Sleep duration: a neglected factor in stroke prevention among individuals with dysglycemia

Stroke imposes a staggering and escalating disease burden in China. Individuals with dysglycemia face a markedly elevated risk of incident cerebrovascular events. This study aimed to characterize global and Chinese trends in stroke burden and examine the association between sleep duration and stroke among individuals with dysglycemia, with a particular focus on the potential utility of sleep duration as an behavioral marker for cerebrovascular risk assessment.

This study utilized the Global Burden of Disease 2023 database to assess the trends of stroke burden globally and in China(1990-2023). In addition, the study was conducted using data from the China Health and Retirement Longitudinal Study (CHARLS), encompassing participants aged ≥ 45 with baseline dysglycemia and no prior stroke. Sleep duration was obtained from the 2020 survey, while incident stroke status was ascertained across the 2013, 2015, 2018, and 2020 follow-up waves. Stepwise logistic regression models were employed to evaluate the independent association between baseline sleep duration and new-onset stroke. Restricted cubic splines (RCS) with four knots were applied to explore non-linear dose-response relationships. Additionally, subgroup analyses were conducted to identify effect modifications, and receiver operating characteristic curves evaluated the discriminatory performance of the prediction model including sleep factors.

Despite declining global age-standardized rate, the absolute stroke burden in China significantly increased from 1990 to 2023, particularly among males and older adults. In the CHARLS cohort, 7,274 patients with dysglycemia were included in 2011. After a ten-year follow-up, 694 participants (9.54%) developed new strokes. Notably, the sleep duration was significantly longer (9.95 ± 1.74 h hours, P < 0.0001) in the new stroke group. The model after including covariates showed that for every additional hour of sleep, the risk of stroke increased by 22% (OR = 1.22, 95% confidence interval (CI): 1.13-1.32). Crucially, RCS analysis uncovered a non-linear inflection point at 7.5-8 h, demonstrating that cerebrovascular risk accelerated rapidly beyond this threshold. The fully adjusted model showed the highest discriminatory performance among the evaluated models, although the overall discrimination remained modest (AUC = 0.666).

China's absolute stroke burden continued to surge. Sleep duration may serve as an easily obtainable behavioral marker that complements conventional factors in stroke risk assessment among individuals with dysglycemia.A non-linear association was observed, with stroke risk increasing more rapidly beyond approximately 7.5-8 h of sleep. These findings suggest that sleep duration may provide complementary information for identifying metabolically vulnerable individuals at elevated cerebrovascular risk, although further prospective validation is required. By synergizing human epidemiological expertise with data-driven modeling, this study may provide a feasible approach to complement stroke risk stratification in metabolically vulnerable populations.