Last week I had a patient with diabetes ask whether he should take aspirin. My patient is a 47-year-old white male with good glycemic control and an A1C around 6.5% since his diagnosis 5 years ago, which occurred shortly after he stopped smoking.

At this examination his systolic blood pressure was 130 mm Hg, and lab tests showed a total cholesterol 200 mg/dl, and an HDL of 40mg/dl. He had no known ischemic vascular disease, nor any known risks for increased bleeding, i.e., no peptic ulcers, GI bleeds, blood thinners, or steroids. He had no known bleeding diathesis and was not a chronic alcohol user. He was not on SSRI antidepressants, which increase bleeding risk, nor routine NSAIDS, which decrease the effectiveness of aspirin.

Although this patient’s risk calculation for cardiovascular disease using UKPDS was 10.2%, the new Pooled Cohort Risk Assessment Equations developed by the Risk Assessment Work Group of the ACC/AHC Task Force on Guidelines put him at an estimated risk of 6.4% over 10 years.ⁱ If he maintains his current health, his 10-year risk for a CV event doesn’t reach 10% until he is 52. What happens now when his risk is between 5 and 10%?

Conflicting Guidelines
It struck me that answering this patient’s question about aspirin therapy used to be easier than it is now. The current American Diabetes Association (ADA), American Heart Association (AHA), and American College of Cardiology Foundation (ACCF) recommendations for aspirin for primary prevention of CV disease have remained basically the same since 2010, although the level of evidence has now been has dropped to a grade C.ⁱⁱ The ADA recommends that low-dose aspirin (75–162 mg/day) should be considered for adults with diabetes at increased CVD risk (10-year risk of CVD events over 10%) and not at increased risk for bleeding. This includes most men aged 50 or women age 60 with at least one additional major risk factor (i.e., family history of CVD, hypertension, smoking, dyslipidemia, or albuminuria).

On the other hand, the European Society for Cardiology Joint Task Force no longer recommends aspirin for primary prevention of CVD in people with diabetes, as reported in the European Guidelines on Cardiovascular Disease Prevention in Clinical Practice. This conclusion sharply differs from the current ADA age/gender/risk-based recommendation and has generated concern in the US about the value of aspirin for the primary prevention of cardiovascular disease (CVD) among people with diabetes.ⁱⁱⁱ

Complicating matters further, both recommendations are in striking contrast to the 2012 Evidence-Based Clinical Practice Guidelines from the American College of Chest Physicians, which recommend low-dose aspirin for everyone over 50 regardless of CV risk profile. Using the same data as the other groups, the Chest Physicians conclude that the presence or absence of diabetes makes no difference in the relative effect of aspirin at all.^iv Confusion was further fueled by the Federal Drug Administration denial of Bayer Pharmaceutical’s request to re-label aspirin as an Over-the-Counter (OTC) drug to prevent a first heart attack.^v The FDA found insufficient evidence to support a change in the aspirin label, however no change in existing recommendations was suggested by the FDA, including specifically the USPSTF recommendations. The FDA found the USPSTF recommendations to be compatible with the FDA position, but considered addition to the OTC aspirin label to require a higher level of evidence than the USPSTF recommendations that were targeted at groups with high cardiovascular risk.

 

How Did We Get To This Point?
In 2006, the American Diabetes Association (ADA) and the American Heart Association (AHA) jointly recommended aspirin therapy for the primary prevention of heart disease for individuals with type 1 or type 2 diabetes older than 40 or older than 30 with cardiovascular risk factors.^vi The American Stroke Association and the AHA further recommended the use of aspirin for diabetic patients with a 10-year CV risk of 6% to 10%.^vii The evidence for the effects of aspirin came predominately from the Physicians’ Health Study, the Hypertension Optimal Treatment (HOT), and meta-analysis of earlier studies.^viii,ix

In the presence of metabolic syndrome or diabetes, aspirin lacks uniform effectiveness in blocking platelet COX-1 activity, a necessary step for antiplatelet activity. Even so, these studies put to rest concerns that aspirin’s protective effect might be diminished in the presence of diabetes, providing strong evidence for secondary prevention, even in the presence of diabetes. In the HOT trial, for example, 75 mg of aspirin significantly reduced the risk of major cardiovascular events in patients with diabetes who had well-controlled hypertension by 15%, reducing all myocardial infarction by 36% but with no effect on stroke. It should be noted that in the HOT study non-fatal major bleeds were significantly more common among patients receiving aspirin.^x

In 2009 the US Preventive Services Task Force (USPSTF) provided a detailed meta-analysis with a high (A) level of evidence recommending aspirin for primary prevention of CV disease in men ages 45 to 79 years when “the potential benefit of a reduction in myocardial infarction outweighs the potential harm of an increase in gastrointestinal hemorrhage” and for women ages 55 to 79 when “the potential benefit of a reduction in ischemic stroke outweighs the potential risk of GI hemorrhage.”^xi The USPSTF highlighted outcome differences between men and women, providing a basis for subsequent US evidence-based recommendations.

Unfortunately, the USPSTF recommendation required calculations of individual risk that were impractical in clinical practices without electronic support. The risk calculator initially provided online has changed, and the URL now automatically redirects to an NHLBI risk calculator not intended for use in patients with diabetes.

More Evidence, More Questions
Serious questions began in 2008 when the POPADAD clinical trial brought into question the effectiveness of aspirin for primary prevention.^xii POPADAD evaluated 1,276 diabetes patients with an ankle brachial pressure index ≤.99 divided into aspirin and placebo groups (with and without antioxidant). With an average follow up of 6.7 years, the study provided a 75% chance of detecting a 25% reduction, and an 80% chance of detecting a 30% reduction, in a composite CV endpoint. The subject cohort was high-risk, with an average age of 60 years and 33% current smokers. Although expecting approximately 392 events, the study recorded a composite endpoint in 233 subjects (2.9 per 100 patient years). The POPADAD trial was underpowered to detect the 7-14% risk reduction now expected for aspirin in primary prevention.

The Low-Dose Aspirin for Primary Prevention of Atherosclerotic Events in Patients with Type 2 Diabetes (JPAD trial) was also published in 2008.^xiii JPAD randomized 2,539 patients from 163 Japanese institutions to low-dose aspirin or placebo and followed them for 4.7 years. Although the 154 composite CV events demonstrated a 20% reduction in the aspirin cohort, the reduction did not reach statistical significance (p=.16). Again, the study was not powered to detect a small reduction.

The 2008 Aspirin for Asymptomatic Atherosclerosis (AAA) trial randomized 28,980 patients without vascular disease and followed them for an average of 1.6 years.^xiv As in both the POPADAD and JPAD studies, no significant difference was found in the composite endpoint of fatal or nonfatal coronary event or stroke or revascularization. In this trial, however, 70% of the participants were women, and only 88 patients had diabetes. The study was powered to have an 80% chance of detecting a 30% reduction in cardiovascular events. The composite endpoint mixed genders and outcomes, diluting the possible cardiovascular effect in men. This study, too, was underpowered to detect the rate of risk reduction expected in primary prevention.

A subsequent meta-analysis of study level data from all available randomized clinical trials by cardiovascular events (total 11,787 subjects from 9 trials ranging from 68 to 3,711 subjects) the ADA and AHA found no statistically significant reduction in the risk of major or all-cause mortality across all people with diabetes and no pre-existing CVD.^xv The Antithrombotic Trialists’ (ATT) Collaboration had suggested gender specific reductions of CHD events in men and stroke in women; however the ADA/AHA meta-analysis did not have access to patient level and did not consider the effects of gender. Adding patients with diabetes from 3 new trials to the 6 trials analyzed previously by the ATT, a 9% reduction in fatal and nonfatal MI and a 15% reduction in stroke was estimated. Although not statistically significant in the meta-analysis, the estimated values approximate the 14% decrease in CV risk in primary prevention suggested by the USPSTF.

Additional compelling evidence for the use of aspirin has been the reduction in cancer deaths seen with aspirin use.^xvi In a meta-analysis of 51 clinical trials, the reduced risk of major vascular events from aspirin was initially offset by an increased risk of major bleeding, but effects on both outcomes diminished with increasing follow-up, leaving only a significantly reduced risk of cancer (absolute reduction 3.13 per 1000 patients per year) from 3 years onwards. Interestingly, case fatality from major extra-cranial bleeds was significantly lower on aspirin than on control (OR 0.32, p=0·009).

The Decision
So whose recommendation do we follow for this patient?

My decision was to follow the Risk Assessment Work Group of the ACC/AHC Task Force on Guidelines lead, split the difference, and go with 7.5%. In this case, the patient hits that risk at age 50. As for his history of smoking, it made me more likely to suggest that he start aspirin. Nevertheless, I recommended that he start aspirin at age 50, and reminded him not to start smoking again.

My recommendation, after another review of the evidence, falls most in line with 2014 ADA recommendations. It closely reflects the 2009 USPSTF recommendation, although the starting age has been raised another 5 years for both men and women. As for the new European recommendations—my patient made the cynical observation that not recommending aspirin may save money in European delivery systems, at least in the short run.

In the end, I am certainly more judicious today about who gets aspirin. An increased risk of bleeding quickly eliminates the small but important 7-14% benefit that may be gained. We have at least two new trials, ASCEND and ASPREE, that are likely to provide a better answer for us in the next 3 years, examining primary prevention in individuals with diabetes and over the age of 70 respectively. Although the USPSTF evaluation is now 5 years old, their analysis has proved to be the most stable, and it still provides a solid foundation for evidence-based treatment.

I hope this case will prompt a wider discussion among primary care physicians and endocrinologists.

• How would you treat this patient?

• What issues are most important to you?

REFERENCES