Aspirin in primary prevention of cardiovascular disease in the elderly

While the use of aspirin for the secondary prevention of cardiovascular (CV) disease is undisputed,1 the decision to use aspirin for primary CV disease prevention is largely based on the balance among aspirin’s antiplatelet/anti-inflammatory properties,2 its side effects,3 and a proper assessment of individual baseline CV risk.4 The recent publication of three randomized controlled trials (RCTs)—comparing low-dose aspirin vs. placebo for primary CV disease prevention in different cohorts of subjects—has underlined a significant increase of bleeding risk not always counterbalanced by clinically relevant reductions of ischaemic events. One of these—ASPREE—enrolled individuals ≥70 years of age (≥65 if American Hispanic or Black) and showed no significant benefit of aspirin on disability-free survival (primary endpoint) nor on CV events (secondary endpoint),5 with an unexpected trend towards increased all-cause mortality.6 Three meta-analyses published thereafter7–9 have stratified individuals according to baseline risk of CV events or to diabetic status, but not according to age. Older adults are increasing in number worldwide; they are typically at higher risk of both ischaemic and bleeding events as compared to younger individuals. Evaluating the risk/benefit ratio of aspirin in primary CV disease prevention in the elderly in the context of all primary prevention trials would help interpret the unexpected findings of ASPREE and contribute in guiding clinical practice in this potentially fragile cohort. Herein, we perform a meta-analysis of RCTs of aspirin in primary CV disease prevention, focusing on the pre-specified subgroup of older subjects. Randomized controlled trials comparing aspirin vs. control (placebo or no aspirin) in adults without a prior history of CV disease, enrolling more than 500 individuals and reporting a pre-specified subgroup analysis by age, were included. Elderly definition ranged from >60 to >70 years old; eight trials had upper age cut-offs (at 80–85 years in five and 69–75 years in three). Details on aspirin dosing, duration of exposure/follow-up, and baseline comorbidities can be found in the references.7–9 Pre-specified ischaemic outcomes were trial defined major adverse CV events (MACE), including myocardial infarction, stroke, CV death, unstable or newly developed angina, transient ischaemic attack, non-fatal coronary event, revascularization, above ankle amputation for limb ischaemia, and critical aortic or peripheral arterial disease. Primary safety outcomes included fatal or life-threatening bleeding, intracranial haemorrhage, gastrointestinal bleeding, bleeding requiring transfusion or hospitalization, and any other serious bleeding. The Q Cochran test and Higgins I2 statistics (with I2 <25%, 25–50%, or >50% indicating low, moderate, or high) were used to determine heterogeneity among trials. Summary random effects risk ratios (RRs) with 95% confidence intervals (CIs) using the random effects model with inverse variance weighting were used when the degree of heterogeneity was high or moderate, whereas the fixed-effect model with Mantel–Haenszel was used when heterogeneity was low. Statistical analysis was performed using RevMan software version 5.3. Among 3584 records included in the initial research, 10 RCTs were selected involving a total of 59 873 individuals. The overall incidence of MACE in older adults did not differ significantly with aspirin vs control (RR = 0.95, 95% CI 0.86–1.05, P = 0.28, I2 = 48%), while the overall incidence of major bleeding increased with aspirin (RR = 1.37, 95% CI 1.19–1.57, P < 0.00001, I2 = 0%) (Figure 1). The absolute rates of MACE were 5.4% with aspirin vs 5.7% with control. The absolute rates of major bleeds were 2.8% with aspirin vs. 2.0% with control. Sensitivity analyses for MACE were performed for adults >70 years old (RR = 0.90, 95% CI 0.78–1.04, P = 0.16, I2 = 51%), for older non-diabetic adults (RR = 0.92, 95% CI 0.78–1.08, P = 0.29, I2 = 66%), for older diabetic adults (RR = 0.96, 95% CI 0.87–1.07, P = 0.51, I2 = 0%), and excluding the ASPREE trial (RR = 0.94, 95% CI 0.83–1.07, P = 0.35, I2 = 54%). Meta-analysis of risk ratios for major adverse cardiovascular events (MACE) and major bleeding in the older subgroups of trials comparing aspirin versus control for primary disease prevention. Meta-analysis of risk ratios for major adverse cardiovascular events (MACE) and major bleeding in the older subgroups of trials comparing aspirin versus control for primary disease prevention. Accepting the limitations of study rather than individual-level data, to the best of our knowledge, this subgroup meta-analysis is the first to pool data from RCTs testing the overall efficacy and safety of aspirin in the primary prevention of CV disease among older subjects. Older, compared to younger, individuals represent a challenge for contemporary medicine, given their increasing number, the high incidence of drug-related adverse effects and their higher CV risk. In the absence of further evidence, our findings support those of ASPREE5 and recent recommendations,10 i.e., to refrain from using aspirin on a routine basis for primary CV disease prevention in adults >70 years of age without known CV disease, given the enhanced bleeding risk in the absence of significant reductions in ischaemic event rates. Future individual participant data analyses should allow improved age-stratification, with focus on single significant outcomes, like all-cause mortality and intracranial haemorrhage. Conflict of interest: F.