What are the benefits and harms of low-dose aspirin in adults with diabetes mellitus?

The 7740 patients who took low-dose aspirin experienced 51 fewer vascular deaths, nonfatal myocardial infarctions (MIs), or nonfatal ischemic strokes; 29 fewer transient ischemic attacks (TIAs); and 44 fewer revascularizations than patients who took placebo over a mean of 7.4 years. This is balanced by an additional 69 major bleeding episodes during that period, with no effect on vascular or all-cause deaths, and no difference in the incidence of cancer. (LOE = 1b)

The ASCEND Study Collaborative Group, Bowman L, Mafham M, et al. Effects of aspirin for primary prevention in persons with diabetes mellitus. N Engl J Med 2018;379(16):1529-1539.  [PMID:30146931]

Randomized controlled trial (double-blinded)

Industry + foundation

Concealed

Outpatient (any)

This British study recruited adults 40 years and older with diabetes mellitus, no known cardiovascular disease, no contraindications to aspirin, and no major comorbidity that would keep them from participating in the study for at least 5 years. After a placebo run-in period to assure adherence, 15,480 participants were randomized to receive aspirin 100 mg once daily or matching placebo. They were also randomized to receive an omega-3 fatty acid capsule or placebo; those results are reported separately. The groups were balanced at the start of the study: the patients had a mean age of 63 years, 63% were men, and 96% were white. Almost all (94%) had type 2 diabetes mellitus. A validated risk score determined that approximately 40% of participants were at low risk of vascular events (< 5% at 5 years), 40% had a 5-year risk of 5% to 10%, and the remainder were at high risk. As the trial was ongoing, the authors added TIA to the original primary composite efficacy outcome of vascular death, nonfatal MI, or nonfatal stroke (excluding intracranial hemorrhage). The primary safety outcome was a composite of intracranial hemorrhage, intra-ocular hemorrhage that threatens sight, gastrointestinal bleeding, or any other serious bleeding event. After a mean follow-up of 7.4 years, 99% of patients had complete follow-up data, with outcomes adjudicated for more than 90% by a committee masked to treatment assignment. The authors also looked at the effect of adding revascularization to the composite efficacy outcome. There was no difference between groups in the original efficacy outcome of vascular death, nonfatal MI, and nonfatal ischemic stroke (7.0% with aspirin vs 7.6% with placebo; hazard ratio [HR] 0.92, 95% CI 0.82 - 1.03). When you add TIA to the composite outcome, the difference between groups is statistically significant (8.5% vs 9.6%; HR 0.88, 0.79 – 97; number needed to treat [NNT] = 90 for 7.4 years). Adding revascularization to the original efficacy outcome had a similar result (10.8% vs 12.1%; HR 0.88, 0.80 - 0.97; NNT = 77 for 7 years). When examining results stratified by vascular risk, those at moderate and higher vascular risk also experienced more major bleeding events (8.9 - 9.6 vs 2.8 per 5000 person-years in the low-risk group). The number of serious vascular events avoided per 5000 person years was 5.7 in the low-risk group, 11.2 in the moderate-risk group, and only 4.9 in the high-risk group. For the composite harm outcome, there was a significantly increased risk of major bleeding, primarily due to more serious gastrointestinal and other bleeds (4.1% vs 3.2%; HR 1.29, 1.09 - 1.52; number needed to treat to harm = 111 over 7 years). However, there was no difference in fatal bleeding events or hemorrhagic strokes. There was no difference in the incidence of cancer (11.6% for aspirin vs 11.5% for placebo), including for gastrointestinal cancers (2.0% vs 2.0%). There was no significant differences between groups in all-cause mortality or in vascular deaths.