Review article: gastrointestinal bleeding with low-dose aspirin what s the risk?

Transcription

1 Alimentary Pharmacology & Therapeutics Review article: gastrointestinal bleeding with low-dose aspirin what s the risk? L. LAINE Department of Medicine, U.S.C. School of Medicine, Los Angeles, CA, USA Correspondence to: Dr L. Laine, Gastrointestinal and Liver Disease Division, Department of Medicine, U.S.C. School of Medicine, 2025 Zonal Ave, Los Angeles, CA 90033, USA. llaine@usc.edu Publication data Submitted 30 March 2006 First decision 3 April 2006 Resubmitted 28 June 2006 Accepted 2 July 2006 SUMMARY This review examines ulcers and gastrointestinal bleeding with low-dose aspirin, focusing on randomized placebo-controlled trials. The single endoscopic trial assessing ulcers showed no significant difference in 12-week ulcer incidence: 6% of 381 given placebo vs. 7% of 387 given 81 mg enteric-coated aspirin. The relative risk of major gastrointestinal bleeding with low-dose aspirin in a meta-analysis of placebo-controlled trials of vascular protection was 2.07 (95% CI: ). The absolute rate increase with aspirin above placebo was 0.12% per year (95% CI: %) with a number-needed-to-harm of 833 patients (95% CI: ). A metaanalysis of aspirin mg daily reported an odds ratio for any gastrointestinal bleeding of 1.68 (95% CI: ) with an numberneeded-to-harm at 1 year of 247. The relative risk of hospitalization for upper gastrointestinal bleeding with low-dose aspirin in a large Danish cohort study was 2.6 (95% CI: ) with an absolute annual incidence of 0.6%. Factors that may increase the risk of gastrointestinal bleeding include prior history of ulcers or gastrointestinal bleeding, corticosteroid use, anticoagulant therapy and addition of a non-aspirin non-steroidal antiinflammatory drug. When determining whether low-dose aspirin is appropriate for an individual patient, the cardiovascular benefit must be weighed against the potential for clinical events such as gastrointestinal bleeding. Aliment Pharmacol Ther 24, ª 2006 The Author 897 doi: /j x

2 898 L. LAINE INTRODUCTION Low-dose aspirin, commonly defined as mg daily, is used for primary and secondary prevention of cardiovascular events such as myocardial infarction (MI). 1 3 Randomized controlled trials document that patients with prior cardiovascular disease have fewer cardiovascular events and deaths with the use of lowdose aspirin (secondary prophylaxis). Furthermore, patients without known cardiovascular disease but with an elevated risk have fewer cardiac events with the use of low-dose aspirin (primary prevention), although a reduction in mortality has not been documented. 4 Aspirin produces its antithrombotic effect via irreversible acetylation of a serine in cyclooxygenase-1 (COX-1) in platelets. 1 This abolishes production of thromboxane A 2 for the life of the platelet. The dose range of aspirin recommended for primary and secondary prevention of coronary artery disease (including from Food and Drug Administration-approved labelling) is mg daily Daily doses of 75 mg of aspirin provide virtually complete inhibition of COX-1 within a few days, suggesting that doses in this range should provide adequate cardiovascular protection. 1, 2 While there are limited comparative data, a recent meta-analysis suggests that doses above mg do not provide additional benefit for the prevention of cardiovascular events, 3 and the American Heart Association 2002 guidelines for primary prevention recommend doses from 75 to 160 mg. 5 Although aspirin doses below 75 mg also inhibit platelet aggregation, 8 a meta-analysis of randomized trials of aspirin compared with no aspirin showed a non-significant reduction in vascular events with daily aspirin doses <75 mg per day (13%), less than the significant decreases in events with mg daily (32%) or mg per day (26%). 3 The authors of this study concluded that the effectiveness of aspirin doses below 75 mg per day is therefore uncertain. However, doses down to 50 mg daily are recommended for the prevention of ischaemic stroke and transient ischaemic attack based on studies of 10, 11 secondary prevention in patients with a history of 12, 13 stroke or transient ischaemic attack. The main factor limiting the use of aspirin is concern about the development of gastrointestinal (GI) events, especially GI bleeding. Many early studies evaluated full-dose aspirin taken as an anti-inflammatory and analgesic medication, and documented GI tract damage comparable with standard non-steroidal anti-inflammatory drugs (NSAIDs). However, damage with NSAIDs, including analgesic doses of aspirin, is dose-related. This narrative review assesses the effect of low-dose aspirin (defined as mg daily), relying primarily on randomized, placebo-controlled trials. The development of ulcers in endoscopic trials is examined, but the major focus will be the most clinically important GI event that occurs with lowdose aspirin, bleeding. A bibliographic data base (MEDLINE via OVID) search from 1966 to 2005 using the term aspirin was conducted to identify randomized, placebo-controlled trials of aspirin in doses of mg daily that provided information on the development of endoscopic ulcers or on upper GI bleeding, and systematic reviews of GI bleeding in randomized controlled trials of aspirin were examined. In addition, to further quantify the risk of bleeding and the risk factors for bleeding with lowdose aspirin, three searches of the terms aspirin combined with gastrointestinal hemorrhage were performed: limited to clinical studies, combined with the term cohort study, and combined with the term case control study. ENDOSCOPIC ULCERS WITH LOW-DOSE ASPIRIN Surprisingly, few data from randomized placebo-controlled trials are available to define the risk of upper GI tract ulcers with low-dose aspirin. Several small trials, usually studying volunteers, have been performed, using a mucosal injury grading system (e.g. grading subepithelial haemorrhage and erosions) as an outcome measure The clinical utility of such scales is uncertain and these studies did not provide specific information on the development of ulcers. Although these studies of low-dose aspirin do not provide completely consistent results, taken together, they suggest that lower doses, such as mg result in a smaller increase in GI mucosal damage than higher doses, such as mg. Furthermore, enteric coating likely decreases acute topical mucosal injury as compared with plain or buffered aspirin. Recently, a large 12-week endoscopic double-blind study in osteoarthritis patients randomly assigned to placebo (n ¼ 381) or 81 mg enteric-coated aspirin (n ¼ 387) revealed no significant difference in the rate

3 REVIEW: GI BLEEDING WITH LOW-DOSE ASPIRIN 899 of ulcers (6% vs. 7%; 95% CI of difference ¼ )4.4% to 6.4%). 18 However, significantly greater increases in the mean number of erosions (mean change: 0.2 vs. 0.9) and in the proportion of patients with an increase in erosions (20% vs. 32%) were seen with low-dose aspirin. The results of this study raise the possibility that even large, longer-term endoscopic studies are not surrogates for an evaluation of clinical events with low-dose aspirin. Reasons for this are uncertain. First, the increase in the rate of clinical events with lowdose aspirin, as discussed below, is small; if the increase in ulcers is similarly small, these differences would not be demonstrated without extremely large endoscopic studies. It may also be that aspirin causes only a very slight increase in ulcers, but the proportion that becomes complicated is much higher due to aspirin s antiplatelet effect. The increased risk of erosions does indicate that some damage to the upper GI tract is occurring. Alternatively, low-dose aspirin might cause complications of pre-existing lesions that occur unrelated to aspirin (e.g. ulcer and vascular ectasia). In any case, the real issue in clinical practice is determining the incidence of clinical GI events such as GI bleeding associated with the long-term use of low-dose aspirin. GASTROINTESTINAL BLEEDING WITH LOW- DOSE ASPIRIN Gastrointestinal bleeding in randomized controlled trials of low-dose aspirin for cardiovascular prevention Several systematic reviews and meta-analyses provide summaries of GI bleeding in randomized controlled trials of low-dose aspirin. 4, However, the clinical utility of some of these reviews may be limited because they include studies with aspirin doses outside the low-dose range ( mg daily), they do not factor out concurrent medications such as anticoagulants in all included studies, and/or they include very short-term studies which may not allow appropriate assessment of relatively rare GI clinical events. Furthermore, some previous systematic reviews include studies for only a single aspirin indication, e.g. primary prevention 4 and secondary prevention. 21 Finally, many reviews provide only relative risks or odds ratios (ORs), while the more clinically relevant measures such as absolute incidences and numbers-needed-to-harm are not available. The number-needed-to-harm (NNH) is a clinically useful result: the number of patients who would need to be treated with low-dose aspirin instead of no therapy (placebo) to cause one additional episode of GI bleeding. Fourteen randomized placebo-controlled trials including over patients present data on major GI bleeding with aspirin dosed at mg daily for cardiovascular indications Major bleeding is based on definitions in individual studies or includes GI bleeding episodes that were fatal or required hospitalization or transfusion. The mean or median duration of 12 studies comprising 97.5% of the patients in the 14 studies ranged from 12 to 76 months, while the other 2.5% of the patients were from two studies with 3 months median duration. Characteristics and results from these 14 trials are shown in Table 1 and the annualized incidences of major GI bleeding for the 12 trials at least 1 year in duration are shown in Table 2. A meta-analysis of these 14 trials has recently been published. 20 The pooled incidence of major GI bleeding in the placebo controlled population of these 14 trials was 0.12% per year. The relative risk of major GI bleeding with the use of low-dose aspirin as compared with placebo was 2.07 (95% CI: ). The appropriate method to determine the increase in risk that is attributable to the medication is to use the placebo rate and the relative risk increase from the meta-analysis to determine the absolute risk increase (placebo rate relative risk increase). These results translate into an absolute rate increase with aspirin above placebo (the incidence of cases of major GI bleeding attributable to low-dose aspirin) of 0.12% per year (95% CI: % per year). 20 Based on this value, 833 patients (95% CI: patients) would need to be treated with low-dose aspirin instead of placebo to cause one major GI bleeding episode during a 1-year period (i.e. the NNH is 833). 20 It should be noted that the annual incidence calculated via this method [0.12% (placebo) % (due to aspirin) ¼ 0.24%] is different from the incidence obtained when the weighted results from the individual studies are merely added together (0.41%). A frequently quoted meta-analysis by Derry and Loke 19 analysed randomized controlled trials of aspirin used as an antiplatelet agent. Twenty-four doubleblind, placebo-controlled studies were included: one used 50 mg daily, 12 employed mg daily and 11 employed doses of mg daily. Their end-

4 900 L. LAINE Table 1. Randomized, placebo-controlled trials of low-dose aspirin ( mg daily dose) for vascular indications that provide information on major gastrointestinal (GI) bleeding outcomes Study Population Aspirin dose (mg) Doubleblind Major GI bleeding definition Mean or median follow-up (months) GI bleeding Aspirin (%) Placebo (%) Thrombosis Men with increased prevention 23 cardiac risk 75 Yes Fatal/life-threatening that required transfusion or surgery 77 6/1268 (0.47) 2/1272 (0.16) HOT 24 Hypertension 75 Yes Major (not defined) 46 77/9399 (0.82) 37/9391 (0.39) 50 11/1009 (1.09) 6/1026 (0.58) SAPAT 25 Chronic stable angina 75 Yes Transfusion, death or serious implications SALT years old with TIA, minor ischaemic stroke, retinal artery occlusion EAFT 27 TIA or minor ischaemic stroke and atrial fibrillation PPP 28 One or more cardiac risk factors 75 Yes Severe or causing discontinuation of study drug 300 Yes Fatal or requiring hospitalization, transfusion or surgery 29 9/676 (1.33) 4/684 (0.58) 28 2/404 (0.50) 1/378 (0.26) 100 No Clinically severe 43 17/2226 (0.76) 5/2269 (0.22) APRICOT 29 Post-thrombolytic therapy 325 Yes Not defined 3 0/102 0/90 Silagy et al years old or older 100 Yes Required hospitalization 12 1/200 (0.50) 0/200 without cardiovascular disease Physicians Health 31 Physicians with no cardiovascular 325 every Yes Required transfusion 60 48/ (0.43) 28/ (0.25) disease other day UK-TIA 32, 33 TIA or minor ischaemic stroke 300 Yes Required hospitalization 48 10/810 (1.23) 2/816 (0.25) 300 No Not defined 12 0/615 0/624 Elwood et al. 34 Less than 65 years old with myocardial infarction Gavaghan et al years old or less with coronary artery bypass graft surgery 324 Yes Active ulcer bleeding 12 2/127 (1.57) 0/110 Lewis et al. 36 Unstable angina 324 Yes Discontinued study drug due to GI bleeding ACBS 37 Carotid stenosis 50% 325 Yes Required hospitalization with transfusion 3 4/625 (0.64) 3/641 (0.47) 23 1/188 (0.53) 1/184 (0.54)

5 REVIEW: GI BLEEDING WITH LOW-DOSE ASPIRIN 901 Table 2. Annualized incidence of major gastrointestinal (GI) bleeding in randomized, placebo-controlled trials of low-dose aspirin ( mg daily dose) of at least 1-year duration that provide information on major GI bleeding outcomes Study Aspirin (%) Placebo (%) Thrombosis prevention HOT SAPAT SALT EAFT PPP Silagy et al Physicians Health UK-TIA 32, Elwood et al Gavaghan et al ACBS point was all GI bleeding without further definition or assessment of major vs. minor bleeding episodes. The pooled OR for bleeding with aspirin was 1.68 (95% CI: ). The number of patients who need to be treated with low-dose aspirin instead of placebo to cause one additional GI bleeding episode over 1 year (NNH) was reported to be 247 (106 patients over average of 28 months). The inclusion of a large number of trials with aspirin doses above the accepted cardioprotective doses and the inclusion of any report of GI bleeding may increase the reported incidence and decrease the NNH in this report. In addition, the authors use of OR rather than relative risk to calculate the NNH results in a small underestimate. 38 When only trials that used lower doses of mg daily were included, the relative risk increase in GI bleeding was similar to the overall results [OR ¼ 1.59 (95% CI: )], but the NNH for this group was not provided. A very large randomized trial that used aspirin in a dose below 75 mg daily and provided information on major GI bleeding was recently published. The Women s Health Study was a placebo-controlled primary cardiovascular prevention trial in over women followed for a mean of 10.1 years. 39 Aspirin was given in a non-standard dose of 100 mg every other day. This trial reported their data on transfusionrequiring GI bleeding. Among patients receiving aspirin, the relative risk of transfusion-requiring GI bleeding vs. placebo was 1.4 (95% CI: ) and the annualized increase in absolute incidence attributable to aspirin was 0.018%, much lower than the 0.12% rate from the meta-analysis cited above. Potential explanations for the much lower absolute risk of major GI bleeding in the Women s Health Study include the very low dose of aspirin, the use of aspirin every other day, the relatively young age of the study population (90% were between 45 and 64 years of age) and the good baseline health status of the study subjects (professional women without a history of cardiovascular disease). Gastrointestinal bleeding with low-dose aspirin in observational studies A potential limitation of randomized controlled trials of low-dose aspirin is that most studies had exclusion criteria that included ulcer disease or intolerance or contraindications to aspirin. Therefore, the populations in these trials may have been at lower risk than individuals in the general population taking low-dose aspirin because the higher-risk patients were excluded by the investigators. Another method to assess the incidence of aspirin-induced GI bleeding in clinical practice is the use of observational cohort studies. Advantages of this type of study are large numbers of patients and longer follow-up in a real-world setting. Disadvantages include the variability in definition and reporting of clinical events, the lack of blinding, and the lack of adjudication of clinical events. Most importantly, comparison of non-randomized study groups can yield biased results due to unequal distribution of confounding factors (e.g. concurrent medications) that influence the development of bleeding in the aspirin and non-aspirin cohorts. A large 5-year observational cohort study from Denmark provided information on major GI bleeding with low-dose aspirin. 40 This study assessed hospitalization for GI bleeding in individuals using mg aspirin when compared with the overall county population of people. The rate of hospitalization for upper GI bleeding in the overall group of patients without use of aspirin, other NSAIDs, anticoagulants, or corticosteroids and without conditions predisposing to GI bleeding was 0.06% per year. The annual incidence of hospitalization for upper GI bleeding was 0.60% for individuals taking aspirin alone. The adjusted relative risk of hospitalization for upper GI bleeding with low-dose aspirin alone vs. an ageand gender-matched population not taking aspirin, other NSAIDs, anticoagulants or corticosteroids and

6 902 L. LAINE without conditions predisposing to GI bleeding was 2.6 (95% CI: ) and not 10 as suggested by looking at the unadjusted data (0.6% vs. 0.06%). These results are consistent with a 0.23% background incidence of upper GI bleeding hospitalization in the group matched to the low-dose aspirin users a value nearly four times the baseline rate of 0.06% noted for the general population without medications or conditions predisposing to bleeding suggesting that the group taking low-dose aspirin was a higher-risk population. No difference in GI bleeding rates was noted between enteric coated and uncoated aspirin, a finding similar to other studies of low-dose aspirin and GI 25, 26 bleeding. Case control studies generally report a significant increase in the odds of low-dose aspirin use in patients with GI bleeding (cases) compared with the odds of low-dose aspirin use in patients without GI bleeding (controls). Odds ratios are in the range of Summary: risk of GI bleeding with low-dose aspirin in clinical trials In summary, meta-analyses of randomized, placebocontrolled trials of low-dose aspirin indicates that aspirin approximately doubles the risk of major GI bleeding compared with placebo. Observational studies provide similar results, generally suggesting a relative risk increase of about 2 to 2 1 2, with a range of approximately to 3. The annualized incidence of major GI bleeding attributable to low-dose aspirin in placebocontrolled trials (i.e. the increase above the bleeding incidence due to placebo) is 0.12% with a NNH of 833; an NNH of 247 was reported with daily doses of aspirin up to 1500 mg and any GI bleeding as an endpoint. The annual incidence of major GI bleeding with low-dose aspirin was somewhat higher in an observational cohort study. The slightly higher rate of major GI bleeding with low-dose aspirin seen in the cohort study compared with the pooled randomized controlled trials may relate to the fact that the randomized trials exclude patients at risk for GI bleeding, while aspirin users in a population-based cohort study may be at higher GI risk. Risk factors for upper GI bleeding A large number of studies have evaluated risk factors for upper GI bleeding with non-aspirin NSAIDs. The characteristics most consistently identified are: (i) prior GI events; (ii) older age; (iii) use of anticoagulants such as warfarin; (iv) use of corticosteroids; and (v) increasing dose or multiple NSAIDs (e.g. NSAID plus aspirin). 49 However, data evaluating these risk factors specifically in low-dose aspirin users are much more limited, and few comes from analyses of randomized controlled trials. Prior upper GI events (e.g. bleeding) A case control study of patients hospitalized with upper GI bleeding identified a prior history of upper GI bleeding [OR ¼ 6.5 (95% CI: )] and a prior history of ulcer [OR ¼ 2.0 ( )] as significant risk factors for hospitalization with upper GI bleeding in patients taking low-dose aspirin. 44 An assessment of 903 patients discharged from a cardiology service on mg of aspirin daily and followed for a mean of 45 months noted a significant association between hospitalization for upper GI bleeding and a history of peptic ulcer or upper GI bleeding (risk ratio ¼ 3.1, 95% CI: ). 50 Age Surprisingly, available data do not support increasing age as a significant risk factor for low-dose aspirininduced bleeding. A large cohort study from Denmark 24 reported that standardized incidence ratios for upper GI bleeding in individuals taking low-dose aspirin without other NSAIDs, corticosteroids or warfarin were 3.1 (95% CI: ) for men years old vs. 2.6 ( ) for men 70 years old. Values for women were 1.5 ( ) at years and 2.3 ( ) for women 70 years old. Multivariable analyses in case control 44 and cohort 50 studies have not identified age as a significant predictor of hospitalization for upper GI bleeding among patients taking lowdose aspirin. Nevertheless, older patients taking low-dose aspirin do have a much greater absolute risk of GI complications than younger patients because older patients have a much higher baseline risk of GI complications. For example, Patrono et al. 51 indicated that the risk of ulcer complications in control subjects under 50 years old was below 0.5% while the risk was nearly 4% in controls aged years and approximately 6% in controls over 80 years of age. Even though an approximately twofold increase in risk with low-dose

7 REVIEW: GI BLEEDING WITH LOW-DOSE ASPIRIN 903 aspirin is consistent across the different age groups, the absolute incidence of complications and the absolute increase in complications with aspirin vs. controls will be dramatically higher in the older population due to their higher baseline risk. Furthermore, concurrent use of other medications (e.g. NSAIDs) that increase the risk of bleeding in low-dose aspirin users also increases with age. Concomitant warfarin use Four randomized controlled trials provide information on major GI bleeding in low-dose aspirin users with and without warfarin (Table 3). The Thrombosis Prevention Trial 23 randomly assigned men to four treatment groups in a double-blind fashion: warfarin only, aspirin only, warfarin plus aspirin or placebo. Major GI bleeding was seen slightly more often in the warfarin plus aspirin group than in the aspirin alone group [9/1277 (0.7%) vs. 6/1268 (0.5%)] but the difference was not significant (95% CI of difference ¼ )0.4% to 0.8%; P ¼ 0.6). Another randomized trial compared aspirin 160 mg daily, warfarin and aspirin 75 mg daily plus warfarin in 3630 patients after MI: 52 major GI bleeding occurred in 0.5% of those taking aspirin alone, 1.5% of those taking warfarin alone and 1.7% of those taking aspirin plus warfarin (95% CI of difference vs. aspirin alone ¼ %; P ¼ 0.006). A third trial, in vascular surgery patients, also noted a greater incidence of major GI bleeding in the combination group [4.1% vs. 1.7% (95% CI of difference ¼ %; P ¼ 0.06)], 53 while a fourth randomized trial in patients with atrial fibrillation noted no suggestion of more frequent major GI bleeding with aspirin plus coumadin vs. low-dose aspirin alone. 54 In summary, concurrent anticoagulation therapy with warfarin does appear likely to increase the risk of GI bleeding among patients taking low-dose aspirin. Concomitant corticosteroid use A cohort study from Denmark examined corticosteroid and low-dose aspirin ( mg) use in relation to hospitalization for upper GI bleeding. 55 The relative risk with both drugs combined was 5.3 ( ). This study did not provide a comparison with a group using low-dose aspirin alone, but a study of the same population during the same time period revealed a relative risk of hospitalization for upper GI bleeding with low-dose aspirin alone of 2.6 ( ). 40 These data Table 3. Randomized, controlled trials of low-dose aspirin ( mg daily dose) vs. warfarin plus low-dose aspirin that provide information on major GI bleeding outcomes GI bleeding Aspirin (%) Combination (%) Mean or median follow-up (months) Warfarin dose Major GI bleeding definition Study Population Aspirin dose 77 6/1268 (0.57) 9/1277 (0.7) 75 mg INR ¼ 1.5 Fatal/life-threatening that required transfusion or surgery Thrombosis Men with increased prevention 23 cardiac risk 48 6/1206 (0.5) 21/1208 (1.7) INR ¼ Non-fatal bleeding necessitating surgery or transfusion 160 mg alone; 75 mg with warfarin Hurlen et al. 52 Hospitalized with acute myocardial infarction 325 mg INR ¼ Transfusion-requiring 72 7/413 (1.7) 17/418 (4.1) Johnson et al. 53 Lower extremity arterial bypass surgery 41 4/169 (2.4) 1/171 (0.6) AFASAK2 54 Atrial fibrillation 300 mg 1.25 mg Fatal, life-threatening, potentially life-threatening GI, gastrointestinal; INR, international normalized ratio.

8 904 L. LAINE therefore suggest that corticosteroids may increase the risk of upper GI bleeding when added to low-dose aspirin. Increasing dose As mentioned above, in many populations, the use of lower low-dose aspirin ( or 160 mg daily) is recommended. Increasing the dose of aspirin above the cardioprotective dose (325 mg) increases the risk of GI bleeding compared with low-dose aspirin. 48 However, the important question related to low-dose aspirin is whether increasing the dose from 75 or 81 mg to 325 mg also increases GI risk. A meta-analysis of randomized controlled trials of aspirin used as an antiplatelet agent did not find a significant increase in GI bleeding with increasing dose. 19 Another metaanalysis found that that the relative risk of major GI bleeding with lower low-dose aspirin ( mg daily) was similar to the relative risk with higher low-dose aspirin (> mg daily): 2.22 (95% CI: ) vs (95% CI: ). 20 However, since these meta-analysis results are not head-to-head comparisons of higher vs. lower doses they do not directly address the question of whether the higher dose (e.g. 325 mg) increases the risk for GI bleeding more than the lower dose (e.g. 75 or 81 mg). Two randomized controlled trials have performed a head-to-head comparison of lower dose (81 mg) and higher dose (325 mg) low-dose aspirin. The first trial was a 3-month study in 1417 patients after carotid endarterectomy. 56 Rates of GI bleeding were comparable, with 1.1% developing bleeding in each treatment group. The second trial was a double-blind colon neoplasm chemoprevention study with 33 months mean follow-up. 57 The proportions with GI bleeding were not significantly different between the 81 mg group [2/377 (0.5%)] and the 325 mg group [4/372 (1.1%); 95% CI of the difference ¼ )0.7% to 1.8%]. Large sample sizes would be required to provide a meaningful comparison of the risks of GI bleeding between lower and higher low-dose aspirin. Epidemiological studies have assessed this question and provide mixed results. Weil et al. 43 in a case control study of patients hospitalized for bleeding ulcer reported adjusted ORs with 75 mg of 2.3 (95% CI: ), with 150 mg of 3.2 ( ) and with 300 mg of 3.9 ( ). No statistical comparison was provided, although the confidence intervals for the 3 groups are widely overlapping. In contrast, de Abajo et al. 48 showed no suggestion of an increase in upper GI complications with increasing doses: OR ¼ 1.9, 2.1 and 1.9 for 75 mg, 150 mg and 300 mg, respectively. Multivariable analysis of a cohort of 903 patients discharged from the cardiology department of a general hospital on mg of aspirin daily did report that dose was an independent predictor of upper GI bleeding [OR (per 100 mg) ¼ 1.8, 95% CI: ]. 50 Low-dose aspirin plus other NSAIDs Several studies document that the addition of a nonaspirin NSAID to low-dose aspirin significantly increases the risk of developing upper GI bleeding by about two- to fourfold. A 5-year cohort study in Denmark 40 showed that the annual incidence of hospital admission for upper GI bleeding was more than doubled when NSAIDs plus low-dose aspirin were compared with low-dose aspirin alone: 1.4% vs. 0.6%. The relative risk for aspirin alone was 2.6 ( ), and it was significantly higher for low-dose aspirin plus NSAID [5.6 ( )]. A case control study of hospitalizations for bleeding ulcer revealed that low-dose aspirin alone was associated with an adjusted OR of 3.3 ( ), while the combination of low-dose aspirin plus an NSAID had an OR of 7.7 ( ). 43 In another case control study 44 of hospitalizations for bleeding peptic lesions, NSAID plus low-dose aspirin use also increased the risk when compared with lowdose aspirin alone [OR ¼ 3.8 ( )]. Choice of non-selective NSAID vs. COX-2 specific drugs with low-dose aspirin An important question in clinical practice is whether prescribing a COX-2 selective inhibitor (coxib) rather than a non-selective NSAID in patients taking lowdose aspirin improves GI safety. In a 12-week double-blind endoscopic trial, the incidence of ulcers among 387 patients taking 81 mg enteric-coated aspirin was 7% when compared with 16% in 377 patients taking 81 mg aspirin plus rofecoxib. 18 The rate of ulcers with ibuprofen 800 mg, three times a day, was similar at 17%. A 1-week doubleblind endoscopic study comparing placebo, celecoxib 200 mg daily, and naproxen 500 mg b.d. in patients taking 325 mg aspirin daily revealed that celecoxib significantly increased the incidence of ulcers vs. placebo (19% vs. 8%), but that naproxen plus aspirin was

9 REVIEW: GI BLEEDING WITH LOW-DOSE ASPIRIN 905 (a) 1.00% 0.80% 0.60% 0.40% 0.20% 0.00% (b) 3.00% 2.00% 1.00% 0.00% RR = 0.21 ( ) RR = 0.79 ( ) No Aspirin Aspirin n = RR = 0.38 ( ) RR = 0.73 ( ) No Aspirin Aspirin n = Lumiracoxib NSAIDs Lumiracoxib NSAIDs Figure 1. (a) Crude incidences of ulcer complications and relative risks (RR; 95% CI) related to use of low-dose aspirin in double-blind gastrointestinal (GI) outcomes trial comparing the COX-2 selective inhibitor lumiracoxib vs. naproxen or ibuprofen in patients with osteoarthritis. 59 (b) Crude incidences of complicated plus symptomatic ulcers and RRs (95% CI) related to use of low-dose aspirin in double-blind GI outcomes trial comparing the COX-2 selective inhibitor lumiracoxib vs. naproxen or ibuprofen in patients with osteoarthritis. 59 significantly worse than celecoxib plus aspirin (27% vs. 19%). 58 A large double-blind outcomes trial (TARGET) comparing a coxib (lumiracoxib) to non-selective NSAIDs (ibuprofen and naproxen) stratified their patients by use of low-dose aspirin, allowing comparison of randomized groups of patients taking combined therapy 59 (Figure 1). Twenty-four per cent of the population (4326 patients) took low-dose aspirin. The relative risk for ulcer complications for the coxib vs. the nonselective NSAIDs was 0.79 ( ) and for complicated plus symptomatic ulcers was 0.73 ( ). In contrast, the risk reduction with the coxib vs. nonselective NSAIDS in patients not taking aspirin was much larger and significant (Figure 1). Interestingly, among non-selective NSAID users, the addition of low-dose aspirin did not lead to a higher incidence of upper GI clinical events or complications compared with non-selective NSAID users not taking low-dose aspirin (Figure 1a,b). This raises the possibility that although addition of a non-selective NSAID to lowdose aspirin increases the risk compared with aspirin alone, addition of low-dose aspirin to a non-selective NSAID may not significantly increase the risk of upper GI complications. However, the comparison of NSAIDs vs. NSAIDs plus aspirin in TARGET is a comparison of non-randomized cohorts, so it can be viewed as only hypothesis-generating. Finally, a large epidemiological cohort study using the Quebec provincial health insurance database assessed the development of upper GI bleeding and perforation related to use of low-dose aspirin, nonselective NSAIDs and coxibs 60 (Table 4). Patients taking coxibs had significantly fewer complications than those using non-selective NSAIDs. When low-dose aspirin was added to coxibs, the rate of complications rose to approach that of the non-selective NSAIDs; however, this combination had significantly fewer complications than non-selective NSAIDs plus lowdose aspirin [hazard ratio 0.6 (95% CI: )]. Thus, endoscopic trials suggest that the GI effect of low-dose aspirin plus coxibs is much more than the additive: the risk of ulcers rises significantly to the level of a non-selective NSAID. 18 However, the combination of low-dose aspirin plus a non-selective NSAID Table 4. Coxibs vs. non-selective NSAIDs with and without low-dose aspirin in elderly patients: Hospitalizations for upper GI bleeding or perforation in a retrospective cohort study of the Quebec Health Insurance Agency Database. 60 Medications Non-selective NSAID only ( prescriptions) Coxib only ( prescriptions) Coxib + aspirin ( prescriptions) Non-selective NSAID + aspirin ( prescriptions) Coxib + aspirin vs. non-selective NSAID + aspirin Hazard ratio (95% CI) 1 (reference) 0.6 ( ) 0.9 ( ) 1.6 ( ) 0.6 ( ) NSAID, non-steroidal anti-inflammatory drugs; GI, gastrointestinal. 60

10 906 L. LAINE appears to increase the risk still further. 58 An epidemiological study assessing GI complications also noted a significantly lower risk with coxibs compared with non-selective NSAIDs in low-dose aspirin users. 60 However, a double-blind outcomes trial showed a non-significant decrease in ulcer complications and ulcer clinical events with a coxib compared with nonselective NSAIDs. 59 This study was not powered to address this question, so it is unclear whether the 21 27% relative risk decrease was due to chance or whether it is real and would be significant in a study of larger sample size. SUMMARY AND CONCLUSIONS Low-dose aspirin, defined as mg daily, offers clearcut benefit in the prevention of cardiovascular events in patients with prior cardiovascular disease and in those whose clinical characteristics predict a moderate-high risk of developing cardiovascular events. The main concern regarding the use of aspirin is the potential for developing serious GI events, such as bleeding. Even at recommended doses, aspirin increases the risk of GI bleeding by approximately twofold. However, the annual incidence of GI bleeding with these doses is relatively low. Pooled estimates from randomized, placebo-controlled trials suggest that the annual incidence of major GI bleeding that can be ascribed to use of low-dose aspirin (i.e. the increment above the placebo or no aspirin group) is 0.12%. 20 This means that 833 patients would need to be treated with low-dose aspirin rather than no aspirin to cause an additional episode of major GI bleeding over 1 year. When the development of any degree of GI bleeding with doses of aspirin up to 1500 mg per day (well above the recommended cardioprotective dose) is assessed, the number needed to harm at 1 year is reported to be The U.S. Preventive Services Task Force concluded that the benefit risk balance with low-dose aspirin is most favourable in patients with a predicted 5-year risk for coronary heart disease 3%. 61 However, their recommendation was based on an assumed annual increased incidence of major GI bleeding of 0.06% from five primary prevention randomized trials a rate somewhat lower than the results from other systematic reviews. Alternatively, the American Heart Association recommends low-dose aspirin be considered in a higher-risk group (10-year risk 10%) to improve the likelihood of a positive balance of benefit vs. risk. 5 The clinical characteristics that portend an increased risk of GI bleeding with low-dose aspirin are not as well defined as they are for full-dose NSAIDs. Limited information does suggest that prior history of ulcers or GI bleeding, the use of corticosteroids or anticoagulant therapy, and the addition of a non-aspirin NSAID all significantly increase the risk. Surprisingly, age has not been documented to increase the risk with lowdose aspirin, in contrast to studies with full-dose NSAIDs. However, older patients taking low-dose aspirin do have a greater risk of GI bleeding due to the marked increase in baseline risk of bleeding with increasing age. The combination of low-dose aspirin plus another NSAID increases the risk of GI bleeding by two- to fourfold. The addition of a COX-2-specific inhibitor also increases the risk of GI injury more than additively to approximately the level of a nonselective NSAID. However, some studies suggest that the combination of a COX-2-specific inhibitor and low-dose aspirin may have lower risk than the combination of a non-selective NSAID and low-dose aspirin. In conclusion, when determining whether low-dose aspirin is appropriate for an individual patient, the risk of clinical events such as GI bleeding must be weighed against the cardiovascular benefit. In cases where the risk of a cardiovascular event is moderate or high based on prior cardiovascular events or an increased cardiovascular risk score, the benefits of treatment are expected to outweigh the risk. 4 In patients with low cardiovascular risk, the use of lowdose aspirin may be anticipated to cause an excess of GI events. ACKNOWLEDGEMENTS The author has received research support from Bayer, a manufacturer of low-dose aspirin; he has also received research support from Merck, Pfizer, and Novartis (makers of coxibs), and he has served as a consultant to these companies in the past.

11 REVIEW: GI BLEEDING WITH LOW-DOSE ASPIRIN 907 REFERENCES 1 Awtry EH, Loscalzo J. Aspirin. Circulation 2000; 101: Antiplatelet Trialists Collaboration. Collaborative overview of randomised trials of antiplatelet therapy-i: prevention of death, myocardial infarction, and stroke by prolonged antiplatelet therapy in various categories of patients. BMJ 1994; 308: Antithrombotic Trialists Collaboration. Collaborative meta-analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. BMJ 2002; 324: Hayden M, Pignone M, Phillips C, Mulrow C. Aspirin for the primary prevention of cardiovascular events: a summary of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med 2002; 136: Pearson TA, Blair SN, Daniels SR, et al. AHA guidelines for primary prevention of cardiovascular disease and stroke: 2002 update. Circulation 2002; 106: Antman EM, Anbe DT, Armstrong PW, et al. ACC/AHA guidelines for the management of patients with ST elevation myocardial infarction executive summary. Circulation 2004; 110: Harrington RA, Becker RC, Ezekowitz M, et al. Antithrombotic therapy for coronary artery disease: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest 2004; 126: 513S 48S. 8 Patrono C, Coller B, FitzGerald GA, et al. Platelet-active drugs: the relationship among dose, effectiveness, and side effects: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest 2004; 126: 234S 64S. 9 Saseen JJ. ASHP therapeutic statement on the daily use of aspirin for preventing cardiovascular events. Am J Health Syst Pharm 2005; 62: Internal analgesic, antipyretic, and antirheumatic drug products for over-thecounter human use. Code Fed Regul 2004; 5: [21CFR343]. 11 Albers GW, Amarenco P, Easton JD, et al. Antithrombotic and thrombolytic therapy for ischemic stroke: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest 2004; 126: 483S 512S. 12 Diener HC, Cunha L, Forbes CJS, et al. European stroke prevention study 2. Dipyridamole and acetylsalicylic acid in the secondary prevention of stroke. J Neurol Sci 1996; 143: The Dutch TIA Trial Study Group. A comparison of two doses of aspirin (30 mg vs. 283 mg a day) in patients after a transient ischemic attack or minor ischemic stroke. N Engl J Med 1991; 325: Dammann HG, Burkhardt F, Wolf N. Enteric coating of aspirin significantly decreases gastroduodenal mucosal lesions. Aliment Pharmacol Ther 1999; 13: Hawthorne AB, Mahida YR, Cole AT, Hawkey CJ. Aspirin-induced gastric mucosal damage: prevention by entericcoating and relation to prostaglandin synthesis. Br J Clin Pharmacol 1991; 32: Cole AT, Hudson N, Liew LC, Murray FE, Hawkey CJ, Heptinstall S. Protection of human gastric mucosa against aspirin-enteric coating or dose reduction? Aliment Pharmacol Ther 1999; 13: Petroski D. A comparison of entericcoated aspirin granules with plain and buffered aspirin: a report of two studies. Am J Gastroenterol 1986; 81: Laine L, Maller ES, Yu C, Quan H, Simon T. Ulcer formation with low-dose enteric-coated aspirin and the effect of COX-2 selective inhibition: a doubleblind trial. Gastroenterology 2004; 127: Derry S, Loke YK. Risk of gastrointestinal haemorrhage with long term use of aspirin: meta-analysis. BMJ 2000; 321: McQuaid KR, Laine L. Systematic review and meta-analysis of adverse events of low-dose aspirin and clopidogrel in randomized controlled trials. Am J Med; 2006; 119: Weisman SM, Graham DY. Evaluation of the benefits and risks of low-dose aspirin in the secondary prevention of cardiovascular and cerebrovascular events. Arch Intern Med 2002; 162: Serebruany VL, Malinin AI, Eisert RM, Sane DC. Risk of bleeding complications with antiplatelet agents: meta-analysis of 338,191 patients enrolled in 50 randomized controlled trials. Am J Hematol 2004; 75: The Medical Research Council s General Practice Research Framework. Thrombosis prevention trial: randomised trial of low-intensity oral anticoagulation with warfarin and low-dose aspirin in the primary prevention of ischaemic heart disease in men at increased risk. Lancet 1988; 351: 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 hypertension optimal treatment (HOT) randomised trial. HOT Study Group. Lancet 1998; 351: Juul-Möller S, Edvardsson N, Jahnmatz B, Rosen A, Sorensen S, Omblus R. Double-blind trial of aspirin in primary prevention of myocardial infarction in patients with stable chronic angina pectoris. The Swedish Angina Pectoris Aspirin Trial (SAPAT) Group. Lancet 1992; 340: The SALT Collaborative Group. Swedish aspirin low-dose trial (SALT) of 75 mg aspirin as secondary prophylaxis after cerebrovascular ischaemic events. Lancet 1991; 338: EAFT (European Atrial Fibrillation Trial) Study Group. Secondary prevention in non-rheumatic atrial fibrillation after transient ischaemic attack or minor stroke. Lancet 1993; 342: Collaborative Group of the Primary Prevention Project. Low-dose aspirin and vitamin E in people at cardiovascular risk: a randomised trial in general practice. Collaborative Group of the Primary Prevention Project. Lancet 2001; 357: Meijer A, Verheugt FW, Werter CJ, et al. Aspirin versus coumadin in the prevention of reocclusion and recurrent ischemia after successful thrombolysis: a prospective placebo-controlled angiographic study. Results of the APRICOT study. Circulation 1993; 87: Silagy CA, McNeil JJ, Donnan GA, Tonkin AM, Worsam B, Campion K. Adverse effects of low-dose aspirin in a healthy elderly population. Clin Pharmacol Ther 1993; 54: Steering Committee of the Physicians Health Study Research Group. Final Report on the Aspirin Component of the Ongoing Physicians Health Study. N Engl J Med 1989; 321:

12 908 L. LAINE 32 The United Kingdom transient ischaemic attack (UK-TIA) aspirin trial: final results. J Neurol Neurosurg Psychiatr 1991; 54: Slattery J, Warlow CP, Shorrock CJ, Langman MJ. Risks of gastrointestinal bleeding during secondary prevention of vascular events with aspirin analysis of gastrointestinal bleeding during the UK-TIA trial. Gut 1995; 37: Elwood PC, Cochrane AL, Burr ML, et al. A randomized controlled trial of acetyl salicylic acid in the secondary prevention of mortality from myocardial infarction. Br Med J 1974; 1: Gavaghan T, Gebski V, Baron D. Immediate postoperative aspirin improves vein graft patency early and late after coronary artery bypass graft surgery: a placebo-controlled, randomized study. Circulation 1991; 83: Lewis HD, Davis JW, Archibald DG, et al. Protective effects of aspirin against acute myocardial infarction and death in men with unstable angina. Results of a Veterans Administration Cooperative Study. N Engl J Med 1983; 309: Cote R, Battista RN, Abrahamowicz M, et al. Lack of effect of aspirin in asymptomatic patients with carotid bruits and substantial carotid narrowing. The Asymptomatic Cervical Bruit Study Group. Ann Intern Med 1995; 123: Page J, Attia J. Using Bayes nomogram to help interpret odds ratios. ACP J Club 2003; 139: A Ridker PM, Cook NR, Lee IM, et al. A randomized trial of low-dose aspirin in the primary prevention of cardiovascular disease in women. N Engl J Med 2005; 352: Sorensen HT, Mellemkjaer L, Blot WJ, et al. Risk of upper gastrointestinal bleeding associated with use of lowdose aspirin. Am J Gastroenterol 2000; 95: Kelly JP, Kaufman DW, Jurgelon JM, Sheehan J, Koff RS, Shapiro S. Risk of aspirin-associated major upper-gastrointestinal bleeding with enteric-coated or buffered product. Lancet 1996; 348: Garcia Rodriguez LA, Hernandez-Diaz S, de Abajo FJ. Association between aspirin and upper gastrointestinal complications: systematic review of epidemiologic studies. Br J Clin Pharmacol 2001; 52: Weil J, Colin-Jones D, Langman M, et al. Prophylactic aspirin and risk of peptic ulcer bleeding. Br Med J 1995; 310: Lanas A, Bajador E, Serrano P, et al. Nitrovasodilators, low-dose aspirin, other nonsteroidal antiinflammatory drugs, and the risk of upper gastrointestinal bleeding. N Engl J Med 2000; 343: Savage R, Moller P, Ballantyne C, Wells J. Variation in the risk of peptic ulcer complications with nonsteroidal antiinflammatory drug therapy. Arthritis Rheum 1993; 36: Henry D, Dobson A, Truner C. Variability in the risk of major gastrointestinal complications from nonaspirin nonsteroidal anti-inflammatory drugs. Gastroenterology 1993; 105: Capet C, Czernichow P, Dupas JL, et al. Upper gastrointestinal bleeding in patients treated by low-dose aspirin. Gastroenterol Clin Biol 2001; 25: de Abajo FJ, Garcia Rodriguez LA. Risk of upper gastrointestinal bleeding and perforation associated with low-dose aspirin as plain and enteric-coated formulations. BMC Clin Pharmacol 2001; 1: Laine L. Approaches to nonsteroidal anti-inflammatory drug use in the highrisk patient. Gastroenterology 2001; 120: Serrano P, Lanas A, Arroyo MT, Ferreira IJ. Risk of upper gastrointestinal bleeding in patients taking low-dose aspirin for the prevention of cardiovascular diseases. Aliment Pharmacol Ther 2002; 16: Patrono C, Garcia Rodriguez LA, Landolfi R, Baigent C. Low-dose aspirin for the prevention of atherothrombosis. N Engl J Med 2005; 353: Hurlen M, Abdelnoor M, Smith P, Erikssen J, Arnesen H. Warfarin, aspirin, or both after myocardial infarction. N Engl J Med 2002; 347: Johnson WC, Williford WO, Department of Veterans Affairs Cooperative Study #362. Benefits, morbidity, and mortality associated with long-term administration of oral anticoagulant therapy to patients with peripheral arterial bypass procedures: a prospective randomized study. J Vasc Surg 2002; 35: Gullov AL, Koefoed BG, Petersen P. Bleeding during warfarin and aspirin therapy in patients with atrial fibrillation: the AFASAK 2 study. Atrial Fibrillation Aspirin and Anticoagulation. Arch Int Med 1999; 159: Nielsen GL, Sorensen HT, Mellemkjoer L, et al. Risk of hospitalization resulting from upper gastrointestinal bleeding among patients taking corticosteroids: a register-based cohort study. Am J Med 2001; 111: Taylor DW, Barnett HJ, Haynes RB, et al. Low-dose and high-dose acetylsalicylic acid for patients undergoing carotid endarterectomy: a randomised controlled trial. ASA and Carotid Endarterectomy (ACE) Trial Collaborators. Lancet 1999; 353: Baron JA, Cole BF, Sandler RS, et al. A randomized trial of aspirin to prevent colorectal adenomas. N Engl J Med 2003; 348: Goldstein JL, Gibofsky A, Fort JG. Incidence of Endoscopic Gastroduodenal Ulcers in Subjects on 325 mg of Aspirin for Cardiovascular Prophylaxis with Placebo, a COX-2 Specific Inhibitor, or a Nonspecific NSAID; Results from a Randomized, Double-blind, Controlled Trial. Orlando, FL: American College of Rheumatology, 2003 (late breaking abstracts). 59 Schnitzer TJ, Burmester GR, Mysler E, et al. Comparison of lumiracoxib with naproxen and ibuprofen in the Therapeutic Arthritis Research and Gastrointestinal Event Trial (TARGET), reduction in ulcer complications: randomised controlled trial. Lancet 2004; 364: Rahme E, Bardou M, Dasgupta K, Tobouti Y, Barkun AN. Gastrointestinal effects of rofecoxib and celecoxib versus NSAIDs among patients on low dose aspirin. Gastroenterology 2004; 126 (Suppl. 2): A U.S. Preventive Services Task Force. Aspirin for the primary prevention of cardiovascular events: recommendation and rationale. Ann Intern Med 2002; 136: