Metabolic signatures of sleep health and stroke risk: A prospective cohort study with mediation analysis

Healthy sleep is associated with lower stroke risk, but the underlying metabolic mechanisms remain poorly understood. We investigated whether a nuclear magnetic resonance-derived metabolic signature mediates this association.

This prospective cohort study included 200,729 UK Biobank participants (mean age 56.48±8.10 years, 53.80% female) free of stroke at baseline. A healthy sleep score (0-6) was derived from 6 behaviors as follows: chronotype, sleep duration, insomnia, snoring, daytime dozing, and ease of getting up. Sleep-related metabolites among 249 plasma nuclear magnetic resonance biomarkers were selected using elastic net regression and combined into a metabolic signature score. Associations with incident stroke were estimated using Cox models, and mediation was assessed using counterfactual analyses over a median 12.6 -year follow-up.

During 2,530,392 person-years, 5140 incident strokes occurred. Elastic net identified 35 sleep-related metabolites, mainly reflecting lower glycoprotein acetyls and VLDL particles and higher polyunsaturated fatty acids, including DHA and linoleic acid. Each 1-point higher healthy sleep score was associated with lower risk of stroke (HR 0.97; 95% CI, 0.95-0.99). Each 1-SD higher metabolic signature score was associated with a reduced risk of stroke (HR 0.93; 95% CI, 0.90-0.96), and the highest versus lowest tertile had lower risk (HR 0.86; 95% CI, 0.79-0.93). The signature mediated 8.63% (95% CI, 3.65%-18.25%) of the sleep-stroke association, with omega-3 fatty acids, HDL subfractions, amino acids, and GlycA as major contributors.

A nuclear magnetic resonance-derived metabolic signature partially mediates the protective effect of healthy sleep on stroke risk through lipid metabolism, amino acid homeostasis, and systemic inflammation pathways, highlighting potential metabolic targets for stroke prevention in individuals with sleep disturbances.