Abstract
BACKGROUND: According to animal and human epidemiologic studies, exposure to outdoor light at night (LAN) may cause circadian disruption, which may disturb sleep quality and lead to incident type 2 diabetes mellitus (T2DM).</p>
METHODS: We followed 283,374 persons from 2006 through 2020. Outdoor LAN exposure was estimated using satellite data for individual address with 500 m2 scale buffer during follow-up. Incidence of T2DM was confirmed by hospital inpatient records. We identified potential confounders by a directed acyclic graph, including demographic, genetic, individual and regional level socioeconomic status, and environmental risk factors, and calculated hazard ratios (HRs) and 95% confidence intervals (CIs) through time-varying Cox proportional hazard model. Furthermore, we examined the association of outdoor LAN with a defined health sleep scores and moderation of genetic predisposition and shift work on the relationship of outdoor LAN and incident T2DM.</p>
RESULTS: We identified 7,775 incident T2DM cases over 3,027,505 person-years. Higher outdoor LAN exposures were significantly associated with higher risk of T2DM. The estimated HR for incident T2DM with an interquartile range (IQR: 11.22 nW/cm2/sr) increase in outdoor LAN was 1.05 (95%CI: 1.01, 1.09) in the fully adjusted model. Participants who lived in the highest quarter of outdoor LAN area were more likely to develop T2DM (HR: 1.14,95%CI: 1.02, 1.27). Besides, those who were exposed to higher levels of outdoor LAN had poorer sleep quality. No moderation role of PRS on outdoor LAN-induced T2DM observed both on the multiplicated and additive scale. The hazards of outdoor LAN were observed in those who never owned a night shift work.</p>
CONCLUSION: Although further work is required to clarify potential mechanisms, our findings indicate that exposure to residential outdoor LAN may contribute to T2DM risk and low sleep quality.</p>