Tipping the Balance: Benefits and Risks of Aspirin in Chemoprevention of Colorectal Cancer
- First Online:
- Cite this article as:
- Thiagarajan, P. & Jankowski, J.A. Curr Colorectal Cancer Rep (2013) 9: 1. doi:10.1007/s11888-012-0151-4
- 134 Views
Convincing evidence now exists to support a clear role for aspirin in the chemoprevention of gastrointestinal malignancies, in particular colorectal cancer. Although much of the data comes from large-scale epidemiological studies, recent randomised controlled trials have added statistically robust evidence to corroborate a beneficial effect of aspirin in the paradigm of chemoprevention, providing a promising avenue for further research into its long-term benefits in this context. The current evidence favours low-dose aspirin: 75-300 mg for 10 years or more in patients at high risk of colorectal cancer is optimal. These putative benefits must be viewed in the light of aspirin’s well-established myriad of side effects, including gastrointestinal and intracerebral haemorrhage. The complex risk–benefit profile of aspirin in general and specific populations is considered in this review.
Worldwide, the rising burden of malignant disease presents a major public health challenge. Cancer is currently the third leading cause of death globally, with an estimated 12.7 million new diagnoses and 7.6 million deaths due to cancer in 2008 .Although mortality rates in Western nations have declined in recent years, the incidence of cancer cases and deaths is expected to double by 2050 . Most of this disease burden will impact on economically developing nations, and is due in part to increased longevity of the general population, as well as the widespread adoption of lifestyle and dietary habits facilitating carcinogenesis, such as smoking and obesity .
Gastrointestinal tumours dominate cancer demographics, with colorectal cancer (CRC) being the third commonest malignancy worldwide (1.24 million new cases diagnosed in 2008) and accounting for over 600,000 deaths per year [4–6]. Despite rapid advances in screening and treatment modalities, CRC still carries a substantial mortality rate, with a 5-year relative survival of 6.6 % for Duke’s stage D disease . With an anticipated upsurge in CRC diagnoses worldwide, a paradigm shift from treatment and screening towards primary chemoprevention has recently gained credence.
To date, aspirin has emerged as a strong candidate for chemoprevention in CRC, with a wealth of evidence supporting its role in this context. Although much data has been derived from epidemiological studies, recent evidence from randomised controlled trials (RCTs) has corroborated the initial findings of an inverse association between aspirin and CRC development. Indeed, long-term use of aspirin (more than 5 years) has been shown to reduce CRC incidence and mortality by 30-40 % .
The chief unanswered question is whether the beneficial properties of aspirin are sufficient in magnitude to justify its routine use in chemoprevention, given major side effects such as gastrointestinal and cerebral haemorrhage. With its widespread use as a cardioprotective agent, the side-effect profile and efficacy data of aspirin are well established, allowing accurate risk–benefit analyses to be undertaken when considering its role in chemoprevention .
History of Aspirin Use
Aspirin (acetylsalicylic acid) is considered to be the most widely consumed medicine in the world [10, 11]. Besides its well-established anti-inflammatory and cardioprotective roles, newer indications for aspirin are the subject of intense research, most notably in chemoprevention of gastrointestinal malignancies.
The analgesic properties of naturally occurring salicylates have been used for centuries, as extracts of the willow bark Salix alba were traditionally used as antipyretic and antirheumatic agents [12, 13]. In 1897, the acetylated form of salicylic acid was synthesised by Felix Hoffman at the Bayer Company in Germany, and was marketed as ‘Aspirin’ in 1899. However, it was not until 1971 that its mechanism of action of was elucidated by the British pharmacologist John Vane, who discovered that aspirin blocks prostaglandin biosynthesis through the irreversible inhibition of the enzyme cyclooxygenase (COX).
In platelets, blockade by aspirin of the constitutive enzyme COX-1 inhibits the conversion of arachidonic acid to prostaglandins and thence to thromboxane A2, a powerful mediator of platelet aggregation. Thus, by inhibiting platelet aggregation, aspirin reduces the development of occlusive vascular events . This discovery spawned several large, multicentre studies worldwide confirming the cardioprotective effects of aspirin and supporting its use in secondary prevention of myocardial infarction and stroke. The second International Study of Infarct Survival (ISIS-2) demonstrated that regular use of orally administered aspirin (160 mg/day) reduced vascular mortality in patients with myocardial infarction by 23 %, and resulted in a 50 % reduction in the incidence of nonfatal re-infarction or stroke . However it remains unclear whether aspirin provides any additional benefit when other cardioprotective agents are already prescribed such as beta-blockers and ACE inhibitors .
The first suggestion of a role for aspirin in cancer prevention arose from observations that neuraminidase-induced thrombocytopenia decreased rates of tumour metastases in mice . Furthermore, Bennett et al.  demonstrated that basal levels of prostaglandin E2 in human colorectal tumour tissue were higher than those in surrounding normal mucosa, thus prompting suggestions that aspirin may obviate colonic carcinogenesis via COX inhibition. Subsequently, consistent evidence supporting a strong chemoprotective effect of aspirin and other non-steroidal anti-inflammatory drugs (NSAIDs) has been garnered from both laboratory studies and epidemiological data. In one such example, Thun et al.  conducted a cohort study of over 600,000 patients who were followed up over 6 years, and demonstrated a significantly reduced incidence of CRC in subjects taking aspirin more than 15 times per month (relative risk 0.66); an effect which persisted even when the data were controlled for body mass index, family history, physical activity and diet.
This analysis supported the observation by Rosenberg et al.  that patients using NSAIDs on a regular basis were 50 % less likely than controls to develop colorectal malignancies.
More recently, RCTs have provided high-quality evidence to demonstrate an antineoplastic effect of aspirin in preventing recurrence of colorectal adenomas and reducing CRC incidence and mortality. Further research is needed to establish the optimal dose of aspirin required to provide benefit whilst mitigating harmful side effects, as well the ideal duration of aspirin use required for effective chemoprevention.
Benefits of Aspirin in Colon Cancer Prevention
Prevention in Colonic Adenomatous Disease
In a French RCT, Benamouzig et al.  randomly assigned 272 patients with a history of colonic adenomas to receive daily aspirin (160 or 300 mg) or placebo for 4 years. At 1 year, colonoscopic follow-up revealed that for adenomas greater than 5 mm in diameter, aspirin was an independent factor associated with lower recurrence rates (P = 0.01).
In another randomised, double-blind trial of over 1,000 patients with colorectal adenomas, low-dose aspirin resulted in a 19 % relative risk reduction in adenoma development . Indeed, patients in the low-dose aspirin group had a reduced risk of advanced lesions compared with the placebo group (40 % risk reduction). However, high-dose aspirin (325 mg) was associated with a non-significant reduction in adenoma recurrence risk (4 %). The reason for this observation is unclear, although Baron et al.  speculate that higher-dose aspirin may paradoxically promote colorectal carcinogenesis through oversuppression of protective prostanoids in the colonic mucosa. However the lack of a characteristic dose–response curve has caused some doubt to be expressed over the ‘proof of principle’ idea of aspirin chemoprevention [23–25].
Primary Prevention of Colorectal Cancer
Recent work by Flossmann et al.  has yielded new evidence regarding the role of aspirin in primary prevention of CRC. In 2007, the group published data from a 20-year follow-up of two large RCTs, demonstrating that daily use of 300 mg aspirin for 5 years or more significantly reduced the incidence of CRC (p = 0.002). However, this effect was observed only after a latency period of 10 years, and no consistent effect on CRC development was shown with lower doses or less frequent administration of aspirin.
Rothwell et al. [26•] subsequently followed up five RCTs over 20 years and reported that allocation to aspirin reduced the 20-year risk of CRC (p = 0.005). Importantly, no net benefit was shown for doses greater than 75 mg daily, although lower doses (i.e. 30 mg/day) carried a higher risk of fatal colonic cancer on long-term follow up (p = 0.15). Aspirin was also found to be most effective in prevention of proximal colonic neoplasms (70 % risk reduction at 5 years, p < 0.0001) and cancers of the rectum (p = 0.01) but not of the distal colon (p = 0.54). There was increased benefit if aspirin was taken for 10 years or more. The statistically robust finding of chemopreventive activity for aspirin at doses of 75 mg (with no increased benefit at higher doses) has favourable implications when considering its potential for long-term primary prophylaxis, as the side effects of aspirin are dose-dependent. However a word of caution is needed as there have been reports where the true effect size of the benefits has been questioned, with a more modest effect being suggested when the same data were reanalysed .
Hassan et al. [28•] recently reported that the addition of low-dose aspirin to regular colonoscopic screening may be effective in reducing the incidence of proximal colonic neoplasms: aspirin increased the CRC death prevention rate from 69 % to 81 % in an analysis of over 100,000 subjects when combined with screening colonoscopy. That the addition of aspirin reduced the financial burden of CRC care (and was associated with a lifetime mortality of 0.1 %) led the group to speculate that aspirin is a cost-effective measure in the primary prevention of CRC.
Effect on Metastatic Disease
One of the main findings from Rothwell’s group was that aspirin had a greater beneficial effect on cancer mortality than on incidence alone, which may reflect the ability of aspirin to inhibit the formation of distant metastases. This hypothesis is supported by the observation that a mortality benefit with regular aspirin use is observed after a delay of several years . A recent systematic review of evidence from both observational studies and RCTs demonstrated that regular aspirin use was associated with a decrease in the proportion of cancers with metastatic disease (odds ratio 0.69) [29•].
This effect has been corroborated in studies of other cancers in which aspirin has been shown to inhibit metastases. In breast cancer, for example, aspirin has been reported to induce the production of maspin, a serine protease inhibitor which in animal models has been reported to reduce growth and metastases of breast cancer, as well as limiting local invasion [30, 31].
Side Effects of Aspirin Use
Serious side effects of aspirin therapy include upper gastrointestinal bleeding, peptic ulcer disease and intracerebral haemorrhage. Other well-known side effects include hepatotoxicity, exacerbation of asthma and hypersensitivity reactions including urticaria and angio-oedema.
Data from the U.S. Preventive Service Task Forces estimates that for every 1,000 patients taking aspirin, two to four would develop major gastrointestinal bleeding and zero to two would develop haemorrhagic stroke . On the basis of these findings and evidence from observational studies and RCTs, the U.S. Preventive Service Task Forces  issued a recommendation statement in 2007 advising that the harm caused by aspirin outweighs its potential benefits in the primary prevention of CRC in individuals at average risk.
Aspirin risks versus benefits
Gastrointestinal haemorrhage is among the most feared complications of aspirin use. Other complications include peptic ulcer formation with the concomitant risks of perforation and death. Gastrointestinal toxicity associated with aspirin use is thought to occur via its inhibition of COX-1, thus blocking the biosynthesis of prostaglandins which protect the gastric mucosal lining, and conferring an increased bleeding risk. This is a dose-dependent effect, with higher concentrations of aspirin associated with a significantly increased risk of severe gastrointestinal side effects (p = 0.03 in one study) . Factors associated with an increased risk of aspirin-induced gastrointestinal side effects include higher dose, history of gastric or duodenal ulceration, previous upper gastrointestinal tract bleeding, age above 70 years, concomitant use of NSAIDs and coexisting H. pylori infection .
In a 2011 meta-analysis, Lanas et al. [37•] demonstrated that low-dose aspirin (75–325 mg/day) was associated with an increased risk of major gastrointestinal haemorrhage compared with placebo (odds ratio 1.55). However, in patients taking concomitant PPI medication, this risk was significantly reduced (odds ratio 0.34), implying that regular PPI co-administration may provide a means of reducing gastrointestinal toxicity. Indeed, there is good evidence from epidemiological and randomised studies to corroborate this assertion . One double-blind RCT demonstrated that aspirin in combination with esomeprazole was superior to clopidogrel alone in prevention of recurrent ulcer bleeding in patients with previously diagnosed peptic ulcers who were H. pylori negative (P = 0.001) .
In view of these potentially serious gastrointestinal complications, the minimum dose of aspirin required for effective CRC prevention is a crucial factor in determining its eligibility for chemoprevention on a large scale, especially in average-risk individuals. The Rothwell group’s findings therefore represent a significant step towards realising the paradigm of aspirin chemoprevention, although further research and formal risk–benefit analyses are warranted to confirm a net benefit of low-dose aspirin therapy in the long term .
The incidence of haemorrhagic stroke is increased in subjects taking aspirin for primary or secondary prophylaxis of cardiovascular events . In a recent population-based cohort study, De Berardis et al.  identified a significantly higher risk of major haemorrhagic events (gastrointestinal or cerebral bleeding) in individuals taking low-dose aspirin versus matched controls over a mean follow-up of 5 years (incidence rate ratio 1.55). Indeed, preceding aspirin use has been shown to be an independent predictor of mortality in individuals following primary intracranial haemorrhage, likely due to rapid haematoma enlargement .
In a comprehensive meta-analysis involving over 55,000 subjects, the summary relative risk of haemorrhagic stroke with aspirin use was 1.84, and the number of patients needed to be treated to cause one excess intracranial haemorrhage was 833 for 3 years [41, 44]. On the basis of these findings, the need to clinically evaluate individual patients in determining the risk–benefit profile of aspirin therapy becomes an increasingly vital measure in the appropriate selection of patients for chemoprevention.
Aspirin has been reported to increase the risk of troublesome epistaxis. In a recent case–control study of 10,241 patients in primary care practices in the UK, the relative risk of epistaxis was increased in patients taking aspirin regularly (relative risk 9.04) compared with controls . Indeed, epistaxis associated with aspirin therapy has been reported to be an independent risk factor for intracranial haemorrhage (P < 0.01) . Although aspirin-associated epistaxis is seldom fatal, the antiplatelet effects of aspirin last for 10 days and cannot be easily antagonised, unlike the effects of coumarin agents. Use of aspirin is also associated with a greater need for surgical intervention to terminate the epistaxis, and a higher recurrence rate, leading some clinicians to argue that use of aspirin should be restricted (especially in elderly populations) to those with atherosclerotic vascular disease .
Benefits and Risks of Aspirin in Specific Populations
Establishing the risk–benefit ratio of aspirin use in elderly populations presents a challenge to practising clinicians. Although elderly persons are more likely to suffer from atherosclerotic vascular disease (and hence benefit from the cardioprotective effects of aspirin), they are also more susceptible to adverse effects of aspirin, such as gastrointestinal haemorrhage, intracerebral bleeding and renal impairment [48, 49].
Administration of even low-dose aspirin (75 mg daily) has been associated with significant deterioration in renal function and reduced clearance of uric acid within a short time (P = 0.009 in one study). Advanced age has also been independently linked to aspirin-associated upper gastrointestinal tract haemorrhage . In view of this evidence, use of NSAIDs as a drug class is generally avoided in elderly patients unless a strong indication for their use exists, such as ischaemic heart disease. If aspirin is required for cardioprotection (in which case it may have a more favourable risk–benefit profile), lower doses are preferred and regular monitoring of renal function is required in order to minimise adverse reactions and detect renal impairment at an early stage. With regard to primary prevention of colorectal malignancy, the risk–benefit balance is equally precarious in elderly individuals, as the incidence of CRC rises with age, and thus the potential benefits of aspirin therapy escalate in concert with its increasingly hazardous side effects.
The risk profile of aspirin in non-Western populations is less well studied. Factors which may influence this balance include a higher prevalence of H. pylori colonisation in developing nations, which could result in increased risk of adverse gastrointestinal consequences .
The paucity of population-specific data regarding aspirin use in non-Western nations is being addressed through randomised trials. In Japan, for example, an estimated 30 % of all deaths are due to cardiovascular or cerebrovascular disease, leading to calls for high-quality evidence to support use of an appropriate preventive agent . In 2005, enrolment began into the Japanese Primary Prevention Project, a large, open-label, multicentre randomised trial seeking to evaluate the risks and benefits of aspirin use compared with placebo in over 14,000 Japanese patients aged 60–85 years with vascular risk factors . The results of this trial are keenly awaited, and will guide clinical practice in the use of aspirin for cardioprotection in the Japanese population. How far the results will extrapolate to the benefit of aspirin in CRC prevention is unclear, although the major adverse effects are likely to be applicable in this context.
Inherited Colorectal Cancer Syndromes
Inherited syndromes such as familial adenomatous polyposis and Lynch syndrome (hereditary non-polyposis colon cancer) confer a high lifetime risk of CRC (95 % and 80 %, respectively) [55, 56]. In such individuals, the net benefit of aspirin in primary prophylaxis may outweigh potential risks, and merits strong consideration. The role of aspirin in CRC prevention among individuals with Lynch syndrome has been evaluated in the recent Colorectal Adenoma/Carcinoma Prevention Programme 2 (CAPP2) study, a double-blind RCT by Burn et al. [57•]. In the trial, 861 carriers of Lynch syndrome were randomly allocated to receive aspirin (600 mg daily) or placebo, with development of CRC as the primary end point. Burn et al. demonstrated that incidence of CRC was significantly reduced in the aspirin group compared with the placebo group (incidence rate ratio 0.56, p = 0.05), and propose that aspirin is as effective as surveillance colonoscopy in reducing the incidence of CRC.
The results of the CAPP2 trial strengthen the argument in favour of aspirin use in individuals with inherited cancer syndromes. The recently initiated CaPP3 trial aims to shed light on the critical questions of optimal dosage and duration of aspirin use required for effective chemoprevention in Lynch syndrome . Recently a potential biomarker of response has been reported—a PIK3CA mutation. However, this will need to be validated in other cohorts .
Aspirin in Other Gastrointestinal Areas
Barrett’s oesophagus, like the colon, has been actively studied to see if aspirin chemoprevention could prevent cancer. To date, the Aspirin Esomeprazole Chemoprevention Trial (AspECT), which is still in follow-up, has not revealed any significant benefit from low-dose aspirin after a mean of 4.5 years’ follow-up (personal communication). Perhaps more importantly a huge task force has recently reported that there is no evidence to support the use of aspirin in even advanced premalignant disease with high-grade dysplasia .
The emergence of aspirin as an effective chemopreventive agent in colorectal malignancy is now backed by compelling evidence from both observational studies and RCTs. That aspirin significantly decreases the incidence of, and mortality from, CRC is unequivocal. However, major side effects such as gastrointestinal and intracranial haemorrhage present a drawback to its universal uptake. Indeed, some authorities recommend against use of aspirin for primary prevention of CRC in average-risk individuals, in view of its complex risk–benefit profile, myriad deleterious effects and the need to establish the optimal dosage and duration of therapy .
Recent evidence has shed light on some of these important questions, paving the way for further studies seeking to optimise the risk–benefit ratio of aspirin use in order to justify its ubiquitous administration on a routine basis for chemoprevention. The demonstration that low doses of aspirin (75 mg daily) are effective may present a stronger argument for its widespread use in CRC prevention, as the side-effect profile is likely to be less onerous with such doses. In addition, recent evidence suggests that risk of gastrointestinal bleeding does not increase with longer duration of aspirin use, and may be in fact maximal upon initiation of aspirin therapy .
These promising findings could tip the balance in favour of aspirin use in selected populations for CRC prevention, although the practising clinician must be judicious in weighing the risks and benefits of long-term aspirin therapy on an individual basis. Certainly in patients with documented cardiovascular disease, in whom antiplatelet therapy is required as secondary prophylaxis against occlusive vascular events, the beneficial effects of aspirin on the colon may lead the clinician to favour its use as a first-line agent.
Further lines of research include the establishment of aspirin-related toxicity in non-Caucasian populations (in whom the incidence of CRC is rapidly rising), as well as additional RCTs to assess the effect of aspirin with CRC development as the primary end point as well an optimal dose, duration of therapy and when to start and stop therapy . In addition, understanding inherited susceptibility to aspirin responsiveness or resistance is key to the process. In this regard, the AspECT trial cohort is being probed for the genes involved .
Although current clinical guidelines for the prevention of colorectal malignancy do not recommend routine use of aspirin, it is hoped that they may be revised in the light of promising new evidence from RCTs.
J.A. Jankowski is supported by a grant from AstraZeneca.
P. Thiagarajan: none; J.A. Jankowski: served as Chief Investigator of the AspECT trial, received compensation from both AstraZeneca and Dr. Falk Pharmaceuticals for serving as a consultant.
Cancer Research UK
Conflict of interest
J.A J Jankowski Chief Investigator of AspECT trial and received funds from AstraZeneca.