Screening colonoscopy and risk for incident late-stage colorectal cancer diagnosis in average-risk adults: a nested case-control study

Abstract

Background: The effectiveness of screening colonoscopy in average-risk adults is uncertain, particularly for right colon cancer.

Objective: To examine the association between screening colonoscopy and risk for incident late-stage colorectal cancer (CRC).

Design: Nested case-control study.

Setting: Four U.S. health plans.

Patients: 1039 average-risk adults enrolled for at least 5 years in one of the health plans. Case patients were aged 55 to 85 years on their diagnosis date (reference date) of stage IIB or higher (late-stage) CRC during 2006 to 2008. One or 2 control patients were selected for each case patient, matched on birth year, sex, health plan, and prior enrollment duration.

Measurements: Receipt of CRC screening 3 months to 10 years before the reference date, ascertained through medical record audits. Case patients and control patients were compared on receipt of screening colonoscopy or sigmoidoscopy by using conditional logistic regression that accounted for health history, socioeconomic status, and other screening exposures.

Results: In analyses restricted to 471 eligible case patients and their 509 matched control patients, 13 case patients (2.8%) and 46 control patients (9.0%) had undergone screening colonoscopy, which corresponded to an adjusted odds ratio (AOR) of 0.29 (95% CI, 0.15 to 0.58) for any late-stage CRC, 0.36 (CI, 0.16 to 0.80) for right colon cancer, and 0.26 (CI, 0.06 to 1.11; P = 0.069) for left colon/rectum cancer. Ninety-two case patients (19.5%) and 173 control patients (34.0%) had screening sigmoidoscopy, corresponding to an AOR of 0.50 (CI, 0.36 to 0.70) overall, 0.79 (CI, 0.51 to 1.23) for right colon late-stage cancer, and 0.26 (CI, 0.14 to 0.48) for left colon cancer.

Limitation: The small number of screening colonoscopies affected the precision of the estimates.

Conclusion: Screening with colonoscopy in average-risk persons was associated with reduced risk for diagnosis of incident late-stage CRC, including right-sided colon cancer. For sigmoidoscopy, this association was seen for left CRC, but the association for right colon late-stage cancer was not statistically significant.

Figure 1

Selection of Cases and Control for the Study, 2006–2008 *Strong family history refers to colorectal cancer diagnosed in one or more first-degree relatives before age 50 or 2 or more relatives of any age, or other familial syndromes. ^†Controls were matched on the reference date to cases according to study site on the calendar year of birth, gender, and length of continuous health plan enrollment (5–10 years) prior to the reference date. ^‡Cancer in segments proximal to the splenic flexure were classified as right colon cancer. ^§Left colon/rectum refer to the splenic flexure, descending colon, sigmoid and rectum.

Figure 2

Sensitivity of the odds ratio estimates for screening colonoscopy on any late-stage colorectal cancer to varying the assumed preclinical period* for excluding tests Note: Odds ratios and 95% confidence intervals (CI) were obtained with conditional logistic regression with screening defined as exposure to a ‘definite or probable’ screening test. Analyses were performed on matched case-controls sets using conditional logistic regression which adjusted for receipt of other screening tests (sigmoidoscopy, barium enema and fecal occult blood test), census block-group poverty levels (as a continuous variable), number of preventive health care visits, family history of colorectal cancer, and comorbidity index at baseline. Missing values of poverty level were imputed using predictive mean matching. *The preclinical period refers to the time from the assumed onset of preclinical late-stage disease to the date of its clinical diagnosis, which we varied from 1 to 15 months.