Screening colonoscopy: option or preference?

Abstract

See related articles on p. 517, 524. A conviction is growing in the endoscopy community that colonoscopy is the best way to screen for colorectal cancer, even in average-risk people.1Lieberman DA Smith FW Screening for colon malignancy with colonoscopy.Am J Gastroenterology. 1991; 86: 946-951PubMed Google Scholar, 2Rogge JD Elmore MF Mahoney SJ Brown ED Troiano FP Wagner DR et al.Low-cost, office-based, screening colonoscopy.Am J Gastroenterol. 1994; 10: 1775-1779Google Scholar, 3Bhattacharya I Sack EM Screening colonoscopy: the cost of common sense.Lancet. 1996; 347: 1744-1745Abstract PubMed Scopus (27) Google Scholar Others are moving in this direction, but slowly. Since 1997, widely accepted guidelines have included colonoscopy as an option for screening, along with fecal occult blood testing, sigmoidoscopy, and double-contrast barium enema.4Winawer SJ Fletcher RH Miller L Godlee F Stolar MH Mulrow CD et al.Colorectal cancer screening: clinical guidelines and rationale.Gastroenterology. 1997; 112: 594-642Abstract Full Text Full Text PDF PubMed Scopus (1817) Google Scholar, 5Byers T Levin B Rothenberger D Dodd GD Smith RA American Cancer Society Guidelines for screening and surveillance for early detection of colorectal polyps and cancer: update 1997.CA Cancer J Clin. 1997; 47: 154-160Crossref PubMed Scopus (448) Google Scholar Medicare has begun to pay for screening colonoscopy with relatively liberal indications: the presence of a family history of colorectal cancer. But no group has said that colonoscopy should be the first choice for screening in average-risk people. How close is colonoscopy to becoming the clear favorite? The strength of the evidence that colonoscopic screening is effective, although convincing to many of us and accepted in current guidelines, is not beyond reproach. The evidence base is not as strong as for other colorectal cancer screening tests, for which controlled studies with colorectal cancer death as the outcome are available. Fecal occult blood test screening is supported by three published randomized controlled trials, all showing effectiveness,6Towler BP Irwig L Glasiou P Weller D Kewenter J Screening for colorectal cancer using the faecal occult blood test, Hemoccult.BMJ. 1998; 317: 559-565Crossref PubMed Scopus (425) Google Scholar and sigmoidoscopy by an excellent casecontrol study7Selby JV Friedman GD Quesenberry Jr, CP Weiss NS A case-control study of screening sigmoidoscopy and mortality from colorectal cancer.N Engl J Med. 1992; 326: 653-657Crossref PubMed Scopus (1560) Google Scholar and others.8Newcomb PA Norfleet RG Storer BE Surawicz TS Marcus PM Screening sigmoidoscopy and colorectal cancer mortality.J Natl Cancer Inst. 1992; 84: 1572-1575Crossref PubMed Scopus (863) Google Scholar, 9Muller AD Sonnenberg A Protection by endoscopy against death from colorectal cancer.Arch Intern Med. 1995; 155: 1741-1748Crossref PubMed Scopus (375) Google Scholar In contrast, the case for colonoscopic screening is built up from several lines of evidence. Perhaps the strongest evidence is from the National Polyp Study.10Winawer SJ Zauber AG Ho MN O'Brien MJ Gottlieb LS Sternberg SS et al.Prevention of colorectal cancer by colonoscopic polypectomy.N Engl J Med. 1993; 329: 1977-1981Crossref PubMed Scopus (3833) Google Scholar Polypectomy and colonoscopic surveillance in patients with polyps (not average-risk people) reduced the incidence of colorectal cancer over the next several years compared with patients in other published series, at other times and places. Differences between the National Polyp Study cohort and the comparison cohorts could be responsible for some of the observed effect. However, the effect is so large (76% to 90% fewer colorectal cancers) that it is implausible that selection bias accounted for all of the apparent effectiveness. Other evidence supports the effectiveness of colonoscopy. It can be argued that colonoscopy must be effective because it is demonstrably better able to detect polyps and localized cancers than other tests, such as fecal occult blood test and sigmoidoscopy, for which strong evidence of effectiveness is available. Similarly, if sigmoidoscopy is effective in preventing colorectal cancer deaths, how could colonoscopy, which uses similar technology and has a longer reach, not be? Moreover, everything we know about the pathogenesis of colorectal cancer suggests that colonoscopy, coupled with excision of clinically important neoplasms when found, would prevent both cancer incidence and deaths. But all this said, colonoscopy earns only a C-level recommendation by the stringent standards of the U.S. Preventive Services Task Force.11Report of the U.S. Preventive Services Task Force.in: Guide to Clinical Preventive Services. 2nd ed. Williams and Wilkins, Baltimore1996Google Scholar Screening policy is not decided by effectiveness alone.4Winawer SJ Fletcher RH Miller L Godlee F Stolar MH Mulrow CD et al.Colorectal cancer screening: clinical guidelines and rationale.Gastroenterology. 1997; 112: 594-642Abstract Full Text Full Text PDF PubMed Scopus (1817) Google Scholar Politics aside, other forms of evidence should also be taken into account when choosing which screening strategy is best. How big is the difference in effect size between strategies? The additional effectiveness of colonoscopic screening, over and above other screening strategies, should be large enough to justify the disadvantages of the test. How safe and convenient is the test? Colonoscopy is suspect on both counts; it causes major complications and deaths, though rarely, and requires an unpleasant bowel preparation and a day off work. Is the workforce in place to accomplish the recommended testing? The endoscopy community would be well advised to undertake a careful analysis of its capacity to meet screening-induced demand for colonoscopy. Cost is an important issue and the cost of colonoscopy is extraordinarily high for a screening test, especially at a time when containing the rising cost of health care is a major public policy issue. Cost-effectiveness, the cost per year of life saved, matters, too. Apparently screening colonoscopy is about as cost effective as other, less expensive screening tests.12Wagner JL Tunis S Brown M Ching A Almeida R The cost effectiveness of colorectal cancer screening in average-risk adults.in: Prevention and early detection of colorectal cancer. : WB Saunders, Philadelphia1996Google Scholar However, the benefits of screening follow the costs by many years. Managed care organizations have a notoriously short time horizon these days. It is not attractive to them that large costs now would be rewarded 10 to 20 years hence, when the patients may have left the organization, the current leaders are no longer in place, and the organization itself may no longer exist. Colonoscopy and the other ways of screening for colorectal cancer differ from each other in all of these dimensions. These differences need to be taken into account in screening guidelines because individual patients place very different values on the various consequences of the tests.13Pignone M Bucholtz D Harris R Patient preference for colon cancer screening.F General Intern Med. 1999; 14: 432-437Crossref PubMed Scopus (134) Google Scholar For example, some patients might want to do everything possible to prevent colorectal cancer, even though they are not themselves at increased risk, and so prefer screening colonoscopy despite its cost, inconvenience, and small risk. Others who are less concerned about their risk of colorectal cancer might prefer a much simpler approach such as fecal occult blood testing if they were told this is a reasonable option. For these reasons, expert groups have chosen to offer all reasonable screening tests as options and not to override individual preferences in the interest of consistency. “One size fits all” is fine as long as there is only one legitimate option, as is the case with mammography or Pap smears. But that is not so for colorectal cancer screening. Two studies published in this issue add to the case for screening colonoscopy. The study by Rex et al.14Rex DK Khan AM Shah P Newton J Cummings OW Screening colonoscopy in asymptomatic average risk African Americans.Gastrointest Endosc. 2000; 51: 524-570Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar calls attention to the possibility that asymptomatic African Americans might have an increased prevalence of colorectal neoplasia and so should be screened differently from other average-risk people. African Americans are already known to have a 10% to 30% higher colorectal cancer incidence and mortality rates than whites.15Cancer Facts and Figures–1999. American Cancer Society, Atlanta1999Google Scholar Although the causes might be partly genetic, it is more likely they are behavioral (for example, resulting from differences in exercise, dietary red meat, folic acid, and obesity)16Cancer Causes Control. 1999; 10: 167-180Crossref PubMed Scopus (100) Google Scholar or related to differences in access to screening. If so, the excess deaths are preventable. Rex et al.14Rex DK Khan AM Shah P Newton J Cummings OW Screening colonoscopy in asymptomatic average risk African Americans.Gastrointest Endosc. 2000; 51: 524-570Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar found that the prevalence of adenomas in asymptomatic, average-risk African Americans was similar to that reported in other studies but that right-sided lesions seemed to be more common. Others have also found proximal lesions more common in African Americans.17Troisi RJ Freedman AN Devesa SS Incidence of colorectal carcinoma in the U.S. An update of trends by gender, race, age, subsite, and stage, 1975-94.Cancer. 1999; 85: 1670-1676Crossref PubMed Scopus (166) Google Scholar The authors' suggestion that, until this issue is resolved, blacks should be screened by colonoscopy rather than sigmoidoscopy is premature or at least not sufficiently supported by the data in their article. The Rex study can, after all, only take us so far in deciding whether African Americans really do have a higher prevalence of right-sided polyps. It is limited, despite extraordinary efforts on the part of the investigators, by a non-representative sample of African Americans and by a small sample size. As a result, observed rates of polyps and their locations might well have resulted from selection bias or chance. As the authors point out, more definitive observation should come from large studies of colonoscopic screening, which are now under way. Although the results of the Rex study are not definitive, it is time to consider whether the risk in African Americans and other increased-risk groups is high enough to justify screening them differently. Emotion, above numbers, often drives screening decisions in people with increased risk. For example, many clinicians and patients believe that people with a family history of colorectal cancer should be screened by colonoscopy. Yet the risk, which in the simple case of a person with a first-degree relative with colorectal cancer is about doubled,18St. John DJ McDermott FT Hopper JL Debney EA Johnson WR Hughes ES Cancer risk in relatives of patients with common colorectal cancer.Ann Intern Med. 1993; 118: 785-790Crossref PubMed Scopus (339) Google Scholar is not very different from the increase in risk that accompanies an average-risk person getting 10 years older (also about double). Similarly, neoplasia occurs more proximally in the colon of average risk people as they age.19Devisa SS Chow W-H Variation in colorectal cancer incidence in the United States by subsite of origin.Cancer. 1993; 71: 3839-3926Crossref PubMed Scopus (18) Google Scholar To my knowledge no one advocates more aggressive screening in average-risk people as they get older. The article by Inadami and Sonnenberg20Inadomi JM Sonnenberg A The impact of colorectal cancer screening on life expectancy.Gastrointest Endosc. 2000; 51: 517-523Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar models the effects of several screening strategies, including colonoscopy, on life expectancy. Colonoscopy is the winner, as expected from its properties and from other models.21Lieberman DA Cost-effectiveness model for colon cancer screening.Gastroenterology. 1995; 109: 1781-1790Abstract Full Text PDF PubMed Scopus (308) Google Scholar Modeling, using generally accepted estimates of sensitivity, specificity, incidence of polyps and cancer, and other rates, is a useful way of estimating the effectiveness of screening strategies when strong, original research, such as randomized trials, are not feasible or will take a long time. This one provides another boost to the assertion that colonoscopic screening is the most effective strategy. But given the vagaries of models, many of which seem to be black boxes (this one included), we still need more original research. Would patients go along with advice to have a screening colonoscopy? Certainly most are willing to have a colonoscopy if another screening test, such as a fecal occult blood test, is positive. But it is another matter to undergo a colonoscopy when there is no other reason to suspect cancer. Few patients, armed with information about their options, now choose colonoscopy as a primary screening test.13Pignone M Bucholtz D Harris R Patient preference for colon cancer screening.F General Intern Med. 1999; 14: 432-437Crossref PubMed Scopus (134) Google Scholar But that could change. Although there is much to learn about determinants of colorectal screening behaviors,22Vernon S Participation in colorectal cancer screening: a review.J Natl Cancer Inst. 1997; 89: 1406-1422Crossref PubMed Scopus (671) Google Scholar it is apparent that screening rates can be high in settings that really try to make them so. Also, rates have increased with time for tests (such as mammography and Pap smears) that are supported by strong evidence of effectiveness. In the future, three issues will be important to the acceptance of screening colonoscopy: cost, safety, and the emergence of new screening tests. The up-front cost of colonoscopy, which is far greater than any other screening test, is a major barrier to its acceptance, cost-effectiveness notwithstanding. In one study in which patients were provided information about the pros and cons of various colorectal cancer screening options, preferences were most sensitive to out-of-pocket expenses.13Pignone M Bucholtz D Harris R Patient preference for colon cancer screening.F General Intern Med. 1999; 14: 432-437Crossref PubMed Scopus (134) Google Scholar But if patients were to accept the higher initial cost, screening colonoscopy might result in no more colonoscopies than widely accepted, simpler screening tests. One model of the effects of screening showed that fecal occult blood test screening starting at age 50 years would, because of the high rate of false-positive tests, result is 2.8 diagnostic colonoscopies by age 85 or death, more than the twice-in-a-lifetime screening advocated by colonoscopy enthusiasts.4Winawer SJ Fletcher RH Miller L Godlee F Stolar MH Mulrow CD et al.Colorectal cancer screening: clinical guidelines and rationale.Gastroenterology. 1997; 112: 594-642Abstract Full Text Full Text PDF PubMed Scopus (1817) Google Scholar Whether or not it is the primary screening test, colonoscopy dominates the cost, as well as the effectiveness and risk, of all other screening strategies. All of them, if they are positive, eventually lead to diagnostic, and in some cases surveillance, colonoscopy. For example, 9.8% of fecal occult blood tests with hydrated specimens are positive, with about 50 false-positives for every cancer found.23Mandel JS Bond JH Church TR Snover DC Bradley GM Schuman LM et al.Reducing mortality from colorectal cancer by screening for fecal occult blood.N Engl J Med. 1993; 328: 1365-1371Crossref PubMed Scopus (2889) Google Scholar Similarly, finding polyps on sigmoidoscopy may lead to full colonoscopy in many cases, although the wisdom of complete structural examination of the colon in patients with small distal tubular adenomas is debated.24Wallace MB Farraye FA Kemp JA Colonoscopy for small adenomas [letter].Ann Intern Med. 1999; 130: 701Crossref PubMed Scopus (2) Google Scholar Although colonoscopy has not been a major target for cost-containment efforts so far, in a marketplace competing mainly on cost it might become so. If colonoscopy charges are driven downward, I hope they will not go so low as to make them economically unfeasible, as appears to be the case in many real-world settings for sigmoidoscopy.25Lewis JD Asch DA Barriers to office-based screening sigmoidoscopy: Does reimbursement cover costs?.Ann Intern Med. 1999; 130: 525-530Crossref PubMed Scopus (75) Google Scholar Much of the effectiveness and efficiency of colonoscopy rests in its ability to stratify patients according to their risk of developing colorectal cancer. No other procedure can provide information on the main predictors of risk, multiple polyps, large polyp size, and villous histology.26Atkins WS Morson BC Cuzick J Long-term risk of colorectal cancer after excision of rectosigmoid adenomas.N Engl J Med. 1992; 326: 658-662Crossref PubMed Scopus (947) Google Scholar This information can be used to create more efficient surveillance strategies and willingness to do so has grown with time, with each successive guideline being less aggressive with low-risk patients.27Bond JH for the Practice Parameters Committee of the American College of Gastroenterology Polyp guideline: diagnosis, treatment and surveillance for patients with nonfamilial colorectal polyps.Ann Intern Med. 1993; 119: 836-843Crossref PubMed Scopus (216) Google Scholar, 28Bond JH Colorectal surveillance for neoplasia: an overview.Gastrointest Endosc. 1999; 49: S35-S40Abstract Full Text Full Text PDF PubMed Scopus (19) Google Scholar Wise use of risk stratification could lead to substantial reductions in the cost of screening at little sacrifice in effectiveness. For example, in models of screening based on existing information about test performance, reducing the interval between screening examinations within the range now considered usual practice resulted in steep rises in cost with little increase in effectiveness.12Wagner JL Tunis S Brown M Ching A Almeida R The cost effectiveness of colorectal cancer screening in average-risk adults.in: Prevention and early detection of colorectal cancer. : WB Saunders, Philadelphia1996Google Scholar As for safety, the major complications, bleeding and perforation, seem to be occurring at about the same rate, about 1/1000, as years ago, despite advances in endoscope technology. Other screening tests do not cause harm directly, so colonoscopy stands out in this regard. True, endoscopists and their patients rarely experience an adverse outcome. But if screening colonoscopy were to be widely practiced and current complication rates continued, many lives would be lost. It appears many more lives would be saved than lost, for a net gain.4Winawer SJ Fletcher RH Miller L Godlee F Stolar MH Mulrow CD et al.Colorectal cancer screening: clinical guidelines and rationale.Gastroenterology. 1997; 112: 594-642Abstract Full Text Full Text PDF PubMed Scopus (1817) Google Scholar But even so, the potential danger of colonoscopy runs counter to the generally accepted belief that one should not harm people who have no complaints.29McCormick J Health promotion: the ethical dimension.Lancet. 1994; 344: 390-391Abstract PubMed Scopus (40) Google Scholar In the coming years, screening colonoscopy will compete with new screening tests now in development. Two seem especially promising. Virtual colonoscopy is a spiral CT combined with graphic reconstruction of the colonic lumen, using powerful computers. At its best, virtual colonoscopy produces strikingly precise images. Moreover, the procedure is no more uncomfortable than fiberoptic colonoscopy or barium enema and there is no risk of perforation or bleeding. An early report suggests that test accuracy is comparable with colonoscopy in patients at high risk for colorectal neoplasia.30Fenlon HM Nunes DP Schroy III, PC Barish MA Clarke PD Ferrucci JT A comparison of virtual and conventional colonoscopy for the detection of colorectal polyps.N Engl J Med. 1999; 341: 1496-1503Crossref PubMed Scopus (666) Google Scholar However, there is still a lot to learn about sensitivity, false-positive rates, cost in average-risk people, and when the study is performed by the average radiologist. Another possibility is testing for DNA abnormalities in stool.31Ahlquist DA Harrington JJ Shuber AP Detection of altered DNA in stool: feasibility for colorectal neoplasia screening [abstract].Gastroenterology. 1999; 116: G1618Google Scholar Carcinogenesis is accompanied by, and presumably caused by, a succession of somatic changes in genetic material and the altered DNA is shed into the bowel lumen. Field testing of this method has begun. When considering these new tests, it is important to remember that neither will replace colonoscopy for diagnosis and (in many cases) treatment. A randomized trial of screening colonoscopy in average-risk people is badly needed. The National Cancer Institute has just funded a 3-year feasibility study of such a trial. Let us hope that the study will be successful and lead to a more definitive study of the effectiveness, safety, and cost of colonoscopy screening.