THE AMERICAN JOURNAL OF GASTROENTEROLOGY
`© 2000 by Am. Coll. of Gastroenterology
`Published by Elsevier Science Inc.
`
`Vol. 95, No. 9, 2000
`ISSN 0002-9270/00/$20.00
`PII S0002-9270(00)01040-6
`
`Risk of Upper Gastrointestinal Bleeding
`Associated With Use of Low-Dose Aspirin
`Henrik Toft Sørensen, M.D., Ph.D., Lene Mellemkjær, Ph.D., William J. Blot, Ph.D.,
`Gunnar Lauge Nielsen, M.D., Flemming Hald Steffensen, M.D., Joseph K. McLaughlin, Ph.D., and
`Jørgen H. Olsen, M.D., Ph.D.
`Department of Clinical Epidemiology, Aarhus University and Aalborg Hospitals, Aarhus, Denmark; The
`Institute of Cancer Epidemiology, The Danish Cancer Society, Copenhagen, Denmark; The International
`Epidemiology Institute, Rockville, Maryland; The Department of Medicine M, Aalborg Hospital, Aalborg,
`Denmark; and the Danish Epidemiology Science Centre at the Department of Epidemiology and Social
`Medicine, University of Aarhus, Aarhus, Denmark
`
`OBJECTIVE: Aspirin products are known to cause irritation
`and injury to the gastric mucosa. We examined the risk of
`hospitalization for upper gastrointestinal bleeding with use
`of low-dose aspirin.
`METHODS: This was a cohort study based on record linkage
`between a population-based prescription database and a
`hospital discharge registry in North Jutland County, Den-
`mark, from January 1, 1991, to December 31, 1995. Inci-
`dence rates of upper gastrointestinal bleeding in 27,694
`users of low-dose aspirin were compared with the incidence
`rates in the general population in the county.
`RESULTS: A total of 207 exclusive users of low-dose aspirin
`experienced a first episode of upper gastrointestinal bleed-
`ing with admission to the hospital during the study period.
`The standardized incidence rate ratio was 2.6 (95% confi-
`dence interval, 2.2–2.9), 2.3 in women and 2.8 in men. The
`standardized incidence rate ratio for combined use of low-
`dose aspirin and other nonsteroidal anti-inflammatory drugs
`was 5.6 (95% confidence interval, 4.4 –7.0). The risk was
`similar among users of noncoated low-dose aspirin (stan-
`dardized incidence rate ratio, 2.6; 95% confidence interval,
`1.8 –3.5) and coated low-dose aspirin (standardized inci-
`dence rate ratio, 2.6; 95% confidence interval, 2.2–3.0).
`CONCLUSIONS: Use of low-dose aspirin was associated with
`an increased risk of upper gastrointestinal bleeding, with
`still higher risks when combined with other nonsteroidal
`anti-inflammatory drugs. Enteric coating did not seem to
`reduce the risk. The findings from this observational study
`raise the possibility that prophylactic use of low-dose aspirin
`may convey an increased risk of gastrointestinal bleeding,
`which may offset some of its benefits. (Am J Gastroenterol
`2000;95:2218–2224. © 2000 by Am. Coll. of Gastroenterol-
`ogy)
`
`INTRODUCTION
`
`Aspirin is one of the oldest and most widely used drugs in
`the world. Standard doses of 750 –1000 mg are used for pain
`relief, whereas the use of low-dose aspirin is increasing in
`secondary prevention of cardiovascular disease (1). The
`most frequent side effects of aspirin are related to the
`gastrointestinal (GI) tract (2, 3). Many animal and human
`studies have shown that aspirin causes gastric mucosal ero-
`sions and inhibition of thromboxane synthesis (4). Standard
`doses of aspirin are associated with a substantial risk of
`upper GI bleeding, but our knowledge about the risk asso-
`ciated with low doses is mainly limited to a few randomized
`trials and case-control studies, some of which suggest that
`no aspirin regimen is free of the risk of upper GI bleeding
`(5–16). Upper GI bleeding is a common reason for admis-
`sion to the hospital, especially among the elderly, and has an
`estimated case fatality rate of 8% (17, 18). Moreover, aspi-
`rin has often not been included in studies of the risk of upper
`GI bleeding associated with use of nonsteroidal anti-inflam-
`matory drugs (NSAIDs), because of underascertainment of
`exposure (19, 20).
`Because aspirin is used so widely, even small risks of side
`effects have major clinical and public health implications.
`On this background, we conducted a population-based co-
`hort study in Denmark to examine: 1) the risk of upper GI
`bleeding associated with use of low-dose aspirin, adjusted
`for use of other drugs that may increase the risk of bleeding;
`2) whether concurrent nonaspirin NSAID treatment affects
`the risk of bleeding; and 3) whether enteric coating reduces
`the risk of bleeding.
`
`MATERIALS AND METHODS
`
`The study was carried out within the population of the
`county of North Jutland in Denmark, which comprised
`about 490,000 inhabitants, approximately 9% of the total
`Danish population, during the study period 1991 to 1995.
`
`Page 1 of 7
`
`Patent Owner Ex. 2011
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`AJG – September, 2000
`
`Risk of Upper GI Bleeding and Low-Dose Aspirin
`
`2219
`
`The National Health Service provides tax-supported health
`care for all inhabitants, guaranteeing free access to general
`practitioners, hospitals, and public clinics and refunding a
`variable proportion of the costs of drugs prescribed by
`physicians. The population-based Pharmaco-Epidemiologi-
`cal Prescription Database of North Jutland (21, 22), initiated
`on January 1, 1991, retains key information on prescriptions
`for refundable drugs dispensed from all 33 pharmacies in the
`county. This includes the personal identification number of
`the customer, type of drug prescribed according to the
`anatomical therapeutical chemical (ATC) classification sys-
`tem (23), and the date of prescription (date of dispensing the
`drug). The personal identification number comprises 10
`digits that encode gender and date of birth. With its use, a
`complete prescription history can be established for each
`individual, and unambiguous linkages with other registers
`can be performed.
`In the Pharmaco-Epidemiological Prescription Database
`we identified 30,952 users of low-dose tablets of 100 mg
`(19.2%) or 150 mg (80.8%) of aspirin during the period
`1991–1995. Regular use of low-dose aspirin for prevention
`of cardiovascular disease is most likely prescribed by doc-
`tors in Denmark, for which the patients get 50% of the cost
`refunded. Record linkage with the Danish mortality files
`resulted in the exclusion of 29 persons (0.1%) because of a
`date of death before the date of aspirin prescription, or to an
`error in the identification number. We also excluded 15
`users (,0.1%) who were aged ,16 yr or .105 yr at the date
`of the aspirin prescription, and 602 users (2.0%) who were
`not residents in the county. The remaining 30,306 persons
`were linked to the Regional Hospital Discharge Register
`(HDR), which, on a permanent basis, retains key informa-
`tion on all patients discharged from somatic hospitals in the
`county during 1977–1995 (24). The files of the HDR include
`information on the identification number of the patient, date
`of discharge, and up to 20 discharge diagnoses (24), coded
`according to the Danish version of the International Classi-
`fication of Diseases, the 8th revision (24) until the end of
`1993, and the 10th revision (24) since. Based on the hospital
`discharge history, we decided to exclude an additional 961
`persons because, before the date of the first notified aspirin
`prescription, they had been diagnosed with an upper GI
`bleeding (n 5 437; 1%), which was the outcome of the
`present investigation, i.e., bleeding caused by a gastric,
`duodenal, or gastrojejunal ulcer as well as hematemesis and
`melena (22) (ICD-8 5 530.98, 531.90, 531.92, 531.95,
`532.90, 533.90, 534.90, 535.01, 784.59, 785.79; ICD-10 5
`K25.0, K25.2, K25.4, K25.6, K26.0, K26.2, K26.4, K26.6,
`K27.0, K27.2, K27.4, K27.6, K28.0, K28.2, K28.4, K28.6,
`K29.0, K92.0-2), or diagnosed with a medical condition
`predisposing to GI bleeding (N 5 524, 2%), i.e., alcoholism
`(ICD-8 5 303; ICD-10 5 F10), esophageal varices (IDC-
`8 5 456.00-09; ICD-10 5 I85, I98.2), Mallory-Weiss syn-
`drome (ICD-8 5 530.97; ICD-10 5 K22.6), or liver cirrho-
`sis (ICD-8 5 571, 573; ICD-10 5 K70, K72-74, K76) (22).
`Another predisposing condition is cancer, so a further 1,651
`
`(5%) subjects were excluded after linkage with the Danish
`Cancer Registry (25) because they had a cancer diagnosis
`after 1980 (all cancer types except nonmelanoma skin can-
`cer) and before the first aspirin prescription. This restriction
`was made to avoid confounding from these conditions. The
`size of our study material did not allow analyses within the
`subgroups of predisposing conditions. Thus, a total of
`27,694 (89.5%) users of low-dose aspirin were included in
`the study cohort.
`
`Person-Years at Risk
`The follow-up for hospitalization of upper GI bleeding
`began at the date of the first notified prescription of low-
`dose aspirin (100 or 150 mg), and ended at the date of a first
`admission to hospital for a GI bleeding or for one of the
`medical conditions that predispose to GI bleeding (alcohol-
`ism, esophageal varices, Mallory-Weiss syndrome, liver
`cirrhosis, or cancer), the date of death, or December 31,
`1995, whichever occurred first. The follow-up period of
`cohort members was then subdivided into periods of current
`exposure to low-dose aspirin, lasting from the date of pre-
`scription until 90 days thereafter (or a censoring date), and
`periods of no exposure extending from 90 days after a
`prescription to the date of next prescription (or a censoring
`date). The period of 90 days was chosen because low-dose
`aspirin is mainly prescribed in packets for 3-month use.
`Furthermore, “exposure” and “no exposure” time segments
`were flagged if cohort members received prescriptions at the
`same time of one or more other drugs suspected to cause GI
`bleeding, i.e., high-dose aspirin, other NSAIDs, vitamin K
`antagonists, or oral corticosteroids, again applying the 90-
`day rule. Periods of low-dose aspirin use, to which all
`27,694 persons contributed follow-up time, were further
`divided into time segments of “only low-dose aspirin” use
`(n 5 26,196), time segments of combined low-dose aspirin
`and other NSAID use (n 5 10,021) and time segments of
`exposure to low-dose aspirin in combinations with one of
`the remaining drugs (n 5 3,976). In addition, periods of
`“only low-dose aspirin” use were further subdivided into
`enteric-coated aspirin use (n 5 6,071) and non– enteric-
`coated aspirin use (n 5 22,678) and into use of 100 mg
`tablets (n 5 6,084) and 150 mg tablets (n 5 22,671).
`In a subanalysis, we restricted the cohort to 17,328 per-
`sons who had no record of a prescription for one of the other
`drugs mentioned in the present study before receiving the
`first prescription for low-dose aspirin. For these persons, we
`investigated the risk of upper GI bleeding during periods of
`low-dose aspirin use only versus subsequent periods after
`cessation of low-dose aspirin (and no use of any of the
`drugs), i.e., for current versus former use of low-dose aspi-
`rin. In all, 84% of the restricted cohort contributed follow-up
`time to the latter time segments.
`
`Standardized Incidence Ratios
`The number of upper GI bleedings observed among cohort
`members allocated to the appropriate exposure category was
`
`Page 2 of 7
`
`Patent Owner Ex. 2011
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`2220
`
`Sørensen et al.
`
`AJG – Vol. 95, No. 9, 2000
`
`Table 1. Descriptive Characteristics of the Study Cohort of Users of Low-Dose Aspirin in the County of North Jutland, Denmark,
`1991–1995
`
`Person-Years of Follow-Up
`Current Aspirin Use
`Aspirin and Other
`Aspirin and
`NSAIDs*
`Other Drugs†
`5,716
`1,778
`3,327
`904
`353
`74
`591
`184
`2,384
`646
`2,389
`874
`381
`119
`566
`214
`1,441
`541
`
`Total Use
`42,054
`21,146
`2,304
`4,449
`14,393
`20,907
`3,624
`5,782
`11,501
`
`Former
`Use
`24,738
`12,594
`1,828
`2,688
`8,078
`12,144
`2,675
`3,241
`6,227
`
`Characteristics
`Total cohort
`Women (yr)
`16–59
`60–69
`701
`Men (yr)
`16–59
`60–69
`701
`Coated aspirin‡
`Noncoated aspirin\
`* Low-dose aspirin and other NSAIDs alone or in combination with one of the other three drugs.
`† Low-dose aspirin and high-dose aspirin on prescription, vitamin K antagonists, or oral corticosteroids, but without other NSAIDs.
`‡ Person-year contributed by 6,071 persons with periods of coated aspirin use only.
`§ Of these, 1,752 persons had used noncoated aspirin only during other periods.
`\ Person-year contributed by 22,678 persons with periods of noncoated aspirin use only.
`# Of these, 1752 persons had used coated aspirin only during other periods.
`
`No. of Users
`27,694
`13,865
`1,979
`3,158
`8,728
`13,829
`2,980
`3,916
`6,923
`6,071§
`22,678#
`
`(%)
`(100)
`(50)
`(7)
`(11)
`(32)
`(50)
`(11)
`(14)
`(25)
`
`Aspirin
`Only
`34,560
`16,915
`1,878
`3,675
`11,362
`17,645
`3,124
`5,001
`9,519
`6,593
`27,966
`
`compared with the number of cases expected on the basis of
`the hospitalization rates of upper GI bleeding of North
`Jutland, and the standardized incidence rate ratio (SIR) was
`calculated as the ratio of the observed to the expected
`number of upper GI bleeding. The SIR is a rate ratio in
`which the incidence rate in the cohort is compared with
`reference rates while adjusting for age, sex, and calendar
`time distribution of person-years in the cohort.
`County-specific incidence rates per 100,000 person-years
`of follow-up for a hospitalization for a first upper GI bleed-
`ing, calculated according to sex, 5-yr age groups, and 1-yr
`calendar periods, were applied to the person-years of obser-
`vation to obtain the number of upper GI bleedings expected
`had the cohort members experienced the same hospitaliza-
`tion rates for this condition as the general population of the
`county. To match the definition of an upper GI bleeding of
`interest that was applied to the study cohort, we calculated
`the population rates of upper GI bleeding after excluding
`from the population the inhabitants of the county who were
`previously hospitalized with GI bleeding, or a medical con-
`dition predisposing to GI bleeding (alcoholism, esophageal
`varices, Mallory-Weiss syndrome, liver cirrhosis, or cancer)
`or who received a prescription for one of the drugs that may
`possibly cause GI bleeding.
`Finally, tests of significance and confidence intervals (CI)
`for the SIR were based on the assumption that the observed
`number of cases in a category followed a Poisson distribu-
`tion.
`
`RESULTS
`
`During the study period, the 27,694 users of low-dose as-
`pirin contributed a total of 42,054 person-years of observa-
`tion for current use and 24,738 person-years after cessation
`
`of low-dose aspirin use (former users) (Table 1). A total of
`85% had received more than one prescription for low-dose
`aspirin. Approximately 82% of these person-years were
`derived from low-dose aspirin use only, 14% from com-
`bined use of low-dose aspirin and other NSAIDs, and 4%
`from time segments of combined use of low-dose aspirin
`and corticosteroids, vitamin K antagonists, or high-dose
`aspirin. About 50% of the users were women.
`In the population of North Jutland during the same period
`of time, we observed a total of 2,475 first time hospitaliza-
`tions with upper GI bleeding, corresponding to an overall
`yearly incidence rate of 126 per 100,000 persons aged $16
`yr (Table 2). Of these, 471 (19%) occurred in individuals
`with a medical condition predisposing to GI bleeding, 308
`(12%) in individuals with current use of low-dose aspirin,
`and 892 (36%) in former users of low-dose aspirin and
`individuals with use of other drugs suspected to induce ulcer
`(namely, high-dose aspirin, other NSAIDs, vitamin K an-
`tagonists, and corticosteroids), leaving a total of 804 (33%)
`cases of GI bleeding in the population of North Jutland,
`1991–1995. This corresponds to an incidence rate of 59.0
`cases of bleeding per 100,000 individuals in the remaining
`population of the county (Table 2).
`The SIR for GI bleeding among current users of low-dose
`aspirin (165 cases among men and 143 among women) was
`3.1 (3.0 in men and 3.2 in women) (Table 3). The overall
`and age-specific risks for bleeding were substantially higher
`during combined use of low-dose aspirin and other NSAIDs
`(overall SIR, 5.6), than those observed during use of low-
`dose aspirin alone (overall SIR, 2.6) and during use of
`low-dose aspirin in combination with other drugs apart from
`NSAID (overall SIR, 4.7) (Table 3). The risk estimates
`among women with combined use of low-dose aspirin and
`other NSAIDs tended to be higher than among men.
`
`Page 3 of 7
`
`Patent Owner Ex. 2011
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`AJG – September, 2000
`
`Risk of Upper GI Bleeding and Low-Dose Aspirin
`
`2221
`
`Table 2. Number (Obs) of First Hospitalizations for Upper Gastrointestinal Bleeding in Defined Segments of the Population of the
`County of North Jutland, Denmark, 1991–1995: Incidence Rate for a Subset of the County Population Not Exposed to Specific Drugs
`and Without Predisposing Conditions
`
`County
`Population
`393,800
`
`Predisposed
`Inhabitants*†
`27,200
`
`Obs
`2,475
`1,154
`211
`164
`779
`1,321
`430
`221
`670
`
`(%)
`(100)
`(100)
`(100)
`(100)
`(100)
`(100)
`(100)
`(100)
`(100)
`
`Obs
`471
`174
`51
`32
`91
`297
`134
`53
`110
`
`(%)
`(19)
`(15)
`(24)
`(20)
`(12)
`(22)
`(31)
`(24)
`(16)
`
`Users of
`Low-Dose
`Aspirin†
`27,700
`
`Obs
`308
`143
`1
`17
`125
`165
`8
`28
`129
`
`(%)
`(12)
`(12)
`(0.4)
`(10)
`(16)
`(12)
`(2)
`(13)
`(19)
`
`County of North Jutland
`Approx. size of
`populations§
`Sex and age (yr)
`Both sexes
`Women (yr)
`16–59
`60–69
`701
`Men (yr)
`16–59
`60–69
`701
`* Category includes individuals with hospitalization for alcoholism, esophageal varices, Mallory-Weiss syndrome, liver cirrhosis, or cancer.
`† The groups overlap partially.
`‡ High-dose aspirin on prescription, other NSAIDs, vitamin K antagonists, oral corticosteroids, and former use of low-dose aspirin.
`§ Population $16 yr of age.
`\ Incidence rate per 100,000 person-years.
`Obs 5 observed.
`
`Users of
`Other
`Specific
`Drugs and
`Former Users
`of Low-Dose
`Aspirin†‡
`158,500
`
`Obs
`892
`514
`79
`65
`370
`378
`103
`57
`218
`
`(%)
`(33)
`(45)
`(37)
`(40)
`(48)
`(29)
`(24)
`(26)
`(33)
`
`Remaining
`Population
`272,700
`
`Obs
`804
`323
`80
`50
`193
`481
`185
`83
`213
`
`(%)
`(33)
`(28)
`(38)
`(30)
`(25)
`(36)
`(43)
`(38)
`(32)
`
`IR\
`59
`48.3
`15.4
`71.4
`247.4
`69.2
`32.6
`124.1
`356.6
`
`Table 4 shows the risk of upper GI bleeding during
`periods with use of low-dose aspirin only, when time seg-
`ments were further subdivided into periods of enteric-coated
`(4,178 person-years accumulated) and non– enteric-coated
`use (18,003 person-years). The risk was similar for non-
`coated aspirin (SIR 5 2.6) and coated tablets (SIR 5 2.6).
`When combined with other NSAIDs, the SIR associated
`with enteric-coated aspirin, based on nine observed cases,
`was 3.7 (95% confidence interval [CI], 1.7–7.0), and the
`corresponding estimated equivalent SIR with noncoated as-
`pirin, based on 71 observed cases, was 6.0 (4.7–7.5) (data
`not shown). The risk was similar for users of tablets of 100
`mg (SIR, 2.6 [1.8 –3.5]) and 150 mg (SIR, 2.6 [2.2–3.0]).
`In the restricted cohort of persons who had never received
`
`a prescription for other NSAID drugs or non-NSAID med-
`ication before receiving the first prescription for low-dose
`aspirin, 137 cases of upper GI bleeding were observed
`during periods of low-dose aspirin use only, which was 2.7
`times more than expected (95% CI, 2.3–3.2). During periods
`of former use of low-dose aspirin among these persons, 45
`cases of upper GI bleeding were seen, yielding a SIR of 1.7
`(95% CI, 1.2–2.3). The risk was increased during the year
`after discontinuation of treatment, when a total of 37 cases
`was observed (SIR, 1.8; 95% CI, 1.3–2.5), whereas the risk
`was quite close to unity beyond the first year (SIR, 1.3; 95%
`CI, 0.6 –2.5).
`The absolute risk of upper GI bleeding in only users of
`low-dose aspirin was 364 per 100,000. Under the assump-
`
`Table 3. Standardized Incidence Ratios (SIRs) of Upper Gastrointestinal Bleeding Among 27,694 Current Users of Low-Dose Aspirin
`in the County of North Jutland, Denmark, 1991–1995
`
`Total, Any Current Use
`
`Low-Dose Aspirin
`
`Aspirin Only
`
`Aspirin and Other
`NSAIDs
`
`Aspirin and Other Drugs
`
`SIR
`3.1
`3.2
`1.3
`5.0
`3.1
`3.0
`3.1
`3.6
`2.9
`
`95% CI
`2.8–3.4
`2.7–3.7
`0.0–7.0
`2.9–8.1
`2.5–3.6
`2.6–3.5
`1.3–6.0
`2.4–5.2
`2.4–3.5
`
`Obs
`207
`82
`1
`7
`74
`125
`7
`24
`94
`
`SIR
`2.6
`2.3
`1.5
`2.5
`2.3
`2.8
`3.1
`3.6
`2.6
`
`95% CI
`2.2–2.9
`1.8–2.9
`0.0–8.5
`1.0–5.2
`1.8–2.9
`2.3–3.3
`1.3–6.4
`2.3–5.3
`2.1–3.2
`
`Obs
`80
`52
`0
`8
`44
`28
`1
`2
`25
`
`SIR
`5.6
`6.9
`
`17.9
`6.3
`4.1
`3.5
`2.6
`4.4
`
`95% CI
`4.4–7.0
`5.2–9.1
`
`7.7–35.2
`4.6–8.5
`2.7–6.0
`0.0–19.6
`0.3–9.5
`2.8–6.4
`
`Obs
`21
`9
`0
`2
`7
`12
`0
`2
`10
`
`SIR
`4.7
`4.6
`
`15.0
`3.9
`4.9
`
`7.0
`4.8
`
`95% CI
`2.9–7.2
`2.1–8.7
`
`1.7–53.1
`1.5–8.0
`2.5–8.5
`
`0.8–25.1
`2.3–8.8
`
`Sex and
`Age (yr) of
`Cohort
`Exp
`Obs
`Members
`99.8
`308
`Both sexes
`45.1
`143
`Women (yr)
`0.8
`1
`16–59
`3.4
`17
`60–69
`701
`41.0
`125
`54.7
`165
`Men (yr)
`2.6
`8
`16–59
`7.8
`28
`60–69
`701
`44.3
`129
`Obs 5 observed; Exp 5 expected.
`
`Page 4 of 7
`
`Patent Owner Ex. 2011
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`2222
`
`Sørensen et al.
`
`AJG – Vol. 95, No. 9, 2000
`
`Table 4. Standardized Incidence Ratios (SIRs) of Upper Gastrointestinal Bleeding Among Cohort Members During Periods With Use
`of Low-Dose Aspirin Only, Stratified by Type of Tablet (Coated vs Noncoated)
`Type of Tablet
`Obs
`Exp
`Coated aspirin
`Both sexes
`Women
`Men
`Noncoated aspirin
`Both sexes
`Women
`Men
`Obs 5 observed; Exp 5 expected.
`
`38
`15
`23
`
`169
`67
`102
`
`14.9
`6.5
`8.4
`
`66.2
`29.1
`37.0
`
`SIR
`
`2.6
`2.3
`2.8
`
`2.6
`2.3
`2.8
`
`95% CI
`
`1.8–3.5
`1.3–3.8
`1.7–4.1
`
`2.2–3.0
`1.8–2.9
`2.2–3.3
`
`tion that the association between low-dose aspirin and upper
`GI bleeding is causal, low-dose aspirin was responsible for
`an extra 153.2 upper GI bleedings in the exposed popula-
`tion, contributing 42,054 person-years (current users of low-
`dose aspirin). This means that 6.2% of all GI bleedings (n 5
`2,475) occurring in North Jutland county over the study
`period were due to the use of low-dose aspirin.
`
`DISCUSSION
`
`Exposure to low-dose aspirin was associated with an in-
`creased risk of admission to the hospital for upper GI
`bleeding, and low-dose aspirin seems to account for a non-
`negligible proportion of all upper GI bleeding in the popu-
`lation over the study period. Although the relative risk
`associated with low-dose aspirin did not vary by age, the
`substantially higher underlying rates of GI bleeding among
`the elderly resulted in nearly all aspirin-associated bleeding
`occurring after age 60 yr. The risk was dose-independent
`and increased further when aspirin use was combined with
`exposure to other NSAIDs.
`The study is consistent with three recent reports from
`case-control studies in which odds ratios were raised for all
`dose levels of aspirin, with odds ratios for low-dose aspirin
`varying between 2.2 and 3.4 (10 –12). By contrast, random-
`ized trials of low-dose aspirin have given mixed results (5,
`7–9). Some studies did not report episodes or differences in
`GI bleeding, whereas others found more episodes in persons
`exposed to prophylactic aspirin than to placebo, although
`differences in definition and detection of GI bleeding hinder
`comparison of the trial results. Randomized trials and case-
`control studies have strengths, but also limitations. Random-
`ized trials may provide unbiased estimates, but the restric-
`tion criteria of
`the study populations may limit
`the
`generalizability (14 –16) and have so far included small
`numbers of events of GI bleeding. Case-control studies may
`be affected by biased or differential recall or reporting of
`prior analgesic use among case and control subjects.
`Secondary prophylactic use of low-dose aspirin in Den-
`mark during 1991–1995 primarily involved prescriptions of
`100 or 150 mg noncoated tablets for daily consumption.
`Endoscopic studies indicate that enteric-coated aspirin has a
`lower risk of gastric erosion and microbleeding (26 –28).
`
`Although ,20% of users took coated products, our results
`suggest that enteric-coated, low-dose aspirin may have no
`substantially reduced risk of upper GI bleeding, and the risk
`estimates correspond well with those of a recent case-con-
`trol study in the United States (12). Newer low-dose prod-
`ucts typically involve doses ,100 mg in enteric-coated
`form. We could not evaluate the risk of GI bleeding among
`such users, but found no difference in the risk for use of
`100-mg and 150-mg tablets. NSAIDs are among the most
`commonly prescribed drugs and, because of their frequent
`use, they often coincide with low-dose aspirin. The risk of
`combined low-dose aspirin and other types of NSAIDs
`seems to be at the same level as NSAID treatment alone (20,
`29).
`The main strengths of our study are its large size, the
`uniformly organized health care system allowing a popula-
`tion-based design (thus avoiding selection bias introduced
`by differential patient recruitment), our ability to adjust for
`intake of other drugs and conditions predisposing to GI
`bleedings, and the completeness of follow-up. However, our
`study did not allow analyses of the subgroups of patients
`with predisposition conditions.
`The weaknesses include our inability to control for smok-
`ing, alcohol intake, and infection with Helicobacter pylori.
`All nonaspirin NSAIDs are only available by prescription,
`except low-dose ibuprofen, which is obtainable over the
`counter in Denmark. However, regular users of low-dose
`ibuprofen are registered in our database, as they receive a
`50% refund when redeeming a prescription for ibuprofen.
`Users of low-dose aspirin also receive a 50% refund if they
`have a prescription. Therefore, we have probably registered
`most of the patients receiving low-dose aspirin and NSAID
`on a regular basis. We also could not control for over-the-
`counter use of 500-mg aspirin tablets that were not pre-
`scribed by physicians. However, we have no reason to
`believe that over-the-counter use of aspirin or ibuprofen
`occurred more often among cohort members that it did in
`our comparison group of persons not having prescriptions
`for low-dose aspirin or the other drugs suspected of causing
`upper GI bleeding. We assume that most of the uncaptured
`use of aspirin will be outside the low-dose aspirin cohort
`members, as they will get their drug cost refunded. Thus,
`
`Page 5 of 7
`
`Patent Owner Ex. 2011
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`AJG – September, 2000
`
`Risk of Upper GI Bleeding and Low-Dose Aspirin
`
`2223
`
`this bias will give a conservative estimate of the risk of
`upper GI bleeding.
`The risk of GI bleeding was reduced after cessation of use
`of low-dose aspirin, but still increased compared with non-
`use, in particular, within the first year after discontinuation.
`Assignment to different exposure categories was based on
`the date of dispensing the drug at the pharmacy and the
`assumption that exposure extended 90 days from each pre-
`scription, as most packets contain tablets for approximately
`a 3-month period. If the compliance was not complete or the
`duration of use varied, the exposure would be misclassified,
`and some periods previously designated as former use could
`have actually included exposure to low-dose aspirin. We
`relied upon the coding of GI diagnoses by the hospital
`doctors at the time of discharging the patient. The validity of
`the diagnoses has previously been found high (30), although
`a nondifferential outcome misclassification of fatal GI
`events of 20 –30% has been reported in the Saskatchewan
`Hospital Service Plan (31). All of these biases may also tend
`to give conservative estimates of the risk.
`If these findings are confirmed, they will have important
`public health implications for the assessment of the overall
`public health benefits of low-dose aspirin. Further research
`is needed to clarify the extent to which low-dose aspirin may
`increase the risk of GI bleeding, to identify potential inter-
`action with alcohol or other substances, and to contrast the
`risk versus benefit of prophylactic aspirin use.
`
`ACKNOWLEDGMENTS
`
`This work was supported by The Danish Medical Research
`Council (grant no. 9700677). The activities of the Danish
`Epidemiology Science Centre are financed by a grant from
`the Danish National Research Foundation. We thank Lars
`Thomassen for help with the data management.
`
`Reprint requests and correspondence: Henrik Toft Sørensen,
`M.D., Ph.D., Department of Clinical Epidemiology, Vennelyst
`Boulevard 6, DK-8000 Aarhus C, Denmark.
`Received Nov. 29, 1999; accepted Apr. 12, 2000.
`
`REFERENCES
`
`1. Orme M. Aspirin all round? Br Med J 1988;296:307– 8 (edi-
`torial).
`2. Faulkner G, Prichard P, Somerville K, et al. Aspirin and
`bleeding peptic ulcers in the elderly. Br Med J 1988;297:
`1311–3.
`3. Levy M, Miller DR, Kaufman DW, et al. Major upper gas-
`trointestinal tract bleeding. Relation to the use of aspirin and
`other non narcotic analgesics. Arch Intern Med 1988;148:
`281–5.
`4. Hawkey CJ. Review article: Aspirin and gastrointestinal
`bleeding. Aliment Pharmacol Ther 1994;8:141– 6.
`5. Petersen P, Boysen G, Godtfredsen J, et al. Placebo-controlled,
`randomised trial of warfarin and aspirin for prevention of
`thromboembolic complications in chronic atrial fibrillation.
`The Copenhagen AFASAK study. Lancet 1989;i:175–9.
`6. Meade TW, Roderick PJ, Brennan PJ, et al. Extra-cranial
`
`bleeding and other symptoms due to low dose aspirin and low
`intensity oral anticoagulation. Thromb Haemost 1992;68:1– 6.
`7. SALT Collaborative Group. Swedish aspirin low-dose trial
`(SALT) of 75 mg aspirin as secondary prophylaxis after ce-
`rebrovascular ischaemic events. Lancet 1991;338:1345–9.
`8. RISC Group. Risk of myocardial infarction and death during
`treatment with low-dose aspirin and intravenous heparin in
`men with unstable coronary artery disease. Lancet 1990;336:
`827–30.
`9. Shorrock CJ, Langman MJS, Warlow CJ. Risks of upper GI
`bleeding during TIA prophylaxis with aspirin. Gastroenterol-
`ogy 1992;102:A165.
`10. Weil J, Colin-Jones D, Langman M, et al. Prophylactic aspirin
`and risk of peptic ulcer bleeding. Br Med J 1995;310:827–30.
`11. Kaufman DW, Kelly JP, Sheehan JE, et al. Nonsteroidal
`anti-inflammatory drug use in relation to major upper gastro-
`intestinal bleeding. Clin Pharmacol Ther 1993;53:485–94.
`12. Kelly JP, Kaufman DW, Jurgelon JM, et al. Risk of aspirin-
`associated major upper-gastrointestinal bleeding with enteric-
`coated or buffered product. Lancet 1996;348:1413– 6.
`13. Roderick PJ, Wilkes HC, Meade TW. The gastrointestinal
`toxicity of aspirin: An overview of randomized controlled
`trials. Br J Clin Pharmacol 1993;35:219 –26.
`14. Hansson L, Zanchetti A, Carruthers SG, et al. Effects of
`intensive blood pressure lowering and low-dose aspirin in
`patients with hypertension: Principal results of the hyperten-
`sion optimal treatment (HOT) randomised trial. Lancet 1998;
`351:1755– 62.
`15. The Medical Research Council’s General Practice Research
`Framework. Thrombosis prevention trial: Randomised trial of
`low-density oral anticoagulation with warfarin and low-dose
`aspirin in the primary prevention of ischaemic heart disease in
`men at increased risk. Lancet 1998;351:233– 41.
`16. Isles C, Norrie J, Paterson J, et al. Risk of major gastrointes-
`tinal bleeding with aspirin. Lancet 1999;353:148 –9.
`17. Rockall TA, Logan RFA, Devlin HB, et al. Incidence of and
`mortality from acute upper gastrointestinal haemorrhage in the
`United Kingdom. Br Med J 1995;311:222– 6.
`18. Blatchford O, Davidson LA, Murray WR, et al. Acute upper
`gastrointestinal haemorrhage in west of Scotland: Case ascer-
`tainment study. Br Med J 1997;315:510 – 4.
`19. Garcia-Rodriguez LA, Jick H. Risk of upper gastrointestinal
`bleeding and perforation associated with individual non-ste-
`roidal anti-inflammatory drugs. Lancet 1994;343:769 –72.
`20. MacDonald TM, Morant SV, Robinson GC, et al. Association
`of upper gastrointestinal toxicity of non-steroidal anti-inflam-
`matory drugs with continued exposure: Cohort study. Br
`Med J 1997;315:1333–7.
`21. Nielsen GL, Sørensen HT, Zhou W, et al. The pharmacoepi-
`demiologic prescription database of North Jutland—A valid
`tool in pharmacoepidemiological research. Int J Risk Safety
`Med 1997;10:203–5.
`22. Steffensen FH, Kristensen K, Ejlersen E, et al. Major haem-
`orrhagic complications during oral anticoagulant therapy in a
`Danish population-based cohort. J Intern Med 1997;242:497–503.
`23. Capella´ D. Descriptive tools and analysis. In: Dukes MNG, ed.
`Drug utilization studies: Methods and uses. Copenhagen:
`World Health Organization Regional Publications, 1993:
`55–78 (European Series no. 45).
`24. Sørensen HT. Regional administrative health registries as a
`resource in clinical epidemiology. Int J Risk Safety Med
`1997;10:1–22.
`25. Storm HH, Michelsen EV, Clemmensen IH, et al. The Danish
`Cancer Registry—History, content, quality and use. Dan Med
`Bull 1997;44:535–9.
`26. Hoftiezer JW, Silvoso GR, Burks M, et al. Comparison of the
`
`Page 6 of 7
`
`Patent Owner Ex. 2011
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`2224
`
`Sørensen et al.
`
`AJG – Vol. 95, No. 9, 2000
`
`effects of regular and enteric-coated aspirin on gastroduodenal
`mucosa of man. Lancet 1980;ii:609 –12.
`27. Lanza FL, Royer GL Jr, Nelson RS. Endoscopic evaluation of
`the effects of aspirin, buffered aspirin, and enteric-coated
`aspirin on gastric and duodenal mucosa. N Engl J Med 1980;
`303:136 – 8.
`28. Hawthorne AB, Mahida YR, Cole AT, et al. Aspirin-induced
`gastr