Incidence of Colorectal Cancer after a Negative Index Colonoscopy: ACG/AstraZeneca Fellow Award

Abstract

Purpose: Colonoscopy has become the preferred colorectal cancer (CRC) screening option in the United States. U.S. guidelines recommend repeating colonoscopy in 10 years after a negative index colonoscopy. However, there are no randomized controlled trials on the efficacy of colonoscopy, and estimates from observational studies on the effect of colonoscopy on CRC incidence vary. In this study we estimated the pooled annual incidence of CRC after a negative index colonoscopy through a systematic review and meta-analysis of observational studies. We also evaluated the factors associated with the variation in estimates of the effect of colonoscopy. Methods: We conducted a systematic search of MEDLINE©, EMBASE, and Web of Science from inception to March 1, 2013, to identify cohort studies that reported the incidence of CRC in average-risk patients after a negative colonoscopy (i.e., no polyps or CRC). We abstracted data on number of interval CRC, number of patients at-risk, and mean follow-up. We calculated pooled incidence rate of CRC (per 1,000 patient years [PY]) and risk ratio based on studies that reported both the observed and expected rates of CRC in their patient cohort and the general population from which their patients were obtained. Results: We identified 15 independent cohort studies (mean age ± SD: 60.5±6 years; range of mean follow-up in individual studies, 1.6-11.2 years). The pooled incidence of CRC after a negative index colonoscopy was 0.5/1,000 PY (95% CI 0.1/1000 to 0.86/1000 PY), equivalent to a 0.05% annual risk. Sub-group analysis revealed stable incidence rates across population- and hospital-based studies (Table 1). The estimated 5- and 10-year CRC risk of 0.25% and 0.5%, respectively, after a negative index colonoscopy compared favorably to the observed 0.6% and 1.5%, 5- and 10-year risk in an average-risk, CRC-free, 60-year-old adult based on 2000-2002 surveillance, epidemiology, and end results (SEER) estimates. On pooled analysis of four studies that reported a standardized incidence ratio, a negative colonoscopy was associated with a 57% lower risk of CRC incidence (OR 0.43; 95% CI 0.22-0.86), similar to the inferred estimates.Table: Table. Subgroup Analyses on the Incidence of Interval CRCConclusion: Based on our meta-analysis, the pooled annual incidence of CRC after a negative index colonoscopy is 0.05%, which is significantly lower than the observed CRC risk in the general population without colonoscopy. This supports the effectiveness of colonoscopy as a screening tool, and its use in CRC prevention programs to decrease the incidence of CRC.