Does Short Sleep Shorten Your Lifespan?

The mortality signal is real, but not clean

The recent attention comes from a 2025 meta-analysis, a statistical summary that combines prior studies, in GeroScience that pooled 79 adult cohort studies, which follow groups of people over time. Compared with adults sleeping about 7 to 8 hours, sleeping fewer than 7 hours per night was associated with a 14% higher risk of death from any cause. Sleeping 9 or more hours was associated with a 34% higher risk.

That short-sleep estimate fits earlier meta-analyses. Gallicchio and Kalesan reported a pooled short-sleep estimate about 10% higher, Cappuccio and colleagues reported about 12% higher, and the 2025 update reported 14% higher. Dose-response reviews, which estimate how risk changes across sleep durations, also often find the lowest mortality near 7 hours. The pattern is not new. The 2025 paper is a larger update, not a discovery that proves sleep duration controls lifespan.

The main problem is causality. These are observational studies, meaning they observe people's usual sleep rather than assigning a sleep schedule, so they can show that sleep duration and mortality travel together, but not which one is driving the other. Chronic disease, pain, depression, sleep apnea, a disorder in which breathing repeatedly stops or becomes shallow during sleep, medications, caregiving stress, shift work, and frailty can all disturb sleep and also raise mortality risk. Statistical adjustment helps, but it cannot remove every hidden factor.

Long sleep makes that caveat especially visible. Its observational association with mortality is larger than the short-sleep association, yet a 2025 UK Biobank mortality analysis found observational links for both short and long sleep while its genetic analysis did not find evidence that genetically predicted sleep duration, short sleep, or long sleep caused all-cause mortality. That does not disprove an effect of sleep behavior. It means lifespan claims remain harder to make than risk-signal claims.

Genetic evidence is stronger for cardiovascular risk than lifespan

Mendelian randomization is not a randomized sleep trial. It asks whether people who inherit genetic variants linked to an exposure, such as shorter sleep, also have different disease risk. Because those variants are inherited before illness develops, this approach reduces reverse causation. It does not eliminate confounding or every genetic assumption.

In a UK Biobank Mendelian randomization analysis of 404,044 participants, genetically predicted short sleep, defined in that study as 6 hours or less, was associated with higher risk of several cardiovascular diseases, meaning heart and blood-vessel diseases. These included arterial hypertension, or high blood pressure; coronary artery disease, or plaque-related narrowing of heart arteries; myocardial infarction, or heart attack; chronic ischemic heart disease, or long-term reduced blood flow to heart muscle; and pulmonary embolism, or a clot in the lung arteries. Genetically predicted long sleep, defined as 9 hours or more, was not associated with any of the 12 cardiovascular diseases tested.

This is relevant because cardiovascular disease is one plausible route by which habitual short sleep could affect health and survival. Still, the method depends on an assumption that the sleep-related genetic variants affect disease only through sleep. This problem is called horizontal pleiotropy, meaning a genetic variant influences an outcome through more than one pathway. The ADRB1 gene is a useful caution. In one study that combined human family genetics with cell and mouse experiments, a rare ADRB1 variant was linked to naturally shorter sleep. The same gene encodes the β1-adrenergic receptor, a major receptor for adrenaline-like stress signals in heart muscle. If a sleep-related gene also affects heart biology through another pathway, the genetic result can overstate the causal role of sleep itself.

Sleep-loss experiments show fast biological strain

Because no ethical trial can assign thousands of adults to decades of short sleep and count deaths, controlled human experiments answer a narrower question: what happens when sleep is sharply reduced for a night, several nights, or a few weeks?

One randomized crossover experiment, meaning each participant completed both normal sleep and sleep deprivation in randomized order, studied 13 healthy adults aged 18 to 35 after a single night of total sleep deprivation. The next day, plasma cortisol increased by 21%, plasma testosterone decreased by 24%, and post-meal skeletal muscle protein synthesis fell by 18%. Cortisol is a stress hormone, testosterone is an anabolic hormone that supports tissue building, and muscle protein synthesis is the process by which muscle protein is built and repaired. This does not mean one sleepless night causes muscle loss, but it shows an acute shift toward a more stress-heavy, less muscle-building state.

The body-composition evidence is also clinically practical. In a randomized crossover trial of 10 overweight adults undergoing calorie restriction, participants lost about 3 kilograms over each 14-day period whether they had 8.5 or 5.5 hours of sleep opportunity. But with 5.5 hours, fat loss fell from 1.4 kilograms to 0.6 kilograms, while fat-free mass loss rose from 1.5 kilograms to 2.4 kilograms. Fat-free mass includes lean tissue, water, bone, and organs, so this was not a direct measurement of muscle loss alone. In plain language, the scale moved either way, but short sleep shifted more of the measured loss away from fat and toward non-fat mass.

Other controlled human studies add a metabolic signal. In 15 healthy young adults, 3 days of moderate sleep restriction reduced insulin sensitivity, which means the body needed more insulin to handle a glucose challenge. Results differ by protocol, but the direction is consistent with sleep loss being biologically active rather than just a next-day fatigue issue.

Inflammation is not an instant alarm bell

The experimental evidence is not uniformly dramatic. An updated meta-analysis first published online in 2025 included 35 human sleep-deprivation studies with 887 participants and found that a single night of total or partial sleep deprivation did not significantly raise peripheral inflammatory markers, meaning inflammatory proteins measured in blood or other body samples. Multiple nights of partial sleep deprivation, with sleep reduced to about 4.3 hours per night for at least 3 nights, did raise interleukin-6 and C-reactive protein, two blood markers of inflammation.

Recovery also matters. Some sleep-loss effects appear small, absent, or reversible, especially after returning to a normal sleep pattern. Other findings suggest recovery may be incomplete for certain markers after repeated restriction. In specific experiments, C-reactive protein and interleukin-17, another immune-signaling protein, remained elevated after recovery sleep, while another randomized crossover study found sleep-time blood pressure remained elevated during recovery and endothelial function, which is how well the blood-vessel lining responds, was attenuated during restriction. The safest interpretation is neither panic nor dismissal: occasional poor sleep is not the same exposure as repeated severe restriction, but repeated restriction is biologically active.

What to do with this evidence

For adults, major sleep-health guidance remains consistent with the evidence: regular sleep of 7 or more hours per night is a reasonable prevention target. The mortality data do not let you calculate a personal lifespan penalty from sleeping 6 hours, and they do not prove that forcing more time in bed will extend life.

The more practical point is pattern recognition. If you are habitually under 7 hours, especially while dieting, training hard, managing blood pressure, or trying to improve metabolic health, sleep is not a soft lifestyle detail. It is a measurable behavior tied to body composition, insulin sensitivity, cardiovascular intermediates, stress hormones, and inflammation in human research.

If you regularly sleep 9 or more hours and still feel unrefreshed, the evidence does not say long sleep itself is the enemy. It suggests long sleep can be a health signal worth interpreting in context: sleep quality, sleep apnea symptoms such as loud snoring or witnessed breathing pauses, depression, medication effects, chronic illness, and recovery demands matter. The longevity-relevant move is not to chase a perfect number. It is to identify a persistent sleep pattern, ask why it exists, and track whether correcting the drivers improves the health markers you care about.