Alimentary Pharmacology & Therapeutics
`
`The impact of low-dose aspirin on endoscopic gastric and
`duodenal ulcer rates in users of a non-selective non-steroidal
`anti-inflammatory drug or a cyclo-oxygenase-2-selective
`inhibitor
`J. L. GOLDSTEIN*, S. C. LOWRY , F. L. LANZAà, H. I. SCHWARTZ§ & W. E. DODGE 
`
`*College of Medicine, University of
`Illinois at Chicago, Chicago, IL, USA;
` Pfizer Inc., New York, NY, USA;
`àSection of Gastroenterology, Baylor
`College of Medicine, Houston, TX,
`USA; §Miami Research Associates,
`Miami, FL, USA
`
`Correspondence to:
`Dr J. L. Goldstein, Department of
`Medicine, University of Illinois at
`Chicago, Room 1020 CSB – 10th
`Floor (M/C 787), 840 S Wood Street,
`Chicago, IL 60612, USA.
`E-mail: jlgoldst@uic.edu
`
`Publication data
`Submitted 18 February 2006
`First decision 21 February 2006
`Resubmitted 2 March 2006
`Accepted 3 March 2006
`
`SUMMARY
`
`Background
`The effect of low-dose aspirin on endoscopic ulcer incidence in cyclo-
`oxygenase-2-selective inhibitor or non-selective non-steroidal anti-
`inflammatory drug users remains controversial.
`
`Aim
`To compare prospectively the incidence of endoscopic ulcers in healthy
`subjects receiving low-dose aspirin plus celecoxib or naproxen.
`
`Methods
`In this double-blind, placebo-controlled, 1-week study, subjects (50–
`75 years) were randomized to receive aspirin 325 mg o.d. plus either
`celecoxib 200 mg o.d., naproxen 500 mg b.d., or placebo. Baseline and
`end of study endoscopies were performed. The primary end point was
`incidence of one or more gastric and duodenal ulcers.
`
`Results
`A lower incidence of gastric and duodenal ulcers was seen in celecoxib/
`aspirin-treated subjects (19%) vs. naproxen/aspirin (27%; RR: 0.63, 95%
`CI: 0.44–0.92). Both naproxen/aspirin and celecoxib/aspirin groups
`demonstrated a higher incidence of gastric and duodenal ulcers vs. pla-
`cebo/aspirin (8%; RR: 3.7, 95% CI: 1.8–7.6 and RR: 2.6, 95% CI: 1.2–
`5.8, respectively).
`
`Conclusions
`Fewer endoscopic ulcers were observed in patients treated with celec-
`oxib/aspirin vs. naproxen/aspirin. However, celecoxib/aspirin was asso-
`ciated with a significantly higher incidence of gastric and duodenal
`ulcers than aspirin alone. Further studies are required to determine the
`generalizability of these findings in the aspirin users and to determine
`the appropriate strategy to minimize risk in susceptible patients.
`
`ª 2006 The Authors
`Journal compilation ª 2006 Blackwell Publishing Ltd
`doi:10.1111/j.1365-2036.2006.02912.x
`
`Aliment Pharmacol Ther 23, 1489–1498
`
`1489
`
`Page 1 of 10
`
`Patent Owner Ex. 2026
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`1490 J . L . G O L D S T E I N et al.
`
`INTRODUCTION
`
`MATERIALS AND METHODS
`
`The antithrombotic effects of so-called ‘low-dose’ aspi-
`rin (£325 mg/day) are well established, and it is widely
`used for primary and secondary prevention of cardio-
`vascular events, particularly in the older population.1–4
`Non-selective non-steroidal anti-inflammatory drugs
`(NSAIDs) are also prescribed frequently in older patients
`for
`the treatment of
`rheumatic diseases
`such as
`(RA).5
`osteoarthritis
`(OA) and rheumatoid arthritis
`However, it is well known that non-selective NSAIDs are
`associated with a heightened incidence of gastric and
`duodenal ulcer complications,6–8 and increasing age is a
`recognized risk factor for these complications.9–11
`Aspirin effects platelet and endothelial
`function
`through thromboxane A2
`(TxA2) and prostacyclin
`(PGI2) pathways respectively.4, 12, 13 It is believed that
`use of low-dose aspirin and its antithrombotic effect is
`beneficial
`for patients at heightened cardiovascular
`risk. However, aspirin also inhibits cyclo-oxygenase
`(COX)-1 activity in the gastric mucosa4, 14, 15 and has
`been found to be associated with an endoscopic gastric
`and/or duodenal ulcer prevalence of 11% at doses of
`75–325 mg o.d.16 Beyond endoscopic ulcers, aspirin is
`also associated with a dose-related propensity for the
`development of clinically significant gastric and duo-
`denal ulcer complications.17–20
`Concurrent administration of aspirin with non-select-
`ive NSAIDs is reported to be associated with substantial
`increases in the risk of upper gastrointestinal (GI) hae-
`morrhage compared with administration of each drug
`alone.18 COX-2-selective inhibitors are associated with
`significantly lower rates of upper GI events and endo-
`scopic ulcers compared with non-selective NSAIDs in the
`absence of aspirin use,21–27 but controversy continues to
`surround the issue of whether their GI safety benefit per-
`sists in the setting of low-dose aspirin co-therapy.24, 25
`Several endoscopic studies28–30 have provided conflicting
`results compared with outcomes studies24, 25, 31 regarding
`the relative and absolute risks associated with co-therapy
`of aspirin and COX-2-selective inhibitors.
`Given the discrepancies between endoscopic data and
`large outcomes trials, the present prospective study was
`designed to further investigate the short-term GI effects
`of a COX-2-selective inhibitor plus aspirin in an age-
`relevant population. We performed a multicentre, dou-
`ble-blind, randomized study to determine the incidence
`of gastric and duodenal endoscopic ulcers in healthy
`subjects aged 50–75 years receiving aspirin (325 mg
`o.d.) plus celecoxib, naproxen, or placebo.
`
`Study design
`
`This was a 7-day, multicentre, randomized, double-
`blind, double-dummy, active comparator and placebo-
`controlled, parallel group study. Volunteers were
`recruited from 39 centres in the United States. The
`protocol was approved by the Institutional Review
`Boards at each participating centre, and written
`informed consent was obtained from each subject prior
`to study entry and before any study-related procedures
`were performed.
`
`Patients
`
`Healthy adults aged 50–75 years were eligible for
`randomization if they had a normal physical examina-
`tion, normal clinical laboratory test results during the
`screening visit and if they had £5 erosions on baseline
`endoscopy.
`Individuals were excluded from the study if they had:
`a positive Helicobacter pylori serological test (FlexSure
`H. pylori Test, SmithKline Diagnostics, San Jose, CA,
`USA) at baseline; a gastric, pyloric or duodenal ulcer, or
`‡6 erosions in the stomach or duodenum at baseline
`endoscopy; any oesophageal ulcer or erosion; a history
`of or active GI disease; or diagnosis or treatment for
`ulcers within 30 days prior to the first dose of study
`medication. Volunteers were also excluded if they had
`used any over-the-counter or prescribed NSAIDs (inclu-
`ding aspirin), analgesics or anti-ulcer medication, ant-
`acids,
`systemic
`steroids or anticoagulants within
`2 weeks of the first dose of study medication. Any indi-
`vidual with a known history of chronic alcohol con-
`sumption, or alcohol or narcotic abuse were not
`permitted entry into the study. Women of child-bearing
`potential were required to have confirmed use of ade-
`quate contraception and a negative urine pregnancy test
`result within 24 h prior to receiving study medication.
`
`Pre-treatment period
`
`At the screening visit, which occurred within 10 days
`of the first dose of study medication, a full medical his-
`tory was obtained and each subject underwent a phys-
`ical
`examination and clinical
`laboratory testing,
`including a complete blood count, alanine amino
`transferase, aspartate amino transferase, blood urea
`nitrogen and creatinine. Serological testing for H. pylori
`
`ª 2006 The Authors, Aliment Pharmacol Ther 23, 1489–1498
`Journal compilation ª 2006 Blackwell Publishing Ltd
`
`Page 2 of 10
`
`Patent Owner Ex. 2026
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`I M P A C T O F A S P I R I N C O X - 2 C O - T H E R A P Y O N E N D O S C O P I C U L C E R R A T E S 1491
`
`was also conducted. Upper GI endoscopy was performed
`within 24 h prior to administration of study medication
`at the baseline/randomization visit (day 1). The investi-
`gator
`followed standard procedures for performing
`upper GI endoscopy and recorded pictures of any
`lesions visualized. Based on the endoscopy findings,
`investigators assigned gastric and duodenal mucosal
`scores according to a predefined 8-point grading scale
`ranging from 0 to 7, where: 0 ¼ normal mucosa; 1 ¼
`1–10 petechiae; 2 ¼ >10 petechiae; 3 ¼ 1–5 erosions;
`4 ¼ 6–10 erosions; 5 ¼ 11–25 erosions; 6 ¼ >25 ero-
`sions and 7 ¼ ulcer. Erosions were defined as a lesion
`with definite discontinuance in the mucosa but without
`depth. Ulcers were classified as any break in the mucosa
`‡3 mm in diameter with unequivocal depth.32
`
`Treatment period
`
`Based on a computer-generated process, subjects were
`randomized according to the order in which they
`enrolled in the study in a 2:2:1 ratio to celecoxib
`200 mg o.d. plus aspirin 325 mg o.d., naproxen
`500 mg b.d. plus aspirin 325 mg o.d. or placebo plus
`aspirin 325 mg o.d. for 7 days. Aspirin was provided
`to all enrolled subjects in an open-label fashion. Indi-
`viduals assigned to celecoxib treatment received celec-
`oxib 200 mg capsules and dummy capsules matching
`naproxen. Subjects randomized to the naproxen arm
`received naproxen 500 mg capsules and dummy cap-
`sules matching celecoxib; subjects randomized to pla-
`cebo received dummy capsules matching celecoxib
`and naproxen. All subjects were instructed to take
`study medication twice daily with a morning and
`evening meal.
`Use of any other drugs in addition to study medica-
`tion was discouraged during the treatment period. The
`following drugs were strictly prohibited: NSAIDs and
`COX-2-selective inhibitors, other than study medica-
`tion; other prescription or over-the-counter anti-
`inflammatory or analgesic drugs,
`including opioids,
`prescription or over-the-counter anti-ulcer medications
`and calcium-channel blockers. If considered necessary,
`subjects were allowed hormone-replacement therapy,
`vitamin supplements and paracetamol/acetaminophen
`(up to 500 mg b.d.).
`
`Post-treatment period
`
`Subjects returned for evaluation at day 7 or upon
`early termination. A second upper GI endoscopy was
`
`ª 2006 The Authors, Aliment Pharmacol Ther 23, 1489–1498
`Journal compilation ª 2006 Blackwell Publishing Ltd
`
`performed 2–4 h after the morning dose of study
`medication. For consistency,
`the same endoscopist,
`blinded to treatment, carried out baseline and post-
`treatment endoscopies using the same equipment and
`measuring devices. Subjects also underwent a physical
`examination and clinical laboratory tests.
`
`Study end points
`
`The primary study end point was the incidence of gas-
`tric and duodenal ulcers, defined as ‡1 gastric, pyloric
`channel or duodenal ulcer, as determined by upper GI
`endoscopy on day 7 of treatment. Secondary study
`end points included the incidence of any gastric or
`duodenal ulcer (grade 7) and any gastric and duode-
`nal, gastric or duodenal erosion/ulcer (grades 4–7).
`Adverse events were self-reported and were collected
`and aggregated for all randomized subjects.
`
`Sample size calculation
`
`In a previous 7-day endoscopy study in healthy sub-
`jects, gastric and duodenal ulcer rates were: placebo
`0%, celecoxib 100 mg b.d. 0% and naproxen 500 mg
`b.d. 19%.33 Using these rates and assuming that aspi-
`rin had no effect on ulcer incidence in the naproxen
`arm, it was determined that a sample size of 164 sub-
`jects per group was required to maintain ‡80% power
`to detect significant differences between celecoxib and
`naproxen, provided the ulcer incidence in the celec-
`oxib arm did not exceed 8%. These calculations were
`based on a two-sided Fisher’s exact test with a 5% sig-
`nificance level.
`
`Statistical analyses
`
`The endoscopy evaluable population, used for analysis
`of
`the endoscopy data,
`included all
`randomized
`patients who received ‡1 dose of study medication
`and had both baseline and post-treatment endoscopies.
`All subjects who were randomized and received ‡1
`dose of study medication were included in the safety
`analyses.
`groups with
`treatment
`between
`Comparability
`respect to gender and race was assessed by Cochran-
`Mantel-Haenszel (CMH) statistics with stratification by
`centre; two-way analysis of variance with centre and
`treatment as fixed factors was used to determine com-
`parability between treatment arms with respect to con-
`tinuous variables such as age and vital signs.
`
`Page 3 of 10
`
`Patent Owner Ex. 2026
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`1492 J . L . G O L D S T E I N et al.
`
`incidence of gastric and
`In the primary analysis,
`duodenal ulcers (grade 7) on day 7 was calculated for
`celecoxib plus aspirin and naproxen plus aspirin
`groups and was compared using CMH statistics strati-
`fied by centre. Secondary analyses involved the com-
`parison of gastric and duodenal ulcer incidence for
`both active treatment arms vs. placebo plus aspirin, as
`well as pairwise group comparisons of incidence of
`gastric and duodenal ulcers (grade 7), gastric and duo-
`denal ulcers/erosions, gastric ulcers/erosions and duo-
`denal ulcers/erosions (grades 4–7) using CMH tests,
`again stratified by centre.
`
`RESULTS
`
`Patient demographics and disposition
`
`Four hundred and sixty-four subjects were randomized
`and 463 of these subjects received at least one dose of
`study medication (187 in the celecoxib plus aspirin
`group, 182 in the naproxen plus aspirin group and 94
`in the placebo plus aspirin group). The single subject
`who did not receive study medication was randomized
`to the naproxen plus aspirin group but was identified
`as having a pre-existing protocol violation (H. pylori-
`positive) and did not proceed in the trial. Across the
`three arms of the trial, a total of 13 subjects prema-
`turely discontinued study medication and 12 of these
`did not undergo a post-baseline endoscopy. These 13
`subjects included five individuals with a pre-existing
`protocol violation, who were found to be H. pylori-
`
`positive, and three who discontinued due to adverse
`events (two GI adverse events and one vision disorder;
`none of which was considered serious; Figure 1).
`At baseline, the treatment arms were well matched
`with respect to demographics and baseline endoscopy
`findings for the randomized population (Table 1). Sub-
`jects ranged in age from 50 to 75 years, with a mean
`age of approximately 57 years, and were predomin-
`antly Caucasian (77–84%) and female (55–65%). Med-
`ical history was similar between treatment arms. Less
`than 3% of subjects in any group had a history of a
`gastric and duodenal ulcer and none had a history of
`upper GI bleeding.
`No statistically significant differences were observed
`in the mean pre-treatment gastric and duodenal endo-
`scopy results across all three groups (P ‡ 0.08, Table 1).
`
`Primary end point
`
`During this 1-week study, analysis of the endoscopy
`evaluable population demonstrated that significantly
`fewer subjects in the celecoxib plus aspirin group
`developed ‡1 gastric and duodenal ulcer compared
`with the naproxen plus aspirin group [relative risk
`(RR): 0.63, 95% CI: 0.44–0.92; P ¼ 0.016; Figure 2a].
`The number-needed-to-treat
`(NNT)
`to prevent one
`ulcer was 12. A significantly higher proportion of sub-
`jects in the celecoxib plus aspirin group developed
`gastric and duodenal ulcers compared with the placebo
`plus aspirin-treated subjects (RR: 2.6, 95% CI: 1.2–5.8;
`
`Figure 1. Subject disposition.
`
`ª 2006 The Authors, Aliment Pharmacol Ther 23, 1489–1498
`Journal compilation ª 2006 Blackwell Publishing Ltd
`
`Page 4 of 10
`
`Patent Owner Ex. 2026
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`I M P A C T O F A S P I R I N C O X - 2 C O - T H E R A P Y O N E N D O S C O P I C U L C E R R A T E S 1493
`
`Table 1. Demographics and baseline characteristics – randomized population
`
`Celecoxib 200 mg o.d.
`plus ASA 325 mg o.d.
`n ¼ 187)
`
`Naproxen 500 mg b.d.
`plus ASA 325 mg o.d.
`(n ¼ 183)
`
`Placebo plus
`ASA 325 mg o.d.
`(n ¼ 94)
`
`56.7  5.8
`119 (63.6)
`
`149 (79.7)
`11 (5.9)
`5 (2.7)
`1 (0.5)
`21 (11.2)
`0 (0)
`
`Age (years; mean  s.d.)
`Female, n (%)
`Race, n (%)
`Caucasian
`Black
`Asian
`Native American
`Hispanic/Latin American
`Other
`Medical history, n (%)
`Gastric and duodenal ulcers
`2 (1.1)
`Angina
`1 (0.5)
`Coronary artery disease
`1 (0.5)
`Myocardial infarction
`1 (0.5)
`Other cardiovascular history
`17 (9.1)
`Musculoskeletal disease
`48 (25.7)
`Mean pre-treatment endoscopy lesions (s.d.)
`Gastric mucosa
`1.0 (3.5)
`Number of petechiae
`Range
`0–30
`0.2 (0.8)
`Number of erosions
`Range
`0–5
`Duodenal mucosa
`0.3 (1.4)
`Number of petechiae
`Range
`0–11
`0.1 (0.3)
`Number of erosions
`Range
`0–3
`
`56.5  5.9
`101 (55.2)
`
`153 (83.6)
`6 (3.3)
`1 (0.5)
`0 (0)
`21 (11.5)
`2 (1.1)
`
`3 (1.6)
`2 (1.1)
`1 (0.5)
`0 (0)
`18 (9.8)
`56 (30.6)
`
`0.6 (1.8)
`0–10
`0.2 (0.7)
`0–4
`0.3 (1.6)
`0–15
`0 (0.1)
`0–1
`
`57.2  7.1
`61 (64.9)
`
`72 (76.6)
`8 (8.5)
`2 (2.1)
`0 (0)
`11 (11.7)
`1 (1.1)
`
`2 (2.1)
`0 (0)
`0 (0)
`0 (0)
`6 (6.4)
`27 (28.7)
`
`1.2 (7.8)
`0–72
`0.2 (0.8)
`0–5
`0.1 (1.0)
`0–10
`0 (0)
`0
`
`P-value
`
`0.738*
`0.094 
`
`0.308 
`
`0.670 
`0.606 
`0.782 
`0.472 
`0.499 
`0.585 
`
`0.487*
`
`0.974*
`
`0.453*
`
`0.079*
`
`ASA, aspirin.
`* P-value from two-way analysis of variance with centre and treatment as fixed effects.
`  P-value from Cochran-Mantel-Haenszel test stratified by centre.
`
`P ¼ 0.008; Figure 2a). A higher incidence of gastric
`and duodenal ulcers was also observed in naproxen-
`treated subjects compared with the placebo group (RR:
`3.7, 95% CI: 1.8–7.6; P < 0.001).
`
`Secondary end points
`
`At the post-treatment endoscopy, more gastric ulcers
`were observed in the naproxen plus aspirin (RR: 3.3,
`95% CI: 1.5–7.3; P ¼ 0.002; Figure 2b) and celec-
`oxib plus aspirin groups (RR: 2.7, 95% CI: 1.2–6.4;
`P ¼ 0.012) compared with the placebo group. The
`incidence of gastric ulcers in the celecoxib treatment
`arm was numerically lower
`than the naproxen
`group, but this difference was not significantly dif-
`
`ª 2006 The Authors, Aliment Pharmacol Ther 23, 1489–1498
`Journal compilation ª 2006 Blackwell Publishing Ltd
`
`ferent (RR: 0.79, 95% CI: 0.52–1.2; P ¼ 0.269; Fig-
`ure 2b). Duodenal ulcers were less common than
`gastric ulcers in all treatment groups and developed
`in significantly fewer subjects treated with celecoxib
`plus aspirin compared with naproxen (RR: 0.40, 95%
`CI: 0.17–0.94; P ¼ 0.027; Figure 2c). Comparisons
`with the placebo plus aspirin group showed that
`subjects treated with naproxen were more likely to
`develop endoscopic duodenal ulcers (RR: 9.9, 95%
`CI: 1.2–83.9; P ¼ 0.006), and a numeric trend was
`apparent for celecoxib (RR: 5.1, 95% CI: 0.59–44.5;
`P ¼ 0.095; Figure 2c).
`Other secondary end points (incidence of gastric and
`duodenal ulcers/erosions) supported the results of the
`primary analysis (Figure 3).
`
`Page 5 of 10
`
`Patent Owner Ex. 2026
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`1494 J . L . G O L D S T E I N et al.
`
`Adverse events
`
`The majority of adverse events were mild-to-moderate
`in severity with GI system disorders and headache
`being the most commonly reported. No serious adverse
`event occurred in the course of the study.
`
`Post hoc analysis
`As subjects with £5 erosions in the stomach or duode-
`num at baseline endoscopy were allowed into the trial,
`an exploratory post hoc analysis was performed on
`subjects with no lesions at baseline (76 subjects in the
`placebo plus aspirin arm, 148 subjects in the celecoxib
`plus aspirin arm and 141 subjects in the naproxen plus
`aspirin arm). Based on the subjects without erosions at
`baseline, the results of the post hoc analysis were con-
`sistent with the primary analysis with an incidence of
`gastric and duodenal ulcers of 17% (25 of 148), 27%
`(38 of 141) and 5% (four of 76) for celecoxib plus
`aspirin, naproxen plus aspirin and placebo plus aspi-
`rin, respectively (P < 0.01 for both active comparators
`vs. placebo and P < 0.05 for celecoxib vs. naproxen;
`Mantel-Haenszel chi-square).
`
`DISCUSSION
`
`It remains controversial as to whether the upper GI
`safety advantage of COX-2-selective inhibitors com-
`pared with non-selective NSAIDs is maintained in the
`setting of low-dose aspirin co-therapy; the clinically
`relevant question is whether or not use of concomitant
`low-dose aspirin, by introducing COX-1 inhibition,
`negates the benefit of a COX-2-selective inhibitor. The
`present placebo-controlled study was conducted to
`compare the incidence of endoscopic ulcers after
`1-week
`of
`therapy
`in
`healthy
`subjects
`(aged
`50–75 years) receiving celecoxib plus aspirin or napr-
`oxen plus aspirin. A significant reduction (37%) in risk
`of gastric and/or duodenal ulcer
`incidence was
`observed in subjects treated with celecoxib plus aspirin
`compared with naproxen plus aspirin (Figure 2a); as
`such, our results indicate that short-term use of aspirin
`does not completely negate the potential GI benefit of
`celecoxib as measured by endoscopic ulcer
`rates.
`However, it is important to recognize that concomitant
`aspirin use was not without impact in COX-2-selective
`
`Figure 2. Incidence of (a) gastric and duodenal ulcers,
`(b) gastric ulcers and (c) duodenal ulcers determined by
`endoscopy in the endoscopy-evaluable population after
`treatment for 1 week with celecoxib 200 mg o.d. plus
`aspirin (ASA) 325 mg o.d., naproxen 500 mg b.d. plus
`ASA 325 mg or placebo plus ASA 325 mg o.d. Treatment
`comparisons based on Cochran-Mantel-Haenszel test
`stratified by centre. * P £ 0.027 vs. naproxen plus ASA;
`  P £ 0.095 vs. placebo plus ASA.
`
`ª 2006 The Authors, Aliment Pharmacol Ther 23, 1489–1498
`Journal compilation ª 2006 Blackwell Publishing Ltd
`
`Page 6 of 10
`
`Patent Owner Ex. 2026
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`I M P A C T O F A S P I R I N C O X - 2 C O - T H E R A P Y O N E N D O S C O P I C U L C E R R A T E S 1495
`
`inhibitor users; subjects treated with celecoxib plus
`aspirin had significantly higher rates of gastric and/or
`duodenal ulcers (RR: 2.6; Figure 2a) and gastric and/or
`duodenal ulcers/erosions (RR: 1.4; Figure 3a) com-
`pared with those receiving placebo plus aspirin.
`
`ª 2006 The Authors, Aliment Pharmacol Ther 23, 1489–1498
`Journal compilation ª 2006 Blackwell Publishing Ltd
`
`Aspirin is commonly prescribed or self-administered
`for the primary and secondary prevention of thrombo-
`embolic cardiovascular events.1–4 Overlap of aspirin
`and anti-inflammatory-using populations is significant;
`a high proportion of aspirin users also require concom-
`itant use of non-selective NSAIDs or COX-2-selective
`inhibitors for the treatment of pain and arthritis. In a
`recent telephone survey of long-term COX-2-selective
`inhibitor users, 45% of respondents reported taking
`aspirin on at least 15 of the last 30 days.34 Thus, the
`question of COX-2-selective inhibitor GI safety in the
`presence of aspirin is of substantial clinical importance
`given the frequency of their co-administration in real
`time practice. In addition, in light of recent concerns
`involving the cardiovascular safety of COX-2-selective
`inhibitors4, 13, 21, 25, 35, 36 and non-selective NSAIDs,
`the incidence of aspirin co-administration with these
`agents for cardioprotection is likely to increase further.
`There are two lines of evidence that address the rel-
`ative or absolute impact of aspirin in non-selective
`NSAID or COX-2-selective inhibitor users. The first
`employs the measurement of endoscopic ulcer rates
`after a predefined exposure period. Use of this model
`implies that endoscopic ulcer rates are surrogate mark-
`ers
`for ulcer
`complications; an assumption that
`remains unproven. The second approach is to directly
`measure absolute gastric and duodenal ulcer complica-
`tion rates as the clinical outcome of exposure. Based
`on existing data,
`there are inconsistencies in the
`findings from these two approaches. For example,
`in subanalyses of aspirin users from two large-scale
`outcomes trials, it has been demonstrated that in the
`presence of aspirin use there was no significant differ-
`ence in the incidence of upper GI ulcer complications
`between non-selective NSAID and the COX-2-selective
`inhibitor users – celecoxib in CLASS24 or lumiracoxib
`in TARGET.25 Neither of these trials was specifically
`powered to definitively address the issue of ulcer com-
`plications in aspirin users. As such, the generalizability
`
`Figure 3. Incidence of (a) gastric and duodenal ulcers/
`erosions, (b) gastric and (c) duodenal ulcers/erosions
`determined by endoscopy in the endoscopy-evaluable
`population after treatment for 1 week with celecoxib
`200 mg o.d. plus aspirin (ASA) 325 mg o.d., naproxen
`500 mg b.d. plus ASA 325 mg or placebo plus ASA
`325 mg o.d. Erosion/ulcer is defined as an upper gastro-
`intestinal mucosal endoscopy score of ‡4. Treatment
`comparisons based on Cochran-Mantel-Haenszel test stra-
`tified by centre. * P £ 0.008 vs. naproxen plus ASA;
`  P £ 0.027 vs. placebo plus ASA.
`
`Page 7 of 10
`
`Patent Owner Ex. 2026
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`1496 J . L . G O L D S T E I N et al.
`
`these results is open to question. The clinical
`of
`importance of this issue needs to be addressed in ade-
`quately powered long-term safety studies evaluating
`ulcer complications. Until such time, the possibility
`remains of a relative benefit of COX-2-selective inhib-
`itors compared with non-selective NSAIDs in aspirin
`users.
`Interestingly,
`in the subpopulation of aspirin
`users in the SUCcessive Celecoxib Efficacy and Safety
`Study (SUCCESS-I), which comprised approximately
`7% of the study population, the incidence of ulcer
`complications was numerically lower in the celecoxib
`group compared with the non-selective NSAIDs group,
`but
`this difference was not statistically significant
`(odds ratio: 2.0, 95% CI: 0.12–31.7).31
`In contrast to the findings from CLASS and TARGET,
`previously reported endoscopic studies offer mixed
`results. Two separate retrospective post hoc analyses of
`pooled endoscopic trials, one with celecoxib and the
`second with valdecoxib, demonstrated that the rate of
`endoscopic ulcers was approximately 50% lower in
`aspirin users (81–325 mg/day)
`receiving a COX-2-
`selective inhibitor compared with those randomized to a
`non-selective NSAID plus aspirin.28, 29 However, in con-
`trast to these results, a recent prospective analysis by
`Laine et al. demonstrated that the use of rofecoxib
`(25 mg o.d.) with low-dose aspirin (81 mg o.d.)
`increased the incidence of endoscopic ulcers (‡3 mm in
`diameter) to a rate not significantly different from that
`of ibuprofen alone. The cumulative endoscopic ulcer
`incidence rate was 16% in the group receiving aspirin
`plus rofecoxib compared with 17% in the subjects
`receiving ibuprofen (P ¼ 0.62).37 Of note, however, is
`that this analysis did not include a non-selective NSAID
`plus aspirin treatment arm. While these endoscopic
`findings are important, they do not provide definitive
`guidance for making therapeutic decisions in patients
`requiring anti-inflammatory therapy for pain and low-
`dose aspirin for cardiovascular prophylaxis.
`The question remains as to the clinical applicability
`of our current findings. While not universally accep-
`ted, endoscopic trials have been designed to investi-
`gate and have been published to evaluate the relative
`effects of anti-inflammatory therapies on the upper GI
`mucosa. Assuming that endoscopic ulcers are surro-
`gate markers, as previously suggested, then we have to
`ask the question of whether a 37% reduction in endo-
`scopic ulcer risk with celecoxib plus aspirin compared
`with naproxen plus aspirin represents a sufficient
`improvement to apply as a single strategy in clinical
`practice. With this in mind, it is not unreasonable to
`
`suggests that the use of proton-pump inhibitors (PPIs)
`may be warranted in patients who require therapy
`with aspirin plus a COX-2-selective inhibitor especially
`in those at high risk for developing non-selective
`NSAID- or aspirin-related gastric and duodenal com-
`plications.38, 39 Chan et al. recently demonstrated that
`the rate of endoscopic gastric and duodenal ulcer
`recurrence in aspirin users was 50% and 31% after
`6 months of therapy in patients receiving celecoxib or
`diclofenac plus omeprazole, respectively, suggesting
`that neither strategy was highly effective for ulcer pre-
`vention in this high-risk patient population.40, 41 Fur-
`thermore, in a recently described preliminary report on
`a 12-week comparative trial, no statistically significant
`difference in the incidence of endoscopic gastric and/
`or duodenal ulcers was observed in OA patients using
`aspirin (81 mg o.d. or 325 mg o.d.) for cardiovascular
`prophylaxis randomized to receive celecoxib 200 mg
`o.d. or naproxen 500 mg b.d. plus lansoprazole 30 mg
`o.d.42 The next logical step may therefore be to pre-
`scribe COX-2-selective inhibitors with a PPI in the set-
`ting of aspirin co-therapy, particularly in patients at
`high risk. Although there is little prospective data sup-
`porting the additive or synergistic benefits of adding a
`PPI
`to a COX-2-selective inhibitor, results from a
`recent retrospective cohort study have demonstrated
`an approximate 40% reduction in the risk of GI hos-
`pitalizations in an elderly population [‡75 years; haz-
`ard ratio (HR): 0.6; 95% CI: 0.4–0.8] and in an
`aspirin-using population (HR: 0.6; 95% CI: 0.4–1.0)
`receiving celecoxib plus a PPI compared with celec-
`oxib alone.43 Given the evolving concerns regarding
`the cardiovascular safety of COX-2-selective inhibitors,
`an alternative approach for patients with (or without)
`cardiovascular risk might be aspirin or other antiplate-
`let therapy and a non-selective NSAID plus a PPI,
`notably at the lowest dose of aspirin and non-selective
`NSAID that are clinically effective. However, it is not
`entirely clear that non-selective NSAIDs are devoid of
`cardiovascular risk44 and if prescribed, one should be
`cognizant of the potential pharmacodynamic interac-
`tion that may exist between simultaneous dosing of
`ibuprofen and aspirin.45, 46 With the ambiguities in the
`absolute and RR of using different anti-inflammatory
`agents in patients at cardiovascular risk, we believe our
`data and the potential generalizability of our data, if
`any, should be cautiously applied in the clinical setting
`based on doctor understanding and patient preference.
`Hopefully, large, outcomes-driven, randomized-con-
`trolled trials will help evaluate the relative benefit (if
`
`ª 2006 The Authors, Aliment Pharmacol Ther 23, 1489–1498
`Journal compilation ª 2006 Blackwell Publishing Ltd
`
`Page 8 of 10
`
`Patent Owner Ex. 2026
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`I M P A C T O F A S P I R I N C O X - 2 C O - T H E R A P Y O N E N D O S C O P I C U L C E R R A T E S 1497
`
`any) of the combination of a COX-2-selective inhibitor
`plus PPI or non-selective NSAID plus PPI in aspirin
`users. Given that aspirin use and cardiovascular dis-
`ease independently increase GI risk, such data are nee-
`ded to help guide practice decisions for patients with
`chronic musculoskeletal conditions necessitating anti-
`inflammatory therapies.
`In addition to the questions and limitations of evalu-
`ating the GI toxicity of anti-inflammatory therapies
`using endoscopic ulcer rates, it should be noted that
`our study utilized 325 mg of aspirin, a dose higher than
`that commonly used in clinical practice for cardiovas-
`cular prophylaxis. As aspirin-associated GI toxicity is
`seemingly dose-dependent,47 similar endoscopic studies
`are needed to evaluate the now more commonly used
`aspirin dose of 81 mg o.d.4
`In conclusion, our study found that the incidence of
`endoscopic gastric and duodenal ulcers and erosions
`was significantly lower with celecoxib 200 mg o.d.
`plus aspirin 325 mg o.d. compared with naproxen plus
`aspirin in healthy subjects treated for 1 week. Subjects
`treated with celecoxib plus aspirin did, however,
`develop significantly more ulcers compared with pla-
`
`cebo plus aspirin. Detection of endoscopic ulcers may
`not be an ideal surrogate marker for complications
`and may provide a poor correlation with clinical out-
`comes. Thus, the question of the GI benefit associated
`with COX-2-selective inhibitors in the setting of aspi-
`rin co-therapy remains open to interpretation and fur-
`ther studies are required to fully resolve this issue. In
`the current environment, it is increasingly important
`that the risk-benefit of all treatments are fully weighed
`and understood in order to provide the most effective
`anti-inflammatory therapy in aspirin-using patients.
`
`ACKNOWLEDGEMENT
`
`This study was sponsored by Pfizer Inc., NY, USA;
`Dr Goldstein is a consultant to and has received travel
`expenses, educational grants, research grants and con-
`tracts, and speaker honoraria from Pfizer Inc., Astra-
`Zeneca, TAP, Novartis, Takeda, POZEN, and Sucampo;
`Dr Lanza has received research grants from Pfizer Inc.,
`Merck, Johnson and Johnson, and TAP; Dr Schwartz
`has no conflicts of interests to declare; and Drs Lowry
`and Dodge are employees of Pfizer Inc.
`
`REFERENCES
`
`1 Antithrombotic Trialists’ Collaboration.
`Collaborative meta-analysis of random-
`ised trials of antiplatelet
`therapy for
`prevention of death, myocardial infarc-
`tion, and stroke in high risk patients.
`BMJ 2002; 324: 71–86.
`2 Meade TW, Brennan PJ. Determination
`of who may derive most benefit from
`aspirin in primary prevention: subgroup
`results from a randomised controlled
`trial. BMJ 2000; 321: 13–7.
`3 Sanmuganatha