Low-Dose Aspirin Use and the Incidence of Colorectal Cancer
Low-Dose Aspirin Use and the Incidence of Colorectal Cancer
In this hybrid study of a cohort with comparison group from an Asian population, regular use of low-dose aspirin (50–150 mg daily) for at least 3.5 years was associated with a 50% reduced risk of CRC in patients with risk factors of cardiovascular disease. This reduction in CRC risk was not obvious until 6 years of follow-up. To our knowledge, this is the first cohort study to show that long-term use of low-dose aspirin among Asia patients at high risk of cardiovascular disease may be an effective form of CRC prevention.
In this study, the overall incidence of the study group (n = 9925) was 160.45 per 100 000 person-years. The mean age of the study group was about 64 years old. When compared to the incidence of Taiwan general population with similar gender, age and calander years, the CRC incidence in this study is slightly higher than that in the general population, which may partly be attributed to more comorbidities such as diabetes mellitus and hyperlipidaemia. The absolute CRC incidence rate among patients with regular aspirin use, 79.42 per 100 000 person-years, was lower than sex- and age-matched Taiwan general population that may indicate the absolute magnitude of preventive effect on CRC by using low-dose aspirin. Previous observational studies of low-dose aspirin differ in many respects such as age, gender (men only or women only), less cardiovascular diseases and time periods, hence, it is difficult to compare their CRC incidence rates with ours.
Evidence from observational studies and randomised controlled trials strongly support the hypothesis that aspirin decreases CRC risk. Recently, a meta-analysis of observational studies and clinical trials published from 1950 to 2011 showed that daily aspirin use was associated with an approximately 20–51% reduction in risk across cohort studies, case-control studies, and trials. This wide range in efficacy might be due to the considerable diversity within and between studies, including heterogeneity of the target population and variation in aspirin dose. Aspirin is known for its protective effect against cardiovascular disease and has been a particularly attractive intervention for CRC chemoprevention, as these diseases share many common risk factors and the incidence rate of CRC increases sharply after age 50 years. In this study, we found that low-dose aspirin might have a protective effect among Taiwanese adults, many of whom had high risk of cardiovascular disease.
The lowest effective dose required for reduction in CRC risk has remained undefined for decades. Two large randomised controlled trials originally designed to investigate cardiovascular prevention – the Physicians' Health Study (PHS) and Women's Health Study (WHS) – failed to show a reduction in CRC incidence with low-dose aspirin prescribed on an alternate-day schedule, i.e. 325 mg or 100 mg every second day respectively. Several cohort studies that specifically investigated the association of low-dose aspirin (average dose 12–165 mg daily) with CRC incidence found no clear benefit in general populations. However, a case-control study by Din et al. found that cumulative daily use of 75 mg of aspirin for 5 years was associated with reduced risk of CRC. A meta-analysis of four randomised controlled trials, including two primary prevention trials (British Doctor's Trial and Thrombosis Prevention Trial) and two secondary prevention trials (Swedish Aspirin Low-dose Trial and UK-Transient Ischaemic Attack), revealed that 75 mg of aspirin taken daily for a mean duration of 5.8 years reduced 20-year risk of colon cancer by 24%, with increasing benefit for longer duration of treatment. A systematic comparison from observational studies vs. randomised trials showed that daily aspirin use for at least 5 years reduced the risk of CRC by 45% (odds ratio 0.55, 95% CI 0.41–0.76) in a pooled analysis of six randomised controlled trials, by 49% (odds ratio 0.51, 95% CI 0.34–0.76) in a nested case-control trial, and by 32% (odds ratio 0.68, 95% CI 0.52–0.9) in a cohort study. Our finding showing a 50% decrease in CRC incidence among adults taking an average daily dose of 100 mg for a mean duration of 8.7 years is consistent with the values of previous studies.
In Taiwan, aspirin can be purchased as an over-the-counter drug. Information on over-the-counter use was not recorded in the prescription database; however, the cost of regular low-dose aspirin, which is used almost exclusively for cardiovascular prophylaxis, is covered by the NHI programme. Any use of aspirin purchased by the subjects would likely be limited to short-term symptomatic relief.
In this study, non-aspirin NSAIDs were used in most patients in both groups (98%). However, these NSAIDs were for short-term use as the mean duration of 8.4 months. Only 3.5% patients have NSAIDs use for more than 3 years, therefore we were unable to draw any conclusions regarding the association of NSAIDs and CRC risk.
Data on the protective effect of statins alone in CRC are conflicting. Hoffmeister et al. reported that individual and joint use of statins and low-dose aspirin may reduce the risk of CRC. The effect was more evident for combined use after five or more years (odds ratio 0.38, 95% CI 0.15–0.97). In this study, 55.6% of the aspirin user group had used statin. We did an analysis to determine whether the effect of aspitin on CRC is related to statin. Participants were classified into 1 of 4 groups: group 1: neither aspirin nor statin use (n = 5138, 94 CRCs), group 2: statin only (n = 2802, 35 CRCs), group 3: aspirin alone (n = 904, 8 CRCs), group 4: combined use of aspirin and statin (n = 1081, 6 CRCs). The age-adjusted HR was 0.72 (95%CI 0.49–1.06) for group 2, 0.48 (95%CI 0.23–0.99) for group 3, and 0.33 (95%CI 0.14–0.75) for group 4 when using group 1 as a reference group. This result shows that aspirin use was independently associated with risk reduction in CRC.
Gastrointestinal bleeding is the most frequent adverse event associated with regular aspirin use, although most episodes are not life-threatening. The effect of aspirin on gastrointestinal bleeding is thought to be largely dose-related. Studies have noted a higher rate of severe gastrointestinal bleeding at higher doses (300–325 mg) as compared with lower doses (75–150 mg). However, some studies found no difference between these two doses in the risk of bleeding. In this study, the similar rate of severe gastrointestinal bleeding, defined as haemorrhage requiring oesophagogastroduodenoscopic examination and blood transfusion (8.4% among users vs. 8.8% among non-users, P = 0.5935) suggests that regular low-dose aspirin for patients with cardiovascular disease is a regimen with adequate safety. On the other hand, FOBT rate was significantly higher in user group (86.25%) than that in non-user group (84.35%) (P = 0.0351). However, the results of FOBT, either positive or negative, were not recorded in the database. As FOBT is a screening procedure and the definite diagnosis of CRC is mainly based on the endoscopy of large intestine with tissue biopsy, the impact of FOBT on the incidence of CRC is not significant.
The strengths of this study include the use of a nationwide population-based data set, use of a comprehensive prescription database rather than self-reported records (thereby minimising recall bias), linkage of administrative databases and use of a national cancer registry database that enabled us to trace prospectively the association between drug exposure and outcome.
However, some limitations of this study warrant mention. First, although we adjusted for several potential confounders, we lacked detailed information on family history of CRC, history of inflammatory bowel disease, smoking habits, alcohol use, physical activity, high-fat diet and body mass index, which have been previously associated with CRC. Second, our use of the prescription database did not permit confirmation of actual usage, as it was impossible to contact patients directly because of the anonymity of the records. The possibility of some degree of treatment noncompliance should be considered. Third, because information on drug prescriptions before 1996 was not available due to the National Health Insurance programme was launched in 1995 in Taiwan, cumulative dose of aspirin use was likely to be underestimated or previous aspirin therapy is not known. We minimised the possible bias by selecting patients with newly prescribed aspirin since 1998. In other words, the user group has no prescription of aspirin for 2 years before the starting date of the study. As the prescription pattern of aspirin is usually continuous and regular, especially for the population with risk of cardiovascular disease, it is less likely that the user group ever used the aspirin for a long time while withheld it for 2 years in our study. Fourth, we were unable to analyse the association of aspirin with site-specific CRC, due to the relatively small number of cases. Finally, there may be unmeasured factors that differ between the two study cohorts.
In conclusion, our findings suggest that regular use of low-dose aspirin for at least 3.5 years was associated with a 50% reduced risk of CRC in patients at high cardiovascular disease risk after adjustments for potential confounding factors like age, gender and underlying comorbidities. Furthermore large-scale randomised clinical trials are necessary to confirm these findings.
Discussion
In this hybrid study of a cohort with comparison group from an Asian population, regular use of low-dose aspirin (50–150 mg daily) for at least 3.5 years was associated with a 50% reduced risk of CRC in patients with risk factors of cardiovascular disease. This reduction in CRC risk was not obvious until 6 years of follow-up. To our knowledge, this is the first cohort study to show that long-term use of low-dose aspirin among Asia patients at high risk of cardiovascular disease may be an effective form of CRC prevention.
In this study, the overall incidence of the study group (n = 9925) was 160.45 per 100 000 person-years. The mean age of the study group was about 64 years old. When compared to the incidence of Taiwan general population with similar gender, age and calander years, the CRC incidence in this study is slightly higher than that in the general population, which may partly be attributed to more comorbidities such as diabetes mellitus and hyperlipidaemia. The absolute CRC incidence rate among patients with regular aspirin use, 79.42 per 100 000 person-years, was lower than sex- and age-matched Taiwan general population that may indicate the absolute magnitude of preventive effect on CRC by using low-dose aspirin. Previous observational studies of low-dose aspirin differ in many respects such as age, gender (men only or women only), less cardiovascular diseases and time periods, hence, it is difficult to compare their CRC incidence rates with ours.
Evidence from observational studies and randomised controlled trials strongly support the hypothesis that aspirin decreases CRC risk. Recently, a meta-analysis of observational studies and clinical trials published from 1950 to 2011 showed that daily aspirin use was associated with an approximately 20–51% reduction in risk across cohort studies, case-control studies, and trials. This wide range in efficacy might be due to the considerable diversity within and between studies, including heterogeneity of the target population and variation in aspirin dose. Aspirin is known for its protective effect against cardiovascular disease and has been a particularly attractive intervention for CRC chemoprevention, as these diseases share many common risk factors and the incidence rate of CRC increases sharply after age 50 years. In this study, we found that low-dose aspirin might have a protective effect among Taiwanese adults, many of whom had high risk of cardiovascular disease.
The lowest effective dose required for reduction in CRC risk has remained undefined for decades. Two large randomised controlled trials originally designed to investigate cardiovascular prevention – the Physicians' Health Study (PHS) and Women's Health Study (WHS) – failed to show a reduction in CRC incidence with low-dose aspirin prescribed on an alternate-day schedule, i.e. 325 mg or 100 mg every second day respectively. Several cohort studies that specifically investigated the association of low-dose aspirin (average dose 12–165 mg daily) with CRC incidence found no clear benefit in general populations. However, a case-control study by Din et al. found that cumulative daily use of 75 mg of aspirin for 5 years was associated with reduced risk of CRC. A meta-analysis of four randomised controlled trials, including two primary prevention trials (British Doctor's Trial and Thrombosis Prevention Trial) and two secondary prevention trials (Swedish Aspirin Low-dose Trial and UK-Transient Ischaemic Attack), revealed that 75 mg of aspirin taken daily for a mean duration of 5.8 years reduced 20-year risk of colon cancer by 24%, with increasing benefit for longer duration of treatment. A systematic comparison from observational studies vs. randomised trials showed that daily aspirin use for at least 5 years reduced the risk of CRC by 45% (odds ratio 0.55, 95% CI 0.41–0.76) in a pooled analysis of six randomised controlled trials, by 49% (odds ratio 0.51, 95% CI 0.34–0.76) in a nested case-control trial, and by 32% (odds ratio 0.68, 95% CI 0.52–0.9) in a cohort study. Our finding showing a 50% decrease in CRC incidence among adults taking an average daily dose of 100 mg for a mean duration of 8.7 years is consistent with the values of previous studies.
In Taiwan, aspirin can be purchased as an over-the-counter drug. Information on over-the-counter use was not recorded in the prescription database; however, the cost of regular low-dose aspirin, which is used almost exclusively for cardiovascular prophylaxis, is covered by the NHI programme. Any use of aspirin purchased by the subjects would likely be limited to short-term symptomatic relief.
In this study, non-aspirin NSAIDs were used in most patients in both groups (98%). However, these NSAIDs were for short-term use as the mean duration of 8.4 months. Only 3.5% patients have NSAIDs use for more than 3 years, therefore we were unable to draw any conclusions regarding the association of NSAIDs and CRC risk.
Data on the protective effect of statins alone in CRC are conflicting. Hoffmeister et al. reported that individual and joint use of statins and low-dose aspirin may reduce the risk of CRC. The effect was more evident for combined use after five or more years (odds ratio 0.38, 95% CI 0.15–0.97). In this study, 55.6% of the aspirin user group had used statin. We did an analysis to determine whether the effect of aspitin on CRC is related to statin. Participants were classified into 1 of 4 groups: group 1: neither aspirin nor statin use (n = 5138, 94 CRCs), group 2: statin only (n = 2802, 35 CRCs), group 3: aspirin alone (n = 904, 8 CRCs), group 4: combined use of aspirin and statin (n = 1081, 6 CRCs). The age-adjusted HR was 0.72 (95%CI 0.49–1.06) for group 2, 0.48 (95%CI 0.23–0.99) for group 3, and 0.33 (95%CI 0.14–0.75) for group 4 when using group 1 as a reference group. This result shows that aspirin use was independently associated with risk reduction in CRC.
Gastrointestinal bleeding is the most frequent adverse event associated with regular aspirin use, although most episodes are not life-threatening. The effect of aspirin on gastrointestinal bleeding is thought to be largely dose-related. Studies have noted a higher rate of severe gastrointestinal bleeding at higher doses (300–325 mg) as compared with lower doses (75–150 mg). However, some studies found no difference between these two doses in the risk of bleeding. In this study, the similar rate of severe gastrointestinal bleeding, defined as haemorrhage requiring oesophagogastroduodenoscopic examination and blood transfusion (8.4% among users vs. 8.8% among non-users, P = 0.5935) suggests that regular low-dose aspirin for patients with cardiovascular disease is a regimen with adequate safety. On the other hand, FOBT rate was significantly higher in user group (86.25%) than that in non-user group (84.35%) (P = 0.0351). However, the results of FOBT, either positive or negative, were not recorded in the database. As FOBT is a screening procedure and the definite diagnosis of CRC is mainly based on the endoscopy of large intestine with tissue biopsy, the impact of FOBT on the incidence of CRC is not significant.
The strengths of this study include the use of a nationwide population-based data set, use of a comprehensive prescription database rather than self-reported records (thereby minimising recall bias), linkage of administrative databases and use of a national cancer registry database that enabled us to trace prospectively the association between drug exposure and outcome.
However, some limitations of this study warrant mention. First, although we adjusted for several potential confounders, we lacked detailed information on family history of CRC, history of inflammatory bowel disease, smoking habits, alcohol use, physical activity, high-fat diet and body mass index, which have been previously associated with CRC. Second, our use of the prescription database did not permit confirmation of actual usage, as it was impossible to contact patients directly because of the anonymity of the records. The possibility of some degree of treatment noncompliance should be considered. Third, because information on drug prescriptions before 1996 was not available due to the National Health Insurance programme was launched in 1995 in Taiwan, cumulative dose of aspirin use was likely to be underestimated or previous aspirin therapy is not known. We minimised the possible bias by selecting patients with newly prescribed aspirin since 1998. In other words, the user group has no prescription of aspirin for 2 years before the starting date of the study. As the prescription pattern of aspirin is usually continuous and regular, especially for the population with risk of cardiovascular disease, it is less likely that the user group ever used the aspirin for a long time while withheld it for 2 years in our study. Fourth, we were unable to analyse the association of aspirin with site-specific CRC, due to the relatively small number of cases. Finally, there may be unmeasured factors that differ between the two study cohorts.
In conclusion, our findings suggest that regular use of low-dose aspirin for at least 3.5 years was associated with a 50% reduced risk of CRC in patients at high cardiovascular disease risk after adjustments for potential confounding factors like age, gender and underlying comorbidities. Furthermore large-scale randomised clinical trials are necessary to confirm these findings.
Source...