Abstract
Objectives: We expressed the combined effect by the ratio of daily sleep time to daily work time. The aim of this study was to discussed the predictive ability of daily sleep hours/work hours (SH/WH) ratio for diabetes risk. Methods: Cox proportional hazards regression was used to calculate the hazard ratios (HRs) of new-onset diabetes. Restricted cubic spline analyses were performed to visualize the influence trend of SH/WH ratio and diabetes risk. Results: The RCS model revealed a non-linear and L-shaped correlation between SH/WH ratio and diabetes risk. Compared with the participates with SH/WH ratio <1, those with a ratio ≥1 had a lower risk of developing diabetes. The multivariable adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) of new-onset diabetes in Q2, Q3, Q4 and Q5 groups compared with Q1 group were 0.82 (0.57, 1.19), 1.05 (0.69, 1.59), 0.57 (0.36, 0.91), 0.66 (0.42, 1.06). The Kaplan-Meier curve showed that Q4 group had lower cumulative incidence. Conclusion: Sleeping longer than working (SH/WH ratio ≥1) can reduce risk for developing diabetes. A minimal risk observed at 1.10-<1.37 (the fourth quintile) of SH/WH ratio.