Aliment Pharmacol Ther 2002; 16: 1945–1953.
`
`doi:10.1046/j.0269-2813.2002.01355.x
`
`Risk of upper gastrointestinal bleeding in patients taking low-dose
`aspirin for the prevention of cardiovascular diseases
`
`P. SERRANO*, A. LANAS  , M. T. ARROY O  & I. J. FERREIRA*
`*Service of Cardiology and  Service of Digestive Diseases, Hospital Clı´nico Universitario ‘Lozano Blesa’, Zaragoza, Spain
`
`Accepted for publication 13 July 2002
`
`SUMMARY
`
`Background: Most patients with vascular-occlusive
`diseases benefit from low-dose aspirin (75–325 mg ⁄ day).
`However, they have an increased risk of upper gastroin-
`testinal bleeding (UGIB).
`Aims: To analyse the incidence and factors influencing
`the occurrence of UGIB in patients taking low-dose
`aspirin for the prevention of cardiovascular diseases
`outside clinical trials.
`Methods: We studied 903 consecutive patients dis-
`charged on low-dose aspirin from the Cardiology
`Department of a general hospital. Data were collected
`from medical charts and structured telephone inter-
`views.
`
`(4.5%) presented with
`Results: Forty-one patients
`UGIB requiring hospitalization during
`follow-up
`(45 ± 22 months). The incidence of UGIB was uniform
`during follow-up (1.2 UGIB per 100 patient years).
`Multivariate analysis showed that a history of peptic ulcer
`or UGIB [risk ratio: 3.1, 95% CI: (1.5–6.5)] and aspirin
`dose (per 100 mg ⁄ day) [1.8 (1.5–2.9)] was associated
`with higher risk of UGIB. On the other hand, antisecretory
`[0.22 (0.07–0.75)] and nitrovasodilator drugs [0.73
`(0.55–0.96)] were associated with a decreased risk.
`Conclusions: Cardiovascular patients on long-term low-
`dose aspirin have a stable risk of major UGIB, which is
`higher than published controlled clinical trials. Antise-
`cretory and nitrovasodilator drugs protect from UGIB,
`whereas previous peptic ulcer or UGIB and higher doses
`of aspirin increase the risk.
`
`INTRODUCTION
`
`In Western countries, more than 40% of deaths are due
`to cardiovascular diseases. It has been shown that
`antiplatelet therapy with low-dose acetyl-salicylic acid
`(ASA) (75–325 mg ⁄ day) reduces the risk of vascular
`death and the risk of nonfatal myocardial infarction and
`stroke in patients with previous myocardial infarction,
`unstable angina, nonhaemorrhagic stroke or a transient
`ischaemic attack.1–3 Low-dose ASA is also beneficial in a
`much wider range of patients,
`including those with
`peripheral vascular disease, those undergoing coronary
`angioplasty or coronary bypass grafting4 and also
`
`Correspondence to: Dr Angel Lanas, Service of Digestive Diseases, Hospital
`Clı´nico Universitario ‘Lozano Blesa’, 50009 Zaragoza, Spain.
`Tel.: 34 976 761238; fax: 34 976 761 236;
`E-mail: alanas@posta.unizar.es
`
`high-risk populations for the primary prevention of
`cardiovascular events. Recent epidemiological data sug-
`gest that aspirin may also be effective in the prevention of
`different forms of gastrointestinal (GI) cancer.5, 6 This
`wide range of indications results in a progressive increase
`of the prevalence of low-dose ASA users with age.
`It has also been reported, however, that low-dose ASA
`use is associated with a small but significant increase in
`the risk of upper and lower GI bleeding.7–9 Weil et al.8
`found that any dose of aspirin was associated with an
`increased risk of GI bleeding and that the risk of GI
`bleeding due to low-dose aspirin was dose-related (Odd’s
`ratio 2.3 for 75 mg ⁄ day; 3.2 for 150 mg ⁄ day; 3.9 for
`300 mg ⁄ day). Despite different attempts to avoid the
`contact of aspirin with the stomach or duodenum,
`similar Odd’s ratios have been reported for plain,
`enteric-coated or buffered aspirin.10 In a prospective
`
`Ó 2002 Blackwell Science Ltd
`
`1945
`
`Page 1 of 9
`
`Patent Owner Ex. 2024
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`1946
`
`P. SERRANO et al.
`
`case-control study,11 it has also been confirmed that,
`after adjusting for confounding factors, low-dose ASA
`use increased the risk of major upper GI bleeding events
`by a factor of 2.4, which was three times lower than
`that found for common NSAID use.
`Unlike NSAIDs, however, both the absolute frequency
`of major upper GI bleeding events requiring hospital-
`ization, and the risk factors associated with complica-
`tions in patients treated with low-dose aspirin are not
`well-defined. Data on the prevention of low-dose aspirin-
`induced GI damage are also scarce and weak. Based on
`indirect evidence from nonaspirin NSAIDs studies,
`misoprostol and omeprazole would provide protection
`against upper GI damage.12–14 The effect of nitric oxide
`releasing drugs administered either orally or parenter-
`ally on the risk of major gastrointestinal damage
`induced by either aspirin or NSAIDs is still under
`investigation.15, 16
`Although two case-control studies have suggested a
`protective effect of nitrates on the risk of upper
`gastrointestinal bleeding,11, 17 we think it should also
`be demonstrated in a cohort study limited to patients
`with cardiovascular disease outside clinical trials, to
`avoid possible selection biases.
`Cardiovascular or cerebrovascular diseases are very
`common and their frequency increases with age. Most
`of these patients are treated with low-dose aspirin, thus
`increasing their risk of GI bleeding. The objective of this
`study was to assess the risk ⁄ incidence of upper GI
`bleeding in a representative cohort of patients who are
`prone to taking low-dose aspirin on a chronic basis
`(patients with cardiovascular diseases) and to determine
`the predictors (risk factors and prophylaxis factors) for
`upper GI bleeding in this population. This was consid-
`ered of great importance in order to design appropriate
`therapeutic guidelines for patients on long-term use of
`low-dose aspirin.
`
`MATERIALS AND METHODS
`
`The Hospital Clı´nico Universitario ‘Lozano Blesa’ of
`Zaragoza is a General Hospital attending an area of
`259 000 inhabitants, with 99% of
`the population
`within the National Health System. We studied all
`consecutive patients diagnosed with cardiovascular
`diseases who were discharged on low-dose aspirin
`regimens (75–325 mg ⁄ day) from the Service of Cardi-
`ology between November 1992 and June 1996. The
`planned follow-up period of observation was 5 years
`
`following hospital discharge. The study population
`represents the standard patient on low-dose aspirin
`regimens discharged from hospital with heart disease.
`Data were collected between November 1997 and July
`1999 by structured telephone interview using a stan-
`dardized questionnaire carried out by the same cardi-
`ologist, contrasting all the information with the medical
`charts from hospital and out-patients clinics at the same
`time as the telephone call. In this way, data on reason
`for aspirin, other medical history besides upper GI
`bleeding, and other medications were confirmed by
`chart
`review in all cases. Because patients were
`telephoned at home, they were able to collect all their
`medical records and present medications, ask for help
`from their relatives, or be telephoned again for more
`accurate information. In all patients who reported a
`possible bleeding complication,
`the event was also
`confirmed by additional data from the medical records,
`in order
`to investigate the characteristics of
`the
`bleeding.
`to note that all cardiologists and
`It
`is important
`gastroenterologists of the area are members of the
`Hospital Services of Cardiology and Gastroenterology.
`The main outcome of the study was the occurrence of a
`major upper GI bleeding event, defined as the presence
`of melena and ⁄ or haematemesis confirmed by hospital
`staff and requiring hospital admission.
`Interviews
`consisted of questions regarding: affiliation data, past
`medical history, cause for anti-aggregation with aspirin,
`diseases appearing during follow-up (directly question-
`ing on gastrointestinal bleeding), medications during
`follow-up, and likely causes for changes in medication.
`After completion of the follow-up, information regard-
`ing Helicobacter pylori infection in patients was collected.
`Some patients had undergone tests
`for H. pylori
`infection (13C-urea breath test, urease test or histology
`test) before or during follow-up.
`In other patients,
`H. pylori status was obtained by serology test17, 18
`carried out in blood samples, if they were available and
`stored from previous clinical visits and there was
`previous oral patient informed consent.
`The study was approved by the ethics committees of
`our hospital, and all the patients gave oral informed
`consent before the telephone interview.
`
`Sample size estimation
`
`For patients on low-dose aspirin therapy, we considered
`a theoretical incidence of the primary end-point (upper
`
`Ó 2002 Blackwell Science Ltd, Aliment Pharmacol Ther 16, 1945–1953
`
`Page 2 of 9
`
`Patent Owner Ex. 2024
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`GASTROINTESTINAL BLEEDING AND LOW-DOSE ASPIRIN
`
`1947
`
`gastrointestinal bleeding) of 1–3% per patient-years.7–9
`Thus, during a follow-up of 4.5 years we expected an
`event rate between 4 and 12% for average patients.
`Based on our previous data,11 we assumed a proportion
`of patients taking antisecretory drugs of 15% and a
`proportion of patients taking nitric oxide donor drugs of
`30–50%, given that a higher use of nitric oxide donor
`drugs should be expected in patients discharged from
`the cardiology service on low-dose aspirin, most of them
`with ischaemic heart disease. With a two-sided
`a ¼ 0.05 the study was expected to have 80% power
`to detect an overall relative-risk reduction of 50% for
`those on antisecretory drug therapy or any other
`protective factors.11 We planned to follow 900 patients
`for 4.5 years, to have a potential follow-up of 4050
`patient-years at risk.
`
`Statistical analysis
`
`All statistical analyses were performed with SPSS 10.0.
`Statistical significance was considered when P < 0.05,
`and all contrasts were bilateral. Results are expressed as
`mean ± standard deviation. We performed survival
`analysis to detect variables influencing or modifying
`the risk of hospitalization due to an upper gastro-
`intestinal haemorrhage during follow-up.
`Survival statistical analyses were performed as follows.
`The observation periods started in the month of hospital
`discharge from the Cardiology Department in all cases.
`Follow-up finished either the month of gastrointestinal
`bleeding, the month of stopping aspirin, the month of
`stopping nitric oxide donor drugs, or the month of the
`telephone interview (if no gastrointestinal bleeding,
`aspirin or nitrovasodilator withdrawal had previously
`occurred).
`With every possibly influencing variable, we performed
`a bivariate analysis, describing the survival
`function
`(Kaplan–Meier curves) without
`taking into account
`other variables in each analysis, using the Mantel–
`Haenszel (log-rank) or Breslow statistic when appropri-
`ate. In a second step, log minus log plot was used to
`check the proportionality assumption. The plot exhib-
`ited constant differences between strata. Afterwards, a
`multivariate analysis was performed by the Cox pro-
`portional hazard model (Cox regression) to estimate a
`model that adjusted the effect of all influencing variables
`together on the risk of upper gastrointestinal haemor-
`rhage. RR is the relative risk of suffering an upper GI
`haemorrhage in people with and without the risk factor
`
`Ó 2002 Blackwell Science Ltd, Aliment Pharmacol Ther 16, 1945–1953
`
`from the Cox proportional hazards model. All variables
`were codified as dummy variables, except aspirin dose
`and nitrovasodilators. Aspirin dose was codified as
`100 mg ⁄ day ¼ 1, 200 mg ⁄ day ¼ 2, 75 mg ⁄ day ¼ 0.75,
`and so on. The standard dose of nitrovasodilator drugs
`was considered 50 mg ⁄ day of transdermal nitroglycer-
`ine or 50 mg ⁄ day of oral isosorbide mononitrate. The
`standard dose of nitrovasodilator was codified as 1,
`the standard dose ¼ 0.5, double the standard
`half
`dose ¼ 2, and so on.
`
`RESULTS
`
`Description of the study population
`
`From the initial population of 1224 consecutive patients
`discharged from the Cardiology Service on low-dose
`aspirin, a total of 321 were excluded from the analysis:
`86 patients had missing or unreliable information, 153
`had died by the time of the telephone call (none of them
`due to upper GI bleeding), 77 were impossible to contact
`after three telephone calls and five refused to attend the
`telephone interview. The remaining 903 patients with
`complete and reliable data were analysed.
`Table 1 describes the baseline characteristics and past
`medical history of the study population. All the study
`population was
`discharged
`on low-dose
`aspirin
`
`Table 1. Baseline characteristics, tobacco use, alcohol intake and
`past medical history of the population studied
`n ¼ 903 (100%)
`
`n (%)
`
`Male sex
`Age (years ± s.d.)
`Active smoking
`Past-smoking history
`Daily alcohol use
`Hypertension
`Hypercholesterolemia
`Diabetes
`Prior angina
`Prior Q myocardial infarction
`Prior non-Q myocardial infarction
`Prior angioplasty
`Prior stent
`Prior coronary bypass grafting
`Stroke
`Peripheral vascular disease
`Atrial fibrillation
`Prior peptic ulcer disease
`Prior upper GI bleeding (peptic lesions)
`Rheumatic disease
`
`667 (73.9)
`65 ± 12
`82 (9.1)
`405 (44.9)
`273 (30.2)
`358 (39.6)
`302 (33.4)
`169 (18.7)
`473 (52.4)
`419 (46.4)
`50 (5.5)
`138 (15.3)
`58 (6.4)
`53 (5.9)
`57 (6.3)
`55 (6.1)
`141 (15.6)
`128 (14.2)
`46 (5.1)
`109 (12.1)
`
`Page 3 of 9
`
`Patent Owner Ex. 2024
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`1948
`
`P. SERRANO et al.
`
`likely indication for anti-
`treatment, and the most
`aggregation was secondary prevention of
`ischaemic
`coronary disease (83.6%, 755 ⁄ 903). Table 2 summar-
`izes all indications for anti-aggregation at discharge and
`Table 3 describes the prevalence of use of cardiovascu-
`lar or gastrointestinal drugs. Of all patients on low-dose
`aspirin, three (0.3%) were on 75 mg ⁄ day, 27 (3%) were
`on 100 mg ⁄ day, 341 (37.8%) were on 125 mg ⁄ day,
`89 (9.9%) were on 150 mg ⁄ day, 416 (46.1%) were on
`200 mg ⁄ day, eight (0.9%) were on 250 mg ⁄ day, and
`19 (2.1%) were on 300 mg ⁄ day. Eighty-five per cent
`(225 ⁄ 265) of transdermal nitroglycerin users were on
`10 mg ⁄ day, and the most common doses of oral
`nitrates were 40 mg ⁄ day (24.3%, 57 ⁄ 235) and
`60 mg ⁄ day (47.7%, 112 ⁄ 235). Of patients taking
`H2-receptor antagonists (86.5%, 109 ⁄ 126 ranitidine
`and 13.5%, 17 ⁄ 126 famotidine), 63.5% (80 ⁄ 126) were
`at standard recommended dose (300 mg ⁄ day for ran-
`itidine and 40 mg ⁄ day for famotidine) and 36.5%
`(46 ⁄ 126) at maintenance dose (half the standard dose).
`However, 95.8% (68 ⁄ 71) of patients on proton pump
`
`Table 2. Indications for anti-aggregation at discharge
`n ¼ 903 (100%)
`
`n (%)
`
`Secondary prevention of ischaemic coronary disease 755 (83.6)
`Primary prevention of ischaemic coronary disease
`47 (5.2)
`Chest pain under study
`23 (2.5)
`Atrial fibrillation
`63 (7)
`Other (dilated cardiomyopathy, atrial flutter,
`15 (1.7)
`syncope…)
`
`Table 3. Prevalence of use of gastrointestinal and cardiovascular
`drugs in the study population
`
`Drugs
`
`NSAIDs
`Antisecretory therapy
`Anti-H2 drugs
`Proton pump inhibitors
`Nitric oxide donor drugs
`Oral nitrates
`Transdermal nitroglycerin
`Calcium channel blockers
`Beta-blockers
`ACE inhibitors
`Diuretics
`Statins
`Amiodarone
`Digoxin
`
`n (%)
`
`19 (2.1)
`197 (21.8)
`126 (13.8)
`71 (7.9)
`500 (55.4)
`235 (26)
`265 (29.4)
`405 (44.9)
`201 (22.3)
`250 (27.7)
`149 (16.5)
`213 (23.6)
`73 (8.1)
`58 (6.4)
`
`inhibitors (92.9%, 66 ⁄ 71 omeprazole and 7.0%, 5 ⁄ 71
`lansoprazole) used
`the
`standard
`dose
`(20 and
`30 mg ⁄ day, respectively) and only a minority took half
`the standard dose (4.2%, 3 ⁄ 71).
`Nine hundred and three patients were followed up for
`a mean of 45 ± 22 months. Forty-one of them (4.5%)
`were hospitalized because of an upper gastrointestinal
`bleeding episode; an incidence of 1.2 upper GI bleedings
`per 100 patient-years, which was uniformly distributed
`during the follow-up period. There were eight additional
`patients who suffered a mild gastrointestinal haemor-
`rhage that did not require hospitalization. Among those
`who had an upper GI haemorrhage, 12 ⁄ 41 (29%) had a
`bleeding gastric ulcer, 10 ⁄ 41 (24%) had a bleeding
`duodenal ulcer, and 19 ⁄ 41 (46%) had acute gastrodu-
`odenal mucosal
`lesions. Figure 1 shows the risk of
`upper GI bleeding requiring hospitalization in the total
`population of the study showing a constant risk in the
`cohort.
`
`Risk factors affecting upper GI bleeding
`in the study population
`
`Of all clinical factors considered (Table 1), the presence
`of a history of peptic ulcer or upper GI bleeding due to
`peptic lesions was the only factor significantly associ-
`ated with the development of an upper GI bleeding
`episode (P ¼ 0.049) (Figure 2). On the other hand,
`antisecretory drugs (P ¼ 0.047) and nitrovasodilator
`drugs (P ¼ 0.046) were found to be associated with a
`decreased risk of serious upper GI bleeding (Figures 3
`and 4). Sex and age (higher vs. lower than 65 years of
`
`Figure 1. Probability of upper GI bleeding requiring hospitaliza-
`tion in the total cohort of patients.
`
`Ó 2002 Blackwell Science Ltd, Aliment Pharmacol Ther 16, 1945–1953
`
`Page 4 of 9
`
`Patent Owner Ex. 2024
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`GASTROINTESTINAL BLEEDING AND LOW-DOSE ASPIRIN
`
`1949
`
`Figure 2. Probability of upper GI bleeding requiring hospitaliza-
`tion in patients with and without a history of peptic ulcer or upper
`GI bleeding (P ¼ 0.049).
`
`Figure 4. Probability of upper GI bleeding requiring hospitaliza-
`tion in nitrovasodilator drug users and nonusers (P ¼ 0.046).
`
`Table 4. Multivariate relative risk of upper GI bleeding requiring
`hospitalization in 903 aspirin users
`n ¼ 903 (100%)
`
`RR (95% CI)
`
`P
`
`History of peptic ulcer or upper GI
`bleeding
`Aspirin dose (per 100 mg ⁄ day)
`Antisecretory therapy
`Nitric oxide donor therapy
`
`3.1 (1.5–6.5)
`
`0.003
`
`1.8 (1.5–2.9)
`0.22 (0.07–0.75)
`0.73 (0.55–0.96)
`
`0.016
`0.015
`0.026
`
`haemorrhage in this population on low-dose aspirin.
`NSAID use was found to be associated with upper GI
`bleeding in this cohort of patients in the univariate
`analysis, but it was not in the multivariate analysis.
`
`Figure 3. Probability of upper GI bleeding requiring hospitaliza-
`tion in antisecretory drug users and nonusers (P ¼ 0.047).
`
`H. pylori infection and antisecretory therapy
`
`age) were not associated with upper GI bleeding in this
`population of low-dose aspirin users. Nineteen patients
`were regular users of nonaspirin NSAID. Four of them
`suffered from upper gastrointestinal bleeding during
`follow-up (P ¼ 0.0005). Over a third of the population
`drank alcohol daily (44.6% of men and 4.4% of
`women). Daily alcohol use was not associated with an
`increased risk of bleeding.
`A multivariate analysis by Cox regression (Table 4)
`showed that aspirin dose and prior peptic ulcer or upper
`GI bleeding were the most
`important
`factors that
`increase the risk of upper GI bleeding. On the other
`hand, antisecretory therapy and nitric oxide donor
`therapy were associated with a lower risk of upper GI
`
`Ó 2002 Blackwell Science Ltd, Aliment Pharmacol Ther 16, 1945–1953
`
`The status of H. pylori infection was determined in 341
`patients of the entire population of the study and was
`positive in 240 patients (70.4%). Of the 174 patients
`who had either a history of peptic ulcer or a history of
`ulcer bleeding at the time of hospital discharge from the
`Cardiology Service, only 20 had undergone a previous
`test for the diagnosis of H. pylori infection and all were
`positive. Of these 20 patients, 11 had received successful
`H. pylori eradication prior to hospital discharge and none
`of these patients developed a bleeding episode. The status
`of H. pylori infection was further determined during or
`after follow-up in another 40 patients, but none of these
`received eradication therapy during the period of
`observation. In all, of 60 patients with a history of
`peptic ulcer, 55 were positive for the infection (91.6%).
`
`Page 5 of 9
`
`Patent Owner Ex. 2024
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`1950
`
`P. SERRANO et al.
`
`Of the 41 patients who developed a major upper GI
`bleeding episode, H. pylori status was determined in 24
`and all (100%) were positive for the infection, but only
`16 (66.6%) received successful eradication therapy.
`Finally, of the 174 patients who had either a history of
`peptic ulcer or a history of ulcer bleeding at the time of
`hospital discharge from the Cardiology Service, 84
`(48.3%) received continuous treatment with antisecre-
`tory therapy (any dose) but 90 (51.7%) did not receive
`any type of prevention therapy. In this subpopulation,
`of
`those who received antisecretory therapy, 2.4%
`(2 ⁄ 84) developed a bleeding episode, which was in
`contrast with the 10% (9 ⁄ 90) observed in the similar
`group of patients who did not receive antisecretory
`therapy.
`
`DISCUSSION
`
`The efficacy of low-dose aspirin in the prevention of
`both cardiovascular and cerebrovascular diseases is
`well-established.1–4 However, low-dose aspirin may be
`toxic to the GI tract. Endoscopically controlled studies
`have shown that common doses of aspirin used in
`vascular occlusive diseases (75–300 mg ⁄ day) induce
`gastric and duodenal mucosal petechia, erosions and
`ulcers, which might be present in more than 60% of
`patients after 1 month of therapy.19–21 These lesions
`are often silent and might disappear with continued use
`of aspirin. Low-dose aspirin use is also associated with
`GI bleeding,8–11 which is a more important side-effect
`because it is a life-threatening complication, especially
`in older patients with concomitant severe diseases.
`Given the prevalence of
`low-dose aspirin use world-
`wide, the relevance of an appropriate assessment of the
`risk of complications in patients taking low-dose aspirin
`is clear. This is of great importance in order to design
`appropriate therapeutic guidelines for this type of
`patient, given that different intervention and epidemi-
`ological studies have identified the presence of cardio-
`vascular disease as an independent risk factor for
`developing an upper GI beeding event in patients taking
`nonsteroidal anti-inflammatory drugs (NSAIDs).11, 12
`Furthermore, low-dose aspirin use seems to increase the
`incidence of major upper GI complications in patients
`taking specific COX-2 inhibitors,22 a combination
`therapy that might be very common in the near future.
`In contrast, however, to the high number of studies on
`the risks and factors affecting the risk of GI complica-
`tions
`in patients
`taking NSAIDs,
`the information
`
`regarding the actual risk and the risk factors associated
`with low-dose aspirin use is scarce and weak.
`In this study we have observed an incidence of upper
`gastrointestinal haemorrhage of 1.2 per 100 patient-
`years associated with low-dose aspirin use in patients
`suffering from cardiovascular diseases, which is a
`higher rate than that reported in another cohort
`study23 or randomized clinical trials.24–26 A recent
`Danish 5-year cohort study23 reported an incidence of
`0.6 UGIB per 100 patient-years, which was lower than
`the rate found in our study. However, the population of
`this study involved an entire country, rather than a
`group of high-risk cardiology patients. The two biggest
`randomized clinical trials on low-dose aspirin are the US
`Physicians’ Health Study25 and the HOT study.26 In the
`first trial, which included people taking a similar dose of
`aspirin and a similar follow-up to that observed in our
`study, the rate of UGIB was 0.7 per 100 patient-years,
`and in the HOT study it was 0.3 per 100 patient-years.
`On the other hand the meta-analyses by Derry and
`Loke24 indicates a rate of 2.47% of bleeding episodes in
`aspirin users for a mean of 28 months of follow-up, but
`it must be observed that 25% of patients included in the
`meta-analysis were on a range of aspirin doses between
`162.5 and 1500 mg ⁄ day, and that most of
`these
`randomized clinical trials did not differentiate between
`upper or lower gastrointestinal bleeding.
`The higher incidence of gastrointestinal bleeding found
`in our study may be explained by several factors. First,
`in the most common clinical setting outside clinical
`trials, the population is not homogeneous and includes
`older patients with more comorbidity and receiving
`more medications than those enrolled in clinical trials.
`Second, in most clinical trials, patients with risk factors
`for gastrointestinal bleeding (e.g. ulcer history) are
`excluded, which is not
`the case in a cohort of
`consecutive patients. Finally, patients discharged from
`a cardiology department may have more comorbidity,
`which may enhance the risk of gastrointestinal bleed-
`ing.11
`An interesting finding of the study was that the rate of
`hospitalization due to major upper gastrointestinal
`bleeding during the entire period of observation was
`constant,
`indicating that
`the risk of complications
`associated with low-dose aspirin use does not decrease
`with time. This finding agrees with previous observa-
`tions reported by Kurata and Abbey in the AMIS study,
`where a constant rate of hospitalization due to peptic
`ulceration was observed over a period of 4 years, with
`
`Ó 2002 Blackwell Science Ltd, Aliment Pharmacol Ther 16, 1945–1953
`
`Page 6 of 9
`
`Patent Owner Ex. 2024
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`GASTROINTESTINAL BLEEDING AND LOW-DOSE ASPIRIN
`
`1951
`
`higher doses of aspirin (1000 mg ⁄ day) prescribed for
`prevention of myocardial infarction 10 years ago.27
`Among the potential
`factors associated with an
`increased risk of upper GI bleeding in cardiovascular
`patients taking low-dose aspirin,
`the history of a
`previous peptic ulcer or upper GI bleeding was the most
`important clinical
`factor. Other risk factors that are
`present in patients taking nonaspirin NSAIDs such as
`age,28 male sex29 and NSAIDs use were not identified in
`this population of low-dose aspirin users. However, it
`must be observed that the frequency of regular NSAID
`use was too small to be detected as a significant factor in
`the multivariate analysis, where the number of poten-
`tial covariates included in the analysis was limited by
`the number of bleeding events.
`This study also confirms that, even at the doses used
`for cardiovascular protection, the higher the dose of
`aspirin,
`the higher the risk of upper GI bleeding,
`indicating that the lowest effective dose should always
`be recommended.
`this study has been the
`An important aspect of
`identification of a number of factors that were associ-
`ated with a decreased risk of bleeding. Among these, the
`most important was the concomitant use of antisecre-
`tory therapy. The use of these drugs reduced the risk of
`upper GI bleeding by 50%, which agrees with a recent
`case-control study in which the use of antisecretory
`drugs was found to be associated with a reduced risk of
`upper GI bleeding in all types of NSAID users.11 This
`may be important
`in order to develop appropriate
`therapeutic strategies for the prevention of upper
`gastrointestinal complications.
`Patients with ischaemic heart disease or heart failure
`are often treated with nitrovasodilators and most of
`them are also under
`low-dose aspirin treatment.
`Although it was hypothesized that nitrovasodilators
`might promote gastrointestinal bleeding by inhibiting
`platelet aggregation,12 in this study the use of nitrova-
`sodilator drugs (nitroglycerin or nitrates) was associated
`with a reduced risk of developing an upper GI bleeding.
`Our results support experimental data indicating that
`nitric oxide releasing drugs reduce the gastric mucosal
`damage induced by NSAIDs.15, 16 In addition, a dose–
`response protective effect was also found for the nitric
`oxide releasing drugs in the prevention of upper GI
`bleeding. This important finding confirms, under a
`different study approach and in the specific population
`of cardiovascular patients, the conclusion of two recent
`case-control studies,11, 17 which suggested that the use
`
`Ó 2002 Blackwell Science Ltd, Aliment Pharmacol Ther 16, 1945–1953
`
`of this drug was associated with a reduction of the risk
`of upper GI bleeding for all types of NSAID users. The
`reduction of risk in this study was, however, lower than
`that found with the use of antisecretory drugs.
`Although the characteristics of the study could not
`define the role of H. pylori infection as a risk factor in the
`development of upper GI bleeding in patients taking
`low-dose aspirin, we collected all the available informa-
`tion in order to provide additional data that could
`illustrate the situation of clinical practice regarding this
`issue. This could be important because Chan et al.30
`reported that H. pylori eradication was as effective as
`omeprazole in the prevention of upper GI bleeding in
`low-dose aspirin users. The main conclusions that could
`be obtained is that H. pylori infection is not considered
`as a potential therapeutic target in the clinical man-
`agement of patients with vascular occlusive diseases
`and that H. pylori infection was present in all patients
`tested who developed a complication event. Although it
`is obvious that H. pylori eradication therapy is not
`indicated in patients taking low-dose aspirin, and
`therefore, at present, there is no indication for testing
`this population, it is also clear that a substantial number
`of patients within this population had a history of peptic
`ulcer that should have been tested and treated. In
`addition, only 50% of patients received some type of
`prevention therapy. Several reasons emerge as potential
`explanations for this, including the general belief among
`doctors who treat patients with vascular occlusive
`diseases
`that
`low-dose aspirin is harmless to the
`digestive tract,31 the still poor expansion of H. pylori
`eradication as the treatment of choice in infected
`patients with ulcer history, and, above all, the absence
`of clear guidelines regarding the risk factors for upper GI
`haemorrhage in patients on long-term low-dose aspirin
`and the best
`therapeutic approach to prevent GI
`complications in this important population of cardio-
`vascular patients, many of whom are elderly and on
`multiple medications.
`Finally, there are several limitations in this study that
`should be taken into consideration when interpreting
`the data. The clinical factors that led some patients to
`receive certain drugs (and others not to receive those
`drugs) or to be hospitalized (e.g. similar disease but
`different basal condition) may influence the outcome,
`despite the use of a multivariate analysis. Also, there
`were patients who were not included in the analyses
`(mainly those who died or were impossible to contact)
`that could have had a different risk and also effect
`
`Page 7 of 9
`
`Patent Owner Ex. 2024
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`1952
`
`P. SERRANO et al.
`
`outcome in some way. In addition, selection biases may
`be present in cohort studies because this type of study is
`different from a randomized controlled trial in order to
`address the questions. Finally,
`the actual rate of
`antisecretory drugs and nitrovasodilators was higher
`than expected, which could be due to either the high
`rate of different drug use (polymedication) or the high
`prevalence of clinically documented ischaemic heart
`disease.
`In conclusion, this study indicates that patients who
`are using low-dose aspirin for
`the prevention of
`cardiovascular diseases have a constant risk of devel-
`oping an upper GI bleeding event of 1.2 per 100 patient-
`years. The study has identified the history of previous
`peptic ulcer or upper GI bleeding events and aspirin dose
`as the main risk factors associated with the increased
`risk of bleeding in this population. On the other hand,
`the concomitant use of antisecretory drugs and nitro-
`vasodilators were associated with a decreased risk of
`bleeding in cardiovascular patients taking low-dose
`aspirin.
`
`ACKNOWLEDGEMENTS
`
`This study was supported by a grant from Schering
`Plough Spain, and the collaboration of
`‘‘Sociedad
`Aragonesa de Patologı´a Digestiva’’. The authors are
`grateful for the collaboration of the Service of Epidemi-
`ology of
`the University of Zaragoza and the Chi-
`Cuadrado group for their inestimable help in the
`statistical analysis of the data.
`
`REFERENCES
`
`1 Antiplatelet Trialist’ Collaboration. Collaborative overview of
`randomised trials of antiplatelet therapy. I. Prevention of
`death, myocardial infarction, and stroke by prolonged anti-
`platelet therapy in various categories of patients. BMJ 1994;
`308: 81–106.
`2 Steering Committee of the Physicians’ Health Study Research
`Group. Final report on the aspirin component of the ongoing
`Physician’s Health Study. N Engl J Med 1989; 323: 1289–98.
`3 The SALT Collaborative Group. Swedish aspirin low-dose trial
`(SALT) of 75 mgr aspirin as secondary prophylaxis after
`cerebrovascular events. Lancet 1991; 338: 1345–53.
`4 Lorenz RL, Schacky CV, Weber M, et al. Improved aortocoro-
`nary bypass patency by low-dose aspirin (100 mgr. daily):
`Effects of platelet aggregation and thromboxane formation.
`Lancet 1984; 1: 1261–4.
`5 Giovannucci E, Rimm EB, Stampfer MJ, Colditz GA, Ascherio
`A, Willett WC. Aspirin use and the risk for colorectal cancer
`
`and adenoma in male health pr