NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Whitlock EP, Lin J, Liles E, et al. Screening for Colorectal Cancer: An Updated Systematic Review [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2008 Oct. (Evidence Syntheses, No. 65.1.)

Cover of Screening for Colorectal Cancer: An Updated Systematic Review

Screening for Colorectal Cancer: An Updated Systematic Review [Internet].

Show details


This review's methods were based primarily on established USPSTF methods for systematic reviews.114 Appendix A includes a more detailed description of our methods.

Under the guidance of the USPSTF, we developed an analytic framework and five key questions (Figure 1), which received final approval from USPSTF liaisons. This report's scope differed from the 2002 USPSTF evidence report in several important ways:


We did not update the direct evidence on standard FOBT screening. We did, however, examine longer-term followup results from the original trials included in the 2002 report, as this evidence was foundational for the last recommendation.


We did not update evidence on CRC screening methods not recommended after the last review (e.g., digital rectal exam) or omitted from this review by the USPSTF during the scoping phase (e.g. DCBE) due to poor test-performance characteristics. A single study (n=580) from the previous 2002 evidence report found that DCBE as a surveillance method after adenomatous polypectomy (with comparison to colonoscopy as the gold standard) showed a sensitivity of only 48 percent (CI: 24, 67) for polyps larger than 10 mm. A more recent study in a high-risk screening and diagnostic-evaluation population compared DCBE to both colonoscopy and CTC. This study found similarly low sensitivity estimates for large polyps.115 Given its confirmed low sensitivity for one of the main targets of screening (lesions 10 mm or larger), DCBE as a primary CRC screening test was excluded from the review.


We did not systematically review screening-test adherence, acceptability, and feasibility. Similarly, the USPSTF judged that a thorough review of cost-effectiveness analyses was beyond the scope of our review, particularly since the USPSTF was conducting a simultaneous decision analysis. Since the separate decision analysis also examined screening intervals and ages to begin and end screening, these were not included in this systematic review.

Figure 1. Analytic framework.


Figure 1. Analytic framework. KQ1: What is the effectiveness of the following screening methods (alone or in combination) in reducing mortality from colorectal cancer? aFlexible SigmoidoscopybColonoscopycCTCdFecal screening (more...)

KQ1 examined direct evidence from RCT, cohort studies, or case-control studies, that screening programs (single or repeated application of screening tests) for colorectal cancer in average-risk adults, aged 40 years and older, reduce mortality. KQ2a examined the accuracy of colonoscopy and/or FS for CRC screening in average-risk persons in the community practice setting. KQ2b examined the accuracy of CTC and fecal screening tests, including high-sensitivity guaiac FOBT, fecal immunochemical test (FIT), and fecal DNA tests in average-risk persons. For KQ2a and 2b, test accuracy was derived from comparison with a valid reference standard (e.g., colonoscopy to all participants) or an acceptable reference standard (e.g., colonoscopy to all positive tests with adequate followup of test negatives). KQ3a examined the adverse effects of colonoscopy or sigmoidoscopy for CRC screening in the community practice setting. KQ3b examined the adverse effects of CTC and fecal screening tests for CRC screening. Summarized results of each key question (text and tables) are presented within the body of the report. Additional study details and corresponding evidence tables can be found in the appendices, along with the corresponding evidence tables.

We searched PubMed, Database of Abstracts of Reviews of Effects (DARE), the Cochrane Database of Systematic Reviews (CDSR), Institute of Medicine (IOM), National Institute for Health and Clinical Excellence (NICE), and Health Technology Assessment (HTA) databases for recent systematic reviews (1999-2006) for all key questions. We used fair- or good-quality existing research syntheses when available, supplemented with primary literature searches bridging the search windows of relevant systematic reviews and meta-analyses. We developed literature search strategies and terms for each KQ (see Appendix A, Table 1), with search dates guided by existing systematic reviews (including the 2002 UPSPTF report) and the timing of screening technology development.

We conducted five separate literature searches through January 2008 in both Medline and Cochrane Central Register of Controlled Trials (CCRCT) (detailed in Appendix A, Table 1). All abstracts were coded for inclusion/exclusion for all key questions.

For KQ1 (mortality outcomes of screening) and KQ2a (accuracy of FS and colonoscopy), we found no systematic reviews conforming to our inclusion and exclusion criteria more recent than the 2002 USPSTF review. Therefore, we searched for newly published primary literature beginning in January, 2000.

KQ2b (test performance characteristics of newer screening tests) required separate approaches for each of the three test types. For CTC, we used a good-quality systematic review published in 2005116 as a foundation, supplemented with additional studies identified through five other systematic reviews (published between 2003 and 2006)117121 and our own search of primary literature beginning in January, 2006. For FIT, we conducted our own searches beginning in 1990, when the early literature was first being published. We confirmed our search results using two technical reports published during the review.122, 123 We used a good-quality Technical Evaluation Center assessment that searched through June 2006124 as the basis for fecal DNA test literature, supplemented by two additional systematic reviews.125, 126 Using this review as our foundation, we searched for new primary literature published since January 2006.

For KQ3a and KQ3b (harms of screening tests) we found no synthesized evidence that could be used as a foundation for the current review. Therefore, we searched Medline and CCRCT beginning in January 2000 for newly published studies conducted in a community setting or, at a minimum, studies that included only asymptomatic individuals, the majority of whom are at average risk for CRC. We developed two search strategies: one comprehensive strategy that yielded many irrelevant abstracts, and one more focused strategy that produced fewer irrelevant abstracts. Although our pilot-testing of the more focused strategy suggested that it was sufficiently comprehensive, we coded all the abstracts from the broader strategy for articles published between January 2006 and January 2007.

Two investigators reviewed all 3948 abstracts identified by these searches. We evaluated 488 articles located through the searchs, the previous 2002 evidence report and outside sources against a set of inclusion/exclusion criteria for each key question, including design-specific quality criteria based on the USPSTF's methods (Appendix A, Table 3). We supplemented these methods with the National Institute for Health and Clinical Excellence (NICE) and Oxman criteria for systematic reviews.127, 128 Detailed inclusion and exclusion criteria are included in Appendix A, Detailed Methods. Two investigators reviewed articles against inclusion and exclusion criteria and critically appraised all studies fulfilling inclusion criteria. Some articles were then excluded for quality reasons. One investigator abstracted data from included studies into evidence tables. A second investigator verified the evidence tables' content.

Due to study design and limitations in reporting for two of the studies evaluating CTC test performance, we calculated point estimates for per person sensitivity and specificity and their respective confidence intervals. For additional details see Appendix A, Detailed Methods.

Because of the stringency of our inclusion criteria for studies to estimate rates of endoscopy harms in the community practice setting (KQ3a), included studies were clinically homogeneous to pool. We conducted full meta-analyses using Stata v9.2 “meta” command for KQ3a to estimate combined complication rates for serious bleeding (with colonoscopy), perforation (with colonoscopy), and any serious complications (with colonoscopy or FS). Several studies reported that their patients experienced no adverse events. Therefore, we used a random-effects logistic model to include studies without adverse events129, 130 and estimate combined complication rates. A description of our model is included in Appendix A, Detailed Methods. Exploratory meta-regressions were conducted using random-effects logistic models to examine the association of the following study-level characteristics: study design; study setting by country; and population characteristics, including age range and indication for endoscopy, with complication rate. The analyses were performed using the NLMIXED procedure in SAS v9.1.

USPSTF Involvement

The authors worked with four USPSTF liaisons at key points throughout the review process to develop and refine the analytic framework questions, to address methodological decisions on applicable evidence, and to resolve issues around scope for the final evidence synthesis. This research was funded by the Agency for Healthcare Research and Quality (AHRQ) under a contract to support the work of the USPSTF. AHRQ staff provided oversight for the project, reviewed the draft report, and assisted in external review of the draft-evidence synthesis.

PubReader format: click here to try


  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...