Colorectal cancer screening: keeping the options open.

The quantity and quality of evidence supporting the effectiveness of colorectal cancer screening now rivals that of Pap smears and mammography screening. Table 1 synthesizes what we now know. Three large randomized trials were highly consistent in showing mortality reductions with screening fecal occult blood test (FOBT) and colonoscopic follow-up of positive results.1–3 One small trial,4 several case–control studies,5–8 and one cohort study9 were consistent in showing both incidence and mortality reductions with sigmoidoscopy screening. Evidence for the effectiveness of the two other approaches, colonoscopy and double-contrast barium enema (DCBE), is less direct and rests primarily on the studies of FOBT and sigmoidoscopy. Given the relative length of the instruments, it is not surprising that colonoscopy is more sensitive than sigmoidoscopy.10, 11 It seems highly plausible to extrapolate the proven benefits of sigmoidoscopy to the entire colon for colonoscopy, particularly when results of the three FOBT trials are considered (studies designed on the assumption that colonoscopic screening could lower colorectal cancer mortality rates). The goal of the FOBT is simply to make colonoscopic screening more efficient by identifying those most likely to benefit. The evidence for DCBE is limited to descriptive studies showing that it has a relatively high sensitivity (50% to 94%) when compared with endoscopy for cancer and larger adenomatous polyps.12, 13 Compared with FOBT, DCBE’s sensitivity is much higher for adenomas and at least as high for cancer. It therefore seems plausible that screening with DCBE should also provide benefit when positive results are followed up with endoscopic polypectomy or surgery. In this issue of ecp, Erban and colleagues14 show that, despite this evidence, levels of colorectal cancer screening remain well below those for Pap smear and mammography. In this 1998 survey of Massachusetts residents, only 51% of adults aged 50 years and older reported being current with any form of colorectal screening. The authors’ approach, of reporting on combined exposure to any of the several choices for screening, is highly useful. It is consistent with current guidelines, as cited by the authors, which recommend that patients and physicians select from a menu of effective screening options; it also provides the best assessment of the coverage of a population by an effective screening method. The authors point out that having a variety of approaches, which differ in recommended screening intervals, complicates measurement of screening performance by providers. This may be unfortunate, in that systematic measurement and reporting has proven to be an effective means of raising mediocre performance levels for other prevention measures. However, deficits in available evidence and even greater deficits in the widespread availability of screening continue to preclude designation of a single, “best,” national approach to colorectal cancer screening. With respect to the evidence, uncertainty about the relative benefits and risks of the available approaches is an important reason to preserve the menu of screening options. Flexible sigmoidoscopy and FOBT screening appear to give roughly equivalent mortality reductions. For sigmoidoscopy, 60% to 90% of screened patients benefit, but only for the 50% to 60% of tumors that lie within reach of the sigmoidoscope. For FOBT, clinical trials find a 15% to 33% mortality reduction for patients with