Of the handful of things you can actually change about how you age, sleep is the one most people treat as negotiable. Diet gets planned, exercise gets scheduled, and sleep gets whatever is left over. Yet it turns up in the research on cognition, cardiovascular health, metabolic function and mood more consistently than almost any other daily habit — which makes it a strange thing to hand over last.
It is also surrounded by a belief, repeated often enough that it now sounds like common sense: that older people simply need less sleep. This is one of the most persistent myths in health, and it carries a cost. It teaches people in their fifties and sixties to accept broken nights as an inevitable part of getting older, rather than as something worth investigating.
The evidence points somewhere else. What falls with age is not how much sleep your body needs. It is how easily you can get it.
The number itself
The most widely used reference point comes from a National Sleep Foundation panel of eighteen specialists drawn from twelve professional organisations, who reviewed the published literature and voted through a formal consensus process. Their conclusion for adults aged 18 to 64 was seven to nine hours a night. For adults over 65, the recommendation was seven to eight.
That is a narrow difference, and it is worth being precise about what it means. The panel was explicit that these are guidelines for healthy people without sleep disorders, that individual requirements vary, and that an hour either side of the range may still be appropriate for a given person. They also noted something important: habitually sleeping well outside the normal range can itself be a signal of an underlying health problem.
So the headline is not that sleep need collapses after 65. It is that it barely moves.
Most of us are not reaching it. In a YouGov survey of British adults, seven hours was the most common answer at 35%, but 22% reported six hours and one in eight reported fewer than six. Half of those surveyed — 49% — said they did not think they were getting enough. This is self-reported data from an online panel rather than laboratory measurement, so treat the precise percentages loosely; the direction of travel is the point.
Need versus ability
A later National Sleep Foundation report on sleep and ageing put the distinction plainly: the need for sleep does not decrease with age, but the ability to sleep can.
Several things change at once. Sleep architecture shifts — there are more awakenings during the night, less deep sleep, and a tendency to fall asleep earlier in the evening and wake earlier in the morning. The result is sleep that feels lighter and less restorative even when the hours on the clock look unchanged.
On top of that sit conditions that become more common with age and interfere with sleep directly: obstructive sleep apnoea, chronic pain, cognitive impairment, and the side effects of medication. Life changes matter too — retirement, bereavement, reduced daytime activity, and sleeping alone for the first time in decades all alter the structure of a night.
This is why “I don’t need as much sleep as I used to” is worth questioning when you hear yourself say it. Often the honest version is “I can’t get as much as I used to.”
Why seven keeps appearing
Two large analyses of the UK Biobank dataset, published within weeks of each other in 2022, landed on the same figure.
The first, in Nature Aging, drew on close to 500,000 participants aged 38 to 73, with brain imaging and genetic data available for roughly 40,000 of them. The relationship between sleep duration and cognition was not a straight line but a curve, with approximately seven hours sitting at the top of it. Both too little and too much sleep were associated with weaker performance on processing speed, visual attention, memory and problem-solving — and, on follow-up, with measurable cognitive decline.
The second, in Communications Biology, analysed 479,420 participants and found the same shape: seven hours associated with the highest cognitive performance, falling away for every hour above and below. The pattern held in the over-60s specifically, a subgroup of more than 200,000 people. Those sleeping between six and eight hours also had greater grey matter volume in 46 of 139 brain regions examined.
Two independent analyses converging on the same number is more persuasive than either alone. But the caveats matter, and honest coverage of this research should state them. Both studies are observational, which means they can show association but cannot establish that short sleep causes cognitive decline rather than early cognitive changes causing disrupted sleep. Sleep duration was self-reported. And the UK Biobank cohort is roughly 94% white and of European ancestry, so its findings should not be assumed to transfer equally to every population.
What happens below six hours
The most useful evidence on long-term risk comes from the Whitehall II study, which has followed British civil servants since the mid-1980s. Researchers led by Séverine Sabia examined 7,959 participants across a 25-year follow-up, during which 521 were diagnosed with dementia.
Compared with people sleeping a normal seven hours, those sleeping six hours or less at age 50 had a hazard ratio of 1.22 for later dementia — a 22% higher rate — with a confidence interval of 1.01 to 1.48. At age 60 the figure rose to 1.37 (1.10 to 1.72). At age 70 the estimate was 1.24, but the confidence interval crossed 1.00 (0.98 to 1.57), meaning the result there was too imprecise to draw conclusions from.
People who slept short consistently across all three ages had a 30% higher dementia risk than those who consistently slept a normal amount, and this held after adjusting for sociodemographic, behavioural, cardiometabolic and mental health factors.
The average age at dementia diagnosis in the study was 77. That gap — decades between the sleep pattern and the diagnosis — is what makes the finding interesting, because it makes reverse causation less likely than in studies with shorter follow-up. It does not eliminate it. Dementia has a long preclinical phase, and disturbed sleep can be an early symptom rather than a cause.
Read the numbers as a signal, not a sentence. A 22% higher rate is a meaningful population-level finding and a poor basis for personal panic.
Working out your own number
The ranges above describe populations. Yours is a single number inside them, and there is a simple way to find it.
Give yourself a stretch of days without an alarm — a holiday works best. Go to bed when you are genuinely sleepy rather than at a fixed time, and let yourself wake naturally. The first few days will run long as you clear accumulated sleep debt. By the end of the week the duration tends to settle, and where it settles is a reasonable estimate of what you actually need.
Two everyday signals are more useful than any tracker. If you fall asleep within minutes of your head touching the pillow every single night, that is often a sign of being under-slept rather than a sign of good sleep. And if you need an alarm to wake up on most days, you are probably going to bed later than your body would choose.
When it isn’t about hours at all
If you are spending eight hours in bed and still waking unrefreshed, adding a ninth will not fix it. That pattern points to sleep quality, and quality problems have specific causes worth naming to a GP rather than managing alone.
Loud snoring, gasping or pauses in breathing witnessed by a partner, and heavy daytime sleepiness are the classic signs of obstructive sleep apnoea, which is both common and treatable, and frequently goes undiagnosed for years.
Difficulty falling or staying asleep on at least three nights a week for at least three months, with real consequences for daytime functioning, meets the standard definition of chronic insomnia. NICE guidance is clear that the first-line treatment is cognitive behavioural therapy for insomnia — CBT-I — rather than medication. It is also clear-eyed about the gap between guidance and reality: face-to-face CBT-I has limited availability across the UK, which is why NICE has assessed digital programmes as a route to the same treatment. If your GP reaches for a prescription first, it is reasonable to ask about CBT-I.
Neither condition belongs in the category of things you simply put up with because you have had a birthday.
This article is for general information and is not medical advice. If sleep problems are affecting your daily life, speak to your GP.
Sources
- Hirshkowitz M, et al. National Sleep Foundation’s sleep time duration recommendations: methodology and results summary. Sleep Health, 2015;1(1):40–43. https://pubmed.ncbi.nlm.nih.gov/29073412/
- National Sleep Foundation. Sleep health and aging: recommendations for promoting healthy sleep among older adults. Sleep Health, 2023. https://www.sleephealthjournal.org/article/S2352-7218(23)00196-1/fulltext
- YouGov. The YouGov Sleep Study: Part one — sleeping patterns, 2022. https://yougov.co.uk/articles/42961-yougov-sleep-study-part-one-sleeping-patterns
- Li Y, Sahakian BJ, Feng J, et al. The brain structure and genetic mechanisms underlying the nonlinear association between sleep duration, cognition and mental health. Nature Aging, 2022. DOI: 10.1038/s43587-022-00210-2
- Tai XY, et al. Impact of sleep duration on executive function and brain structure. Communications Biology, 2022;5:201. https://www.nature.com/articles/s42003-022-03123-3
- Sabia S, Fayosse A, Dumurgier J, et al. Association of sleep duration in middle and old age with incidence of dementia. Nature Communications, 2021;12:2289. https://www.nature.com/articles/s41467-021-22354-2
- NICE. Sleepio to treat insomnia and insomnia symptoms. Medical technologies guidance MTG70, 2022. https://www.nice.org.uk/guidance/mtg70/chapter/1-Recommendations
This article is for general information and is not medical advice. If a health problem is affecting your daily life, speak to your GP.
Comments
We read every comment. Be kind, stay on topic, and please don't ask for medical advice — we can't give it.