Tie AMERICAN JOURNAL OF GASTROENTEROLOGY
`Copyright © 1998 by Am.Coll. of Gastroenterology
`Published by Elsevier Science Inc.
`
`Vol. 93, No. 11, 1998
`ISSN 0002-9270/98/819.00
`PLL $0002-9270(98)00469-9
`
`A Guideline for the Treatment and Prevention of
`NSAID-Induced Ulcers
`
`Frank L. Lanza, M.D., F.A.C.G.
`Baylor College of Medicine, Houston, Texas
`
`and the Members of the Ad Hoc Committee on Practice Parameters of the American College of Gastroenterology*
`
`PREAMBLE
`
`INTRODUCTION
`
`Guidelines for clinical practice are intendedto indicate pre-
`ferred approaches to medical problemsas established by sci-
`cntifically valid research. Double-blind, placcbo-controlled
`studics are preferable, but compassionate usc reports and ex-
`pert review articles are used m a thorough reviewof the
`literature conducted through Medline with the National Library
`of Medicine. When only datathat will not withstand objective
`scrutiny are available, a recommendation is identified as a
`consensus of experts. Guidelines are applicable to all physi-
`cians who address the subject without regard to specialty train-
`ing or interests and are mtcnded to indicate the preferable, but
`not necessarily the only, acceptable approach to a specific
`problem. Guidelines are intended to be flexible and must be
`distinguished from standards of care, whichare inflexible and
`rarely violated. Given the wide range of specifics in any health
`care problem, the physician must always choosethe course best
`suited to the mdividualpaticnt and the variables in cxistence at
`the moment of decision.
`
`Guidelines are developed under the auspices of the Ameri-
`can College of Gastroenterology and its Practice Parameters
`Committee and approved by the Board of Trustees. Each has
`been intensely reviewed and revised by the Committee, other
`experts in the field, physicians who will use them, and spe-
`cialists in the science of decision analysis. The recommenda-
`tions of cach guideline are therefore considered valid at the
`time of their production based on the data available. New
`developments in medical research and practice pertinent to
`each guideline will be reviewed at a time established and
`indicated at publication to assure continued validity.
`
`* J. Patrick Waring, M.D., Atlanta, GA; James T.. Alchord, M.D.,
`Jackson, MS; Todd H. Baron, M.D., Birmingham, AL; Eugene M. Bozym-
`ski, M.D., Chapel Hill, NC; Patrick G. Brady, M.D., Tampa, FL; W. Scott
`Brooks, Jr., M.D., Atlanta, GA; William D. Carey, M.D., Cleveland, OH;
`Kenneth R. DeVault, M.D., Jacksonville, FL; Norman D. Grace, M.D.,
`Boston, MA: John I. Ilughes, M.D., [louston, TX; Simon K. Lo, M.D.,
`Torrance, CA; Jack A. DiPalma, M.D., Schenectady, NY; George W.
`Meyer, M.D., Atlanta, GA; John F. Reinus, M.D., Bronx, NY; Marvin M.
`Schuster, M.D., Baltimore, MD; Douglas M. Simon, M.D., New Rochelle,
`NY; Robert H. Squires, Jr. M.D., Dallas, TX; Rosalind U. Van Stolk,
`M.D., Cleveland, OH; John Wo, M.D., Atlanta, GA; and Gregory Zuccaro,
`Jr., M.D., Cleveland, OH.
`Received June 24, 1998; accepted Aug. 12, 1998.
`
`The relationship between nonsteroidal anti-inflammatory
`drugs (NSAIDs) and gastroduodenal
`injury is now well
`established (1, 2). Patients with rheumatoid arthritis (RA)
`and osteoarthritis (OA) taking NSAIDs have an ulcer inci-
`dence of approximately 15-20% (3, 4). Complications of
`ulcer disease, i.e, hemorrhage and perforation, occur far
`more often in patients taking these agents than in compa-
`rable control groups (5-7). The overall risk for serious
`adverse gastrointestinal
`(GI) events in patients taking
`NSAIDsis about three times greater than that ofcontrols. In
`elderly patients (>60 yr), this risk riscs to more than five
`times that of controls, whereas the risk for younger patients
`is only slightly more than one-and-one-half times (6). In
`elderly patients taking NSAIDs the relative risk of GT sur-
`gery is ten times, and for GI cause of death, about four-
`and-one-half times greater than in control groups (6). Ap-
`proximately 20 million patients in the US take NSAIDs on
`a regular basis; the risk for hospitalization for serious GI
`adverse effects is 1-2%, resulting in approximately 200,000
`to 400,000 hospitalizations per ycar at an average cost of
`$4,000 perpatient, or 0.8—1.6 billion dollars annually (8, 9).
`The economic impact of this problem is increased by mul-
`tiple other factors, including lost wages, postop care, efc.
`Two major issues confront clinicians using these agents:
`1) the prevention of NSAID-induced ulcers, especially im
`high risk groups, and 2) their treatment, often when under-
`lying disease mandates continued NSAID use. A third prob-
`lem is the recent recognition of small bowel and colon
`NSAID-related mucosal injury. The purpose of this guide-
`line is to make recommendations based on the pertinent
`medical literature addressing these problems.
`
`PREVENTION OF NSAID-INDUCED ULCERS
`
`Recommendation
`
`Patients at high risk for hemorrhage and perforation
`from aspirin and other NSAID-induced ulcers should be
`consideredfor prophylaxis with misoprostol. Proton pump
`inhibitors are an acceptable alternative for prevention of
`NSAID-related complications. H2 receptor antagonists
`have been shown to prevent only duodenal ulcer, and there-
`
`2037
`
`Page 1 of 10
`
`MYL-EN000041037
`
`Patent Owner Ex. 2064
`Mylan v. Pozen
`IPR2017-01995
`
`Page 1 of 10
`
`Patent Owner Ex. 2064
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`2038
`
`LANZA ef al.
`
`44
`12.7
`
`TABLE 1
`Factars Related to Increased Risk ofNSAID Induced Gi Complications
`Risk Factor Ref # Relative Risk 95% CI
`
`
`
`2.74
`2.54-2.97
`Overall
`6
`Age (>60)
`5.52
`4.63-6.60
`6
`Prior GI event
`4.76
`4.05-5.59
`6
`High dosage (>2 X
`10.1
`4.6-22.0
`43
`normal)
`Concurrent corticosteroids
`Concurrent anticoagulants
`
`2.0-9.7
`6.3-25.7
`
`44
`45
`
`Jore cannot be recommended for prophylaxis. Factors that
`have been identified as placingpatients at increasedriskfor
`NSAID-related GI complications include the following:
`
`1. Prior history ofgastrointestinal event (ulcer, hem-
`orrhage)
`i) . Age >60 yr
`
`3. High dosage
`
`4. Concurrent use of corticosteroids
`
`5. Concurrent use ofanticoagulants
`
`Although a direct link has not been established between
`NSAID-induced ulcers and gastromtestinal hemorrhage and
`perforation,
`these complications occur significantly more
`often in patients taking NSAIDs when compared with con-
`trol groups. As noted above, these serious adverse events
`also tend to occur more often in high risk groups which
`include patients with a prior history of a GI event, advanced
`age (60 yr), high NSAID dosage, and concomitant use of
`corticosteroids or anticoagulants (Table 1). Several random-
`ized controlled trials (RCTs) have been published attempt-
`ing to showthat various therapeutic mancuvers can prevent
`the development of gastric ulcer (GU) and duodenal ulcer
`(DU)in patients taking NSAIDs. However, only one study
`has been published linking any cotherapy with prevention of
`the complications ofulcer disease, i.e., bleeding and perfo-
`tation (10). It is not unreasonable, however, to assume that
`prevention of NSAID-induced ulcer should be associated
`with a similar prevention of the complications of ulcer
`disease.
`Numcrous RCTs have been carricd out, both in normal
`volunteers and in patients with arthritic disorders, to eval-
`uate the efficacy of coadministration of various agents for
`the prevention of both NSAID-related, nonulcer gastropathy
`and GU and DU (11-28). Generally, these have shown that
`concomitant admimuistration of the prostaglandin El analog
`misoprostol along with various NSAIDs can prevent both
`GUand DU(11-23); that the proton pumpinhibitor (PPT)
`(24-26) omeprazole reduces GU and prevents DU;and that
`H2 receptor antagonists are cffective in preventing DU but
`not GU (27, 28).
`Prophylaxis with prostaglandins or PPIs forall patients
`taking NSAIDsis unnecessary and cost-prohibitive. Studies
`
`AJG — Vol. 93, No. 11, 1998
`
`with misoprostol have shownthat in high risk groups, pro-
`phylaxis may be cost-effective (7). The numerous factors
`involved in these analyses,
`i.e., the time lost from work,
`expense of hospitalization with or without surgery, varying
`costs of prophylactic and therapeutic drugs, and quality of
`life issues make it very difficult to determine the cost
`effectiveness of prophylactic therapy for NSAID-related
`ulcer disease and its complications.
`
`Risk Factors
`
`All patients taking NSAIDs do not require prophylaxis.
`These drugs are used extensively as treatment for limited
`illnesses. Anti-inflammatory doses of numerous NSAIDs
`have been administered for short periods of time (up to 7
`days) to large numbers of young, healthy volunteers without
`any reports of significant GI bleeding or other serious event
`(29). Manypatients of all ages, however, take anti-inflam-
`matory NSAID doses for longer periods, and many case-
`control studies have shownthat, in these patients, significant
`GI bleeding and other severe adverse events occur more
`commonly than in matched control patients (30-37). An
`increased risk of GI bleeding has even been noted in patients
`taking low dose aspirin therapy for cardiovascular prophy-
`laxis (38-41). Obviously, asall patients taking NSAIDs do
`not develop serious complications, risk factors must exist in
`some patients that inercase the incidence of GI blecding,
`perforation, surgery, and even death.
`A series of nested case-controlled studies based on hos-
`
`pitalization for GI hemorrhage of Medicaid recipients aged
`>65 yr in the state of Tennessee showed an inercased
`bleeding risk for patients >65 yr (odds ratio 4.7), imcreased.
`dose (odds ratio 8.0), concomitant corticosteroid (odds ratio
`4.4), or anticoagulant therapy (odds ratio 12.7), and short
`term onset of use (<1 month) (7.2) (42-45). A large au-
`topsy scrics on patients with a history of NSAID usc showed
`that gastric and duodenal lesions were more common in
`patients who took NSAIDsfor <3 months, whereas patients
`with a longer duration of therapy tended to have more
`lesions in the small bowel and colon (46).
`A large retrospective cohort study, also based on data
`from Medicaid patients, confirmed the overall increased risk
`for GT bleeding in patients taking NSAIDs, especially over
`the aged >60 yr (47). Similar data have been obtained from
`other large cohort studics (40, 48, 49). In a large prospcec-
`tive, multicenter study in patients with RA (N = 2747) and
`OA (N = 1091), the principalrisk factors for hospitalization
`for serious GI events were age, prior NSAID GIside effect,
`prior Gl hospitalization, corticosteroid use, and debility
`expressed as a level of disability. The overall risk in this
`study for hospitalization during NSAID therapy in patients
`with RA was 1.58%(5).
`A large meta-analysis of these and other studies showed
`an overall odds ratio for G1 blecding of 2.74 in all paticnts
`taking NSAIDs. Patients with a prior gastrointestinal event
`had an odds ratio of 4.76. Age (>60 yr), concurrent corti-
`costeroids, and shorter duration (>3 months) of NSAID
`
`Page 2 of 10
`
`MYL-EN000041038
`
`Patent Owner Ex. 2064
`Mylan v. Pozen
`IPR2017-01995
`
`Page 2 of 10
`
`Patent Owner Ex. 2064
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`AJG — November 1998
`
`NSAID-INDUCED ULCER TREATMENT AND PREVENTION
`
`2039
`
`16
`
`17
`19
`
`323
`
`465
`
`25
`
`47
`
`V7
`
`10.3
`
`TABLE 2
`NSAID Protection Studies With Misoprastol for Gastric Ulcer
`
`Misoprostol
`Placebo
`Ref.
`
`Dose
`N
`Ulcers
`%
`N
`Ulcers
`%
`PB
`100 pg g.id.
`143
`8
`5.6
`138
`30
`21.7
`0.001
`200 pe y.id.
`139
`2
`19
`0.001
`200 pe g.id.
`320
`6
`1.9
`0.001
`200 pg bid.
`465
`27
`58
`0.05
`200 pg tid.
`474
`15
`3.2
`0.01
`200 pg g.id.
`229
`7
`3.1
`0.01
`21
`200 pg tid. or gid.
`2
`4
`12.5
`38
`1]
`28.9
`0.05
`
`Sucralfate 1 qm g.id.
`18
`200 pg g.id.
`122
`2
`16
`131
`21
`16.0
`0.001
`
`Ranitidine 150 mg rie.
`22 0.01 200 pe g.id. 180 1 0.5 194 11 5.7
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`consumption were all associated with an increased relative
`risk ratio for serious adverse GI events (Table 1). Gender,
`smoking, and alcohol were not found to be independentrisk
`factors (6).
`A large, double blind, randomized, controlled treatment
`preventiontrial in >8000 patients with RA also identified a
`history of cardiovascular disease as a risk factor for UGI
`complications of NSAID use (odds ratio 1.84). In this same
`trial, age >75 yr (odds ratio 2.48), prior peptic ulecr (odds
`ratio 2.29), and prior GI bleeding (odds ratio 2.56) were
`again associated with increased risk (10).
`There appears to be some difference between the various
`NSAIDs with reference to the incidence of significant Gl
`bleeding and other adverse events. Four large cohort studies
`have been published comparing the risk of these complica-
`tions associated with the various NSAIDs. Overall, these
`studies show an increased toxicity for ketorolac and piroxi-
`cam, and intermediate toxicity for naproxen, indomethacin,
`ketoprofen, and diclofenac. Ibuprofenin all studies was less
`toxic than the other agents, but this is probably related to the
`generally lower doses employed with this agent, which is
`available over the counter (50-54).
`No good prospective controlled data for GI bleedmg and
`other ulcer complications is available for the recently intro-
`duced, newer NSAIDs purported to be less toxic to the
`upper GI mucosa (nabumetone, etodolac, oxaproxin). Sev-
`cral large postmarkcting, open label studics involving thou-
`sands of patients m Europe suggest that bleeding rates with
`these agents are in the range of 0.5% (55-59). These studies
`have recently been reviewed in the American literature (60).
`Ofthese agents, nabumetone has been the most extensively
`studied. A 12-wk endoscopic study compared nabumetone
`(1000 mg g.d.), ibuprofen (600 mg g.i.d), and the same dose
`of ibuprofen plus misoprostol
`in 171 patients with OA.
`There was no difference in the numberofulcers found in the
`
`nabumetone and nabumctone/misoprostol groups (one and
`zeTo, respectively), which were significantly less (p < 0.01)
`than the eight ulcers found in the ibuprofen group (61).
`Another endoscopic study compared nabumetone 1000 mg
`
`q.d. with naproxen 500 mg d.i.¢. in RA patients for 4 wk.
`One ulcer was found in the 22 patients taking nabumetone
`and eight in the 30 patients treated with naproxen (p = 0.01)
`(62). Ina third study, 27 paticnts with cither RA or OA were
`followed for 5 yr taking either naproxen 250 mg b.i.d. or
`nabumetone 1000 mg g.d. Ulcers were found in eight of 12
`patients taking naproxen and in one of 15 taking nabum-
`etone (p = 0.02). No bleeding was found in either group
`(63).
`The combination of advanced age with any of the other
`noted risk factors has also been noted to further increase the
`
`probability of ulcer complications in NSAID users, espe-
`cially in the first month of therapy. The relative risk of
`requiring GI surgery or GI cause of death increases dramat-
`ically in patients >60 yr (6). Although a 15-25% incidence
`of gastric and/or duodenal ulcer has been demonstrated inall
`patients taking NSAIDs(3, 4), the bleeding rate is estimated
`by most experts at only 2-4%. In a meta-analysis of 100
`trials of NSAID therapy, bleeding occurred in 24 of 1157
`patients taking active drug (2%), compared with only eight
`of 1103 taking placebo (0.6%) (64). It can be derived from
`these data that about only one in 10 NSAID-induced ulcer
`bleeds. A recent large, multicenter, double-blind RCT in
`>800 arthritic patients taking NSAIDs showed that the
`incidence of perforation, outlet obstruction, and hemorrhage
`was reduced by 40%in those subjects taking misoprostol
`compared with placebo. In this study, these complications
`were seen in 22 of 4404 patients on misoprostol compared
`with 42 of 4439 taking placebo (p = 0.049) (10).
`
`Prostaglandins
`Currently, the only available prostaglandin is misopros-
`tol, a synthetic prostaglandin E1 analog. A series of RCTs in
`normal volunteers have shown that this agent can prevent
`acute NSAID-related erosion and ulceration (11-15). Sub-
`scqucnt RCTs in patients with OA and RA have demon-
`strated that misoprostol was significantly better than pla-
`cebo,
`sucralfate, and ranitidine for
`the prevention of
`NSAID-related GU (Table 2). Misoprostol was also signif-
`
`Page 3 of 10
`
`MYL-EN000041039
`
`Patent Owner Ex. 2064
`Mylan v. Pozen
`IPR2017-01995
`
`Page 3 of 10
`
`Patent Owner Ex. 2064
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`2040
`
`LANZA ef al.
`
`AJG — Vol. 93, No. 11, 1998
`
`TABLE 3
`NSAID Protection Studies With Misoprostal for Duadenal Ulcer
`
`Misoprostol
`Placebo
`Ref.
`
`Dose
`N
`Ulcers
`%
`N
`Ulcers
`%
`PB
`200 pg gid.
`320
`2
`0.6
`323
`15
`46
`0.002
`200 wg bid.
`465
`10
`2.2
`0.05
`200 pg tid.
`474
`12
`2.5
`0.05
`200 pg gid.
`229
`2
`0.9
`0.01
`
`Ranitidine 150 mg bid.
`2
`1d
`
`26
`
`3.7
`
`465
`
`185
`
`NS
`
`17
`19
`
`22
`
`200 wg g.id.
`
`181
`
`2
`
`1]
`
`TABLE 4
`NSAID Protection Studies With Omeprazole
`
`24
`
`O 20 gd.
`Placebo
`
`
`Ulcers (%)
`Ref.
`Regimen
`
`GU
`DU
`pr
`2/85 (2.4)
`2/85 (2.4)
`>0.005 vs placebo
`6/90 (6.6)
`15/90 (16.7)
`GU
`DU
`35/274 (12.0)
`7/274 (3)
`31/296 (9.5)
`30/296 (10)
`50/155 (30.3)
`19/155 (12)
`11/210 (5.2)
`1/210 (0.5)
`35/215 (16.3)
`7/215 (4.2)
`
`25
`
`O 20 gd.
`M200 bid.
`Placebo
`O 20 gd.
`R 150 bid.
`* All ulcers combined (GU and DU).
`O = omeprazole (in mg); M = misoprostol {in 4g): R = ranitidine (in mg).
`
`26
`
`O .001 vs M and placebo
`
`0.004 vs R
`2/215 (0.9)
`
`icantly better than placebo for the prevention of duodenal
`ulcer. Ranitidine and misoprostol were equally effective in
`preventing DU (Table 3). Lower doses of misoprostol were
`also effective in preventing GU and DU with a side effect
`profile indistinguishable from that of placebo (16-19, 21,
`22). In another double blind RCT, two groups ofarthritic
`patients requiring chronic NSAID therapywith either endo-
`scopically normal (N = 223) or nonulccr-injured gastrodu-
`odenal mucosa (N = 778) were treated with NSAIDs for 2
`wk with concurrently administered misoprostol (400—800
`ug/day) or placebo. The incidence of severe mucosal dam-
`age, including ulcer, was significantly reduced by misopros-
`tol in the previously endoscopically normal subjects and
`also in the group with preexisting nonulcer damage (p <
`0.001) (20). A recent large meta-analysis of RCTs ofpre-
`vention of NSAID-induced gastric mucosal injury by miso-
`prostol or H2 receptor antagonists, published between 1970
`and 1994, showed that misoprostol (but not H2 receptor
`antagonists) was beneficial in the prevention of NSAID-
`induced GUs. It was also found that the numberofpatients
`to be treated to prevent one GU with short and long term
`NSAIDtherapyis 11 and 15, respectively (23).
`
`Proton pump inhibitors
`Omeprazole,
`the most extensively studied PPI, has a
`protective cffect against NSAID-related mucosal injury. Not
`unexpectedly, because of its potent acid-inhibiting property,
`it prevents DU in patients taking NSAIDs. There is evidence
`that omeprazole also protects agamst GU. In acrossover,
`
`double-blind RCT, 20 normal volunteers were given aspirin
`650 mg g.i.d. with either placebo or omeprazole 40 mg/day
`for 14 days, with endoscopy before and after each treatment
`period. Omeprazole significantly decreased aspirin-induced
`gastric mucosal injury (p < 0.01) by protecting 85% of the
`subjects from extensive erosions or ulcer, whereas 70% of
`the subjects developed severe injury (rate 3 or 4 on 0-4
`scalc) on aspirin and placcbo. No duodenal injury was scen
`in any grade or any subject on omeprazole, whereas 50% on
`placebo developed erosions and 15% had DU (p < 0.001)
`(65).
`Three large RCTs have been carried out in patients with
`OA and RA comparing omeprazole with placcbo, misopros-
`tol, and ranitidine for the prevention of GU and DU (Table
`4) (24-26). Overall, omeprazole significantly reduced the
`total number of NSAID-related ulcers when compared with
`placebo and ranitidine (26). It was more effective than
`misoprostol in preventing DU, and equally so m reducmg
`GU(25). It should be noted that the lowest effective dose of
`misoprostol was used in this study, and that most of the
`overall prevention in NSAID-related ulcer in the placebo-
`controlled studics was duc to a reduction in the numbers of
`duodenal ulcers.
`
`AA2 receptor antagonists
`Although commonly coadministered with NSAIDs, H2
`receptor antagonists (H2RAs) have not been shownto pre-
`vent gastric ulcer, the most common NSAID-related lesion,
`but do prevent DU. Ranitidine has been the most extensively
`
`Page 4 of 10
`
`MYL-EN000041040
`
`Patent Owner Ex. 2064
`Mylan v. Pozen
`IPR2017-01995
`
`Page 4 of 10
`
`Patent Owner Ex. 2064
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`AJG — November 1998
`
`NSAID-INDUCED ULCER TREATMENT AND PREVENTION
`
`2041
`
`studied H2RA. In two large, multicenter RCTs, ranitidine
`(150 mg, 4.i.d.) or matching placebo was given to patients
`with OA and RA in conjunction with various NSAIDs. In
`both studies there was no difference in the occurrence of
`
`gastric ulcer between the two groups; however, the inci-
`dence of duodenal ulcer was significantly reduced in both
`studies in the ranitidine group (27, 28). Nizatidine (150 mg
`b.i.d.), in an RCT of 496 patients with OA, did not lower the
`overall incidence of NSAID-related ulcers. However, a sep-
`arate analysis of high risk groups (patients with ulcer history
`and patients >65 yr of age) showedstatistically less GU and
`DU in patients recciving nizatidine (p = 0.035 and p =
`0.042, respectively) (66). In a nonplacebo-controlled RCT
`of 221 arthritic patients with recently healed NSAID-related
`ulcers, cumulative relapse rates for nizatidine 150 mg HS
`and 150 mg b.i.d. were 5.5% and 1.8%, respectively (67). A
`double-blind RCT of famotidine for the prevention of gas-
`tric and duodenal ulcers caused by NSAIDs was performed
`in arthritic patients receiving placebo, famotidine 20 mg
`b.i.d., and famotidine 40 mg 6.i.d., concurrently with the
`usual NSAID. The cumulative incidence of GU was 20% in
`
`the placebo group (n = 93), 13%in the patients receiving
`famotidine 20 mg b.i.d., (NS), and 8% in the group receiv-
`ing famotidine 40 mg b.i.d., (» = 0.03 vs placebo) (68). This
`study, however, has been criticized because of the small
`minimum ulcer size of 0.3 cm. In a similar study from the
`same group of investigators, patients with RA or OA with
`GU or DU were treated with famotidine 40 mg b.i.d. The
`subjects with healed ulcers were then randomized to 6
`months oftherapy with famotidine 40 mg 4.i.d. or placcbo.
`Ulcer recurrence was seen in 26% ofthe subjects on famo-
`tidine, compared with 54% with placebo (yp = 0.01). Despite
`the superiority over placebo, the recurrence rate of ulcer in
`patients on famotidine was unacceptably high (69). High
`dose ranitidine (300 mg b.i.d.) was found to be ineffective
`for the prevention of gastric ulcers, but was protective
`against duodenal ulcer. Patients with rheumatoid arthritis
`and a past history of peptic ulcer disease were followed for
`1 yr on NSAIDs with either high dose (300 mg 4.i.d.)
`ranitidine (n = 10) or placebo (n = 10). Fourpatients in the
`placebo group had recurrent DU and nonein the ranitidine
`group (p = 0.04). Six recurrent GUs were found in the
`placebo group and three in the ranitidine group (p = 0.18)
`(70). Recurrenee of DU in paticnts taking NSAIDs is more
`likely when there is a past history of that disease.
`
`Other agents
`Sucralfate has not been shownto be effective in prevent-
`ing NSAID-related ulcers (13, 18, 71). In a recent study
`from two sites in Europe, 107 arthritic patients were ran-
`domly allocated to receive diclofenac 200 mg/day or
`naproxen | g/day, plus sucralfate gel | g b.i.d. (mn = 53) or
`identical placebo (n = 54) for 14 days in an RCT. Although
`there was an unexplamed difference in the mcidence ofulcer
`betweenthe two centers, there was an overall decrease in the
`occurrence of ulcer between patients receiving sucralfate
`
`(8%) compared with placebo (28%) (p < 0.05). Both GU
`and DU were analyzed togetherin this study, and the proven
`ctficacy of sucralfate for DU mayaccountfor these results
`(72). Studies have shown that antacids and buffered tablets
`do not protect against NSAID injury (73-75). Enteric coat-
`ing maybe helpful in reducing aspirin-related gastric and
`duodenal ulcer (74, 75), but does not reduce the risk of
`NSAID-related ulcer complications (76).
`Several new compounds have shown promise in both
`animal studies and patients. In a double-blind RCT, sulgly-
`cotide 200 mg ¢.i.¢. or placebo was coadministered with an
`NSAID to paticnts with rhcumatoid arthritis. Gastric or
`duodenal ulcer was seen in six of 42 (18%) in the sulgly-
`cotide group and in 15 of 44 (34%) in the placebo group
`(p = 0.02) (77). Another newagent understudy is zinc
`acexonite (ZAC). Either ZAC 300 mg/day or placebo was
`randomly assigned in a double blind mannerto 276 arthritic
`patients receiving NSAIDs. Of 141 patients receiving ZAC,
`no GU and only one DU (0.9%) was found, whereas 12 of
`135 subjects (6.0%) on placebo developed an ulcer. Unfor-
`tunatcly, this study suffers from a high withdrawal and loss
`to followup rate (67/276) (78). Because of lack of confir-
`matory studies, no recommendations can be made at this
`time concerning either of these agents.
`More promising is the development of new, purportedly
`nontoxic, NSAIDs. These fall
`imto two groups: COX-2-
`selective
`inhibitors,
`and nitric
`oxide
`(NO)-releasing
`NSAIDs. Studies of these agents are very limited. Meloxi-
`cam (15 mg/day), a weak COX-2 inhibitor (COX 1: COX-2
`ratio 0.33), has been studied in an uncontrolled manner in
`357 patients with rheumatoid arthritis. Severe side effects
`(bleeding, perforation, and ulcer) were seen in only three
`patients (0.8%) (79). The GI effects of more potent COX-2
`inhibitors have thus far only been studied in normal volun-
`teers. These studics have shown a lesser degree of overall
`erosive injury and ulcer when compared with other NSAIDs
`(80, 81). At this time,
`there are no published RCTs of
`NO-NSAIDsin patients or human volunteers.
`
`TREATMENT OF NSAID-INDUCED ULCERS
`
`Recommendation
`
`NSAID-induced ulcer disease may be treated with any
`approved therapyfor ulcer disease. It is preferable to stop
`NSAID therapy when ulcer disease occurs. A proton pump
`inhibitor is the agent of choice when NSAIDs must be
`continued in the presence of ulcer disease and for large
`ulcers. Treatmentfor H. pylori is recommendedfor patients
`taking NSAIDs who have ulcers and are infected with this
`organism.
`NSAID-related ulcers may be treated effectively with any
`approved therapyfor peptic ulcer disease. Healing generally
`is more rapid when the NSAID is discontinued and com-
`pares favorably with healing rates for peptic ulcer disease
`not associated with NSAID intake. Several recent RCTs
`
`provide what are probablythe best data on the healmg rates
`
`Page 5 of 10
`
`MYL-EN000041041
`
`Patent Owner Ex. 2064
`Mylan v. Pozen
`IPR2017-01995
`
`Page 5 of 10
`
`Patent Owner Ex. 2064
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`2042
`
`LANZA ef al.
`
`AJG — Vol. 93, No. 11, 1998
`
`TABLE 5
`Treatment afNSAID-Related Ulcers
`% Healed
`Regimen TT Pp
`Ref.
`Based on Treated.
`4 wk
`8 wk
`
`
`25
`
`26
`
`84
`
`$2
`
`All ulcers
`NSAIDs con’t
`
`All Ulcers
`NSAIDs con’t
`
`GU only
`NSAIDs con’t
`
`O 20 g.d.
`O 40 gd.
`R 150 bid.
`O 20 g.d.
`O 40 gd.
`M 200 bid.
`020 g.d.
`O 40 gd.
`R 150 bid.
`
`GU
`
`DU
`
`83
`
`R 150 bid.
`NSAIDs DC’d
`NSAIDs Con’t
`NSAIDs DC’d
`NSAIDs con’t
`
`
`8 wk
`95
`63
`100
`84
`
`R 150 bid.
`
`65
`63
`52
`56
`60
`56
`61
`81
`332
`
`
`12 wk
`100
`79
`100
`92
`
`80
`79
`63
`75
`75
`71
`82
`75
`63
`
`<0.001 vs R 150 bid.
`<0.001 vy R 150 bid.
`
`NS vs O 40 and M 200
`NS vs O 20 and M 200
`NS vs O 20 and O 40
`<0.001 vs R150 bid.
`<0.03 vs R 150 bid.
`
`
`12 wk
`—0.004
`
`8 wk
`
`0.001
`DC’d vs con’t NSAIDs
`0.006
`DC’d vs con’t NSAIDs
`
`4 wk
`
`NSAIDs Con’t
`NSAIDs DC’d
`Placebo
`NSAIDs DC’d
`R 150 bid.
`NSAIDs Con’t
`NSAIDs DC’d
`Placebo
`
`NSAIDs DC’d 42
`
`NS vs placebo
`NS vs placebo
`
`NS vs placebo
`0.02 vs Placebo
`
`GU
`
`DU
`
`67
`68
`
`47
`
`61
`81
`
`O = omeprazole (in mg}; M = misoprostol {in hg); R = raniudime Gin mg).
`
`of NSAID-related ulcers treated with H2 RAs or PPIs
`
`(Table 5). These studies showed that good healing rates
`could be obtained at both 4 and 8 wk withall of the agents
`employed. Generally, healing rates were betterfor all treat-
`ments when NSAIDs were discontinucd, but satisfactory
`healing still occurred with prolonged therapy when it was
`necessary to continue NSAID treatment. Healing rates in
`patients taking omeprazole were significantly better than
`those in patients taking ranitidine when NSAIDs were con-
`tinued (25, 26, 82-84). In an uncontrolled study, famotidine
`40 mg/day was given to 71 patients with endoscopically-
`proven GUfor 8 wk. The healing rate for patients with
`NSAID-related ulcers was compared with that for idiopathic
`ulcer. It was found that healing occurred in 46 of 48 (96%)
`of patients with NSAID-related ulcer, which was signifi-
`cantly greater than the 17 of 23 (74%) seen for idiopathic
`ulcer (p = 0.01) (85). In a nonplacebo-controlled study,
`three doses ofnizatidine (150 mg HS,150 mg 6.i.d.,, 300 mg
`b.i.d.) were given to patients with active GU or DU while
`they continued their original NSAID. After 8 wk of therapy.
`>90%of the ulcers ofall three groups were healed. There
`was a tendency to higher healing rates for GU after four
`wecksin the high dose (300 mg 4.i.d.) nizatidine group (67).
`In another RCT, substitution of enteric-coated aspirm in
`patients with ASA-related ulcers did not increase healing
`rates in patients treated with cimetidme (400 mg/day) and
`
`in fact, seven patients given enteric-coated
`antacids and,
`aspirin failed to heal their ulcers, whereas 15 of 16 healed
`after aspirin was discontinued in all forms (86).
`Helicobacter pylori infection has been strongly associ-
`ated with peptic ulccr discase, especially duodenal ulccr.
`However,
`this association has not been demonstrated in
`patients with NSAID-related ulcer (87-90). In one study,
`however, gastropathy was more severe in //. pylori-positive
`patients (88). Some recent studies have shown that the
`incidence of DU is increased in 7. pylori-positive patients
`taking NSAIDs (91, 92). However,
`in another study 7.
`pylori eradication did not increase the healing of GU and
`DUassociated with long-term NSAID use (93). It has been
`recently reportedthat cradication ofH. pylori before NSAID
`therapy strikingly reduces the incidence of ulcer disease m
`patients being treated with naproxen (750 mg/day). Standard
`triple therapy (bismuth subcitrate 120 mg q.i.d., tetracycline
`500 mg g.i.d., and metronidazole 400 mg q.i.d) was given to
`45 H. pylori-positive patients before 8 wk of naproxen
`therapy. Naproxen alone was given to 47 patients over the
`sameperiod. In the triple therapy group, only three patients
`(7%) developed ulcer, two of whom were found to have
`failed H. pylori cradication.
`In the naproxen group, 12
`patients (26%) had ulcers, one of which bled (94). Another
`recent case-control study of ulcer bleeding in 487 elderly
`patients revealed that NSAID usage (odds ratio 4.93) and H.
`
`Page 6 of 10
`
`MYL-EN000041042
`Patent Owner Ex. 2064
`Mylan v. Pozen
`IPR2017-01995
`
`Page 6 of 10
`
`Patent Owner Ex. 2064
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`AJG — November 1998
`
`NSAID-INDUCED ULCER TREATMENT AND PREVENTION
`
`2043
`
`pylori status (odds ratio 2.80) increased risk substantially,
`but there was no evidence ofinteraction (95). In view ofthe
`conflicting data, serecning for H. pylori infection cannot be
`recommended for all patients receiving NSAIDs. There is
`some recent evidence that H. pylori-negative patients heal
`more slowly than those who are positive for this organism
`(26). However, experts still agree that H. pylori-positive
`patients with a past or current history of ulcer requiring
`NSAID therapy should be treated for the infection, as it
`cannot be determined whether the prior ulcer was due to
`NSAID therapy or to //. pylori infection.
`
`NSAID INJURY TO THE SMALL BOWLL
`AND COLON
`
`Recommendations
`
`NSAID-related injury to both the small and large howel
`has includes occult and frank bleeding, perforation, ob-
`struction, an acute colitis, and exacerbation of existing
`colon disease. Physicians prescribing NSAIDs should be
`aware of these potential complications.
`NSAID-related injury to the small bowel and colon has
`only recently been described and has been documented
`primarily by anecdotal case reports. However,
`these are
`numerous and constitute asignificant body ofliterature. The
`overall incidence of NSAID-related injury to the gastroin-
`testinal tract distal to the duodenum is much less than that
`
`seen in the stomach and proximal duodenum. Nevertheless,
`it occurs frequently enough to warrant serious consideration
`byphysicians treating patients with NSAIDs.
`
`Small bowel injury
`Tnjury to the small intestine can be manifested by perfo-
`ration, ulceration and stricture, or by an NSAID-induced
`enteropathy (31, 46, 96-98). Perforation of the small bowel,
`although a rare complication of NSAID use, was found in
`one retrospective review to be significantly more common
`among patients taking NSAIDs when compared with con-
`trols (31). More commonly seen are small intestinal ulcers
`and strictures (95). Diaphragm-like strictures, often with a
`very small lumen