`©. VSP I099.
`
`Management of' NSAID-related gastrointestinal
`
`mucosa] injury-
`
`ADAM F. BARRISON and M. MICHAEL W'OLFE*
`
`S£'L‘!fr)It ufGctsrru¢+nterr:lag_v. Botmm Ultil'£‘)1\'ff_\’ Sr:;‘zoui qfM'edirf1z9 and Burma Mect't't'm' (Tm-rig-;;
`86’ East Newrran Street, B0.\‘FtJl'1. Ma.t'.t'urrhtrsett;r MA 02! lb’-2393. USA
`
`Received 10 August I999‘: accepted I8 August I999
`
`Abstract—The three therapeutic goals in patients with NSAID-induced gastrotiuodenopath-y are
`treatment of dyspept-ic syniptoms. n'1anager11ent of NSAlD—relate(l ulcers and their complications.
`and prophylaxis against rec-urrcrtt gtt'stroi'ntestinal toxicity-. Both H3-receptor antagonists and proton
`pump inhibitors (PPls) appear to be helpful in relieving the symptoms associated with NSAID use.
`while lt'e:tt'n'1enl -of NSAID-induced _t_r,ast'roduodenal ulcers. whether the NS‘./-\lD is continued or not.
`is best achiever! by the use of PPIS. However. because symptonts do not often predict the .p1'.ese11t:e
`ol' gastroduodenal ulcers. the goal of prevention has become paramount in the treatment of patients
`with an increased likelihood of gastrointestinal toxicity. The best prophylaxis against NSAID-related
`toxicity is the use of an alternative agent such as salsalate or paracetamol‘ (acetam_intJpl'ten). I-Io_weve.r_.
`if an NSAID is to be used. prophylaxis is best accomplished with a PP] or misoprostol. a prostaglandin
`E1 analogue. The use of mistiprostol is limited by its frequent dosing, at least 200 pg three times a day.
`and its own gastrointestinal side effects. Future therapy will include -NSAIDS that tnai-nu-tin their arm'-
`infiamtnatory effect-5. while possessing superior safety profiles. ttntl include pre-l’erent_ial and highly
`selective COX-2 itthihitors and nitric oxide releasing compounds.
`
`gt-JSLPCJ-l.|l1E!ill.l‘tE.ll
`Kev wmrlsz Gastrointestinal: NSAlDs; COX-2',
`hemorrhage; proton pump inhibitors; proslaglandins.
`
`toxicity: ulcers: gastrointesti'na.|
`
`1. INTRODUCTION
`
`Non-steroidal attti—taflamm.ato.ry drugs (NSAIDS) have become one of the most
`commonly used m_e_dict_tt'ions. Over 70 million NSAID prescriptions and 30 billion
`over-the—counter preparations are sold annually in the United States (Wolfe et at.
`1999). The majority of N S-AID users tolerate "these medications without untoward
`effects. but adverse gastrointestinal effects are seen with relative frequency. In 1984.
`the Center for Disease Control reported that 100,000 to 200.000 hospitalizations
`
`"To whom correspondence should be addressed. E-n1ai.l: michael.wolfe@hmc.org
`
`
`
`CFAD EXHIBIT 1019
`
`
`
`278
`
`A. F.” u’3t!I‘l‘f.\'!»!Ii and M. M. lliilfe
`
`and 10.000 to 20.000 deaths occur annually in the United States due to the
`gastrointestinal side effect of these medications. Side effects range from dyspepsia
`to complicated gastroduodenal Lt-lcers associated with hemorrhage or perforation.
`NSAlDs thus constitute a class of drugs th-at can be best C-ha1‘acte1‘i7:ed as a ‘double-
`edged sword‘. medication that
`is very effective. yet carries a substantial
`risk
`potential.
`
`2. PATHOGENESIS OF GAS'l‘ROI'N'TESTINAL DAMAGE
`
`Topical and systemic elfects contribute to the mucosa] damage associated with
`NSAID use. and although g'en.erally considered a primary cause of gastroduode-
`nal ulcers. NS/-‘tIDs are capable of inducing widespread injury throughout the gas-
`trointestinal tract and liver. Lanes er al.. 1992 recen-tly reported that approximately
`35% of NSAlD—induced gastrointestinal bleeding occurred below the ligament of
`Treitz. with a spectrum of small intestinal and colonic injury ranging from a protein-
`and blood-losirtg enteropathy and colitis resembling inllamtnatoty bowel disease to
`colonic perforation and bleeding.
`ln the esoph-agus.
`injury- induced by NSAIDS has been reported and is thought
`to result from local acidic injury to the esophageal mucosa, a condition termed pill
`esophagitis (Minoch at £11.. 199]).
`[n the stomach and duodenum.
`topical
`injury
`results from both direct and indirect mechanisms. In the presence of highly acidic
`gastric contents, nonionized lipophilic NSAIDS are easily absorbed and migrate
`through gastric mucus to_ the surface epithelium, where NSAIDS are dissociated
`into the ionized forth, resulting in I-1+ ion tr-apping (Somasundaran1.erciI., 1995).
`In addition. NSAIDs reduce the hydrophohicity of gastric mucosa and render the
`surface e.pitheli.u-m susceptible to injury by gastric acid (Lichtenberger. 1995).
`Indirectly, mucosa] erosion may result from duodenogastric reflux of bile containing
`active NSAID metabolites (Wolfe er a!.. I-999)-.
`
`Although topical effects of NSAIDs are a significant cause of mucosal damage.
`the fact that NSAID-induced -gastroduodenal
`injury occurs with equal frequency
`using enteric-coated preparations and following either rectal or parenteral admin-
`istration suggests that the systemic effects of these compounds may play an even
`greater role. This contention is further supported by the observation that prodrugs
`such as sulindac. a compound whose effect is mediated by an active metabolite,
`have been associated with gastroduodenal ulceration.
`The mucosal production of prostaglandins (PGs)' appears to play an integral role.
`in the stimulation of several muco-sal components necessary for maintaining normal
`mucosa] integrity and possibly reducing gastric acid secretion (Table I) (Wolfe
`er u!.. I999)-. PCs are derived from arachidonic acid by the action of cyclooxygenase
`(COX). Two related cyclooxygenase isoenzymes. COX-1 and COX-2. are expressed
`in mammalian cells. COX—l
`is expressed conslitutively in most tissues. including
`the gastric mucosa. and is thought to function as a ‘housekeeping’ enzyme. The
`expression of COX-2, especi-ally in macrophages and synovial cells.
`is induced
`
`
`
`l'VImtt'tgt'm(4ttf r;fN.S'.=lID-i'c’l:iml' GI l'o.t:it‘it_\'
`
`279
`
`'l'al:Ilt' I.
`
`Pmteclive properties of the gztstrttdttodenitl mttcosa
`
`Mucus secretion
`Bicarbonate secretion
`Mucosal blood How
`
`lntercellular tight junctions
`Maintenance oftissue acid—_ha-se status
`
`Epitlielial restitution
`
`Table 2.
`
`Risk l'actors tin.‘ d'cvelop.mcnt of NSAlD~rela'tetl ulcers
`
`Definite
`
`Advanced age
`Prior history of ulcer
`Coticoniitattl corticosteroid therapy
`C'.o11comitant anlicoagtilzltion therapy
`High doses of NSA-.lDs
`Short clttration of therapy ( < two -weeks}
`Serious systemic diseases
`
`Possible
`
`Concotnitant H. p_\-'i'm'i ial'eclion
`Snioking
`Alcohol
`
`it has been hypothesized
`by inflammation or mitogen stimulation-. Therefore.
`that
`the anti—inll—a1nn1atoty action of .NSAIDs are secondary to their inhibitory
`effect on COX-2, w'hi"l.e their adverse properties are the result of the inhibition of
`COX-I. Despite strong evidence that PGs play an important role in NSAID-induced
`gastrointestinal toxicity. recent studies performed in COX-l (—I—) mice did not
`demonstrate spontaneous ulcer forrnatiort. suggesting that other mechanisms may
`be involved as well (Wolfe at al.. I999}.
`
`3. RISK FACTORS FOR NSAID-RELATED GASTRODUODENAL TOXICITY
`
`Patients who use NSAlDs develop gastrointestinal complications at a rate three
`times higher than those patients who avoid these medications. Unfortunately.
`the presence or absence of gastrointestinal symptoms does not correlate with
`.gastrodu'od_enal pathology, and for this reason.
`it is importzuit to identify patients
`at an increased risk of developing complications (Table 2).
`Although "corticosteroids alone do not appear to "cause gastroduodenal ulcers.
`their cornbination with NSAlDs is associated with a more than ten fold increase in
`
`gastrointestinal complicat.ions. Age also seems to play a significant role in NSAlD-
`associated complications, with data revealing a 56-fold increase in gastrointestinal
`complications in patients over the age of 70. A prior history of gastroduodenal
`ulcers is another important risk factor for the development of NSAID~re-lated ulcers
`
`
`
`280
`
`A. E Bcn'r.i.\'rm trim’ M. M.
`
`lrl-'ol_['e
`
`[Wolfe er m'.. 1999). Wh.ether a history of Helicobrmi-er p_rlort'—related ulcers further
`increases this risk is not known. In a group of H. p_.v-lm-1' infected individuals. Chan
`et al..
`[997 recently reported a significant decrease in naproxen-induced ulcers
`in individuals whose H. p_vl'ori infection was eradicated. Further investigation is
`needed to determine whether screening for H. p_vt'.m'i' infection pri.o.r to i-nstituting
`NS-AID therapy is indicated.
`In addition. the type. dose. and length of therapy appear to influence the rate of
`complications. Several studies have reported a greater risk" of cornplications with
`p-iroxicam a-nd a relatively lower risk with the u.se of ibuprofen. While the risk of
`complications is also proportional to the dose of the NSAID given,
`it appears to
`-be inversely proportional to the duration of therapy (Wolfe re: ul'.. 1999). Therefore.
`those patients at greatest‘ risk for complications may be the elderly who require short
`or intermittent courses of NSAIDs at high doses.
`
`4. TREATMENT OF NSAID-ASSOCIATED DYSPEPSIA
`
`-Dyspepsia is a broad term used to describe an array of upper abdominal symptoms
`including pain, cramping, _distenti_on, nausea, anorexia andlor heartburn. At least
`
`5~20% of patients taking NSAIDS in large epidemiological studies encountered
`dyspeptic symptoms (Wolfe et ul.. 1999). Several studies have evaluated the effect
`of Hg-receptor antagonists in treating NSAID-related gastrointestinal symptoms.
`Bijlsr-n‘-a at at‘. (1988) performed a _prospective. double-blinded study evaluating the
`effectiveness of cimetidine in treating NSAl'D—relate_d dyspeptic symptoms. They
`found that 72% of patients receiving eimetidine 400 mg BID reported resolution
`of their symptoms. while only 49% of those receiving placebo became symptom
`free. Taha er al‘., 19.96 examined famotidine 20 mg BID and 40 mg BID in patients
`with arthritis receiving NSAIDS and found that abdominal symptom-s decreased
`by 36.6 % and 43.3%. respectively. These studies thus support the use of Hg’
`antagonists in in-dividuals with NSAID associated dyspepsia. As discussed below,
`proton pump inhibitors (PPls) also provide excellent symptomatic relief in addition
`to their capacity to decrease the incidence of NSAID-induced ulcers.
`
`5. TREATMENT OF NSAII)-ASSOCIATED ULCERS
`
`"Since abdominal symptoms do not reliably predict the presence of NSAID-related
`gastroduodenal. ulcers. patients with acute ulcers should have their NSAID therapy
`discontinued whenever possible or should have their current NSAID substituted
`with a nontoxic analgesic such as acetaminophen or salsalate. a non-acetylated
`salicylatc that does not
`inhibit prostaglandin synthesis.
`If NSAID therapy is
`discontinued, treatment -directed at healing the ulcer can be instituted with one of
`several agents discussed below. and at rates that compare favorably to those patients
`with 'idiopathic' peptic ulcers (Lancaster—Smith_et at‘-.. 1990).
`
`
`
`Mitt-r:‘.'go-:li.t':ri Q,’'NS/1. ID-ns*la_tt='rl' GI m.w't'r'r_\-‘
`
`281
`
`5. )7 . SrtL'r'aJfi:I.te
`
`Sucralfate. -a basic aluminum salt of sucrose octasulfate, is effective in the treatment
`and prophylaxis ofduodenal ulcers and appears to be as "effective as Hyantagonists
`in the healing of gastric ulcers. In patients who remain on NSAI[)s, sucralfate has
`been found to heal duodenal ulcers as effectively as Hg—antagonists; however, its
`benefit in healing gastric ulcers in this setting has not been proven (Wolfe at al.,
`1999};
`
`5.2. H3 — 'r'er.'eprm' czrttcagcurisrs‘.
`
`Several open. uncontrolled, non-randomized studies (Croker etal.. 1980) and
`prospective,
`randomized studies {Davies erul., 1986) have demonstrated that
`treatment with conventional doses of l-Ig—recepto1' antagonists for 6 to I2 weeks
`results in healing ‘of approximately 75% 60-88%)" of gastric ulcers and 87%
`(67—l0()%) of duodenal ulcers despite the continued use of NSATDs. When
`NSAIDS are continued. healing appears to be delayed and is largely dependent
`on the initial ulcer size with ulcers greater than 5 mm in diameter healing at a
`substantially lower rate than those less than 5 mm.
`
`5.3. Pro.mrgIand1'ns
`
`The role of PGs- in the treatment of NSAlD.—associated ulcers is not -well known. In
`I998. Hawkey at 4:11.. 1998 published a double—blind study of935 patients comparing
`the efficacy of misoprostol
`to omeprazole in the healing of NSAID-associated
`gast.roduode.nal ulcers or erosions (greater than 10 erosions). The number ofpatients
`successfully treated after e_ighI weeks was somewhat superior in those receiving
`omeprazole cornpa-red to those administered 200 gig QID of misoprostol. Healing
`of gastric ulcers occurred in 89%, 89%, and 77%, in those treated with 20 mg of
`omeprazole. 40 mg of omeprazole. and misoprostol, respectively. while healing of
`duodenal ulcers was 80%. 87%, and 73%, respectively, in these groups. Of note.
`omeprazole was better "tolerated than misoprostol.
`
`5.4. Pmmn pump i'n_l1ibt'l‘.or's
`
`that PPls possess an improved capacity to heal NS.-MD.-
`Recent data suggests.
`related gastrodu-odenal ulcers independent of whether the NSAID is continued
`or not. A recent study comparing omeprazole to ranitidine in 541 patients with
`gastroduodenal ulcers demonstrated healing rates of 80% and 79% .in patients
`treated with -omeprazole 20 mg and 40 mg. respectively, while the healing rate with
`r'anitidine' was 63% (Yeomans er a1.. 1998). Ag-rawal at al‘..
`I998 compared the
`efficacy of lansoprazole and ranitidine in the healing ofgastric ulcers greater than
`0.5 cm in diameter in patients continuing NSAID therapy. After 8 weeks. ulcers
`were healed in 73% and 75% of those patients treated with lansoprazole 15 mg -and
`30 mg, respectively. while the healing rate in the individuals receiving. ranitidine
`was 57%.
`
`
`
`282
`
`A. E 3c!.f‘.!'i.\'r)H and M. M.
`
`lrl/«rift»
`
`6. PROPHYLAXIS AGAINST NSAID-INDUCED GASTRODUODENAL ULCERS
`
`Because of the potential risks associated with NSAID use. efforts have recently
`focused on the prevention of mucosal injury. As mentioned earli.er. prophylaxis is
`best accomplished by the avoidance of these medications i.n patients at increased
`risk from their toxicity. Alternative strategies t.o be considered -include the co-
`administration of an agent with protective properties or the .deve.lopment of new
`anti-inflammatory agents with improved safety profiles.
`
`6. I . Stt.ct'aIfitte
`
`The exact mechanism by which sucralfate protects gastroduoden-al mucosa is not
`known. Sucralfate has been shown by Caldwell er nt!'., 1987 to reduce gast__roduo—
`denal mucosa] injury associated with NSAIDs. However. -Agrawal at m'.. 1991 re—
`ported no significant beneficial effect of sucralfate in the prevention of gastric ulcers
`in osteoarthritis patients receiving NSAIDs.
`
`6.2. H1-I'€C€pI0r
`
`a.rt.mgm1isI.s‘
`
`In two placebo—contr0l]ed. prospective studies investigating the protective effects of
`ranitidine in arthritis patients treated with NSAIDS. an eight-week course of rani—
`tidine 150 mg BID was demonstrated to be effective in preventing duodenal ulcers
`with rates of 0% and l.S% in the treated patients, compared to 8% in the placebo-
`trented patients (Robinson et ul., 1989', Ehsanullah er al.. 1988).. In contrast. raniti-
`dine was ineffective in preventing gastric ulcers in both studies. Furthe-nnore. Taba
`c=taI.. I996) recently reported a beneficial effect of high does famotidine (40 mg
`BID) in preventing both gastric and duodenal ulcers in arthritis patients receiving
`NS,AlDs for 24 weeks. Although I-Iyreceptor antagonists appear to be effective
`in reducing NSAID—related ulcers and dyspeptic symptoms, Singh er at‘.. 1996 re-
`cently found that asymptomatic rhea-tnatoid arthritis patients receiving H3-receptor
`antagonist-s had a significantly higher risk for developing GI coniplications than
`those‘ not taking these drugs. Thus. the murt'm' use of H_3~recept'or antagonists in the
`prevention of NSAID-associated ulcers cannot be recommended.
`
`6. 3. Mi.s'opms!u.l
`
`I988 examined the prevalence of gastric ulcers in patients with
`Graham eru!..
`osteoarthritis who had abdominal pain and were receiving NSAIDS. Patients were
`prescribed rnisoprostol I00 tag QID, misoprostol 200 pg QID. or placebo. Ulcers
`were detected in 1.4%, 5-6%. and 21.7%. respectively.
`In a subsequent study of
`638 patients with chronic arthritis, Graham er al.
`(1993) further demonstrated
`that misoprostol significantly reduced the incidence of duodenal ulcers from 4.6%
`in those taking placebo to 0.6% in those taking misoprostol. Despite these
`encouraging results. dyspeptic symptoms did not improve in these s-tudies, and
`diarrhea developed in 39% of patients taking 200 [L g of rnisoprostol.
`In a study
`
`
`
`Mtriirmt-riitertr nj'NSA1D—i-elated G1 m.\'ic-ity
`
`2-83
`
`by Raskin at at. ([995) evaluating the prophy'l.actie effect of misoprostol at doses
`of 200 pg BID. TID and QID. the authors found that lower doses of m_isopro'stol
`were better tolerated. but that a significant protective effect was only seen with
`the TID and QID regimens. Although the results of the above studies suggest
`that tTliS0pt‘OSlOi is indeed effective in preventing NSAID—induced gastroduodenal
`ulcers. it is associated with significant side effects which include not only diarrhea
`and abdominal pain, but also increased uterine corttractility that can lead to
`spontaneous abortion. Misoprostol is thus contraindic-ated in women of childbearing
`age who are sexually active.
`
`6.4. Proton pump -fr:-I'ribiror.r
`
`Yeomans er ai., 1998 evaluated the prophylactic effect o'l'omepra'zole compared to
`that ofranitidine in 432 patients who were randomly assigned maintenance therapy
`with either 20 mg QD of omeprazole or 150 rng BID of ranitidine. After six months.
`16.3% and 4.2% of those given runitidine developed gastric and duodenal -ulcers.
`respectively. but only 5.2% developed a gastric ulcer and 0.5% a duodenal ulcer in
`the omeprazole group (Fig. I). Another study compared the efficacy of omeprazole
`and inisoprostol in the prevention of recurrent ulcers in 732 patients who continued
`to receive NSAIDS. Patients were randomly assigned either placebo. 20 mg QD
`of omeprazole, or 200 pg BID of misoprostol. At six months, duodenal ulcers
`were detected i-n 12% and 10% of those treated with placebo and misoprostol,
`respectively. while only 3% of those treated with omeprazole developed a duodenal
`ulcer. Gastric ulcers recurred in 32%. 10%. and I3% of the individuals receiving
`placebo, misoprostol. and omeprazole. respectively. These studies suggest that
`omeprazole is effective in preventing both initial ulcer formation as well as ulcer-
`recurrence in patients who continue N SAID use.
`
`hiC3
`
`4. U1
`
`RELAPSEmy 8 D
`6-MONTH
`
`5
`
`GU
`
`DU
`
`Figure 1,. R=esulls-of it randomized. double-blind. plucebwcontrolled trial that contpared the ability
`of ranitidine 150 mg BID and omepruzole 20 mg QD to prevent the cleveloprneht of NSAID-iriduced
`gastrodnodenal ulcers. ‘Data adapted from Yeornuns er all ([993). *p -< 0.00].
`
`
`
`
`
`284
`
`A. E Brzrrisrirz and‘ M. M. Wmffe
`
`fig
`35
`hi)5
`NU’!
`
`
`
`
`
`
`5-mmulcers("M AM(II9
`
`
`
`Placebo
`
`I Rofeeoxib 25 mg
`
`Rofecoxib 50 mg
`1:: Ibuprofen 2.4 2
`
`24 Weeks
`
`
`Figure 2. Incidence ol'ga.-umduodena] ulcers t:- 5 mm in diameter). The placebo group was observed
`For [2 weeks only. Data adapted from Laine rt ml. (in press). *p < ().()l)l L‘£JI'I‘lp:it’i:‘(l to both doses ol‘
`rofecoxib.
`
`6.5. Preferential and .\'elec'ii'i!e ‘COX-2 iriliihirrarts
`
`Attempts have been made to reduce NSAID toxicity through novel changes in their
`fonnulation. Recent surv.eiIlan'cc -and endoscopic studies have confirmed a lower
`incidence of gastroduodenal mucosa] injury with the use of nabumetone. etodolac,
`and meloxicam (Wolfe er al., 1999). These drugs preferentially inhibit COX-2, with
`less effect on COX-1, which appears to account for their improved safety profile
`New highly selective COX-2 inhib.itors_. celecoxib and rofccoxib, recently entered
`the market and been shown to reduce the incidence of NSAID-—relate_d endoscopic.
`
`ulcers (Fig. 2}.
`
`7. FUTURE DEVELOPMENT OF SAFER NSAIDS
`
`Several other -compounds are in development with hopes of increasing NSAID
`safety. These include nitric oxide (NO) releasing NSAIDS and NSAIDS prea.sso-
`ciated with zwitterionic phospholipids. Trefoil peptides and basic fibroblast growth
`factor ('bFG_F) are also being examined for their potential clinical role in protecting
`against NSAID—induced mucosa] injury.
`
`8. CONCLUSION
`
`that
`is important
`it
`Because of the pervasive use "of NS/*\ID_s in our society,
`physicians be able to identify and manage patients affected by NSAID-related
`gastrointestinal toxicity. Dyspepsia is a relatively common cornplic-ation associated
`with NSAID use which is be best managed with either an Hyreceptor antagonist or
`a PPI. Treatment of confirmed gastroduodeiial ulcers should include withholding the
`NSAID when possible-and providing a PPI. Once ulcer healing has been -achieved.
`if NSAID therapy is to be re—instituted, prophylaxis against gastroduodenal toxicity
`
`
`
`Ma:-iictg_c'::ievr-I 'ofN.S‘A1D-relotert GI rrm‘z'i!_v
`
`285
`
`Table 3.
`
`Recommendations for nmnagemenl of NSAID-ztssociated ;g:Jstrr.)du_odent{l mucosa]
`in'jun_.!
` jjj.
`
`C‘lin,it:aJ pres_entati.on
`
`Recommendation-
`
`Dyspepsia
`
`Empirical H3-—reeeplo-r antagonists t-e.g. cimetidinc -400
`mg; BID. ranitiilinefnizat-idine LSO mg BID. or famo-
`tidine 20 mg, BID) or proton pump inhibitor te.g,.
`oineprazzole 20 mg ()1).
`lansopruzule 30 mg QD-.
`ritbeprttzole 20 mg QD. or panloprnzole 40 mg OD)
`
`Active ulcer
`
`NSAID discontinued: Hg-receptor antagonist or proton
`pump inhibitor
`
`NSAI-D continued: proton pump inhibitor
`
`Mztintenztrtee -therapy
`
`C‘oatlmini::tra..ti'(-In of" proton pump inhibitor
`
`Coarlministnttion Oi-‘t'l‘ll.50pf05lUl (at least 200 [1 g TlD'l
`
`COX-2 ptel’erenItiatl- ‘agent (e.g. mtbumetone. etodol-aw.
`or meloxicam)
`
`COX-2 ‘selective ngent (e.g. eelecoxib nnd.1'oli:emtib)
`
`is best accomplished through .the use of a PPI, misoprostol (at least 200, pig TID). or
`an NSAID that preferentially or selectively inhibits COX-2 (Table 3).
`
`REFERENCES
`
`~t I991). Misopmntol compared with sucralfate
`Agrawal. N. M.. Roth. S. 14.. Graham. D. Y.. «I cu’.
`in the prevention of non-steroidal ztnti-int'ia_tmm'ntory dmgeindueed gztrstrit: ulcer:
`tl randomized.
`controlled trial. Arm. frrtmrn. Med. 115. 9-1 1—_3.
`
`t 1998). Efl'ectivene.~:.-a of lansopmzole in the healing of
`Agrawnl. N. M.. Safadi. M.. Wruble. L.. at ail.
`NSAID-induced gaittric ulcers in patients continuing to take NSAIDS. Grmrrmuzmilug)‘ 114. A52.
`-Caldwell. .l. R.. Roth, 5.
`I"-I.. Wu, W. C., at rt}. (I987). Sucralfztte treatment of non-ste1'oidai tutti-
`iuflitmmatory drugvindueetl gzutroirttestinztl sy.n1pmms and mucosztl ‘damage. Ainc-n J. Med. 83
`(suppl 313). 744312.
`
`Cliuug. S. C. S.. at at-. (‘I997-). Randonnized trial of eradication of
`Chan. F.. K. L.. Sung. J. J.
`Hc=lic'a!>nt‘r:e:- pyl.-arr‘ helfore non—st_eroid -ant.i-inflammatory drug therapy to prevent peptic ulcers.
`:‘..w.tc'er 350. 975-9.
`
`Croker. J. R.. Cotten. P. B.. Boyle. A.
`arthritis. Ann. Rlmmn. iD:'.r. 39. 275-8.
`
`er (:1. (W80). Ci1netid'i=11e for peptic ulcer in patients with
`
`Dav-ies. J.. Collins. A. J. and Dixon. A. S. J. (1986). The influence ‘oi’ cimetidine on peptic ulcer in
`patients with arthritis taking anti-inflammatory drugs. Br.iL- J. Rhemmrml. 25. 544%.
`Ehsetnulluh. R. S. B.. Page. M. C.. Tildesly. (3.. er al. {I988}. Preventi'on of gastrointestinal daninge
`induced by non-steroidal ztnti-ittflumtitntory drugs: coittrollecl lriztl oi"ran'i-ti'd'ine. Brit. Med. J’. 297.
`l()l7'—2t.
`
`Graham. D. Y.. Agruwal, N. M. ztnd Roth. S. H. (1988,). Prevention of NSA[DAinc[uecd gastric ulcer
`with misopmstol: multicenter. double-blind. platebo-con'troll'e(li trial. I.a.u.r*et 2.. IZT/-80.
`
`
`
`286
`
`A. E Bm'ri.m:z «ml M. M.
`
`‘Wz:i_'}"e
`
`G'r'.tha=tm. D. Y.. White. R. H.. Moreland. L. W.. at all ([993). Duodenal and gastric ulcer prevention
`with tnisoproeatol in t1rtI'1t‘itis taking NSAIDS. Aim. lrmrrr. Marti. 119. ’.!57—62-.
`Grifhn. M. R.. Roy. W. A. and Schatiner. W. ([988). Non.~'.t-eroid m1_l'i-inliantmzttury drug use and death
`from peptic ulcer‘ in elderly persons. Ann. Intern. Mm‘. 109. 359433.
`I-iawkey, C. 1.. Kurrnsch. J. A.. Szczepunski. I;., 1-Jul‘. ( 1993). Omcpraznle compared with misoprostol
`for ulcer.-t associated with non-steroidal anti-ittflatmimttory clrugs. Nr!t1‘Ertgl. J. Mm’-. 338. 727-34.
`Laine. 1... Harper. 5.. Simon. T.. at £11. (in press]. A randomized trial comparing the e'l"l'ect-of refecoxib,
`at COX.-2 specific inhibitor. to ib:.1pr0'l"en on the gastrocluodettal mucn.-«;:1 of osteoarthritis patients.
`G£L\‘!."tI€r2I(:'l'r}."0g_\‘.
`Lani-15, A__ 5-gt;-mg 1V[_ C_ -“mi. Hirshuwitz. B. I. (1992). Objective evidenceof i-tspiri'n use in both‘ ulcer
`and nomticer upper and lower gastreintestinzti bleeding. G(I.ttmc*nl't.'J‘m'ug_\= I03. 862- 9.
`L_i1nc_21ster-Stttith. M. J.. Juclerberg, M. E-. and Jzickson. D. A. (I 990), Rzt.nit'idine' in the treatment of
`non-steroid anti-infi:1m111atory drug associated gmitric and dutidenttil ulcers. Cut. 32. 252-5.
`Lictitenberger. L. M. (1995)-. The 3hydroph'oE.1ic barrier properties 0l'.gastroit1.testitnal muctls, Aimrmi
`Rev. FJ'i_v.w‘nI. 57. 565-8.
`t 1991). Pill-esopha-gitis caused by nun-steroithtl
`Minnch. PL. Greenbaum. D. S.
`drugs. Amm: J’. Grtstt'cw.rtter-nlagy. -86. 1086-9.
`Ruskin. .F. B.. White. R. H.. Jackson}. J. '13.. at all. (I995). Misoprotatui tlusttge for the prevention of
`nsah-stem-idai anti-'i'nfia1nmzttory' druig-itntluced gastric and dttodenztl ulcers: a cnntpuris-sen of three
`regimen.-5. Amt. Imcrti. Med. 123. 34-4~5().
`Robins(_m. M. G.. Griffin. J. W.. Bowers. J.. e! m’. (I989). Effect of rzmitidine on gatstroduodenitl
`mucosa] dainage induced by non-.~a'teroiclal untieinflztntmzttnry drug-5. Dig. Dis. .S't'i'. 34. 4244}.
`Singh. (3.. Rztmey. D. R.. Merfeld. D.. at as’. (1996): G':mrointesti‘nztl tract contplications of non~
`steroidal anti—inflan1matory drug treatment
`in rheunttttoid a.rth'r-it_i.~;: a prospective ehserviitional
`cohort study. Arch. litre:-rt. Med. 156. 1530-6.
`Somasundar-tun. S._. Hayllar. I-1.. Rafi. 5.. anti. (1995). The biochemieztl basis of n(iI1.~slereith1~| anti-
`infiztmrnzttory drug-induced damage to the ga.~ttr.uintestinal tract: a review and a hypoléhesis. .Sfc‘ctrrri-.
`J. Grimm:-:1re*r'ulag_v 30. 289-99.
`Tztha. A. 5.. H'uds.nn.
`H-awkey. C. J., at til. (I996). F'.tm.0t'idine for the prevention of gastric and
`duodenal ulcers catised by nort-stcroidai uuti-inflammatory drugs. New Engll J. Med; 334. I4-3549.
`Wolfe-. M. M.. Lichtenstein. D. R. and Singh. G. (1999). Gztstrointestittal toxicity of non-tstereidztl
`ttnti-iufianlm-zttory drugs. New Etigl. J. Med. 340 (24). |888+99.
`Yetirnans. N..D.. Tulassay. 2.. Juhasz. I... et ul. (1998). A cmnparison of qmeprazole with tanitidine
`for ulcers associated with non-.~;teroid'al ztnti-infia'mma_1ury drugs. New Engl. J. Med. 338; 719-26.
`
`flltli-illii£tt.t117l1ui-()t’)l