An official website of the United States government
NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Authors; Sara D. Khangura, and Carolyn Spry.
Colorectal cancer (CRC) has been described as 1 of the most common types of cancer and causes of death from cancer in the world.1-3 In Canada, CRC is the third most common cancer, with an estimated 24,100 Canadians expected to be diagnosed in 2023.4
CRC is usually asymptomatic in its early stages, with symptoms developing as the disease progresses.5 Symptoms and signs occurring in later stages of the disease may include abdominal pain, anemia, rectal bleeding, and bowel obstruction.5 It is believed that CRC generally develops across what is known as the adenoma-carcinoma sequence, during which initially noncancerous lesions, known as adenomatous polyps, develop into cancer.6-8 This sequence can take years or decades.8 It is during this time that an opportunity for detection of precancerous lesions in the colon and rectum is available, allowing for early treatment, which can reduce morbidity and mortality.9 Survival rates for CRC are closely associated with the stage of cancer at detection, with earlier stages resulting in higher survival and later-stage cancers resulting in lower survival rates.10-12
There are 2 types of tests available to screen for CRC: stool based (including fecal immunochemical testing) and direct visualization techniques (including colonoscopy).13 The fecal immunochemical test (FIT) is a measure of hemoglobin detected in stool samples,14 which can indicate the presence of adenomatous polyps or CRC. A positive FIT is followed up with a diagnostic colonoscopy.15 Colonoscopy is an endoscopic procedure used for diagnosis and treatment (i.e., colonoscopy can be used to identify precancerous polyps or lesions and remove them from the colon).15 The benefit of colonoscopy is the opportunity to both identify and remove adenomatous polyps before they develop into invasive CRC.6,8,12,16 However, due to its invasiveness, colonoscopy can also cause harm, such as perforation or bleeding.15,17,18 Although other screening tests for CRC are available,6 this report is limited to a focus on FIT and colonoscopy.
Many current recommendations — including those in Canada19 — advise that people 50 years and older be screened for CRC.20-22 Generally, this is supported in part by evidence indicating that the incidence of CRC is higher in individuals 50 years and older compared with younger individuals.19
The incidence of CRC in people younger than 50 years has been on the rise in Canada and other countries, prompting reconsideration of the recommended age for initiation of CRC screening to include those younger than 50 years.6,7,17,18,23-27 CRC in younger populations has been characterized as more aggressive and resulting in more deleterious outcomes compared with disease identified in older populations.26 Nonetheless, the potential for harm caused by CRC screening, as well as the cost-benefit implications, give pause when considering the optimal age for initiation of CRC screening in younger populations.11,17,18,26
To help inform decisions concerning policy and practice for CRC screening, this report sought to assemble and summarize available evidence describing the clinical benefits, harms, and recommendations concerning CRC screening in individuals of average risk younger than 50 years.
An information specialist conducted a literature search on key resources, including MEDLINE, the Cochrane Database of Systematic Reviews, the International HTA Database, the websites of Canadian and major international health technology agencies, as well as a focused internet search. The search approach was customized to retrieve a limited set of results, balancing comprehensiveness with relevancy. The search strategy comprised both controlled vocabulary, such as the National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. Search concepts were developed based on the elements of the research questions and selection criteria. The main search concepts were fecal immunochemical test, colonoscopy, and colorectal cancer screening.
CADTH-developed search filters were applied to limit retrieval to health technology assessments, systematic reviews, meta-analyses, indirect treatment comparisons, any type of clinical trial, observational studies, economic studies, and guidelines. Comments, newspaper articles, editorials, and letters were excluded. The search was completed on September 5, 2023, and limited to English-language documents published since January 1, 2017. The search was supplemented by reviewing bibliographies of key papers and through contacts with experts.
One reviewer screened citations and selected studies. In the first level of screening, titles and abstracts were reviewed and potentially relevant articles were retrieved and assessed for inclusion. The final selection of full-text articles was based on the inclusion criteria presented in Table 1.
Selection Criteria.
Articles were excluded if they did not meet the selection criteria outlined in Table 1, were duplicate publications, or were published before 2017. Guidelines with unclear methodology were also excluded.
Although studies addressing issues of health equity and/or implementation that were not otherwise eligible according to the selection criteria were not included for full summary in this report, they were given consideration, with some information of relevance being summarized.
The eligible and included publications were critically appraised by 1 reviewer using the following tools as a guide: the Downs and Black checklist28 for nonrandomized studies, the Drummond checklist29 for economic evaluations, and the Appraisal of Guidelines for Research and Evaluation (AGREE) II instrument30 for guidelines. Summary scores were not calculated for the included studies and guidelines; rather, the strengths and limitations of each included publication were described narratively.
Modelling studies for which some data were summarized were not scrutinized using formal critical appraisal tools; rather, key limitations relevant to the methods were highlighted. Sources describing health equity and/or implementation issues of relevance that were not otherwise eligible for this report did not undergo critical appraisal or consideration of methodological approach.
A total of 1,220 citations were identified in the literature search. Following screening of titles and abstracts, 1,162 citations were excluded and 58 potentially relevant reports from the electronic search were retrieved for full-text review. Thirteen potentially relevant publications were retrieved from the grey literature search and hand searches for full-text review. Of these potentially relevant articles, 54 publications were excluded for various reasons, and 17 publications met the inclusion criteria and were included in this report.31-47 The included publications comprised 2 nonrandomized studies,34,46 4 modelling studies,32,36,44,45 1 economic evaluation,33 and 8 evidence-based guidelines (2 of which had 2 related publications: the original full guideline published in 2017 as well as updated recommendations of relevance to this report published in 2022 or 2023).31,35,37-43,47 Appendix 1 presents the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)48 flow chart of the study selection outlining the inclusion of eligible studies.
Additional references of potential interest are provided in Appendix 5.
Two nonrandomized studies (NRS),34,46 4 microsimulation modelling studies,32,36,44,45 1 economic evaluation,33 and 8 evidence- based guidelines (2 of which had original guidelines published in 2017 with focused updates published in 2022 and 2023) met the eligibility criteria and were summarized in this report.31,35,37-43,47
Details regarding the characteristics of these eligible publications are provided in Appendix 2.
Two nonrandomized studies met the review’s eligibility criteria, both of which were retrospective cohort studies published in 202146 and 2020.34
The 4 modelling studies used similar methods, including modifications to existing models. One modelling study used the OncoSim model,45 1 used the CRC Simulated Population Model for Incidence and Natural History (CRC-SPIN),32 1 used the Microsimulation Screening Analysis-Colon (MISCAN-Colon),44 and 1 used the CRC-SPIN, MISCAN-Colon, and Simulation Model of CRC (SimCRC).36 Inputs for the extant Canadian model were based on data from Canada,45 and based on data from the US for the other 3 models.32,36,44 The modelling studies aimed to estimate benefits and/or harms of CRC screening given estimated incidence, natural history, and/or birth cohort effects, with updated incidence data incorporated to account for more recent trends in populations younger than 50 years.32,36,44,45 Two modelling studies were published in 2023,32,45 1 in 2021,36 and 1 in 2018,44 with all using a lifetime time horizon. Three studies assumed complete or perfect adherence to screening protocols and/or follow-up,32,36,44 whereas the Canadian study assumed screening participation based on published data describing this variable.45
One economic evaluation met the review's eligibility criteria, describing a cost-utility analysis published in 2021 and conducted from a societal perspective.33 Data were sourced from reviews of published literature, national statistical data sources, and registries. The time horizon of the cost-utility study was not reported, and a decision tree model was used. The analyses assumed a willingness-to-pay threshold of €39,760 and assumed that screening for CRC is cost-effective for individuals aged 50 years and older.33
The 8 evidence-based guidelines were produced by the American College of Physicians,43 Austrian National Committee for Cancer Screening,31 Kaiser Permanente (2022),42 US Multi-Society Task Force on Colorectal Cancer (which published a full guideline in 2017 and an updated recommendation of relevance in 2022),35,40 American College of Gastroenterology,38 US Preventive Services Task Force,37 American Cancer Society,39 and Cancer Council Australia (i.e., which published a full guideline in 2017 and an updated chapter in 2023).41,47 Guidelines were published between 2017 and 2023, with 3 published in 2023,31,43,47 1 published in 2022,35 2 published in 2021,37,38 1 published in 2018,39 and 2 published in 2017.40,41 All the evidence-based guidelines reported literature searches with some systematic methods used to identify evidence.31,35,37-43,47 All of the guidelines report critical appraisal of the evidence used to inform development of the recommendations,31,37-43,47 with 3 guidelines including a standardized method to assign the quality level of the evidence supporting their recommendations,31,35,38 whereas 3 of the guidelines include some acknowledgement of the quality of the evidence supporting the relevant recommendations,37,39,43 and 2 guidelines did not make an explicit acknowledgement concerning quality of the supporting evidence.42,47 Five of the guidelines assigned strength to the relevant recommendation(s),31,35,37-39 whereas 3 did not.42,43,47
One of the 2 retrospective cohort studies was conducted in the US46 and the other in Greece.34 The economic evaluation was undertaken in Portugal.33 One of the modelling studies was conducted in Canada45 and the remaining 3 were conducted in the US.32,36,44
Six of the guidelines (1 with a full published guideline in 2017 and updated recommendations of relevance published in 2022) were developed by groups based in the US,35,37,38,40,42,43 1 was developed by an Austrian group,31 and 1 guideline (with a separately published updated chapter) by an Australian group.41,47
One retrospective cohort study included a large cohort of residents in the state of Florida who had a colonoscopy and had no evidence of a history of inflammatory bowel disease (IBD) or CRC, including 1 subgroup (some of whom underwent screening) between the ages of 45 and 49 years and another subgroup who underwent screening between the ages of 50 and 54 years.46 The other retrospective cohort study included asymptomatic individuals of average risk (i.e., excluding those with a personal or family history of CRC or adenomas, a positive screen for CRC, digestive symptoms [i.e., persistent abdominal pain, rectal blood, chronic diarrhea, weight loss], chronic IBD, or iron-deficiency anemia). The mean age of participants who underwent CRC screening was 63 years, including subgroups of those younger than 50 years (mean age = 42.5 years) and 50 years and older (mean age = 65.9 years).34
The Canadian modelling study considered 4 hypothetical birth cohorts, including those born from 1973 to 1977, 1978 to 1982, 1983 to 1987, and 1988 to 1992. However, it was unclear whether these individuals were all of average risk.45 The 3 US-based modelling studies considered hypothetical cohorts of individuals of average risk from 40 years.32,36,44 Although “average risk” was stated by these 3 modelling studies, 1 did not define average risk,32 1 defined average risk as asymptomatic and unscreened,36 and 1 described average risk as free from CRC.44
The economic evaluation’s base case examined population-level data from individuals between the ages of 45 and 50 years who had not previously undergone screening for CRC.33 No other information concerning the definition of average risk was reported.33
All the included guidelines were clear in their description of the target population of interest, which specified individuals at average risk for CRC. However, the definition of average risk was reported variably.31,35,37-43,47 One guideline did not provide a definition of average risk,38 the others included definitions that described no personal and/or family history of CRC,31,35,37,39,40,42,43,47 no diagnosis and/or clinical symptoms of CRC,31,35,37,41,43 no symptoms or history of IBD,31,35,37,39,42,43 and/or no known genetic disorders that may increase risk of CRC.31,37,39,43 Intended users of the guidelines were often described as clinicians and/or health care providers,35,38,39,41-43 although 2 of the guidelines were not explicit concerning the intended users.31,37
Most of the included publications evaluated CRC screening in individuals younger than 50 years using colonoscopy or FIT as the initial screening test.31-40,42-46 One guideline (with an original version published in 2017 and an updated chapter published in 2023) presented relevant recommendations that did not specify the screening intervention, but referred only to “screening.”47 Exceptions related to the indications for colonoscopy (i.e., indications for colonoscopy beyond screening; unclear indications for colonoscopy) are further described for individual studies as applicable.
One retrospective cohort study included subgroups of patients between the ages of 45 and 49 years or between the ages of 50 and 54 years for whom the colonoscopy indication was presumed to be screening (but also included those with other indications such as polyps or a history of polyps, or benign neoplasms).46 The other compared colonoscopy screening as the initial test for individuals younger than 50 years versus initial screening with colonoscopy in those who were 50 years and older.34
The 4 modelling studies considered CRC screening beginning at the age of 45 years,32,36,44,45 and 2 modelling studies also considered screening beginning at 40 years.44,45 Two of the modelling studies considered both FIT and colonoscopy as screening tests,36,44 whereas 1 considered biennial FIT only45 and 1 considered colonoscopy only.32
Of the 3 studies that modelled FIT screening:
No screening was the hypothetical comparator for 3 of the modelling studies,32,36,44 with 2 modelling studies including comparisons with individuals initiating screening at 50 years.36,45
The economic evaluation compared CRC screening, with either FIT or colonoscopy as the initial test, to no screening in individuals starting at 45 years.33 Analyses were mutually exclusive (i.e., FIT as initial test versus no screening, or colonoscopy as the initial test versus no screening).
All the evidence-based guidelines made recommendations concerning CRC screening in individuals younger than 50 years.31,35,37-43,47 Although some guidelines addressed multiple testing interventions that were not relevant to this report, colonoscopy and FIT were explicitly considered in 7 of the included guidelines,31,35,37-40,42,43 with 1 guideline not specifying which intervention was recommended in the recommendations of relevance to this report (although, other recommendations that are not relevant to this report describe the use of the immunochemical fecal occult blood test).47
One retrospective cohort study investigated incidence rates of CRC per 100,000 person-years across a 13-year time frame from 2005 to 2017.46 The other investigated CRC cases that were detected cumulatively across the span of 1 year.34
Three of the modelling studies reported estimates of life-years gained (LYG).32,36,44 Other estimated benefits included CRC cases avoided and CRC deaths avoided, which were reported by 2 studies.36,45 Three of the modelling studies reported burdens and/or harms of screening, with all 3 describing estimates of lifetime number of colonoscopies,32,36,44 and 1 describing estimates of complications from screening tests.36 Complications from this 1 modelling study were reported aggregately, and described as serious and other gastrointestinal events (including perforations, bleeding, need for transfusion, paralytic ileus, nausea, and vomiting) and cardiovascular events (including myocardial infarction, angina, arrhythmia, congestive heart failure, cardiac or respiratory arrest, syncope, hypotension, and shock).36 Three of the modelling studies described a lifetime horizon across which the analyses were conducted,32,36,44 whereas the Canadian modelling study reported the use of a 40-year time horizon.45
The economic evaluation measured cost-utility, expressed using Euros per quality-adjusted life-year (€/QALY). Time horizon was not reported. Incremental cost-effectiveness ratios (ICERs) were also reported as outcomes of the cost-utility analysis.33
The evidence-based guidelines considered CRC incidence, CRC mortality, and/or harms of screening as major outcomes informing development of their recommendations.31,35,37-43,47 Two guidelines (1 with 2 publications) also considered LYG with screening younger than 50 years.39,41,47
Strengths of the retrospective cohort studies included clarity of reporting, appropriate statistical methods, and valid and reliable outcome measures, which support the internal validity of the findings.34,46 One retrospective cohort study included a large sample size, drawing from a state-wide data source across a 13-year time frame, which supported its external validity.46 Limitations included a lack of reporting on harms of screening and limited clarity concerning the representativeness of the study population, intervention, and/or setting.34,46 In particular, it was not clear whether the study populations were representative of the source population from which they were selected. In 1 study, it was unclear whether the group of patients younger than 50 years was representative of individuals of average risk because indications for colonoscopy included those other than screening (i.e., history of polyps or benign neoplasm).46 Similarly, in the other nonrandomized study, it was reported that screening was provided at the request of the patient, which may not support the representativeness of the group to an average-risk population.34 Both the studies described information on the race or ethnicity of study participants, but no other details that may be relevant to health equity considerations (e.g., socioeconomic status or rurality of residence). However, outcome data were not reported by race or ethnicity.34,46 Notably, 1 study reported that all participants were of “white-Caucasian” ethnicity, which may limit the representativeness of the findings outside of these groups.34 In addition, it was not clear whether the 1 study centre at which the screening was provided was representative of the care received among the source population.34 This lack of clarity concerning representativeness may indicate limitations to the external validity of the studies, which is critical to demonstrating that the findings are likely to be applicable across populations and settings.49
The research objectives and their importance were made clear in the 3 modelling studies.32,36,44,45 Each of these studies used and modified existing models, describing the rationale and methods for doing so and including published sources for incidence and mortality data that were incorporated.32,36,44,45 One of the modelling studies assumed screening participation rates in accordance with published data,45 whereas 3 of the modelling studies assumed perfect or complete adherence to screening and/or follow-up. This is a notable limitation given that adherence to screening protocols is not perfect in real-world scenarios.43 Three of the studies addressed both estimated benefits and burdens of screening32,36,44 and 1 also reported on estimated harms of screening,36 but 1 reported on estimated benefits of screening only.45 The limited investigation and reporting of estimated harms of screening may have limited the interpretation of the findings (i.e., emphasis on the estimated benefits of screening in the analyses may not have adequately characterized the balance against estimated harms from screening). One of the modelling studies accounted for uncertainty in its estimates using credible intervals,32 but the other 3 did not.36,44,45 Characterizing uncertainty statistically is important for clarifying the accuracy and potential validity of reported estimates.29
Given the research objectives and methods described, as well as the estimates generated, the conclusions reported for the 3 modelling studies appeared balanced and reasonable.32,36,44,45
Strengths of the economic evaluation included a clearly stated research objective, viewpoint, and main outcomes. Data inputs for incidence, mortality, cost, and utilities were drawn from published sources.33 Key limitations included the lack of a reported time horizon, preventing the reader from understanding the duration over which the model estimates were generated. Although variables and transition probabilities were provided in a supplementary file to the published report, details describing the methods used for the decision tree model were not reported.33 An important limitation was the source of incidence estimates, which authors explained were only available from 1993 until 2010.33 This limits the model estimates from consideration of more recent trends in CRC incidence and could affect the cost-utility findings and conclusions (i.e., incorporation of more recent data that may demonstrate increasing incidence may generate estimates that favour cost-utility).
The scope and purpose of the included guidelines are generally clear,31,35,37-43 although the overall objectives of the guidelines are not stated or are not clear in 3 guidelines,37,38,42 and the questions being addressed by the guidelines are not clearly stated in 2 of the guidelines (1 with 2 publications).35,40,43 Stakeholder involvement is well represented in 4 guidelines, which describe involvement of relevant professional groups,31,39,42,43 whereas 4 guidelines (2 with 2 publications each) do not make this clear.35,37,38,40,41,47 Efforts to consult with members of the target population are described in 5 of the guidelines (1 with 2 publications),31,37,39,41,43,47 but not clear in the remaining 3 (1 with 2 publications).35,38,40,42
Rigour of development was supported by a systematic literature review in 7 guidelines,37-43 whereas details of the systematic review methods were not reported in 1 guideline.31 A link between the evidence and the recommendations is made clear in 7 of the guidelines,31,35,37-39,43,47 whereas the association is unclear in 1 guideline.42 There is evidence of external peer review for 7 of the guidelines,31,37-41,43 although external review is not clear for 1 guideline.42
The recommendations are clear and easily identifiable in all 8 included guidelines.31,35,37-39,42,43,47 The applicability of the recommendations is generally not clearly reported, with 6 (1 with 2 publications) not describing barriers and facilitators to applying the guidelines31,35,37,38,40,42,43 and 4 of the guidelines (1 with 2 publications) not providing advice or tools to support putting recommendations into practice.38,41-43,47 Similarly, the potential influence of funding on the development of the guidelines is not reported or not clearly reported in 6 of the guidelines (1 with 2 publications),37-39,41-43,47 and potential conflicts of interest are not reported or not clearly reported in 4 of the guidelines (1 with 2 publications).37,38,41,42,47
Additional details regarding the strengths and limitations of included publications are provided in Appendix 3.
In 1 retrospective study, 38,586 individuals between the ages of 45 and 49 years were analyzed across a 13-year time frame. Some of these individuals underwent colonoscopy for an indication of screening producing a CRC incidence rate of 23.7 per 100,000 person-years.46 In the same study, 365,152 individuals between the ages of 50 and 54 years were analyzed, producing a CRC incidence rate of 15.7 per 100,000 person-years.46 No statistical comparisons were made between these 2 groups; however, the 95% confidence intervals (CIs) for the 2 rates did not overlap (45 to 49 years: 95% CI, 18.4 to 30.2; 50 to 54 years: 95% CI, 14.1 to 17.5).46 This finding suggests there is a significant difference between the groups (i.e., a higher incidence rate of CRC observed in the younger group).
Of 47 individuals younger than 50 years (mean = 42.5 years; SD = 5.9 years) who were screened for CRC in the other retrospective cohort study, 1 CRC case (2.1%) was detected over the course of 1 year.34 In 333 individuals who were 50 years or older (mean = 65.9 years; SD = 8.3 years), 7 cases (2.1%) of CRC were detected by first-time colonoscopy screening. The odds ratio (OR) between the groups indicated a non–statistically significant difference in CRC cases detected between groups (OR = 1.01; 95% CI, 0.12 to 8.4; P = 0.9).34
Table 9 provides tabulated details for these findings.
One modelling study estimated that 27 LYG may be observed with colonoscopy every 10 years initiated at 45 years compared to 50 years in 1,000 individuals of average risk across a lifetime horizon.36
One modelling study estimated that 26 LYG may be observed with annual FIT as the initial screening test initiated at 45 years compared to 50 years in 1,000 individuals of average risk across a lifetime horizon.36
One modelling study estimated that 3 CRC cases would be avoided across a lifetime horizon with colonoscopy every 10 years initiated at 45 years compared to 50 years in 1,000 individuals of average risk.36
The Canadian modelling study estimated that 18,135 fewer CRC cases may occur across a 40-year interval among 4 birth cohorts if biennial FIT screening was initiated at 40 years compared to 50 years.45 The same study estimated that if screening was initiated at 45 years versus 50 years, 12,188 CRC cases may be avoided.45
Another modelling study estimated that 3 CRC cases would be avoided across a lifetime horizon with annual FIT initiated at 45 years compared to 50 years in 1,000 individuals of average risk.36
One modelling study estimated that 1 death caused by CRC would be avoided across a lifetime horizon with colonoscopy every 10 years initiated at 45 years compared to 50 years in 1,000 individuals of average risk.36
The Canadian modelling study estimated that 7,988 fewer deaths caused by CRC may occur across a 40-year interval among 4 birth cohorts if biennial FIT screening was initiated at 40 years compared to 50 years.45 The same study estimated that if biennial FIT screening was initiated at 45 years versus 50 years, 5,261 deaths caused by CRC may be avoided.45
Another modelling study estimated that 1 death caused by CRC would be avoided across a lifetime horizon with annual FIT initiated at 45 years compared to 50 years in 1,000 individuals of average risk.36
One modelling study estimated that an additional 784 lifetime colonoscopies would occur with colonoscopy every 10 years initiated at 45 years compared to 50 years in 1,000 individuals of average risk.36
One modelling study estimated that an additional 186 colonoscopies would be performed with annual FIT as the initial screening test initiated at 45 years compared to 50 years in 1,000 individuals of average risk across a lifetime horizon.36
One modelling study estimated that more than 2 complications (which could include gastrointestinal and/or cardiovascular events but were reported aggregately only) may occur with colonoscopy every 10 years initiated at 45 years compared to 50 years in 1,000 individuals of average risk across a lifetime horizon.36
With annual FIT as the initial screening test initiated at 45 years compared to 50 years in 1,000 individuals of average risk, 1 modelling study estimated that more than 0.2 complications may occur across a lifetime horizon. Complications were reported aggregately but may have included gastrointestinal and/or cardiovascular events.36
Table 10 provides tabulated details for these findings.
No eligible studies were identified, so no summary of the evidence could be provided.
The 3 modelling studies32,36,44 estimated a range of 33736 to 43844 LYG with colonoscopy screening every 10 years starting at age 40 or 45 versus no screening in 1,000 individuals of average risk across a lifetime horizon.36,44
With annual FIT as the initial screening test, 3 modelling studies32,36,44 estimated a range of 31836 to 41744 LYG per 1,000 individuals of average risk across a lifetime horizon.
Compared to no screening:
Compared to no screening:
Compared to no screening:
Compared to no screening:
The 3 modelling studies32,36,44 estimated expected lifetime colonoscopies for 1,000 individuals aged 40 years with average risk. These individuals received either colonoscopy testing every 10 years beginning at either age 40 or 45 years or no screening, with an estimated range of expected lifetime colonoscopies from between 3,96132 to 6,08344 for the screened groups receiving colonoscopy.
Two modelling studies estimated expected lifetime colonoscopies for 1,000 individuals aged 40 years with average risk who received either annual FIT initiated at 45 years or no screening, with estimates of between 1,68236 and 2,69844 lifetime colonoscopies expected in the screened groups.
One modelling study estimated that more than 16 lifetime complications (reported aggregately but may have included gastrointestinal and/or cardiovascular events) may occur from CRC screening with colonoscopy testing every 10 years in 1,000 individuals of average risk younger than 50 years versus no screening.36
One modelling study estimated that more than 10.2 lifetime complications (reported aggregately but may have included gastrointestinal and/or cardiovascular events) may occur from CRC screening with annual FIT screening in 1,000 individuals of average risk younger than 50 years versus no screening.36
Table 10 provides tabulated details for these findings.
Assuming a willingness-to-pay threshold of €39,760, neither a CRC screening strategy using FIT nor colonoscopy as the initial test in adults starting at 45 years demonstrated cost-utility across an unreported time horizon. Assuming the cost of colonoscopy at €150, FIT generated an ICER of €84,304, with an assumed 50% screening participation rate, and colonoscopy had an ICER of €3,112,244 with an assumed 38% participation rate, compared with no screening until age 50 years.33 Analyses assuming the cost of colonoscopy at €397 generated ICERs of €176,213 for FIT screening and €6,620,987 for colonoscopy screening.33 Authors emphasized that the findings were most sensitive to the estimated incidence of CRC, which would have to rise from 30 to 47.5 cases per 100,000 individuals of average risk younger than 50 years to demonstrate cost-utility.33
Table 11 provides tabulated detail for these findings.
Seven of the guidelines (1 developed by an Austrian group, 5 developed by groups in the US, and 1 developed by an Australian group) make recommendations in favour of screening individuals of average risk beginning at the age of 45 years,31,35,37-39,42,47 whereas 2 guidelines make recommendations against the initiation of screening in individuals at 40 years 47 or 45 years.43 Recommendations with supporting evidence and rationale were summarized, when provided in the guidelines, and are reported in Table 12.
Of the 7 guidelines that make recommendations in favour of CRC screening for individuals from 45 years, 2 indicated that the evidence is of low or very low quality,35,38 1 indicated that the evidence is of moderate quality,31 and 4 were not explicit about the quality of the evidence informing the relevant recommendation(s).37,39,42,47 Three of the guidelines assigned a grade to the recommendations, with weak, conditional or qualified recommendations favouring screening in younger populations.35,38,39
Of the 2 guidelines that make recommendations against screening for individuals younger than 50 years (1 Australian and 1 from the US), 1 did not describe the quality of the evidence used to inform the relevant recommendation against screening initiation at age 40 (but did specify an unfavourable “benefits-to-burden balance” for this age group in its rationale).47 The other evidence-based guideline, which recommends against the initiation of screening at 45 years, was clear about important limitations of the evidence reviewed, which was deemed to not support a recommendation in favour of screening in this population.43 Neither of these guidelines are explicit about assigning a grade to the relevant recommendations indicating their strength.43,47
Appendix 4 presents tabulated details about the study findings and evidence-based guidelines.
As has been highlighted in many published sources on the topic, empirical data describing the clinical benefits and harms of CRC screening in individuals of average risk younger than 50 years are scarce.50 Whereas 4 modelling studies reporting microsimulated estimates were included and summarized, 2 eligible retrospective cohort studies and 1 economic evaluation published since 2017 were identified for inclusion. The limitations of available empirical data are also highlighted in the 9 included guidelines,31,35,37-43,47 several of which are explicit about the scarcity of relevant, available evidence to inform decisions and practice regarding CRC screening in populations younger than 50 years.38,40,43 The dearth of available evidence may have resulted in no eligible studies of relevance identified using observed, empirical data in answer to the second of this report’s research questions regarding the effectiveness of colorectal cancer screening versus no screening in individuals of average risk younger than 50 years.
Eligible primary clinical and cost-effectiveness data informing this report are scarce, although modelling studies were included which provided estimates of relevance to the research questions regarding effectiveness of CRC screening.32,36,44,45 However, because these studies are not based on observed outcomes following CRC screening in individuals of average risk, the inferences and conclusions that can be drawn from their findings are limited to the validity and quality of the assumptions and inputs of the models. For instance, 3 of the modelling studies assumed perfect adherence to screening protocols, which is a limitation because perfect adherence to screening and follow-up testing is not observed in real-world settings.43 This limitation could overestimate the projected benefits, burdens, and harms of CRC screening. The Canadian modelling study reported on estimated benefits without accounting for the potential burdens and/or harms of screening and was not clear about analyzing an average-risk population,45 which could overestimate the estimated benefits of screening. Regarding complications, 1 of the modelling studies reported on this outcome, but did not specify types of harms. It reported complications aggregately as gastrointestinal and/or cardiovascular events.36 This lack of specificity limits the interpretation of the findings regarding complications from colonoscopy that can occur because the severity of different types of complications is variable. Several of the evidence-based guidelines included in this report relied on some of these and other modelling studies,31,35,39,41,47 most of which emphasize the limitations of relying on modelled data for informing decision-making and clinical practice.
The potential utility of some, or all, of the recommendations in the included evidence-based guidelines may also be limited. Seven of the guidelines made recommendations favouring CRC screening in individuals of average risk younger than 50 years,31,35,37-39,42,47 whereas 2 made recommendations that do not favour CRC screening in this age group, specifically in individuals starting at the age of 45 years43 or in individuals starting at the age of 40 years.47 The discordance among some of the recommendations made in the evidence-based guidelines is particularly notable given the similar evidence sources used across the included guidelines, and the agreement among them concerning the availability and quality of evidence. For instance, it is not clear whether other factors in addition to the available evidence may have accounted for the lack of agreement across all the included guidelines.
Of the eligible and included studies summarized in this report, some data may be limited in their potential applicability, utility, and/or generalizability, including that of relevance to the Canadian context. Four of the included sources did not clearly state or define whether the population of interest was of average risk,32,33,38,45 which may limit the applicability of their findings and/or recommendations to the population of interest. One retrospective cohort study was conducted using health administrative data in a US context,46 and the other was conducted on a limited sample size in a Greek population and health system,34 which may bear limited representativeness to Canadian populations and health care systems. In addition, 1 retrospective cohort study may have included study patients who were not of average risk.,46 The other study compared younger study participants to those older than 50 years, including those who were 75 years and older,34 neither of which are relevant to this report or to the Canadian context. Further, because the Canadian modelling study did not clearly limit its analyses to individuals of average risk,45 its findings may be limited in their generalizability to an average-risk population.
The cost-utility analyses were conducted in Portugal and incorporated incidence data from 1993 until 2010.33 This may also bear limited relevance to the Canadian context and may not adequately consider more recent trends in CRC incidence among individuals of average risk younger than 50 years. In addition, the lack of clarity about the CRC risk of included participants and the lack of a reported time horizon limited the interpretation of the findings.
Six of the included guidelines (1 with 2 publications) were developed in the US, which has a health system that is distinct from Canada’s health systems.35,37-40,42,43 The remaining 2 guidelines (1 with 2 publications) were produced in Austria31 and Australia,41,47 which have distinct social, cultural, economic, and geographic contexts, thus may have limited generalizability within Canada.
This review identified 2 retrospective cohort studies, 1 economic evaluation, 3 modelling studies, and 8 evidence-based guidelines (2 of which also published updates in 2022 and 2023) reporting data and information relevant to CRC screening in individuals of average risk younger than 50 years.31-44,47
Empirical data describing observations of the effectiveness of screening regarding CRC incidence, mortality, and harms were limited. One retrospective cohort study reported rates that are suggestive of a higher incidence of CRC in individuals between the ages of 45 and 49 years compared with those between the ages of 50 and 54 years (although the authors concede that the younger group may not be representative of an average-risk population).46 The other retrospective cohort study found no statistically significant difference in the cumulative incidence in Greece of CRC among 380 individuals of average risk who underwent colonoscopy for the first time, 47 of whom were younger than 50 years, compared with 333 who were 50 years or older.34 Differences in the methods and other study features may account for this apparent difference in observed direction of effect (e.g., the US study did not clearly limit study participants to those of average risk, or to colonoscopy for screening alone,46 whereas the Greek study used a small sample size that may not be generalizable to a larger population34).
It is likely that the limited empirical data available that answers the research questions posed in this report are a function of the widespread and longstanding recommendations in most jurisdictions limiting CRC screening to those aged 50 years and older (i.e., where screening for individuals of average risk younger than 50 years is not practised, generating data on its effectiveness is challenging to produce). Consequently, this report and many of its included evidence-based guidelines have looked to modelled data to help inform questions about CRC screening in individuals of average risk younger than 50 years. The estimated benefit generated by modelled data highlight opportunities for increased LYG, reductions in CRC cases, and mortality, alongside estimated burdens and harms that include increased testing and potential adverse events that screening in individuals of average risk younger than 50 years are anticipated to produce.32,36,44 These limited findings may be interpreted as supportive of CRC screening strategies that initiate in populations younger than 50 years, and are corroborated by observed and broadly acknowledged increases in CRC incidence among individuals younger than 50 years across the world.6,7,24,25,51 Nonetheless, modelled screening scenarios that assumed the initiation of screening in individuals at 40 or 45 years compared to 50 years and older also produce higher estimates of lifetime colonoscopies and potential harms from this increased number of endoscopic interventions across the lifespan.32,36,44 Finally, 1 modelling study included in this report accounted for imperfect participation in screening,45 but 3 did not.32,36,44 Future modelling studies may also consider incorporating assumptions and inputs that are more consistent with real-world observations (e.g., that adherence to CRC screening protocols are not perfect, which will impact the estimates of benefits and harms).
One economic evaluation determined there was no cost-utility of CRC screening in individuals of average risk younger than 50 years in Portugal based on current estimates of incidence.33 Nonetheless, incidence data used in the decision tree model covered the time period up until 2010,33 rendering its findings limited in their applicability to more current trends in CRC incidence.33 Further, sensitivity analyses demonstrated that CRC incidence would have to rise from its current rate of 30 to 47.5 per 100,000 individuals of average risk younger than 50 years to render cost-utility, assuming the lowest cost test option (i.e., FIT at €150).33 Although screening for CRC in populations younger than 50 years has been described elsewhere in the literature as not being cost-effective,26 the rising trends in incidence of the disease among younger people may impact the findings for this outcome. In 2 economic evaluations that did not meet the eligibility criteria for this review, the authors concluded that CRC screening for those younger than 50 years compared to those 50 years and older demonstrated cost-effectiveness.45,52 However, 1 of these also reported greater benefits at lower costs were estimated with CRC screening strategies that targeted uptake among higher-risk and older populations.52
Seven of the 8 included evidence-based guidelines made recommendation(s) in favour of CRC screening for individuals of average risk younger than 50 years, specifically starting at age 45.31,35,37-39,42,47 Nonetheless, a recurring assertion among the evidence-based guidelines included in this report was the lack of empirical data available to inform recommendations for CRC screening in individuals of average risk younger than 50 years.31,35,37-39,41-43,47 Evidence describing the effectiveness of CRC screening relies almost solely on nonrandomized methods, which are arguably unavoidable due to the nature of the disease and the diagnostic process, but which render less robust findings than randomized studies (e.g., CRC screening studies may be vulnerable to self-selection and other biases).12 Similarly, a recent review of Canadian recommendations and practices for CRC screening in individuals of average risk younger than 50 years concluded that the balance of costs and benefits remains unclear.19
Further, a preponderance of data from modelling studies has been used to inform recommendations in favour of CRC screening for younger individuals,31,35,37-40,42 And although modelling studies are common and can be useful for estimating long-term outcomes of CRC screening, the data are also vulnerable to variability and uncertainty according to model inputs. For example, various existing models assume variable durations for the adenoma-carcinoma sequence that range from 6 to 23 years.8 These and other differences in model assumptions and inputs can account for important differences in the estimates that they generate concerning the benefits of CRC screening. Lastly, a more extensive and robust assessment of the estimated harms of screening may provide a more fulsome perspective on the balances of benefits and harms, which could better inform decision-making concerning policy and practice for CRC screening in individuals of average risk younger than 50 years.
In addition to the evidence relevant to clinical effectiveness and cost-effectiveness, considerations that address health equity, implementation, and broader impacts of CRC screening for younger individuals of average risk are critical. One Canadian guidance document highlighted the importance of prioritizing CRC screening outreach interventions for adults between the ages of 45 and 75 years who are experiencing disadvantage.53 Another analysis concluded that CRC screening in younger individuals would likely be cost-effective, but a strategy focused on increasing uptake among those 50 years and older, as well as among those at highest risk, would produce greater societal benefits.52
Some information addressed the need for increased uptake among higher-risk and/or underserved communities as a strategy for improving the effectiveness of CRC screening. One of the modelling studies summarized in this report generated estimates by race and sex subgroups, reporting that life expectancy, CRC cases, and deaths were unlikely to differ by race in a hypothetical population of 40 year olds that had not previously undergone CRC screening.36 Based on this, and their supporting work of a review of the literature, the authors posited that disparities by race and sex in CRC outcomes are more likely to be caused by differences in screening uptake than by differences in natural history of the disease by racial group.36 Accordingly, 1 of the evidence-based guidelines summarized in this report made a qualifying statement concerning race (i.e., that while CRC incidence rates are higher in Black, American Indian, and Alaskan Native adults, screening should be offered to all adults beginning at the age of 45 years).37 Similarly, another of the evidence-based guidelines incorporated good practice points highlighting the importance of health care human resources availability and upskilling to support the uptake of screening — particularly among potentially disadvantaged communities.47 CRC screening strategies that prioritize improving uptake among higher-risk and/or underserved communities have been studied as well, with 2 studies identified by this review investigating implementation techniques for screening programs in racialized populations.54,55 One US-based study of a FIT mailed outreach intervention found that an enhanced envelope with what was intended to be a more “eye catching” appearance produced statistically significantly higher return rates than a plain envelope in adults between the ages of 45 and 49 years.54 Return rates were not statistically significantly different between envelope types among non-Hispanic white and Hispanic groups, a statistically significant higher return rate for enhanced envelopes was observed among the non-Hispanic Black population in the study.54 Another US-based study describing health equity and implementation considerations found that the completion rate of a mailed FIT kit in a previously unscreened African American population who were between the ages of 45 and 50 years was statistically significantly higher than those of previously unscreened African American, white, and Hispanic groups between the ages of 51 and 56 years (although, there was no statistically significant difference with Asian and Pacific Islanders between the ages of 51 and 56 years).55 Completion rates of colonoscopy among FIT-positive African Americans aged 45 to 50 years did not differ versus those of African American, white, and Hispanic groups (although rates were higher compared to the Asian and Pacific Islander group aged 51 to 56 years).55 CRC cases among FIT-positive African American, white, and Asian and Pacific Islander groups between the ages of 51 and 56 years were statistically significantly higher than those among African Americans aged 45 to 50 years (with no cases observed among FIT-positive Hispanics aged 51 to 56 years).55 These findings appear to support analyses indicating that strategies focused on increasing CRC screening uptake in older and/or higher-risk populations may yield greater overall benefit than strategies focused on universal screening in individuals of average risk younger than 50 years.52
Finally, while considering implementation of CRC screening programs in individuals of average risk younger than 50 years and/or optimizing uptake among higher-risk and disadvantaged communities is important, barriers and facilitators to buy-in among health care providers is also essential. One US-based survey of health care providers reported that a majority of respondents did not believe that CRC screening (using colonoscopy or FIT) was effective for individuals younger than 50 years.56 Similarly, a large majority of respondents indicated that they would not routinely recommend CRC screening with colonoscopy or FIT in individuals younger than 50 years.56For CRC screening programs to be optimally effective in any population, the support of clinicians who provide health care and requisition CRC screening tests is critical. Changes to CRC screening recommendations and policy must carefully consider outreach and education for health care providers.
Findings from relevant sources describing health equity and/or implementation considerations are provided in Table 13.
Given the limited empirical data available to inform policy and practice for CRC screening among individuals of average risk, future research efforts may benefit from a focus on rigorous observational studies of populations in which CRC screening for younger populations has been recommended (e.g., in the US). Real-world observations — using pilot programs, for instance — and comparisons with no screening in younger individuals as well as comparisons with outcomes in older individuals, will better support consideration of the benefits, harms, costs, and implementation considerations of a CRC screening approach in individuals of average risk younger than 50 years. In addition, adjustments for potentially confounding factors (e.g., sex, race and/or ethnicity, as well as lifestyle risk factors such as overweight or obesity, excessive alcohol, and/or tobacco use), will support more accurate estimates of the effectiveness of screening in younger individuals of average risk. Importantly, broader considerations addressing the potential for other screening strategies — including increasing uptake in older, higher-risk, and disadvantaged populations — will also be essential to informing decisions concerning optimal approaches to CRC screening in the future.57
colorectal cancer
fecal immunochemical test
life-years gained
Contributors: Elizabeth Carson, Shannon Hill, Camille Santos, Anusree Subramonian
Selection of Included Studies.
Note that this appendix has not been copy-edited.
Characteristics of Included Primary Clinical Studies.
Characteristics of Relevant Modelling Studies.
Characteristics of Included Economic Evaluation.
Characteristics of Included Guidelines.
Note that this appendix has not been copy-edited.
Strengths and Limitations of Clinical Studies Using the Downs and Black Checklist.
Strengths and Limitations of Economic Evaluation Using the Drummond Checklist.
Strengths and Limitations of Guidelines Using AGREE II.
Note that this appendix has not been copy-edited.
Colorectal Cancer Incidence.
Findings From Relevant Modelling Studies.
Summary of Findings of Included Economic Evaluation.
Summary of Recommendations in Included Guidelines.
Health Equity and Implementation Considerations From Eligible and Otherwise Relevant Studies.
Note that this appendix has not been copy-edited.
Disclaimer: The information in this document is intended to help Canadian health care decision-makers, health care professionals, health systems leaders, and policy- makers make well-informed decisions and thereby improve the quality of health care services. While patients and others may access this document, the document is made available for informational purposes only and no representations or warranties are made with respect to its fitness for any particular purpose. The information in this document should not be used as a substitute for professional medical advice or as a substitute for the application of clinical judgment in respect of the care of a particular patient or other professional judgment in any decision-making process. The Canadian Agency for Drugs and Technologies in Health (CADTH) does not endorse any information, drugs, therapies, treatments, products, processes, or services.
While care has been taken to ensure that the information prepared by CADTH in this document is accurate, complete, and up-to-date as at the applicable date the material was first published by CADTH, CADTH does not make any guarantees to that effect. CADTH does not guarantee and is not responsible for the quality, currency, propriety, accuracy, or reasonableness of any statements, information, or conclusions contained in any third-party materials used in preparing this document. The views and opinions of third parties published in this document do not necessarily state or reflect those of CADTH.
CADTH is not responsible for any errors, omissions, injury, loss, or damage arising from or relating to the use (or misuse) of any information, statements, or conclusions contained in or implied by the contents of this document or any of the source materials.
This document may contain links to third-party websites. CADTH does not have control over the content of such sites. Use of third-party sites is governed by the third- party website owners’ own terms and conditions set out for such sites. CADTH does not make any guarantee with respect to any information contained on such third-party sites and CADTH is not responsible for any injury, loss, or damage suffered as a result of using such third-party sites. CADTH has no responsibility for the collection, use, and disclosure of personal information by third-party sites.
Subject to the aforementioned limitations, the views expressed herein are those of CADTH and do not necessarily represent the views of Canada’s federal, provincial, or territorial governments or any third-party supplier of information.
This document is prepared and intended for use in the context of the Canadian health care system. The use of this document outside of Canada is done so at the user’s own risk.
This disclaimer and any questions or matters of any nature arising from or relating to the content or use (or misuse) of this document will be governed by and interpreted in accordance with the laws of the Province of Ontario and the laws of Canada applicable therein, and all proceedings shall be subject to the exclusive jurisdiction of the courts of the Province of Ontario, Canada.
The copyright and other intellectual property rights in this document are owned by CADTH and its licensors. These rights are protected by the Canadian Copyright Act and other national and international laws and agreements. Users are permitted to make copies of this document for non-commercial purposes only, provided it is not modified when reproduced and appropriate credit is given to CADTH and its licensors.
About CADTH: CADTH is an independent, not-for-profit organization responsible for providing Canada’s health care decision-makers with objective evidence to help make informed decisions about the optimal use of drugs, medical devices, diagnostics, and procedures in our health care system.
Funding: CADTH receives funding from Canada’s federal, provincial, and territorial governments, with the exception of Quebec. Questions or requests for information about this report can be directed to ac.HTDAC@stseuqeR
This document was externally reviewed by a content expert who has granted permission to be cited.
Linda Rabeneck, MD MPH FRCPC
Professor of Medicine
University of Toronto
Copyright © 2023 - Canadian Agency for Drugs and Technologies in Health. Except where otherwise noted, this work is distributed under the terms of a Creative Commons Attribution-NonCommercial- NoDerivatives 4.0 International licence (CC BY-NC-ND).
Your browsing activity is empty.
Activity recording is turned off.
See more...
