Screening for colorectal cancer: confuting the refuters

Abstract

Confute: To overwhelm in argument; to disprove what is false or erroneous (Funk and Wagnalls 1947).The U.S. Preventive Services Task Force for the first time in December 1995 recommended screening of the average-risk asymptomatic population for colorectal cancer.1Department of Health and Human Services Guide to clinical preventive services.Second edition. Report of the U.S. Preventive Services Task Force, Department of Health and Human Services,, Washington, DC1995Google Scholar The task force is an independent panel charged by the U.S. Department of Health and Human Services with recommending preventive services for primary care clinicians. They conduct impartial assessments of scientific evidence and base recommendations solely on the strength of available data. During the release of its new guideline, Dr. Harold Sox, Chairman of the Task Force, stressed that there is strong new evidence of efficacy for both flexible sigmoidoscopy and fecal occult blood test screening in reducing the great mortality and morbidity of colorectal cancer.Although the data supporting screening for colorectal cancer are as compelling as that for any of the major cancers, a number of recent publications and presentations continue to question the value of this approach. 2Alquist DA Wieand HS Moertel CG et al.Accuracy of fecal occult blood screening for colorectal neoplasia: a prospective study using Hemoccult and HemoQuant tests.JAMA. 1993; 269: 1262-1267Crossref PubMed Scopus (309) Google Scholar, 3Lang CA Ransohoff DF Fecal occult blood screening for colorectal cancer: is mortality reduced by chance selection for screening colonoscopy?.JAMA. 1994; 271: 1011-1013Crossref PubMed Scopus (129) Google Scholar, 4Toribara NW Sleisenger MH Screening for colorectal cancer.N Engl J Med. 1995; 332: 861-867Crossref PubMed Scopus (168) Google Scholar I will briefly review these supporting data, and then, using the evidence, will objectively confute those who continue to argue against implementation of standard colorectal cancer screening.Colorectal cancer definitely is a major health problem in the United States. We see about 155,000 new cases each year, and the disease causes nearly 55,000 deaths.5Wingo PA Tons T Bolden S Cancer statistics, 1995.CA Cancer J Clin. 1995; 45: 8-30Crossref PubMed Scopus (1795) Google Scholar The lifetime risk in this country is now over 6%. In spite of all the advances that have been made in medical and surgical care over the past 30 years, survival from the disease is still only about 52%. Colorectal cancer thus causes great personal pain and suffering, substantial loss of productivity, and the expenditure of billions of health care dollars. It is a tragic fact that the disease causes this much loss and yet is actually one of the most preventable or curable cancers when detected early.By its purest definition, screening is the use of a simple, affordable, and acceptable test to identify a subgroup of the at-risk population more likely to have a clinically significant lesion or abnormality in which it would be justified to perform more complex, expensive, and possibly invasive diagnostic tests. Compared with most malignancies, colorectal cancer is uniquely suitable for a screening approach. 6Bond JH Fecal occult blood testing for colorectal cancer: can we afford not to do this?.Gastroenterol Clin North Am. 1997; (in press)PubMed Google Scholar The disease is common and lethal. It has a relatively long preclinical phase. Safe and accurate diagnostic tests are available (mainly colonoscopy) and, without question, early detection improves survival. Two simple screening tests are available and have been extensively studied: the fecal occult blood test (FOBT) and flexible sigmoidoscopy.The University of Minnesota FOBT trial, involving over 47,000 volunteers, was the first long-term randomized trial of occult stool blood screening to report definitive endpoint results. 7Mandel JS Bond JH Church TR et al.Reducing mortality from colorectal cancer by screening for fecal occult blood.N Engl J Med. 1993; 328: 1365-1371Crossref PubMed Scopus (2898) Google Scholar In 1993, after 13 years of screening and follow-up, the trial observed a highly statistically significant 33% reduction in mortality from colorectal cancer of the group screened annually with rehydrated Hemoccult slides (Smith Kline Diagnostics, San Jose, Calif.). A second group screened every other year is showing a similar trend and is still being followed-up.Efficacy of flexible sigmoidoscopy remains unproven simply because there has not yet been a completed controlled trial of this modality. However, the classic cohort study by Gilbertsen and Nelms 8Gilbertsen VA Nelms JM The prevention of invasive cancer of the rectum.Cancer. 1978; 41: 1137-1139Crossref PubMed Scopus (229) Google Scholar at the University of Minnesota along with two case-control studies, one by Selby et al.9Selby JV Friedman GD Quesenberry Jr, CP et al.A case-control study of screening sigmoidoscopy and mortality from colorectal cancer.N Engl J Med. 1992; 326: 653-657Crossref PubMed Scopus (1564) Google Scholar from Kaiser Permanente in northern California and one by Newcomb et al.10Newcomb PA Norfleet RG Storer BE et al.Screening sigmoidoscopy and colorectal cancer mortality.J Natl Cancer Inst. 1992; 84: 1572-1575Crossref PubMed Scopus (864) Google Scholar from the Marshfield Clinic in Wisconsin, indicate a reduction in mortality from distal colorectal cancer of 60% to 85% as the result of sigmoidoscopy screening.Based on these impressive data, current recommendations for screening the asymptomatic average-risk population call for annual FOBTs and flexible sigmoidoscopy about every 5 years, beginning at 50 years of age. A positive screen is an indication for colonoscopy. Screening should be individualized according to age and comorbidity. These recommendations are now endorsed not only by the U.S. Preventive Services Task Force, but also by the American Cancer Society, the World Health Organization U.S. Collaborating Center for the Prevention of Colorectal Cancer, and by virtually all of the medical and surgical societies concerned with digestive diseases.11Levin B Murphy B Revision in American Cancer Society recommendations for the early detection of colorectal cancer.CA Cancer J Clin. 1992; 42: 296Crossref PubMed Scopus (136) Google Scholar, 12Winawer SJ St John DJ Bond JH et al.Prevention of colorectal cancer: guidelines based on new data.Bull World Health Organ. 1995; 73: 7PubMed Google ScholarRegrettably, implementation and widespread adoption of these recommendations by both the public and the medical community are being delayed, I believe, by critics of this approach who continue to raise a number of theoretic and real limitations of these two screening tests. For example, they repeatedly assert that the FOBT is insufficiently sensitive for detecting colorectal neoplasia, or that the mortality reduction in the Minnesota trial is simply due to random colonoscopy, not the discriminating power of the fecal occult blood screening test itself. Many still take the position that, although the results of the Minnesota trial are promising, more data are needed or the tests must be improved before a public health recommendation can be made. Others say that the cost of screening is prohibitive, or that this approach will not be cost-effective. In the following discussion, I will attempt to confute these refuters of standard colorectal cancer screening.FIRST ISSUE: FOBT screening for colorectal cancer is insufficiently sensitive for detecting cancersStudies of highly selected groups of patients, often using unorthodox screening techniques, have concluded that the Hemoccult test detects only a small fraction of existing cancers. 2Alquist DA Wieand HS Moertel CG et al.Accuracy of fecal occult blood screening for colorectal neoplasia: a prospective study using Hemoccult and HemoQuant tests.JAMA. 1993; 269: 1262-1267Crossref PubMed Scopus (309) Google Scholar However, published sensitivity calculations from the Minnesota trial actually indicate that very few cancers are missed using this approach.13Mandel JS Bond JH Bradley M et al.Sensitivity, specificity, and positive predictivity of the Hemoccult test in screening for colorectal cancers: the University of Minnesota's Colon Cancer Control Study.Gastroenterology. 1989; 97: 597Abstract PubMed Google Scholar Of the first 281 colorectal cancers detected in screened patients in that trial, all but 21 were detected as the direct result of screening, giving a sensitivity of 92%. Sensitivity calculations made in ongoing screening trials have been questioned because of the obvious fact that FOBT-negative subjects do not undergo complete structural examination of their large bowel, and false-negative screens therefore may not be accurately ascertained. However, follow-up in the Minnesota trial is virtually complete over 18 years, and very few unsuspected cancers have been diagnosed after subjects had negative screening tests. Furthermore, the high sensitivity found in the Minnesota trial recently has been confirmed by a report of 89% sensitivity for cancer in the Swedish trial, the only other of the five controlled trials to also employ the more sensitive rehydrated Hemoccult slides.14Kewenter J Engaras B Haglind E et al.Value of retesting subjects with a positive Hemoccult in screening for colorectal cancer.Br J Surg. 1990; 77: 1349Crossref PubMed Scopus (21) Google Scholar, 15Kewenter J Bjork S Haglind E et al.Screening and rescreening for colorectal cancer: a controlled trial of fecal occult blood testing in 27,700 subjects.Cancer. 1988; 62: 645Crossref PubMed Scopus (227) Google ScholarIt should also be emphasized that a simple, indirect screening test repeated at defined intervals does not necessarily have to be highly sensitive to be efficacious, as long as it reliably detects developing neoplasms before they become incurable. In other words, application of the test may have relatively low sensitivity at one time, whereas the screening program of repetitive testing over time may be highly sensitive and efficacious.SECOND ISSUE: In the Minnesota trial, a large number of cancers were detected by random colonoscopyThis assertion was first made by Lang and Ransohoff3Lang CA Ransohoff DF Fecal occult blood screening for colorectal cancer: is mortality reduced by chance selection for screening colonoscopy?.JAMA. 1994; 271: 1011-1013Crossref PubMed Scopus (129) Google Scholar in an analytical piece published in 1994. From a mathematical model using a number of rather arbitrary assumptions, they concluded that 33% to 50% of the mortality reduction reported by the Minnesota trial could have resulted from random colonoscopy performed in patients with false-positive Hemoccult tests who, by chance, happened to also have nonbleeding cancers.Investigators from the Minnesota trial responded to this criticism with actual experimental data that show the substantial discriminating power of fecal occult blood screening. 16Mandel JS Ederer F Church TR Bond JH Screening for colorectal cancer: which test is best? [letter].JAMA. 1994; 272: 1099Crossref PubMed Scopus (14) Google Scholar For example, they reported that there was a strong, positive correlation between the likelihood of detecting cancer and the fraction of six slides positive for each positive screening episode. If a positive screening test was merely a random excuse to do colonoscopy, obviously no such correlation could exist. In addition, the Minnesota trial's chief biostatistician, Dr. Fred Ederer re-ran this mathematical modeling analysis using the actual data from the trial rather than Lang and Ransohoff's estimates or assumptions. He found that a small independent colonoscopy effect did appear to exist, but it accounted for only 6% to 11% of the observed reduction in colorectal cancer mortality (personal communication with F. Ederer, MD, 1995).Whenever an indirect, simple screening test is followed by a definitive diagnostic test in a large enough population, there will be some random or accidental benefit from the diagnostic test itself. This is not only true of Hemoccult tests followed by colonoscopy for colon cancer, but is also true for prostate-specific antigen (PSA) testing followed by prostate biopsy, and mammography followed by breast biopsy. Although the majority of the benefit is directly due to the screening test, some additional benefit results from accidental application of the diagnostic test in a large population at risk for the given disease.THIRD ISSUE: The results of the Minnesota trial are promising, but more data are needed, or the tests must be improved, before recommendations can be made for screening of the average-risk publicTable 1Results of FOBT trials21Winawer SJ Bond JH Fecal occult blood test screening trials.in: Cancer of the colon, rectum and anus. McGraw-Hill, New York1995: 279Google ScholarMortality reduction (%)SurvivalStageControlled trials Minnesota36ImprovedShifted New York43ImprovedShifted Denmark19ImprovedShifted United KingdomPendingImprovedShifted SwedenPendingImprovedShiftedCase-control studies Kaiser Permanente31NANA Germany57NANA Open table in a new tab There are two conclusions, therefore, that we can now make with great confidence on the basis of these data. First, all of the preliminary and endpoint reports from trials and studies are consistently in favor of this approach to screening for colorectal cancer. Second, no additional data in the near future are either needed or likely that would appreciably modify the conclusions of the U.S. Preventive Services Task Force and other groups that recommend screening NOW.Are better fecal occult blood tests needed? The answer is yes. Now that screening has been shown to be efficacious, work should and will proceed to develop tests with similar sensitivity to the rehydrated Hemoccult slide, but with better specificity and other performance characteristics. The newer guaiac methods, such as the HemeSensa test (Smith Kline Diagnostics, San Jose, Calif.) and the emerging immunochemical tests, are very promising and are undergoing comparison field trials.22Allison JE Tekawa IS Ransom LJ et al.A comparison of fecal occult blood tests for colorectal cancer screening.N Engl J Med. 1996; 334: 155Crossref PubMed Scopus (523) Google Scholar, 23Van Dam J Bond JH Sivak MV Fecal occult blood screening for colorectal cancer.Arch Intern Med. 1995; 155: 2389Crossref PubMed Scopus (26) Google Scholar Because the greatest cost of screening is for the diagnostic evaluation prompted by a positive screen, any improvement in this method that will improve specificity without sacrificing sensitivity is likely to be cost-effective.FOURTH ISSUE: Screening for colorectal cancer is too costly and is not cost-effectiveScreening the average-risk population for colorectal cancer admittedly is very expensive. Ransohoff and Lang 24Ransohoff DF Lang CA Screening for colorectal cancer.N Engl J Med. 1991; 325: 37Crossref PubMed Scopus (148) Google Scholar estimated the cost of screening the entire 60 million Americans over the age of 50 years would total $1.5 billion per year. This estimate does not, however, consider cost-saving strategies and realistic predictions of screening compliance. The Congressional Budget Office and the Administration's Office of Management and Budget have estimated the actual cost of screening Medicare patients for colorectal cancer over 7 years to be $750 million and $1.1 billion, respectively.25Bond JH Levin B Colorectal cancer screening: the battle may be lost, but the war must be won.ASGE News. 1995; 3: 6Google ScholarBoth medical care payers and the public should realize that the cost of missing an early curable cancer, or of failing to prevent the disease, may be greater than the cost of screening. For example, the absolute cost of caring for a patient with incurable advanced cancer substantially exceeds that of colonoscopic polypectomy or of uncomplicated curable surgical resection of a cancer still localized to the bowel.Table 2Cost-effectiveness of lifesaving interventions cost/year life saved)Mandatory motorcycle helmets$2,000Colorectal cancer screening$25,000Breast cancer screening$35,000Dual airbags in cars$120,000Smoke detectors in homes$210,000Seat belts in school buses$2,800,000Data compiled by the Harvard Center for Risk Analysis. 28Tengs TO Adams ME Pliskin JS et al.Five-hundred life-saving interventions and their cost-effectiveness.Risk Anal. 1995; 15: 369-390Crossref PubMed Scopus (1037) Google Scholar Open table in a new tab Another cost-effectiveness analysis recently published by Lieberman30Lieberman DA Cost-effectiveness model for colon cancer screening.Gastroenterology. 1995; 109: 1781Abstract Full Text PDF PubMed Scopus (308) Google Scholar used a simple and very useful mathematical model for comparing several different colorectal cancer screening programs and for determining the key variables that affect program cost-effectiveness. Of five different screening strategies, annual fecal occult blood screening alone was the most cost-effective program. Adding flexible sigmoidoscopy every 5 years to annual fecal occult blood tests increased cancer prevention 2.2-fold because of its special ability to detect precancerous adenomas. Not surprisingly, the most important variable affecting cost and cost-effectiveness was compliance.This analysis by Lieberman 30Lieberman DA Cost-effectiveness model for colon cancer screening.Gastroenterology. 1995; 109: 1781Abstract Full Text PDF PubMed Scopus (308) Google Scholar is unique in that it also estimated for the first time the cost of doing no screening. This cost, incurred as the result of medical care required for cancers that would not be prevented or cured, was $677 to $1000 per potential screenee for each 10-year period.Summary and conclusionsAlthough some time in the future reliable genetic screening tests or direct screening colonoscopy may prove in clinical trials to be feasible, acceptable, and effective, today and for the foreseeable future, the best hope of reducing the great morbidity and mortality of the second most common cancer killer of Americans is to implement widespread fecal occult blood and flexible sigmoidoscopy screening. The combination of these two tests is feasible, highly effective, affordable, and has the potential to reduce colorectal cancer mortality by at least 50%. At the present time, about 52% of the 155,000 people who develop these cancers in the United States are cured of their disease. If, as the result of screening, we can increase this survival rate to 75%, it would result in the salvage of over 35,000 lives each year.What additional action is now required to accomplish this objective? Even if we have adequately sensitive screening tests, an effective evaluation for those with a positive screen, and proper follow-up surveillance, the critical final link or prerequisite to the success of screening is adequate compliance by both the lay public and by the medical community, especially primary-care physicians, who are in the best position to deliver preventive services. To this end, during Digestive Disease Week 1996 the American Digestive Health Foundation officially launched the Digestive Health Initiative (DHI) Colorectal Cancer Campaign. Following the very successful DHI Ulcer Campaign, this effort is designed to increase the awareness of both the public and managed care groups of the importance of colorectal cancer and the value of screening. Using a creative public promotion program plus the assistance of knowledgable volunteers from our membership, this campaign will not only benefit the 60 million Americans at risk for this disease, but will also reinforce the leadership position of our subspecialty in digestive health in this era that emphasizes primary care. Confute: To overwhelm in argument; to disprove what is false or erroneous (Funk and Wagnalls 1947). The U.S. Preventive Services Task Force for the first time in December 1995 recommended screening of the average-risk asymptomatic population for colorectal cancer.1Department of Health and Human Services Guide to clinical preventive services.Second edition. Report of the U.S. Preventive Services Task Force, Department of Health and Human Services,, Washington, DC1995Google Scholar The task force is an independent panel charged by the U.S. Department of Health and Human Services with recommending preventive services for primary care clinicians. They conduct impartial assessments of scientific evidence and base recommendations solely on the strength of available data. During the release of its new guideline, Dr. Harold Sox, Chairman of the Task Force, stressed that there is strong new evidence of efficacy for both flexible sigmoidoscopy and fecal occult blood test screening in reducing the great mortality and morbidity of colorectal cancer. Although the data supporting screening for colorectal cancer are as compelling as that for any of the major cancers, a number of recent publications and presentations continue to question the value of this approach. 2Alquist DA Wieand HS Moertel CG et al.Accuracy of fecal occult blood screening for colorectal neoplasia: a prospective study using Hemoccult and HemoQuant tests.JAMA. 1993; 269: 1262-1267Crossref PubMed Scopus (309) Google Scholar, 3Lang CA Ransohoff DF Fecal occult blood screening for colorectal cancer: is mortality reduced by chance selection for screening colonoscopy?.JAMA. 1994; 271: 1011-1013Crossref PubMed Scopus (129) Google Scholar, 4Toribara NW Sleisenger MH Screening for colorectal cancer.N Engl J Med. 1995; 332: 861-867Crossref PubMed Scopus (168) Google Scholar I will briefly review these supporting data, and then, using the evidence, will objectively confute those who continue to argue against implementation of standard colorectal cancer screening. Colorectal cancer definitely is a major health problem in the United States. We see about 155,000 new cases each year, and the disease causes nearly 55,000 deaths.5Wingo PA Tons T Bolden S Cancer statistics, 1995.CA Cancer J Clin. 1995; 45: 8-30Crossref PubMed Scopus (1795) Google Scholar The lifetime risk in this country is now over 6%. In spite of all the advances that have been made in medical and surgical care over the past 30 years, survival from the disease is still only about 52%. Colorectal cancer thus causes great personal pain and suffering, substantial loss of productivity, and the expenditure of billions of health care dollars. It is a tragic fact that the disease causes this much loss and yet is actually one of the most preventable or curable cancers when detected early. By its purest definition, screening is the use of a simple, affordable, and acceptable test to identify a subgroup of the at-risk population more likely to have a clinically significant lesion or abnormality in which it would be justified to perform more complex, expensive, and possibly invasive diagnostic tests. Compared with most malignancies, colorectal cancer is uniquely suitable for a screening approach. 6Bond JH Fecal occult blood testing for colorectal cancer: can we afford not to do this?.Gastroenterol Clin North Am. 1997; (in press)PubMed Google Scholar The disease is common and lethal. It has a relatively long preclinical phase. Safe and accurate diagnostic tests are available (mainly colonoscopy) and, without question, early detection improves survival. Two simple screening tests are available and have been extensively studied: the fecal occult blood test (FOBT) and flexible sigmoidoscopy. The University of Minnesota FOBT trial, involving over 47,000 volunteers, was the first long-term randomized trial of occult stool blood screening to report definitive endpoint results. 7Mandel JS Bond JH Church TR et al.Reducing mortality from colorectal cancer by screening for fecal occult blood.N Engl J Med. 1993; 328: 1365-1371Crossref PubMed Scopus (2898) Google Scholar In 1993, after 13 years of screening and follow-up, the trial observed a highly statistically significant 33% reduction in mortality from colorectal cancer of the group screened annually with rehydrated Hemoccult slides (Smith Kline Diagnostics, San Jose, Calif.). A second group screened every other year is showing a similar trend and is still being followed-up. Efficacy of flexible sigmoidoscopy remains unproven simply because there has not yet been a completed controlled trial of this modality. However, the classic cohort study by Gilbertsen and Nelms 8Gilbertsen VA Nelms JM The prevention of invasive cancer of the rectum.Cancer. 1978; 41: 1137-1139Crossref PubMed Scopus (229) Google Scholar at the University of Minnesota along with two case-control studies, one by Selby et al.9Selby JV Friedman GD Quesenberry Jr, CP et al.A case-control study of screening sigmoidoscopy and mortality from colorectal cancer.N Engl J Med. 1992; 326: 653-657Crossref PubMed Scopus (1564) Google Scholar from Kaiser Permanente in northern California and one by Newcomb et al.10Newcomb PA Norfleet RG Storer BE et al.Screening sigmoidoscopy and colorectal cancer mortality.J Natl Cancer Inst. 1992; 84: 1572-1575Crossref PubMed Scopus (864) Google Scholar from the Marshfield Clinic in Wisconsin, indicate a reduction in mortality from distal colorectal cancer of 60% to 85% as the result of sigmoidoscopy screening. Based on these impressive data, current recommendations for screening the asymptomatic average-risk population call for annual FOBTs and flexible sigmoidoscopy about every 5 years, beginning at 50 years of age. A positive screen is an indication for colonoscopy. Screening should be individualized according to age and comorbidity. These recommendations are now endorsed not only by the U.S. Preventive Services Task Force, but also by the American Cancer Society, the World Health Organization U.S. Collaborating Center for the Prevention of Colorectal Cancer, and by virtually all of the medical and surgical societies concerned with digestive diseases.11Levin B Murphy B Revision in American Cancer Society recommendations for the early detection of colorectal cancer.CA Cancer J Clin. 1992; 42: 296Crossref PubMed Scopus (136) Google Scholar, 12Winawer SJ St John DJ Bond JH et al.Prevention of colorectal cancer: guidelines based on new data.Bull World Health Organ. 1995; 73: 7PubMed Google Scholar Regrettably, implementation and widespread adoption of these recommendations by both the public and the medical community are being delayed, I believe, by critics of this approach who continue to raise a number of theoretic and real limitations of these two screening tests. For example, they repeatedly assert that the FOBT is insufficiently sensitive for detecting colorectal neoplasia, or that the mortality reduction in the Minnesota trial is simply due to random colonoscopy, not the discriminating power of the fecal occult blood screening test itself. Many still take the position that, although the results of the Minnesota trial are promising, more data are needed or the tests must be improved before a public health recommendation can be made. Others say that the cost of screening is prohibitive, or that this approach will not be cost-effective. In the following discussion, I will attempt to confute these refuters of standard colorectal cancer screening. FIRST ISSUE: FOBT screening for colorectal cancer is insufficiently sensitive for detecting cancersStudies of highly selected groups of patients, often using unorthodox screening techniques, have concluded that the Hemoccult test detects only a small fraction of existing cancers. 2Alquist DA Wieand HS Moertel CG et al.Accuracy of fecal occult blood screening for colorectal neoplasia: a prospective study using Hemoccult and HemoQuant tests.JAMA. 1993; 269: 1262-1267Crossref PubMed Scopus (309) Google Scholar However, published sensitivity calculations from the Minnesota trial actually indicate that very few cancers are missed using this approach.13Mandel JS Bond JH Bradley M et al.Sensitivity, specificity, and positive predictivity of the Hemoccult test in screening for colorectal cancers: the University of Minnesota's Colon Cancer Control Study.Gastroenterology. 1989; 97: 597Abstract PubMed Google Scholar Of the first 281 colorectal cancers detected in screened patients in that trial, all but 21 were detected as the direct result of screening, giving a sensitivity of 92%. Sensitivity calculations made in ongoing screening trials have been questioned because of the obvious fact that FOBT-negative subjects do not undergo complete structural examination of their large bowel, and false-negative screens therefore may not be accurately ascertained. However, follow-up in the Minnesota trial is virtually complete over 18 years, and very few unsuspected cancers have been diagnosed after subjects had negative screening tests. Furthermore, the high sensitivity found in the Minnesota trial recently has been confirmed by a report of 89% sensitivity for cancer in the Swedish trial, the only other of the five controlled trials to also employ the more sensitive rehydrated Hemoccult slides.14Kewenter J Engaras B Haglind E et al.Value of retesting subjects with a positive Hemoccult in screening for colorectal cancer.Br J Surg. 1990; 77: 1349Crossref PubMed Scopus (21) Google Scholar, 15Kewenter J Bjork S Haglind E et al.Screening and rescreening for colorectal cancer: a controlled trial of fecal occult blood testing in 27,700 subjects.Cancer. 1988; 62: 645Crossref PubMed Scopus (227) Google ScholarIt should also be emphasized that a simple, indirect screening test repeated at defined intervals does not necessarily have to be highly sensitive to be efficacious, as long as it reliably detects developing neoplasms before they become incurable. In other words, application of the test may have relatively low sensitivity at one time, whereas the screening program of repetitive testing over time may be highly sensitive and efficacious. Studies of highly selected groups of patients, often using unorthodox screening techniques, have concluded that the Hemoccult test detects only a small fraction of existing cancers. 2Alquist DA Wieand HS Moertel CG et al.Accuracy of fecal occult blood screening for colorectal neoplasia: a prospective study using Hemoccult and HemoQuant tests.JAMA. 1993; 269: 1262-1267Crossref PubMed Scopus (309) Google Scholar However, published sensitivity calculations from the Minnesota trial actually indicate that very few cancers are missed using this approach.13Mandel JS Bond JH Bradley M et al.Sensitivity, specificity, and positive predictivity of the Hemoccult test in screening for colorectal cancers: the University of Minnesota's Colon Cancer Control Study.Gastroenterology. 1989; 97: 597Abstract PubMed Google Scholar Of the first 281 colorectal cancers detected in screened patients in that trial, all but 21 were detected as the direct result of screening, giving a sensitivity of 92%. Sensitivity calculations made in ongoing screening trials have been questioned because of the obvious fact that FOBT-negative subjects do not undergo complete structural examination of their large bowel, and false-negative screens therefore may not be accurately ascertained. However, follow-up in the Minnesota trial is virtually complete over 18 years, and very few unsuspected cancers have been diagnosed after subjects had negative screening tests. Furthermore, the high sensitivity found in the Minnesota trial recently has been confirmed by a report of 89% sensitivity for cancer in the Swedish trial, the only other of the five controlled trials to also employ the more sensitive rehydrated Hemoccult slides.14Kewenter J Engaras B Haglind E et al.Value of retesting subjects with a positive Hemoccult in screening for colorectal cancer.Br J Surg. 1990; 77: 1349Crossref PubMed Scopus (21) Google Scholar, 15Kewenter J Bjork S Haglind E et al.Screening and rescreening for colorectal cancer: a controlled trial of fecal occult blood testing in 27,700 subjects.Cancer. 1988; 62: 645Crossref PubMed Scopus (227) Google Scholar It should also be emphasized that a simple, indirect screening test repeated at defined intervals does not necessarily have to be highly sensitive to be efficacious, as long as it reliably detects developing neoplasms before they become incurable. In other words, application of the test may have relatively low sensitivity at one time, whereas the screening program of repetitive testing over time may be highly sensitive and efficacious. SECOND ISSUE: In the Minnesota trial, a large number of cancers were detected by random colonoscopyThis assertion was first made by Lang and Ransohoff3Lang CA Ransohoff DF Fecal occult blood screening for colorectal cancer: is mortality reduced by chance selection for screening colonoscopy?.JAMA. 1994; 271: 1011-1013Crossref PubMed Scopus (129) Google Scholar in an analytical piece published in 1994. From a mathematical model using a number of rather arbitrary assumptions, they concluded that 33% to 50% of the mortality reduction reported by the Minnesota trial could have resulted from random colonoscopy performed in patients with false-positive Hemoccult tests who, by chance, happened to also have nonbleeding cancers.Investigators from the Minnesota trial responded to this criticism with actual experimental data that show the substantial discriminating power of fecal occult blood screening. 16Mandel JS Ederer F Church TR Bond JH Screening for colorectal cancer: which test is best? [letter].JAMA. 1994; 272: 1099Crossref PubMed Scopus (14) Google Scholar For example, they reported that there was a strong, positive correlation between the likelihood of detecting cancer and the fraction of six slides positive for each positive screening episode. If a positive screening test was merely a random excuse to do colonoscopy, obviously no such correlation could exist. In addition, the Minnesota trial's chief biostatistician, Dr. Fred Ederer re-ran this mathematical modeling analysis using the actual data from the trial rather than Lang and Ransohoff's estimates or assumptions. He found that a small independent colonoscopy effect did appear to exist, but it accounted for only 6% to 11% of the observed reduction in colorectal cancer mortality (personal communication with F. Ederer, MD, 1995).Whenever an indirect, simple screening test is followed by a definitive diagnostic test in a large enough population, there will be some random or accidental benefit from the diagnostic test itself. This is not only true of Hemoccult tests followed by colonoscopy for colon cancer, but is also true for prostate-specific antigen (PSA) testing followed by prostate biopsy, and mammography followed by breast biopsy. Although the majority of the benefit is directly due to the screening test, some additional benefit results from accidental application of the diagnostic test in a large population at risk for the given disease. This assertion was first made by Lang and Ransohoff3Lang CA Ransohoff DF Fecal occult blood screening for colorectal cancer: is mortality reduced by chance selection for screening colonoscopy?.JAMA. 1994; 271: 1011-1013Crossref PubMed Scopus (129) Google Scholar in an analytical piece published in 1994. From a mathematical model using a number of rather arbitrary assumptions, they concluded that 33% to 50% of the mortality reduction reported by the Minnesota trial could have resulted from random colonoscopy performed in patients with false-positive Hemoccult tests who, by chance, happened to also have nonbleeding cancers. Investigators from the Minnesota trial responded to this criticism with actual experimental data that show the substantial discriminating power of fecal occult blood screening. 16Mandel JS Ederer F Church TR Bond JH Screening for colorectal cancer: which test is best? [letter].JAMA. 1994; 272: 1099Crossref PubMed Scopus (14) Google Scholar For example, they reported that there was a strong, positive correlation between the likelihood of detecting cancer and the fraction of six slides positive for each positive screening episode. If a positive screening test was merely a random excuse to do colonoscopy, obviously no such correlation could exist. In addition, the Minnesota trial's chief biostatistician, Dr. Fred Ederer re-ran this mathematical modeling analysis using the actual data from the trial rather than Lang and Ransohoff's estimates or assumptions. He found that a small independent colonoscopy effect did appear to exist, but it accounted for only 6% to 11% of the observed reduction in colorectal cancer mortality (personal communication with F. Ederer, MD, 1995). Whenever an indirect, simple screening test is followed by a definitive diagnostic test in a large enough population, there will be some random or accidental benefit from the diagnostic test itself. This is not only true of Hemoccult tests followed by colonoscopy for colon cancer, but is also true for prostate-specific antigen (PSA) testing followed by prostate biopsy, and mammography followed by breast biopsy. Although the majority of the benefit is directly due to the screening test, some additional benefit results from accidental application of the diagnostic test in a large population at risk for the given disease. THIRD ISSUE: The results of the Minnesota trial are promising, but more data are needed, or the tests must be improved, before recommendations can be made for screening of the average-risk publicTable 1Results of FOBT trials21Winawer SJ Bond JH Fecal occult blood test screening trials.in: Cancer of the colon, rectum and anus. McGraw-Hill, New York1995: 279Google ScholarMortality reduction (%)SurvivalStageControlled trials Minnesota36ImprovedShifted New York43ImprovedShifted Denmark19ImprovedShifted United KingdomPendingImprovedShifted SwedenPendingImprovedShiftedCase-control studies Kaiser Permanente31NANA Germany57NANA Open table in a new tab There are two conclusions, therefore, that we can now make with great confidence on the basis of these data. First, all of the preliminary and endpoint reports from trials and studies are consistently in favor of this approach to screening for colorectal cancer. Second, no additional data in the near future are either needed or likely that would appreciably modify the conclusions of the U.S. Preventive Services Task Force and other groups that recommend screening NOW.Are better fecal occult blood tests needed? The answer is yes. Now that screening has been shown to be efficacious, work should and will proceed to develop tests with similar sensitivity to the rehydrated Hemoccult slide, but with better specificity and other performance characteristics. The newer guaiac methods, such as the HemeSensa test (Smith Kline Diagnostics, San Jose, Calif.) and the emerging immunochemical tests, are very promising and are undergoing comparison field trials.22Allison JE Tekawa IS Ransom LJ et al.A comparison of fecal occult blood tests for colorectal cancer screening.N Engl J Med. 1996; 334: 155Crossref PubMed Scopus (523) Google Scholar, 23Van Dam J Bond JH Sivak MV Fecal occult blood screening for colorectal cancer.Arch Intern Med. 1995; 155: 2389Crossref PubMed Scopus (26) Google Scholar Because the greatest cost of screening is for the diagnostic evaluation prompted by a positive screen, any improvement in this method that will improve specificity without sacrificing sensitivity is likely to be cost-effective. There are two conclusions, therefore, that we can now make with great confidence on the basis of these data. First, all of the preliminary and endpoint reports from trials and studies are consistently in favor of this approach to screening for colorectal cancer. Second, no additional data in the near future are either needed or likely that would appreciably modify the conclusions of the U.S. Preventive Services Task Force and other groups that recommend screening NOW. Are better fecal occult blood tests needed? The answer is yes. Now that screening has been shown to be efficacious, work should and will proceed to develop tests with similar sensitivity to the rehydrated Hemoccult slide, but with better specificity and other performance characteristics. The newer guaiac methods, such as the HemeSensa test (Smith Kline Diagnostics, San Jose, Calif.) and the emerging immunochemical tests, are very promising and are undergoing comparison field trials.22Allison JE Tekawa IS Ransom LJ et al.A comparison of fecal occult blood tests for colorectal cancer screening.N Engl J Med. 1996; 334: 155Crossref PubMed Scopus (523) Google Scholar, 23Van Dam J Bond JH Sivak MV Fecal occult blood screening for colorectal cancer.Arch Intern Med. 1995; 155: 2389Crossref PubMed Scopus (26) Google Scholar Because the greatest cost of screening is for the diagnostic evaluation prompted by a positive screen, any improvement in this method that will improve specificity without sacrificing sensitivity is likely to be cost-effective. FOURTH ISSUE: Screening for colorectal cancer is too costly and is not cost-effectiveScreening the average-risk population for colorectal cancer admittedly is very expensive. Ransohoff and Lang 24Ransohoff DF Lang CA Screening for colorectal cancer.N Engl J Med. 1991; 325: 37Crossref PubMed Scopus (148) Google Scholar estimated the cost of screening the entire 60 million Americans over the age of 50 years would total $1.5 billion per year. This estimate does not, however, consider cost-saving strategies and realistic predictions of screening compliance. The Congressional Budget Office and the Administration's Office of Management and Budget have estimated the actual cost of screening Medicare patients for colorectal cancer over 7 years to be $750 million and $1.1 billion, respectively.25Bond JH Levin B Colorectal cancer screening: the battle may be lost, but the war must be won.ASGE News. 1995; 3: 6Google ScholarBoth medical care payers and the public should realize that the cost of missing an early curable cancer, or of failing to prevent the disease, may be greater than the cost of screening. For example, the absolute cost of caring for a patient with incurable advanced cancer substantially exceeds that of colonoscopic polypectomy or of uncomplicated curable surgical resection of a cancer still localized to the bowel.Table 2Cost-effectiveness of lifesaving interventions cost/year life saved)Mandatory motorcycle helmets$2,000Colorectal cancer screening$25,000Breast cancer screening$35,000Dual airbags in cars$120,000Smoke detectors in homes$210,000Seat belts in school buses$2,800,000Data compiled by the Harvard Center for Risk Analysis. 28Tengs TO Adams ME Pliskin JS et al.Five-hundred life-saving interventions and their cost-effectiveness.Risk Anal. 1995; 15: 369-390Crossref PubMed Scopus (1037) Google Scholar Open table in a new tab Another cost-effectiveness analysis recently published by Lieberman30Lieberman DA Cost-effectiveness model for colon cancer screening.Gastroenterology. 1995; 109: 1781Abstract Full Text PDF PubMed Scopus (308) Google Scholar used a simple and very useful mathematical model for comparing several different colorectal cancer screening programs and for determining the key variables that affect program cost-effectiveness. Of five different screening strategies, annual fecal occult blood screening alone was the most cost-effective program. Adding flexible sigmoidoscopy every 5 years to annual fecal occult blood tests increased cancer prevention 2.2-fold because of its special ability to detect precancerous adenomas. Not surprisingly, the most important variable affecting cost and cost-effectiveness was compliance.This analysis by Lieberman 30Lieberman DA Cost-effectiveness model for colon cancer screening.Gastroenterology. 1995; 109: 1781Abstract Full Text PDF PubMed Scopus (308) Google Scholar is unique in that it also estimated for the first time the cost of doing no screening. This cost, incurred as the result of medical care required for cancers that would not be prevented or cured, was $677 to $1000 per potential screenee for each 10-year period. Screening the average-risk population for colorectal cancer admittedly is very expensive. Ransohoff and Lang 24Ransohoff DF Lang CA Screening for colorectal cancer.N Engl J Med. 1991; 325: 37Crossref PubMed Scopus (148) Google Scholar estimated the cost of screening the entire 60 million Americans over the age of 50 years would total $1.5 billion per year. This estimate does not, however, consider cost-saving strategies and realistic predictions of screening compliance. The Congressional Budget Office and the Administration's Office of Management and Budget have estimated the actual cost of screening Medicare patients for colorectal cancer over 7 years to be $750 million and $1.1 billion, respectively.25Bond JH Levin B Colorectal cancer screening: the battle may be lost, but the war must be won.ASGE News. 1995; 3: 6Google Scholar Both medical care payers and the public should realize that the cost of missing an early curable cancer, or of failing to prevent the disease, may be greater than the cost of screening. For example, the absolute cost of caring for a patient with incurable advanced cancer substantially exceeds that of colonoscopic polypectomy or of uncomplicated curable surgical resection of a cancer still localized to the bowel. Data compiled by the Harvard Center for Risk Analysis. 28Tengs TO Adams ME Pliskin JS et al.Five-hundred life-saving interventions and their cost-effectiveness.Risk Anal. 1995; 15: 369-390Crossref PubMed Scopus (1037) Google Scholar Another cost-effectiveness analysis recently published by Lieberman30Lieberman DA Cost-effectiveness model for colon cancer screening.Gastroenterology. 1995; 109: 1781Abstract Full Text PDF PubMed Scopus (308) Google Scholar used a simple and very useful mathematical model for comparing several different colorectal cancer screening programs and for determining the key variables that affect program cost-effectiveness. Of five different screening strategies, annual fecal occult blood screening alone was the most cost-effective program. Adding flexible sigmoidoscopy every 5 years to annual fecal occult blood tests increased cancer prevention 2.2-fold because of its special ability to detect precancerous adenomas. Not surprisingly, the most important variable affecting cost and cost-effectiveness was compliance. This analysis by Lieberman 30Lieberman DA Cost-effectiveness model for colon cancer screening.Gastroenterology. 1995; 109: 1781Abstract Full Text PDF PubMed Scopus (308) Google Scholar is unique in that it also estimated for the first time the cost of doing no screening. This cost, incurred as the result of medical care required for cancers that would not be prevented or cured, was $677 to $1000 per potential screenee for each 10-year period. Summary and conclusionsAlthough some time in the future reliable genetic screening tests or direct screening colonoscopy may prove in clinical trials to be feasible, acceptable, and effective, today and for the foreseeable future, the best hope of reducing the great morbidity and mortality of the second most common cancer killer of Americans is to implement widespread fecal occult blood and flexible sigmoidoscopy screening. The combination of these two tests is feasible, highly effective, affordable, and has the potential to reduce colorectal cancer mortality by at least 50%. At the present time, about 52% of the 155,000 people who develop these cancers in the United States are cured of their disease. If, as the result of screening, we can increase this survival rate to 75%, it would result in the salvage of over 35,000 lives each year.What additional action is now required to accomplish this objective? Even if we have adequately sensitive screening tests, an effective evaluation for those with a positive screen, and proper follow-up surveillance, the critical final link or prerequisite to the success of screening is adequate compliance by both the lay public and by the medical community, especially primary-care physicians, who are in the best position to deliver preventive services. To this end, during Digestive Disease Week 1996 the American Digestive Health Foundation officially launched the Digestive Health Initiative (DHI) Colorectal Cancer Campaign. Following the very successful DHI Ulcer Campaign, this effort is designed to increase the awareness of both the public and managed care groups of the importance of colorectal cancer and the value of screening. Using a creative public promotion program plus the assistance of knowledgable volunteers from our membership, this campaign will not only benefit the 60 million Americans at risk for this disease, but will also reinforce the leadership position of our subspecialty in digestive health in this era that emphasizes primary care. Although some time in the future reliable genetic screening tests or direct screening colonoscopy may prove in clinical trials to be feasible, acceptable, and effective, today and for the foreseeable future, the best hope of reducing the great morbidity and mortality of the second most common cancer killer of Americans is to implement widespread fecal occult blood and flexible sigmoidoscopy screening. The combination of these two tests is feasible, highly effective, affordable, and has the potential to reduce colorectal cancer mortality by at least 50%. At the present time, about 52% of the 155,000 people who develop these cancers in the United States are cured of their disease. If, as the result of screening, we can increase this survival rate to 75%, it would result in the salvage of over 35,000 lives each year.