ArticlesEffects of regular aspirin on long-term cancer incidence and metastasis: a systematic comparison of evidence from observational studies versus randomised trials
Introduction
Globally, there were 12·7 million new cancer cases and 7·6 million cancer deaths in 2008,1 with colorectal, lung, and breast cancer most common in the developed world and lung, stomach, and breast cancer most common in the developing world.1 Incidence of some cancers is projected to increase further,2 partly because of ageing populations, and so effective and inexpensive strategies for prevention of cancer are urgently needed in the developed world (5·6 million new cases in 2008) and the developing world (7·1 million cases). One strategy would be chemoprevention with aspirin in some risk groups. There has been laboratory evidence for many years that aspirin might reduce the risk of cancer or restrict spread via effects on platelet function,3, 4 inhibition of COX-2,5, 6 or via other proapoptotic effects.7, 8, 9, 10 Building on previous evidence from observational studies,11 long-term follow-up of randomised trials of aspirin versus control with mean duration of at least 4 years has shown that daily aspirin reduces incidence and mortality due to colorectal cancer, after a latent period of about 8–10 years,12, 13, 14 and that it reduces total cancer mortality from 5 years to 20 years, with reductions in deaths due to several other cancers.12, 13 The accompanying analyses of all trials of aspirin versus control in prevention of vascular disease now also identify reductions in shorter-term incidence and mortality,15 due mainly to a reduction in metastasis.16
However, statistical power was inadequate in these recent analyses of randomised trials to establish effects on less common cancers and on cancers occurring only in women. More data for the effects of aspirin are urgently needed, but new randomised trials with follow-up of 10–20 years are impractical and cohort studies would also need a substantial period of follow-up. Case–control studies could be done much more quickly, but are of little use unless they can be shown to produce reliable estimates of the effects of aspirin. Many such studies have been reported during the past two decades,10, 11, 14 but the results have not been regarded as being sufficiently reliable to affect clinical guidelines.17, 18 Although observational studies can sometimes reliably identify causal associations, they have proven less reliable in establishing the effects of possible treatments such as vitamins or hormone replacement therapy in prevention of vascular disease.19, 20 Nevertheless, we showed previously that the effect of aspirin on risk of colorectal cancer identified during long-term follow-up of two large randomised trials was consistent with the associations reported in observational studies when stratified by frequency and duration of aspirin use.14, 21 We now aimed to assess the reliability of estimates of the effect of aspirin use on risk and outcome of all types of cancer in case–control and cohort studies. First, we updated our previous comparison of risk associations for colorectal cancer,14 including new data from randomised and observational studies.10, 11 Second, we compared risk associations from observational studies of all other common cancers with comparable data from our most recent follow-up of randomised trials.12, 13 Third, we also compared data from observational studies reporting the effects of aspirin on risk of metastasis with those from the accompanying reports from randomised trials.15, 16
Section snippets
Search strategy and selection criteria
We aimed to identify published observational studies of the association between use of aspirin or non-steroidal anti-inflammatory drugs (NSAIDs) and risk of all types of cancer. We searched PubMed, National Library of Medicine, from Jan 1, 1950, to Jan 15, 2011, with the following search terms: (“neoplasms”[MeSH Terms] OR cancer[Text word]) AND (aspirin OR salicyl* OR “anti-inflammatory agents, non-steroidal”[MeSH Terms] OR “anti-inflammatory agents, non-steroidal” [Pharmacological Action] OR
Results
We identified 13 728 articles with our electronic search (figure 1). Review of titles and abstracts identified 323 potentially eligible studies, of which 148 were found to report original data on review of the full papers. We found an additional 160 papers after checking review articles, reference lists, and other sources. Of these 308 articles, 60 were duplicates or were otherwise ineligible. Of the 248 reports remaining, 187 described a total of 150 case–control studies and 61 described 45
Discussion
There is good evidence from randomised trials that daily aspirin prevents colorectal polyps25 and reduces the long-term risk of sporadic colorectal cancer12, 14 and colorectal cancer in Lynch syndrome.41 This evidence is now sufficiently robust to use as a test of the reliability of associations in observational studies.42 We have now shown that case–control studies yield estimates of the effect of aspirin on risk of colorectal cancer that are in close agreement with those from the trials.
References (45)
- et al.
Anti-metastatic effect of aspirin
Lancet
(1972) Anti-metastatic action of non-steroidal anti-inflammatory drugs
Kaohsiung J Med Sci
(2008)- et al.
Aspirin, salicylates, and cancer
Lancet
(2009) - et al.
Long-term effect of aspirin on colorectal cancer incidence and mortality: 20-year follow-up of five randomised trials
Lancet
(2010) - et al.
Effect of daily aspirin on long-term risk of death due to cancer: analysis of individual patient data from randomised trials
Lancet
(2011) - et al.
Effect of aspirin on long-term risk of colorectal cancer: consistent evidence from randomised and observational studies
Lancet
(2007) - et al.
Aspirin and non-steroidal anti-inflammatory drugs for cancer prevention: an international consensus statement
Lancet Oncol
(2009) - et al.
Those confounded vitamins: what can we learn from the differences between observational versus randomised trial evidence?
Lancet
(2004) - et al.
Aspirin dose and duration of use and risk of colorectal cancer in men
Gastroenterology
(2008) - et al.
Risk of colorectal cancer in patients prescribed statins, nonsteroidal anti-inflammatory drugs, and cyclooxygenase-2 inhibitors: nested case-control study
Gastroenterology
(2007)