In 200,729 UK Biobank participants followed for a median 12.6 years, each one-point improvement on a six-component healthy sleep score (covering chronotype, duration, insomnia, snoring, daytime dozing, and ease of waking) was associated with a 3% lower stroke hazard. More strikingly, a composite metabolic signature derived from 35 plasma NMR biomarkers — dominated by lower glycoprotein acetyls (GlycA, a systemic inflammation marker), reduced VLDL particle load, and higher omega-3 polyunsaturated fatty acids including DHA and linoleic acid — carried an independent 14% stroke risk reduction from top to bottom tertile, mediating 8.6% of the sleep-stroke association.
This large prospective cohort adds mechanistic granularity to a well-established epidemiological relationship. The identification of GlycA, omega-3 fatty acids, and HDL subfractions as the dominant mediators is biologically coherent: poor sleep chronically elevates inflammatory tone and disrupts lipid clearance, both established stroke risk drivers. The 8.6% mediated fraction is modest, however, reminding us that the metabolic pathway is only one of several channels — autonomic dysfunction, blood pressure, and platelet reactivity likely account for the remainder. Observational confounding cannot be fully excluded despite the large sample and lengthy follow-up. Practically, the findings reinforce that improving multiple sleep dimensions simultaneously — not just duration — may reduce vascular risk through measurable metabolic improvements, and they flag omega-3 status and inflammatory load as potentially useful biomarkers in sleep-disturbed adults. Confirmatory intervention trials manipulating sleep quality and tracking these metabolites are now warranted.