An aspirin a day? Aspirin use across a spectrum of risk: cardiovascular disease, cancers and bleeds

Abstract

Aspirin or acetylsalicylic acid (ASA) is commonly used in the general population for primary prevention of cardiovascular disease (CVD). Strong evidence supports the use of ASA in secondary prevention of CVD; however, for primary prevention, potential benefits are offset by potential harms (primarily major bleeds), with no benefit in overall mortality. Anti-platelet agents, including ASA, are one of the most commonly implicated medications for hospital admissions related to adverse medication events. Studies of primary prevention in patients with risk factors for CVD also fail to show a benefit with ASA. Finally, evidence supporting ASA use for cancer prevention is limited. Health care providers should be aware of the benefits and risks associated with ASA use in primary and secondary prevention and discuss these with their patients in the context of individual patient values and preferences.

Keywords::

• acetylsalicylic acid
• aspirin
• cancer
• cardiovascular disease
• primary prevention

1. Current rates of acetylsalicylic acid use by patients

About 40% of people in the USA and Canada over the ages of 40 or 50 are regularly using acetylsalicylic acid (ASA), mostly for cardiovascular disease (CVD) prevention, and more commonly for primary than secondary CVD prevention [1-3].

When ASA is used for primary CVD prevention, the potential benefits of fewer cardiovascular events (CVE) are offset by the potential harms of bleeding [4,5]. On the contrary, when ASA is used for secondary CVD prevention, the potential benefits outweigh the potential harms [6,7]. Despite the evidence of benefit in secondary prevention, and although patients appear to be compliant with their anti-platelet therapy early after their CVE [8], other studies report that up to 40% of patients with previous CVD are not using ASA [1-3,9].

2. Evidence for ASA in CVD prevention

2.1 Secondary CVD prevention

For patients with existing CVD, ASA at ≤ 325 mg per day reduces future CVE with a number needed to treat (NNT) of 30 over about 3 years [7]. This includes the reduced risk of myocardial infarction (NNT = 83), stroke (NNT = 40) and all-cause mortality (NNT = 71) [5]. When ASA is used in this population, the risk of major bleeding increases [number needed to harm (NNH) = 111], but overall the benefits outweigh the potential risks [7]. In primary prevention the numbers are quite different.

2.2 Primary CVD prevention

Evidence of ASA in primary CVD prevention comes from six randomized-controlled trials (RCTs) [10-15]. There were issues with these trials that should have limited the resultant widespread promotion of ASA for primary CVD prevention. These issues included: using non-generalizable study subjects (physicians or health care workers) [10,11,15], pre-trial run ins (where patients with adverse events or non-compliance were excluded) [15,16], unblinded [11,14], early terminations (one when positive findings from other studies were released) [10,14] and all were either supplied with medications or supported financially by the manufacturers of ASA.

Four trials demonstrated a reduction in CVE [10,12-14], however, despite the numerous biases, only the Primary Prevention Project [14] showed statistically significant (albeit borderline) reduction in cardiovascular mortality [relative risk 0.56 (95% CI: 0.31 – 0.99)] and no trial showed a significant reduction in overall mortality.

Two early meta-analysis were performed on these studies [4,5], one separating findings based on patient sex [4]. For women, the NNT to prevent one CVE (primarily strokes) was 330 over about 6 years. This benefit was offset by the increased risk of major bleeding (NNH = 400). Similarly, 270 men would need to take ASA for about 6 years to prevent one CVE (primarily myocardial infarction), and for every 303 men taking ASA, one additional major bleed would occur. There was no difference in cardiovascular or all-cause mortality in either sex [4,5], but for males there was a statistically significant increase in hemorrhagic strokes (NNH = 770) [4]. Therefore, when ASA is used for primary CVD prevention, the potential benefits are offset by potential harms without altering mortality.

2.2.1 Diabetics and hypertensive patients

Even in diabetic patients, who may have a greater risk of CVE and cardiovascular death [17], a systematic review of six RCTs failed to show a reduction in CVE, cardiovascular mortality, or all-cause mortality in diabetic patients on ASA [18]. Similarly, another systematic review of ASA use in hypertensive patients demonstrated that the reduction in myocardial infarction was outweighed by the increase in bleeding [19].

2.2.2 Confusing systematic reviews and guidelines on ASA for primary CVD prevention

Two groups recently published updated systematic reviews on ASA for primary CVD prevention. They both included the same nine primary prevention studies including patients with diabetes, peripheral vascular disease and hypertension [20,21]. Despite deriving the exact same point estimate [0.94 (95% CI: 0.88, 1.00)], one group concluded that ASA decreased all-cause mortality [20], while the other group did not [21].

Different guidelines also provide discrepant interpretations of the literature. European guidelines do not recommend ASA at all for primary prevention [22], Canadian guidelines recommend using in special circumstances [23], and one American guideline encourages ASA if the predicted benefits outweigh the harms [24]. Unfortunately, patients at increased risk of a future CVE are often also at increased risk of bleeding [5,25] and predicting who may achieve a net clinical benefit is difficult. Finally, another American guideline suggests that all males and females > 50-years old should use ASA, regardless of cardiovascular risk [26].

3. Potential adverse events with ASA use

All medications have potential adverse events. In a UK study of two large general hospitals, one in 15 admissions was related to medication adverse events. ASA was the commonest medication implicated, and was used in 61% of the patients who died as a result of the medication adverse event [27]. In the USA, there are an estimated 99,628 annual hospital admissions in patients ≥ 65-years old for medication adverse events. Four drug classes, including anti-platelet agents, are responsible for 67% of these admissions [28].

Compared to placebo, ASA approximately doubles one's relative risk of overall bleeding, irrespective of major, minor or intracranial [29]. However, the absolute risks (and, therefore, NNH) depend on the baseline risk of these different events. A systematic review of 65,987 predominantly middle-aged males found that, irrespective of dose, ASA increased overall gastrointestinal bleeding from a baseline of 1.4% to 2.5% over about two and a half years (NNH = 91) [30].

For major gastrointestinal bleeding (bleeding resulting in hospital admission, transfusion or death), another systematic review found the annual baseline risk was 0.12% with the attributable risk of using ASA 0.12% [31]. Therefore, for every 833 patients using ASA, every year one additional patient will experience a major GI bleed. Finally, for every approximately every 1000 patients who use ASA for 3 – 6 years, one additional hemorrhagic stroke will occur [4,31].

Recently published cohort data found baseline and ASA associated hospitalizations for gastrointestinal and intracerebral bleeding rates five times higher than rates seen in RCTs [25]. This may reflect real-world bleeding rates, the older study population, or potentially residual confounding.

4. ASA and cancers

Using studies on ASA for primary and secondary CVD prevention, Rothwell et al. have published a number of meta-analyses demonstrating a reduction in the relative risk of colorectal cancer [32] and general cancer mortality [33]. These results should be interpreted with caution for two main reasons. First, two large studies [10,15] were excluded from the analyses on the basis that ASA was used every other day. These excluded studies had twice as many patients as the number of patients in the meta-analyses, and initial reports did not show a reduction in colorectal cancer incidence [34,35], colorectal cancer death [35] or overall cancer incidence or mortality [35]. Long-term follow-up (mean of 18 years) of the Women's Health Initiative did show that ASA reduced colorectal cancer incidence rates, but no difference in overall cancer mortality was found [36]. In addition, multiple comparisons and sub-group analysis were performed in the meta-analyses which increase the likelihood that some findings were due to chance. Benefit remains controversial, and if present, duration of daily aspirin use required to obtain this benefit is unclear.

ASA may, however, be beneficial in patients who already have colorectal cancer [37], especially in those expressing certain genetic mutations [38], but these findings require confirmation with RCTs. Until the ASA and cancer mortality systematic review is re-analyzed with proper study inclusion or additional research is available, ASA cannot be recommended for cancer prevention.

5. Expert opinion

ASA is being used by millions of patients worldwide, most commonly for primary CVD prevention. For these patients, the potential benefit of decreased CVE is offset by the potential harms of bleeding, with cardiovascular or overall mortality being unchanged.

On the contrary, many patients with established CVD, where the evidence demonstrates a net clinical benefit, are not using ASA. Some patients start (and stop) ASA on their own, whereas others commence ASA on the recommendation of their family physician [3]. To address this ‘aspirin gap,’ both health care providers and patients need to be better educated about the benefits and harms of ASA therapy for primary and secondary CVD prevention.

Shared clinical decision making with individual patients should occur after a conversation between the patient and their primary health care provider. This dialogue should outline the potential benefits and harms of therapy in understandable terms (Table 1) as well as taking into account patients values and preferences.

Table 1. Numbers needed to treat (NNT) and harm (NNH) pertaining to aspirin or acetylsalicylic acid therapy.

Patients	Cardiovascular events (NNT)	All-cause mortality (NNT)	Major GI bleeds* (NNH)
Primary prevention females (over ∼ 6 years)	303	NS	400
Primary prevention males (over ∼ 6 years)	270	NS	303
Secondary prevention (females and males) (over ∼ 3 years)	30	71	111

Number of patients needed to treat over a specific time frame to prevent (NNT) or cause (NNH) one additional event.

*Major GI bleeds commonly defined as admission to hospital, blood transfusions or death.

NNH: Numbers needed to harm; NNT: Numbers needed to treat.

For most patients in primary care who do not have CVD, discussions surrounding lifestyle modification (Mediterranean diet, smoking cessation and exercise), statin therapy or hypertension management should occur before considering ASA therapy. Given the current level of evidence, ASA should not be recommended for cancer prevention.

Declaration of interest

The authors state no conflict of interest and have received no payment in preparation of this manuscript.

Acknowledgement

The authors wish to thank Sharon Nickel (Evidence Based Medicine Programme, Alberta College of Family Physicians) for manuscript formatting and editing for grammar.