Low Rate of Cancer Detection by Colonoscopy in Asymptomatic, Average-Risk Subjects with Negative Results From Fecal Immunochemical Tests

George F Longstreth,Daniel S Anderson,Daniel S Zisook,Jiaxiao M Shi,Jane C Lin
DOI: https://doi.org/10.1016/j.cgh.2020.01.029
Abstract:Background & aims: Screening colonoscopies are of uncertain benefit for persons with negative results from a fecal immunochemical test (FIT). We investigated detection of CRC by colonoscopy in asymptomatic, average-risk, FIT-negative subjects. Methods: We conducted a retrospective, population-based cohort study of 96,804 subjects with an initial negative result from a FIT at ages 50-75 years, from 2008 through 2014, who then underwent colonoscopy, using the Kaiser Permanente California databases. We identified participants diagnosed with CRC from January 1, 2008 through December 31, 2015 from a cancer registry. Subjects were followed until initial colonoscopy, health plan disenrollment, death, or December 31, 2015. We reviewed records from 400 randomly selected persons without CRC (controls) for risk features to estimate the proportion who underwent screening colonoscopy. We performed logistic regression to identify variables associated with CRC detection. Results: Of 257 subjects with a diagnosis of CRC, 102 did not have a record of CRC risk factors; 86 of these patients (84.3%) had non-advanced-stage CRC (no regional node spread/distant metastases). Of the 400 controls, 299 (74.75%; 95% CI, 70.49%-79.01%) lacked CRC risk features, enabling estimation that 72,263 (mean age, 57.5 ± 7.0 y; 54.5% female) had undergone screening colonoscopy. CRC was detected in 1.4 per 1000 persons after 1 FIT, without association with increasing FITs (P = .97). CRC was detected in 1.3 per 1000 persons in 2 y or less after the last FIT and in 4.4 per 1000 persons more than 2 y after the last FIT (P < .001). When the last FIT was 2 y earlier or less, CRC increased from 0.7 per 1000 persons age 50-59 y to 3.1 per 1000 persons older than 70 y. Age and time from the last FIT were associated with CRC, with adjusted odds ratios of 1.08 (95% CI, 1.05-1.11) and 2.76 (95% CI, 1.28-5.95), respectively. Conclusions: In asymptomatic, average-risk persons with a negative result from a FIT, CRC is infrequent within 2 y after the last FIT (especially for persons younger than 60 y), usually non-advanced, and unrelated to the number of FITs performed.
What problem does this paper attempt to address?