Indications for Colonoscopy Revisited

 

 Buy Article Permissions and Reprints

Population-based surveys show that scant bright red rectal bleeding is very common in adults of all ages. Dent et al. reported a positive history of this type of bleeding in 14 % of an apparently well population [1]. Tally et al. conducted a population-based survey in Olmsted County, Minnesota [2]. Scant bright red blood per rectum was reported by 13 % of those surveyed, and was more common among younger adults. Only 14 % of people with any type of rectal bleeding had consulted their physician about the problem. Since unexplained rectal bleeding in adult patients - overt or occult - is highly predictive of the presence of advanced colorectal neoplasia, all patients with this history require evaluation. However, there is considerable disagreement among both primary-care physicians and gastroenterology specialists on whether all patients with hematochezia require full colonoscopy, or whether flexible sigmoidoscopy is sufficient, in selected patients, to rule out important neoplastic disease.

In the USA and many European countries, an increasing number of experts advocate full colonoscopy for all patients with hematochezia, regardless of their age or the nature of their bleeding. They argue that the only reliable way to rule out advanced colorectal neoplasia is to examine the entire large bowel endoscopically. Admittedly, the present author until now belonged to the same school of thought [3]. It is, however, the right, and indeed the obligation, of all medical scientists and practitioners to be willing to change their approach to a clinical problem when presented with new, persuasive, high-quality scientific evidence. The paper by Eckardt et al. in this issue of Endoscopy is, I believe, an epiphanic event of this type [4]. It presents compelling new evidence that flexible sigmoidoscopy may represent sufficient work-up for most patients who have experienced small amounts of bright red rectal bleeding. This excellent study employed structured preprocedural interviews in 4265 patients referred for colonoscopy. Of these, 468 patients had scant hematochezia, defined as small amounts of bright red blood on top of their stool and/or on the toilet paper. An additional 299 patients had occult stool blood only, and 57 reported dark or burgundy-colored blood, and/or blood that was mixed with their stool. Patients were excluded from this analysis if they had inflammatory bowel disease, a personal or family history of colorectal neoplasia, or an incomplete colonoscopy. At colonoscopy, patients with scant hematochezia had no increased prevalence of neoplasms located above 50 cm in the colon, compared to an equivalent sex-matched and age-matched group of patients undergoing colonoscopy who had no rectal bleeding or other risk factors for colorectal neoplasia (OR 1.2). In contrast, this risk was significantly increased in patients with either occult bleeding (OR 3.1) or maroon-colored blood in their stool (OR 4.8).

Previous studies of this issue have provided conflicting conclusions, usually based on smaller numbers of patients without a comparable control group with no symptoms or risk factors for colorectal cancer. Fine et al. determined the yield of colonoscopy in 312 consecutive patients presenting with rectal bleeding [5]. Of 217 patients with passage of small amounts of bright red blood, 181 bled from lesions in the distal 60 cm of the large bowel, and 20 had more proximal neoplasms, including eight with cancer. The authors concluded that the diagnostic evaluation of patients with hematochezia of all types should begin with colonoscopy. A retrospective study by Guillem et al. of 372 consecutive patients undergoing colonoscopy for rectal bleeding arrived at the same conclusion [6]. They found that the yield at colonoscopy was similar, regardless of the pattern or nature of the rectal bleeding, and that a third of detected neoplasms were proximal to the splenic flexure.

In contrast, a smaller study by Segal et al. of 103 patients with hematochezia, defined as the passage of bright red blood, concluded that very few clinically important lesions are missed when flexible sigmoidoscopy is used as the primary diagnostic method [7]. Of these patients, 28 were found to have clinically significant neoplasia - four cancers and 31 adenomatous polyps > 7 mm in diameter. However, only six of the polyps and none of the cancers were located proximal to the sigmoid colon. The authors concluded that flexible sigmoidoscopy would have detected most (95 %) of the important neoplasms in these patients. One of the most convincing earlier studies of this question was published in 1991 by Church of the Cleveland Clinic [8]. A total of 269 patients referred for colonoscopy for evaluation of rectal bleeding were prospectively categorized according to a careful determination of the pattern of their bleeding. In 115 patients with “outlet bleeding” (defined as bright red blood seen during or after defecation, on the toilet paper or in the toilet bowl) and no other symptoms or special risk factors for colorectal cancer, only one adenomatous polyp and no cancers would have been missed by flexible sigmoidoscopy. However, those patients who passed dark red blood, or had blood mixed with or streaked on stool, larger volumes of bleeding, or occult stool blood, had a high prevalence of proximal neoplasia that would have been missed by flexible sigmoidoscopy. Church’s conclusions are in agreement with those of the present study by Eckardt et al. It is sufficient to begin the evaluation of outlet-type bleeding with flexible sigmoidoscopy; all those with other patterns of rectal bleeding should undergo immediate colonoscopy.

Eckardt and co-workers correctly emphasize that the inappropriate use of colonoscopy wastes a valuable, expensive, and in many areas a scarce resource that is critically needed in our attempts to reduce the great mortality from colorectal cancer. Whether it is used as a direct screening test (as is now strongly advocated by some groups) or as a diagnostic response to a positive simpler screening test, colonoscopy is the linchpin of our efforts to control this deadly disease [9]. Some argue that any reason that persuades a patient to have a colonoscopy is worthwhile, because it accomplishes this screening objective. Others, however, correctly point out that the current limited capacity to perform colonoscopy examinations means that we must carefully ration this resource if we are to accomplish our screening goals. Recent large outcomes studies have shown that colonoscopy performed to evaluate abdominal symptoms or a change in bowel movement patterns, or for too frequent postpolypectomy surveillance, does not detect an increased prevalence of advanced colorectal neoplasia. Neugut et al., in a multicenter prospective study, found that only 7.1 % of patients with abdominal pain or change in bowel habits had an advanced adenoma or a cancer at colonoscopy, a prevalence equal to that of an asymptomatic population [10]. In a review of reported colonoscopy series, Rex found that colorectal cancer was detected in only one of 109 colonoscopies performed to evaluate nonbleeding symptoms, and in only one of 317 when performed for postpolypectomy surveillance [11]. Chak et al., in an analysis of 1223 consecutive colonoscopies performed at the University Hospitals of Cleveland, reported no association between colorectal cancer and symptoms in 298 patients presenting with abdominal pain, constipation, or diarrhea [12]. Lastly, Lieberman et al. used a large multicenter computer database to determine findings in 20 745 patients undergoing colonoscopy in 31 practice sites over an 18-month period [13]. Patients with symptoms of abdominal pain, constipation, or diarrhea did not have a higher prevalence of clinically important colorectal neoplasia than did asymptomatic patients. Rex and Lieberman recently have estimated that 27 - 50 % of postpolypectomy follow-up surveillance colonoscopies are performed inappropriately and unnecessarily, wasting resources that could be better used for screening [14]. Based on Eckardt’s findings in the paper under discussion here, scant hematochezia can now be added to the list of indications for which colonoscopy is not routinely indicated.

Caveats are important when dealing with the evaluation of any sign or symptom possibly due to colorectal neoplasia. If no cause of scant bright red rectal bleeding is found by anoscopy and flexible sigmoidoscopy, or if bleeding persists in spite of treatment of a benign lesion, colonoscopy should usually be performed, regardless of the patient’s age. Even if the colonoscopic examination is negative, this approach will substantially relieve both physician and patient anxiety about cancer, and will obviate the need to do any further colorectal cancer screening for up to 10 years, according to current guidelines [15]. Lastly, the findings of Eckardt et al. are in agreement with those of numerous previous studies that show the substantial prevalence of clinically serious proximal colonic neoplasms in patients with occult stool blood or dark blood in the stool, which can only be detected accurately and promptly by performing immediate full colonoscopy.

References

• 1 Dent O F, Goulston K J, Zubrzycki J, Chapulis P H. Bowel symptoms in an apparently well population.  Dis Colon Rectum. 1986;  29 234-237
  CrossrefPubMedGoogle Scholar
• 2 Tally N J, Jones M. Self-reported rectal bleeding in a United States community: Prevalence, risk factors and health care seeking.  Am J Gastroenterol. 1998;  93 2179-2183
  CrossrefPubMedGoogle Scholar
• 3 Bond J H. Rectal bleeding: Is it always an indication for colonoscopy?.  Am J Gastroenterol. 2002;  97 223-225
  CrossrefPubMedGoogle Scholar
• 4 Eckardt V F, Schmitt T, Kanzier G, Eckardt A J, Bernhard G. Does scant hematochezia necessitate the performance of total colonoscopy?.  Endoscopy. 2002;  34 599-603
  Article in Thieme ConnectPubMedGoogle Scholar
• 5 Fine K D, Nelson A C, Ellington R T. et al . Comparison of the color of fecal blood with the anatomic location of gastrointestinal bleeding lesions: Potential misdiagnosis using only flexible sigmoidoscopy for bright red blood per rectum.  Am J Gastroenterol. 1999;  94 3202-3210
  CrossrefPubMedGoogle Scholar
• 6 Guillem J G, Forde K A, Treat M R. et al . The impact of colonoscopy on the early detection of colonic neoplasms in patients with rectal bleeding.  Ann Surg. 1987;  206 606-611
  PubMedGoogle Scholar
• 7 Segal W N, Greenberg P D, Rockey D C. et al . The outpatient evaluation of hematochezia.  Am J Gastroenterol. 1998;  93 179-182
  CrossrefPubMedGoogle Scholar
• 8 Church J M. Analysis of the colonoscopic findings in patients with rectal bleeding according to the pattern of their presenting symptoms.  Dis Colon Rectum. 1991;  34 391-395
  PubMedGoogle Scholar
• 9 Rex D K, Johnson D A, Lieberman D A, Burt R W, Sonnenberg A. Colorectal cancer prevention 2000: Screening recommendations of the American College of Gastroenterology.  Am J Gastroenterol. 2000;  95 868-877
  PubMedGoogle Scholar
• 10 Neugut A I, Garbowski G C, Waye J D. et al . Diagnostic yield of colorectal neoplasia with colonoscopy for abdominal pain, change in bowel habits, and rectal bleeding.  Am J Gastroenterol. 1993;  88 1179-1183
  PubMedGoogle Scholar
• 11 Rex D K. Colonoscopy: A review of its yield for cancers and adenomas by indication.  Am J Gastroenterol. 1995;  90 353-365
  PubMedGoogle Scholar
• 12 Chak A, Post A B, Cooper G S. Clinical variables associated with colorectal cancer on colonoscopy: A prediction model.  Am J Gastroenterol. 1996;  91 2483-2488
  PubMedGoogle Scholar
• 13 Lieberman D A, de Garmo P L, Fleischer D E. et al . Colonic neoplasia in patients with nonspecific GI symptoms.  Gastrointest Endosc. 2000;  51 647-651
  CrossrefPubMedGoogle Scholar
• 14 Rex D K, Lieberman D A. Feasibility of colonoscopy screening: Discussion of issues and recommendations regarding implementation.  Gastrointest Endosc. 2001;  54 662-667
  CrossrefPubMedGoogle Scholar
• 15 Smith R A, von Eschenbach A C, Wender R. American Cancer Society guidelines for the early detection of cancer: Update of early detection guidelines for prostate, colorectal, and endometrial cancers.  CA Cancer J Clin. 2001;  51 38-75
  CrossrefPubMedGoogle Scholar

J. H. Bond, M.D.

Gastroenterology Section (111D) · VA Medical Center

One Veterans Drive · Minneapolis, MN 554l7 · USA

Fax: + 1-12-725-2248

Email: john.bond@med.va.gov