Longevity writing usually centers on exercise, diet and lately supplements. Sleep has stayed oddly in the background — something everyone agrees matters but few discuss with actual numbers. Over the last five years, though, several large cohort studies have started measuring sleep's relationship to aging in much more specific terms.
The gap is even more visible in Turkish-language coverage, where most results repeat generic, unsourced lines like "cells regenerate during sleep." Which study, how many people, what risk ratio — in-depth sourced content answering those questions is hard to find.
This piece looks at sleep along three separate axes: how much you sleep (duration), when you sleep and wake (regularity), and how much of that sleep is deep sleep. Each has been studied independently, and each points to a different kind of evidence.
A systematic review by Ungvari and colleagues, published in 2025 in GeroScience, pooled 79 cohort studies with at least one year of follow-up. The reference range used was 7-8 hours of sleep.
The relationship is U-shaped: both too little and too much sleep are associated with risk, with the middle band looking safest. One distinction matters here, though. Long sleep is often a consequence, not a cause — depression, chronic fatigue, untreated sleep apnea or an undiagnosed illness can make someone both sleep longer and carry a higher baseline mortality risk. The meta-analysis doesn't rule out this reverse causation; it only establishes that the association exists.
A 2023 study in eLife, led by Cribb and colleagues, tested a different question: if sleep duration is held constant, does going to bed and waking at the same time every day make a difference on its own? Using 7 days of wrist-accelerometer data from 88,975 people in the UK Biobank cohort, the authors calculated a Sleep Regularity Index (SRI) scored from 0 to 100 — 100 meaning perfectly consistent sleep-wake timing, 0 meaning entirely random.
The value of this study is that it tries to separate irregularity's effect from total sleep duration. A clean separation isn't fully possible, though: someone with irregular sleep timing usually also sleeps less, more fitfully and at lower quality overall. Shift work, social jet lag or an irregular lifestyle can disrupt both timing and general health at once. By the authors' own framing, this is an observational association — it shares the same logic as the circadian-rhythm research covered in our circadian eating article: living out of sync with your body clock carries a cost, but whether that cost should be attributed to sleep timing, meal timing, or the combination of the two isn't fully settled.
The third axis is sleep quality — specifically slow-wave sleep, the deep stage thought to be when the brain clears out waste accumulated during the day. A 2023 study in JAMA Neurology, led by Himali and colleagues using the Framingham Heart Study cohort, tested this directly.
The small sample (52 dementia cases in total) and observational design put real limits on this study. More importantly, the direction of causation isn't settled: does losing deep sleep trigger dementia, or do dementia's earliest, undiagnosed stages already disrupt sleep architecture? The researchers note both are plausible, and that the two could reinforce each other in a loop.
The evidence here is earlier and messier. A study by Carroll and colleagues, published in 2016 in Biological Psychiatry, used Women's Health Initiative data on 2,078 women (average age 64.5) and drew an interesting distinction: self-reported short or long sleep duration was not related to epigenetic age. But insomnia symptoms — trouble falling asleep, waking at night, difficulty getting back to sleep — were related to an older epigenetic age even after adjusting for covariates (B=1.02, SE=0.37, p=0.005).
In other words, what was measured wasn't "how many hours did you sleep" but "how broken and unsatisfying was your sleep." UCLA's press coverage of the study noted that postmenopausal women reporting all five insomnia symptoms looked biologically about two years older than peers reporting none.
A separate 2024 study in Clinical Epigenetics, covering 692 Korean adults (441 good sleepers, 251 poor sleepers), pointed to a similar pattern: the association between sleep quality score (PSQI) and epigenetic age acceleration was statistically significant only within the poor-sleeper subgroup; it did not hold among good sleepers. In the same study, each 1-point rise in PSQI among poor sleepers was linked to a 15% higher likelihood of metabolic syndrome (odds ratio 1.15, 95% CI 1.03-1.29).
The shared message of both studies: what tracks with epigenetic aging is probably not total sleep hours but how broken and insufficient the sleep feels. Both are cross-sectional or partly cross-sectional designs, though — no intervention trial has yet shown that improving sleep reverses epigenetic age.
Lined up together, these four studies point to a practical picture: aiming for 7-8 hours is more defensible than "the more I sleep, the better" — the highest risk in the meta-analysis showed up on the long-sleep side, not the short-sleep side. If you consistently sleep past 9 hours, asking why (sleep apnea, depression, another health issue) is probably more useful than trying to sleep even more.
The UK Biobank data suggests keeping your bed and wake times consistent, including on weekends, may carry value independent of total hours — a concrete, actionable recommendation, more specific than most generic "sleep hygiene" advice.
The epigenetic-age data suggests the real target isn't a nightly hour count but insomnia symptoms. If you consistently struggle to fall asleep, wake frequently at night, or wake early and can't get back to sleep, addressing that directly — including seeing a sleep specialist if needed — looks like a more precise target than simply trying to log more hours.
If you're over 60 and concerned about deep sleep loss, it's worth remembering this isn't something you can measure on your own — slow-wave sleep percentage is only determined in a sleep lab (polysomnography); wearable-device deep sleep estimates are not the same measurement used in these studies.
Observational data shows a strong association. A 2025 meta-analysis of 79 cohort studies found both short sleep (under 7 hours) and long sleep (9 hours or more) were linked to higher mortality. A UK Biobank study of 88,975 people found irregular sleep timing raised risk independent of total duration. None of this is a randomized trial, so no study shows that changing your sleep extends human lifespan.
The 2025 meta-analysis used 7-8 hours as its reference range; both less and more were linked to higher mortality (hazard ratio 1.14 for short sleep, 1.34 for long sleep). This is a cohort average, not a personal prescription — age, health status and individual needs vary.
A 2023 UK Biobank analysis published in eLife found the most irregular sleepers had 53% higher mortality than the median group, while the most regular sleepers had 10% lower risk. The researchers found a similar pattern for cardiovascular and cancer deaths. This is still an observational finding; irregular sleepers may carry other risk factors too.
"The sturdiest finding in sleep science isn't about sleeping more or less — it's about going to bed and waking at the same time every night. That's a longevity intervention that costs nothing and needs no supplement or device."
— Vivelong Longevity Science Team