Digestive Diseases and Sciences, Vol. 36, No. 7 (July 1991), pp. 924-928
Flexible Sigmoidoscopy May Be Ineffective for Secondary Prevention of Colorectal Cancer in Asymptomatic, Average-Risk Men P. GREGORY FOUTCH, DO, MS, FACP, HUGH MAI, MD, KATHLEEN PARDY, RN, JAMES A. DISARIO, MD, R.K. MANNE, MD, and DAVID KERR, MD
Asymptomatic men (N = 114) 50 years of age or older had screening for colorectal neoplasia with flexible sigmoidoscopy followed by colonoscopy regardless of the sigmoidoscopic result. Our study objective was to determine the prevalence o f patients having isolated adenomatous polyps in a proximal colonic segment in the absence of a distal index neoplasm within reach of the sigmoidoscope. Through the combined use of sigmoidoscopy and colonoscopy, adenomatous polyps were detected in 47 of 114 individuals (41%). A total of 88 adenomas was found. Seventeen patients had isolated neoplasms in proximal colonic segments in the absence of distal adenomas. These patients represented 15% of screened subjects (17 of 114) and 20% of individuals who lacked adenomas on sigmoidoscopy (17 of 84). The majority of proximal neoplasms were small (50 years; (2) absence of symptoms or signs suggestive of colorectal neoplasms including pain, change in bowel habits, or bleeding; (3) absence of risk factors indicating screening colonoscopy, such as ulcerative colitis or evidence of a polyposis or nonpolyposis hereditary colon cancer syndrome; (4) absence of a familial history for colorectal cancer in a first-degree relative. Patients having prior barium enema, sigmoidoscopy, colonoscopy, or polypectomy were excluded. A complete blood count was obtained if a recent result was not available in the medical record, and anemic individuals were not enrolled in the study. One hundred fourteen men (age range 50-79 years, mean 64 years) met entry criteria and consented to participate. Hemoccult testing of stool was performed on 105 subjects, and results were negative in all instances. Nine men did not return their hemoccult cards but none of these individuals had neoplasms on FS or colonoscopy, so the undetermined hemoccult result was presumed to be negative. The mean hemoglobin concentrations for the group was 14.8 g/100 ml. Methods. After preparing the bowel with two enemas, FS was performed by second-year gastroenterology fellows using a 60-cm instrument. The depth of insertion of the sigmoidoscope and absence or presence and location of all polyps was recorded. Colonoscopy, using conscious sedation, was performed after FS at a later date. Colyte was used to clean the colon. Size of polyps was estimated by comparison to the open span of standard biopsy forceps. All diminutive polyps were biopsied and larger lesions were removed with snare cautery. Tissue specimens were fixed in formalin prior to routine histological evaluation. Two pathologists with an interest in colorectal neoplasia reviewed all specimens. Number and histology of adenomas detected within and beyond the extent Digestive Diseases and Sciences, Vol. 36, No. 7 (July 1991)
TABLE 1. PRESENCE OR ABSENCE OF ADENOMAS IN 114 PATIENTS DETECTED BY SIGMOIDOSCOPY OR COLONOSCOPY PERFORMED ABOVE EXTENT OF SIGMOIDOSCOPIC EXAMINATION
Adenomas detected
Subjects N (%) 67 21 9 17
(59) (18) (8.0) (15)
By sigmoidoscopy Absent Present* Presentt Absent
By colonoscopy Absent Absent Present Present
*An invasive c a r c i n o m a in a tubulovillous a d e n o m a was detected in this group. t A n invasive carcinoma in a tubular a d e n o m a and cancer in situ in a tubulovillous a d e n o m a were detected in this group.
of the FS examination were recorded for each patient. For purposes of this study only, cancer, tubular, tubulovillous, and villous adenomas were regarded as positive results. Criteria established by the World Health Organization were used to classify adenomas (14). Lipomas, lymphoid nodules, and hyperplastic polyps were biopsied or removed but were considered negative findings. The study was approved by the Investigational Review Board and Human Subjects Subcommittee. RESULTS The flexible sigmoidoscope was inserted 30-60 cm (mean, 56 cm), and the cecum was reached with the colonoscope in all instances. Bowel preparations were judged to be adequate, and there were no complications associated with either procedure. Through the combined use of sigmoidoscopy and colonoscopy, adenomatous polyps were identified in 47 of 114 individuals (41%) (Table 1). A total of 88 adenomas was detected and sizes ranged from 2 to 25 mm (mean, 6.2 ram). Ten percent were -> 1.0 cm. N u m b e r and location of neoplasms included the following: rectum, 13 (15%); sigmoid colon, 24 (27%); descending colon, 11 (13%); transverse colon, 12 (14%); ascending colon, 16 (18%); cecum, 12 (14%). Seventy-one adenomas were tubular (81%), 16 tubulovillous (18%), and one was villous (1.0%). Endoscopic results for detection of colorectal adenomas are shown in Table 1. Seventeen of 114 subjects (15%) had all neoplasms detected by colonoscopy beyond the reach of the sigmoidoscope. Eighty-four subjects (67 plus 17, Table 1) did not have adenomas on FS, yet 17 of these (20%) had proximal neoplasms diagnosed by colonoscopy. Nine of 30 patients (30%) with adenomas detected by FS had one or more synchronous proximal neoplasm identified by c o l o n o s c o p y (Table 1). These individuals served as a reference for compar-
925
FOUTCH ET AL TABLE 2. COMPARISONOF GROUPS OF PATIENTS WITH PROXIMAL ADENOMAS* ASSOCIATED(GROUP 1) AND NOT ASSOCIATED(GRouP 2) WITH SYNCHRONOUSNEOPLASIAIN DISTAL COLON AND RECTUM
Characteristics
Total number of proximal adenomas Size of proximal adenomas (mm) Range Mean Patients with proximal adenomas ->1.0 cm Patients with histological types of adenomas Tubular Tubulovillous Villous Cancer Patients in whom adenomas were solitary multiple?
Group 1 (N=9 patients)
Group 2 (N=17 patients)
17
29
2-25 6.6
2-25 6.1
1 (11%)
2 (12%)
7 (78%) 2 (22%) l (11%) 0
15 (88%) 4 (24%) 0 0
6 (67%) 3 (33%)
11 (65%) 6 (35%)
*Refers to neoplasms detected beyond the extent of the sigmoidoscopic examination. ?->2 adenomas.
ison to patients (N = 17) who did not have similar lesions on the sigmoidoscopic examination. Comparison of characteristics of groups of patients with adenomas in proximal sites associated and not associated with syDchronous index neoplasms in the distal colon and rectum is shown in Table 2. No important differences between the groups were observed. DISCUSSION We used colonoscopy as a screening tool and prospectively compared results to fiberoptic sigmoidoscopy in average-risk, asymptomatic men. This approach renders a true approximation of prevalence and distribution of neoplasia in this important cohort. Forty-one percent of patients had adenomatous polyps, and this figure is consistent with findings from autopsy studies that show an overall adult prevalence of 21-47% percent (5, 6, 8, 9, 15-17). The majority of adenomas were tubular, and most were small in size. Only 10% of neoplasms were - 1 . 0 cm, and this finding correlates well with the observation of others. Results from analysis of necropsy specimens show that 4-13% of adults with adenomas have neoplasms 1.0 cm or larger in diameter (6, 8). A surprising dearth of information exists regarding the efficacy of FS for screening asymptomatic
926
individuals for neoplasia (18, 19). However, cumulative results from studies suggest that adenomas are detected in 3-20% and cancers in 0.3-4.3% of cases (19-23). Sigmoidoscopy alone identified adenomas and cancer in 26% and 2.6% of our patients, respectively. This high diagnostic yield for neoplasia may have occurred because of selection bias from patients requesting screening and because our study population consisted of elderly men (mean age 64 years). Results from studies show that adenomatous polyps are more common in men than women and increase in frequency with increasing age (5, 6, 24). Arminski and McLean determined the incidence and location of 797 adenomatous polyps in 330 colons reviewed at autopsy (6). Approximately 40% of men between ages 60 and 69 had adenomas, and 41% were located in the sigmoid colon or rectum. Our high detection rate for cancer might be explained by the concomitant high prevalence of adenomas, which are important in the pathogenesis of malignancy. These findings emphasize the need for aggressive screening for neoplasia in this group of individuals. Thirty percent of our patients (nine of 30) with distal adenomas detected by FS had concomitant proximal neoplasms diagnosed by colonoscopy. Previous studies have shown that synchronous polyps occur in 20-50% of cases and it is on the basis of this association that colonoscopy is recommended when an index adenoma is detected by sigmoidoscopy (1, 4, 25). The important finding in our study is that 15% of screened subjects (17 of 114) and 20% of individuals with a normal FS examination (17 of 84) have undiagnosed adenomas in a proximal location. These persons may be at increased risk for cancer, yet they cannot be easily identified by our current screening practices. It is possible that distal hyperplastic polyps (HPP) can serve as markers for proximal adenomas, but this point is controversial (26). Our study design did not consider HPP a significant finding because the Health and Public Policy Committee of the American College of Physicians does not recommend colonoscopy for purposes of detecting proximal adenomas when nonneoplastic polyps are detected by FS (27). However, prevalence for HPP in this study was 31%. If HPP as well as adenomas were considered valid markers for proximal neoplasms, then only 12% of screened subjects and 17% with a totally normal FS examination would have had undiagnosed proximal neoplasms. We have initiated a prospective controlled trial designed to deDigestive Diseases and Sciences, Vol. 36, No. 7 (July 1991)
F L E X I B L E SIGMOIDOSCOPY IN ASYMPTOMATIC MEN
termine the significance of distal HPP, and at present our results are preliminary (28). Previous studies have shown that cancer risk increases with size, number, and increasing villous components of the adenoma (1, 16, 29, 30). The majority of isolated proximal polyps in our series were small tubular adenomas and were of low risk. Yet some adenomas were large, multiple, and contained villous elements (Table 2), and the future rate of growth of the small lesions is unpredictable. There are data to suggest that small polyps cannot be ignored. O'Brien et al detected high-grade dysplasia in nearly 6.0% of adenomatous polyps -
Flexible Sigmoidoscopy May Be Ineffective for Secondary Prevention of Colorectal Cancer in Asymptomatic, Average-Risk Men P. GREGORY FOUTCH, DO, MS, FACP, HUGH MAI, MD, KATHLEEN PARDY, RN, JAMES A. DISARIO, MD, R.K. MANNE, MD, and DAVID KERR, MD
Asymptomatic men (N = 114) 50 years of age or older had screening for colorectal neoplasia with flexible sigmoidoscopy followed by colonoscopy regardless of the sigmoidoscopic result. Our study objective was to determine the prevalence o f patients having isolated adenomatous polyps in a proximal colonic segment in the absence of a distal index neoplasm within reach of the sigmoidoscope. Through the combined use of sigmoidoscopy and colonoscopy, adenomatous polyps were detected in 47 of 114 individuals (41%). A total of 88 adenomas was found. Seventeen patients had isolated neoplasms in proximal colonic segments in the absence of distal adenomas. These patients represented 15% of screened subjects (17 of 114) and 20% of individuals who lacked adenomas on sigmoidoscopy (17 of 84). The majority of proximal neoplasms were small (50 years; (2) absence of symptoms or signs suggestive of colorectal neoplasms including pain, change in bowel habits, or bleeding; (3) absence of risk factors indicating screening colonoscopy, such as ulcerative colitis or evidence of a polyposis or nonpolyposis hereditary colon cancer syndrome; (4) absence of a familial history for colorectal cancer in a first-degree relative. Patients having prior barium enema, sigmoidoscopy, colonoscopy, or polypectomy were excluded. A complete blood count was obtained if a recent result was not available in the medical record, and anemic individuals were not enrolled in the study. One hundred fourteen men (age range 50-79 years, mean 64 years) met entry criteria and consented to participate. Hemoccult testing of stool was performed on 105 subjects, and results were negative in all instances. Nine men did not return their hemoccult cards but none of these individuals had neoplasms on FS or colonoscopy, so the undetermined hemoccult result was presumed to be negative. The mean hemoglobin concentrations for the group was 14.8 g/100 ml. Methods. After preparing the bowel with two enemas, FS was performed by second-year gastroenterology fellows using a 60-cm instrument. The depth of insertion of the sigmoidoscope and absence or presence and location of all polyps was recorded. Colonoscopy, using conscious sedation, was performed after FS at a later date. Colyte was used to clean the colon. Size of polyps was estimated by comparison to the open span of standard biopsy forceps. All diminutive polyps were biopsied and larger lesions were removed with snare cautery. Tissue specimens were fixed in formalin prior to routine histological evaluation. Two pathologists with an interest in colorectal neoplasia reviewed all specimens. Number and histology of adenomas detected within and beyond the extent Digestive Diseases and Sciences, Vol. 36, No. 7 (July 1991)
TABLE 1. PRESENCE OR ABSENCE OF ADENOMAS IN 114 PATIENTS DETECTED BY SIGMOIDOSCOPY OR COLONOSCOPY PERFORMED ABOVE EXTENT OF SIGMOIDOSCOPIC EXAMINATION
Adenomas detected
Subjects N (%) 67 21 9 17
(59) (18) (8.0) (15)
By sigmoidoscopy Absent Present* Presentt Absent
By colonoscopy Absent Absent Present Present
*An invasive c a r c i n o m a in a tubulovillous a d e n o m a was detected in this group. t A n invasive carcinoma in a tubular a d e n o m a and cancer in situ in a tubulovillous a d e n o m a were detected in this group.
of the FS examination were recorded for each patient. For purposes of this study only, cancer, tubular, tubulovillous, and villous adenomas were regarded as positive results. Criteria established by the World Health Organization were used to classify adenomas (14). Lipomas, lymphoid nodules, and hyperplastic polyps were biopsied or removed but were considered negative findings. The study was approved by the Investigational Review Board and Human Subjects Subcommittee. RESULTS The flexible sigmoidoscope was inserted 30-60 cm (mean, 56 cm), and the cecum was reached with the colonoscope in all instances. Bowel preparations were judged to be adequate, and there were no complications associated with either procedure. Through the combined use of sigmoidoscopy and colonoscopy, adenomatous polyps were identified in 47 of 114 individuals (41%) (Table 1). A total of 88 adenomas was detected and sizes ranged from 2 to 25 mm (mean, 6.2 ram). Ten percent were -> 1.0 cm. N u m b e r and location of neoplasms included the following: rectum, 13 (15%); sigmoid colon, 24 (27%); descending colon, 11 (13%); transverse colon, 12 (14%); ascending colon, 16 (18%); cecum, 12 (14%). Seventy-one adenomas were tubular (81%), 16 tubulovillous (18%), and one was villous (1.0%). Endoscopic results for detection of colorectal adenomas are shown in Table 1. Seventeen of 114 subjects (15%) had all neoplasms detected by colonoscopy beyond the reach of the sigmoidoscope. Eighty-four subjects (67 plus 17, Table 1) did not have adenomas on FS, yet 17 of these (20%) had proximal neoplasms diagnosed by colonoscopy. Nine of 30 patients (30%) with adenomas detected by FS had one or more synchronous proximal neoplasm identified by c o l o n o s c o p y (Table 1). These individuals served as a reference for compar-
925
FOUTCH ET AL TABLE 2. COMPARISONOF GROUPS OF PATIENTS WITH PROXIMAL ADENOMAS* ASSOCIATED(GROUP 1) AND NOT ASSOCIATED(GRouP 2) WITH SYNCHRONOUSNEOPLASIAIN DISTAL COLON AND RECTUM
Characteristics
Total number of proximal adenomas Size of proximal adenomas (mm) Range Mean Patients with proximal adenomas ->1.0 cm Patients with histological types of adenomas Tubular Tubulovillous Villous Cancer Patients in whom adenomas were solitary multiple?
Group 1 (N=9 patients)
Group 2 (N=17 patients)
17
29
2-25 6.6
2-25 6.1
1 (11%)
2 (12%)
7 (78%) 2 (22%) l (11%) 0
15 (88%) 4 (24%) 0 0
6 (67%) 3 (33%)
11 (65%) 6 (35%)
*Refers to neoplasms detected beyond the extent of the sigmoidoscopic examination. ?->2 adenomas.
ison to patients (N = 17) who did not have similar lesions on the sigmoidoscopic examination. Comparison of characteristics of groups of patients with adenomas in proximal sites associated and not associated with syDchronous index neoplasms in the distal colon and rectum is shown in Table 2. No important differences between the groups were observed. DISCUSSION We used colonoscopy as a screening tool and prospectively compared results to fiberoptic sigmoidoscopy in average-risk, asymptomatic men. This approach renders a true approximation of prevalence and distribution of neoplasia in this important cohort. Forty-one percent of patients had adenomatous polyps, and this figure is consistent with findings from autopsy studies that show an overall adult prevalence of 21-47% percent (5, 6, 8, 9, 15-17). The majority of adenomas were tubular, and most were small in size. Only 10% of neoplasms were - 1 . 0 cm, and this finding correlates well with the observation of others. Results from analysis of necropsy specimens show that 4-13% of adults with adenomas have neoplasms 1.0 cm or larger in diameter (6, 8). A surprising dearth of information exists regarding the efficacy of FS for screening asymptomatic
926
individuals for neoplasia (18, 19). However, cumulative results from studies suggest that adenomas are detected in 3-20% and cancers in 0.3-4.3% of cases (19-23). Sigmoidoscopy alone identified adenomas and cancer in 26% and 2.6% of our patients, respectively. This high diagnostic yield for neoplasia may have occurred because of selection bias from patients requesting screening and because our study population consisted of elderly men (mean age 64 years). Results from studies show that adenomatous polyps are more common in men than women and increase in frequency with increasing age (5, 6, 24). Arminski and McLean determined the incidence and location of 797 adenomatous polyps in 330 colons reviewed at autopsy (6). Approximately 40% of men between ages 60 and 69 had adenomas, and 41% were located in the sigmoid colon or rectum. Our high detection rate for cancer might be explained by the concomitant high prevalence of adenomas, which are important in the pathogenesis of malignancy. These findings emphasize the need for aggressive screening for neoplasia in this group of individuals. Thirty percent of our patients (nine of 30) with distal adenomas detected by FS had concomitant proximal neoplasms diagnosed by colonoscopy. Previous studies have shown that synchronous polyps occur in 20-50% of cases and it is on the basis of this association that colonoscopy is recommended when an index adenoma is detected by sigmoidoscopy (1, 4, 25). The important finding in our study is that 15% of screened subjects (17 of 114) and 20% of individuals with a normal FS examination (17 of 84) have undiagnosed adenomas in a proximal location. These persons may be at increased risk for cancer, yet they cannot be easily identified by our current screening practices. It is possible that distal hyperplastic polyps (HPP) can serve as markers for proximal adenomas, but this point is controversial (26). Our study design did not consider HPP a significant finding because the Health and Public Policy Committee of the American College of Physicians does not recommend colonoscopy for purposes of detecting proximal adenomas when nonneoplastic polyps are detected by FS (27). However, prevalence for HPP in this study was 31%. If HPP as well as adenomas were considered valid markers for proximal neoplasms, then only 12% of screened subjects and 17% with a totally normal FS examination would have had undiagnosed proximal neoplasms. We have initiated a prospective controlled trial designed to deDigestive Diseases and Sciences, Vol. 36, No. 7 (July 1991)
F L E X I B L E SIGMOIDOSCOPY IN ASYMPTOMATIC MEN
termine the significance of distal HPP, and at present our results are preliminary (28). Previous studies have shown that cancer risk increases with size, number, and increasing villous components of the adenoma (1, 16, 29, 30). The majority of isolated proximal polyps in our series were small tubular adenomas and were of low risk. Yet some adenomas were large, multiple, and contained villous elements (Table 2), and the future rate of growth of the small lesions is unpredictable. There are data to suggest that small polyps cannot be ignored. O'Brien et al detected high-grade dysplasia in nearly 6.0% of adenomatous polyps -
