Long-Term Efficacy of Sigmoidoscopy in the Reduction of Colorectal Cancer Incidence

Polly A. Newcomb 0 1 Barry E. Storer 0 1 Libby M. Morimoto 0 1 Allyson Templeton 0 1 John D. Potter 0 1 0 Journal of the National Cancer Institute , Vol. 95, No. 8, April 16, 2003 1 Journal of the National Cancer Institute , Vol. 95, No. 8, Oxford University Press 2003, all rights reserved - Screening sigmoidoscopy is associated with a reduction in both the incidence and mortality of colorectal cancer. Although current guidelines recommend sigmoidoscopy screening every 5 years, the duration of risk reduction is not known. We conducted a population-based casecontrol study to examine the association between sigmoidoscopy screening and colorectal cancer incidence. We collected information on screening history and risk factors from case patients with distal (n = 1026) and proximal (n = 642) colorectal cancer and from 1294 control subjects from October 1998 through February 2002. Screening sigmoidoscopy was associated with a statistically significant reduction in the incidence of distal colorectal cancer (odds ratio [OR] = 0.24, 95% confidence interval [CI] = 0.17 to 0.33). These reductions were sustained for up to 16 years with little attenuation. We also observed strong inverse associations between cancer incidence and sigmoidoscopy in analyses that included subjects with symptom-related tests. Current recommendations regarding the frequency of sigmoidoscopy screening may be unnecessarily aggressive. [J Natl Cancer Inst 2003;95:6225] Sigmoidoscopy screening for colorectal cancer has been shown to be efficacious in reducing the mortality (13) and probably the incidence (310) of this common disease. Risk reductions for distal disease appear to be as much as 60%80% for mortality and as much as 50%70% for incidence. Although the optimum sigmoidoscopy screening interval for individuals at average risk of colorectal cancer is not known, current guidelines recommend a 5-year screening interval (1113). However, such a period may be overly aggressive, given that the duration of the progression from adenoma to carcinoma may be as long as 15 years (14,15). Indeed, some have advocated once-in-a-lifetime sigmoidoscopy screening (16,17). Here we evaluate the efficacy of sigmoidoscopy in relation to screening interval in a population-based casecontrol study of colorectal cancer. We used an institutionally approved protocol to identify eligible case patients, which included all male and female residents of King, Snohomish, and Pierce counties (WA) who were newly diagnosed with invasive colorectal adenocarcinoma [International Classification of Diseases for Oncology codes C18.0, C18.2.9, and C20.0.9 (18)] from October 1998 through February 2002, as identified through the Puget Sound Surveillance, Epidemiology, and End Results (SEER) Program1 registry, and who were aged 2074 years at diagnosis. SEER reports include information on cancer stage and grade, the patients first course of treatment, and demographics. All eligible subjects had a publicly available telephone number. Of the 2185 eligible case patients identified, 131 (6%) were deceased, 66 (3%) had physicians who refused permission to contact them, 22 (1%) could not be located, and 240 (11%) refused to participate, resulting in a final sample size of 1726 case patients (overall response rate of 79%). Community-based control subjects were randomly selected according to the age and sex distribution of the case patients by using Washington State drivers license data for individuals aged 2064 years and Health Care Financing Administration files for individuals older than 64 years. Of the 1891 potential control subjects identified, 38 (2%) had died, 19 (1%) could not be located, and 510 (27%) refused to participate. The final study sample included 1324 control subjects (overall response rate of 70%). We used a structured 50-minute telephone interview to obtain information from the study subjects on known or suspected risk factors for colorectal cancer, including their screening histories prior to 1 year before diagnosis (for case patients) or before interview date (for control subjects). Information on screening tests (fecal occult blood test, sigmoidoscopy, and colonoscopy) included the date of (or subjects age at) first and last tests, number of tests, and the reason for the test; we also collected information about the subjects demographics, personal medical history, family history of cancer, medication use, and lifestyle factors such as level of physical activity, occupation, alcohol consumption, and diet. Subjects were classified as having undergone colorectal cancer screening (i.e., screening-only sigmoidoscopy) if they had sigmoidoscopy without having had prior symptoms, regardless of their family history of colorectal cancer. We considered the associations between screening-only sigmoidoscopy and colorectal cancer incidence and between any sigmoidoscopy (including symptom-related) and colorectal cancer incidence. To eliminate the possibility of bias that might arise from the selection of individuals who were at reduced risk of colorectal cancer because they had had a previous screen that was negative (19), subjects who had undergone more than one test were excluded from the analysis of the association between single-screenonly sigmoidoscopy and colorectal cancer incidence. It is possible, however, that some individuals who had multiple tests may have been at higher than average risk of disease due to the fact that they had frequent sigmoidoscopies because they were previously diagnosed with polyps, which are a precursor of colorectal cancer (20). Our analyses included only those tests performed more than 1 year prior to diagnosis or interview date, to avoid the clustering of screening tests that may have occurred shortly before diagnosis (21). Odds ratios (ORs) and 95% confidence intervals (CIs) for the association between screening and colorectal cancer incidence were estimated from a logistic regression model. Covariates were age (in 5-year intervals), sex, family history of colorectal cancer, postmenopausal hormone use (females), level of education, smoking history, body mass index (BMI), and the number of previous tests (for individuals who had more than one sigmoidoscopy). We excluded case patients with missing information about the affected subsite within the colon (n 10). We also excluded subjects with incomplete information on screening (case patients, n 48; control subjects, n 30). All statistical tests were two-sided. The mean age was 60.6 years (range 2075 years) for case patients and 62.0 years (range 2075 years) for control subjects. Overall, case patients were more likely than control subjects to report having a family history of colorectal cancer (26% versus 15%), to have a higher BMI (mean BMI, 27.8 kg/m2 versus 26.7 kg/m2), and to be current or former smokers (62% versus 57%). Among the women in our study, case patients were less likely than control subjects to have used postmenopausal hormones (45% v (...truncated)