Background: Prophylactic aspirin has been considered to be beneficial in reducing the risks of heart disease and cancer. However, potential benefits must be balanced against the possible harm from side effects, such as bleeding and gastrointestinal (GI) symptoms. It is particularly important to know the risk of side effects when aspirin is used as primary prevention - that is when used by people as yet free of, but at risk of developing, cardiovascular disease (CVD) or cancer. In this report we aim to identify and re-analyse randomised controlled trials (RCTs), systematic reviews and meta-analyses to summarise the current scientific evidence with a focus on possible harms of prophylactic aspirin in primary prevention of CVD and cancer. Objectives: To identify RCTs, systematic reviews and meta-analyses of RCTs of the prophylactic use of aspirin in primary prevention of CVD or cancer. To undertake a quality assessment of identified systematic reviews and meta-analyses using meta-analysis to investigate study-level effects on estimates of benefits and risks of adverse events; cumulative meta-analysis; exploratory multivariable meta-regression; and to quantify relative and absolute risks and benefits. Methods: We identified RCTs, meta-analyses and systematic reviews, and searched electronic bibliographic databases (from 2008 September 2012) including MEDLINE, Cochrane Central Register of Controlled Trials, Database of Abstracts of Reviews of Effects, NHS Centre for Reviews and Dissemination, and Science Citation Index. We limited searches to publications since 2008, based on timing of the most recent comprehensive systematic reviews. Results: In total, 2572 potentially relevant papers were identified and 27 met the inclusion criteria. Benefits of aspirin ranged from 6% reduction in relative risk (RR) for all-cause mortality [RR 0.94, 95% confidence interval (CI) 0.88 to 1.00] and 10% reduction in major cardiovascular events (MCEs) (RR 0.90, 95% CI 0.85 to 0.96) to a reduction in total coronary heart disease (CHD) of 15% (RR 0.85, 95% CI 0.69 to 1.06). Reported pooled odds ratios (ORs) for total cancer mortality ranged between 0.76 (95% CI 0.66 to 0.88) and 0.93 (95% CI 0.84 to 1.03). Inclusion of the Women's Health Study changed the estimated OR to 0.82 (95% CI 0.69 to 0.97). Aspirin reduced reported colorectal cancer (CRC) incidence (OR 0.66, 95% CI 0.90 to 1.02). However, including studies in which aspirin was given every other day raised the OR to 0.91 (95% CI 0.74 to 1.11). Reported cancer benefits appeared approximately 5 years from start of treatment. Calculation of absolute effects per 100,000 patient-years of follow-up showed reductions ranging from 33 to 46 deaths (all-cause mortality), 60-84 MCEs and 47-64 incidents of CHD and a possible avoidance of 34 deaths from CRC. Reported increased RRs of adverse events from aspirin use were 37% for GI bleeding (RR 1.37, 95% CI 1.15 to 1.62), between 54% (RR 1.54, 95% CI 1.30 to 1.82) and 62% (RR 1.62, 95% CI 1.31 to 2.00) for major bleeds, and between 32% (RR 1.32, 95% CI 1.00 to 1.74) and 38% (RR 1.38, 95% CI 1.01 to 1.82) for haemorrhagic stroke. Pooled estimates of increased RR for bleeding remained stable across trials conducted over several decades. Estimates of absolute rates of harm from aspirin use, per 100,000 patient-years of follow-up, were 99-178 for non-trivial bleeds, 46-49 for major bleeds, 68-117 for GI bleeds and 8-10 for haemorrhagic stroke. Meta-analyses aimed at judging risk of bleed according to sex and in individuals with diabetes were insufficiently powered for firm conclusions to be drawn. Limitations: Searches were date limited to 2008 because of the intense interest that this subject has generated and the cataloguing of all primary research in so many previous systematic reviews. A further limitation was our potential over-reliance on study-level systematic reviews in which the person-years of follow-up were not accurately ascertainable. However, estimates of number of events averted or incurred through aspirin use calculated from data in study-level meta-analyses did not differ substantially from estimates based on individual patient data-level meta-analyses, for which person-years of follow-up were more accurate (although based on less-than-complete assemblies of currently available primary studies). Conclusions: We have found that there is a fine balance between benefits and risks from regular aspirin use in primary prevention of CVD. Effects on cancer prevention have a long lead time and are at present reliant on post hoc analyses. All absolute effects are relatively small compared with the burden of these diseases. Several potentially relevant ongoing trials will be completed between 2013 and 2019, which may clarify the extent of benefit of aspirin in reducing cancer incidence and mortality. Future research considerations include expanding the use of IPD meta-analysis of RCTs by pooling data from available studies and investigating the impact of different dose regimens on cardiovascular and cancer outcomes.
