Medicine For People!

September 2010 Appendix: On Colonoscopy

Note to Physician Readers

While Medicine for People! aims at a lay audience, I've included copious footnotes so that you may evaluate my reasoning. Clearly all-cause mortality is the best criterion for any medical intervention, including screening. Unfortunately, there are no such data for iFOBT, but neither are there for colonoscopy.

My goal is to increase screening rates and decrease mortality and cost. Many people who would never have undertaken colonoscopy will be appropriately endoscoped once enrolled in a two-step ifobt-colonoscopy program. While specialty groups hold out colonoscopy as the sine qua non of colon cancer prevention, the American Cancer Society, the US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology, and the USPFTF point out in various guidelines that the iFOBT, done annually, detects over 90 percent of colon neoplasms.1

Faulty studies overstate success rates

Not all of the literature in this field withstands careful analysis. As an example, a study (Gut 2001; 48:812–ndash;815) took a group of people, gave them a colonoscopy with or without an additional double contrast barium enema, then dropped those who were negative on both tests from consideration. The positive group (any colon lesion greater than 5 mm) was followed for about ten years, three quarters of them receiving additional colonoscopies. Nine developed colon cancer, three of which they dropped from consideration as the cancers occurred early and these were considered "failed colonoscopies." Then they compared the remaining colon cancer count with a reference population of unscreened people. What's wrong with this study?

  • First, they really are not studying colonoscopy, but colonoscopy sometimes combined with barium enema.
  • Second, they don't count all the cancers that occurred in their screened group.
  • Third, they violate one of the primary rules of scientific proof. They assume that which they set out to prove. In this case, they assume de novo colon cancer does not occur, (hence the omission of three cancers from their count) then at the end take the data so derived and say "In conclusion, our data support the concept of the adenoma-carcinoma sequence as the major pathway of colon carcinogenesis" page 815. While other studies tell us that they are probably correct that adenoma-carcinoma is the major pathway, their conclusions are tainted by their premises.

Fourth, the study data best applies to the group studied, people positive on one or both of a colonoscopy and double-contrast barium enema, a group of whom 74 percent had at least one follow-up colonoscopy within a ten year interval. The study data do not demonstrably apply to a general population screened with colonoscopy alone and followed for ten years.

Other studies on the sensitivity of screening colonoscopy

Estimates of colonoscopy' reduction of colon cancer deaths run from 97 percent2 to 90 percent3 to 60 percent45 of what it would otherwise be.

Additional information re "missed polyp" study

During the second colonoscopy in both groups, an attempt was made to keep the withdrawal examination time equal to that of the first colonoscopy. To facilitate this, the assistant announced accumulated inspection time every 30 to 60 seconds.

No dysplastic polyps were missed. See table below for details.

Results of "missed polyp" study


Found on first cscope

Found on second cscope

Percent missed on first exam



Number with positive colonoscopy


Number of polyps found





* polyps 10 mm or greater in size





* polyps 6 to 9 mm in size





* polyps 1 to 5 mm in size





Table data from "Cap-fitted colonoscopy: a randomized, tandem colonoscopy study of adenoma miss rates." Hewett DG, Rex DK. Gastrointest Endosc. 2010 Jun 23; [Epub ahead of print]

Colonoscopy Complication Rates

Below are three large studies, the most recent reported in 2008. In that study, reported in the Archives of Surgery, records were reviewed for over 250,000 colonoscopies. During those colonoscopies, the colon wall was inadvertently perforated 180 times, or once every 1435 procedures.

J Natl Cancer Inst. 2003 Feb 5; 95(3):230-6. 

Number of procedures to result in a perforation

Procedures done


Perforations of colon wall



Deaths resulting from procedure


Deaths per 10,000 colonoscopies


Can J Gastroenterol. 2004 Apr; 18(4):221-6. 

Procedures done


Perforations of colon wall



Deaths resulting from procedure


Deaths per 10,000 colonoscopies


Arch Surg. 2008;143(7):701-707

Procedures done


Perforations of colon wall



Deaths resulting from procedure


Deaths per 10,000 colonoscopies


Note that more recent studies show reduced complication rates. This is not an accident. Much work has gone into lowering those complication rates.

Complication rates are lower when compared to control group

The US Preventive Services Task Force estimated in 2008 that 25 serious complications occur per 10,000 procedures6. For a period of 30 days, investigators compared people who had just had a colonoscopy with a matched group of people who had not. Because certain conditions increase the risk of colonoscopy, they did not include people with known disease of the colon or those for whom they did not have complete information, leaving 53,000 in the final analysis. Only complications requiring an Emergency Department visit or hospitalization were counted.

The beauty of this approach, of course, is that there are always a few people out there who develop GI bleeding, or heart disease, or whatever, so if you want to really know what the sequelae of a procedure are, you need to compare your colonoscopy group with the man on the street.

Here are the problems occurring in this group of 53,000 people, compared with 53,000 people of similar age, gender, education, income, health status, etc. (For the control group, a random date was chosen as the date they "didn't have their colonoscopy," and ER visits and hospitalizations counted for the next 30 days.)

In each case, problems were severe enough to warrant an ER visit or hospitalization. The group was not large enough to determine if there was any difference in the death rate due to colonoscopy.

Events per 1000 people within 30 days of outpatient colonoscopy

Control group

Screening colonoscopy, no polyp removal

Colonoscopy with polyp removal

Excess with polyp removal

Perforation, GI bleeding, transfusion

about 2 or 3 in each group



Paralytic ileus, severe pain, nausea, vomiting, dehydration

about 6 in each group



Irregular heartbeat, heart failure

about 14 in each group



Complications per 1000 people, above baseline rate, due to polyp removal


Screening colonoscopy frequently occurs more than once in ten years

Frequently people will be asked to return for a repeat colonoscopy within 3 to 5 years if they had a polyp removed during their colonoscopy. Physicians vary widely in their recommendations as to when a repeat screen should occur when a polyp is found the first time.

To help answer this question, doctors in Japan recorded results7 of over 5000 patients who underwent colonoscopy and had all identified colon growths were removed. Five years later, on average, each had a second colonoscopy. Each row gives the findings for the subdivisions of those original 5000 patients, from those with no polyps found (the third row) all the way down to those with a cancer on the first study (the bottom row.) The columns show what was found at the second colonoscopy.

Number of cancers

Percent with cancer

Results of original colonoscopy

Number of people







No growths found








Polyps 1 to 5 mm in size








Polyps 6 mm or larger








Any localized cancer








Key to above table:

  • Adenoma- a benign lesion 10 mm in size or over
  • Mucosal cancer –ndash; the earliest stage of colon cancer, confined to the innermost lining of the colon
  • Submucosal cancer –ndash; the second stage, the cancer still confined to the colon wall, but one layer deeper
  • Advanced cancer- stages beyond the above two

More about cancer staging.

The study shows that polyps 6 mm or over, removed at the first screening, heralded a greater likelihood of cancer on the second screening.

These cancers were found for one of several reasons.

  1. The cancer (or the aggressive polyp which then developed into a cancer) was missed on the initial colonoscopy. Some precursor cancers are flat and not easy to distinguish from normal colon wall.
  2. The cancer was not present previously, but developed "de novo" as mentioned in our last newsletter
  3. The person had an increased predisposition to cancer, as evidenced by the larger number of cancers in people with polyps, especially larger polyps.

This information poses a dilemma for patient and doctor. Someone with a polyp on a colonoscopy is at higher risk for another, and if so, at higher risk for colon cancer. Patient and doctor need to develop a rational method for future screening. Too little screening lead to cancer for the patient and a lawsuit for the doctor. Too much brings the hazard of colonoscopy complications.


1 Levin B, Lieberman DA, McFarland B, et al. Screening and surveillance for early detection of colorectal cancer and adenomatous polyps, 2008: a joint guideline from the American Cancer Society, the US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology. Gastroenterology 2008;134:1570-95.

The sensitivity for cancer with 3 FIT samples with a hemoglobin threshold set at 75 ng/mL was 94.1%. Specificity for cancer was 87.5%. Allison and colleagues recently published results of a comparison of a sensitive gFOBT (Hemoccult SENSA) with an FIT (Hemoccult ICT) for cancer and advanced adenomas in the distal colon in nearly 6000 average-risk subjects who had undergone FSIG. Both tests showed superior sensitivity for cancer compared with the single-test performance of an unrehydrated gFOBT. The sensitivity for CRC of the FIT and the sensitive gFOBT was 81.8% and 64.3%, respectively.

Annual screening with FIT that have been shown in the published peer-reviewed literature to detect a majority of prevalent CRC in an asymptomatic population at the time of testing is an acceptable option for colorectal screening in average-risk adults aged 50 years and older. Page 1577

CSPY has several limitations. It requires one or more days of dietary preparation and bowel cleansing, usually a day dedicated to the examination, and because of sedation, a chaperone is needed for transportation. It is an invasive procedure, and surveys indicate that a significant percentage of adults prefer other noninvasive options for CRC screening. Effective performance of the procedure is dependent on thorough bowel preparation, which is often perceived as the most unpleasant part of the CSPY process by those who have undergone the test. Limitations with regard to detection of neoplasia have been previously discussed, and the fact that CSPY is operator skill dependent is another significant limitation. Patients are generally poorly informed about the problem of variable performance of the procedure and are unaware of the skill level of their endoscopists. Formal quality assurance programs do not exist, and the current reimbursement system for CSPY does not reward careful examination but tends to reward rapidly performed examinations and repeated examinations at unnecessarily short intervals.117 Polypectomy is sometimes ineffective in eradicating polyps, a factor that has been implicated as the cause of up to 25% of interval cancers.118,119 Finally, CSPY is not an infallible "gold standard." Controlled studies have shown the CSPY miss rate for large adenomas (10 mm) to be 6% to 12%.120,121 The reported CSPY miss rate for cancer is about 5% pg 1582

2 They make a lot of assumptions, including that the colonoscopic detection rate for CRC is 100%. They estimate a 97 to 99% reduction in CRC through use of colonoscopy, 80% of that from the initial procedure.

Stat Methods Med Res 2009; 18; 163

A deterministic model for estimating the reduction in colorectal cancer incidence due to endoscopic surveillance

There is evidence that the removal of adenomas, by endoscopy, from the large bowel can prevent the occurrence of colorectal cancer (CRC). However, the reduction in cancer incidence due to endoscopic surveillance is difficult to estimate. Studies of cohorts of adenoma patients typically rely on comparisons with groups of historical controls. We present a model for disease progression which enables estimation of this quantity without direct comparison to a reference group. Models are applied to data from the National Polyp Study. Rates of adenoma recurrence and progression to carcinoma are estimated based on study data and relevant literature. This allows calculation of the number of cancers expected in the absence of surveillance and, thus, the number of cancers prevented. Results are compared with the original analysis. Models estimate that surveillance reduced CRC incidence by at least 97% in this cohort. The majority of the effect was due to the initial removal of adenomas rather than the follow-up surveillance. These results are similar to those produced in the original analysis when using the most appropriate reference groups. They indicate that polypectomy and follow-up surveillance can lead to large reductions in cancer incidence which may have been under-estimated in previous studies.

3 "we must realize that current evidence is indirect and does not support a claim of 90% effectiveness." Ann Intern Med. 2009;150:50-52.

4 "and what we should probably tell our patients— might be closer to a 60% to 70% reduction of the risk for death from CRC with high-quality colonoscopy. A 60% to 70% mortality reduction is not as good as 90%, but it should not be considered disappointing. It would be remarkably high compared with screening for other types of cancer, such as breast (a 25% cancer mortality reduction at best) or prostate cancer" Ann Intern Med. 2009;150:50-52.

5 "colonoscopy with polypectomy decreases the incidence of CRC by about 75–ndash;90%; recently, a mortality decrease of about 69% (95% confidence interval, CI, 46–ndash;84%) was demonstrated, based on 48,743 person-years at risk, from the National Polyp Study cohort" Digestion 2007;76:20–ndash;25 Colonoscopy screen.pdf

6 Screening for Colorectal Cancer: U.S. Preventive Services Task Force Recommendation Statement

Ann Intern Med. 2008;149:627-637.

Evidence is adequate to estimate the harms of colonoscopy. In the United States, perforation of the colon occurs in an estimated 3.8 per 10 000 procedures (4). Serious complications—defined as deaths attributable to colonoscopy or adverse events requiring hospital admission, including perforation, major bleeding, diverticulitis, severe abdominal pain, and cardiovascular events—are significantly more common, occurring in an estimated 25 per 10 000 procedures

7 Neoplasia after neg cscope 435.pdf

Incidence of Advanced Neoplasia after Initial Colonoscopy in Japan

Objective: The National Polyp Study is used as the basis of recommendations for colonoscopic surveillance after polypectomy, establishing an interval of 3 years after removal of newly diagnosed adenomas. The aim of this retrospective cohort study was to estimate the incidence of advanced neoplasia after initial colonoscopy and compare the differences among risk groups. Methods: Patients over 40 years who were referred for initial colonoscopy at six institutes were selected. They were classified into four groups based on the initial colonoscopy: A, patients without any adenoma; B, with adenomas of ,6 mm only; C, with adenomas of 6 mm; D, with any intramucosal cancer. The index lesion (IL) at follow-up colonoscopy was defined as large adenoma 10 mm, intramucosal/invasive cancer. Results: A total of 5309 patients were enrolled in this study. Overall, median follow-up period was 5.1 years. The numbers of eligible patients in the various subgroups were A, 2006; B, 1655; C, 1123; D, 525. A total of 379 ILs were newly diagnosed during follow-up colonoscopy. The cumulative incidence of ILs in each group was A, 2.6%; B, 6.7%; C, 13.4%; and D, 12.6%. Conclusions: Patients with any adenomas .6 mm or intramucosal cancer at the initial colonoscopy have a higher risk of advanced neoplasia during follow-up colonoscopy.

Jpn J Clin Oncol 2009;39(7)435–442



Feedback Welcome

Comment Title:
Your Name:
Your Email Address:
Notify me of new comments to this item
This is a captcha-picture. It is used to prevent mass-access by robots. (see:
please type the characters you see

Medicine for People! is published by Douwe Rienstra, MD at Port Townsend, Washington.