THE STOMACH
`
`1052-5157/96 $0.00 + .20
`
`NONSTEROIDAL
`ANTI-INFLAMMATORY
`DRUG GASTROPATHY
`
`Loren Laine, MD
`
`Billions of doses of aspirin5 and other nonsteroidal anti-inflammatory
`drugs (NSAIDs) are consumed each year in the United States; even 10
`years ago, 100 million prescriptions for nonaspirin NSAIDs were written
`in the United States. 35 Because 40% to 60% of patients regularly taking
`NSAIDs have gastric erosions and approximately 10% to 30% have gastric
`ulcers/0
`63 NSAIDs are probably the most common cause of gross
`43
`23
`•
`•
`•
`gastric injury in the United States today.
`
`ENDOSCOPIC FINDINGS OF NSAID GASTROPATHY
`
`The endoscopically visible lesions induced by NSAID use may be
`labeled NSAID gastropathy and include subepithelial hemorrhages, ero(cid:173)
`sions, and ulcers. Subepithelial hemorrhages have the appearance of pete(cid:173)
`chiae or bright red areas of mucosa without any visible break ("blood
`under clear plastic wrap"). Although many have used the term submucosal
`hemorrhage to describe this finding, subepithelial is a much better term
`because it leaves open the possibility that the hemorrhage is located
`either in the mucosa or the submucosa. Furthermore, as discussed later,
`histologic examination of these lesions reveals blood in the mucosa, just
`beneath the epithelium. 32
`Erosions are visible breaks in the mucosa that are flat or minimally
`depressed, often surrounded by a halo of erythema. The base is generally
`
`From the Department of Medicine, University of Southern California School of Medicine,
`Los Angeles, California
`
`GASTROINTESTINAL ENDOSCOPY CLINICS OF NORTH AMERICA
`
`VOLUME 6 • NUMBER 3 • JULY 1996
`
`489
`
`Page 1 of 16
`
`Patent Owner Ex. 2043
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`490
`
`LAINE
`
`white with a fibrinous exudative appearance. Differentiation between an
`ulcer and an erosion is a histologic one. An ulcer is a defect that penetrates
`the muscularis mucosae into the submucosa or deeper, whereas an erosion
`is a break that remains confined to the mucosa.
`A singl dose of aspirin (650 mg) leads to evidence of gastric damage
`within 10 minutes. Baskin et aP found· that the gastric potential difference
`was decreased and approximately a quarter of surface epithelial cells
`were damaged on light and electron microscopic examination of blind
`gastric biopsies.
`On endoscopic examination, multiple subepithelial hemorrhages de(cid:173)
`velop within 15 to 30 minutes after a one-time ingestion of aspirin. 15
`50
`19
`•
`•
`These lesions appear to progress to their maximal extent by 1 to 2 hours
`after ingestion, and they may begin to decrease at 24 hours. Erosions, on
`the other hand, seem to be unusual after a single dose of an NSAID.
`Hemorrhagic lesions may occur in the body or the antrum of the stomach.
`Continued aspirin ingestion for 1 day (650 mg four times a day)
`leads to development of gastric erosions at 24 hours. Unlike subepithelial
`50 and seem to
`hemorrhages, erosions occur primarily in the antrum19
`43
`•
`•
`occur in the areas that show the most severe mucosal hemorrhage after
`one dose of aspirin. 19
`After 1 week of regular NSAID ingestion, ulcers may be identified
`in a significant minority of subjects. Lanza37 examined his data from a
`variety of studies of over 900 volunteers and reported an 8% incidence
`of ulcers developing after 7 days of NSAID use (aspirin 9%, nonaspirin
`NSAIDs 7%)? Evaluation of placebo groups in large trials of medical
`prophylaxis for NSAID-induced ulcers provides the best information for
`estimating the proportion of patients who develop new ulcers during 1
`to 3 months of NSAID ingestion. The incidence of new ulcer forma(cid:173)
`tion is quite variable, however, from study to study. For example, two
`similarly designed studies evaluating the efficacy of misoprostol in arthri(cid:173)
`tis patients taking NSAIDs reported an intent-to-treat 3-month incidence
`of gastric ulcers in 30 of 138 (22%) patients assigned to placebo in the
`first study18 and in 25 of 323 (8%) patients in the placebo group in the
`second study. 20
`Graham et aP9 have suggested that gastric adaptation occurs with
`prolonged use of NSAIDs. Thus, ulcers and erosions might be seen more
`commonly at endoscopy soon after initiation of NSAID use and might
`decrease in frequency with prolonged use. As mentioned, virtually all
`subjects given aspirin develop erosions within 1 to 2 days and yet only
`half of patients regularly using aspirin or other NSAIDs have erosions
`fmmd on endoscopy.
`
`CLINICAL SIGNIFICANCE OF NSAID GASTROPATHY
`
`Hemorrhagic and erosive gastropathy is of minor clinical significance.
`Gastritis is frequently implicated as a cause of upper gastrointestinal
`
`Page 2 of 16
`
`Patent Owner Ex. 2043
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`NSATD GASTROPATHY
`
`491
`
`bleeding. For example, the large 1981 American Society for Gastrointesti(cid:173)
`nal Endoscopy (ASGE) survey of 2225 patients reported that gastric ero(cid:173)
`sions were the cause of bleeding in 23% of patients.61 TI1e more recent
`experience of myself and others, however, is that hemorrhagic and erosive
`g1;1stropathy is rarely a cause of serious upper gastrointestinal bleeding.
`For example, a prospective evaluation of 445 patients with major upper
`gastrointestinal hemorrhage at Los Angeles County + University of
`Southern California Medical Center revealed that hemorrhagic or erosive
`gastropathy was considered the cause of bleeding in 3% of cases, and
`virtually never led to life-threatening hemorrhage or large transfusion
`requirements.27
`Subepithelial hemorrhages and erosions are strictly mucosal lesions,
`whereas all blood vessels of significant size are in the submucosa or
`deeper. Thus, unlike ulcers, which can induce severe bleeding when they
`erode into arteries below the mucosa, subepithelial hemorrhages and
`erosions generally do not cause major bleeding. When severe bleeding
`occurs in association with NSAID use, it indicates that an ulcer, rather
`than an erosion, is present as the source of the hemorrhage.
`Dyspepsia is common in patients taking NSAIDs. Larkai et al42
`studied 245 rheumatic patients taking NSAIDs and found that 16% had
`daily dyspepsia, 29% had symptoms in the preceding week, and 37% had
`dyspeptic symptoms in the preceding 2 months. No studies, however,
`have documented that NSAID gastropathy correlates with abdominal
`symptoms. The lack of association between gross NSAID-induced lesions
`identified at endoscopy and NSAID-associated symptoms is well demon(cid:173)
`strated in studies of the prevention of NSAID-induced gastric ulcers.
`Graham et aP8 showed· that although misoprostol was significantly better
`than placebo in preventing gastric ulcers (discussed later), the frequency
`of abdominal symptoms, such as dyspepsia, abdominal pain, and nausea
`was similar in the misoprostol and placebo groups.
`A number of epidemiologic studies (case control and cohort) have
`demonstrated an increased risk of complicated ulcers (e.g., bleeding, per(cid:173)
`foration, hospitalization, death) and overall gastrointestinal complications
`in patients taking NSAIDs. 64 A 1991 meta-analysis13 reported that the odds
`ratio of risk for adverse gastrointestinal events related to NSAID use was
`2.7 (95% CI, 2.5 to 3). Age over 60 years, prior gastrointestinal event, use
`of steroids, and less than 1 month of NSAID use all significantly increased
`the risk of adverse events. Griffin et aP found the relative risk for develop(cid:173)
`ment of a gastric ulcer leading to hospitalization with NSAID use in
`Tennessee was 5.5 (95% CI, 4.4 to 6.9) (duodenal ulcer relative risk was
`4.3 [95% CI, 3.5 to 5.2]). The risk of hospitalization for peptic ulcer rose
`with increasing dose and with shorter (:530 days) duration. In the United
`Kingdom, Langman et aP6 reported a relative risk of 4.5 (95% CI, 3.6 to
`5.6) for peptic ulcer bleeding with NSAID use; the risks of bleeding from
`gastric ulcers and duodenal ulcers were similar.
`Experimental studies provide a much more reliable estimate of the
`risk of NSAID-associated gastrointestinal complications. Very few large-
`
`Page 3 of 16
`
`Patent Owner Ex. 2043
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`492
`
`LAINE
`
`scale prospective, placebo-cor;trolled trial~, ho"':ever, have h:cluded care(cid:173)
`ful evaluation of gastrointestmal events m their study des1gn. The best
`estimate of the risk of regular NSAID use probably comes from a recently
`published prospective study in patients with rheumatoid arthritis. Silver(cid:173)
`stein et al62 followed 4439 patients taking NSAIDs for rheumatoid arthritis
`and a placebo for up to 6 months. The incidence of gastrointestinal compli(cid:173)
`cations is shown in Table 1. Importantly, 42% were taking steroids, poten(cid:173)
`tially increasing· the risk of complications. Four clinical characteristics were
`identified as independent predictors of NSAID-associated gastrointestinal
`complications: (1) age greater than or equal to 75 years (odds ratio: 2.5
`[95% CI, 1.5 to 4.1]); (2) history of peptic ulcer (2.3 [95% CI, 1.3 to 4.1]);
`(3) history of gastrointestinal bleeding (2.6 [95°/o CI, 1.3 to 5]); and (4)
`history of heart disease (1.8 [95% CI, 1.1 to 3.2]).
`With the increasing use of aspirin for vascular prophylaxis, the risk
`of complications in this group of patients also must be assessed. Although
`many large-scale studies of aspirin prophylaxis are available, careful atten(cid:173)
`tion to gastrointesth1al events is not usually a primary aim of the study.
`Kurata and Abbey,Z6 in a study of 4524 subjects receiving 0.5 g of aspirin
`twice a day or placebo for at least 3 years, reported a relative risk for
`gastric ulcer hospitalization of 9.1 (95% CI, 1.2 to 71) higher for the aspirin
`group than the placebo group (risk for duodenal ulcer hospitalization
`was 10.7 [95% CI, 2.5 to 45.5]). Sharrock et al,60 in a study of 2435 patients
`for transient ischemic attack prophylaxis, found that 300 mg of aspirin
`every day was associated with a 7.7-fold increased risk of upper gastroin(cid:173)
`testinal bleeding (from all causes) as compared with a placebo (95% CI,
`1.7 to 33.8) and 600 mg of aspirin twice a day was associated with an
`odds ratio of 14.4 (95% CI, 3.4 to 60).
`
`Table 1. RISK OF SERIOUS NSAID-ASSOCIATED GASTROINTESTINAL
`COMPLICATIONS IN PATIENTS WITH RHEUMATOID ARTHRITIS: A 6-MONTH
`DOUBLE-BLIND, PLACEBO-CONTROLLED TRIAL OF MISOPROSTOL
`
`Gastrointestinal Event
`
`Perforated ulcer
`Bleeding ulcer or erosion
`Perforated ulcer, gastric outlet
`obstruction, or bleeding
`ulcer or erosion
`Clinical evidence of upper
`gastrointestinal bleeding
`Perforation, gastric outlet
`obstruction, or clinical
`evidence of upper
`gastrointestinal bleeding
`
`Misoprostol
`(N = 4404)
`1 (0.02%)
`15 (0.34%)
`16 (0.36%)
`
`Placebo
`(N = 4439)
`7 (0.16%)
`23 (0.52%)
`33 (0.74%)
`
`Odds Ratio
`(95% Cl)
`
`0.14 (0.02-1.15)
`0.66 (0.33-1.31)
`0.49 (0.27-0.89)
`
`32 (0.73%)
`
`42 (0.95%)
`
`0.77 (0.47-1.24)
`
`33 (0.75%)
`
`52 (1.17%)
`
`0.64 (0.40-1.01)
`
`From Silverstein FE, Graham DY, Senior JR, et al: Misoprostol reduces serious gastrointestinal
`complications in patients with rheumatoid arthritis receiving nonsteroidal anti-inflammatory drugs. A ran(cid:173)
`domized, double-blind, placebo-controlled trial. Ann Intern Med 123:241, 1995; with permission. The
`American College of Physicians is not responsible for the accuracy of the translation.
`
`Page 4 of 16
`
`Patent Owner Ex. 2043
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`NSAID GASTROPATHY
`
`493
`
`As the daily dose of aspirin decreases, the risk of gastrointestinal
`complications also should decrease. The double-blind, placebo-controlled
`Physicians' Health Study evaluating 325 mg of aspirin every other day
`fotmd that 38 (0.34%) of the 11,037 participants taking aspirin developed
`a bleeding ulcer as compared with 22 (0.20%) of the 11,034 receiving
`placebo (relative risk, 1.8; 95% CI, 1.1 to 2.9) during a mean follow-up
`period of 60.2 months. 65 Daily use of 75 mg of aspirin was associated with
`a nearly significant relative risk of gastrointestinal bleeding of 2.8 (95%
`CI, 0.9 to 8.7) in a Swedish trial of prophylaxis for cerebrovascular events. 58
`Doses as low as 30 mg of aspirin have been shown to be effective
`for vascular prophylaxis. 10 A Dutch study comparing 30 mg with 283 mg
`of aspirin in 3131 patients followed for a mean of 2.6 years revealed a
`nonsignificant relative risk of major gastrointestinal bleeding for 30 mg
`compared with 283 mg of 0.7 (95% CI, 0.4 to 1.3).10 Although any risk of
`gastrointestinal complications should be lower at these doses, Cryer et
`al8 have recently shown that doses as low as 10 mg of aspirin a day still
`significantly decrease gastric prostaglandin production to levels that are
`similar to the inhibition seen with 81 mg and 325 mg of aspirin. These
`data suggest that any dose of aspirin has the potential to induce gastric
`lesions and gastrointestinal complications.
`
`PATHOGENESIS OF NSAID GASTROPATHY
`
`NSAIDs cause gastric damage by both topical and systemic effects,
`although inhibition of prostaglandin synthesis (a systemic effect) is
`thought to be the major mechanism of action.4•64 Most NSAIDs are weakly
`acidic (pka ranging froin 3 to 5), and in gastric juice they are relatively
`nonionized and lipophilic. The NSAIDs move rapidly across gastric muco(cid:173)
`sal cell membranes into the neutral environment within the celt where
`the neutral pH leads to conversion of the nonionized NSAID to the ionized
`form of the NSAID and a hydrogen ion. The NSAID thus accumulates
`within the cell at a higher concentration than outside the cell leading to
`direct local damage.
`Although this topical damage may explain the endoscopically visual(cid:173)
`ized lesions seen soon after the initiation of NSAIDs, this topical effect
`does not appear to be of primary importance in the development of gastric
`injury and complications. Parenteral NSAIDs, given via intravenous, in(cid:173)
`tramuscular, or rectal routes, are well documented to cause ulcers. 6• 39• 64
`Furthermore, prodrugs, such as sulindac, which require absorption and
`metabolism before they exert their pharmacologic effect, induce relatively
`little initial endoscopic damage37 but are still associated with a significant
`risk of bleeding ulcers. 21
`Several lines of reasoning indicate that inhibition of prostaglandin
`synthase H (cyclooxygenase) and thus prostaglandin synthesis is a major
`mechanism of NSAID-induced gastric damage. NSAIDs do inhibit prosta(cid:173)
`glandin synthesis and induce gastric injury, although correlation of gastric
`mucosal prostaglandin production and gastric injury is relatively weak
`
`Page 5 of 16
`
`Patent Owner Ex. 2043
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`494
`
`LAINE
`
`at best.31 More dramatically, antiprostaglandin antibodies produce severe
`gastrointestinal ulceration in rabbits. 5
`3 Finally, prostaglat;din-replac~ment
`therapy significantly reduces the development of gastnc damage m pa(cid:173)
`tients taking NSAIDs. 18•19 Nevertheless, a variety of other mechanisms
`have been proposed to explain NSAID-induced gastrointestinal damage. 4•6
`Strong evidence exists il1. the animal model that leukocyte (primarily
`neutrophil) adherence to the endothelium of postcapillary venules is im(cid:173)
`portant in the development of NSAID-induced gastrointestinal injury.
`NSAID-induced damage can be prevented by neutropenia or prevention
`72
`74
`75 Neutrophil adherence, however, appears
`of neutrophil adherence.71
`•
`•
`•
`to be a direct consequence of the inhibition of prostaglandin production
`by NSAIDs. 73
`75
`•
`
`HISTOLOGIC FINDINGS OF NSAID GASTROPATHY
`
`Based on the grossly visible gastric injury induced by NSAID inges(cid:173)
`tion, NSAIDs are commonly said to cause gastritis. The subepithelial
`hemorrhages, erosions, and ulcers associated with NSAID use, however,
`have not been shown to be characterized by gastritis; that is, microscopic
`evidence of a diffuse inflammatory cell infiltration. In a 1973 retrospective
`study of patients with gastric ulcers, MacDonald44 noted that the only
`patients without histologic evidence of gastritis were those taking aspirin.
`More recently, we reported that patients with NSAID-associated gas(cid:173)
`tric ulcers have a significantly lower prevalence of Helicobacter pylori infec(cid:173)
`tion than patients with gastric ulcers who are not taking NSAIDs.3° Fur(cid:173)
`thermore, we found no evidence of worsening gastric mucosal histologic
`injury with NSAID il1.gestion when comparing pretreatment biopsies with
`biopsies taken after 1 and 4 weeks of NSAID treatment in a controlled,
`double-blind triaU8 The presence of histologic gastritis was related to
`the presence of H. pylori. Mean histology scores (e.g., inflammatory cell
`infiltrate, epithelial abnormality) were zero or near zero in H. pylori(cid:173)
`negative subjects and did not increase with 4 weeks of NSAID ingestion.28
`Interestingly, Metzger et al49 studied the effects of aspirin ingestion at 3
`and 14 days in the pre-H. pylori era and also found no significant changes
`in histologic gastric injury in endoscopically uninvolved areas; the degree
`of histologic gastritis seen at baseline was unchanged by aspirin adminis(cid:173)
`tration. Thus, it appears that NSAIDs cause gastric injury via a mechanism
`distinct from H. pylori-induced gastritis and that any underlying diffuse
`histologic gastritis in patients with NSAID-associated ulcers is due to
`coincident H. pylori infection.
`The focal macroscopic lesions seen with NSAID use, such as subepi(cid:173)
`thelial hemorrhage and erosions, do have associated microscopic features.
`We have shown that subepithelial hemorrhages are characterized by hem(cid:173)
`orrhage (red blood cells) in the superficial portion of the gastric mucosa. 32
`Erosions are characterized histologically by marked epithelial changes at
`the site of the erosion. 29 Our studies have not identified similar histologic
`changes in uninvolved mucosa adjacent to the subepithelial hemorrhages
`
`Page 6 of 16
`
`Patent Owner Ex. 2043
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`NSAID GASTROPATHY
`
`495
`
`or erosions, and, as mentioned previously, diffuse inflammatory cell infil(cid:173)
`tration (gastritis) is generally present only in patients with H. pylori infec(cid:173)
`tion. A number of authors have suggested that NSAID use causes a diffuse
`reactive .or chemical gastritis distinct from the typical gastritis associated
`with H. pylori infection. The term chemical gastritis was initially used to
`·describe changes in the postgastrectomy stomach, which included foveo-
`lar hyperplasia, edema, and vasodilatation.9 This term has subsequently
`been expanded by some to include similar changes in patients on
`67
`NSAIDs.52
`76 Although these changes have been reported in biopsy
`•
`•
`studies from patients taking NSAIDs, they have not been documented to
`develop diffusely in most studies of NSAID ingestion with pre- and
`28
`47
`49
`66
`posttreatment biopsies.17
`•
`•
`-
`•
`In our double-blind, placebo-controlled trial we found no increase
`in epithelial abnormalities (including foveolar hyperplasia); edema; vaso(cid:173)
`dilatation; or congestion after 4 weeks of NSAID ingestion. McCarthy et
`al/6 however, recently suggested that 1 week of naproxen ingestion did
`cause chemical gastritis in some patients when comparing pre- and post(cid:173)
`treatment biopsies. Nineteen subjects were assessed using a 0 to 15 grading
`scale, with a score of 10 being considered chemical gastritis. One patient
`had chemical gastritis at baseline. Six of the 19 subjects had an increase
`in chemical gastritis score of three to five points; all other subjects had
`posttreatment scores within two points of baseline.
`
`INTERACTION OF NSAIDs AND H. PYLORI
`
`Of greater clinical importance is the question of whether underlying
`H. pylori infection increases the risk of gross gastric injury and ulcer
`formation in patients taking NSAIDs. At least six prospective studies of
`1-week,38• 46 1-month,17· 28• 69 and 3-months24 duration have indicated that
`the presence of H. pylori does not alter the gross gastrointestinal injury
`that develops with NSAID ingestion. In the largest study, Kim et aF4
`found that H. pylori infection was present in half of patients who developed
`a gastric or duodenal ulcer over 3 months of NSAID use and half of those
`who did not develop an ulcer. When just gastric ulcers were analyzed,
`however, there was a trend toward an increased prevalence of H. pylori
`infection in those who developed an ulcer (P = 0.06). More gastric ulcers
`also developed in patients who had erosions present at study entry.
`More recently, Taha et al68 reported that patients with H. pylori infec(cid:173)
`tion were more likely to develop gastroduodenal ulcers or erosions than
`patients without H. pylori infection (separate information was not pro(cid:173)
`vided for gastric lesions). Ulcers were more likely to develop in patients
`who had erosion at baseline (patients were on NSAIDs at enrollment and
`were included if erosions were present), however, and erosions were
`more common in those with H. pylori infection. Finally, a preliminary
`report from Kim et aFS indicated that 2 weeks of NSAID ingestion induced
`greater mucosal injury (on a 0 to 4 grading scale) in patients with H. pylori
`infection than in those without H. pylori.
`
`Page 7 of 16
`
`Patent Owner Ex. 2043
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`496
`
`LAINE
`
`Thus, at the present time, most evidence does not indicate that pa(cid:173)
`tients with H. pylori infection are at increased risk of developing ulcer
`disease when started on NSAID therapy, and we do not suggest that
`physicians check a patient's H. pylori ~tatus prior to beghming NSAID
`therapy. Larger studies of longer dt~rahon, how.ever, ~re neede~ to settle
`this question. Patients with tmderlymg H. pylon-associated eroswns may
`68
`be at greater risk of developing ulcers.24
`•
`Furthermore, those with underlying H. pylori-associated ulcers do
`appear to be at greater risk of developing clinically manifest or compli(cid:173)
`cated ulcers when beginning NSAIDs. A prior history of ulcer disease is
`documented to be a risk for development of complications in patients
`taking NSAIDs. 14•
`62 Complications appear to be most frequent immediately
`after starting NSAIDs (during the first 30 days),13
`21 suggesting the possibil(cid:173)
`•
`ity that NSAIDs cause complications in ulcers that were already present
`but clinically silent. In addition, although NSAIDs appear to cause endo(cid:173)
`scopically visualized ulcers in the stomach more commonly than in the
`duodenum/0•23•43•64 NSAID-associated complicated ulcers occur with simi(cid:173)
`36
`lar frequenf=y in the stomach and duodenum/6
`64 suggesting that NSAIDs
`•
`•
`may induce complications in duodenal ulcers that were present but clini(cid:173)
`cally silent at the time NSAIDs were started.
`
`TREATMENT OF NSAID-ASSOCIATED ULCERS
`
`NSAID-associated ulcers heal with standard antiulcer therapy in most
`cases. Healing rates are improved, however, by stopping NSAID therapy,
`especially with gastric ulcers.33
`45 Lancaster-Smith et aP3 performed a pro(cid:173)
`•
`spective trial in which 190 patients were randomly assigned to continue
`or stop their NSAID therapy and all received ranitidine, 150 mg twice a
`day (Table 2). The rate of gastric ulcer healing was significantly better at
`
`Table 2. RANDOMIZED COMPARISON OF ULCER HEALING RATES WITH
`H2-RECEPTOR ANTAGONIST THERAPY WHEN STOPPING VERSUS
`CONTINUING NSAIDs
`
`Gastric Ulcer
`
`4-week healing
`8-week healing
`12-week healing
`
`NSAID Stopped (N = 47)
`54%
`63%*
`79%*
`
`NSAID Continued (N - 37)
`71%
`95%
`100%
`
`Duodenal Ulcer
`
`4-week healing
`8-week healing
`12-week healing
`
`NSAID Stopped (N = 43)
`57%
`84%*
`92%
`
`NSAID Continued (N - 55)
`74%
`100%
`100%
`
`* P < 0.01, NSAID slopped versus NSAID continued
`Data from Lancaster-Smith MJ, Jaderberg ME, Jackson DA: Ranitidine in the treatment of non(cid:173)
`steroidal anti-inflammatory drug-associated gastric and duodenal ulcers. Gut 32:252, 1991; with per(cid:173)
`mission.
`
`Page 8 of 16
`
`Patent Owner Ex. 2043
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`NSAID GASTROPATHY
`
`tl97
`
`8 weeks (95% versus 63%) and 12 weeks (100% versus 79%) in those
`stopping the NSAIDs. The differences in duodenal ulcer healing were
`less marked and were significant at 8 weeks (100% versus 84%) but not
`at 12 weeks (100% versus 92%).
`Patients with gastric ulcers who must remain on NSAID therapy
` appear to have improved healing rates with the use of the proton-pump
`inhibitor omeprazole. Walan et aF0 reported 8-week healing rates of 53%
`with ranitidine, 150 mg twice a day, 82% with omeprazole, 10 mg once
`a day, and 95% with omeprazole, 40 mg once a day (P = 0.02 for ranitidine
`versus 40 mg omeprazole).
`Thus, NSAIDs should be discontinued in patients who develop ulcers,
`and any form of standard antiulcer therapy can be used in this group. In
`patients who must continue their NSAIDs, use of a proton-pump inhibitor,
`such as omeprazole, should be strongly considered, especially in patients
`with gastric ulcers.
`
`PREVENTION OF NSAID-ASSOCIATED
`GASTROINTESTINAL DAMAGE
`
`Strategies to decrease the risk of NSAID-associated gastrointestinal
`complications include the following:
`Use analgesics, such as acetaminophen
`Use lowest possible dose of NSAID
`Avoid concomitant corticosteroids
`Give cotherapy with misoprostol (high-risk patients)
`Consider potentially less injurious NSAIDs
`Nonacetylated salicylates
`Etodolac
`Nabumetone
`
`The first step in preventing NSAID-associated gastrointestinal injury
`is to avoid the use of NSAIDs whenever possible. Many patients have
`adequate symptomatic relief with non-NSAID analgesics, such as acet(cid:173)
`aminophen. Bradley et aV in a 4-week, double-blind, randomized trial
`of 184 patients with chronic knee pain owing to osteoarthritis, found that
`the efficacy of acetaminophen was similar to low-dose (1200 mg per day)
`and high-dose (2400 mg per day) ibuprofen.
`If NSAIDs must be used, the lowest dose possible should be tried
`initially. Physicians should also avoid the use of combined NSAIDs and
`steroids if possible, because concomitant use of corticosteroids increases
`the risk of gastrointestinal complications.13
`51 Physicians also need to assess
`'
`for high-risk clinical characteristics before initiating NSAID therapy.
`Treatment options must be considered carefully before beginning NSAIDs
`in patients with a history of peptic ulcer disease or a history of gastrointes(cid:173)
`tinal bleeding, or in elderly patients.
`Cotherapy to prevent ulcer complications must be considered in a
`small group of patients taking NSAIDs. I generally recommend cotherapy
`
`Page 9 of 16
`
`Patent Owner Ex. 2043
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`498
`
`LAINE
`
`in patients who have had clinically significant ulcers (e.g., bleeding) who
`require continued NSAID therapy. Histamine-2 receptor antagonists pre(cid:173)
`vent development of duodenal ulcers but not gastric ulcers with NSAID
`therapy. 11• 54 Therefore, these agents cam1ot generally be recommended
`for prophylaxis. Although more potent antisecretorytherapy with proton(cid:173)
`pump inhibitors, such as omeprazole, theoretically may prevent develop(cid:173)
`ment of both duodenal and gastric ulcers, studies clearly documenting
`this concept ar~ not available at present.
`The sole medication available that has been documented to prevent
`both gastric and duodenal ulcers is misoprostol. In a 12-week study,
`Graham et aF0 found gastric ulcers developing in 7.7% of patients receiving
`placebo as compared with 1.9% of those taking misoprostol, and duodenal
`ulcers in 4.6% of those in the placebo group versus 0.6% of those in the
`misoprostol group. Because of the high rate of gastrointestinal complica(cid:173)
`tions (primarily diarrhea) and the potential problem of compliance with
`full-dose misoprostol (200 /hg four times a day), I generally recommend
`beginning therapy at 200 /hg twice a day and increasing to three times a
`day if the drug is well tolerated. Clinical studies suggest that lower dose
`misoprostol should decrease gastrointestinal side effects while still main(cid:173)
`taining much of the gastrointestinal protective effect.18
`• 40
`The studies cited perviously all are merely endoscopic studies assess(cid:173)
`ing the presence of endoscopically visualized ulcers. We know, however,
`that only a small proportion of patients with ulcers seen at endoscopy
`ever develop clinical problems from their ulcers. Our aim is to prevent
`important clinical manifestations of ulcers, such as bleeding. Unfortu(cid:173)
`nately, all prospective experimental studies performed assessing the ef(cid:173)
`fects of NSAIDs on the gastrointestinal tract have assessed only endo(cid:173)
`scopic damage, not clinically significant ulcers, with the exception of
`one recent study. 62 This 6-month double-blind, placebo-controlled trial of
`misoprostol in 8843 patients with rheumatoid arthritis demonstrated a
`significant reduction in "serious ulcer complications" (perforation, gastric
`outlet obstruction, or bleeding from ulcers or erosions) (see Table 1). The
`risk reduction with misoprostol was approximately 50%, although the
`difference between the two groups was only 17 patients (0.38% of the
`placebo group). In addition, the decreased risk for ulcer bleeding or any
`gastrointestinal bleeding was not significant. Nevertheless, this is the first
`and only study to indicate that the results of endoscopic studies may be
`used to predict the risk of gastrointestinal complications, at least qualita(cid:173)
`tively.
`The most attractive way to prevent NSAID-associated gastrointestinal
`damage is to develop NSAIDs that are less injurious to the gut. Compari(cid:173)
`sons of the relative rates of complications associated with different
`NSAIDs in epidemiologic studies are fraught with hazard; however, ibu(cid:173)
`profen consistently shows the lowest risk and piroxicam is among the
`highest-risk NSAIDs.14
`36 Indomethacin and naproxen appear to have
`21
`22
`•
`•
`•
`risks in the middle ground of NSAIDs, although 95% confidence intervals
`of their risk usually overlap both the low- and high-risk NSAIDs.
`Three commercially available agents appear to induce less gastric
`injury than the standard NSAIDs. The nonacetylated salicylate, salsalate,
`
`Page 10 of 16
`
`Patent Owner Ex. 2043
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`NSAID GASTROPATHY
`
`499
`
`appears to cause no significant increase in gastric or duodenal damage;
`this appears to be due to a lack of inhibition of mucosal prostaglandin
`production?· 55
`59 The drug nabumetone also appears to cause less gastric
`•
`damage than standard NSAIDs. Roth et al56 found that 7 (13%) of 53
`patients taking ibuprofen, 600 mg four times a day, for up to 12 weeks
`developed a gastric ulcer (another patient developed a duodenal ulcer),
`whereas 0 of 58 receiving nabumetone, 1000 mg once a day, developed
`a gastric ulcer (one patient developed a duodenal ulcer).
`Finally, four separate endoscopic studies all confirm that etodolac,
`given at doses of 200, 300, or 500 mg twice a day for up to 4 weeks,
`produces levels of gastrointestinal damage not significantly greater than
`placebo and significantly less than standard NSAIDs, such as naproxen,
`ibuprofen, or indomethacin.2
`57 This appears to be due to the fact that
`31
`41
`•
`•
`•
`etodolac does not inhibit gastric or duodenal prostaglandin production. 31
`•
`66 Although we cannot be certain that these short-term endoscopic studies
`can definitely be used to predict a lower rate of gastrointestinal complica(cid:173)
`tions, it seems reasonable to consider the use of etodolac, nabumetone,
`or a nonacetylated salicylate in patients who have a history of NSAID(cid:173)
`induced gastrointestinal injury.
`New NSAIDs are now being developed that hold the promise of
`virtually no gastrointestinal injury. Recently, the enzyme cyclooxygenase
`(prostaglandin H synthase) has been found to have two isozymes16• 34 : (1)
`COX -1 and (2) COX-2. COX -1 is the constitutive form, which is responsible
`for prostaglandin production for maintenance of homeostatic or physio(cid:173)
`logic functions in tissues, such as the stomach and kidney. COX-2, on the
`other hand, is the inducible isozyme, induced by inflammatory stimuli
`and considered to be responsible for proinflammatory prostaglandin pro(cid:173)
`duction in cells, such as macrophages and synoviocytes. It now appears
`that COX-2 is the target in attempting to decrease inflammation with
`NSAID therapy. Theoretically, NSAIDs that are COX-2 selective should
`provide excellent