ASCEND (Aspirin)

Published

Bowman L, et al. "Effects of aspirin for primary prevention in persons with diabetes mellitus". The New England Journal of Medicine. 2018. epub 2018-08-26:1-11.
PubMed • Full text • PDF

Clinical Question

In patients with diabetes with no established atherosclerotic disease, does the use of low-dose aspirin reduce vascular events compared to placebo? Does low-dose aspirin significantly increase the risk of bleeding?

Bottom Line

In patients with well-controlled diabetes (hgb a1c < 8 in ~51% of patients), at 7.4 years primary prevention aspirin use was associated with a 1.1% absolute reduction in serious vascular events. Aspirin use was also associated with a 0.9% absolute increase in major bleeding. Aspirin use was not associated with any change in the incidence of gastrointestinal cancers or all cancers.

Major Points

Although it is well-accepted that aspirin is beneficial in reducing recurrent vascular events in patients with known vascular disease, the role of aspirin as primary prevention therapy (i.e., therapy designed to reduce vascular events among patients without established atherosclerotic disease) is unclear despite being the focus of multiple studies. The 2009 Antithrombotic Trialists' Collaboration meta-analysis involving 95,000 patients in six primary prevention trials showed that aspirin use was associated with a 12% relative reduction in serious vascular events, but was also associated with a 50% relative increase in bleeding. The prevailing interpretation of these findings was that any benefit seen with aspirin as primary prevention was likely largely, if not completely, counterbalanced by increased bleeding risk. As a result of these data, the USPSTF guidelines for primary prevention provide a grade B recommendation for aspirin use to prevent vascular events in a very circumscribed group of patients (age 50-59, estimated 10-yr ASCVD risk > 10%) at particularly high risk for vascular events but at low risk for bleeding due to younger age.

Patients with diabetes, however, are at 2-3 times greater risk of vascular events than patients without diabetes, raising interest in whether aspirin may be beneficial as primary prevention specifically in this population. Two randomized controlled trials were conducted to study aspirin as primary prevention for patients with diabetes. The POPADAD and JPPP trials showed no reduction in all-cause mortality or cardiovascular events, and did not report detailed information regarding bleeding events. Furthermore, more recent evidence suggests that aspirin use may be associated with a reduced incidence of cancers (particularly gastrointestinal cancers), which may add to the possible benefit of aspirin use as a primary prevention therapy. Thus, a subsequent, larger RCT was undertaken in patients with diabetes to better assess for the effect of primary prevention aspirin on vascular events and bleeding, as well as possibly cancer.

The 2018 A Study of Cardiovascular Events in Diabetes (ASCEND) trial randomized 15,480 patients with diabetes to primary prevention aspirin or placebo and assessed for a primary outcome of serious vascular events. During mean follow-up 7.4 years, there was a 1.1% absolute reduction in serious vascular events in patients randomized to aspirin therapy (NNT=91 at 7.4 years). Aspirin therapy was also associated with a 0.9% absolute increase in major bleeding (NNH=112 at 7.4 years). There was no significant difference between groups in the incidence of cancer or gastrointestinal cancer. Notably, diabetes was well-controlled in the majority of patients (mean hgb a1c < 8 in 80% of patients at enrollment) and the majority of patients were on a statin at baseline.

In summary, the ASCEND trial suggests that aspirin may have a slight benefit in reducing vascular events in patients with well-controlled diabetes, although this is at the cost of an increase in major bleeding of similar magnitude. Previous data suggesting improved cancer-related morbidity/mortality were not re-demonstrated, although these effects may not be observed until over 10 years of follow-up, which is ongoing. In 2019, AHA/ACC guidelines were updated to recommend against blanket use of aspirin (see Guidelines section, following).

Guidelines

ACC/AHA Primary Prevention of ASCVD (2019, adapted)

• Low-dose aspirin (75-100 mg po qday) for primary prevention of ASCVD:
  • Consider use if 40-70 years old with elevated ASCVD risk and no increased risk of bleeding (LOE A, COR IIb)
  • Avoid use if age ≥70 years (LOE B-R, COR III/harm)
  • Avoid use at any age if increased risk of bleeding (LOE C-LD; COR III/harm)

Design

• Multicenter, double-blind, randomized, controlled trial
• N=15480
  • Aspirin (N=7740)
  • Placebo (N=7740)
• Setting: United Kingdom
• Enrollment: June 2005 - July 2011
• Duration of follow-up: 30 months
• Analysis: Intention-to-treat
• Primary Efficacy Outcome: Serious vascular event (nonfatal MI, nonfatal stroke, transient ischemic attack, or death from any vascular cause)
• Primary Safety Outcome: Any major bleeding (intracranial hemorrhage, sight-threatening eye bleeding, gastrointestinal bleeding, any other bleeding requiring hospitalization, transfusion, or death)

Population

Inclusion Criteria

• Age ≥ 40 years
• Diagnosis of diabetes
• No known cardiovascular disease

Exclusion Criteria

• Clear indication for aspirin
• Contraindication to aspirin
• Presence of other clinically significant conditions that might limit adherence to the trial regimen for at least 5 years

Baseline Characteristics

From the placebo group.

• Demographics: Age 63.3 years, male 62.5%, white 96.5%,
• Comorbidities: BMI 30.6, current smoker 8.3%, former smoker 45.5%, HTN 61.6%, systolic BP 136.2
• Diabetes: type 2 DM 94.1%, duration of diabetes 7 years, hgb a1c > 8 12%, hgb a1c 6.5-8 30%, hgb a1c < 6.5 21%, hgb a1c unknown 37%
• Vascular risk score: low 40.5%, moderate 42.0%, high 17.4%
• Medications: aspirin 35.8%, statin 74.9%

Interventions

• Patients randomized 1:1 to aspirin or placebo
  • Aspirin 100mg (N=7740)
  • Placebo (N=7740)
• Potential participants were identified from regional diabetes centers or general practices from around the UK and were sent a questionnaire
• Eligible participants were sent placebo aspirin for a prerandomization run-in period, their general practitioners were notified of study involvement, and participants were sent a kit to obtain blood and urine samples and to record blood pressure, height, and weight
• After the 8-10 week run-in, participants remained eligible if they returned a randomization questionnaire confirming their willingness to continue, they still met eligibility criteria, and they had adhered to the trial regimen
• Using minimized randomization, patients were randomized to receive 100mg aspirin or matching placebo
  • Participants were also assigned 1g capsules of n-3 fatty acid or matching placebo as part of a separate RCT
• After randomization, participants were sent follow-up questionnaires and appropriate tablets and capsules every 6 months until the end of the trial
• Questionnaires sought information regarding all serious adverse events (including potential trial outcomes), adherence to the trial regimen, use of nontrial antiplatelet or anticoagulant therapy, nonserious adverse events resulting in discontinuation of the trial regimen, and any symptomatic bleeding episodes for which patients saw a doctor
• After mean follow-up 2.5 years, participants were sent blood and urine sample collection kits, along with measures of blood pressure and weight, in 1,800 randomly selected participants
• All reports of possible primary or secondary outcomes were adjudicated centrally by clinicians who were unaware of trial group assignment

Outcomes

Comparisons are aspirin versus placebo

Primary Efficacy Outcomes

Serious vascular event (composite of vascular death, non-fatal MI, non-fatal stroke and TIA)

658 (8.5%) vs. 743 (9.6%); HR 0.88 (95% CI 0.79-0.97); p = 0.01

ARR 1.10%; NNT 92

Vascular death

197 (2.55%) vs. 217 (2.80%); HR 0.91 (0.75-1.10)

Non-fatal myocardial infarction

191 (2.47%) vs. 195 (2.52%); HR 0.98 (0.80-1.19)

Non-fatal stroke

202 (2.61%) vs. 229 (2.96%); HR 0.88 (0.73-1.06)

Transient ischemic attack (TIA)

168 (2.17%) vs. 197 (2.55%); HR 0.85 (0.69-1.04)

Primary Safety Outcomes

Major bleeding

314 (4.1%) vs. 245 (3.2%); HR 1.29 (95% CI 1.09-1.52); p = 0.003

ARI 0.89%; NNH 112

Secondary Outcomes

Any serious vascular event or revascularization

833 (10.8%) vs. 936 (12.1%); HR 0.88 (95% CI 0.80-0.97)

Gastrointestinal tract cancer

157 (2.0%) vs. 158 (2.0%); HR 0.99 (95% CI 0.80-1.24)

Any cancer

897 (11.6%) vs. 887 (11.5%); HR 1.01 (95% CI 0.92-1.11)

Criticisms

• Use of 8-10 week run-in period prior to enrollment (with exclusion for nonadherent participants) introduces selection bias and limits generalizability
• Diabetes was well-controlled in the majority of patients (mean hgb a1c < 8 in 80% of patients at enrollment) and the majority of patients were on a statin at baseline. Thus, results may not be generalizable to patients with more severe diabetes, who would be expected to have a higher risk for vascular events
• Despite the fact that nearly half of the major bleeding events observed were gastrointestinal bleeds, only 25% of patients were prescribed proton pump inhibitors (PPI) by the end of the trial. Excess bleeding due to aspirin may have been mitigated by more routine PPI use.

Funding

• Study funded by the British Heart Foundation
• Study drug provided by Bayer
• Bayer provided comment on the manuscript draft but played no other role in the study
• Authors with multiple ties to industry