Abstract
BACKGROUND: Interim 10-year results from the Nordic-European Initiative on Colorectal Cancer (NordICC), a randomized controlled trial (RCT) of screening colonoscopy, demonstrated a statistically significant reduction in colorectal cancer (CRC) incidence but not mortality, contrary to results from 4 flexible sigmoidoscopy RCTs. METHODS: We constructed CRC incidence and mortality Kaplan-Meier curves through 10 years to standardize comparisons across RCTs and examined CRC screen detection and stage. Novel analyses of 1 flexible sigmoidoscopy RCT (Prostate, Lung, Colorectal, and Ovarian cancer screening trial [PLCO]) assessed year-by-year mortality in screen-detected CRCs. RESULTS: At 10 years, all RCTs demonstrated statistically significant CRC incidence reductions with screening (ratio = 0.77, 95% confidence interval [CI] = 0.70 to 0.84, to ratio = 0.82, 95% CI = 0.69 to 0.97, vs control arm; P ≤ .011). Two flexible sigmoidoscopy RCTs and NordICC showed no statistically significant CRC mortality reduction (ratio = 0.84, 95% CI = 0.64 to 1.10, to ratio = 0.90, 95% CI = 0.69 to 1.18; P = .10-0.23). In 3 flexible sigmoidoscopy RCTs and NordICC, relative reductions were greater in CRC incidence than CRC mortality, but only NordICC reported higher CRC mortality with screening vs the control arm for the first 7 years. In contrast, PLCO observed fewer CRC deaths with screening by year 2 (ratio = 0.59; P = .03), and screen-detected CRCs were less often advanced (odds ratio = 0.26; P < .001) or fatal (ratio = 0.50; P < .001). CONCLUSIONS: After 10 years, NordICC is similar to 2 flexible sigmoidoscopy RCTs in observing statistically significant reductions in CRC incidence but not CRC mortality. However, only NordICC observed greater CRC mortality with screening vs the control arm for 7 years. Granular analyses of CRC cases and deaths in NordICC, paralleling our PLCO analyses, could provide insight into why CRC mortality results differ in NordICC vs flexible sigmoidoscopy RCTs.