Cardiovascular disease (CVD) is the leading cause of death and colorectal cancer (CRC) is the third leading cause of death in the United States.
To systematically review evidence for the effectiveness of aspirin to prevent myocardial infarction (MI), stroke, cardiovascular death, and all-cause mortality in those without a history of CVD. In addition, to review evidence for CRC incidence and mortality associated with aspirin use in primary and secondary CVD populations. To further review harms associated with aspirin use.
We searched MEDLINE, PubMed, and the Cochrane Collaboration Registry of Controlled Trials to identify literature that was published between January 2014 and January 14, 2021. We supplemented our searches with reference lists from the previous review, relevant existing systematic reviews, suggestions from experts, and Clinicaltrials.gov to identify ongoing trials. We conducted ongoing surveillance for relevant literature through January 21, 2022.
Two investigators independently reviewed identified abstracts and full text articles against a set of a priori inclusion and quality criteria.
One investigator abstracted data into an evidence table and a second investigator checked these data. We conducted Peto fixed effects meta-analyses to estimate the effect size of aspirin in preventing MI, stroke, CVD-related death and all-cause mortality, CRC incidence and mortality, major bleeding, major gastrointestinal (GI) bleeding, intracranial bleeding, hemorrhagic stroke, and extracranial bleeding. Additionally, we conducted sensitivity analyses using Mantel-Haenszel fixed effects and Restricted Maximum Likelihood random effects.
We included 13 fair- to good-quality randomized, controlled trials (RCTs) (N=161,680) examining the effectiveness of aspirin for the primary prevention of CVD. Based on pooled analysis of 11 primary CVD prevention trials using aspirin ≤100 mg/day, low-dose aspirin reduces the risk of major CVD events (total MI, total stroke, CVD mortality) by 10 percent (k=11, N=134,470; Peto odds ratio [OR], 0.90 [95% confidence interval (CI), 0.85 to 0.95]), MI by 11 percent (k=11, N=134,470; Peto OR, 0.89 [95% CI, 0.82 to 0.96]), and ischemic stroke by 18 percent (k=5, N=79,334; Peto OR, 0.82 [95% CI, 0.72 to 0.92]) with no differences in CVD mortality (k=11, N=134,470; Peto OR, 0.95 [95% CI, 0.86 to 1.05]) or all-cause mortality (k=11, N=134,470; Peto OR 0.98 [95% CI, 0.93 to 1.03]). Absolute risk reductions in major CVD events in the trials ranged from 0.08 to 2.5 percent. Aspirin’s benefits were similar when trials of all doses were pooled. Sensitivity analyses restricted to more recent trials where usual care includes aggressive risk factor modification including statin therapy show diminished effects of aspirin for major CVD events and total MI but larger effects for total ischemic stroke compared to older trials. A small subset of the trials reporting CVD outcomes also reported CRC outcomes. Based on 4 low-dose aspirin trials (N=86,137) recruiting primary CVD prevention populations, there was no statistically significant association between aspirin and CRC incidence when analyzing randomized trial periods (Peto OR 1.07 [95% CI, 0.92 to 1.24]; trial period 5-10 years). Analysis including post-trial observation periods up to 20 years and including trials with high-dose aspirin up to 500 mg/day (k=2; N=45,015) in primary prevention populations show statistically significant reductions in CRC incidence (0.70 [95% CI, 0.50 to 0.98] and 0.82 [95% CI, 0.69 to 0.98]). Two low-dose aspirin RCTs (N=59,020) in primary CVD prevention populations report CRC mortality during the trial period (5-10 years) showing results concerning for possible harm with one trial demonstrating a statistically significant increase in CRC mortality in older adults. At 18 years of followup, including post-trial observational periods, three primary CVD prevention trials with mean daily aspirin doses ranging from 75 to 500 mg showed aspirin was associated with a decreased risk of CRC mortality (Peto OR 0.76 [95% CI, 0.62 to 0.94]). Low-dose aspirin is associated with a 31 percent increase in intracranial bleeding events (k=11; N=134,470; Peto OR, 1.31 [95% CI, 1.11 to 1.54]), and 53 percent increase in extracranial bleeding events (k=10; N=133,194; Peto OR 1.53 [95% CI, 1.39 to 1.70]). The absolute increases ranged from −0.2 to 0.4 percent for intracranial bleeding events and 0.2 to 0.9 percent for extracranial bleeding events. There is no compelling evidence to suggest that aspirin has a different relative CVD benefit or bleeding risk in specific populations defined by age, sex, race and ethnicity, diabetes status, or baseline 10-year CVD risk. Aspirin’s CVD benefits appear to begin within the first 1-2 years of administration and the bleeding harms begin soon after aspirin initiation; there are limited data for more precise time increments or longer durations.
Primary CVD prevention trials used different aspirin doses in heterogeneous populations with relatively short study followup, with duration mostly ranging from 4-6 years. Trials reporting CRC incidence and mortality outcomes are limited by short trial duration and multiple comparisons; observational followup of trials are limited by heterogeneity of aspirin doses, duration, indications, and populations with risk of biases and confounding. Estimates of rare bleeding harms are imprecise.
In primary prevention populations, low-dose aspirin reduces major CVD events, MI and ischemic stroke, but also increases major GI bleeding, extracranial bleeding, and intracranial bleeding. Our evidence suggests aspirin is associated with a possible long-term reduction in CRC incidence and mortality based on post-trial period observation, but the results are limited for low-dose aspirin among primary CVD prevention populations. More precise real-world U.S.-based estimates for bleeding events in the general population and specific populations with elevated CVD risk are necessary to accurately estimate the net benefit. Depending on CVD risk, this absolute CVD benefit in specific populations could potentially outweigh the bleeding risks. Models to identify these populations are needed.
