- Email: [email protected]
Association of rectal prolapse with colorectal cancer
Association of rectal prolapse with colorectal cancer Zaihan Rashid, MBBS, FRCS, and Marc D. Basson, MD, PhD, New Haven, Conn.
Background. Although screening for colorectal cancer facilitates earlier detection and improves survival, cost-effective screening requires identification of patients at high risk far colorectal cancer. Rectal prolapse has not been clearly linked to colorectal carcinoma. Whether patients with rectal prolapse should be screenedfor colorectal cancer is therefore unclear. Methods. We retrospectively identified 70 consecutive patients treated for rectal prolapse at a community hospital during a period of 16 years and monitored for an average of 4.4 +- 2. 7 years and 350 patients of similar age treated for chronic unrelated conditions in the same institution during a similar period. We determined their incidence of colorectal cancer and sought demographic correlates. Results. The prevalence of rectosigmoid carcinoma among patients with prolapse was 5. 7% during the study. Only 1.4 % of the comparative group had colorectal cancer. Thus patients with rectal prolapse exhibited a 4.2fold (95% confidence interval, 1.1 to 16.0) relative risk for colorectal cancer over the comparative group (p < O.02). Conclusions. To the extent to which these patients represent the population of patients with rectal prolapse, routine initial screening of patients with symptomatic rectal prolapse by use of flexible sigmoidoscopy may be appropriate. (SURGERY1996;119:51-5.) From the Departments of Surgery, Yale University School of Medicine and Bridgeport Hospital, New Haven and Bridgeport, Conn.
RECTALPROLAPSEIS COMMONa m o n g older patients (60 to 90 years o f age) a n d may be p r o m o t e d by constipation and o t h e r causes of increased intraabdominal pressure that lead to laxity o f the pelvic tissues.1 Because new onset constipation in the older population itself may p r o m p t investigation a n d screening for colorectal cancer, one might question whether rectal prolapse may be a presenting sign for colon cancer in a significant number of patients with symptomatic rectal prolapse. Review of the literature suggests that this is an unanswered question. Although some texts or reviews suggest that patients with rectal prolapse should be screened for colorectal cancer, 24 others I suggest screening only in selected patients with unspecified characteristics, whereas similar suggestions do not a p p e a r at all in o t h e r major reviews o f the subject. 5-7 T h e English-language literature lacks primary data to answer the question o f whether rectal prolapse may be a sign for colorectal cancer. The question is n o t only of epidemiologic interest b u t also o f potential clinical significance because of r e c e n t demonstrations that screening for colorectal cancer
Accepted for publication May 1, 1995. Reprint requests: Marc D. Basson, MD, PhD, Department of Surgery, Yale UniversitySchool of Medicine, 333 Cedar St., Box 208062, New Haven, CT 06520-8062. Copyright 9 1996 by Mosby-YearBook, Inc. 0039-6060/96/$5.00 + 0 11/56/67809
may result in increased detection of the disease and the detection o f earlier lesions in high risk patients, s-l~ Screening is costly, however, a n d places a considerable b u r d e n on the providers o f this service a n d the health care system. It is therefore essential that those at high risk o f colorectal cancer be identified. We hypothesized that an increased prevalence of colorectal cancer would occur a m o n g patients with symptomatic rectal prolapse. We c o n d u c t e d a retrospective study of all patients admitted with diagnosis of rectal prolapse to one community primary care hospital during a p e r i o d of 16 years, deliberately choosing to study patients in a primary care institution to avoid the potential referral bias implicit in tertiary care referral practices. O u r patient selection p r o c e d u r e encore-. passed only those patients who were symptomatic e n o u g h to warrant hospital admission a n d surgical repair, so that data from this study may n o t necessarily be extrapolated to patients with asymptomatic or only mildly symptomatic disease. We c o m p a r e d the incidence of colorectat carcinoma in these patients with patients o f similar age and length o f follow-up being managed for chronic diseases u n r e l a t e d to colorectal carcin o m a in the same institution.
MATERIAL A N D METHODS We retrospectively identified 70 patients with an admitring diagnosis of rectal prolapse at a community hospital d u r i n g a p e r i o d o f 16 years (1978 to 1994). SURGERY
51
52
Rashid and Basson
Surgery January 1996
T a b l e I. Patient demographics
Rectal prolapse Comparative group group No. of patients Age (yr) Gender (% female) Year of entry into study Years of followup Incidence of colorectal cancer (%)
70 77.2 _+ 1.2 74 1984 • 0.5
350 75.4 • 0.8 63 1984 • 0.2
4.4 2 2.7 5.7
4.9 • 0.2 1.4
Chart review yielded information on principal symptoms at admission, reducibility, size of prolapse, and the nature of any operative intervention. We also recorded information on any colorectal cancer screening procedures performed at admission for rectal prolapse. Follow-up for all patients was obtained either from treating physicians or from current hospital records, and the identification of colorectal neoplasms during this period was recorded together with the site and stage of disease and survival data. For purposes of comparison we chose an equivalnt group of 350 patients closely matched for age, year of entry into the study population, and length of available follow-up data. These patients were selected randomly by the hospital medical records computer in a two-stage r a n d o m process. We first generated a list of patients who had sought medical attention at the same institution during the same time period for any of four chronic conditions that merited long-term follow-up. These other medical conditions used to generate the initial list of patients were diabetes mellitus, hypertension, ischemic heart disease, and valvular heart disease. We then selected randomly from that list five patients who were similar to each one of the 70 patients with prolapse in age, year of entry into the hospital medical record system, and their length of available follow-up data. Past history of colorectal cancer was not a contraindication to inclusion in either the rectal prolapse or comparative patient group. The identification of colorectal neoplasms in these patients during equivalent follow-up periods was also recorded. Careful assessment was made of the time of entry of comparative group patients to coincide with the entry of rectal prolapse group patients into the study. Statistical analysis was performed to compare occurrence ofcolorectal cancer a m o n g rectal prolapse group and the comparative group by means of a t test for proportions. Comparison was made between the occurrence of colorectal cancer in patients with rectal prolapse and the patients' age, gender, presenting symptoms, and reducibility and size of prolapse by means of a stepwise regression analysis. The validity of the heter-
ogeneous comparative group was further tested by use of a t test for proportions comparing the incidence of colorectal carcinoma a m o n g patients with each individual diagnosis.
RESULTS The patients with rectal prolapse and patients in the comparative group were similar in age, date of entry into the study, and length of available follow-up data (Table I). Slightly more men were noted a m o n g the comparative group (prolapse, 24%; comparative, 33%;p < 0.05). Four colorectal cancers (5.7%) were identified a m o n g the 70 patients with rectal prolapse. Three were Duke's stage A and one was stage C. Three were rectal and one was sigmoid. All colorectal cancers in the patients with rectal prolapse were diagnosed at admission for operative treatment of rectal prolapse. Fifty-seven (81%) of 70 patients with rectal prolapse underwent some type of colorectal screening procedure at the index admission to the hospital. These included rigid proctoscopy (14%), flexible sigmoidoscopy (26%), colonoscopy (20%), and barium enema (21%). No screening procedure had been performed in 20% of these patients. There was no record of any previous screening procedures having been performed in these patients. The predominant symptoms of the rectal prolapse in the patients studied here were rectal mass (52%), rectal bleeding (29%), or a combination of the two (18%). No patients in this group were identified as asymptomatic. A m o n g the prolapse group 33 (47%) had a full thickness rectal prolapse, and 37 (53%) had a partial or mucosal prolapse. The full thickness rectal prolapse was on the average 8.8 -+ 7.1 cm in size, and the partial thickness prolapse was on average 3.5 -+ 1.2 cm. The patients with full thickness prolapse underwent the following surgical procedures for correction of prolapse: Ripstein's procedure 11 (33%), Delorme's procedure 10 (30%), mesh rectopexy 6 (18%), and low anterior resection 6 (18%). The patients with partial thickness underwent the following procedures: Delorme's procedure 20 (54%), Thiersch's wire 11 (30%), anterior resection 1 (3%), abdominoperineal resection 1 (3%), amputation of rectal sleeve 1 (3%), and no procedures 3 (8%). Thus the patients with partial thickness prolapse studied here were not simply patients with hemorrhoids. One of the colorectal cancers identified here was in a patient with a full thickness prolapse, whereas three cancers were identified in patients with mucosal prolapse only, and the size of the prolapse did not correlate with the presence or otherwise of a colorectal cancer. Furthermore, no particular symptoms could be specifically associated with cancer in the rectal prolapse group, although a trend for higher risk for colorectal cancer was noted a m o n g women with prolapse as compared with men with rectal prolapse (Table II). Simi-
,Surgery Volume 119, Number 1 larly, patients in the comparative group with colorectal carcinoma had symptoms that were n o t distinguishable from those of patients with rectosigmoid c a r c i n o m a a n d rectal prolapse except for the prolapse itself. No statistical difference was n o t e d in the incidence o f colorectal cancer a m o n g the f o u r different groups o f patients according to a t test o f proportions. In contrast to the observed 5.7% prevalence o f colorectal c a r c i n o m a a m o n g the patients with rectal prolapse, four colonic (right colon 2, d e s c e n d i n g colon 1, and sigmoid colon 1) a n d one rectal cancers were diagnosed a m o n g the 350 patients in the comparative g r o u p (1.4%). Thus a comparison of the prevalence o f colorectal neoplasms a m o n g the patients with rectal prolapse and the comparative group yielded a relative risk of 4.18 (95% confidence interval, 1.1 to 16.0) A m o n g the 350 patients randomly selected for the comparative group, there were 154 patients with hypertension (44%), 126 patients with ischemic h e a r t disease (36%), 7 patients with valvular heart disease (2%), a n d 63 patients with diabetes mellitus (18%). T h e relative prevalence of cancer a m o n g the patients with each of these diseases was 1.25%, 1.57%, 0%, and 1.6%, respectively. No statistically significant difference was n o t e d in the prevalence of colorectal cancer a m o n g these four groups o f patients. Two o f the colorectal cancers a m o n g the comparative group o f patients were Duke's stage A, one was stage B, and two were stage D. O n e carcinoma in the comparative group was in the rectum, two were in the left colon, and two were in the right colon. T h e a p p a r e n t trends toward m o r e advanced disease, decreased survival, a n d less rectal disease a m o n g the comparative g r o u p than a m o n g patients with rectal prolapse could not be assessed statistically because of the relatively small number of tumors (Table II). Patients with four different diseases had been studied in the comparative g r o u p to minimize ascertainment bias related to any specific disease endty. I n d e e d no statistical difference in the incidence of colorectal cancer was n o t e d a m o n g the four different groups o f patients according to a t test o f proportions. This study included patients with both partial a n d full thickness prolapse. T h a t the patients with partial thickness prolapse did n o t simply have h e m o r r h o i d s is supp o r t e d by the size o f these lesions and the fact that all patients u n d e r w e n t p r o c e d u r e s to correct rectal prolapse. Reanalysis by subgroups (full versus partial thickness) suggests that the incidence of colorectal cancer a m o n g the patients with full thickness prolapse was 3% (1 of 33), whereas the incidence o f colorectal cancer a m o n g the patients with partial thickness prolapse was 6.4% ( n = 3 of 47). T h e former p r o p o r t i o n did n o t achieve statistical significance as c o m p a r e d with the 1.4% incidence in the comparative g r o u p because of the
Rashid and Basson
53
T a b l e II. Characteristics of patients with tumors
No. of patients Gender (% female) Site
% Resectable Mean smwival (yr)
Rectal prolapse group
Comparative group
4 75
5 40
Rectal 3, sigmoid colon 1
Rectal 1, sigmoid colon 2, descending 1, ascending 1 100 3.42 -+ 1.27
100 4.93 _+ 1.58
relatively small numbers. The latter p r o p o r t i o n remains significant (p < 0.02). DISCUSSION Rectal prolapse is most c o m m o n a m o n g elderly a n d constipated persons. 1'11 Complete prolapse is four times m o r e c o m m o n a m o n g women, primarily a m o n g the nulliparous. 1' 11 Complications of rectal prolapse include inflammation a n d inflammatory polyps, bleeding, excoriation, and strangulation. Rectal tumors may prolapse,l, 11 b u t an i n d e p e n d e n t link between rectal prolapse and colorectal cancer has not b e e n established. O f consecutive patients presenting to a single community hospital with symptomatic rectal prolapse, 5.7% h a d colorectal cancer (95% confidence interval, 0.1 _+ 11.3). A m o n g a comparative g r o u p the prevalence was 1.4% (95% confidence interval, 0.2 -+ 2.8), consistent with historical data suggesting the expected prevalence o f colorectal cancer in this age c o h o r t to be 0.6% to 1.7%.1214 Symptomatic rectal prolapse may thus be a sign o f colorectal cancer in a significant fraction of patients. Prolapse in these patients may reflect a response of the lax elderly p e r i n e u m to increased intraabd o m i n a l pressure e n g e n d e r e d by constipation from an occult colorectal malignancy. Several investigators have similarly associated inguinal h e r n i a with asymptomatic colorectal cancer, postulating that subclinical constipation may u n m a s k a patent processus vaginalis. 15-2~Others have questioned this theory. 21 These results suggest that the prevalence o f colorectal cancer may be increased a m o n g patients with symptomatic rectal prolapse. Not all patients with rectal prolapse are hospitalized. N o n e of the patients studied here was asymptomatic. Thus the present study applies to patients with rectal prolapse and symptoms sufficient to require hospitalization or surgical correction but not necessarily to less symptomatic patients. In addition, this study was not designed to distinguish patients with full or partial thickness prolapse. Subgroup analysis sug-
54
Rashid and Basson
gested higher cancer prevalence in patients with partial thickness prolapse. The increased prevalence in patients with full thickness prolapse did n o t differ statistically from comparative patients because of small n u m bers. However, retrospective data reanalysis of small subgroups is problematic a n d requires prospective clarification. Two problems can be adduced with regard to the comparative group. First, more m e n were present a m o n g comparative group than a m o n g rectal prolapse group. Because m e n may be at higher risk for colorectal cancer, 12 this gender inequality should have biased the study against d o c u m e n t i n g increased risk a m o n g patients with prolapse a n d does not invalidate the present findings. Second, colorectal neoplasms could have b e e n missed in comparative patients n o t screened for colorectal cancer, either because colorectal neoplasms were present but did n o t appear d u r i n g the follow-up period or because death from colorectal neoplasm might have excluded patients from the comparative group in which a m e a n 4.9-year follow-up was required. These possibilities are implicit in the retrospective design of the study a n d the tack of a wellscreened asymptomatic population in the community. However, the prevalence of colorectal cancer in asymptomatic patients by use of screening endoscopy has been reported to be 0.6% to 1.7%. 12-14It seems unlikely that a routine screen of a hospital-based asymptomatic population would yield a prevalence as high as the 5.7% f o u n d in the patients with rectal prolapse. Furthermore, both comparative patients and those with rectal prolapse were monitored for a m e a n 4.4 years. Many "missed" neoplasms should have presented d u r i n g this period. An American's cumulative lifetime risk of dying of colon cancer is 3%. 22 Colonoscopic screening detects asymptomatic polyps a n d colon cancer. 2~-25 Repeated screening colonoscopy reduces colon cancer mortality by one haft. 9' 10, 23 If patients with symptomatic rectal prolapse are at even higher risk, screening would seem reasonable. The costs of screening for colorectal cancer can be estimated assuming a cost of $191 per flexible sigmoidoscopy a n d $734 per colonoscopy. 26 O u r data suggest that screening colonoscopy for patients with symptomatic prolapse would cost $12,845 per cancer diagnosed. However, all cancers in patients with rectal prolapse were rectosigmoid. Flexible sigmoidoscopy could therefore reduce the cost to $3343 per cancer diagnosed without missing many lesions. By comparison, calculation from published data 1214 suggests that detecting a cancer in asymptomatic patients costs $31,578 to $11,365 by use of flexible sigmoidoscopy and $121,354 to $43,673 by use of colonoscopy. Diagnosing a cancer in patients with hernias 16' 18-20 by means of flexible sig-
Surgery January 1996 moidoscopy costs $4028 to $9550. Screening patients with symptomatic rectal prolapse for rectosigmoid cancer thus appears more cost-effective than screening asymptomatic patients or those with hernias. This study suggests that patients with symptomatic rectal prolapse exhibit a substantially increased prevalence of rectosigmoid cancer. Straining from constipation may explain this synchronous presentation of colorectal cancer a n d rectal prolapse. Complex issues of cost-effectiveness require prospective study to determine the impact of screening o n life expectancy. Differences in the prevalence of colorectal cancer a m o n g patients with symptomatic partial or full thickness prolapse and the prevalence of colorectal cancer a m o n g asymptomatic or mildly symptomatic patients also await further prospective study. However, these data suggest that routine screening of symptomatic patients with rectal prolapse by use of flexible sigmoidoscopy is warranted. The assistance of Emanuel Lerner, MA, with statistical analysis is greatly appreciated. REFERENCES
1. KaranjiaND, Rutter KRP. Solitaryrectal ulcer and rectal prolapse. Current Practice in Surgery 1994;6:59-64. 2. Gordon PH. Rectal procidentia. In: Gordon PH, NivatvongsS. Principlesand practiceof surgeryfor the colon,rectum and anus. 2nd ed. St Louis: QualityMedical PublishingInc, 1992:449-81. 3. Kodner IJ, Fry RD, FleshmanJW, BirnbaumEH. Colon, rectum and anus.In: Schwartz,S. Principlesof surgery.6th ed. NewYork: McGraw Hill, Inc, 1994:1191-306. 4. Britten-JonesR. Complete rectal prolapse: anterior wrap (Ripstein operation). In: DudleyH, Carter DC, RussellRCG.Atlasof generalsurgery.2nd ed. London:Butterworths,Inc, 1986:575-80. 5. Hicks TC, Opelka FG. Rectal and perianal complaints.In: Polk HC Jr, Gardner B, Stone HH. Basic surgery. 4th ed. St. Louis: Quality Medical PublishingInc, 1993:545-60. 6. Fry RD, Kodner IJ. Anorectaldiseases. In: LevineVA, Copeland EM III,HowardRJ, SugermanHJ, WarshawAL. Current practice of surgery. New York: ChurchillLivingstone,1993:1-23. 7. Dent TL, KukoraJS, NejmanJH. The colon, rectum and anus. In: HardyJD. Hardy'stextbookof surgery. 2nd ed. Philadelphia: JB LippincottCompany, 1988:582-636. 8. FleischerDE, GoldbergSB, BrowningTH, et al. Detection and surveillanceof colorectal cancer.JAMA1989;261:580-6. 9. GilbertsonVA, NelmsJM. The preventionof invasivecancer of the rectum. Cancer 1978;41:1137-9. 10. Friedman GD, Collen MF, Fireman BH. Multiphasic health checkup evaluation:a 16yearfollow-up.JChron Dis 1986;39:45363. 11. KeighlyMRB,Matheson DM, DuncanMM. Resultsof treatment for rectal prolapse and fecal incontinence. Dis Colon Rectum 1981;24:449-53. 12. DisarioJA,Foutch P. Prevalenceand malignantpotential of colorectal polypsin asymptomatic,average risk men. AmJ Gastroenterol 1991;86:941-5. 13. GuptaTP,JaszewskiR, Luk GD. Efficacyof screeningflexiblesigmoidoscopy for colorectal neoplasia in asymptomaticsubjects. AmJ Med 1989;86:547-50. 14. Rex KD, Lehman GA, Ulbright TM, et al. Colonicneoplasiain asymptomatic personswith negativefecal occult blood tests: in-
Surgery Volume 119, Number 1 fluence of age, gender, and family history. Am J Gastroenterol 1993;6:825-31. 15. Day T, Ferrara MD. Preoperative barium contrast enema in patients with inguinal hernia. South MedJ 1986;79:1339-41. 16. Rubin BG, Ballantyne GH, Zdon MJ, Zucker KA, Modlin IM. The role of flexible sigmoidoscopy in the preoperative screening of patients with inguinal hernia. Arch Surg 1987;122:296-9. 17. Maccini MM, Miller RM. The yield of barium enema in patients undergoing inguinal hernia repair or abdominal hysterectomy. Surg Gynecol Obstet 1991;172:391-3. 18. Pratt SM, Weaver FA, PottsJR. Preoperative evaluation of patients with inguinal hernia for colorectal disease. Surg Gynecol Obstet 1987;165:5~6. 19. Wheeler WE, Wilson SL, Kurucz J, Flanigan J, Ratliff D, ScottConner C. Flexible sigmoidoscopy screening for asymptomatic colorectal disease in patients with and without inguinal hernia. South MedJ 1991;84:876~.
Rashid and Basson
55
20. LovettJ, Kirgan D, Mcgregor B. Inguinal herniajustifies sigmoidoscopy. Am J Surg 1989;158:615-7. 21. Brendel TH, Kirsh IE. Lack of association between inguinal hernia and carcinoma of the colon. N EnglJ Med 1971;284:369-70. 22. Seldman H, Mushinski MH, Gelb SK, et al. Probabilities of eventually developing or dying of cancer. United States, 1985. CA 1985;35:36-56. 23. Stryker SJ, Wolff BG, Culp CE, Libbe SD, Ilstrup DM, MacCarty RL. Natural history of untreated colonic polyps. Gastroenterology 1987;93:1009-13. 24. Muto T, Bussey HJR, Morson BC. Cancer 1975;36:2251-70. 25. Connecticut HealthPlan, Inc. Fee schedule information (1994). Meriden, Connecticut: Connecticut HealthPlan Inc, 1995. 26. Wheeler WE, Scott-Conner CEH, Stone RA. Flexible sigmoidoscopy as a screening procedure for asymptomatic colorectal carcinmna in patients with inguinal hernia. South MedJ 1985; 78: 1417-20.
SURGERYiS abstracted a n d / o r i n d e x e d in Index Medicus, Science Citation Index, Current Contents/ Clinical Medicine, Current Contents~Life Sciences, Excerpta Medica, a n d MEDI,INE. This JOURNALhas b e e n registered with Copyright Clearance Center, Inc., 222 Rosewood Dr., Danvers, MA 01923. C o n s e n t is given for the copying o f articles for personal or internal use o f specific clients. This c o n s e n t is given o n the condition that the copier pay directly to the C e n t e r the per-copy fee stated o n the first page o f each article for copying b e y o n d that p e r m i t t e d by U.S. Copyright Law. This c o n s e n t does n o t e x t e n d to o t h e r kinds o f copying, such as for general distribution, resale, advertising a n d p r o m o t i o n a l purposes, or for creating new collective works. All inquiries regarding c o p y r i g h t e d material f r o m this publication o t h e r than those that can be h a n d l e d t h r o u g h Copyright Clearance C e n t e r should be d i r e c t e d to Associate J o u r n a l Publisher, Permissions, Mosby-Year Book, Inc., 11830 Westline Industrial Dr., St. Louis, MO 63146; p h o n e (314)579-2808; fax (314)432-1380.
Background. Although screening for colorectal cancer facilitates earlier detection and improves survival, cost-effective screening requires identification of patients at high risk far colorectal cancer. Rectal prolapse has not been clearly linked to colorectal carcinoma. Whether patients with rectal prolapse should be screenedfor colorectal cancer is therefore unclear. Methods. We retrospectively identified 70 consecutive patients treated for rectal prolapse at a community hospital during a period of 16 years and monitored for an average of 4.4 +- 2. 7 years and 350 patients of similar age treated for chronic unrelated conditions in the same institution during a similar period. We determined their incidence of colorectal cancer and sought demographic correlates. Results. The prevalence of rectosigmoid carcinoma among patients with prolapse was 5. 7% during the study. Only 1.4 % of the comparative group had colorectal cancer. Thus patients with rectal prolapse exhibited a 4.2fold (95% confidence interval, 1.1 to 16.0) relative risk for colorectal cancer over the comparative group (p < O.02). Conclusions. To the extent to which these patients represent the population of patients with rectal prolapse, routine initial screening of patients with symptomatic rectal prolapse by use of flexible sigmoidoscopy may be appropriate. (SURGERY1996;119:51-5.) From the Departments of Surgery, Yale University School of Medicine and Bridgeport Hospital, New Haven and Bridgeport, Conn.
RECTALPROLAPSEIS COMMONa m o n g older patients (60 to 90 years o f age) a n d may be p r o m o t e d by constipation and o t h e r causes of increased intraabdominal pressure that lead to laxity o f the pelvic tissues.1 Because new onset constipation in the older population itself may p r o m p t investigation a n d screening for colorectal cancer, one might question whether rectal prolapse may be a presenting sign for colon cancer in a significant number of patients with symptomatic rectal prolapse. Review of the literature suggests that this is an unanswered question. Although some texts or reviews suggest that patients with rectal prolapse should be screened for colorectal cancer, 24 others I suggest screening only in selected patients with unspecified characteristics, whereas similar suggestions do not a p p e a r at all in o t h e r major reviews o f the subject. 5-7 T h e English-language literature lacks primary data to answer the question o f whether rectal prolapse may be a sign for colorectal cancer. The question is n o t only of epidemiologic interest b u t also o f potential clinical significance because of r e c e n t demonstrations that screening for colorectal cancer
Accepted for publication May 1, 1995. Reprint requests: Marc D. Basson, MD, PhD, Department of Surgery, Yale UniversitySchool of Medicine, 333 Cedar St., Box 208062, New Haven, CT 06520-8062. Copyright 9 1996 by Mosby-YearBook, Inc. 0039-6060/96/$5.00 + 0 11/56/67809
may result in increased detection of the disease and the detection o f earlier lesions in high risk patients, s-l~ Screening is costly, however, a n d places a considerable b u r d e n on the providers o f this service a n d the health care system. It is therefore essential that those at high risk o f colorectal cancer be identified. We hypothesized that an increased prevalence of colorectal cancer would occur a m o n g patients with symptomatic rectal prolapse. We c o n d u c t e d a retrospective study of all patients admitted with diagnosis of rectal prolapse to one community primary care hospital during a p e r i o d of 16 years, deliberately choosing to study patients in a primary care institution to avoid the potential referral bias implicit in tertiary care referral practices. O u r patient selection p r o c e d u r e encore-. passed only those patients who were symptomatic e n o u g h to warrant hospital admission a n d surgical repair, so that data from this study may n o t necessarily be extrapolated to patients with asymptomatic or only mildly symptomatic disease. We c o m p a r e d the incidence of colorectat carcinoma in these patients with patients o f similar age and length o f follow-up being managed for chronic diseases u n r e l a t e d to colorectal carcin o m a in the same institution.
MATERIAL A N D METHODS We retrospectively identified 70 patients with an admitring diagnosis of rectal prolapse at a community hospital d u r i n g a p e r i o d o f 16 years (1978 to 1994). SURGERY
51
52
Rashid and Basson
Surgery January 1996
T a b l e I. Patient demographics
Rectal prolapse Comparative group group No. of patients Age (yr) Gender (% female) Year of entry into study Years of followup Incidence of colorectal cancer (%)
70 77.2 _+ 1.2 74 1984 • 0.5
350 75.4 • 0.8 63 1984 • 0.2
4.4 2 2.7 5.7
4.9 • 0.2 1.4
Chart review yielded information on principal symptoms at admission, reducibility, size of prolapse, and the nature of any operative intervention. We also recorded information on any colorectal cancer screening procedures performed at admission for rectal prolapse. Follow-up for all patients was obtained either from treating physicians or from current hospital records, and the identification of colorectal neoplasms during this period was recorded together with the site and stage of disease and survival data. For purposes of comparison we chose an equivalnt group of 350 patients closely matched for age, year of entry into the study population, and length of available follow-up data. These patients were selected randomly by the hospital medical records computer in a two-stage r a n d o m process. We first generated a list of patients who had sought medical attention at the same institution during the same time period for any of four chronic conditions that merited long-term follow-up. These other medical conditions used to generate the initial list of patients were diabetes mellitus, hypertension, ischemic heart disease, and valvular heart disease. We then selected randomly from that list five patients who were similar to each one of the 70 patients with prolapse in age, year of entry into the hospital medical record system, and their length of available follow-up data. Past history of colorectal cancer was not a contraindication to inclusion in either the rectal prolapse or comparative patient group. The identification of colorectal neoplasms in these patients during equivalent follow-up periods was also recorded. Careful assessment was made of the time of entry of comparative group patients to coincide with the entry of rectal prolapse group patients into the study. Statistical analysis was performed to compare occurrence ofcolorectal cancer a m o n g rectal prolapse group and the comparative group by means of a t test for proportions. Comparison was made between the occurrence of colorectal cancer in patients with rectal prolapse and the patients' age, gender, presenting symptoms, and reducibility and size of prolapse by means of a stepwise regression analysis. The validity of the heter-
ogeneous comparative group was further tested by use of a t test for proportions comparing the incidence of colorectal carcinoma a m o n g patients with each individual diagnosis.
RESULTS The patients with rectal prolapse and patients in the comparative group were similar in age, date of entry into the study, and length of available follow-up data (Table I). Slightly more men were noted a m o n g the comparative group (prolapse, 24%; comparative, 33%;p < 0.05). Four colorectal cancers (5.7%) were identified a m o n g the 70 patients with rectal prolapse. Three were Duke's stage A and one was stage C. Three were rectal and one was sigmoid. All colorectal cancers in the patients with rectal prolapse were diagnosed at admission for operative treatment of rectal prolapse. Fifty-seven (81%) of 70 patients with rectal prolapse underwent some type of colorectal screening procedure at the index admission to the hospital. These included rigid proctoscopy (14%), flexible sigmoidoscopy (26%), colonoscopy (20%), and barium enema (21%). No screening procedure had been performed in 20% of these patients. There was no record of any previous screening procedures having been performed in these patients. The predominant symptoms of the rectal prolapse in the patients studied here were rectal mass (52%), rectal bleeding (29%), or a combination of the two (18%). No patients in this group were identified as asymptomatic. A m o n g the prolapse group 33 (47%) had a full thickness rectal prolapse, and 37 (53%) had a partial or mucosal prolapse. The full thickness rectal prolapse was on the average 8.8 -+ 7.1 cm in size, and the partial thickness prolapse was on average 3.5 -+ 1.2 cm. The patients with full thickness prolapse underwent the following surgical procedures for correction of prolapse: Ripstein's procedure 11 (33%), Delorme's procedure 10 (30%), mesh rectopexy 6 (18%), and low anterior resection 6 (18%). The patients with partial thickness underwent the following procedures: Delorme's procedure 20 (54%), Thiersch's wire 11 (30%), anterior resection 1 (3%), abdominoperineal resection 1 (3%), amputation of rectal sleeve 1 (3%), and no procedures 3 (8%). Thus the patients with partial thickness prolapse studied here were not simply patients with hemorrhoids. One of the colorectal cancers identified here was in a patient with a full thickness prolapse, whereas three cancers were identified in patients with mucosal prolapse only, and the size of the prolapse did not correlate with the presence or otherwise of a colorectal cancer. Furthermore, no particular symptoms could be specifically associated with cancer in the rectal prolapse group, although a trend for higher risk for colorectal cancer was noted a m o n g women with prolapse as compared with men with rectal prolapse (Table II). Simi-
,Surgery Volume 119, Number 1 larly, patients in the comparative group with colorectal carcinoma had symptoms that were n o t distinguishable from those of patients with rectosigmoid c a r c i n o m a a n d rectal prolapse except for the prolapse itself. No statistical difference was n o t e d in the incidence o f colorectal cancer a m o n g the f o u r different groups o f patients according to a t test o f proportions. In contrast to the observed 5.7% prevalence o f colorectal c a r c i n o m a a m o n g the patients with rectal prolapse, four colonic (right colon 2, d e s c e n d i n g colon 1, and sigmoid colon 1) a n d one rectal cancers were diagnosed a m o n g the 350 patients in the comparative g r o u p (1.4%). Thus a comparison of the prevalence o f colorectal neoplasms a m o n g the patients with rectal prolapse and the comparative group yielded a relative risk of 4.18 (95% confidence interval, 1.1 to 16.0) A m o n g the 350 patients randomly selected for the comparative group, there were 154 patients with hypertension (44%), 126 patients with ischemic h e a r t disease (36%), 7 patients with valvular heart disease (2%), a n d 63 patients with diabetes mellitus (18%). T h e relative prevalence of cancer a m o n g the patients with each of these diseases was 1.25%, 1.57%, 0%, and 1.6%, respectively. No statistically significant difference was n o t e d in the prevalence of colorectal cancer a m o n g these four groups o f patients. Two o f the colorectal cancers a m o n g the comparative group o f patients were Duke's stage A, one was stage B, and two were stage D. O n e carcinoma in the comparative group was in the rectum, two were in the left colon, and two were in the right colon. T h e a p p a r e n t trends toward m o r e advanced disease, decreased survival, a n d less rectal disease a m o n g the comparative g r o u p than a m o n g patients with rectal prolapse could not be assessed statistically because of the relatively small number of tumors (Table II). Patients with four different diseases had been studied in the comparative g r o u p to minimize ascertainment bias related to any specific disease endty. I n d e e d no statistical difference in the incidence of colorectal cancer was n o t e d a m o n g the four different groups o f patients according to a t test o f proportions. This study included patients with both partial a n d full thickness prolapse. T h a t the patients with partial thickness prolapse did n o t simply have h e m o r r h o i d s is supp o r t e d by the size o f these lesions and the fact that all patients u n d e r w e n t p r o c e d u r e s to correct rectal prolapse. Reanalysis by subgroups (full versus partial thickness) suggests that the incidence of colorectal cancer a m o n g the patients with full thickness prolapse was 3% (1 of 33), whereas the incidence o f colorectal cancer a m o n g the patients with partial thickness prolapse was 6.4% ( n = 3 of 47). T h e former p r o p o r t i o n did n o t achieve statistical significance as c o m p a r e d with the 1.4% incidence in the comparative g r o u p because of the
Rashid and Basson
53
T a b l e II. Characteristics of patients with tumors
No. of patients Gender (% female) Site
% Resectable Mean smwival (yr)
Rectal prolapse group
Comparative group
4 75
5 40
Rectal 3, sigmoid colon 1
Rectal 1, sigmoid colon 2, descending 1, ascending 1 100 3.42 -+ 1.27
100 4.93 _+ 1.58
relatively small numbers. The latter p r o p o r t i o n remains significant (p < 0.02). DISCUSSION Rectal prolapse is most c o m m o n a m o n g elderly a n d constipated persons. 1'11 Complete prolapse is four times m o r e c o m m o n a m o n g women, primarily a m o n g the nulliparous. 1' 11 Complications of rectal prolapse include inflammation a n d inflammatory polyps, bleeding, excoriation, and strangulation. Rectal tumors may prolapse,l, 11 b u t an i n d e p e n d e n t link between rectal prolapse and colorectal cancer has not b e e n established. O f consecutive patients presenting to a single community hospital with symptomatic rectal prolapse, 5.7% h a d colorectal cancer (95% confidence interval, 0.1 _+ 11.3). A m o n g a comparative g r o u p the prevalence was 1.4% (95% confidence interval, 0.2 -+ 2.8), consistent with historical data suggesting the expected prevalence o f colorectal cancer in this age c o h o r t to be 0.6% to 1.7%.1214 Symptomatic rectal prolapse may thus be a sign o f colorectal cancer in a significant fraction of patients. Prolapse in these patients may reflect a response of the lax elderly p e r i n e u m to increased intraabd o m i n a l pressure e n g e n d e r e d by constipation from an occult colorectal malignancy. Several investigators have similarly associated inguinal h e r n i a with asymptomatic colorectal cancer, postulating that subclinical constipation may u n m a s k a patent processus vaginalis. 15-2~Others have questioned this theory. 21 These results suggest that the prevalence o f colorectal cancer may be increased a m o n g patients with symptomatic rectal prolapse. Not all patients with rectal prolapse are hospitalized. N o n e of the patients studied here was asymptomatic. Thus the present study applies to patients with rectal prolapse and symptoms sufficient to require hospitalization or surgical correction but not necessarily to less symptomatic patients. In addition, this study was not designed to distinguish patients with full or partial thickness prolapse. Subgroup analysis sug-
54
Rashid and Basson
gested higher cancer prevalence in patients with partial thickness prolapse. The increased prevalence in patients with full thickness prolapse did n o t differ statistically from comparative patients because of small n u m bers. However, retrospective data reanalysis of small subgroups is problematic a n d requires prospective clarification. Two problems can be adduced with regard to the comparative group. First, more m e n were present a m o n g comparative group than a m o n g rectal prolapse group. Because m e n may be at higher risk for colorectal cancer, 12 this gender inequality should have biased the study against d o c u m e n t i n g increased risk a m o n g patients with prolapse a n d does not invalidate the present findings. Second, colorectal neoplasms could have b e e n missed in comparative patients n o t screened for colorectal cancer, either because colorectal neoplasms were present but did n o t appear d u r i n g the follow-up period or because death from colorectal neoplasm might have excluded patients from the comparative group in which a m e a n 4.9-year follow-up was required. These possibilities are implicit in the retrospective design of the study a n d the tack of a wellscreened asymptomatic population in the community. However, the prevalence of colorectal cancer in asymptomatic patients by use of screening endoscopy has been reported to be 0.6% to 1.7%. 12-14It seems unlikely that a routine screen of a hospital-based asymptomatic population would yield a prevalence as high as the 5.7% f o u n d in the patients with rectal prolapse. Furthermore, both comparative patients and those with rectal prolapse were monitored for a m e a n 4.4 years. Many "missed" neoplasms should have presented d u r i n g this period. An American's cumulative lifetime risk of dying of colon cancer is 3%. 22 Colonoscopic screening detects asymptomatic polyps a n d colon cancer. 2~-25 Repeated screening colonoscopy reduces colon cancer mortality by one haft. 9' 10, 23 If patients with symptomatic rectal prolapse are at even higher risk, screening would seem reasonable. The costs of screening for colorectal cancer can be estimated assuming a cost of $191 per flexible sigmoidoscopy a n d $734 per colonoscopy. 26 O u r data suggest that screening colonoscopy for patients with symptomatic prolapse would cost $12,845 per cancer diagnosed. However, all cancers in patients with rectal prolapse were rectosigmoid. Flexible sigmoidoscopy could therefore reduce the cost to $3343 per cancer diagnosed without missing many lesions. By comparison, calculation from published data 1214 suggests that detecting a cancer in asymptomatic patients costs $31,578 to $11,365 by use of flexible sigmoidoscopy and $121,354 to $43,673 by use of colonoscopy. Diagnosing a cancer in patients with hernias 16' 18-20 by means of flexible sig-
Surgery January 1996 moidoscopy costs $4028 to $9550. Screening patients with symptomatic rectal prolapse for rectosigmoid cancer thus appears more cost-effective than screening asymptomatic patients or those with hernias. This study suggests that patients with symptomatic rectal prolapse exhibit a substantially increased prevalence of rectosigmoid cancer. Straining from constipation may explain this synchronous presentation of colorectal cancer a n d rectal prolapse. Complex issues of cost-effectiveness require prospective study to determine the impact of screening o n life expectancy. Differences in the prevalence of colorectal cancer a m o n g patients with symptomatic partial or full thickness prolapse and the prevalence of colorectal cancer a m o n g asymptomatic or mildly symptomatic patients also await further prospective study. However, these data suggest that routine screening of symptomatic patients with rectal prolapse by use of flexible sigmoidoscopy is warranted. The assistance of Emanuel Lerner, MA, with statistical analysis is greatly appreciated. REFERENCES
1. KaranjiaND, Rutter KRP. Solitaryrectal ulcer and rectal prolapse. Current Practice in Surgery 1994;6:59-64. 2. Gordon PH. Rectal procidentia. In: Gordon PH, NivatvongsS. Principlesand practiceof surgeryfor the colon,rectum and anus. 2nd ed. St Louis: QualityMedical PublishingInc, 1992:449-81. 3. Kodner IJ, Fry RD, FleshmanJW, BirnbaumEH. Colon, rectum and anus.In: Schwartz,S. Principlesof surgery.6th ed. NewYork: McGraw Hill, Inc, 1994:1191-306. 4. Britten-JonesR. Complete rectal prolapse: anterior wrap (Ripstein operation). In: DudleyH, Carter DC, RussellRCG.Atlasof generalsurgery.2nd ed. London:Butterworths,Inc, 1986:575-80. 5. Hicks TC, Opelka FG. Rectal and perianal complaints.In: Polk HC Jr, Gardner B, Stone HH. Basic surgery. 4th ed. St. Louis: Quality Medical PublishingInc, 1993:545-60. 6. Fry RD, Kodner IJ. Anorectaldiseases. In: LevineVA, Copeland EM III,HowardRJ, SugermanHJ, WarshawAL. Current practice of surgery. New York: ChurchillLivingstone,1993:1-23. 7. Dent TL, KukoraJS, NejmanJH. The colon, rectum and anus. In: HardyJD. Hardy'stextbookof surgery. 2nd ed. Philadelphia: JB LippincottCompany, 1988:582-636. 8. FleischerDE, GoldbergSB, BrowningTH, et al. Detection and surveillanceof colorectal cancer.JAMA1989;261:580-6. 9. GilbertsonVA, NelmsJM. The preventionof invasivecancer of the rectum. Cancer 1978;41:1137-9. 10. Friedman GD, Collen MF, Fireman BH. Multiphasic health checkup evaluation:a 16yearfollow-up.JChron Dis 1986;39:45363. 11. KeighlyMRB,Matheson DM, DuncanMM. Resultsof treatment for rectal prolapse and fecal incontinence. Dis Colon Rectum 1981;24:449-53. 12. DisarioJA,Foutch P. Prevalenceand malignantpotential of colorectal polypsin asymptomatic,average risk men. AmJ Gastroenterol 1991;86:941-5. 13. GuptaTP,JaszewskiR, Luk GD. Efficacyof screeningflexiblesigmoidoscopy for colorectal neoplasia in asymptomaticsubjects. AmJ Med 1989;86:547-50. 14. Rex KD, Lehman GA, Ulbright TM, et al. Colonicneoplasiain asymptomatic personswith negativefecal occult blood tests: in-
Surgery Volume 119, Number 1 fluence of age, gender, and family history. Am J Gastroenterol 1993;6:825-31. 15. Day T, Ferrara MD. Preoperative barium contrast enema in patients with inguinal hernia. South MedJ 1986;79:1339-41. 16. Rubin BG, Ballantyne GH, Zdon MJ, Zucker KA, Modlin IM. The role of flexible sigmoidoscopy in the preoperative screening of patients with inguinal hernia. Arch Surg 1987;122:296-9. 17. Maccini MM, Miller RM. The yield of barium enema in patients undergoing inguinal hernia repair or abdominal hysterectomy. Surg Gynecol Obstet 1991;172:391-3. 18. Pratt SM, Weaver FA, PottsJR. Preoperative evaluation of patients with inguinal hernia for colorectal disease. Surg Gynecol Obstet 1987;165:5~6. 19. Wheeler WE, Wilson SL, Kurucz J, Flanigan J, Ratliff D, ScottConner C. Flexible sigmoidoscopy screening for asymptomatic colorectal disease in patients with and without inguinal hernia. South MedJ 1991;84:876~.
Rashid and Basson
55
20. LovettJ, Kirgan D, Mcgregor B. Inguinal herniajustifies sigmoidoscopy. Am J Surg 1989;158:615-7. 21. Brendel TH, Kirsh IE. Lack of association between inguinal hernia and carcinoma of the colon. N EnglJ Med 1971;284:369-70. 22. Seldman H, Mushinski MH, Gelb SK, et al. Probabilities of eventually developing or dying of cancer. United States, 1985. CA 1985;35:36-56. 23. Stryker SJ, Wolff BG, Culp CE, Libbe SD, Ilstrup DM, MacCarty RL. Natural history of untreated colonic polyps. Gastroenterology 1987;93:1009-13. 24. Muto T, Bussey HJR, Morson BC. Cancer 1975;36:2251-70. 25. Connecticut HealthPlan, Inc. Fee schedule information (1994). Meriden, Connecticut: Connecticut HealthPlan Inc, 1995. 26. Wheeler WE, Scott-Conner CEH, Stone RA. Flexible sigmoidoscopy as a screening procedure for asymptomatic colorectal carcinmna in patients with inguinal hernia. South MedJ 1985; 78: 1417-20.
SURGERYiS abstracted a n d / o r i n d e x e d in Index Medicus, Science Citation Index, Current Contents/ Clinical Medicine, Current Contents~Life Sciences, Excerpta Medica, a n d MEDI,INE. This JOURNALhas b e e n registered with Copyright Clearance Center, Inc., 222 Rosewood Dr., Danvers, MA 01923. C o n s e n t is given for the copying o f articles for personal or internal use o f specific clients. This c o n s e n t is given o n the condition that the copier pay directly to the C e n t e r the per-copy fee stated o n the first page o f each article for copying b e y o n d that p e r m i t t e d by U.S. Copyright Law. This c o n s e n t does n o t e x t e n d to o t h e r kinds o f copying, such as for general distribution, resale, advertising a n d p r o m o t i o n a l purposes, or for creating new collective works. All inquiries regarding c o p y r i g h t e d material f r o m this publication o t h e r than those that can be h a n d l e d t h r o u g h Copyright Clearance C e n t e r should be d i r e c t e d to Associate J o u r n a l Publisher, Permissions, Mosby-Year Book, Inc., 11830 Westline Industrial Dr., St. Louis, MO 63146; p h o n e (314)579-2808; fax (314)432-1380.