Sleep and longevity: how many hours you actually need (and what happens underneath)

Apparently we spend about a third of our lives asleep, which sounds like a waste until you find out what’s actually happening during those hours. Your brain is power-washing itself. Your heart rate is drifting downward in patterns that turn out to predict how long you’ll live. Growth hormone is quietly rebuilding tissue. Memories — the important ones — are being sorted into long-term storage while the emotional charge gets gently skimmed off the top. It is, genuinely, the busiest your body gets without you.

And yet sleep is the one longevity intervention that gets skipped first. We optimise VO2 Max, we argue about protein targets, we track HRV down to the millisecond — and then we sit up until 1 AM reading about biohacks on our phones. Bit ironic, that.

Here’s what the research actually says about sleep and ageing, what’s happening underneath while you’re asleep, and how much sleep most adults actually need before the body starts keeping score.

What’s going on while you’re not using it

Bit nerdy, but stay with me. Sleep isn’t one thing — it’s a coordinated cycle of very different brain and body states, each doing different work.

Slow-wave sleep (deep sleep) is where your body does most of its physical maintenance. Growth hormone pulses are largest during slow-wave sleep, skeletal muscle repair ramps up, and cerebrospinal fluid starts flushing waste products out of the brain through a plumbing system called the glymphatic system. Xie and colleagues published a startling paper in Science in 2013 showing that the interstitial space in a sleeping mouse brain expanded by about 60% compared to wakefulness, and clearance of toxic proteins — including beta-amyloid, the one implicated in Alzheimer’s — roughly doubled. The brain, essentially, takes itself offline to clean up. You cannot clean during normal business hours.

REM sleep is where emotional processing, memory consolidation, and learning happen. Your brain is almost as active as when you’re awake, but your skeletal muscles are paralysed so you don’t act out your dreams. (This is a useful design decision.) REM is where new skills get bedded in, language gets stitched together, and the emotional sharpness of yesterday’s argument quietly gets filed into “things that happened” rather than “things still happening.”

Light sleep connects the two and makes up roughly half of a night. It isn’t wasted time — it’s where memory encoding continues and where your body transitions between the deeper states.

You cycle through all of these every 90 minutes or so, with more slow-wave sleep in the first half of the night and more REM in the second. Which is why cutting your sleep short in the morning disproportionately costs you REM, and going to bed late disproportionately costs you deep sleep. The hours aren’t interchangeable.

The mortality research — and why “just sleep more” is the wrong takeaway

The question everyone actually cares about: does sleep duration predict how long you’ll live? Turns out, yes, but not in a straight line.

The landmark paper here is Cappuccio and colleagues’ 2010 meta-analysis in the journal Sleep, which pooled 16 prospective studies covering 1,382,999 adults across multiple countries. They found a U-shaped relationship between self-reported sleep duration and all-cause mortality. Short sleepers (less than 7 hours a night) had a 12% increased risk of dying over the follow-up period. Long sleepers (more than 8 or 9 hours, depending on the study) had a 30% increased risk. The sweet spot — lowest mortality — sat around 7 to 8 hours for most adults.

That “long sleep is worse” finding is the one that confuses people, and it’s worth being careful with. Sleeping 10 hours a night does not directly kill you. But needing 10 hours to feel rested tends to be a marker of something else — depression, undiagnosed sleep apnoea, chronic illness, or disrupted sleep quality that forces your body to spend more time in bed to extract the same amount of restorative sleep. The duration is the symptom; the underlying issue is the risk. Meta-analyses that adjust heavily for health conditions shrink the long-sleep effect, though some residual association tends to persist.

The short-sleep effect is more direct. A 2017 meta-analysis by Itani and colleagues in Sleep Medicine combined 153 studies and found short sleep was associated with increased risk of diabetes, obesity, hypertension, cardiovascular disease, and all-cause mortality across every category they examined. Effect sizes varied, but the direction was consistent. Habitually under 6 hours is where the curve starts bending sharply upward.

The practical takeaway from all this: the target is 7 to 8 hours of actual sleep for most adults, with some individual variation. Not 6. Not 10. Not “I’ll sleep when I’m dead” — that one turns out to be self-fulfilling.

What happens to the rest of your body on too little sleep

The reason sleep duration predicts mortality so reliably is that short sleep messes with almost every system your longevity depends on. A short, honest tour:

Cardiovascular. Sleep is when your blood pressure is supposed to dip — healthy adults show a 10-20% drop in systolic BP overnight. Chronic short sleep blunts that dip, which is associated with increased risk of stroke and heart disease. Your HRV is highest during slow-wave sleep and lowest on nights you’ve had five hours and a glass of red wine. Sleep is where recovery actually happens.

Metabolic. A single week of five-hour nights is enough to meaningfully reduce insulin sensitivity in healthy young adults. Eve Van Cauter’s group at the University of Chicago has been publishing on this since the late 1990s, and the findings are robust: short sleep pushes healthy bodies toward a pre-diabetic metabolic profile within days. It usually recovers once you sleep properly again, but the cumulative effect of years of this is not kind.

Hormonal. Leproult and Van Cauter published a study in JAMA in 2011 showing that restricting healthy young men to 5 hours of sleep a night for a week dropped their daytime testosterone by 10-15% — roughly the effect of ageing 10-15 years in a fortnight. Cortisol, the stress hormone, runs higher for longer when you’re underslept. Growth hormone drops. The overall pattern shifts from “build and repair” toward “conserve and worry.”

Immune. Prather and colleagues at Carnegie Mellon ran a beautifully concrete 2015 study in Sleep: they tracked people’s habitual sleep for a week, then dripped rhinovirus into their noses in a controlled chamber. People sleeping less than six hours a night were more than four times as likely to develop a cold as those sleeping more than seven. Four times. Irwin’s 2016 meta-analysis in Biological Psychiatry found that both sleep deprivation and chronic sleep disturbance were associated with elevated inflammatory markers — IL-6 and CRP — both of which are themselves linked with accelerated biological ageing.

Cognitive and brain health. Pase and colleagues analysed data from the Framingham Heart Study in a 2017 Neurology paper and found that adults with lower REM sleep proportions had a substantially increased risk of developing dementia over the following 12 years. Every 1% reduction in REM was associated with a roughly 9% increase in dementia risk. This is quite a grim sentence to read at night, I’ll admit, but the mechanistic story is at least coherent: REM sleep is deeply involved in the emotional and memory processing that makes a cognitively resilient brain.

Any one of these would be reason enough to take sleep seriously. All of them together start to explain why it shows up so reliably in mortality data.

Sleep changes with age — and not in the direction you’d hope

One of the quietly frustrating things about sleep is that your ability to sleep deeply declines faster than your need for sleep does. Van Cauter and colleagues published a landmark 2000 paper in JAMA tracking sleep architecture across the adult lifespan and found that slow-wave sleep — the deepest, most physically restorative kind — drops by roughly 50% between the ages of 16 and 50 in healthy men. REM declines more gradually but steadily. By age 60, many otherwise healthy adults are getting half the deep sleep they had in their teens, even if total time in bed looks similar.

This is part of why sleep quality starts mattering more than sleep duration as you get older. You can be in bed for eight hours and still be running a deficit on the restorative parts of sleep. It’s also why interventions that protect deep sleep — consistent schedules, temperature control, not drinking alcohol too close to bed — tend to pay off more and more as you age.

Muscle mass and cardiovascular fitness also decline with age, but most of those declines respond surprisingly well to training. Sleep architecture is more stubborn. The cleanest defence is to actually use the sleep you can still get.

How much sleep do you actually need?

The National Sleep Foundation’s 2015 expert consensus (Hirshkowitz and colleagues, published in Sleep Health) landed on 7-9 hours for adults aged 18-64, and 7-8 hours for adults 65 and older. These are population averages; real people vary by maybe an hour in either direction for genuinely physiological reasons. But the number of adults who truly thrive on 5 hours is vanishingly small — the research on natural “short sleepers” suggests it’s well under 1% of adults, most of whom carry a specific genetic variant of the DEC2 gene. If you haven’t had that confirmed, you are almost certainly not one of them. Sorry.

The more useful question is usually: how much sleep do you get on a week when nothing external forces you awake? That’s a reasonable approximation of your actual need. Most adults who try this experiment find they sleep about 30-60 minutes more than they assumed they should. Which is mildly annoying but also informative.

Common misconceptions

A few things that come up constantly:

• “I can train myself to need less sleep.” You can train yourself to tolerate less sleep, in the sense that you stop noticing how tired you are. You cannot train your body to need less. Performance on reaction-time and cognitive tests keeps degrading linearly with short sleep even after subjective feelings of tiredness level off, which is the bit that’s genuinely scary.
• “I’ll catch up on the weekend.” Partial recovery is possible, but weekend sleep-ins don’t fully reverse the metabolic and cognitive effects of weekday short sleep. The cleaner approach is not running a deficit in the first place.
• “Alcohol helps me sleep.” It helps you fall asleep, then wrecks the second half of the night. Alcohol suppresses REM sleep and fragments sleep architecture — you spend more time in bed and less time in the stages that matter. I wish this weren’t true, but the evidence is very clear.
• “Naps are cheating.” They aren’t. A 20-30 minute nap improves cognitive performance and doesn’t dent overnight sleep for most people. Longer naps (90+ minutes) can work but risk grogginess if you wake mid-cycle.
• “More sleep is always better.” See above — it isn’t. The curve has a floor, not a direction.

How to measure it

The good news is sleep is one of the easier longevity biomarkers to track, because you’re already carrying a sleep tracker on your wrist or next to your bed. The options, roughly ranked:

• Polysomnography (PSG) — the clinical gold standard, done in a sleep lab with EEG and other sensors. Fantastic accuracy, not practical for regular tracking. Worth doing once if you suspect sleep apnoea.
• Apple Watch / Oura / Whoop — consumer wearables that estimate sleep stages from heart rate, movement, and in some cases pulse oximetry. They’re genuinely quite good at estimating total sleep time and wake periods, and reasonable at detecting stages in aggregate. They’re worse at predicting any single night’s stage distribution, so treat weekly trends more seriously than last night’s deep sleep percentage.
• Sleep diaries — unfashionable but surprisingly useful. Writing down bedtime and wake time for two weeks often reveals patterns the numbers don’t catch.

Sarvita pulls sleep data from Apple Health — Apple Watch, third-party sleep apps, manual entries — and folds it into the recovery picture alongside HRV and activity. The point isn’t to obsess over a single night’s score. It’s to spot whether your restorative sleep is trending up, flat, or quietly sliding downward against the age curve.

How to actually sleep better

The boring part is the most effective part. Honestly:

• Consistent schedule. Bed and wake times within about an hour of each other, including weekends. Your circadian system loves predictability and rewards it.
• Morning daylight. 10-15 minutes of outdoor light in the first hour after waking anchors your circadian clock. This is free and one of the most evidence-backed interventions you can do.
• Cool, dark, quiet. Sleep depth improves noticeably below 19°C for most people. Blackout curtains are unglamorous and effective. Earplugs if you live somewhere like central Munich where the trams run early.
• Alcohol trade-off. Knowing it wrecks sleep doesn’t mean you can’t have a glass of wine. It means being honest with yourself about what it costs and not drinking it three hours before bed.
• Caffeine cutoff. Caffeine has a half-life of roughly 5-6 hours. A coffee at 4 PM is still 25% active in your system at 10 PM. If you’re having trouble sleeping, the espresso you had after lunch is suspect.
• Wind-down ritual. The transition matters. Screens are fine if you’ve made peace with them, but “reading a novel in bed for 20 minutes” has more evidence behind it than most of the expensive sleep tech on the market.

I’m not going to pretend I do all of this. I do most of it most of the time, and my sleep score is, quietly, quite good actually. That’s the honest version.

The practical bit

If you wanted the one-page summary of this whole post, it would read like this:

1. Aim for 7-8 hours of actual sleep for most adults. Not 6, not 10.
2. Quality matters as much as quantity, especially after 40 — protect deep sleep with consistent bedtimes and no alcohol close to bed.
3. Short sleep is more dangerous than long sleep; chronic under-6-hour nights reliably harm metabolic, cardiovascular, immune, and cognitive health.
4. Measure the trend, not any single night. Wearables are good enough for that job.
5. Morning light, consistent schedule, cool room, honest caffeine limits. Boring, unglamorous, effective.

The thing I find genuinely reassuring about sleep is that, unlike most longevity interventions, the timeline is short. You don’t need six months to see the effect of sleeping properly this week. Your HRV will notice in days. Your mood will notice in two. Your metabolic markers in a fortnight. It’s one of the few places in longevity where the feedback loop is tight enough to actually feel.

Anyway. Link’s there if you want Sarvita to track sleep alongside HRV, VO2 Max, and biological age. No pressure — though, fair warning, it does quietly reward you for having boring sleep habits, which is genuinely my favourite kind of gamification.