Six Hours of Sleep in Midlife: What the Research Does and Does Not Show

Checked against primary sources by the Editorial Team. Every figure in this article was traced to the study it came from, including the ones that contradict what we previously published. How we review.

This article contains affiliate links. If you buy something through one of them, we may earn a commission at no extra cost to you. It never changes what we recommend or what we say about it. Advertiser Disclosure.

Somewhere in your 40s or 50s, many of us make peace with six hours. You function. You drink coffee. You get through the day. The question is what the research says about that bargain, and how much of it we can trust.

What the large study found

In 2021, Séverine Sabia and colleagues published an analysis of the Whitehall II cohort in Nature Communications. It followed 7,959 British civil servants for up to 25 years and recorded 521 dementia cases.

The authors report that sleeping six hours or less at age 50 and at age 60, compared with seven hours, went with a higher dementia risk. The hazard ratios were 1.22 at age 50 (95% confidence interval 1.01 to 1.48) and 1.37 at age 60 (1.10 to 1.72). At age 70 the ratio was 1.24, but its confidence interval (0.98 to 1.57) included no difference, so that result is less certain.

The authors also report that persistent short sleep across those ages went with a 30% higher dementia risk. They say this held after adjusting for sociodemographic, behavioral, cardiometabolic, and mental health factors.

What this does not show

This is an observational study, so it cannot show that short sleep causes dementia. The authors themselves note that studies with follow up of 10 years or less are likely to suffer from reverse causation, where the disease process affects sleep. They tried to reduce that problem with a 25 year follow up and by looking at sleep at ages 50, 60, and 70 separately. They also point out that the long preclinical phase of dementia makes randomized trials hard to run.

Adjustment for other factors narrows the explanations but cannot remove them. A fair reading: short sleep in midlife is a marker worth taking seriously, and the study cannot tell us whether lengthening sleep would lower anyone's risk.

Why a possible biological link gets attention

Two lines of research are often cited. In mice, Lulu Xie and colleagues reported in Science in 2013 that natural sleep and anesthesia went with a 60% increase in the space between brain cells, and that β-amyloid cleared faster than in waking animals. That is animal work.

In people, Bryce Mander, Matthew Walker, and colleagues reported in Nature Neuroscience in 2015 that, in cognitively normal older adults, more β-amyloid in the medial prefrontal cortex correlated with weaker deep (slow wave) sleep. Their modeling suggested that weaker slow wave sleep statistically accounted for the link between amyloid and poorer memory consolidation. The study is correlational, so it cannot tell us which way the arrow points.

Taken together, these findings make a plausible story. They do not prove that trimming sleep leads to Alzheimer's disease.

Six hours can feel fine and still cost you

Hans Van Dongen and colleagues, writing in Sleep in 2003, assigned 48 healthy adults to 4, 6, or 8 hours in bed per night for 14 days, or to total sleep deprivation. The authors report that restricting sleep to 4 or 6 hours produced cumulative, dose dependent deficits in cognitive performance. Subjective sleepiness did not track the decline well: volunteers stayed largely unaware of how much their performance had slipped.

That matters for the "I function fine on six" belief. The study was short and run in a lab, so it says nothing about dementia. It does suggest that feeling fine is a poor test of whether you are getting enough.

How much sleep counts as enough

The American Academy of Sleep Medicine states that adults should sleep seven or more hours per night on a regular basis to promote optimal health, productivity, and daytime alertness. The National Institute on Aging says older adults need about the same amount as other adults, seven to nine hours, and notes that sleep tends to become shorter and lighter with age, with more waking during the night.

So the shorter, lighter sleep many of us notice is common. Needing less is a different matter, and the guidance says we do not.

If you cannot get there on your own

If you spend weeks struggling to fall or stay asleep, talk to your doctor. Many conditions and medications can interfere with sleep, and a clinician can check them.

The American College of Physicians guideline, by Amir Qaseem and colleagues (2016), recommends that all adults with chronic insomnia disorder receive cognitive behavioral therapy for insomnia, CBT-I, as the initial treatment. The guideline grades this a strong recommendation based on moderate quality evidence. CBT-I combines cognitive techniques, behavioral methods such as sleep restriction and stimulus control, and education about sleep habits, and it is considered before medication.

Based on that guideline, one way to act on it is to work through a CBT-I program in structured form, either with a clinician or with a workbook. The book below is one workbook option. It is a commercial link, and we may earn a commission if you buy it.

Online therapy services built on cognitive behavioral therapy exist too. We name no service here because we could not verify that any of them delivers CBT-I as a named, structured program.

The bottom line

The evidence supports a measured claim. Short sleep in midlife travels with higher dementia risk in a large cohort, lab work shows six hour nights erode performance without us noticing, and professional bodies advise seven or more hours. The evidence does not show that six hours causes dementia, and it does not show that sleeping more will prevent it. It does give you a good reason to stop treating six hours as settled.

This article is for educational purposes only and is not medical advice. If you have persistent sleep problems, please consult a healthcare professional.