Colonoscopy Reported to be Better for Polyp Surveillance Than Barium Enema
Colonoscopy Reported to be Better for Polyp Surveillance Than Barium Enema
Abstract & Commentary
Synopsis: In a prospective study of follow-up (surveillance) examinations in high-risk patients, colonoscopy found many more polyps (and many more adenomas) than did double contrast barium enema. Though not without its faults, this study is likely to become a major citation arguing for increased use of colonoscopy over the barium enema.
Source: Winawer SJ, et al, for the National Polyp Study Work Group. A comparison of colonoscopy and double-contrast barium enema for surveillance after polypectomy. N Engl J Med 2000;342:1766-1772.
A long-standing competition between the double-contrast barium enema and colonoscopy for their relative roles in evaluating patients for colonic lesions is updated in an article from the New England Journal of Medicine comparing the two techniques for polyp surveillance following initial polypectomy.
Under the auspices of The National Polyp Study, patients who had an adenoma removed at colonoscopy were eligible to participate in a study of polyp surveillance strategies, specifically a randomized comparison of colonoscopy at one and three years after initial adenoma removal vs. at three years alone. All participants were subsequently offered a colonoscopy at six years. Prior to each colonoscopy, patients were offered an additional double-contrast barium enema (DCBE), which was performed on average 16 days (range, 2-164) before the colonoscopy, with findings recorded at eight anatomic segments (e.g., cecum, splenic flexure) within the colon. At the post-DCBE colonoscopy, the endoscopist performed colonoscopy blinded to the DCBE segmental results. After the endoscopist reported the findings for a given segment from cecum to rectum, the study coordinator told the endoscopist of the results of the DCBE for that segment. If there was a finding on DCBE not seen on initial colonoscopy, the endoscopist reexamined the segment several times until a lesion was found, or the endoscopist was satisfied that no lesion was present (unblinded colonoscopy). Endoscopically detected polyps were removed.
Of the total 1418 patients entered into the original polyp surveillance study, 973 underwent at least one surveillance colonoscopy, for a total of 1762 examinations, of which 881 were paired with DCBE examinations. Nineteen paired examinations were excluded: the cecum was not reached by colonoscopy in 12, the time between examinations was greater than six months in six, and both reasons were true for one patient. That left 862 paired examinations in 580 patients that form the core of the report. The examinations were performed from 1981 through 1990. (I wonder why it took a decade to report the results of a study that doesn’t require patient follow-up?) There were no major complications after any of the paired examinations.
In the 862 colonoscopic examinations, polyps of any kind were found in 392 (45%), adenomas were found in 242 (28%), and adenomas more than 1 cm in size were found in 23 (3%). Polyps were reported in 222 (26%) of the DCBE examinations. DCBE examinations were positive (at least 1 polyp seen) in 139 (35%) of the 392 colonoscopies that found polyps, in 94 (39%) of the 242 colonoscopies that found adenomas, and in 11 (48%) of the 23 colonoscopies that found adenomas greater than 1 cm. Further size analysis for the adenomatous polyps showed that DCBE was positive in 32% of colonoscopies finding adenomas of 0.5 cm or less, 53% of colonoscopies finding adenomas of 0.6-1 cm, and, as above, 48% of colonoscopies finding adenomas of greater than 1 cm.
On a per-polyp basis, colonoscopy detected 791 total polyps, of which 160 (20%) were detected by DCBE. DCBE found 103 (27%) of the 375 polyps that were adenomas. The rate of detection of adenomas was significantly related to size, with DCBE finding 21% of adenomas 0.5 cm or less, 42% of adenomas 0.6-1 cm, and 46% of adenomas of greater than 1 cm. DCBE found 34% of the endoscopically detected polyps on the left side of the colon compared to 22 of right-sided polyps, a statistically significant difference. Only two malignant polyps were found at colonoscopy. Both were 1.5 cm and cecal in location. One was detected by DCBE and one was not.
Polyps were found by DCBE in 83 (18%) of the 470 colonoscopies that were negative at the initial examination blinded to the DCBE results, leading to unblinded colonoscopies reexamining the areas in question. (The text states 139 unblinded colonoscopies, but I think the table has the correct figure.) Nineteen polyps (12 adenomas) were found during the unblinded colonoscopies. Winawer and colleagues calculate a rate of missed adenomas by colonoscopy of 20% as follows: 47 adenomas were found during the blinded portion of these examinations and 12 additional adenomas were found during the unblinded portion. Thus, 12 of the total 59 adenomas represents 20% missed by blinded colonoscopy.
Winawer and associates concluded that their study supports what is already happening in practice: that "colonoscopic examination has become the preferred way of examining the colon for both the detection and removal of polyps, replacing diagnostic barium enema as a means of surveillance." They also concluded that "the low rate of detection of large adenomas with barium enema is a drawback to the use of this radiologic technique as the primary surveillance tool."
Comment by James H. Ellis, MD
I saw this article comparing diagnostic results of colonoscopy with DCBE covered in the newspaper, and perhaps it was reported in other of the popular media as well. Because it was published in a prestigious journal, it has taken on a certain cachet. An accompanying editorial lauds it as "at last there is a good study of the accuracy of barium enema."1 Yet, in my opinion, this article is flawed in many ways despite it being one of the few prospective comparisons of colonoscopy and the barium enema. Some of the limitations of the research are noted by the authors, but many are not.
One of the most critical problems is the way in which the two examinations were compared: the colonoscopy was always considered the gold standard. Granted, the barium enema found many fewer polyps than colonoscopy; but it would seem that this would allow Winawer et al to be magnanimous about those polyps found by the DCBE and not by colonoscopy, and count all those as false-negatives for colonoscopy. In 83 of the paired examinations, the DCBE found at least one polyp, and the blinded colonoscopy found none. The exact number of such DCBE-detected polyps is not even reported, but on re-inspection, colonoscopy found 19 polyps. In the remaining cases, the negative repeat colonoscopy was assumed to be "truth." We do not know how hard the endoscopist searched for a lesion that had just been "missed" by that very physician. A second endoscopist, who might have used somewhat different technique, was not at hand to look for the missing lesion. No mention is made of endoscopic "blind spots."2,3 Additional inspection would never succeed in finding those polyps that cannot be seen because of a limitation of the colonoscopic technique rather than because of an observational error. Location was critical in determining whether DCBE was credited with finding a polyp. Winamer et al admit that telescoping of the bowel can make it difficult to know where the tip of the endoscope is. Though they did attempt to match similar sized polyps in the same or adjoining segments, it is unknown how many times the endoscopist may have been more than one segment off in determining the location of the tip of the scope. Yet these data are used to state that colonoscopy has a miss rate of 20%, whereas it could potentially be much higher.
Other biases in the study are not pointed out. Though not clearly stated, it is intimated that a complete colonoscopy (to the cecum) was a prerequisite for entry into the study, thus providing a group of patients in which the most difficult anatomy for complete colonoscopy has already been weeded out. This bias is compounded by excluding the baker’s dozen of patients who later underwent both examinations but in whom colonoscopy was incomplete. Some DCBE had poor mucosal coating, making detection of polyps more difficult, but these patients were not excluded from analysis.
The rate of detection of polyps was very low for the barium enema, only 46% for polyps larger than 1 cm. This is a far cry from other estimates that the barium enema may be 70-90% sensitive for detecting polyps of this size.2 No explanation for this discrepancy is given. Indeed, Winawer et al do not give the results of prior studies, nor even mention their existence except to dismiss them.
Granted this was not the point of the project, but Winawer et al studiously avoid much mention of cost or complications of the two examinations, other than to say that no patients in the study experienced any major complications. Even the laudatory editorial by Fletcher allows that cost and complication rates favor the DCBE.1 Whether the polyps missed by DCBE are clinically important, and therefore worth the increased cost and risk of colonoscopy to detect, is only obliquely touched by Winawer et al. However, even in these high-risk patients of the National Polyp Study, fewer than 3% of the surveillance colonoscopic examinations found adenomas larger than 1 cm.
Winawer et al’s conclusions are broad and don’t make specific reference to this high-risk population, though they do use the word surveillance to distinguish the intent of these examinations from screening. This may lead the unsophisticated to think that the article can be applied without critical thinking to screening for colon cancer in the general population. This question remains to be answered because of the increased importance of cost, complications, and outcomes when the issue is screening rather than diagnosis.3 Perhaps CT colonography, rather than barium enema, will prove to be the test against which colonoscopy should be measured as a screening test for colorectal cancer. CT colonography may find significant extracolonic abnormalities in 10% or more examinations,2 which may contribute to its value as a screening test.
Fletcher’s editorial cites a British colonoscopy study in which more than a third of the adenomas detected by colonoscopy had a flat morphology.1 If confirmed, this does not bode well for imaging examinations that rely primarily on detecting mass lesions that project into the lumen, including the barium enema and virtual colonoscopy. Even without this concern, the facts of the Winawer et al study remain: the barium enema misses a lot of polyps found by endoscopy. With the insatiable need for certainty at any cost that our medical training instills in us and our legal system reinforces, this may be one of the last nails in the coffin of the DCBE.
References
1. Fletcher RH. The end of barium enemas? N Engl J Med 2000;342:1823-1824.
2. Johnson CD, Dachman AH. CT colonography: The next colon screening examination? Radiology 2000;216:331-341.
3. Nelson RL. Screening for colorectal cancer. J Surg Oncol 1997;63:249-258.
Subscribe Now for Access
You have reached your article limit for the month. We hope you found our articles both enjoyable and insightful. For information on new subscriptions, product trials, alternative billing arrangements or group and site discounts please call 800-688-2421. We look forward to having you as a long-term member of the Relias Media community.