Lifelong Aspirin for All in the Secondary Prevention of Chronic Coronary Syndrome: Still Sacrosanct or Is Reappraisal Warranted?

Alan P. Jacobsen, MB, BCh, BAO; Inbar Raber, MD; Cian P. McCarthy, MB, BCh, BAO; Roger S. Blumenthal, MD; Deepak L. Bhatt, MD, MPH; Ronan W. Cusack, BM, BS, BSc; Patrick W.J.C. Serruys, MD, PhD; William Wijns, MD, PhD; John W. McEvoy, MB, BCh, BAO, MHS


Circulation. 2020;142(16):1579-1590. 

In This Article

Historical Trials Testing the Efficacy of Aspirin for Secondary Prevention of CCS

Participants With a History of Previous MI

Early trials of aspirin for secondary prevention showed some evidence of efficacy among adults with a history of previous MI (Table 1).[26–31] In 1974, in the Cardiff I trial,[26] 1239 men with a discharge diagnosis of MI were randomized to receive aspirin 300 mg daily or placebo for the duration of follow-up. There was a relative reduction in mortality of 25% at 12 months with aspirin, but this was not statistically significant. The Cardiff II trial,[27] published in 1979, included women (248 out of 1682 participants) and randomized participants with MI to aspirin 300 mg 3 times daily or placebo. Fifty percent of participants were enrolled within 7 days of the index MI. There was a 2.5% absolute risk reduction in the primary outcome of all-cause death, which was not statistically significant. A 6.2% absolute risk reduction in the secondary composite outcome of all-cause mortality and nonfatal ischemic heart disease morbidity was observed in the aspirin group (P<0.05).

Subsequent trials evaluated the effect of aspirin initiation at later times after the index MI event. The 1976 CDPA (Coronary Drug Project Aspirin Study) randomized 1529 men with a history of MI to receive aspirin 324 mg 3 times daily or placebo.[28] Three-quarters of participants were randomized >5 years after the diagnosis of MI. The risk of mortality in the aspirin group was 5.8% compared with 8.3% in the placebo group at an average follow-up time of 22 months (although this difference in favor of aspirin did not achieve statistical significance).

The 1980 PARIS (Persantine–Aspirin Reinfarction Study) was a double-blind, randomized trial that included 2026 persons (1759 men and 267 women) with MI ≈8 weeks to 60 months before enrollment.[29] Participants were randomized to aspirin 324 mg with dipyridamole (Persantine) 75 mg 3 times daily, aspirin 324 mg 3 times daily, or placebo for the duration of the study. Three coprimary endpoints were tested: all-cause mortality, coronary death, and total coronary events (coronary death and nonfatal MI). After an average follow-up of 41 months, there was no statistically significant difference in mortality among the dipyridamole–aspirin, aspirin-only, and placebo groups (10.7% versus 10.5% versus 12.8%). The other coprimary end points of coronary death and total coronary events were both significantly lower in the dipyridamole–aspirin group and aspirin-only group when compared with placebo.

The 1980 AMIS (Aspirin Myocardial Infarction Study) assigned 4524 individuals with a history of at least 1 MI to aspirin 500 mg twice daily or placebo.[30] Randomization occurred 8 weeks to 60 months after the MI. After 3 years of follow-up, there was no statistically significant difference in mortality (9.6% in the aspirin group and 8.8% in the placebo group). A secondary end point of definite nonfatal MI was significantly lower (6.3% in the aspirin group compared with 8.1% in the placebo group).

The 1980 GAMIS (German–Austrian Aspirin Trial) randomized 946 patients (203 women) to aspirin 1500 mg daily, phenprocoumon (a vitamin K–dependent anticoagulant), or placebo 30 to 42 days after a diagnosis of MI.[31] At a 2-year follow-up, the primary end point of coronary deaths (fatal MI and sudden deaths) was numerically lower in the aspirin group compared with placebo (13 versus 22), but this did not reach statistical significance.

A number of factors limit the application of the historic trials to clinical practice in the contemporary era. First, the majority of participants in the trials were white men, limiting ability to assess whether differences in sex and race might affect outcomes. Second, participants included in these studies tended to be younger, limiting the generalizability of these results to individuals older than 70 years. Third, the dose of aspirin was higher (300 to 1500 mg daily) than contemporary dosing (81 to 325 mg daily). Fourth, only PARIS and AMIS reported use of β-blockers or cholesterol-lowering medications, which was <20% of patients in both trials. Fifth, in contrast to the generally chronic and lifelong use of aspirin recommended for secondary prevention in clinical practice, the longest duration of these trials was 4 years (PARIS). Sixth, the timing of enrollment in these trials varied substantially from weeks to years after the incident MI and those trials that enrolled individuals within 6 months of the index event (Cardiff I, Cardiff II, PARIS, and GAMIS) were the ones that tended to suggest benefit for aspirin, whereas trials that enrolled participants who were further out in time from their index MI tended to show more neutral results, suggesting a diminishing benefit over time (a finding that appears to undermine the need for lifelong aspirin in all patients with CCS).

Putting it all Together: The ATT Collaboration

Most of the early trials examining aspirin use in CCS reported trends toward a benefit for aspirin, but when considering each trial in isolation, none of the 6 after-MI trials in the ATT meta-analyses consistently demonstrated a statistically significant reduction in the primary end point. Therefore, the individual trials themselves can be considered inconclusive as to the benefit for aspirin for secondary prevention, even without extrapolating these historical results to modern patients. This is a nuance that is often underappreciated.

The first ATT meta-analysis in 1988 included 25 randomized trials on antiplatelet treatment, with a collective 29 000 patients who had a history of acute MI, unstable angina, ischemic stroke, or transient ischemic attack.[1] When the results of these individual historical trials were combined, allocation to an antiplatelet treatment convincingly decreased risk of vascular death by 15% (P<0.001). Although the majority of studies used aspirin (dosed at 300 to 1500 mg daily), a number of trials used sulphinpyrazone or a combination of aspirin with dipyridamole.

The second ATT meta-analysis was published in 1994 in an effort to examine antiplatelet effect across a wider range of clinical indications.[2] The meta-analysis included 145 studies, and analyses were divided into an acute MI group and a previous MI group (the latter being more representative of CCS). Among 20 000 patients with a history of previous MI, there was a 4% absolute risk reduction in vascular events over 2 years with antiplatelet therapy (P<0.001). The most widely tested antiplatelet regimen was aspirin.

The third ATT study (2002) assessed lower doses of aspirin.[3] Eleven trials compared different aspirin doses, although none of these trials included patients with CCS. The authors concluded that aspirin is protective in most patients at risk of occlusive vascular events and that low-dose aspirin is an effective antiplatelet regimen for long-term prevention.