Long-Term Colorectal Cancer Incidence After Negative Colonoscopy in the State of Utah: The Effect of Family History.

OBJECTIVES: Colonoscopy is widely recommended for colorectal cancer (CRC) screening, but evidence to guide the optimal frequency of repeat screening examination is limited. We examined the duration and magnitude of the risk of developing CRC, following a negative colonoscopy in those at average risk and those with a first-degree family history of CRC.

METHODS: A cohort of Utah residents aged 50-80 years who had a negative colonoscopy between 1 January 2001 and 31 December 2011 was identified using the Utah Population Database. Patients were followed from the time of the index colonoscopy until diagnosis of CRC, death, migration out of state, repeat colonoscopy, or end of the study period. CRC incidence after the index colonoscopy was compared with that of the state population by standardized incidence ratios (SIRs).

RESULTS: A cohort of 131,349 individuals at average risk with a negative colonoscopy was identified. Compared with the state population, a negative colonoscopy was associated with SIRs of 0.15 (95% confidence interval (CI): 0.08-0.23) at 1 year, 0.26 (95% CI: 0.19-0.32) at 2-5 years, 0.33 (95% CI: 0.22-0.43) at 5-6 years, and 0.60 (95% CI: 0.44-0.76) at 7-10 years for CRC following the index colonoscopy. In a secondary analysis involving only patients with a first-degree relative with CRC, patients had a significantly lower incidence of CRC only for the first 5 years of follow-up (SIR 0.39, 95% CI: 013-0.64). There was also a difference in the risk of proximal (SIR 0.72, 95% CI: 0.45-0.98) and distal (SIR 0.51, 95% CI: 0.30-0.72) colon cancers at 7-10 years following a negative colonoscopy.

CONCLUSIONS: The risk of developing CRC remains decreased for at least 10 years following the performance of a negative colonoscopy. However, the lower incidence of CRC in those with a family history of CRC differed in magnitude and timing being limited primarily to the first 5 years of follow-up and of lesser magnitude than that in the overall cohort.