ABSTRACT
BACKGROUND@#Prospective analyses have yet to identify a consistent relationship between sleep duration and the incidence of gastrointestinal (GI) cancers. The effect of changes in sleep duration on GI cancer incidence has scarcely been studied. Therefore, we aimed to examine the association between baseline sleep duration and annual changes in sleep duration and GI cancer risk in a large population-based cohort study.@*METHODS@#A total of 123,495 participants with baseline information and 83,511 participants with annual changes in sleep duration information were prospectively observed from 2006 to 2015 for cancer incidence. Cox proportional-hazards models were used to calculate hazard ratios (HRs) and their confidence intervals (CIs) for GI cancers according to sleep duration and annual changes in sleep duration.@*RESULTS@#In baseline sleep duration analyses, short sleep duration (≤5 h) was significantly associated with a lower risk of GI cancer in females (HR: 0.31, 95% CI: 0.10-0.90), and a linear relationship between baseline sleep duration and GI cancer was observed (P = 0.010), especially in males and in the >50-year-old group. In the annual changes in sleep duration analyses, with stable category (0 to -15 min/year) as the control group, decreased sleep duration (≤-15 min/year) was significantly associated with the development of GI cancer (HR: 1.29; 95% CI: 1.04-1.61), especially in the >50-year-old group (HR: 1.32; 95% CI: 1.01-1.71), and increased sleep duration (>0 min/year) was significantly associated with GI cancer in females (HR: 2.89; 95% CI: 1.14-7.30).@*CONCLUSIONS@#Both sleep duration and annual changes in sleep duration were associated with the incidence of GI cancer.