Randomized trial of different screening strategies for colorectal cancer: patient response and detection rates

Abstract

Background: Although there is general consensus concerning the efficacy of colorectal cancer screening, there is a lack of agreement about which routine screening strategy should be adopted. We compared the participation and detection rates achievable through different strategies of colorectal cancer screening.

Methods: From November 1999 through June 2001 we conducted a multicenter, randomized trial in Italy among a sample of 55-64 year olds in the general population who had an average risk of colorectal cancer. People with previous colorectal cancer, adenomas, inflammatory bowel disease, a recent (< or =2 years) colorectal endoscopy or fecal occult blood test (FOBT), or two first-degree relatives with colorectal cancer were excluded. Eligible subjects were randomly assigned, within the roster of their general practitioner, to 1) biennial FOBT (delivered by mail), 2) biennial FOBT (delivered by general practitioner or a screening facility), 3) patient's choice of FOBT or "once-only" sigmoidoscopy, 4) "once-only" sigmoidoscopy, or 5) sigmoidoscopy followed by biennial FOBT. An immunologic FOBT was used. Participation and detection rates of the strategies tested were compared using multivariable logistic regression models that adjusted for age, sex, and screening center. All statistical tests were two-sided.

Results: Of 28 319 people sampled, 1637 were excluded and 26 682 were randomly assigned to a screening arm. After excluding undelivered letters (n = 427), the participation rates for groups 1, 2, 3, 4, and 5 were 30.1% (682/2266), 28.1% (1654/5893), 27.1% (970/3579), 28.1% (1026/3650), and 28.1% (3049/10 867), respectively. Of the 2858 subjects screened by FOBT, 122 (4.3%) had a positive test result, 10 (3.5 per 1000) had colorectal cancer, and 39 (1.4%) had an advanced adenoma. Among the 4466 subjects screened by sigmoidoscopy, 341 (7.6%) were referred for colonoscopy, 18 (4 per 1000) had colorectal cancer, and 229 (5.1%) harbored an advanced adenoma.

Conclusions: The participation rates were similar for sigmoidoscopy and FOBT. The detection rate for advanced neoplasia was three times higher following screening by sigmoidoscopy than by FOBT.