IntroductionVarious modalities have been proposed for population screening in colorectal cancer, including fecal occult blood (FOB) testing, digital rectal examination, barium enema, rigid and flexible sigmoidoscopy, and colonoscopy. Of these, the only tests that have been studied sufficiently to warrant consideration are the guaiac-based FOB test, flexible sigmoidoscopy, and colonoscopy. In this report we consider the evidence relating to these three approaches and present an economic analysis of them, and, lastly, draw some general conclusions.The biases inherent in screening can only be reliably eliminated by population-based randomized trials where the diseasespecific mortality in the whole population offered screening (including refusers and those who develop interval cancers after a negative test) is compared with that in the population not offered screening. Thus, for the purposes of examining the efficacy of the different screening methods, emphasis has been placed on evidence from such trials.
FOB TestingMost of the research into FOB test screening has been carried out using a guaiac-based test that detects the peroxidase activity of hemoglobin and is thus an indirect measure of blood loss in stool [1]. Immunological tests that are specific for human hemoglobin are available [1, 2], but, to date, these have not been evaluated in a population screening randomized trial.Three randomized trials of FOB test screening have reported mortality data: one from Minnesota, USA that was essentially a volunteer study [3], and two true population-based trials, one from Funen, Denmark [4] and the other from Nottingham, England [5]. All these trials showed a significant reduction in colorectal cancer mortality. A meta-analysis, which included additional data from French and Swedish studies, has indicated that the likely magnitude of the reduction in mortality is 16 %, rising to 23 % when adjusted for compliance [6]. Furthermore, evidence from the Minnesota study, which is the longest running of the randomized trials, indicates that FOB test screening may reduce the incidence of colorectal cancer presumably because of the detection and removal of adenomatous polyps [7]. On the basis of these results, the UK government commissioned a demonstration pilot of FOB test screening to establish whether or not the introduction of a national screening program would be feasible [8]. This is now complete, and the results of the independent evaluation are favorable [9].However, the FOB test is far from the perfect screening test. The interval cancer rates in the Funen and Nottingham trials suggest that the sensitivity of FOB testing in a screening context is only in the region of 50 %, and although the specificity is high, there is still an appreciable false-positive rate such that around half of all colonoscopies done following a positive FOB test will show no neoplasia [4,5]. Thus there is a need for a more sensitive and specific test, but one which will still have a sufficient acceptability and economic viability for population sc...