere with OXPHOS at micromolar nificantly increase intestinal permeability in concentrations in vitro.56 The carboxylic group rats above that seen in untreated animals.44 acts as a proton translocator resulting in lower Seventy five mg of indomethacin for one day adenosine-triphosphate (ATP) generation.57 significantly increases the permeability of the Intercellular tight junctions of mucosal cells small intestine in healthy volunteers.61 All four are damaged increasing gastrointestinal per­ drugs contain a carboxyl group. Other NSAIDs meability. Cells depleted of ATP are vulner­ of this kind, for instance sulindac, etodolac, able to oxidant stress.31 The DNA synthesis is and flurbiprofen, are all associated with gas­ reduced, and mucosal cell proliferation is trointestinal toxicity.57,62 From the gastric impaired. Mucosal cells are less able to gener­ lumen these NSAIDs rapidly penetrate the ate components of the protective gastric bar­ hydrophilic lipid-protective layer and reach a rier resulting in backflow of acid and pepsin high level of concentration in the superficial into the mucosa. The NSAID-induced con- cells of the mucosa.58,62 ADVERSE EFFECTS OF NSAIDS 7 3

Prostaglandins consist of a group of potent dermal growth factor (EGF) induces both lipid mediators affecting gastrointestinal secre­ COX-2 messenger ribonucleic acid (mRNA) tion, blood flow, and motility.63 Cytoprotective and protein, but has no effect on COX-1 prostaglandins preserve the gastric mucosa by expression.71 Benzo[a]pyrene up-regulates inhibiting acid secretion, raising bicarbonate COX-2 expression,72 possibly explaining why output and mucus secretion, lowering mucosal tobacco smoking is a risk factor for NSAID permeability to H+ ions, and maintaining induced ulcers. In rat intestinal cells, trans­ mucosal blood flow. The NSAID-induced forming growth factor beta l(TGF-fJl) inhibition of COX-1 reduces synthesis of gas­ strongly induces COX-2 at both the mRNA tric cytoprotective prostaglandin. This results and protein level, downregulating cyclin D1 in increased acid production, decreased mucus and inhibiting cell growth. 3 Human adenovi­ production, back-diffusion of H+ ions into the rus E4 promoter binding protein (E4BP4) gastric mucosa, reduction in mucosal blood type elements, inducible by dexamethasone, flow and delayed cellular repair. Consequently are located in the COX-2 promoter, explaining inhibition of COX-1 by NSAIDs results in an a possible mechanism for glucocorticoid injured mucosa less able to cope with acid. repression of COX-2.74 The COX-2 is induced Back-diffusion of acid from the lumen in wounds, granulomas, ulcers, osteoclasts, increases mucosal acidosis and enhances proliferative phase endometrium, ovulation drug absorption. and parturition, and in colon carcinomas.75 The importance of reduced cytoprotective Transgenic COX-2 mice are able to mount prostaglandin synthesis in the pathogenesis of an inflammatory response.76 They show no mucosal injury is supported by experimental and innate gastrointestinal pathology. However, clinical evidence. Oral indomethacin adminis­ they suffer from serious renal development tered to rats (2 mg/kg daily for 4 days) induces anomalies and progressive deterioration of kid­ intestinal adhesions, perforations, and neutrophil ney function with age. Their life span is adhesion in mesenteric venules and inflamma­ reduced. Such findings in COX-2 gene knock­ tory cell infiltration in the mesenteric intersti- out mice raise significant questions of potential tium.64 In patients, exogenous prostaglandin adverse renal effects of long-term COX-2 inhi­ analogues such as misoprostol reduces the inci­ bition in humans. However, gene disruption dence of reactive gastritis65 and protects against during the development in utero is probably both gastric and duodenal ulcers.66 different from COX-2 inhibition in an adult However, mice homozygous for COX-1 patient with fully developed tissues.58 deficiency survive well and without gastric An imbalance between the production of pathology.67 In spite of lack of measurable prostaglandins and vasoconstrictive leukot- amounts of COX-1 in the gastric mucosa the rienes is an important factor in the loss of gene knockout mice showed no increase in mucosal integrity during NSAIDs absorp­ gastroduodenal ulcer development and sur­ tion.40 The NSAID-induced inhibition of COX prisingly developed less indomethacin- metabolism diverts arachidonic acids into the induced gastric ulceration when compared 5-lipoxygenase pathway that produces vaso­ with mice harboring intact COX-1 genes. constrictor leukotrienes and generates oxyradi- These results in COX-1 null mice challenges cals.64,77,78 Leukotrienes attract inflammatory the concept of cytoprotective prostaglandins cells to local sites of inflammation and produce and of NSAID-induced COX-1 inhibition in ulcerations. Neutrophil adherence to the vas­ the pathogenesis of gastric injury.68 cular endothelium increases and neutrophils In cell models IL-169 and hypoxia70 induce release tissue-damaging mediators.79 Mucosal COX-2 expression via nuclear factor-kappa B perfusion is reduced. Indomethacin, for (NF-kB) p65 binding to matching sites in the example, increases leukotriene C-4 in the gas­ COX-2 promoter region. Oncogenes src and tric efferent circulation in rats and pigs and ras up-regular COX-2 promoter activity.52 Epi­ induces mucosal lesions.80 However, oral dos­ 7 4 FOSSLIEN

ing of the selective 5-lipoxygenase inhibitor, and etodolac of 400 mg with twice daily inges­ MK-886, prevents development of both gastric tion for 4 weeks versus a placebo given to 52 and intestinal mucosal lesions.80 healthy volunteers with normal baseline Conventional NSAIDs injure the mucosal endoscopy did not cause diffuse mucosal endothelium within minutes of administration injury of the gastric mucosa. The NSAID use by inducing neutrophil adherence to the did not alter H. pylori-induced gastritis. mucosal vascular endothelium. Subsequently, Development of NSAID-induced gastro­ neutrophils release oxygen-derived free radical duodenal damage was not influenced by and proteolytic enzymes. Reduced prostaglan­ underlying H. pylori infection. dins synthesis through inhibition of COX-1 combined with an increase in lipoxygenase products such as leukotriene B4 contribute to Therapeutic Alternatives the damage.79 Prevention of neutrophil adher­ ence or depletion of circulating neutrophils Younger patients without risk factors may results in reduced susceptibility to NSAID- adapt to NSAID use. However, NSAIDs induced experimental mucosal injury.81 should be used cautiously in elderly patients Sodium salicylate, indomethacin, and piroxi- and in patients who smoke or have a history of cam all inhibit stimulated neutrophil adher­ peptic ulcer, or who use oral corticosteroids ence to the endothelium.82 or anticoagulants.38 Strategies to reduce the Different NSAIDs differ in the mechanism side effects of NSAIDs include develop­ of their effect on the interaction of neutrophils ment of new drug classes, enteric coating, with the vascular endothelium. Indomethacin non-acidic drugs, and pro-drugs such as droxi- induces margination of circulating neutrophils cam (a prodrug of piroxicam) and nabume- (PMN) in the gastric microcirculation via up- tone.87 A lack of cytoprotective prostaglandin regulation of beta-2-integrin on the surface of due to NSAID-induced COX-1 inhibition the PMNs.83 Sodium salicylate inhibits stimu­ can be ameliorated by administration of syn­ lated neutrophil adhesion to endothelium thetic prostaglandins.88 without inhibiting prostaglandin synthesis. The Patients given the prostaglandin El ana­ drug inhibits oxidative phosphorylation, logue misoprostol concomitantly with NSAID thereby making more ADP available. therapy have reduced gastric pathology Adenosine inhibits stimulated neutrophil compared to patients given NSAIDs without adhesion to endothelium. Adding adenosine exogenous prostaglandin supplementation.89 deaminase (ADA) converts ADP to its inactive Misoprostol protects against both gastric and metabolite, inosine. Thus, ADA inhibits neu­ duodenal ulcers and reduces the risk of serious trophil adhesion induced by sodium salicylate complications by 40 percent.66 However, but not by indomethacin and piroxicam. Pep­ NSAID-induced bleeding is not affected by sinogen is a possible growth factor promoting concomitant oral misoprostol treatment. One healing of NSAID-induced gastric ulcer­ to two years of misoprostol treatment of 90 ations.84 Indomethacin increases pepsinogen patients receiving NSAIDs significantly production in isolated guinea pig gastric chief reduced the prevalence of reactive gastritis.65 cells by enhancing LTB4 release from the Ornoprostil, a PGE1 analogue simulta­ cells.8 Pretreating the cells with a 5-LO neously administered with indomethacin pre­ inhibitor abolishes pepsinogen generation. vents indomethacin-induced intestinal perme­ Infection with the bacterium Helicobacter ability.90 Synthetic prostaglandin analogues pylori does not aggravate NSAID-induced gas­ may have their own side effects. Mild diarrhea tric ulcers. They heal with routine treatment.66 and gastrointestinal intolerance are prominent Presence of H. 'pylori in the stomach is asso­ adverse reactions experienced by patients ciated with diffuse histological injury of the receiving arbaprostil and enprostil,91 miso­ gastric mucosa.86 Daily naproxen of 500 mg prostol, on the other hand, is well tolerated.92 ADVERSE EFFECTS OF NSAIDS 7 5

Cimetidine and ranitidine (H2 blockers) rat.94 The selective 5-lipoxygenase (5-LO) showed minimal protective value in both short inhibitor nordihydroquaiaretic acid (NDHA) term (<2 weeks) and long term trials.88 reduces the severity or indomethacin-induced Meloxicam, a relatively selective COX-2 ulcer formation in rats.95,96 No correlation was inhibitor, combines anti-inflammatory efficacy evident between the antioxidant properties of with improved tolerability.93 Results from a NDHA and the ability to reduce the severity of double-blind, randomized, 28 day trial in more gastric damage.96 than 9,000 patients showed that meloxicam in In another study, oral indomethacin (100 therapeutic dosages causes less dyspepsia, rng/kg) was employed to produce elevated lev­ abdominal pain, nausea, vomiting and diarrhea els of leukotriene (LT) B4 (LTB4) in rat gas­ than diclofenac. Meloxicam caused fewer pep­ tric mucosa 90 minutes after administration; tic ulcers and gastrointestinal bleeds than pretreatment with the selective 5-LO inhibi­ naproxen, diclofenac, or prioxicam and without tor zileuton and the COX/5-LO inhibitor any increase in renal or liver abnormalities tepoxalin [5-(4-chlorophenyl)-N-hydroxy-(4- compared to other NSAIDs. methoxyphenyl)-N-methyl-lH-pyrazole-3- The NSAIDs that are more COX-2 selective propanamide]97 prevented the increase in (from 3- to 10-fold more selective for COX-2 LTB4 levels as well as indomethacin-induced than for COX-1) have less gastrointestinal tox­ neutrophil adhesion.64 The LTD4 receptor icity associated with their use. Highly selective antagonists, such as MK-571, significantly inhibitors of COX-2 (300 fold or more selec­ reduce indomethacin-induced mucosal per­ tive for COX-2 over COX-1)89 reduce the risk meability increases.96 Tebufelone is an NSAID of adverse effects from NSAIDs. These thera­ of the di-tert-butylphenol class that inhibits peutically promising compounds with a high both PGE2 and LTB4 generation.98 Long selectivity for COX-2 still require well- term use of single doses up to 800 mg was designed large clinical trials to adequately generally well tolerated.99 evaluate advantages versus potential draw­ The carboxyl group can be chemically modi­ backs that may result from prolonged selective fied, such as in dimero-flurbiprofen and nitro- COX-2 inhibition.50 butyl-flurbiprofen (figure 4). The latter Studies on selective 5-lipoxygenase inhibi­ NSAIDs, as well as nabumetone and highly tion using MK-886 (3-[l-(4-chlorobenzyl)-3-t- selective COX-2 inhibitors such as flosulide,100 butylthio-5-isopropylindol-2-yl]-2,2-dimethyl- do not cause uncoupling in vitro and are propanoic acid) failed to demonstrate94 or reported to show increased gastrointestinal demonstrated80 significant prevention of indo- tolerability (figure 5). Some enteric-coated or methacin-induced gastroenteropathy in the prodrug formulations induce more side effects

Uncoupler group Flurbiprofen

F igure 4. Carboxyl moi­ ety of flurbiprofen acts as a proton translocator interfer­ ing with mitochondrial 2 1 OXPHOS, altered structures (CH X— O — N02 NO-flurbiprofen such as NO-flurbiprofen and dimero-flurbiprofen do not uncouple OXPHOS.62 O CH, C------O ------C —CH (O)—(O Dimero-flurbiprofen 7 6 FOSSLIEN 20 18 16 □ Flosulide 14 ■ Naproxen 12 N 10 8 6 4 2 0 u

1 2 3 F ig u r e 5. Bargraph of endoscopic data100 obtained Lanza score from patients (N) given either naproxen (solid bars) or the selective COX-2 inhibitor flosulide (open bars) graded either by Lanza scores or by gastroscopic damage scores. Higher scores represent more seri­ ous lesions in both scoring systems. Flosulide use resulted in fewer serious endoscopic findings com­ pared to naproxen.

0 1 2 3 4 5 6 7 Gastroscopic damage score than others. Kelly et al reported that use of low proinflammatory leukotrienes and prostaglan­ doses of enteric-coated aspirin carried a 3-fold dins. In a one year, placebo-controlled study of increase in risk of major upper-gastrointestinal 64 patients with stable rheumatoid arthritis bleeding.101 However, administration of requiring NSAID therapy, EPA supplementa­ enteric coated naproxen to patients with either tion (171 mg/day) led to significantly reduced osteoarthritis or rheumatoid arthritis signifi­ use of NSAIDs without any deterioration in cantly reduced gastrointestinal complaints the clinical and laboratory parameters of dis- • • 1 O'} compared to ingestion of standard immediate ease activity. release naproxen.102 Avoidable NSAID-induced adverse effects Fish oil has anti-inflammatory properties. may be iatrogenic or due to self-medication by Supplementation in the form of eicosapentae- over-the-counter (OTCD) NSAID containing noic acid (EPA) as an alternative substrate to drugs. Patient ignorance may lead to ingestion arachidonic acid leads to the formation of less of NSAIDs in OTCD in addition to prescrip­ ADVERSE EFFECTS OF NSAIDS 7 7 tion drugs. Physician prescribing patterns may growth during healing of the gastric mucosa.84 contribute to avoidable gastrointestinal mor­ Patients with gastric body ulcers have a higher bidity. In a Canadian study, exposure to poten­ frequency of allele 4 of the pepsinogen gene, tially inappropriate drug combinations (PIDC) possibly reducing the rate of mucosal healing. in 51,587 elderly patients receiving NSAID An indomethacin-induced increase in pep­ therapy was 4 percent.104 About one quarter of sinogen production in isolated guinea pig gas­ the incidents resulted from contemporaneous tric chief cells probably represents cellular prescribing by different physicians. The use of adaptation to the toxic effects of the drug.85 a single prescribing pharmacy but not the Indomethacin enhances LTB4 release from presence of a single prescribing physician low­ the cell and pretreating the cells with a 5-LO ered the risk of PIDC involving NSAIDs. inhibitor abolishes pepsinogen generation. A separate prospective cohort study of The COX-1 gene is induced by endotoxin NSAID prescriptions by 112 physicians con­ administration in the rat. This induction is cluded that unnecessary NSAID prescribing important in gastric adaptation and healing, and less than optimal management of NSAID- and is not related to H. pylori infection.108 related side effects were sufficiently common Gliostatin is a protein factor related to rheu­ to cause concern about the appropriateness of matoid arthritis disease activity.109 Gliostatin NSAID use in the general population.105 Inap­ infusion delays experimental ulcer healing in propriate prescription of NSAIDs was more rats. The precise mechanism of the interaction common when contraindications to NSAID of these and other cytokines and growth fac­ therapy were incompletely assessed. While tors in NSAID-induced gastrointestinal lesions NSAID-related gastropathy was diagnosed and their healing still remain to be elucidated. correcdy in 93 percent of office visits, they were acceptably managed only in 77 percent of Conclusion the visits. The NSAID therapy should not be used or No NSAID that lacks the potential for seri­ continued in elderly, high risk patients with a ous gastrointestinal toxicity is currently avail­ history of recent major gastric ulcer activity or able. Generally, some newer drugs are better bleeding ulcer. Age may represent a risk factor tolerated than many conventional NSAIDs. for damage to mitochondrial function caused Still, substantial morbidity and mortality owing by NSAIDs. Paracetamol generally does not to NSAID-induced adverse effects impart a impair kidney mitochondrial energy in young high cost both to the patients and society. animals, however, in aged animals administra­ Long-term use of NSAIDs in high-risk tion of the drug leads to impaired mitochon­ patients should be avoided whenever possible. drial energy metabolism.106 Similar studies of When NSAIDs are prescribed, the lowest the role of age in the impact of NSAIDs on effective dose of NSAID should be selected. gastrointestinal cell mitochondria have not yet The NSAIDs which do not affect mitochon­ been reported. However, NSAIDs which drial function and are not exposed to entero- decouple OXPHOS should probably be hepatic recirculation are preferable. Weak avoided in the elderly. COX-1 and strong COX-2 inhibitors are desir­ Gastroduodenal ulcer disease consists of a able. Patients at increased risk for upper gas­ heterogeneous group of different multifacto­ trointestinal complications should be given a rial etiologies. Some patients harbor genetic prostaglandin analogue such as misoprostol risk factors. For instance, blood groups A and concomitantly with NSAID treatment to O are associated with gastric and duodenal restore mucosal cytoprotective functions. ulcers, respectively, and non-secretor pheno­ All gastric ulcers require biopsy and histo­ types predispose to both type of ulcers.1 7 Pep­ logical examination. Selective COX-2 inhibi­ sinogen C may play a role in epithelial cell tors appear to provide an improved gastroin­ 7 8 FOSSLIEN testinal safety profile compared to older induction of apoptosis in colon cancer cells by a pros­ NSAIDs. However, the COX-2 gene is induc­ taglandin-independent pathway. Biochem Pharma­ col 1996;52:237-45. ible by several cytokines and growth factors. It 14. Breitner JC. Inflammatory processes and antiinflam­ codes for a multifunctional enzyme involved in matory drugs in Alzheimer’s disease: a current wound healing, proliferative stage endome­ appraisal. Proc Natl Acad Sci USA. 1996;93:11091-6. 15. Polisson R. Nonsteroidal anti-inflammatory drugs: trium, osteoclasts, ovulation and parturition. Practical and the theoretical considerations in their The safety of long term use in patients of selec­ selection. AJM 1996;100:Suppl 31S-6S. tive inhibitors of COX-2 and 5-lipoxygenase 16. Colin R. Digestive complication of non-steroid anti­ inflammatory drugs. Schweiz Med Wochenschr must therefore be carefully established. 1991;121:716-21. Adverse effects on the gastrointestinal tract 17. Mohamed AH, Salena BH, Hunt RH. NSAID- due to non-prescription use of high dose aspi­ induced gastroduodenal ulcers: exploring the silent rin and other conventional NSAIDs remain a dilemma. J Gastroenterol 1994;29:Suppl 7:34—8. 18. Hawkey CJ. Non-steroidal anti-inflammatory drug considerable problem. gastropathy: causes and treatment. Scand J Gastro­ enterol 1994;29:Suppl 220 M:124-7. 19. Peltier AP. Practical management of NSAID gas­ References tropathy in rheumatic patients. Scand J Rheumatol Suppl 1989;78:12-7. 20. Blower AL. Consideration for nonsteroidal anti­ 1. Garcia J, Altman RD. Chronic pain states: patho­ inflammatory drug therapy: safety. Scand J Rheuma­ physiology and medical therapy. Semin Arthritis tol 1996;105:Suppl 13-24. Rheum 1997;27:1-16. 21. Carson JL, Willett LR. Toxicity of nonsteroidal anti­ 2. Amadio P. NSAIDs revisited. Selection, monitoring, inflammatory drugs. An overview of the epidemio­ and safe use. Postgrad Med 1997;101:257-76. logical evidence. Drugs 1993;46:Suppl 1:243-8. 3. Fries JF, Williams CA, Bloch DA, Michel BA. Non­ 22. Miwa LJ, Jones JK, Pathiyal A, Hatoum H. Value of steroidal anti-inflammatory drug-associated gas- epidemiologic studies in determining the true inci­ tropathy: incidence and risk factor models. Am J dence of adverse events. The nonsteroidal anti­ Med 1991;91:213-22. inflammatory drug story. Arch Intern Med 1997;157: 4. Siebert K, Zhang Y, Leahy K, Hauser S, Masferrer J, 2129-36. Perkins W, Lee L, Isakson P. Pharmacological and 23. Hawkey CJ, Cullen DJ, Greenwood DC, Wilson JV, biochemical demonstration of the role of cyclooxy- Logan RF. Prescribing of nonsteroidal anti­ genase 2 in inflammation and pain. Proc Natl Acad inflammatory drugs in general practice: determi­ Sci USA 1994;91:12013-17. nants and consequences. Aliment Pharmacol Ther 5. Brigden M, Smith RE. Acetylsalicylic-acid-contain- 1997;11:293-8. ing drugs and nonsteroidal anti-inflammatory drugs 24. Wilcox CM, Alexander LN, Cotsonis GA, Clark WS. available in Canada. Can Med Assoc J 1997;156: Nonsteroidal antiinflammatory drugs are associated 1025-8. with both upper and lower gastrointestinal bleeding. 6. Hawkey CJ. Gastroduodenal problems associated Dig Dis Sci 1997;42:990-7. with non-steroidal, anti-inflammatory drugs 25. Laine L. Nonsteroidal anti-inflammatory drug gas­ (NSAIDs). Scand J Gastroenterol Suppl 1993;200: tropathy. Gastrointest Endosc Clin N Am 1996;6: 94-5. 489-504. 7. Urquhart E. Analgesic agents and strategies in the 26. Svanes C, Overebo K, Soreide O. Ulcer bleeding and dental pain model. J Dent 1994;22:336-41. perforation: non-steroid anti-inflammatory drugs or 8. Cashman J, McAnulty G. Nonsteroidal anti-inflam­ Helicobacter pylori. Scand J Gastroenterol matory drugs in perisurgical pain management. 1996;Suppl 220:128-31. Mechanism of action and rationale for optimum use. Drugs 1995;49:51-70. 27. Lanas A, Serrano P, Bajador E, Esteva F, Benito R, 9. Montalescot G. [Use of aspirin in coronary disease]. Sainz R. Evidence of aspirin use in both upper and Presse Med 1995;24:925-7. lower gastrointestinal perforation. Gastroenterology 10. Ridker PM, Cushman M, Stampfer MJ, Tracy RP, 1997;112:683-9. Hennekens CH. Inflammation, aspirin, and the risk 28. Bar Dayan Y, Amital H, Levy Y, Shoenfeld Y. Low of cardiovascular disease in apparently healthy men. dose aspirin in patients with ischemic heart disease N Engl J Med 1997;336:973-9. may precipitate secondary myocardial infarction. Isr 11. Giardiello FM. NSAID-induced polyp regression in J Med Sci 1996;32:288-91. familial adenomatous polyposis patients. Gastroen­ 29. Fries JF, Miller SR, Spitz PW, Williams CA, Hubert terol Clin North Am 1996;25:349-62. HB, Bloch DA. Identification of patients at risk for 12. Gustafson-Svard C, Lilja I, Hallbook O, Sjodahl R. gastropathy associated with NSAID use. J Rheuma­ Cyclo-oxygenase and colon cancer: clues to the aspi­ tol 1990;20(Suppl):12—9. rin effect? Ann Med 1997;29:247-52. 30. Wolfe MM. NSAIDs and the gastrointestinal mucosa. 13. Hanif R, Pittas A, Feng Y, Koutsos MI, Qiao L, Stai- Hosp Ptract 1996Dec 15;31:37-44. ano-Coico L, Shiff SI, Rigas B. Effects on nonsteroi­ 31. Hayllar J, Macpherson A, Bjamason I. Gastroprotec- dal anti-inflammatory drugs on proliferation and on tion and nonsteroidal anti-inflammatory drugs ADVERSE EFFECTS OF NSAIDS 7 9

(NSAIDs). Rationale and clinical implications. Drug 49. Vane JR, Botting RM. Mechanism of action of aspi­ Saf 1992;7:86-105. rin-like drugs. Semin Arthritis Rheum 1997;26(6 32. Blower AL, Brooks A, Fenn GC, Hill A, Pearce MY, Suppl 1):2-10. Morant S, Bardhan KD. Emergency admissions for 50. Jouzeau JY, Terlain B, Abid A, Nedelec E, Netter P. upper gastrointestinal disease and their relation to Cyclo-oxygenase isoenzymes. How recent findings NSAID use. Aliment Pharmacol Ther 1997;11:283- affect thinking about nonsteroidal anti-inflammatory 91. drugs. Drugs 1997;53:563-82. 33. Aly A. NSAID-induced ulcer and its complications. 51. Ferraz JG, Wallace JL, McKnight W, Brown ML. Nord Med 1997;112:195-7. Induction of cyclooxygenase 1 and 2 in the rat stomac 34. Kurata JH, Nogawa AN, Noritake D. NSAIDs increase during endotoxemia: role in resistance to damage. risk of gastrointestinal bleeding in primary care Gastroenterology 1997; 113:195-204. patients with dyspepsia. J Fam Pract 1997;45:227- 52. Subbaramaiah K, Telang N, Ramonetti JT, Araki R, 35. DeVito B, Weksler BB, Dannenberg AJ. Transcrip­ tion of cyclooxygenase-2 is enhanced in transformed 35. Bertschinger P, Zala GF, Fried M. Effect o f non­ mammary epithelial cells. Cancer Res 1996;56: steroidal antirheumatic agents on the gastrointestinal 4424-9. tract: clinical aspects and pathophysiology. Schweiz 53. Zhang Y, Shaffer A, Portanova J, Seibert K, Isakson Med Wochenschr 1996;126:1566-8. PC. Inhibition of cyclooxygenase-2 rapidly reverses 36. Roth SH. From peptic ulcer disease to NSAID gas- inflammatory hyperalgesia and prostaglandin E2 tropathy. An evolving nosology. Drugs Aging 1995; production. J Pharmacol Exp Ther 1997;28:1069-75. 6:358-67. 54. Reuter BK, Davies NM, Wallace JL. Nonsteroidal 37. Guttmann SP, Garcia Rodriguez LA, Raiford DS. anti-inflammatory drug enteropathy in rats: role of Individual nonsteroidal antiinflammatory drugs and permeability, bacteria, and enterohepatic circulation. other risk factors for upper gastrointestinal bleeding Gastroenterology 1997;112:109-17. and perforation. Epidemiology 1997;8:18-24. 55. Sato N, Kawano S, Tsuji S, Ogihara T, Yamada S. 38. Garcia Rodriguez LA, Jick H. Risk of upper gastro­ Gastric blood flow in ulcer diseases. Scand J Gastro­ intestinal bleeding and perforation associated with enterol Suppl 1995;208:14-20. individual non-steroidal anti-inflammatory drugs. 56. Banos G, Reyes PA. A comparative study of the Lancet 1994;343:769-72. effect of ten non-steroid anti-inflammatory drugs 39. Davies NM. Toxicity of nonsteroidal anti-inflam­ (NSAIDs) upon some mitochondrial and platelet matory drugs in the large intestine. Dis Colon Rec­ functions. Int J Biochem 1989;21:1387-94. tum 1995;38:1311-21. 57. Mahmud T, Wrigglesworth JM, Rafi S, Scott DL, 40. Bjamason I, Macpherson A. The changing gastroin­ Bjamason I. Uncoupling of oxidative phosphoryla­ testinal side effect profile of non-steroidal anti­ tion by non-steroidal anti-inflammatory drugs inflammatory drugs. A new approach for the preven­ (NSAIDs) is related to ionizable moiety and can be tion of a new problem. Scand J Gastroenterol Suppl altered by modification of carboxylic group. Arthritis 1989;163:56-64. Rheum 1995;38 (Suppl):S207. 41. El-Serag HB, Sonnenberg A. Association of esopha­ 58. Simon LS. Biologic effects of nonsteroidal anti­ gitis and esophageal strictures with diseases treated inflammatory dmgs. Curr Opin Rheumatol 1997;9: with nonsteroidal anti-inflammatory drugs. Am J 178-82. Gastroenterol 1997;92:52-6. 59. Mingatto FE, Santos AC, Uyemura SA, Jordani MC, 42. Hayllar J, Smith T, Macpherson A, Price AB, Curti C. In vitro interaction of nonsteroidal anti­ Gumpel M, Bjamason I. Nonsteroidal antiinflamma­ inflammatory dmgs on oxidative phosphorylation of tory drug-induced small intestinal inflammation and rat kidney mitochondria: respiration and ATP syn­ blood loss. Arthrit Rheumat 1994;37:1146-50. thesis. Arch Biochem Biophys 1996;334:303-8. 60. Somasundaram S, Rafi S, Hayllar J, Sigthorsson G, 43. Lesur G. Non-gastroduodenal digestive toxicity of Jacob M, Price AB, Macpherson A, Mahmod T, Scott non-steroidal anti-inflammatory agents. Presse Med D, Wrigglesworth JM, Bjamason I. Mitochondrial 1996;25:1376-80. damage: a possible mechanism of the “topical” phase 44. Davies NM, Wright MR, Fussel AS, Jamah F. Effect of NSAID induced injury to the rat intestine. Gut of the enantiomers of flurbiprofen, ibuprofen, and 1997;41:344-53. ketoprofen on intestinal permeability. J Pharm Sci 61. Bjamason I, Smethurst P, Fenn CG, Lee CE, Men- 1996;85:1170-3. zies IS, Levi AJ. Misoprostol reduces indomethacin- 45. Carratu R, Parisis P, Agozzino A. Segmental isch­ induced changes in human small intestinal perme­ emic colitis associated with nonsteroidal antiinflam­ ability. Dig Dis Sei 1989;34:407-11. matory drugs. J Clin Gastroenterol 1993;16:3H 62. Mahmud T, Scott DL, Bjamason I. A unifying hypothe­ 46. Cashman JN. The mechanism of action of NSAIDs sis for the mechanism of NSAID-related gastrointes­ in analgesia. Drugs 1996;52 Suppl 5:13-23. tinal toxicity. Ann Rheum Dis 1996;55:211-3. 47. Davies NM, Wallace JL. Nonsteroidal anti-inflam­ 63. Wallace JL. Prostaglandins, NSAIDs, and cytopro- matory drug-induced gastrointestinal toxicity: new tection. Gastroenterol Clin North Am 1992;21:631— insight into an old problem. J Gastroenterol 1997;32: 41. 127-33. 64. Kirchner T, Aparicio B, Artentieri DC, Lau CY, 48. Vane JR. Inhibition of prostaglandin synthesis as a Ritchie DM. Effects of tepoxalin, a dual inhibitor of mechanism of action for aspirin-like drugs. Nature cycIooxygenase/5-lipoxygenase, on events associated New Biol 1971;231:232-5. with NSAID-induced gastrointestinal inflammation. 8 0 FOSSLIEN

Prostaglandins Leukot Essent Fatty Acids 1997;56: 80. Rainsford KD. Leukotrienes in the pathogenesis of 417-23. NSAID-induced gastric and intestinal mucosal dam­ 65. Shah K, Price AB, Talbot IC, Bardhan KD, Fenn age. Agents Actions 1993;39:C24—6. CG, Bjamason I. Effect of long-term misoprostol 81. Scheiman JM, Tillner A, Phol T, Oldenburg A, coadministration with non-steroidal anti­ Angermuller S, Gorlach E, Engel G, Usadel KH, inflammatory drugs: a histological study. Gut 1995; Kusterer K. Reduction of non-steroidal anti­ 37:195-8. inflammatory drug induced gastric injury and leuco­ 66. Raskin JB, White RH, Jaszewski R, Korsten MA, cyte endothelial adhesion by octreotide. Gut 1997; Schubert TT, Fort JG. Misoprostol and ranitidine in 40:720-5. the prevention of NSAID-induced ulcers: a prospec­ 82. Cronstein BN, Van de Stouwe M, Druska L, Levin tive double-blind, multicenter study. Am J Gastro­ RI, Weissmann G. Nonsteroidal antiinflammatory enterol 1996;91:223-7. agents inhibit stimulated neutrophil adhesion to 67. Langebach R, Morham SG, Tiano HF, Loftin CD, endothelium: adenosine dependent and independent Ghanayem BI, Chulada PC. Prostaglandin synthase 1 mechanisms. Inflammation 1994;18:323-35. gene disruption in mice reduces arachidonic acid- 83. Fiorucci S, Santucci L, Gerli R, Brunori PM, Federici induced inflammation and indomethacin-mduced B, Ugolini B, Fabbri C, Morelli A. NSAIDs upregu- gastric ulceration. Cell 1995;83:483-92. late beta 2-integrin expression on human neutrophils 68. Morgan G. NSAID treatment of gastrointestinal abnor­ through a caleium-dependent pathway. Aliment malities: A challenge to ‘cytoprotection’. Scand J Pharmacol Ther 1997;11:619-30. Gastroenterol 1997;32:288. 84. Kishi K, Kinoshita Y, Matushima Y, Okada A, Mae- 69. Crofford LJ, Tan B, McCarthy CJ, Hla T. Involve­ kawa T, Kawanami C, Watanabe N, Chiba T. Pep­ ment of nuclear factor kappa b in the regulation of sinogen C gene product is a possible growth factor cyclooxygenase-2-expression by interleukin-1 in during gastric mucosal healing. Biochem Biophys rheumatoid synoviocytes. Arthritis Rheum 1997;40: Res Commun 1997;238:17-20. 2226-36. 70. Schmedtje JF Jr, Ji YS, Liu WL, DuBois RN, Runge 85. Fiorucci S, Santucci L, Gresele P, Luinetti O, Morelli MS. Hypoxia induces cyclooxygenase-2 via the NF- A. Effect of NSAIDs on pepsinogen secretion and kappaB p65 transcription factor in human vascular calcium mobilization in isolated chief cells. Am J endothelial cells. J Biol Chem 1997;272:601-8. Physiol 1995;268:G968-78. 71. Perkins DJ, Kniss DA. Rapid and transient induction 86. Koch M, Dezi A, Ferrario F, Capurso L. Prevention of cyclo-oxygenase 2 by epidermal growth factor in of nonsteroidal anti-inflammatory drug-induced gas­ human amino-derived WISH cells. Biochem J trointestinal mucosal injury. Arch Intern Med 1996; 1997;321(Pt 3):677-81. 156:2321-32. 72. Kelley DJ, Mestr JR, Subbaramaiah K, Sacks PG, 87. Lancaster C. Effective nonsteroidal anti-inflam- Schantz SP, Tanabe T, Inoue H, Ramonetti JT, Dan- matory drugs devoid of gastrointestinal side effects: nenberg AJ. Benzo[a]pyrene up-regulates cyclooxy­ do they really exist? Dig Dis 1995. 13 Suppl 1:40-7. genase-2 gene expression in oral epithelial cells. Car­ 88. Laine L, Cominelli F, Sloane R, Casini-Raggi V, cinogenesis 1997;18:795-9. Marin-Sorensen M, Weinstein WM. Interaction of 73. Sheng H, Shao J, Hooton EB, Tsujii M, DuBois, NSAIDs and Helicobacter pylori on gastrointestinal Beauchamp RD. Cyclooxygenase-2 induction and injury and prostaglandin production: a controlled transforming growth factor beta growth inhibition in double-blind trial. Aliment Pharmacol Ther 1995;9: rat intestinal epithelial cells. Cell Growth Differ 127-135. 1997;8:463-70. 89. Simon LS. Biologic effects of nonsteroidal anti­ 74. Wallace AD, Wheeler TT, Young DA. Inducibility of inflammatory drugs. Curr Opin Rheumatol 1997;9: E4BP4 suggests a novel mechanism of negative gene 178-82. regulation by glucocorticoids. Biochem Biophys Res 90. Nagase K, Hiwatashi N, Ito K, Maekawa H, Noguchi Commun 1997;232:403-6. M, Kinouchi Y, Toyota T. Effects o f NSAIDs and 75. Katori M, Majima M: [Multiple roles o f inducible PGE1 analogue on the permeability of human small cyclooxygenase-2 and its selective inhibitors]. Nip­ intestine. Nippon Shokakibyo Gakkai Zasshi 1997;94: pon Yakurigaku Zasshi 1997;109:247-58. 469-74. 76. Morham SG, Langebach R, Loftin CD, Tiano HF, 91. Sontag SJ, Schnell TG, Budiman-Mak E, Adelman Vouloumanos N, Jennette C. Prostaglandin synthase K, Fleischman R, Cohen S, Roth SH, Ipe D, 2 gene disruption causes severe renal pathology in the mouse. Cell 1995;83:473-82. Schwartz KE. Healing of NSAID-induced gastric ulcers with a synthetic prostaglandin analog (enpros- 77. Kirchner T, Aparicio B, Artentieri DC, Lau CY, til). Am J Gastroenterol 1994;89:1014-20. Ritchie DM. Effects of tepoxalin, a dual inhibitor of cycIooxygenase/5-lipoxygenase, on events associated 92. Dajani EZ, Agrawal NM. Prevention of nonsteroidal with NSAID-induced gastrointestinal inflammation. anti-inflammatory drug-induced gastroduodenal Prostaglandins Leukot Essent Fatty Acids 1997;56: ulcers: role of mucosal protective and gastric antise- 417-23. cretory drugs. Dig Dis 1995;13 Suppl 1:48-61. 78. Rainsford KD. Mechanisms of gastrointestinal toxic­ 93. Furst DE. Meloxicam: selective COX-2 inhibition in ity of non-steroid anti-inflammatory drugs. Scand J clinical practice. Semin Arthritis Rheum 1997;26: Gastroenterol Suppl 1989;163:9-16. 21-7. 79. Wallace JL. Gastric ulceration: critical events at the 94. Ford-Hutchinson AW, Tagari P, Ching SV, Ander­ neutrophil-endothelium interface. Can J Physiol son CA, Coleman JB, Peter CP. Chronic leukotriene Pharmacol 1993;71:98-102. inhibition in the rat fails to modify the toxicological ADVERSE EFFECTS OF NSAIDS 8 1

effects of a cyclooxygenase inhibitor. Can J Physiol intolerant patients with rheumatoid arthritis and Pharmacol 1993;71:806-10. osteoarthritis. J Rheumatol 1994;21:689-95. 95. Kapui Z, Boer K, Rozsa I, Blasko G, Hermecz I. 103. Lau CS, Morley KD, Belch JJ. Effects of fish oil Investigations of indomethacin-induced gastric ulcer supplementation on non-steroidal anti-inflammatory in rats. Arzneimittelforschung 1993;43:767-71. drug requirement in patients with mild rheumatoid 96. Vaananen PM, Keenan CM, Grisham MB, Wallace arthritis-a double-blind placebo controlled study. Br JL. Pharmacological investigation of the role of leu- J Rheumatol 1993;32:982-9. kotrienes in the pathogenesis of experimental 104. Tamblyn RM, McLeod PJ, Abrahamowisz M, Laprise NSAID gastropathy. Inflammation 1992;16:227-40. R. Do too many cooks spoil the broth? Multiple phy­ 97. Artentieri DC, Ritchie DM, Ferro MP, Kirchner T, sician involvement in medical management of elderly Wächter MP, Anderson DW, Rosenthale ME, Cape- patients and potentially inappropriate drug combina­ tola RJ. Tepoxalin: a dual cyclooxygenase/5- tions. Can Med J 1996;154:1177-84. lipoxygenase inhibitor of arachidonic acid metabo­ 105. Tamblyn R, Berkson L, Dauphine WD, Gayton D, lism with potent anti-inflammatory activity and a Grad R, Huang A, Isaac L, McLeod P, Snell L. favorable gastrointestinal profile. J Pharmacol Exp Unnecessary prescribing of NSAIDs and the man­ Ther 1994;27:1399-409. agement of NSAID-related gastropathy in medical 98. Weisman SM, Doyle MJ, Wehmeyer KR, Hynd BA, practice. Ann Intern Med 1997;127:429-38. Eichhold TH, Clear RM, Coggeshall CW, Kuhlen- 106. Satav JG, Bhattacharya RK. Respiratory functions in beck DL. Effects of tebufelone (NE-11740), a new kidney mitochondria following paracetamol adminis­ anti-inflammatory drug, on arachidonic acid metabo­ tration to young-adult and old rats. Ind J Med Res lism. Agents Actions 1994;41:156-63. 1997;105:131-5. 99. Depre M, Van Hecken A, Verbesselt R, Verpooten 107. Correa P, Schmidt BA. The relationship between GA, Amout J, Brunner F, Jurgens A, Pousset V, gastric cancer frequency and the ratio of gastric to Chow A, Baldauf C, Vermylen J, De Broe M, De duodenal ulcer. Aliment Pharmacol Ther 1995;9 Schepper PJ. Biochemical activity, pharmacokinetics Suppl 2:13-19. and tolerability of tepoxalin, a cyclooxygenase/5- 108. Ohtaka Y, Azuma T, Konishi J, Ito S, Kuriyama. lipoxygenase inhibitor, in man. Int J Clin Pharmacol Association between genetic polymorphism of the Res 1996;16:1-8. pepsinogen C gene and gastric body ulcer: the 100. Bjamason I, Macpherson A, Rotman H, Schupp J, genetic predisposition is not associated with Helico­ Hayllar J. A randomized, double-blind, crossover bacter pylori infection. Gut 1997;41:469-74. comparative endoscopy study on the gastroduodenal 109. Kasugai K, Joh T, Kataoka H, Sasaki M, Tada T, Asai tolerability of a highly specific cyclooxygenase-2 K, Kato T, Itoh M. Evidence for participation of inhibitor, Aosulide, and naproxen. Scand J Gastroen­ gliostatin/platelet-derived endothelial cell growth terol 1997;32:126-30. factor in gastric ulcer healing. Life Sci 1997;61:1899- 101. Kelly JP, Kaufman DW, Jurgelon JM, Sheehan J, 906. Koff RS, Shapiro S. Risk of aspirin-associated major 110. Smedley FH, Taube M, Leach R, Wastell C. Non­ upper-gastrointestinal bleeding with enteric-coated steroidal anti-inflammatory drug ingestion: retro­ or buffered aspirin. Lancet 1996;348:1413-6. spective study of 272 bleeding or perforated peptic 102. Caldwell JR, Roth SH. A double blind study com­ ulcers. Postgrad Med J 1989;65:892-5. paring the efficacy and safety of enteric coated 111. Hayllar J, Bjamason I. NSAIDs, Cox-2 inhibitors, naproxen to naproxen in the management of NSAID and the gut. Lancet 1995;346:521-2.