Sleep Optimization

Sleep duration's relationship to mortality is not linear, it is U-shaped: both too little and too much sleep are associated with higher death rates, with the lowest risk clustering around 7 hours per night across most large cohorts.

A large NIH-AARP cohort study published in the American Journal of Epidemiology found both shorter and longer self-reported sleep duration were associated with increased cardiovascular disease mortality, with the effect compounding in people who were also overweight or obese [1]. A Chilean cohort study (SantOSA) similarly found both short (6 hours or less) and long (9 hours or more) total sleep time increased all-cause mortality and cardiovascular risk relative to 6-9 hours [2]. A meta-analysis by Cappuccio and colleagues, pooling multiple cohorts, found the lowest all-cause mortality risk clustered around 7 hours of sleep per night, with risk rising on both sides of that point [3].

A large-scale analysis from the Semmelweis Study quantified the asymmetry precisely: short sleep (under 7 hours) carried a 14% increased mortality hazard (HR 1.14, 95% CI 1.10-1.18), while long sleep (9 hours or more) carried a substantially larger 34% increased mortality hazard (HR 1.34, 95% CI 1.26-1.42) [4]. That asymmetry, long sleep associated with a bigger mortality increase than short sleep, is counterintuitive relative to popular sleep-optimization messaging that treats 'more sleep' as an unambiguous good.

Who this does not work for, or where the evidence is weakest: the long-sleep-mortality association is widely interpreted as reverse causation, meaning people who are already sick sleep longer because illness increases sleep need or reduces daytime activity, rather than long sleep itself causing death. Grandner's review of age-stratified sleep-mortality data found the associations differ by age group, and the studies are almost entirely observational and rely on self-reported sleep duration, which correlates only moderately with objectively measured sleep via actigraphy or polysomnography [5]. No large RCT has randomized people to different sleep durations and tracked mortality, for practical and ethical reasons that make such a trial unlikely to ever exist, so causal claims about sleep duration and mortality are inherently indirect.

The practical target most consistently supported across these different cohorts, despite the different countries and populations studied, converges on 7 to 8 hours per night, with the risk curve rising measurably on both the short and long side of that range, and rising faster on the long side according to the largest recent hazard-ratio estimates [4].

Mechanistically, chronic short sleep is linked to elevated cortisol, impaired glucose regulation, and increased systemic inflammation, plausible pathways to cardiovascular and metabolic disease; the mechanism behind the long-sleep association is less clear and more consistent with sleep duration as a marker of underlying illness than as an independent causal driver.

Plain takeaway: sleep duration follows a U-shaped mortality curve with the lowest risk around 7 hours a night, both short and long sleep increase mortality risk, and long sleep's association is larger in magnitude but more likely confounded by underlying illness than short sleep's more plausibly causal link to metabolic and cardiovascular stress. A 2026 randomized study (ENLIGHTENme) tested 12 weeks of self-administered daytime light from lamps above 8,000 melanopic EDI in 202 adults aged 63 to 92. Starting sessions before 10:00 tied to greater daytime activity and lower rest-activity fragmentation. Daytime light exposure is now a tier 1 tested lever for sleep quality in older adults.