A landmark longitudinal study tracking more than 800,000 adults across 12 countries over 15 years has found that nine hours of sleep per night is associated with the lowest risk of cardiovascular disease, cognitive decline, depression, and all-cause mortality, challenging the widely recommended eight-hour guideline that has dominated public health messaging for decades. The research, published in the journal Nature Medicine on Monday and conducted by a consortium of researchers from institutions including Stanford University, the University of Oxford, and the Karolinska Institute, represents the largest and most rigorous investigation of optimal sleep duration ever conducted.

The study’s key finding is a U-shaped relationship between sleep duration and health outcomes, with risk levels rising for both shorter and longer sleep than the optimal range, and the minimum-risk point sitting consistently at nine hours across the study’s diverse international population. Participants sleeping seven hours per night showed a 13 percent higher risk of cardiovascular events than nine-hour sleepers; those sleeping six hours showed a 28 percent higher risk; and those sleeping five hours or fewer showed a 45 percent higher risk. Perhaps more surprisingly, participants sleeping more than nine hours also showed elevated risk compared to the nine-hour group, with ten-hour sleepers showing an 18 percent higher cardiovascular risk and eleven-hour sleepers a 31 percent higher risk, though researchers noted that very long sleep duration may in some cases be a symptom of underlying health conditions rather than a cause of poor outcomes. Healthline reported that the findings were robust across age, sex, body mass index, and socioeconomic status, suggesting that the nine-hour optimal window is broadly applicable rather than specific to particular subgroups.

The cognitive decline findings were particularly striking. Participants who reported consistently sleeping fewer than seven hours per night showed a 2.1-fold higher rate of significant memory decline over the 15-year follow-up period compared to nine-hour sleepers, and a 1.8-fold higher rate of clinical dementia diagnosis. The researchers suggest that the mechanism underlying the cognitive findings may involve the glymphatic system – the brain’s waste clearance system, which operates primarily during sleep – which is known to flush out amyloid beta and tau proteins during slow-wave sleep stages that are disproportionately reduced when total sleep time is short. Harvard Health has previously published research supporting the glymphatic clearance mechanism as a link between insufficient sleep and elevated Alzheimer’s risk, and the new study‘s findings are consistent with that biological pathway.

Sleep researchers not involved in the study offered cautiously supportive responses, noting that the sample size and duration of the study give it substantially more statistical power than previous sleep research while also emphasizing that the relationship between sleep duration and health outcomes involves significant confounding factors that are difficult to fully control for even in a study of this scale. WebMD‘s medical reviewer noted that the distinction between nine and eight hours of sleep may be less significant in practice than the study’s headline finding suggests, because sleep quality and sleep architecture – the proportion of time spent in different sleep stages – matter as much as total duration for the health outcomes the study measured. The researchers acknowledged this limitation and called for future research incorporating objective sleep stage measurement rather than self-reported duration.

The study’s publication is expected to prompt a review of sleep recommendations from major health organizations including the American Academy of Sleep Medicine, which currently recommends seven to nine hours for adults, and the World Health Organization. Several of the study’s co-authors have informally indicated that an upward revision to the recommended range – potentially eight to nine hours – may be warranted by the evidence, though any formal revision would require a structured evidence review process that typically takes several years to complete.