This document is the first update of the American College of Gastroenterology (ACG) colorectal cancer (CRC) screening recommendations since 2000. The CRC screening tests are now grouped into cancer prevention tests and cancer detection tests. Colonoscopy every 10 years, beginning at age 50, remains the preferred CRC screening strategy. It is recognized that colonoscopy is not available in every clinical setting because of economic limitations. It is also realized that not all eligible persons are willing to undergo colonoscopy for screening purposes. In these cases, patients should be offered an alternative CRC prevention test (flexible sigmoidoscopy every 5–10 years, or a computed tomography (CT) colonography every 5 years) or a cancer detection test (fecal immunochemical test for blood, FIT).
The members of the writing committee carried out a systematic literature review and developed the updated guideline recommendation document. Only peer-reviewed English language articles were included. The criteria used for evaluation of studies and assessment of the category of evidence and strength of recommendation are shown in Table 1 (1). These guidelines have also been reviewed and approved by the Practice Parameters Committee of the American College of Gastroenterology (ACG) and by the ACG Board of Trustees.
The ACG is an organization of more than 10,000 clinical gastroenterologists and related health professionals. In 2000, the ACG issued colorectal cancer (CRC) screening recommendations that endorsed colonoscopy every 10 years, beginning at age 50, as the preferred CRC screening strategy (2). The ACG was the first organization to recommend colonoscopy as the preferred strategy for the CRC screening; and the American Society for Gastrointestinal Endoscopy (3) and National Comprehensive Cancer Network (4) subsequently endorsed this recommendation.
Other guidelines for CRC screening often utilize an approach called the “menu of options.” In this approach, multiple options for screening are presented which differ with regard to their effectiveness, risk, and degree of invasiveness (and, therefore, potentially their acceptability to patients). The menu-of-options approach was first formalized by the “GI consortium” in May 1997 (5), endorsed by the American Cancer Society in 1997 (6), revised by the US Multisociety Task Force in 2003 (7), and revised by a joint committee of the US Multisociety Task Force, the American Cancer Society, and the American College of Radiology in 2008 (8). The ACG participated in and endorsed the menu-of-options approach in 1997, 2003, and 2008. The ACG continues to endorse the menu-of-options approach as appropriate to CRC screening. Publication of this guideline does not rescind the ACG's endorsement of the joint guideline (8). New recommendations, which differ from the earlier ACG guideline, are highlighted in Table 2. The rationale for a separate ACG screening guideline is discussed below.
Rationale for a preferred strategy
As in 2000, the ACG recommends that clinicians have access to a “preferred” strategy for making CRC screening recommendations, as an alternative to the “menu of options” approach, if warranted by the performance characteristics of one of the tests. The ACG recommends colonoscopy every 10 years based on the evidence of colonoscopy effectiveness, cost-effectiveness, and acceptance by patients. A “preferred” strategy simplifies and shortens discussions with patients and could also increase the likelihood that screening is offered to patients. One randomized trial showed that patients were more likely to undergo screening with the “preferred” strategy approach compared with the “menu of options” (9). Another study found no improvement in screening rates when multiple options were presented (10). Maintaining simplicity in guidelines may have value, in that recent evidence has suggested that practitioners often do not follow recommended intervals for post-polypectomy surveillance, which may in part be because of their complexity (11,12,13). The ACG acknowledges that listing quality colonoscopy as a “preferred” CRC prevention strategy places greater emphasis on effectiveness than on risk. Current trends in procedure use in the United States reflect and are consistent with the ACG's recommendation of colonoscopy as the preferred strategy for CRC screening, in that colonoscopy procedure volumes have risen dramatically, whereas flexible sigmoidoscopy and double-contrast barium enema (DCBE) procedure volumes have decreased precipitously, and fecal occult blood test (FOBT) has decreased modestly (14).
Cancer prevention tests vs. cancer detection tests
The recent joint guideline (8) groups CRC screening tests into cancer prevention and cancer detection tests. Cancer prevention tests have the potential to image both cancer and polyps, whereas cancer detection tests have low sensitivity for polyps and typically lower sensitivity for cancer compared with that in cancer prevention tests (imaging tests). The ACG supports the division of screening tests into cancer prevention and cancer detection tests, but recommends a preferred cancer prevention test––colonoscopy every 10 years (Grade 1 B) and a preferred cancer detection test––annual fecal immunochemical test (FIT) to detect occult bleeding (Grade 1 B). All recommendations in this guideline are provided in Table 3.
Preferred CRC prevention test: colonoscopy every 10 years (Grade 1 B)
The ACG recommends that quality colonoscopy be offered first to patients aged ≥50 years (Table 3). A background discussion of screening colonoscopy, including discussion of quality in technical performance (which is deemed critical to screening colonoscopy) is found in Appendix B. Alternative CRC prevention tests are discussed in Appendix C. In clinical settings, in which economic issues preclude primary screening with colonoscopy, or for patients who decline colonoscopy, one of the alternative cancer prevention tests (Table 3, Appendix C) or the preferred cancer detection test, occult blood detection through the FIT (Table 3) should be offered.
Preferred cancer detection test: annual FIT (Grade 1 B)
The preferred cancer detection test is annual FIT. This test has superior performance characteristics when compared with older guaiac-based Hemoccult II cards (15,16,17); additionally, there were 10 and 12% gains in adherence with the FIT in the first two randomized controlled trials comparing the FIT with guaiac-based testing (18,19), The overall result of superior performance and improved adherence was a doubling in the detection of advanced lesions, with little loss of positive predictive value (18,19). The ACG supports the joint guideline recommendation that older guaiac-based fecal occult blood testing be abandoned as a method for CRC screening. Alternatives, such as the higher sensitivity guaiac-based Hemoccult Sensa and the fecal DNA test (Table 3), are discussed in Appendix D. However, because of more extensive data (compared with Hemoccult Sensa), and the high cost of fecal DNA testing, the ACG recommends the FIT as the preferred cancer detection test (Grade 1 B).
Age to begin screening in average-risk persons
The ACG continues to recommend that screening begin at age 50 years in average-risk persons (i.e., those without a family history of colorectal neoplasia) (Grade 1 B), except for African Americans. The ACG recommends that screening begin at age 45 years in African Americans (Grade 2 C). The rationale for this recommendation has been presented elsewhere (20).
The “average risk” population is large and complex with regard to risk. Certain other subgroups of the average-risk population might warrant initiation of screening at an earlier or later age, depending on their risk. For example, the age-adjusted risk of incident cancers (21) and prevalent adenomas (22,23,24,25) is greater in men than in women. However, delaying the onset of screening in women could result in a greater loss of life years in women who develop CRC in their 50s compared with that in men, as women on average live longer than men. Pending further study and evaluation of this issue, the ACG recommends that screening begin at age 50 years for both the genders (at age 45 years for African-American men and women).
In reviewing the literature, the writing committee also identified heavy cigarette smoking and obesity as linked to an increased risk of CRC and to the development of CRC at an earlier age. The clinical evidence supporting the increased risk in these groups is given in Appendix A. The current evidence supports a decision by clinicians in individual patients with an extreme smoking history or obesity to begin screening at an age earlier than 50 years and perhaps as early as 45 years. A formal recommendation to begin screening at an earlier age in smokers and obese patients will be re-evaluated as additional evidence appears.
Family history screening
Single first-degree relative with CRC or advanced adenoma (adenoma ≥1 cm in size, or with high-grade dysplasia or villous elements) diagnosed at age ≥60 years.
Recommended screening: same as average risk (colonoscopy every 10 years beginning at age 50 years) (Grade 2 B).
Single first-degree relative with CRC or advanced adenoma diagnosed at age <60 years or two first-degree relatives with CRC or advanced adenomas.
Recommended screening: colonoscopy every 5 years beginning at age 40, or 10 years younger than age at diagnosis of the youngest affected relative (Grade 2 B).
The ACG recommendations for modification of the screening approach, according to family histories of colorectal polyps and cancer that are not suggestive of the Hereditary Non-polyposis Colorectal Cancer, are summarized in Table 3. Justification for these recommendations was outlined in the 2000 guideline (2). The major change in this guideline is that an increased level of screening is no longer recommended for a simple family history of adenomas in a first-degree relative. The earlier ACG recommendations were that adenomas and cancer in first-degree relatives be treated equally in modifying the family history. Many studies purporting to describe the risk of CRC in first-degree relatives of patients with adenomas could be considered to have evaluated the reverse risk, i.e., the risk of adenomas in first-degree relatives of patients with CRC. In particular, case–control studies addressing this issue have often delivered an odds ratio (rather than a true risk ratio) that describes the “risk of adenomas among relatives of a patient with colorectal cancer” instead of the “risk of colorectal cancer among relatives of a patient with adenoma(s).” A single study carried out colonoscopies in first-degree relatives of patients with large adenomas, and found these relatives to have an increased risk of either large adenomas or cancer (26). There are no similar studies carried out in first-degree relatives of patients with small tubular adenomas. It is well known that persons with only small tubular adenomas (<1 cm) and only low-grade dysplasia are not at an increased risk for developing CRC (27). From a genetic perspective, it makes little sense that their relatives should be considered at an increased risk. Recently, some studies have identified an extremely high prevalence of small tubular adenomas in screening populations (28). Continuation of the recommendation to screen first-degree relatives of patients with only small tubular adenomas could result in most of the population being screened at age 40, with doubtful benefit. From a practical perspective, many clinicians have found that patients are often not aware of whether their first-degree relatives had advanced adenomas vs. small tubular adenomas, or whether their family members had non-neoplastic vs. neoplastic polyps. Given these difficulties, the ACG now recommends that adenomas only be counted as equal to a family history of cancer when there is a clear history, or medical report containing evidence, or other evidence to indicate that family members had advanced adenomas (an adenoma ≥1 cm in size, or with high-grade dysplasia, or with villous elements). Patients without information on the nature of polyps in a family member can be encouraged to pursue such information, but because of confidentiality requirements, pursuit of such information by the treating physicians is typically not feasible.
Familial adenomatous polyposis
Patients with features of an inherited CRC syndrome should be advised to pursue genetic counseling and, if appropriate, genetic testing. Genetic counseling and informed consent are preferred over direct genetic testing, as current laws may not provide adequate protection with regards to life insurance, long-term care insurance, or disability insurance. Individuals with familial adenomatous polyposis (FAP) should undergo APC mutation testing and, if negative, MYH mutation testing. Patients with FAP or at risk of FAP based upon family history should undergo annual flexible sigmoidoscopy or colonoscopy, as appropriate, until such time when colectomy is deemed by both physician and patient as the best treatment (29). Patients with a retained rectum after total colectomy and ileorectal anastomosis, ileal pouch, after total proctocolectomy and ileal-pouch anal anastomosis, or stoma after total proctocolectomy and end ileostomy, should undergo endoscopic assessment approximately every 6–12 months after surgery, depending on the polyp burden seen (Grade 2 B). Individuals with oligopolyposis (<100 colorectal polyps) should be sent for genetic counseling, consideration of APC and MYH mutation testing, and individualized colonoscopy surveillance depending on the size, number, and pathology of polyps seen (Grade 2 C). Upper endoscopic surveillance is recommended in individuals with FAP or MAP (MYH-associated polyposis).
Hereditary non-polyposis colorectal cancer
Patients who meet the Bethesda criteria for hereditary non-polyposis colorectal cancer (30) should undergo microsatellite instability testing of their tumor, or an affected family member's tumor, and/or tumor immunohistochemical staining for mismatch repair proteins. Patients with positive tests can be offered genetic testing and when genetic testing is positive in a proband, at risk family members can be offered genetic testing. Those patients with positive genetic testing, or those at risk when genetic testing is unsuccessful in an affected proband, should undergo colonoscopy every 2 years beginning at age 20–25 years, until age 40 years, then annually thereafter (Grade 2 B).
SUMMARY OF CURRENT GUIDELINE UPDATES
Owing to its potential for a high level of effectiveness in CRC prevention and extensive study of outcomes associated with its use, quality colonoscopy every 10 years beginning at age 50 remains the preferred CRC screening strategy. Patients who decline colonoscopy, or for whom colonoscopy is unavailable, or not feasible should be offered one of the alternative CRC prevention tests (flexible sigmoidoscopy every 5–10 years or computed tomography, CT, colonography every 5 years) or the preferred CRC detection test (FIT). The CRC screening in average-risk persons should begin at age 50, except that in African Americans, screening should begin at age 45 years. A family history of polyps need not invoke earlier onset of screening or other adjustment in screening, unless there is convincing evidence that the polyps were advanced adenomas.
Risk factors under consideration for more intense screening in future guidelines (smokers and obese patients)
The ACG recommends that clinicians be aware of an increased risk of CRC in cigarette smokers and obese patients. This evidence is summarized below. The ACG does not recommend that screening be initiated earlier in these groups at this time. Clinicians should make special efforts to ensure that screening takes place in these groups. The ACG recommends additional study to characterize the potential benefits, harms, and cost-effectiveness of earlier screening in these groups.
Smoking is associated with up to 20% of all CRCs in the United States (31) and was one of the strongest predictors of CRC in the Physician's Health Study (32). As over 20% of Americans currently smoke (33), the increase in risk for CRC may be yet another major medical consequence of tobacco use within the United States and worldwide. Literature review reveals that people who have more than 20 pack-years of smoking have over 2–3 times the risk for colorectal adenomas as non-smokers (31). There is as much as a 30% increased risk for colon and rectal cancer in male and female smokers (34,35,36,37,38,39,40,41) and smoking may account for 12% of deaths from CRC (42,43). Smokers have perceptions which may decrease their likelihood to be screened (44).
An important observation that underscores the potential value of screening smokers earlier is the younger age at which smokers are diagnosed with CRC. Although there may be other factors that explain this observation, an age difference of at least 5 years between smokers and non-smokers with CRC has been noted in four separate populations over two decades (45,46,47). Smokers may also be more likely to present with an advanced stage of CRC than non-smokers (48). Two studies of patients undergoing screening colonoscopy showed that smoking was associated with a two-fold increase in risk for advanced neoplasia, similar or greater than that of having a first-degree relative with CRC (49,50). Although many studies show a predilection for distal colorectal neoplasia in smokers (34,35,47), the Iowa Women's Health Study showed that female smokers had a higher risk for proximal CRCs (51). This observation may be explained by an increase in microsatellite instability in smokers (52). Anderson et al. (53) observed that smokers are at a risk for advanced isolated proximal neoplasia, underscoring the need for complete colonic evaluation in smokers during colonoscopy.
Smoking can be measured by duration, intensity, and number of years since cessation. It has been shown that smokers recall details of their exposure quite accurately (54). Several studies have suggested that smoking one pack per day or more significantly increases the risk and mortality for CRC (38,39,40,41,42,43). It has also been observed that the risk of CRC (40) and mortality (42) may be increased after 20 pack-years or less of smoking exposure. The impact of quitting is as yet unclear, but it appears that the risk may continue to increase, perhaps as long as 20 years after smoking cessation (34,35,37,38,39,42).
Based on these data, the ACG recommends that special efforts be made to ensure that screening takes place in active smokers and those who have smoked for more than 20 pack-years. Initiation of screening at a younger age (as early as 45 years) may be shown to be beneficial and cost-effective in persons with more 20 pack-years of smoking. These recommendations, however, may be tempered by the presence of medical complications of smoking that reduce the impact of CRC screening on overall life expectancy. Additional study is warranted.
A consistent body of evidence supports the concept that both overweight and obese statuses are associated with an increased risk of CRC. The risk of CRC for obese patients compared with that for non-obese patients is increased by 1.5–2.8 fold (55,56,57,58,59,60).
Recent data from the NIH–AARP cohort found that body mass index (BMI) was related to CRC risk for younger (age 50–66 years) but not older (age 67–71 years) persons (60). The BMI was associated with an increased incidence of colon cancer in men and women but not with rectal cancer. For men, the relative risks for overweight (BMI 25–30) ranged from 1.44 to 1.53 and for obese (BMI>30–<40) from 1.57 to 2.39, respectively. Corresponding relative risks for women were 1.29–1.31 and 1.13–1.49, respectively. A meta-analysis of six studies estimated a 3% increase (95% CI, 2–4) in CRC risk per one unit increase in BMI (59). The pattern of fat distribution is important as it relates to the reported CRC risk. Abdominal obesity is a stronger risk factor than truncal obesity or BMI (59,61).
Obesity is also associated with colon adenomas (presence and size) (62,63,64). Overall, obesity approximately doubles the relative risk of adenomas, and is particularly associated with high-risk adenomas (≥1 cm, tubulovillous). The mechanisms by which obesity may promote colon carcinogenesis are discussed elsewhere (65,66,67,68,69,70).
Based on the apparent increased relative risks for CRC and adenomas, the ACG recommends that special efforts are warranted to ensure the screening takes place in obese and overweight patients. Initiation of screening at an earlier age (as early as 45 years) may be beneficial and cost-effective in obese patients. These recommendations, however, may be tempered by the presence of medical complications of obesity, which reduce the impact of CRC screening on overall life expectancy. Additional study is warranted.
Discussion of screening colonoscopy
The evidence that colonoscopy prevents incident CRCs and reduces the consequent mortality from CRC is indirect but substantial. No prospective randomized controlled trial, comparing colonoscopy with no screening, has been carried out. However in a randomized controlled trial, involving only 800 patients, in which flexible sigmoidoscopy with colonoscopy carried out for any polyp detected was compared with no screening, the screening strategy resulted in an 80% reduction in the incidence of CRC (71). In addition, at the University of Minnesota, a randomized controlled trial was carried out comparing annual vs. biennial fecal occult blood testing with rehydration with no screening. Screening resulted in a 20% incidence reduction in CRC, which appeared to have resulted from detection of large adenomas by fecal occult blood testing and subsequent colonoscopy and polypectomy (72). Cohort studies involving patients, who have undergone colonoscopy and polypectomy with apparent clearance of colonic neoplasia, have shown a 76–90% reduction in the incidence of CRC in comparison with reference populations (73,74). Case–control studies of colonoscopy showed a 50% reduction in mortality from CRC in a US Veterans Administration population (75) and there was an 80% reduction in the CRC incidence in the German population (76). Population-based studies in the United States have associated increases in the use of colonoscopy with earlier and more favorable stages in CRC presentation (77), and with reductions in the incidence of CRC (78). Additional evidence for a benefit from colonoscopy screening is extrapolated from case–control studies of sigmoidoscopy, which have shown mortality and incidence reductions of distal CRC of 60 (79) and 80% (80), respectively, in screening populations.
Major advantages of colonoscopy as a screening test include that it is widely available (81), examines the entire colon, allows single-session diagnosis and treatment, is comfortable when carried out with sedation, and is the only test recommended at 10-year intervals (2,3,4,5,6,7,8). The incremental benefit of colonoscopy over sigmoidoscopy is the detection of patients with proximal colon neoplasia (particularly advanced adenomas), as well as large hyperplastic polyps that are not associated with distal neoplasia (82,83). Overall, sigmoidoscopy detects 60–70% of the significant neoplasia detected by complete colonoscopy (23). The preference of most American patients is for highly effective strategies (84), as well as for strategies that provide high levels of comfort and thereby increase the chance that patients will return for additional testing (85). These are important rationales for the use of colonoscopy rather than sigmoidoscopy.
Screening colonoscopy can be associated with significant harm, particularly colonic perforation (86,87). Many perforations are related to polypectomy and because small polyps are so numerous, small polyp polypectomy perforations contribute substantially to the overall perforation risk (87). Perforations associated with removal of small polyps are unfortunate, because the overwhelming majority of these polyps will not harm patients. Effective removal of these polyps by cold snare polypectomy or biopsy techniques is possible, at least for very small polyps (88), and is not associated with either bleeding or perforation. In general, there are insufficient data available from randomized controlled trials to guide or mandate particular polypectomy techniques (89). Pending such trials, the ACG recommends that colonoscopists consider carefully the polypectomy techniques they utilize for small polyps with an aim to reduce the burden of perforation. On the other hand, the ACG acknowledges that use of effective polypectomy techniques is critical for adequate resection of larger polyps. Two studies have suggested that about one-quarter of incident cancers occurring after colonoscopy result from ineffective polypectomy (90,91). Overall, the perforation risk and the requirement for thorough bowel preparation are the major downsides of colonoscopy.
The ACG continues to recommend that colonoscopy be carried out at 10-year intervals in average-risk persons with normal initial examinations. The evidence to support the 10-year interval is indirect but substantial. First, the protective effect for distal CRC provided by sigmoidoscopy and polypectomy in case–control studies, although imperfect, has been shown to be prolonged (79,80). In the Kaiser Permanente case–control study (this study first established the benefit of endoscopic screening), the duration of mortality reduction was 10 years (79). In a recent study of flexible sigmoidoscopy, the duration of protection was 16 years (80). Observational data, in which colonoscopy has been carried out at an initial baseline examination and then was repeated 5 years later, showed a very low yield of advanced adenomas (92,93,94,95). Cost analyses of colonoscopy as a screening test for CRC have found cost-effectiveness at equal or greater levels than other screening strategies with a 10-year interval (5). Recent studies in which follow-up sigmoidoscopies were carried out after initial negative examinations (96,97) and population-based studies of symptomatic individuals with negative colonoscopies (98,99) have established that some patients present shortly after negative examinations with cancers or advanced adenomas. What is not clear is the interval at which a second examination would have to be carried out in order to alter the outcome in these cases. Thus, in the population-based study of symptomatic patients with negative colonoscopies in Manitoba, many patients with interval cancers presented in the first few years after the negative colonoscopy, and it is not clear that a second planned examination at 5 years would have altered the outcome. In addition, some biologic variation in the growth rates of tumors, (which is best established for tumors with microsatellite instability or the CpG Island Methylator Phenotype), contributes to the appearance of cancers shortly after negative examinations (100,101). There is little evidence that performing a second examination at 5 years can impact substantially the incidence of these cancers.
Despite these caveats, there is little doubt that the overall impact of colonoscopy depends critically on high-quality baseline examinations. Therefore, the ACG recommends that screening colonoscopies be carried out by appropriately trained and skilled examiners, who are dedicated to consistent performance of high-quality examinations and employ programmatic measurements to optimize the outcomes through continuous quality improvement processes (88,102).
The ACG has both endorsed (102) and developed (88) quality indicators for colonoscopy. Readers can consult these documents (88,102) for a full description of quality indicators for colonoscopy. A major focus of these quality indicators that bears importantly on the impact of colonoscopy at 10-year intervals, are those directed to the quality of mucosal inspection. In addition to using an appropriate technique and time for mucosal inspection, colonoscopists must have expertise in safe and effective bowel preparation. Mucosal inspection during screening colonoscopy should be meticulous. The examiner should perform a slow and obsessive examination, designed to expose all of the colonic mucosa and identify and remove the smallest and flattest adenomas and proximal colon hyperplastic polyps. Several studies have shown that colonoscopists vary dramatically in their detection rates of adenomas (103), and in two recent studies, colonoscopists were shown to differ substantially in their detection of large adenomas (104,105). Colonoscopists in clinical practice should measure their individual adenoma detection rates in the continuous quality improvement process. One or more adenomas should be detected in at least 25% of men aged ≥50 years and 15% of women aged ≥50 years (88,102). These recommendations are derived from screening colonoscopy studies (88,102). In addition, endoscopists should measure their withdrawal times by noting the time of cecal intubation and termination of the examination. These withdrawal times should average at least 6 min in normal colonoscopies, in which no biopsy or polypectomy is carried out. This recommendation is not meant to imply that every colonoscopic withdrawal must last 6 min, as some colons can be examined effectively in <6 min. Furthermore, future research may revise the optimal mean withdrawal time that represents quality colonoscopy. The ACG also recommends that in institutions in which endoscopists from multiple specialties practice, that clinical gastroenterologists should establish institution-wide continuous quality improvement programs, designed to enhance the mucosal inspection performance of all specialties. In particular, three major studies have now identified that colonoscopy by primary care physicians is more likely to result in missed CRC compared with the performance by gastroenterologists (106,107,108).
The rationale and importance of the continuous quality improvement programs is emphasized by recent studies, showing lower than anticipated rates of protection against CRC by colonoscopy and polypectomy. Thus, adenoma cohorts participating in dietary intervention trials in the United States (109,110) and chemoprevention trials (111) have experienced little or no reduction in CRC incidence, compared with that in general population risk. Although the risk in these cohorts might be anticipated to be higher than the general population, the observed incidence of cancer clearly exceeds that anticipated based on earlier cohort studies (73,74). Population-based studies have confirmed a reduction in the incidence of CRC associated with negative colonoscopy, but the reduction in incidence has been less than anticipated (98,99). In the Manitoba study, the reduction in incidence was <50% for the first 5 years after the index negative colonoscopy and increased to 72% at 10 years (98). This suggests that significant numbers of lesions present at the index colonoscopy were not detected.
Inadequate bowel preparation is common in the United States (112), and inadequate preparation has been shown to impair the detection of both small (112,113) and large (113) polyps, and has also been shown recently in prospective colonoscopy studies to correlate with polyp detection (114,115,116). Although several commercial bowel preparations are available, certain principles of preparation will enhance the effectiveness of each of these commercial preparations. Best established is the principle of “splitting,” in which at least half of the preparation is given on the day of the colonoscopy (116,117,118). When all of the bowel preparation is given on the day before examination and the interval between the last dose of preparation and the performance of colonoscopy is prolonged, the probability of poor preparation increases dramatically, particularly in the cecum and ascending colon (116,117,118). Splitting can be carried out with oral dosing of either polyethylene glycol (116,118) or sodium phosphate (116,117) preparations. The practice guidelines of the American Society of Anesthesiologists allow ingestion of clear liquids until 2 h before sedation (119). Recent guidelines for an effective and safe preparation are available (120) and have particularly emphasized the importance of aggressive hydration before and during the preparation, during the procedure, and after the procedure, especially when using oral sodium phosphate preparations (120).
Several recent technical developments can enhance the mucosal inspection process during colonoscopy. Pancolonic chromoendoscopy is effective for enhancing adenoma detection, but impractical for routine use (103). Narrow band imaging does not enhance mucosal inspection by endoscopists with high adenoma detection rates, but may be a useful teaching tool for enhancement of flat lesion detection by endoscopists with low adenoma detection rates (103). Wide-angle colonoscopy, cap-fitted colonoscopy, and the Third Eye Retroscope (Avantis Medical Systems, Sunnyvale, CA) are all under development as techniques to improve exposure of hidden mucosa during colonoscopy (103). The ACG recommends that clinical gastroenterologists follow actively the technical developments pertaining to mucosal inspection enhancement techniques and incorporate such techniques into practice, as they are proven to be both effective and practical. However, endoscopists should understand that no enhancement technique replaces the need for a meticulous inspection. Elements critical to high-quality mucosal inspection during colonoscopy and which should be incorporated into all colonoscopy practices are detailed in Table 4.
Although colonoscopy is widely available and reimbursed as a strategy for CRC prevention, in some health care systems economic factors place limits on the feasibility of screening colonoscopy. In such cases, or when patients decline colonoscopy, alternative CRC prevention tests or FIT are very acceptable alternatives (Table 3).
Alternative cancer prevention tests
Alternative CRC prevention tests are listed in Table 3. The rationale for flexible sigmoidoscopy as a CRC screening test was reviewed in the 2000 guideline. Since that time, the use of flexible sigmoidoscopy has declined dramatically in the United States (14), though its use is still prevalent in certain settings. Flexible sigmoidoscopy is fundamentally similar to colonoscopy, except that less of the colon is examined, bowel preparation on average is less effective, and patients are not sedated. Flexible sigmoidoscopy can be offered at either 5-year or 10-year intervals. In the past, flexible sigmoidoscopy has typically been recommended at 5-year intervals, and this approach may be best if the extent of the examination is limited, or if the examination is carried out by an individual with limited endoscopic skills. However, the protective effect of sigmoidoscopy is long (79,80). Furthermore, colonoscopy may have more protection against left-sided compared with right-sided colon cancers (99,101). Therefore, flexible sigmoidoscopy is carried out by highly skilled practitioners, it may be recommended at 10-year, rather than 5-year intervals (8).
Double contrast barium enema is no longer recommended as an alternative CRC prevention test, because its use has declined dramatically and also as its effectiveness for polyp detection is less than computed tomography (CT) colonography. The ACG considers that the DCBE could be used as a CRC screening test that is within the standard of care, if it is carried out by high volume operators with special interest and expertise in the technique. The rationale for DCBE over CT colonography is its low cost, but patients clearly prefer CT colonography (121,122). Only a few centers in the United States still perform sufficient volumes of screening DCBE to warrant its continued use.
CT colonography, every 5 years, is endorsed as an alternative to colonoscopy every 10 years because of its recent performance in the American College of Imaging Network Trial 6664 (also known as the National CT Colonography Trial) (123). Results from earlier multicenter trials in the United States ranged from excellent (124) to poor (121,125). The principle performance feature that justifies inclusion of CT colonography as a viable alternative in patients who decline colonoscopy, is that the sensitivity for polyps ≥1 cm in size in the most recent multicenter US trial was 90% (123). In this study, 25% of radiologists who were tested for entry into the trial but performed poorly were excluded from participation, and thus lower sensitivity might be expected in clinical practice. The CT colonography probably has a lower risk of perforation than colonoscopy in most settings, but for several reasons it is not considered the equivalent of colonoscopy as a screening strategy. First, the evidence to support an effect of endoscopic screening on prevention of incident CRC and mortality is overwhelming compared with that for CT colonography (see Appendix B). Second, the inability to detect polyps 5 mm and smaller, which constitutes 80% of colorectal neoplasms, and whose natural history is still not understood, necessitates performance of the test at 5-year, rather than 10-year intervals (8). This is likely to increase overall costs, if CT colonography is used as a primary strategy. Although management of polyps <1 cm in size is controversial, the ACG continues to recommend that patients with polyps 6 mm or larger be referred for polypectomy, as should patients with three or more polyps of any size read with high confidence (126). Polyps ≤5 mm in size interpreted with high confidence should be described in the CT colonography report (126). Unfortunately, false positives are common, and the specificity for polyps ≥1 cm in size in the National CT Colonography Trial was only 86%, with a positive predictive value of 23% (123). Thus, colonoscopy for polyps detected on CT colonography will often require long procedures, in order to verify absence of other polyps. False positives diminish cost-effectiveness by increasing follow-up colonoscopies and repeat CT colonographies to verify false positive status. The ACG recommends that asymptomatic patients be informed of the possibility of radiation risk associated with one or repeated CT colonography studies, though the exact risk associated with radiation is unclear (127,128). The value of extracolonic findings detected by CT colonography is mixed, with substantial costs associated with incidental findings, but occasional important extracolonic findings are detected such as asymptomatic cancers and large abdominal aortic aneurysms. As a final point, the ACG is also concerned about the potential impact of CT colonography on adherence and thus on polypectomy rates. Thus, if CT colonography substantially improves adherence, it should improve polypectomy rates and thereby reduce CRC, even if only large polyps are detected and referred for colonoscopy. On the other hand, if CT colonography largely displaces patients who would otherwise be willing to undergo colonoscopy, then polypectomy rates will fall substantially, which could significantly increase the CRC incidence (129). Thus, for multiple reasons, and pending additional study, CT colonography should be offered to patients who decline colonoscopy.
Alternative cancer detection tests
The alternative cancer detection tests are listed in Table 3. Hemoccult Sensa is an improved guaiac-based card for fecal occult blood testing. It has superior sensitivity to older guaiac-based cards, but the overall evidence is less than that supporting the FIT. Furthermore, the FIT resulted in improved adherence for CRC screening over card-based tests in two randomized controlled trials (18,19). Therefore, FITs are preferred over Hemoccult Sensa.
Fecal DNA testing has been evaluated in three different versions. The first (Version 1.0) included tests for point mutations in k-ras, APC, P53, mutations in the BAT26 microsatellite instability marker, and the DNA integrity assay. The sensitivity for cancer was superior to traditional guaiac-based occult blood testing, but the absolute sensitivity was 52% and disappointing considering the high cost of the test (130). After completion of the trial, it was learned that the DNA integrity assay, which had appeared to be the most promising element in the assay in early studies (131), was non-informative because of the instability of DNA during shipment. Subsequently, Version 1.1 has been commercialized, which includes the same DNA test used in Version 1.0, but includes technical improvements of gel-based DNA capture and buffer stabilization of long or redundant DNA critical to the DNA integrity assay. No screening test using Version 1.1 has been reported, but a trial in established CRCs identified 70% sensitivity and specificity of ∼95%, (specificity similar to Version 1.0) (132). Version 2.0 utilizes a simplified assay consisting of the DNA integrity assay and hypermethylation of the vimentin gene. No screening trial with Version 2.0 has been carried out, but a study in established CRCs shows sensitivity of 87% for cancer, but specificity fell to 82% (133). The latter specificity limits the frequency with which the test can be carried out reasonably. Given that the performance characteristics of the FIT are approximately equal to Versions 1.0, and 1.1, and superior to Version 2.0 with regard to specificity, and that FIT costs much less than fecal DNA testing, there is no rationale for primary use of fecal DNA testing as a CRC detection test. The value of combining FIT and fecal DNA testing is unknown. Additional disadvantages of fecal DNA testing include no established data on which to determine an optimal interval, and the lack of clinical recommendations on how to respond to patients who have positive DNA tests and negative colonoscopies. Although the recent guideline endorsing fecal DNA testing declined to recommend an interval for DNA testing, the ACG considers that testing at intervals <3 years would be cost prohibitive.