How Much Sleep Do You Need? Norms by Age

Adults should sleep seven or more hours a night on a regular basis — that is the wording of the joint consensus statement from the American Academy of Sleep Medicine and the Sleep Research Society. In practice the working range is 7 to 9 hours for adults aged 18 to 64, 8 to 10 for teenagers and 9 to 12 for school-age children. Note that these are hours of sleep, not hours in bed: to get eight hours of sleep, most people need roughly eight and a half hours of opportunity.

Sleep norms by age#

Age-based norms are not prescriptions. They are the ranges into which the sleep of healthy people falls, derived by expert panels reviewing the published evidence — the AASM with the Sleep Research Society for adults, the AASM for children and adolescents, and a parallel National Sleep Foundation panel. The three sets of figures agree almost entirely, and the NIH's National Heart, Lung, and Blood Institute publishes the same numbers for patients.

Age	Recommended	May be appropriate	Notes
0–3 months	14–17 hours	11–19 hours	Sleep is fragmented; no consolidated rhythm yet
4–11 months	12–16 hours	10–18 hours	Including daytime naps
1–2 years	11–14 hours	9–16 hours	Usually consolidating to a single nap
3–5 years	10–13 hours	8–14 hours	Napping typically stops between ages 5 and 7
6–12 years	9–12 hours	7–12 hours	Shortfall shows up as attention and school problems
13–18 years	8–10 hours	7–11 hours	Body clock shifts later during puberty
18–25 years	7–9 hours	6–11 hours	Same requirement as older adults
26–64 years	7–9 hours	6–10 hours	Habitually under 7 hours is a risk factor
65+ years	7–8 hours	5–9 hours	Need barely changes; sleep architecture does

Two rows deserve comment, because intuition usually gets them wrong.

Older adults do not need less sleep. The belief that five or six hours suffices after 65 is not supported by any of the expert panels; the recommendation is 7 to 8 hours, essentially the same as for middle age. What changes is architecture: slow-wave sleep shrinks, awakenings multiply, and the circadian phase advances so both sleep and waking come earlier. Older people sleep less well, which is not the same as needing less sleep.

Teenagers are not lazy. During puberty the circadian phase delays by one to three hours and sleep pressure accumulates more slowly. A 16-year-old who genuinely cannot fall asleep before midnight is behaving normally. The problem is arithmetic: an 8 to 10 hour requirement against a 7 a.m. school start guarantees a structural deficit five days a week.

Why 7 to 9 hours for adults#

The number seven in the AASM/SRS statement is a threshold, not an average. The panel reviewed more than 5,000 published articles and asked, for each family of outcomes — cardiovascular, metabolic, immune, cognitive, mental health, safety — at what duration risk begins to rise measurably. Habitually below seven hours, risk rises across most of those domains. At seven and above, it does not. A later meta-analysis and meta-regression by Itani and colleagues reached the same conclusion from the other direction, linking short sleep to mortality, diabetes, hypertension, cardiovascular disease and obesity.

The upper bound of nine works differently. The consensus statement is explicit that the evidence on sleeping more than nine hours in healthy adults is insufficient to call it harmful. The figure is statistical rather than biological. If you naturally sleep nine and a half hours and wake refreshed, nothing is wrong.

Individual variation within the range is real and substantially heritable. One person is genuinely fine on seven hours, another needs nine, and willpower has nothing to do with it. Treat the range as the map and your own experience as the territory — there is a method for finding your number at the end of this article.

The U-shaped curve and the reverse causation trap#

Plot sleep duration against all-cause mortality in large observational cohorts and you get a U: risk is elevated both below about six hours and above about nine. Meta-analyses reproduce this reliably. The review of long sleep duration by Jike and colleagues found associations with mortality, diabetes, cardiovascular disease, obesity and depression.

The tempting conclusion — that sleeping a lot is dangerous and should be curtailed — is almost certainly wrong.

The two arms of the U differ in what kind of evidence supports them.

The short arm is backed by experiments. You can restrict healthy volunteers' sleep in a laboratory and watch attention, glucose tolerance and immune response deteriorate. Causation is demonstrated directly.

The long arm is not. Nobody has shown that extending a healthy person's sleep harms them. What is well documented is the long list of conditions that lengthen sleep: depression, chronic inflammatory disease, undiagnosed cancer, heart failure, hypothyroidism, sedating medication, and obstructive sleep apnea — where sleep is so fragmented that the body compensates with quantity. Add the fact that "sleep duration" in most cohorts is self-reported time in bed, and the long arm looks less like a cause than a signal.

There is exactly one practical takeaway. If you routinely sleep nine or ten hours and still feel wrecked, the answer is not discipline. It is a work-up: thyroid function, iron studies, mood, and a look at whether you are breathing properly at night.

Читайте также: Sleep Hygiene: 12 Rules That Actually Work

What chronic sleep debt actually costs#

Not the all-nighter before a deadline, but the ordinary version: six hours instead of seven and a half, for months.

Attention — and the fact that you cannot feel it#

The landmark experiment here is Van Dongen and colleagues' 2003 study. Healthy volunteers were restricted to 4, 6 or 8 hours in bed for two weeks while reaction time and self-rated sleepiness were measured daily. Two findings came out of it.

First, the deficit accumulates in a dose-dependent way: after two weeks, the six-hour group performed on lapse-of-attention tests about as badly as people who had gone one to two nights entirely without sleep. Second, and more consequentially, subjective sleepiness plateaued after a few days while objective performance kept falling. Participants stopped being able to perceive how impaired they were.

That gap explains the entire population of people who insist six hours is fine for them. It is fine as an experience. It is not fine as a level of function.

Metabolism#

Sleep restriction reduces insulin sensitivity and glucose tolerance in healthy young adults, a result reproduced repeatedly since Spiegel and colleagues' 1999 study in the Lancet. Appetite regulation shifts at the same time: ghrelin rises, leptin falls, and preference tilts towards energy-dense food. Epidemiologically, short sleep tracks consistently with obesity and type 2 diabetes.

Immunity#

Prather and colleagues measured a week of sleep with wrist actigraphy in 2015, then exposed volunteers to rhinovirus under controlled conditions. Those sleeping under six hours were roughly four times more likely to develop a cold than those sleeping more than seven. Related work shows blunted antibody responses to vaccination when sleep is short around the time of the shot.

The road#

The sharpest and most underrated risk. Analyses of real crash data show a dose-response relationship between hours slept in the preceding 24 hours and the probability of being the driver at fault — with the risk at under four hours of sleep several times that of drivers who slept 7 to 9 hours.

The mechanism is the microsleep: an intrusion of sleep lasting one to ten seconds that the person does not register. At 60 mph, four seconds of microsleep is a hundred metres of road driven blind.

The short sleeper myth#

Natural short sleepers are real. They are also vanishingly rare.

In 2009 Ying-Hui Fu's group described a mother and daughter who thrived on about 6.25 hours of sleep, and traced it to a mutation in the DEC2 gene (BHLHE41); introducing the same mutation into mice shortened their sleep too. In 2019 the same laboratory reported a mutation in ADRB1, the beta-1 adrenergic receptor, in a family whose carriers slept around 5.5 to 6 hours with no ill effects. Variants in NPSR1 and other genes have since been added.

Three things matter about these discoveries.

• They are rare. Prevalence estimates for true natural short sleep are a fraction of a per cent. The prior probability that you are a carrier is very low.
• They are congenital and lifelong. A real short sleeper never trained for it. They slept little as a child, do not use an alarm, do not sleep in on holiday and never feel sleepy during the day. If you catch up at weekends, you are not one.
• They cannot be acquired. Programmes promising to train you down to four or five hours do not create short sleepers. They create sleep-deprived people whose self-perception has gone numb.

Can you catch up at the weekend?#

Partly, and with caveats.

What recovers: subjective alertness, much of the attention deficit, and some hormonal measures. Epidemiological work even suggests that people who sleep short on weekdays but catch up at weekends fare better than those who sleep short all seven days. Catching up beats not catching up.

What does not recover: metabolic function. Depner and colleagues restricted participants' sleep for five days, allowed two nights of unrestricted recovery sleep, then restricted them again. The weekend recovery did not prevent the loss of insulin sensitivity, nor the late-night snacking and weight gain — and after the return to restriction, some markers were worse than in the group that never got the recovery weekend at all.

Then there is the cost of the phase shift. Waking three hours later on Saturday is roughly equivalent to flying three time zones west, with the return flight on Sunday night. That is where the Sunday insomnia and the wretched Monday morning come from; the phenomenon has a name, social jet lag.

The workable compromise: allow yourself at most one extra hour in the morning at weekends. If you still need more sleep, take it as an afternoon nap rather than a later wake-up — naps do not move your body clock. And treat catch-up sleep as an emergency measure, not a scheduled feature of the week.

Читайте также: How to Manage Stress: Methods That Actually Work

Napping: how long and when#

Naps are a legitimate tool when used with some precision.

Ten to twenty-five minutes. A short nap stays in the light stages, which are easy to exit, and buys two to three hours of improved alertness and reaction time.

Before mid-afternoon. Later naps spend sleep pressure you need for the evening and directly delay falling asleep at night.

Not 30 to 60 minutes. This is the worst window: long enough to reach deep slow-wave sleep, too short to finish the cycle. Waking from that stage produces sleep inertia — 20 to 40 minutes of grogginess and the distinct feeling of having made things worse.

Ninety minutes is the alternative if you have the time, since a full cycle ends in light sleep and inertia is minimal. But a nap that long will noticeably reduce your sleep drive that night.

Two exceptions. If you have trouble falling asleep in the evening, cut napping entirely — it is working against you. And if you need to nap daily despite a full night's sleep, that is a symptom worth investigating rather than a habit to optimise.

Quality versus quantity#

Eight hours of bad sleep is worse than seven hours of good sleep. Duration is only one dimension, and AASM documents are careful to state that continuity, regularity and timing matter alongside it.

• Continuity. Sleep broken by repeated arousals does not restore you the same way. Obstructive sleep apnea is the extreme case: nine hours in bed, hundreds of micro-arousals, and a person who wakes exhausted.
• Regularity. A stable wake time affects how you feel as much as total duration does, and in several cohort studies irregular sleep timing predicted cardiovascular events independently of how long people slept.
• Timing. The same seven hours taken from 3 a.m. to 10 a.m. sit badly against the internal clock and are tolerated worse. This is the core problem of shift work.

The sensible order of operations: get the duration right first, then the regularity, and only then start fine-tuning the bedroom.

How to find your own number#

The reliable method is two weeks without an alarm — a holiday is ideal. For the first three to five nights you will oversleep, repaying accumulated debt. After that, duration settles onto a plateau. That plateau is your requirement.

Without a holiday, use four proxy signs of sufficient sleep:

1. You wake at roughly the same time on your own, often a few minutes before the alarm.
2. You are not fighting sleepiness during the morning. The brief post-lunch dip is normal and does not count.
3. You sleep in by no more than an hour at weekends.
4. You do not need caffeine to function, as distinct from drinking coffee because you enjoy it.

If two or more of those fail, run an experiment: for two weeks, go to bed 30 to 45 minutes earlier while keeping your wake time fixed. That is usually enough to feel the difference — and to realise that what you had been calling normal was a chronic deficit.

Читайте также: How Much Protein Do You Need Per Day?

The summary is unglamorous. Seven to nine hours for adults, and the overwhelming majority of people who consider themselves exceptions are not. The upper end should not alarm you: nine hours followed by a refreshed morning is fine. Nine hours followed by exhaustion is not a sleep quantity problem at all, and it belongs in a doctor's office.