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


Diagnosis

Mass screening for colorectal cancer is not recommended

ACP J Club. 1994 Sept-Oct;121:48. doi:10.7326/ACPJC-1994-121-2-048


Source Citation

Solomon MJ, McLeod RS, Canadian Task Force on the Periodic Health Examination. Periodic health examination, 1994 update: 2. Screening strategies for colorectal cancer. Can Med Assoc J. 1994 Jun 15;150:1961-70.


Abstract

Objective

To make recommendations for the screening of colorectal cancer in asymptomatic patients > 40 years of age.

Data Sources

English-language articles assessing screening for colorectal cancer were identified by MEDLINE (1966 through June 1993) using the keywords screening and colorectal neoplasia; articles from the bibliographies of review articles were retrieved; and content experts were contacted.

Study Selection

Studies were selected if they concerned Hemoccult testing as the first step in a multiphase secondary prevention strategy or sigmoidoscopy and colonoscopy in a single-phase secondary prevention strategy, and both asymptomatic and high-risk patients were included. Articles were classified according to task force levels of evidence.

Data Extraction

4 randomized controlled trials (RCTs) and 1 controlled trial in which patients were allocated by calendar period assessed Hemoccult testing and provided death rates and sensitivity and specificity. 1 case-control study and 2 case series examined screening with sigmoidoscopy, and 3 case series examined screening with colonoscopy.

Main Results

Among trials of multiphase screening with the Hemoccult test, 1 study reported a reduction in deaths in patients screened annually compared with control patients (82 deaths vs 121 deaths), but no difference was observed between patients screened biennially and control patients. In the 5 trials, sensitivity ranged from 48% to 74% and specificity ranged from 90% to 99%. The studies assessing sigmoidoscopy suggested a reduction in the risk for death from colorectal cancer, but the study designs were subject to volunteer, lead-time, and length-time bias. No data on the accuracy of the test were given. The 3 case series describing colonoscopy had no survival or accuracy data; the test was reported as effective in detecting adenomas and carcinomas but had poor compliance and high cost.

Conclusions

A small benefit from annual screening for colorectal cancer with the Hemoccult test was found in patients > 40 years of age; encouraging results were shown with sigmoidoscopy, but the evidence was subject to study design biases; insufficient evidence exists to support screening with colonoscopy. Hemoccult testing and sigmoidoscopy were neither included nor excluded from the periodic health examination and colonoscopy was not recommended.

Sources of funding: Health Services and Promotion Branch, Health Canada, and the National Health Research and Development Program.

For article reprint: Health Services Directorate, Health Services and Promotion Branch, Health Canada, Tunney's Pasture, Ottawa, Ontario K1A 1B4.


Commentary

It is in the nature of governmental assessments of clinical practice guidelines to be conservative and, like their American counterparts (1), the Canadian Task Force on the Periodic Health Examination concluded that the evidence for colorectal cancer screening was still insufficient to recommend it. The Task Force's method of evaluation essentially demands an RCT in order to conclude otherwise. Most clinical decisions, however, must be made in settings where RCTs have not been completed and, in these settings, the clinician must rely on observational evidence. The strength of the effect of the interventions, their consistency, and their biological plausibility in these less rigorous studies may often have sufficient weight to recommend an intervention, at least until an RCT is completed.

The conservatism of the Task Force is especially noticeable in its recommendation regarding sigmoidoscopy. Although it notes the 70% to 80% reduction in mortality observed in a recent case-control study (2), the Task Force does not note that the reduction in mortality was limited to distal malignancies, as is expected, with no effect on the proximal colon, and that a second case-control study duplicated these findings (3). In conjunction with earlier evidence regarding the efficacy of sigmoidoscopy, the observational evidence seems compelling. The Task Force is correct in desiring an RCT but, until then, the American Cancer Society seems to be on target (4).

Alfred I. Neugut, MD, PhD
Columbia University New York, New York, USA