Scientific Evidence
In 1974 the first randomised trial of aspirin in the reduction of vascular deaths was reported. That report went global and seems to have stimulated a global interest in aspirin, including an interest in aspirin and cancer.
Thirty years later the Welsh Aspirin Group commenced a carefully planned series of studies to investigate key issues relevant to aspirin and the treatment of cancer. The work was largely based on systematic literature searches, followed by meta-analyses of all the available published data relevant to each key issue. The following summarises these investigations:
1. THE SAFETY OF LOW-DOSE ASPIRIN
Bleeds caused by aspirin are much less serious than the bleeds which occur from stomach ulcers or a stomach infection and are rarely – if ever– fatal. A systematic literature search identified eleven randomised trials of aspirin within which bleeding, and fatal bleeds had been recorded over an average of 2.8 years.
Table GI bleeding in a meta-analysis of data from 11 RCTs
| Bleeding | Risk of a bleed (in 2.8 years) | Relative risk for aspirin |
| Incidence of a GI bleed – in 54,625 subjects randomised to aspirin – in 52,583 subjects randomised to placebo | 8 per1,000 5 per 1,000 | RR 1.55 (1.32, 1.83) |
| Proportion of bleeds that were fatal – in subjects on aspirin – in subjects on placebo | 4% 8% | RR = 0.45 (0.25, 0.80) |
| Risk of a fatal bleed in trial participants – randomised to aspirin – randomised to placebo | 3.7/10,000 4.7/10,000 | RR = 0.77 (0.41, 1.43) |
2. THE ‘early’ DIAGNOSIS OF CANCER, AND ASPIRIN
The definition ‘early’ depends largely upon whether or not the cancer has spread beyond the site of its origin to any other organ(s) of the body, This is of importance because the ‘satellite’ growths are responsible for much of the pain and complications of cancer, and many of the deaths are attributable to the metastases, rather than to the primary tumour.
SEE: Chatting to the King about cancer. British Medical Journal BMJ 2024;384:q342
Studies of patients with cancer have shown that perhaps around 10% of the deaths have been due to vascular disease, that is, heart attacks, ischaemic strokes and venous thrombosis (DVT). Aspirin has been shown repeatedly to reduce the risk of these thromboembolic complications.
A research study based upon all the available published observational studies of deaths in patients with cancer, showed that the survival of 250,000 patients who were taking aspirin was more than 20% longer than the survival of patients not taking aspirin. Furthermore, the average increase in survival with aspirin in a range of 15 less common cancers was similar to the survival associated with aspirin taking in the common cancers. Table: Aspirin taking and mortality: meta-analysis of results in 118 published observational reports. COMMENT: This evidence is from real life and it is highly impressive. But there could be bias. The most important is ‘publication bias’: that is – most researchers do all they can to get ‘positive’ papers published, while ‘negative ‘ papers tend to be dismissed and there is much less desire to publish them The statistician, Christine Delon, who worked on this study, conducted a detailed examination of publication bias using a programme called ‘Trim and Fill’. Detail findings are in ‘Supplementary File 6: following the published report (eCancer ttp://dx.doi.org/10.3332/ecancer.2021.1258). The effect is to reduce the associations with aspirin, but the relationship between aspirin and cancer mortality is still significant with ‘trim and fill’: (HR 0.85 (0.79, 0.91) and for all-cause mortality HR 0.94 (0.87, 1.02) Thus, although powerful… observational evidence has inherent uncertainty. GROUP Cancer mortality All-cause mortality No. of studies HRs (95% CIs) ORs (95% CIs) No. of studies HRs (95% CIs) ORs (95% CIs) Colon cancer 24 One HR 0,71 (0,62, 0.80) OR 0.78 (0.66, 0.93) 21 One HR 0.81 (0.73, 0.91) OR 0.78 (0.65, 0.92) Breast cancer 13 4 HR 0.84 (0.72, 0.98) OR 0.75 (0.36, 1.57) 9 None HR 0.94 0.70, 1.25) – Prostate cancer 15 One HR 0.89 (0.78, 1.02) OR 1.02 (0.78, 1.34) 6 One HR 1.00 (0.78, 1.27) OR 1.06 0.94, 1.19) 15 other cancers 18 5 HR 0.79 (0.70, 0.83) OR 0.49 (0.26, 0.95 21 5 HR 0.67 0.60, 0.75) OR 0.47 (0.26, 0.83) All 18 cancers 70 11 HR 0.77 (0.72, 0.83) OR 0.67 (0.45, 1.00) 56 7 HR 0.79 (0.74, 0.86) OR 0.57 (0.36, 0.89)
Aspirin has effect upon a wide range of cellular and the biological mechanisms in cancer, including the growth of cancer, its metastatic spread to other organs in the body, thrombotic complications of cancer, and cancer death. SEE: Aspirin and cancer: biological mechanisms and clinical outcomes’ Royal Society Open Biology 2022
A summary of the evidence in 2024 ‘for’ and ‘against’ the use of aspirin in patients with cancer.
In order to explore the attitudes by the public to medical research the Welsh Aspirin Group conducted a ‘citizens’ Jury’ under the title: ‘My Health – whose responsibility?’ Sixteen members of the general public in Cardiff served as ‘jurors’ and over three days there was lively debate about many issues… including aspirin as a possible treatment of cancer. The jury clearly wanted a greater involvement of the public in the evaluation of research findings, and their relevance to clinical practice and to public policy… and to all this they added the unanimous verdict: ‘even before there is agreement between doctors.’ SEE: My health: whose responsibility? A jury decides. J epidemiol Community Health. 2010;64(9):761
The situation with cancer in the poorer countries is clearly ethically unjust. About one in every six deaths worldwide is due to cancer, giving an estimated 9.6 million deaths in 2018, and around 70% of the deaths in low- and middle-income countries. WHO points out that most cancers in the poorer countries are diagnosed at a very late stage, when most treatments are no longer effective – even if treatments were available, which they are not in many countries.
There are two main objectives in medical research: These differ enormously, and to achieve success of failure in ones does not imply success in the other. while all the above relate clearly to the treatment of cancer, the evidence on prevention is inconsistent and further evidence is required on aspirin and prevention of cancer. Nevertheless. we understand that some oncologists have decided not to recommend aspirin to cancer patients aged over 70 yrs. SEE: Effect of aspirin in disability-free survival in the healthy elderly NEJM2018:379:1499-15088
A situation that urgently needs sorting out is the relevance of the mutation PIK3CA to the effect of aspirin on colon, and possibly other cancers. This mutation produces a protein that increases Cox-2 and prostaglandin activity, and has been shown to enhance the response of the tumour to aspirin. The population prevalence of this mutation is stated in several studies to be around 15–20% and in an overview of four studies, its presence appears to reduce cancer mortality attributable to aspirin by over 50% (HR 0,45; 0.28-0.71) and all-cause mortality by about 30%. SEE: Aspirin in the treatment of cancer: reductions in metastatic spread and in mortality: a systematic review and meta-analysis of published studies. Public Library of Science 2016. http://dx.doi.org/10.1371/journal.pone.0152402. In a recent randomised trial in PIK3CA-mutant colon cancer the primary endpoint was disease-free survival and aspirin was associated with a relative reduction of over 40% (HR 0.57). Due to financial constraints, the trial was prematurely closed. Overall, the main weakness in the body of evidence on aspirin and the treatment of cancer, is that very few randomised trials have been conducted so far, and the results of those which have reported are inconsistent. A most notable randomised trial by Sir John Burn showed that aspirin reduced the mortality of colon cancer in patients with the Lynch Syndrome, in which there is a defect in in DNA mismatch repair genes. This led NICE to recommend aspirin treatment of the syndrome – implying incidentally that aspirin is a relatively safe drug!. However, it has been estimated that the population prevalence of the syndrome is about 1 in every 300 individuals in the general population. In a recent interim report on the ‘ASCOT’ randomised trial, 1,550 patients with colon cancer had been randomised to aspirin or placebo. At 60 months follow-up 791 patients on aspirin had survived, compared with 759 on placebo (HR 0.91; 0.73, 1.13) and overall, 58 and 72 events in the aspirin and the placebo groups (HR 0.75; 0.53, 1.07) had occurred. Both these estimates of reduction could well have been due to chance. Another recent randomised trial of aspirin (300 mg) was conducted in 3020 participants who had high-risk nonmetastatic breast cancer (Chen et al. JAMA. 2024 May 28;331(20):1714-1721). 46% of the patients discontinued the trial treatment early in the trial, and in the remaining 54% compliance was estimated to be around 91%. At a median follow-up time of 34 months an ‘intention to treat’ analysis showed that aspirin was associated with a small and non-significant excess of ’invasive disease-free survival events’ (141 vs 112: HR 1.27; 0.99,1.63 SEE: JAMA 2024;331(20):1714-1721. Doi:10,1001/jama.2024.4840 An alternative to randoms allocation of treatment, is ‘Mendelian Randomisation’ based on the fact that a mutation which mimics the effect(s) of a drug such as aspirin occur at random within a populations. Several studies using this strategy have given answers favourable to aspirin treatment of cancer. SEE: Aspirin, salicylates and cancer. Lancet 2009; vol 373:9671:1301-9, Further randomised trial have been set up (‘ADD-Aspirin’; ‘ASPIRIN NCT02647099’; ‘0SAKK 41/13’; ‘ACAC’ and ‘EPISODE-111’) but it will be some years before any of these will have results to report. So, in the meantime…
….given the relative safety and the favourable effects of aspirin, its use in cancer treatment seems justified, and ethical implications of this imply that cancer patients should be informed of the present evidence and encouraged to raise the topic of aspirin with their healthcare team.
Low-dose aspirin is not an alternative to any other treatment. Patients with cancer should however raise the topic of aspirin with their healthcare team and enquire whether or not low-dose aspirin is likely to be helpful to them. The evidence presented in this website is drawn from studies of adult cancer patients.