If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Research, Cancer Registry of Norway, Oslo, NorwayOslo Centre for Biostatistics and Epidemiology, Department of Biostatistics, Institute of Basic Medical Sciences, University of Oslo, Oslo, Norway
Department of Research, Cancer Registry of Norway, Oslo, NorwayNorwegian Research Centre for Women's Health, Women's Clinic, Oslo University Hospital, Oslo, Norway
Section for Colorectal Cancer Screening, Cancer Registry of Norway, Oslo, NorwayDepartment of Research and Development, Telemark Hospital Trust, Skien, NorwayInstitute of Clinical Medicine, University of Oslo, Oslo, Norway
Section for Colorectal Cancer Screening, Cancer Registry of Norway, Oslo, NorwayDepartment of Medicine, Vestre Viken Hospital Trust Bærum, Gjettum, Norway
The possible protective effect of aspirin on risk of colorectal cancer (CRC) is still highly debated.
Methods
We used data from Bowel Cancer Screening in Norway, a trial randomizing individuals from general population, aged 50-74 years, to flexible sigmoidoscopy or faecal immunochemical test (FIT), to study the association between aspirin use and detection of CRC and two CRC precursors: adenomas and advanced serrated lesions (ASL). Prescriptions of low-dose aspirin were obtained from Norwegian prescription database. Logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs).
Results
Among 64,889 screening participants (24,159 sigmoidoscopy, 40,730 FIT), 314 (0.5%) had CRC, 6,208 (9.6%) adenoma and 659 (1.0%) ASL. Overall and short-term use (<3 years) of low-dose aspirin, versus no use, were not associated with any colorectal lesion. Long-term use (≥3 years) was associated with lower detection of CRC (overall OR 0.66, 95%CI 0.46-0.93; sigmoidoscopy: 0.56, 0.33-0.97; FIT: 0.72, 0.45-1.15), adenomas in sigmoidoscopy arm (overall OR 0.95, 95%CI 0.87-1.03; sigmoidoscopy: 0.89, 0.80-0.99; FIT: 1.03, 0.89-1.18), but not ASLs. We did not observe significant differences in the effect of aspirin according to the location of colorectal lesions.
Conclusion
Our results suggest that long-term use of aspirin might have a protective effect against adenomas and colorectal cancer, but not ASLs.
Annual report to the nation on the status of cancer, 1975-2006, featuring colorectal cancer trends and impact of interventions (risk factors, screening, and treatment) to reduce future rates.
Chemoprevention is an additional strategy that has been long studied to lower the incidence of CRC, with aspirin being the most studied agent for chemoprevention of CRC and its precursor lesions [
. In the last three decades, several observational studies have shown promising results on the association between aspirin use and primary prevention of colorectal adenomas [
. A possible explanation for these conflicting findings might be the large variations in the administration plans of aspirin in the different studies, including dosage and duration of use [
. Notably, while in 2016 the U.S. Preventive Services Task Force recommended the initiation of low-dose aspirin for CRC prevention in average-risk adults aged 50-59 years [
Only a few studies investigated aspirin use as primary prevention for CRC and colorectal polyps taking into account the histopathological and molecular features of these lesions. There is promising evidence of a protective effect of aspirin on adenomas [
, but the effect of aspirin on serrated lesions is largely unknown. Though one study found a protective effect of aspirin against large serrated polyps [
The possible protective effect of aspirin against colorectal neoplasia, optimal administration plan and mechanism of action remain highly debated. Moreover, the evidence of aspirin's effects according to the molecular and histopathological features, stage, and anatomical site of lesions is still unclear. In a large Norwegian CRC screening trial, we investigated the association between the use of aspirin and detection of CRC and colorectal polyps according to the lesions’ stage, site, and histopathological characteristics.
2. Methods
2.1 Design and population
This cross-sectional study uses data from the randomized pilot trial of Bowel Cancer Screening in Norway (BCSN) [
]. For the present study, we extracted data in 2017, before the enrolment in the sigmoidoscopy arm was completed, and included 24,159 participants who attended sigmoidoscopy and 40,930 who attended the first round of FIT between 2012 and 2017. Participants with a positive FIT (> 15 microgram haemoglobin per gram faeces) or positive sigmoidoscopy (any polyp ≥ 10 mm, ≥3 adenomas, any advanced adenomas, or CRC) were invited to attend a follow-up colonoscopy [
]. Two hundred FIT-positive attendees did not undergo a colonoscopy and were excluded from this study.
Sociodemographic data were retrieved from Statistics Norway (SSB) and included marital status (recategorized as married or having a cohabitant/not married or not having a cohabitant), immigration status (Norwegian/immigrant), education (primary school/secondary (high school)/≤4 years university/>4 years university), employment status (employed/unemployed/ outside of workforce or retired), and household income (recategorized in quartiles). Only individuals born outside of Norway to non-Norwegian parents were defined as immigrants. Age, sex, and screening centre were retrieved from the BCSN trial database. Data on medication prescriptions covering four years before screening invitation were obtained from the Norwegian Prescription Database (NorPD), which does not include over-the-counter use of medications. To be used as a proxy for co-morbidities, we included drugs used in diabetes (anatomical therapeutic chemical (ATC) code A10), antithrombotic agents (group B01, drugs other than aspirin), cardiac therapy (C01), diuretics (C03), antihypertensives including beta-blockers (C07), calcium channel blockers (C08), agents acting on the renin-angiotensin system (C09), psycholeptics (N05), antidepressants (N06A), drugs for obstructive airway diseases (R03). Where prescriptions included a combination of different drug/drug classes, it was included in both agents’ categories. The data from the two registries, SSB and NorPD were linked to the screening data using 11-digit unique Norwegian personal identification numbers.
2.2 Main exposure
Data on exposure to aspirin (ATC codes B01AC06, B01AC56) were available only for the four years before invitation. All the retrieved aspirin prescriptions had low-dose formulations (<325 mg). We divided the exposure period into four distinct years. If an individual received two prescriptions in one of the four years, then he/she was considered a user in that specific year. If an individual was a user for three or four years, then he/she was considered a long-term user (≥3 years). If an individual was a user for one or two years, then he/she was considered a short-term user (<3 years). All other individuals were considered non-users. In some analyses, we combined short and long-term users in the broader category of users.
2.3 Outcome
Using the most advanced lesion for each participant, findings were categorized as: no lesions, non-advanced adenomas (any adenoma measuring <1 cm, without high-grade dysplasia, and less than 25% villous features), advanced adenomas (any adenoma with size ≥1 cm, with at least 25% villous features, or high-grade dysplasia), non-advanced serrated polyps (hyperplastic polyps <1 cm in diameter and sessile serrated polyps/adenomas <1 cm in diameter and without dysplasia), advanced serrated lesions (ASL; traditional serrated adenomas or any serrated polyp with diameter size ≥1 cm, or with dysplasia) [
In participants with multiple lesions, the most severe one was chosen as the outcome of interest. However, in some cases, the choice of the most severe lesion was arbitrary (e.g., in the presence of both advanced adenoma and advanced serrated polyp). Therefore, we ran different multinomial logistic regression models according to the lesion of interest. For CRC and adenoma, we used a multinomial model with the following outcome categories, ordered by severity: CRC, adenoma, any serrated polyps, and no lesion. For serrated polyps, we used a multinomial model with the following outcome categories, ordered by severity: CRC, any serrated polyps, adenoma, and no lesion. For non-advanced and advanced adenomas, we used a multinomial model with the following outcomes, ordered by severity: CRC, advanced adenoma, non-advanced adenoma, any serrated polyps, and no lesion. For non-advanced and advanced serrated polyps, we used a multinomial model with the following outcomes, ordered by severity: CRC, advanced serrated polyps, non-advanced serrated polyps, adenomas, and no lesion. We estimated the association between aspirin use (or long-term aspirin use) and the detection of colorectal lesions, with no-use as the reference exposure category and no lesion as the reference outcome category, reporting odds ratios (OR) with 95% confidence intervals (CI). In the model for CRC and adenoma, we further separated CRCs into distal and proximal CRCs, and into stage 1 and stage ≥ 2 CRCs. We studied the association between use of aspirin and characteristics of adenomas in participants who were not diagnosed with CRC, including proximal location (from caecum to and including splenic flexure), distal location (distal to splenic flexure including rectum) [
], size ≥1 cm, presence of ≥25% villous features, and high-grade dysplasia in five distinctive multinomial models. In each model, we considered adenomas with that feature, adenomas without that feature, any serrated polyps, and no lesion as the outcome categories.
All estimates for aspirin use were adjusted for continuous age in years (integer), sex, screening centre, marital status, immigration status, education, employment status, household income, and current use (≥2 prescriptions in the 12 months before the screening invitation) of certain medications (ATC codes A10, B01, C01, C03, C07, C08, C09, N05, N06A, R03) as proxy for the presence of comorbidities at the time of invitation. We had missing values only for education and income, and we used a separate category for them in the multivariable models.
We investigated the heterogeneity of the estimates between screening arms and sex strata using the Cochran's Q test, and between CRC and adenomas’ sites, and CRC stages using the Wald test. All tests were 2-sided, and p-values <0.05 were considered statistically significant. Statistical analyses were performed with R software, version 3.6.1 and the VGAM package was used for multinomial regression.
2.5 Ethical statement
The randomized trial was approved by the Regional Committee for Medical Research Ethics in South East Norway (2011/1272). The CRC screening participants provided written informed consent on attendance at the screening centre (the sigmoidoscopy arm) or by return of the faecal sample (FIT arm). The Regional Committee for Medical Research Ethics in South East Norway approved use of registry-based data on everyone who were invited, regardless of consent. Trial registration number and date of registration are NCT 01538550 (clinicaltrials.gov), February 24, 2012.
3. Results
Of the 64,889 attenders in the screening, 24,159 were in the sigmoidoscopy arm, and 40,730 were in the FIT arm. Among sigmoidoscopy attenders, 2,372 (9.8%) had a positive sigmoidoscopy. Among FIT attenders, 3,100 (7.6%) had a positive FIT. Of all attenders, 11,608 (17.9%) individuals were aspirin users (Table 1). All users were prescribed low-dose aspirin, with 117,254 (85.1%) of prescriptions being 75 mg aspirin and 20,553 (14.9%) 160 mg aspirin. In both arms, use of aspirin was more common in males, individuals who were older, retired, and had lower education and income. In total, 314 (0.5%) participants had CRCs, 3,800 (5.9%) adenomas, 4,170 (6.4%) serrated lesions, and 2,408 (3.7%) both adenomas and serrated lesions.
Table 1Attenders’ characteristics and detected lesions by use of aspirin overall and by arm.
Compared to no use, use of aspirin (OR 0.82; 95%CI 0.61-1.11) or short-term use of aspirin (<3 years) (OR 1.39; 95%CI 0.90-2.14) were not associated with the detection of CRC (Fig. 1). However, long-term use of aspirin (≥3 years) was associated with a lower detection of CRC (OR 0.66; 95%CI 0.46-0.93) compared to no use. When comparing long-term use to no use, no significant differences were observed between CRC location (Pheterogeneity = 0.59), CRC stage (P = 0.76), screening arm (P = 0.49) and sex (P = 0.67).
Fig. 1Association between aspirin and detection of colorectal cancer. CRC: colorectal cancer; FIT: faecal immunochemical test; U: user; NU: non-user. Odds ratios (OR) and 95% confidence interval (CI) were adjusted for age, sex, screening centre, marital status, immigration status, education, employment status, household income, and use of antidiabetics, psycholeptics, antidepressants, cardiac therapy, antithrombotic agents, drugs for obstructive airway diseases, beta-blockers, agents acting on the renin-angiotensin system, calcium channel blockers, and diuretics. Reference for each estimate is no use of aspirin.
In the overall cohort, long-term use of aspirin, compared to no use, was not associated with a lower detection of adenomas (OR 0.95; 95%CI 0.87-1.03) (Table 2), but it was associated with a lower detection of serrated polyps (OR 0.86; 95%CI 0.79-0.93). No significant differences were observed between locations of adenomas (OR 0.94; 95%CI 0.86-1.02 for distal adenoma, OR 0.98; 95%CI 0.86-1.12 for proximal adenoma, Pheterogeneity= 0.60).
Table 2Association between long-term use of aspirin (≥3 years) and detection of colorectal lesions in screening attenders, overall and according to gender.
P-value from Cochran's Q test comparing ORs between subgroups of the population.
CRC
OR (95% CI)
0.66 (0.46-0.93)
0.72 (0.38-1.38)
0.61 (0.40-0.93)
0.67
CasesU/TotalU-CasesNU/TotalNU
46/8715-243/53281
12/3269-100/29397
34/5446-143/23884
Adenoma
OR (95% CI)
0.95 (0.87-1.03)
0.89 (0.76-1.03)
0.97 (0.87-1.07)
0.34
CasesU/TotalU-CasesNU/TotalNU
1005/8715-4899/53281
252/3269-2114/29397
753/5446-2785/23884
Non-advanced adenoma
OR (95% CI)
0.95 (0.86-1.05)
0.88 (0.73-1.06)
0.98 (0.86-1.11)
0.35
CasesU/TotalU-CasesNU/TotalNU
625/8715-3227/53281
161/3269-1429/29397
464/5446-1798/23884
Advanced adenoma
OR (95% CI)
0.95 (0.83-1.08)
0.90 (0.71-1.15)
0.95 (0.81-1.11)
0.71
CasesU/TotalU-CasesNU/TotalNU
380/8715-1672/53281
91/3269-685/29397
289/5446-987/23884
Serrated polyp
OR (95% CI)
0.86 (0.79-0.93)
0.86 (0.74-0.99)
0.84 (0.76-0.94)
0.80
CasesU/TotalU-CasesNU/TotalNU
871/8715-5415/53281
256/3269-2588/29397
615/5446-2827/23884
Non-advanced serrated polyp
OR (95% CI)
0.84 (0.77-0.92)
0.87 (0.75-1.01)
0.81 (0.72-0.90)
0.44
CasesU/TotalU-CasesNU/TotalNU
774/8715-4932/53281
229/3269-2330/29397
545/5446-2602/23884
Advanced serrated polyp
OR (95% CI)
1.02 (0.80-1.31)
0.74 (0.48-1.13)
1.25 (0.91-1.70)
0.05
CasesU/TotalU-CasesNU/TotalNU
97/8715-483/53281
27/3269-258/29397
70/5446-225/23884
CRC: colorectal cancer; U: user; NU: non-user. Odds ratios (OR) and 95% confidence interval (CI) were adjusted for age, sex, screening centre, marital status, immigration status, education, employment status, household income, and use of antidiabetics, psycholeptics, antidepressants, cardiac therapy, antithrombotic agents, drugs for obstructive airway diseases, beta-blockers, agents acting on the renin-angiotensin system, calcium channel blockers, and diuretics. Reference for each estimate is no use of aspirin.
a P-value from Cochran's Q test comparing ORs between subgroups of the population.
In the sigmoidoscopy arm, long-term aspirin use, compared to no use, was associated with lower detection of any adenoma (OR 0.89; 95%CI 0.80-0.99; non-advanced adenomas 0.89; 95%CI 0.79-1.01, advanced adenomas 0.88; 95%CI 0.74-1.06; Table 3), serrated polyps (OR 0.80; 95%CI 0.72-0.89), non-advanced serrated polyps (OR 0.79; 95%CI 0.71-0.88), but not ASLs (OR 0.97; 95%CI 0.72-1.32). In the sigmoidoscopy arm, the odds of detection of any adenoma and ASL for long-term versus no use were lower in women than men (Pheterogeneity = 0.07 and 0.03, respectively).
Table 3Association between long-term use of aspirin (≥3 years) and detection of colorectal lesions in the sigmoidoscopy arm, overall and according to gender.
P-value from Cochran's Q test comparing ORs between subgroups of the population.
CRC
OR (95% CI)
0.56 (0.33-0.97)
0.75 (0.30-1.89)
0.47 (0.24-0.91)
0.42
CasesU/TotalU-CasesNU/TotalNU
19/3243-110/19877
6/1168-47/10647
13/2075-63/9230
Adenoma
OR (95% CI)
0.89 (0.80-0.99)
0.76 (0.62-0.92)
0.94 (0.82-1.07)
0.07
CasesU/TotalU-CasesNU/TotalNU
662/3243-3546/19877
156/1168-1549/10647
506/2075-1997/9230
Non-advanced adenoma
OR (95% CI)
0.89 (0.79-1.01)
0.76 (0.61-0.96)
0.94 (0.81-1.10)
0.14
CasesU/TotalU-CasesNU/TotalNU
461/3243-2597/19877
110/1168-1160/10647
351/2075-1437/9230
Advanced adenoma
OR (95% CI)
0.88 (0.74-1.06)
0.74 (0.53-1.05)
0.94 (0.75-1.16)
0.25
CasesU/TotalU-CasesNU/TotalNU
201/3243-949/19877
46/1168-389/10647
155/2075-560/9230
Serrated polyp
OR (95% CI)
0.80 (0.72-0.89)
0.78 (0.66-0.93)
0.80 (0.70-0.91)
0.82
CasesU/TotalU-CasesNU/TotalNU
709/3243-4716/19877
211/1168-2279/10647
498/2075-2437/9230
Non-advanced serrated polyp
OR (95% CI)
0.78 (0.70-0.87)
0.80 (0.67-0.95)
0.76 (0.67-0.88)
0.66
CasesU/TotalU-CasesNU/TotalNU
645/3243-4384/19877
196/1168-2103/10647
449/2075-2281/9230
Advanced serrated polyp
OR (95% CI)
0.97 (0.72-1.32)
0.60 (0.34-1.06)
1.26 (0.87-1.84)
0.03
CasesU/TotalU-CasesNU/TotalNU
64/3243-332/19877
15/1168-176/10647
49/2075-156/9230
CRC: colorectal cancer; U: user; NU: non-user. Odds ratios (OR) and 95% confidence interval (CI) were adjusted for age, sex, screening centre, marital status, immigration status, education, employment status, household income, and use of antidiabetics, psycholeptics, antidepressants, cardiac therapy, antithrombotic agents, drugs for obstructive airway diseases, beta-blockers, agents acting on the renin-angiotensin system, calcium channel blockers, and diuretics. Reference for each estimate is no use of aspirin.
a P-value from Cochran's Q test comparing ORs between subgroups of the population.
In the overall cohort of screening attenders, we found no associations between long-term aspirin use and presence of ≥25% villous features (OR 0.95; 95%CI 0.80-1.13), high-grade dysplasia (OR 1.11; 95%CI 0.89-1.38) or size ≥1cm (OR 0.96; 95%CI 0.83-1.11) in adenomas. Likewise in the sigmoidoscopy arm, the estimates were non-significant for presence of ≥25% villous features (OR 0.87; 95%CI 0.68-1.11), high-grade dysplasia (OR 1.19; 95%CI 0.87-1.61) or size ≥1cm (OR 0.93; 95%CI 0.76-1.14) in adenomas. In FIT participants, we found no association between aspirin use and polyp detection and no differences between men and women (Table 4).
Table 4Association between long-term use of aspirin (≥3 years) and detection of colorectal lesions in the FIT arm, overall and according to gender.
P-value from Cochran's Q test comparing ORs between subgroups of the population.
CRC
OR (95% CI)
0.72 (0.45-1.15)
0.68 (0.27-1.67)
0.72 (0.41-1.24)
0.91
CasesU/TotalU-CasesNU/TotalNU
27/5472-133/33404
6/2101-53/18750
21/3371-80/14654
Adenoma
OR (95% CI)
1.03 (0.89-1.18)
1.12 (0.88-1.43)
0.97 (0.81-1.15)
0.34
CasesU/TotalU-CasesNU/TotalNU
343/5472-1353/33404
96/2101-565/18750
247/3371-788/14654
Non-advanced adenoma
OR (95% CI)
1.07 (0.87-1.31)
1.16 (0.83-1.63)
1.02 (0.79-1.31)
0.55
CasesU/TotalU-CasesNU/TotalNU
164/5472-630/33404
51/2101-269/18750
113/3371-361/14654
Advanced adenoma
OR (95% CI)
0.99 (0.82-1.20)
1.08 (0.76-1.52)
0.93 (0.74-1.17)
0.48
CasesU/TotalU-CasesNU/TotalNU
179/5472-723/33404
45/2101-296/18750
134/3371-427/14654
Serrated polyp
OR (95% CI)
1.02 (0.84-1.24)
1.13 (0.79-1.60)
0.96 (0.75-1.22)
0.45
CasesU/TotalU-CasesNU/TotalNU
162/5472-699/33404
45/2101-309/18750
117/3371-390/14654
Non-advanced serrated polyp
OR (95% CI)
1.01 (0.81-1.26)
1.21 (0.80-1.81)
0.92 (0.70-1.20)
0.27
CasesU/TotalU-CasesNU/TotalNU
129/5472-548/33404
33/2101-227/18750
96/3371-321/14654
Advanced serrated polyp
OR (95% CI)
1.04 (0.68-1.60)
0.94 (0.48-1.83)
1.16 (0.66-2.03)
0.67
CasesU/TotalU-CasesNU/TotalNU
33/5472-151/33404
12/2101-82/18750
21/3371-69/14654
CRC: colorectal cancer; FIT: faecal immunochemical test; U: user; NU: non-user. Odds ratios (OR) and 95% confidence interval (CI) were adjusted for age, sex, screening centre, marital status, immigration status, education, employment status, household income, and use of antidiabetics, psycholeptics, antidepressants, cardiac therapy, antithrombotic agents, drugs for obstructive airway diseases, beta-blockers, agents acting on the renin-angiotensin system, calcium channel blockers, and diuretics. Reference for each estimate is no use of aspirin
a P-value from Cochran's Q test comparing ORs between subgroups of the population.
In this large screening population, we found a significant association between long-term use (≥3 years) of aspirin and lower detection of CRC. The reduced detection of CRC was unrelated to CRC location, stage, screening arm, and sex. In the sigmoidoscopy arm, we found an association between long-term use of aspirin and lower detection of adenomas, but not ASLs. In the FIT arm, we did not observe any association between aspirin use and any polyp detection.
Our results showed a 34% reduction in detection of CRC in long-term users of low-dose aspirin, quite consistent with the findings of a 2020 meta-analysis of observational studies that reported a 27% reduction (OR 0.73; 95%CI 0.69-0.78) in regular aspirin users [
], and with the findings of a 2021 meta-analysis of RCTs that reported a 26% reduction (OR 0.74; 95%CI 0.56-0.97) in the individuals randomized to the aspirin arm [
]. In the Nurses’ Health Study and Health Professionals Follow-up Study, there was a protective effect of aspirin only if initiated before 70 years of age [
] years of use. In the Women's Health Initiative cohort and the Aspirin in Reducing Events in the Elderly (ASPREE) trial, no protective effect of aspirin against CRC was observed [
. It was argued that a possible explanation for the lack of a protective effect of aspirin in the ASPREE trial was due to the fact that the majority of participants had not used aspirin regularly before the age of 70 [
]. In the ASCEND (A Study of Cardiovascular Events in Diabetes) trial, researchers did not find an association between aspirin use and the risk of gastrointestinal cancers, but they did not report an estimate separately for CRC [
We found an association between use of three years or more of low-dose aspirin and CRC detection. Consistent with the present study, three years of aspirin use was used as a cut-off for definition of regular aspirin use in other published studies [
. In our study, all aspirin users used low-dose aspirin (75-160 mg daily). There is great variation between published studies in aspirin dosage, from approximately 50 mg to ≥500 mg daily [
]; the latter being consistent with our results indicating no significant difference between association of long-term aspirin use and detection of distal or proximal CRC. In a large observational study, current use of aspirin showed no risk reduction for Duke A CRC, while significant risk reduction was reported for more advanced stages [
]. In our study, we did not detect any significant difference between the estimates for stage 1 and the higher stages of CRC.
Two major distinctive pathways are proposed for colorectal tumorigenesis, one originating from serrated and the other from adenomatous polyps, harbouring BRAF and KRAS/APC mutations, respectively [
, are consistent with our results indicating an association between long-term aspirin use and lower detection of adenomas but not ASL. We found an indication of an association between long-term use of aspirin and the detection of advanced adenomas in the sigmoidoscopy arm, which is consistent with the findings of three large pooled cohorts that reported a reduction in risk of advanced adenomas (OR 0.88; 95% CI, 0.82–0.94) with regular use of aspirin [
]. We found no association between use of aspirin and detection of ASLs, which is in support of the results from other published studies that investigated the association between use of aspirin and risk of BRAF-mutated CRC [
], both indicating no association between use of aspirin and risk of tumorigenesis through the serrated pathway. However, the pooled estimate of three cohort studies showed an association between use of aspirin and lower incidence of serrated polyps with a diameter of >1 cm [
]. One reason for the inconsistency between that study and ours could be due to the fact that we used a definition of ASL which was not only based on the polyp size. However, results did not change when we used size as the only criterion for defining ASL (data not shown). It is also important to notice that, since ASLs are usually found in the proximal colon [
Sporadic and syndromic hyperplastic polyps and serrated adenomas of the colon: classification, molecular genetics, natural history, and clinical management.
], the two screening methods in this study might be inadequate to detect ASLs, making it suboptimal for investigating the association between aspirin use and ASLs.
We found a significant inverse association between aspirin use and CRC detection in the sigmoidoscopy participants but not in the FIT participants, even though the heterogeneity analysis did not show a significant difference between the two arms. Moreover, we found a significant inverse association between use of aspirin and the detection of adenomas in the sigmoidoscopy arm, while we found no association in the FIT arm. One reason may be that only a small proportion of adenomas bleed [
]. Also, the use of aspirin as an antiplatelet medication leads to a general increased risk of bleeding events and referrals to follow-up colonoscopy in users, higher incidental detection of adenomas, and consequently to a biased estimate of the association between aspirin and detection of adenomas towards the null [
This study is among the few studies evaluating the association between aspirin and colorectal lesions according to their characteristics, including location, stage, and histopathological features. This allowed us to not only supply new evidence on the protective effect of aspirin on the risk of CRC but also provide new insights into the association between aspirin use and detection of polyps arising from distinctive pathways. Linkage of high-quality population registries, which collects data on drug prescriptions as well as on important sociodemographic confounders, and screening outcomes is a major strength of this study. Considering that long-term use of aspirin is only possible by prescription in Norway, acquiring the medication history from the prescription database avoids bias due to self-reported use of drugs and represents another major strength of this study.
Having only four years of medication history was possibly the most important limitation of our study. However, we were able to classify the use of aspirin in short- and long-term use, using the same definition as in some previous studies [
. Another limitation is that our study is based on a selected population of individuals aged 50-74 years who decided to participate in a CRC screening trial, and we know that different sociodemographic and health-related factors influence screening participation [
. This hampers the generalizability of our study findings. In addition, since only the screening attenders with a positive FIT or sigmoidoscopy were invited to a colonoscopy, many lesions could have remained undetected, leading to possible bias in our estimates. Finally, we did not have any data regarding the potential side effects of aspirin, such as gastrointestinal ulcer and bleeding, in our population. From a public health point of view, the possibility of widespread use of aspirin as a primary prevention agent for CRC could be limited due to its adverse effects [
In conclusion, our study, conducted in a large screening population, strengthens the hypothesis that long-term use of low-dose aspirin can lower the risk of CRC, and supplies new evidence that aspirin might exert its protective effect through the prevention of adenomatous lesions.
Declaration of Competing Interest
The authors declare no potential conflict of interests.
This study was funded by the Norwegian Research Council (project number: 301628). The funders had no role in the design of the study, the data collection, the analysis or interpretation of the data, or the writing of the article.
References
Sung H.
Ferlay J.
Siegel R.L.
et al.
Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries.
Annual report to the nation on the status of cancer, 1975-2006, featuring colorectal cancer trends and impact of interventions (risk factors, screening, and treatment) to reduce future rates.
Sporadic and syndromic hyperplastic polyps and serrated adenomas of the colon: classification, molecular genetics, natural history, and clinical management.
00167-6&id=gr1.jpg){kind=link}