Friday October 22 2010
Taking daily aspirin has risks, especially bleeding complications
Research suggests that “a small daily dose of aspirin can significantly cut the risk of getting or dying from bowel cancer”, reported The Guardian.
This review gathered evidence from four large trials that compared taking daily aspirin for several years to placebo. As reported, aspirin reduced the risk of getting or dying from colorectal cancer, compared to placebo. The study also indicated that taking a low dose (75 to 300mg daily) was as beneficial as a high dose.
However, for individuals, the absolute reduction in cancer risk was quite small (the absolute risk of colorectal cancer was reduced from approximately 4% to approximately 2.5%). Regularly taking aspirin is known to increase the risk of internal bleeding, particularly in the elderly. As the review did not look at whether the people in these trials developed bleeding, we can't judge whether the potential benefits outweigh the potential harms.
Aspirin is known to benefit people who are at risk of cardiovascular disease as it thins the blood. However, it is still unclear if healthy people should take it as a preventative medicine.
Where did the story come from?
The study was carried out by researchers from the University of Oxford, the Karolinska Institute and other academic institutions in Sweden, the Netherlands and the UK. Individual researchers received honoraria from several pharmaceutical companies with an interest in antiplatelet agents. The cost of cancer registry and death certificate follow-up of the UK-TIA Aspirin Trial was met by unrestricted research funds from the Stroke Prevention Research Unit, Oxford, UK.
The study was published in the peer-reviewed medical journal The Lancet.
Generally, the newspapers have reflected the findings of this study accurately, though the study itself does not recommend that people start taking aspirin to protect against colon cancer.
What kind of research was this?
Worldwide, there are about 1 million cases of colorectal cancer every year, with 600,000 deaths. Two previous trials of high-dose aspirin (more than 500mg daily) have indicated that it might reduce the risk of colorectal cancer. However, taking high-dose aspirin daily is often not feasible in the long term because of the high risk of bleeding complications.
This review compiled the data from these two earlier trials, as well as data from three large, long-term trials of low-dose aspirin (75 to 300mg daily). The researchers aimed to establish how the dose and duration of aspirin treatment affect the development of colorectal cancer and deaths.
This review cannot be categorised as a systematic review as it does not appear to have conducted a search across the global literature, but instead identified trials from the UK or Sweden during the 1980s and 90s. These countries were chosen as they both had centralised death certification and cancer registration, making it possible to follow these outcomes. It is unclear if there are other unidentified trials relevant to the question of how aspirin affects cancer risk.
What did the research involve?
Eligible trials came from the UK and Sweden, each of which treated at least 1,000 people with aspirin for a minimum of 2.5 years, and compared them to a control group who did not. Four trials met this criteria, two of which used aspirin for the primary prevention of vascular events, and two used aspirin as a secondary prevention in people who had already suffered a vascular event (such as a stroke or heart attack).
The two trials of primary prevention were:
- Thrombosis Prevention Trial (TPT). This trial compared both aspirin and warfarin against placebo in men aged 45-69 who were at increased cardiovascular risk. The current study assessed only the aspirin aspect of the trial. A total of 5,085 high-risk men were recruited to this arm of the trial. Of these men, 2,545 were allocated to 75mg of aspirin a day, and 2,540 allocated to inactive placebo. Treatment lasted an average of 6.9 years, with all notifications of cancer development or death (flagged through the UK NHS Central Register) obtained over a further 12 years of follow-up.
- British Doctor’s Aspirin Trial (BDAT). This trial randomised 5,139 healthy male doctors (average age 61 years) to either high-dose aspirin (500mg daily, which could be reduced to 300mg upon later request) or no treatment (i.e. an inactive placebo tablet was not given). Treatment was for an average of six years. Participants were flagged with the National Cancer Registry and the Office of the Registrar General to identify all cancers and deaths up to 2001 (average 17 years later).
The two trials of secondary prevention were:
- Swedish Aspirin Low Dose Trial (SALT). This trial randomised people (average age 66) who had suffered a stroke or transient ischaemic attack (TIA or mini-stroke) within the past four months: 676 were allocated to 75mg of aspirin a day and 684 to placebo. The trial duration was for an average 2.7 years. The participants were flagged through the Swedish Medical Board for deaths for another 17 years (1990 through to 2007). Therefore, only fatal cancers could be identified.
- In the UK-TIA trial, 2,449 people over the age of 40 who had suffered a stroke or transient ischaemic attack were randomised to either high-dose aspirin (1,200mg a day), low-dose aspirin (300mg a day) or inactive placebo. The average treatment period was 4.4 years. Deaths and cancers occurring up to 20 years later were identified using national registries.
The researchers pooled patient data on the four trials of aspirin versus placebo to look at how aspirin affected outcomes of colorectal cancer and deaths due to cancer. They stratified their results by dose of aspirin.
Also included was one additional Dutch TIA Aspirin Trial, which had examined the long-term effect of a varying dose of aspirin (no inactive control group). This trial randomised 3,131 patients (average age 65) who had suffered a stroke or mini-stroke in the past three months to either 30mg of aspirin or 283mg aspirin a day. The average duration of treatment was 2.6 years, with further follow-up of an additional 10-13 years.
What were the basic results?
The four trials of aspirin versus control had a combined average treatment period of six years, with cancer and death follow-up over an average 18.3 years. Over the follow-up time, 391 of 14,033 patients (2.8%) developed colorectal cancer. Aspirin at any dose reduced the 20-year risk of dying from colorectal cancer by 34% (odds ratio 0.66, 95% CI 0.51 to 0.85). Aspirin reduced the risk of developing colon cancer by 24% (HR 0.76, 0.60 to 0.96) and the risk of dying from colon cancer by 35% (HR 0.65, 0.48 to 0.88). Aspirin had no significant effect on the risk of developing rectal cancer. Subgroup analysis also revealed that aspirin reduced the risk of developing cancer of the first section of the colon and of dying from this cancer, but aspirin did not affect the risk of developing or dying from cancer of the lower part of the colon leading to the rectum.
The effect of aspirin increased with increasing duration of treatment. Taking aspirin for five or more years (compared to taking it for less than five years) reduced the risk of developing any colorectal cancer (HR 0.68, 0.54 to 0.87) and of dying from colorectal cancer (0.57, 0.42 to 0.78). Subgroup analysis revealed that taking aspirin for five or more years reduced the risk of cancer in the first part of the colon (HR 0.35, 0.20 to 0.63) and dying from this cancer (HR 0.24, 0.11 to 0.52); and also reduced the risk of developing rectal cancer (HR 0.58, 0•36 to 0•92) and dying from this cancer (HR 0.47, 0.26 to 0.87).
Pooled analysis of the four trials revealed that, when taken for five or more years, high-dose aspirin was no more effective than low-dose aspirin in reducing the risk of developing fatal colorectal cancer over the next 20 years.
How did the researchers interpret the results?
The researchers conclude that aspirin taken for several years at doses of at least 75mg a day reduces the long-term risk of developing colorectal cancer and dying from this cancer. The benefits were greatest for cancers of the first part of the colon (proximal colon).
This well-conducted research pooled the results of four large trials that investigated how aspirin or inactive placebo affect the development of colorectal cancer or deaths from this cancer over up to 20 years of follow-up. The review is thorough and has collected individual patient data from all of these trials. The trials themselves benefit from high and complete follow-up rates.
A couple of points to bear in mind:
- Although this was a well-conducted review, it does not appear to be systematic. The results of large trials with extended follow-up conducted in the UK, Sweden and the Netherlands were identified, but the explicit methods used to locate or appraise these trials are not given. It is unclear whether trials from other countries could have contributed to this research.
- None of the trials were originally designed to investigate how aspirin affects the risk of colorectal cancer. Assessing outcomes that were not a planned study outcome has less statistical reliability than those defined up front.
- Aspirin increases the risk of bleeding complications, particularly in the elderly. The research does not supply information on any adverse events associated with aspirin use, so it is difficult to judge how the reduced risk of colorectal cancer in these trials weighed against the risks of bleeding complications or stomach irritation in these people.
- The absolute risk of colorectal cancer was still relatively small, with only 2.8% of the patient population in these trials (391 of 14,033) developing cancer. In subgroup analyses of these cases by site of cancer or duration of aspirin use, the numbers become smaller still, which increases the possibility of chance findings in statistical comparisons. For example, although taking aspirin for more than five years reduced the risk of developing cancer of the first part of the colon or dying from this cancer, only 61 people taking aspirin for more than five years developed cancer at this site, and only 37 died from it.
- There is the possibility that a small number of cancers were included that were already developing at the start of the trial when aspirin was assigned. It would not be possible to say how aspirin affects the development of cancer in such cases.
- As the researchers acknowledge, it is possible that people assigned to aspirin could have had more invasive investigations during follow-up caused by adverse bleeding events. This could have led to camera investigations, which could in turn have led to earlier diagnosis of cancers or developing cancers, thereby reducing the cancer mortality risk.
As the researchers say, any person in these trials (average age 60) had about a 4% absolute risk of developing colorectal cancer during the next 20 years. This is consistent with the estimated lifetime risk of about 5% in the general population. These findings indicate that taking aspirin for more than five years reduces the risk of colorectal cancer developing or of death from this cancer, but the risks versus benefits for an otherwise healthy person need careful consideration.