Abstract
BACKGROUND: The current study on sleep regularity, weekend recovery sleep (WRS), and the potential risk of developing incident type 2 diabetes (T2D) is both limited and lacks comprehensiveness. METHODS: This study included 72 562 participants without diabetes who wore an accelerometer for 1 week. Sleep regularity was assessed using standard deviation (SDs) of sleep duration, onset, midpoint, wake-up timing, and sleep regularity index (SRI). Cox models evaluated associations of sleep regularity, WRS, and T2D risk. Elastic net regression identified key sleep characteristics, and a joint analysis differentiated individual and combined effects of sleep regularity indicators. Subgroup analyses explored the effects of weekday and weekly sleep duration, weekday SRI, and sleep alignment on the WRS-T2D relationship. RESULTS: During the average follow-up of 5.1 years, 1087 incident cases of T2D were documented. Irregular sleep patterns were associated with an increased risk of T2D. The elastic net regression model identified the SD of sleep onset timing and SD of sleep duration as key characteristics of sleep regularity. Joint analyses showed that irregular sleep onset timing, but not sleep duration, independently contributed to an increased risk of T2D morbidity, with hazard ratios (95% confidence intervals) of 1.30 (1.07, 1.58) and 1.18 (0.97, 1.44), respectively. The risk of T2D was comparable between participants with and without WRS, regardless of WRS duration. Subgroup analyses showed that the relationship between WRS and T2D did not significantly differ when considering adherence to the recommended weekly sleep duration, achieving adequate sleep duration on weekdays, maintaining regular sleep patterns during weekdays, and aligning sleep schedules on weekdays and weekends. CONCLUSIONS: Irregular sleep patterns were associated with an increased risk of T2D. In contrast to previous studies, WRS was not associated with a reduced risk of T2D.