The clinical efficacy and long-term safety of aspirin as an antithrombotic agent has been well established by more than 50 randomized clinical trials in a variety of vascular disorders. 1,2 Overall, aspirin reduces the risk of nonfatal as well as fatal clinical manifestations of atherothrombosis by approximately 25% in secondary prevention settings. 3 The saturability of the antithrombotic effect of aspirin at low doses (that is, 100 mg daily) is consistent with this effect being primarily related to permanent inactivation of the cyclooxygenase (COX) activity of platelet prostaglandin H-synthase. 1,2 Nevertheless, the effect of aspirin in the primary prevention of vascular events is less convincingly established, with a protective effect against nonfatal myocardial infarction on the one hand, but by no clear reduction in stroke and no effect on vascular mortality. 2 Moreover, there is even less information on the primary prevention of coronary heart disease in women. 4
A variety of non-aspirin nonsteroidal antiinflammatory drugs (NSAIDs) can inhibit thromboxane (TX) A2-dependent platelet function through competitive, reversible inhibition of platelet COX-1. However, whether partial and reversible inhibition of platelet COX-1 by non-aspirin NSAIDs is sufficient to produce clinically detectable cardiovascular protection, comparable with that achieved by complete and irreversible blockade of this enzyme by low-dose aspirin, is unknown.
We decided to test this hypothesis in a large, population-based cohort study in women 50–74 years of age, free of a history of coronary heart disease. The main objective of our study was to estimate and compare the effects of aspirin and non-aspirin NSAIDs in the primary prevention of myocardial infarction.
Subjects and Methods
Approximately 3 million residents in England and Wales are registered with general practitioners (GPs) who participate in the General Practice Research Database. This scheme comprises more than 2,000 GPs who use computers in their offices for the purpose of recording medical patient information in a standard manner. 5 They have agreed to provide the information anonymously to the Office of National Statistics and to allow the information to be used for research projects. The GPs generate prescriptions directly from the computer, and these are recorded in each patient’s computerized file. As a result, information on over-the-counter drug use is limited. The information recorded includes demographics, clinical diagnoses from outpatient visits, consultant referrals, and hospitalizations. A previous validation study has documented that more than 90% of all referrals present in the manual records in the GPs’ offices are entered into the GPs’ computers with a code that reflects the clinical diagnosis. 6,7 An additional requirement of this data resource is that the indication for any new course of therapy be entered in the computer. In addition, the GP may record laboratory results and other medical data in a free text comment field. A modification of the Oxford Medical Information System (OXMIS) classification is used to code specific diagnoses, and a drug dictionary based on the Prescription Pricing Authority drug dictionary is used to record drugs. 5
We identified a cohort of women 50–74 years of age registered in the General Practice Research Database between January 1, 1991, and December 31, 1995. We excluded women with a history of myocardial infarction, other coronary heart diseases, stroke, neoplasms, coagulopathies, vasculitis, and alcohol-related diseases from the source population. The resulting study cohort comprised 164,769 women who were followed until the date of the first recorded code of myocardial infarction, any of the exclusion criteria (as above), reaching 75 years of age, death, or end of the study period (December 31, 1995).
Ascertainment and Validation of Cases
We identified a total of 1,242 potential cases of first myocardial infarction. We sent a questionnaire to their corresponding GPs and requested all medical records related to the episode of myocardial infarction (including hospital discharge letters, death certificates, or autopsies when available). Overall methods and validation of cases have been described in detail elsewhere. 8 In summary, we considered as a case of acute myocardial infarction a woman presenting at least two of the following criteria: characteristic chest pain, raised levels of serum creatinine kinase, an electrocardiogram diagnostic of myocardial infarction, an arteriogram documenting a recent coronary occlusion, or treatment with fibrinolytic drugs. We also considered as cases those women in whom a consultant cardiologist confirmed the diagnosis of myocardial infarction. We applied the same criteria to fatal cases of myocardial infarction (death in the first month after occurrence of myocardial infarction). For patients who died before reaching the hospital or for whom hospital records were not available, we used the diagnosis recorded on death certificates and autopsies, as available, to determine the cause of death. Women were excluded if another cause of death was recorded. After review of all available information by two independent reviewers (without knowledge of drug exposure), 1,013 women were confirmed with a first episode of myocardial infarction (22% were fatal cases). The remaining 229 were excluded because the diagnosis of myocardial infarction was not established or an exclusion criterion was identified. The index date used was the earliest of the following: date of onset of symptoms, date of hospital admission, or date of death.
Selection of Controls
A random date within the study period was generated for each of the 164,769 women of the study cohort. All women with a random date included in their period of observation (from study entry to end of follow-up) were eligible as controls. 9 A group of 5,000 controls frequency-matched by age was randomly sampled among the study cohort. We applied to the controls the same computer-based exclusion criteria as those applied to cases, using each woman’s random date as her index date.
Assessment of Exposure
We identified aspirin and NSAID prescriptions before the index date and categorized women as non-users when no prescription was ever recorded before the index date. A woman was defined as a current user when she was prescribed aspirin/NSAIDs during the month before the index date and as a past user when the supply of aspirin/NSAIDs ended more than 1 month before the index date. Duration of therapy was defined as the period for which “continuous” supply was prescribed. We considered consecutive prescriptions those with a maximum interval of 60 days between two prescriptions.
Aspirin daily dose was classified into two categories: equal or less than 150 mg and more than 150 mg (300 mg accounting for close to 90% of this dose category).
Non-aspirin NSAID daily dose was classified into low-/medium-dose and high-dose categories. We used similar cutoff values for individual NSAIDs as in a recent study examining the risk of upper gastrointestinal bleeding associated with NSAIDs. 10
The risk of myocardial infarction was examined for exposure to aspirin, NSAIDs, and other risk factors. For each drug regimen, the risk associated with current use was compared with risk in non-users. The effect of duration of use was assessed among current users. The effect of dose was assessed in long-term current users of NSAIDs (duration greater than 1 year).
We calculated person-time of follow-up for all cohort members and used it as the denominator to obtain incidence rates. We also performed a nested case-control analysis.
Sixty per cent of our cases were under the age of 65 years. The overall incidence of myocardial infarction was 1.6 per 1,000 person-years in our study cohort. Table 1 shows the relative risk associated with major risk factors. Smoking and diabetes, in that order, were the two major risk factors. Hormone replacement therapy was associated with a 28% reduction in risk of myocardial infarction. 8
The relative risk of myocardial infarction associated with current use of aspirin was 0.80 [95% confidence interval (95% CI) = 0.41–1.53]. Table 2 shows that prophylactic aspirin use (defined as more than 1 month of treatment duration in current users) was associated with a relative risk of 0.56 (95% CI = 0.26–1.21). We observed the greatest reduction in relative risk among nonfatal cases of myocardial infarction (0.28; 95% CI = 0.08–0.91). There was no indication that aspirin use was associated with a reduced risk of acute coronary deaths, but the confidence limits were wide owing to the small number of exposed cases and controls. The beneficial effect was observed among women using doses equal to or less than 150 mg (83% of these women used 75 mg) of aspirin with a relative risk of 0.30 (95% CI = 0.09–1.00). We observed a slightly smaller effect at doses of 300 mg (data not shown).
Table 3 provides corresponding estimates of risk for NSAIDs. The estimates of relative risk for any NSAID use showed no evidence of protection. There was also a lack of beneficial effect when we stratified by treatment duration or daily dose or when we examined separately fatal and nonfatal cases of myocardial infarction (data not shown).
We examined the association between use longer than 2 months, as well as long-term use (more than 1 year) of individual NSAIDs and the risk of myocardial infarction (data not shown). All of the estimates for individual NSAIDs were compatible with no effect on risk. Long-term use of paracetamol was not associated with a protective effect (data not shown).
The distribution of risk factors among controls, who were long-term users of aspirin and NSAIDs, as a proxy for the distribution in the study cohort, is presented in Table 4. Hypertension was more than twice as frequent among aspirin users than NSAID users, and there were 30% more smokers among women taking aspirin. Ever-use of hormone replacement therapy was 50% less common among aspirin users than NSAID users. The vast majority of long-term aspirin treatment indication was for prophylaxis of vascular events, whereas the main indication for use of NSAIDs was osteoarthritis.
We found that aspirin offers protection against the development of first nonfatal myocardial infarction, but apparently not against fatal coronary heart disease in women. In the same population, we did not observe an association between non-aspirin NSAID use and reduced risk of myocardial infarction. Our finding of a reduction in the risk of a first myocardial infarction among women currently taking aspirin should be viewed within the context of randomized trial evidence and is consistent with the results of primary prevention trials of aspirin in men 11–13 and in men and women, 14 showing a protective effect on the outcome of nonfatal myocardial infarction, but no effect on vascular-related mortality. In one of these studies, aspirin reduced all ischemic heart disease by 20% in men at high risk recruited from 108 practices in the United Kingdom. 12 This effect was the net result of no detectable effect on fatal ischemic heart disease but a large reduction of 32% in nonfatal episodes. This result is similar to the finding of the U.S. Physicians’ Health Study in low-risk men. 11 In the Hypertension Optimal Treatment trial, aspirin reduced all myocardial infarctions 14 by 36%, with no effect on cardiovascular mortality among hypertensive men and women. 14 Although there might be other explanations, 15 overall the effect of aspirin on fatal events in all randomized trials as well as in our observational study remains inconclusive for subjects with a relatively low cardiovascular risk, owing to the small event rate among those in these studies.
There is only one nonexperimental study published in the last decade that examined the role of aspirin in the primary prevention of myocardial infarction in women. 16 The annual incidence risk of nonfatal myocardial infarction was 0.5 per 1,000 women in the Nurses’ Health Study (NHS) cohort and 1.3 in our cohort. The estimate of the reduction for a first myocardial infarction found in the NHS was smaller than in the present study. This difference could be explained in part by a younger, self-selected, and healthier population in the NHS than our study cohort of women drawn from the general population. Moreover, exposure data were elicited through questionnaires and were not ascertained at the time of the occurrence of myocardial infarction, thus possibly leading to some degree of misclassification in the NHS. In our study, we did not have information on over-the-counter drug use and so use of short-term aspirin (pain relief) is missing. On the other hand, it is unlikely that in the United Kingdom patients take prophylactic aspirin for primary cardiovascular prevention without a prescription from a GP. More definitive evidence on the benefit/risk ratio of primary prevention with aspirin in women is needed and is currently being sought in an ongoing randomized, placebo-controlled trial, the Women’s Health Study. 17
The increase in relative risk observed with short-term, current use of either aspirin or other NSAIDs was likely related to recognized or unrecognized prodromic symptoms of heart attack. In general, patients taking short-term medications at the time of onset of acute myocardial infarction are a high-risk subgroup. 18
We are not aware of any published epidemiologic study that specifically examined whether non-aspirin NSAIDs share the same antithrombotic effect of aspirin through competitive, reversible inhibition of platelet COX-1. The apparent lack of cardiovascular protection associated with long-term NSAID use in the present study should be considered within the context of limited randomized trial evidence. 2 Thus, the only reversible COX inhibitors that have been tested for their antithrombotic efficacy are sulfinpyrazone, indobufen, and triflusal, with conflicting results in trials of limited sample size. 2 None of these drugs is approved as an antiplatelet agent in the United Kingdom, and it is unclear under which circumstances they are prescribed in other countries.
Our null findings for non-aspirin NSAIDs are mechanistically consistent with the knowledge that more than 95% suppression of platelet COX-1 activity is required to inhibit TXA2 biosynthesis in vivo and TXA2-dependent platelet function. 19 Low-dose aspirin is associated with virtually complete blockade of this metabolic pathway because of the cumulative nature of platelet COX-1 acetylation upon repeated daily dosing. 20 In contrast, non-aspirin NSAIDs, when used at conventional antiinflammatory dosages, inhibit platelet COX activity by only 70–85% in a time-dependent reversible fashion. 21 Thus, the present results argue against the hypothesis that primary prevention of myocardial infarction by low-dose aspirin is related to its antiinflammatory effect, as suggested recently by Ridker et al22 on the basis of a subgroup analysis of the Physicians’ Health Study. Furthermore, our data support the notion that the pattern of COX-1 inhibition associated with non-aspirin NSAID use is insufficient to block TX-mediated amplification of platelet activation at the site of coronary plaque rupture. Alternatively, inhibition of vascular prostacyclin synthesis—largely a COX-2-driven process 23 —by non-aspirin NSAIDs, but not by low-dose aspirin, might offset any protective effect resulting from incomplete TXA2 inhibition. In view of this, the potential cardiovascular effects of selective COX-2 inhibitors remain to be evaluated after a large enough number of patients will have been exposed over a long term to this new class of agents.
We thank the general practitioners for their excellent collaboration and the Boston Collaborative Drug Surveillance Program for providing access to the General Practice Research Database. Validation of cases was supported by a grant of Novartis. We also thank Susanne Pérez Gutthann for helpful comments and the review of the manuscript, and Marta Marcet for her support in data entry and the editing of this report.
1. Patrono C. Aspirin
as an antiplatelet drug. N Engl J Med 1994; 330:1287–1294.
2. Patrono C, Coller B, Dalen JE, Fuster V, Gent M, Harker LA, Hirsh J, Roth G. Platelet-active drugs: the relationships among dose, effectiveness, and side effects. Chest 1998; 114:470S–88S.
3. 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; 308:81–106.
4. Rich-Edwards JW, Manson JE, Hennekens CH, Buring JE. Primary prevention
of coronary heart disease
in women. N Engl J Med 1995; 332:1758–1766.
5. García Rodríguez LA, Pérez Gutthann S. Use of the UK General Practice Research Database for pharmacoepidemiology. Br J Clin Pharmacol 1998; 45:419–425.
6. Jick H, Jick SS, Derby LE. Validation of information recorded on general practitioner based computerised data resource in the United Kingdom. BMJ 1991; 302:766–768.
7. Van Staa TP, Abenhaim L. The quality of information recorded on a UK database of primary care records: a study of hospitalizations due to hypoglycemia and other conditions. Pharmacoepidemiol Drug Safety 1994; 3:15–21.
8. Varas-Lorenzo C, García-Rodríguez LA, Pérez-Gutthann S, Duque-Oliart A. Hormone replacement therapy and incidence of acute myocardial infarction
: a population-based nested case-control study. Circulation (in press).
9. Walker AM. Observation and Inference. Newton Lower Falls, MA: Epidemiology Resources Inc, 1991.
10. García Rodríguez LA, Cattaruzzi C, Troncon MG, Agostinis L. Risk of hospitalization for upper gastrointestinal tract bleeding associated with ketorolac, other nonsteroidal anti-inflammatory drugs, calcium antagonists and other antihypertensive drugs. Arch Intern Med 1998; 158:33–39.
11. 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; 321:129–135.
12. 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 ischemic heart disease in men at increased risk. Lancet 1998; 351:233–241.
13. Peto R, Gray R, Collins R, Wheatley K, Hennekens C, Jamrozik K, Warlow C, Hafner B, Thompson E, Norton S, Gilliland J, Doll R. Randomised trial of prophylactic daily aspirin
in British male doctors. BMJ 1988; 296:313–316.
14. Hansson L, Zanchetti A, Carruthers SG, Dahlof B, Elmfeldt D, Julius S, Menard J, Rahn KH, Wedel H, Westerling S. 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; 351:1755–1762.
15. Born G. Effects of aspirin
and warfarin on fatal and non-fatal heart attacks. Lancet 1999; 354:1472.
16. Manson JE, Stampfer MJ, Colditz GA, Willett WC, Rosner B, Speizer FE, Hennekens CH. A prospective study of aspirin
use and primary prevention
of cardiovascular disease in women. JAMA 1991; 266:521–527.
17. The Women’s Health Initiative Study Group. Design of the Women’s Health Initiative clinical trial and observational study. Control Clin Trials 1998; 19:61–109.
18. Thomson PL, Nidorf SM, Parsons RW, Jamrozik KD, Hobbs MST. The benefits of β-blockade at the time of myocardial infarction
. J Hypertens 1991;S9: 3:5–7.
19. Reilly IA, FitzGerald GA. Inhibition of thromboxane formation in vivo
and ex vivo: implications for therapy with platelet inhibitory drugs. Blood 1987; 69:180–186.
20. Patrignani P, Filabozzi P, Patrono C. Selective cumulative inhibition of platelet thromboxane production by low-dose aspirin
in healthy subjects. J Clin Invest 1982; 69:1366–1372.
21. Patrono C, Ciabattoni G, Patrignani P, Pugliese F, Filabozzi P, Catella F, Davi G, Forni L. Clinical pharmacology of platelet cyclooxygenase inhibition. Circulation 1985; 72:1177–1184.
22. Ridker PM, Cushman M, Stampfer MJ, Tracy RP, Hennekens CH. Inflammation, aspirin
, and the risk of cardiovascular disease in apparently healthy men. N Engl J Med 1997; 336:973–979.
23. McAdam BF, Catella-Lawson F, Mardini IA, Kapoor S, Lawson JA, FitzGerald GA. Systemic biosynthesis of prostacyclin by cyclooxygenase (COX)-2: the human pharmacology of a selective inhibitor of COX-2. Proc Natl Acad Sci USA 1999; 96:272–277.