Current issues of ACP Journal Club are published in Annals of Internal Medicine


Experienced colonoscopists missed 24% of adenomas

ACP J Club. 1997 Jul-Aug;127:16. doi:10.7326/ACPJC-1997-127-1-016

Source Citation

Rex DK, Cutler CS, Lemmel GT, et al. Colonoscopic miss rates of adenomas determined by back-to-back colonoscopies. Gastroenterology. 1997 Jan;112:24-8.



To determine the accuracy of colonoscopy for identifying colonic neoplasms.


Blinded comparison of the diagnostic accuracy of colonoscopy with a second colonoscopy done on the same day.


A hospital in the United States.


183 patients (mean age 62 y, 62% men) with indications for colonoscopy. Exclusion criteria were previous surgical resection of the colorectum, inability to clear the colon of polyps because of anticoagulation, or poor health that would make 2 colonoscopies on the same day potentially harmful.

Description of test and diagnostic standard

Colonoscopies were done by attending physicians (all had done ≥ 500 colonoscopies). The first examination was done in either the left lateral decubitus or supine position. The initial colonoscopist passed the instrument tip to the cecum and removed all polyps during withdrawal. For the second examination (the diagnostic standard), patients were randomly allocated to 1 of 4 combinations: same position, same examiner; different position (prone), same examiner; same position, different examiner; or different position (prone), different examiner.

Main outcome measures

Miss rates of colonoscopy for adenomas.

Main results

289 adenomas were detected during the initial colonoscopy, and 89 were detected during the second colonoscopy (overall miss rate 24%). Miss rates increased as the size of the adenomas decreased (miss rates for adenomas ≤ 5 mm, 6 to 9 mm, and ≥ 10 mm were 27%, 13%, and 6%, respectively). {The sensitivity of the initial colonoscopy for detecting adenomas was 76% (95% CI 62% to 87%), specificity was 51% (CI 42% to 60%), and likelihood ratios of a positive or negative test result were 1.55 and 0.47, respectively.}* Miss rates did not differ among the 4 groups or in individual colonic segments. 2 large adenomas (≥ 1 cm) were missed in the first examination by first-year faculty members.


Colonoscopy was sensitive for the detection of adenomas, and miss rates were lower for larger adenomas. Up to 27% of small adenomas (≤ 5 mm) were not detected during first colonoscopy.

Source of funding: Not stated.

For article reprint: Dr. D.K. Rex, Division of Gastroenterology, Indiana University School of Medicine, Room 2300, 550 North University Boulevard, Indianapolis, IN 46202, USA. FAX 317-274-3106.

*Numbers calculated from data in article.


If the doors of perception were cleansed everything would appear to man as it is, infinite.

William Blake, 1793

Before the introduction of fiberoptic endoscopy, studies on the prevalence of colorectal adenomas were done on autopsy material. These studies showed that the prevalence of adenoma increased with age to > 50% for some middle-aged and older populations (1). Prevalence rates based on colonoscopy in symptomatic populations have ranged from 20% to 40%. The difference in reported prevalence rates probably reflects the intensity of scrutiny. In the study by Rex and colleagues, the adenoma prevalence rate of 57% confirmed that examinations were truly intense and meticulous.

In medicine, the harder you look, the more you are likely to find. Similarly, axillary nodal metastasis is found in patients with breast cancer whose axillary nodes are sliced more finely (2). The good news in the study by Rex and colleagues is that skill may increase with experience; given the practical realities of colonoscopic examination today, we are probably not missing much that is important.

On the other hand, enhanced sensitivity with 2 colonoscopies may lead to a Will Rogers-like phenomenon (3) in which the number of polyps detected increases but the number of patients who are truly at risk for colorectal cancer does not. We will, however, have to deal with surveillance for more polyp-bearers. Further, many chemoprevention trials use adenoma recurrence as an end point; most of these trials to date have had negative results. If the recurrent adenomas in these trials do not represent new growths but rather old misses, the negative trials may not be truly negative (4).

Alfred I. Neugut, MD, PhD
Columbia-Presbyterian Medical CenterNew York, New York, USA


1. Neugut AI, Jacobson JS, DeVivo I. Epidemiology of colorectal adenomatous polyps. Cancer Epidemiol Biomarkers Prev. 1993; 2:159-76.

2. de Mascarel I, Bonichon F, Coindre JM, Trojani M. Prognostic significance of breast cancer axillary lymph node micrometastases assessed by special techniques: reevaluation with longer follow-up. Br J Cancer. 1992; 66:523-7.

3. Feinstein AR, Sosin DM, Wells CK. The Will Rogers phenomenon. Stage migration and new diagnostic techniques as a source of misleading statistics for survival in cancer. N Engl J Med. 1985;312:1604-8.

4. Schatzkin A, Freedman LS, Dawsey SM, Lanza E. Interpreting precursor studies: what polyp trials tell us about large-bowel cancer. J Natl Cancer Inst. 1994;86:1053-7.