Lead Time Bias May Contribute to the Shorter Life Expectancy in Post-colonoscopy Colorectal Cancer.

Abstract

The long-term outcomes of post-colonoscopy colorectal cancer have varied in previous studies. Our nationwide cohort analysis estimated expected years of life lost to adjust for lead time bias. We recalculated the long-term outcomes for post-colonoscopy and detected colorectal cancer. Patients with colorectal cancer registered in the Taiwan Cancer Registry between 2002 and 2009 were enrolled. The detected group included 22,169 cases of colorectal cancer confirmed within 6months after a colonoscopy. The post-colonoscopy group included 1653 cancer patients who received a colonoscopy 6-60months before diagnosis. Patients were followed up until 2011. We simulated age-, sex-, and calendar year-matched referents from life tables in the Taiwan National Vital Statistics using a Monte Carlo method. The life expectancy and expected years of life lost of the cancer patients were obtained from extrapolation of the logit transformation of the survival ratio between the cancer cohorts and the referent groups. Post-colonoscopy colorectal cancer had shorter life expectancies than detected cancer (stages 2-4: 13.6 vs. 16.1years; 8.7 vs. 12.6years; 2.1 vs. 4.1years, p < 0.001). The loss-of-life expectancy did not show this trend after adjusting for lead time bias. Post-colonoscopy colorectal cancer was found at an older age, more often proximal, and was associated with previous endoscopic polypectomy procedures (p < 0.001). Post-colonoscopy colorectal cancer leads to a shorter life expectancy, which appears partially explained by the presence of lead time bias. Quality assurance for colonoscopy and close surveillance for high risk groups would reduce post-colonoscopy colorectal cancer.