Can poor sleep cause Alzheimer’s Disease?

Sleep is a hot topic in fitness and wellness at the moment, and rightly so. Thanks to the efforts of scientists like Dr Matthew Walker, the importance of sleep is becoming much more widely appreciated. Sleep is nature’s performance enhancing drug and it’s also a powerful tool for disease-prevention. If you haven’t read Dr Walker’s book yet, I strongly recommend that you do.

My intention with this post, along with a number of future posts, is to look in a bit more detail at the evidence supporting the absolutely critical role of sleep. Today, to start the campaign with some shock and awe, I’ll be addressing the links between poor sleep and Alzheimer’s disease. My hope is that I may be able to scare some of you into taking this seriously!

Normally, I would now warn you of my lack of credentials and discourage you from taking anything I say here as advice since I’m not a doctor or scientific professional. Notwithstanding that, in this case I’m happy to offer one piece of advice: before you think of any other aspect of your lifestyle, fix your sleep. Ok, if you smoke, take care of that first. But otherwise, sleep is primary.

Alzheimer’s disease (AD) affected some 30 million people in 2015 and it’s predicted that by 2040, 81 million patients will suffer from it. Roughly 6% of people over 65 have the disease. If there is any chance sleep can help avoid this terrible disease, I know I would take notice.

A paper by Lim et al. in Sleep in 2013 examined the association between sleep fragmentation and Alzhemeir’s disease in 737 older individuals (average age of 82) in a care home. They measured sleep fragmentation over a 10-day period using wrist-watch actigraphy (movement trackers) and then followed up over a 6 year period to record diagnoses of AD.

The team found that those in the 90th percentile for sleep fragmentation were 50% more likely to develop AD than those in the 10th percentile. In addition, general cognitive decline was 22% more rapid in the 90th percentile group than in the 10th percentile group. A 1 standard deviation increase in sleep fragmentation was associated with a 22% increase in AD risk.

I can already hear some of you chomping at the bit to call me out for hypocrisy… Just last week I outlined all the limitations of observational studies in order to challenge the hypothesis that red meat is killing us. And here I am 1 week later using observational data to make a case for sleep.

There are a few important differences between this study and the red meat study I looked at last week, however. One strength of this study is that the measurement of sleep is objective. Food frequency questionnaires are notoriously susceptible to bias and subjectivity. Wrist-watch measures of movement clearly are not. They may not measure sleep as precisely as polysomnography but nobody could claim that the measurements produced by these devices hide some bias.

Another strength is that the effect sizes under consideration are fairly large. In this study, the average risk of developing AD was 3.94% per year. For a subject with a sleep fragmentation score 1 standard deviation above the mean, they are looking at a risk of 4.73%, an increase of 0.79%. That’s not huge by any means, but not a number I would ignore given the objectivity of the data.

That being said, there are limitations with this study. The chain of causality is always at issue with epidemiology and this paper is no different. As you can see from the chart below, those in the top 90th percentile for sleep fragmentation had lower baseline composite global cognition scores than those in the 10th percentile. In other words, those who sleep more poorly already have greater levels of cognitive impairment so it’s possible that the cognitive decline comes first, the poor sleep second. You’re probably also wondering how relevant 80 year old subjects are to you. If poor sleep in your 80s is associated with AD risk, does it necessarily follow that poor sleep when you’re younger is a risk factor?

Had this been the only form of evidence supporting the link between poor sleep and AD, I probably wouldn’t have written this post. Luckily (or unluckily depending on your current sleep philosophy) there is also good data from a randomised controlled trial (RCT) supporting the hypothesis.

If you want to prove causality, you need a theory for the mechanism by which a risk factor leads to a disease. In other words, what physiological process mediates between poor sleep and AD? What is it about sleep deprivation that increases AD risk?

One theory of AD, known as the Amyloid cascade hypothesis, conjectures that Alzheimer’s disease is caused by a build of protein plaques in the brain. Amyloid-beta and Tau proteins aggregate and bring about the onset of dementia. Some scientists theorise different primary causal factors (and indeed some autopsies show brains with significant plaque build-up but no AD), but most agree that a build-up of amyloid-beta is a significant risk factor and, at any rate, is associated with AD.

Amyloid-beta is a metabolite created in the normal operation of neurones. It has also been discovered that one of the primary functions of sleep is to clear away potentially toxic metabolites through the glymphatic system. During sleep, glial cells, which fill the space between neurones, shrink by up to 60% in size, allowing cerebrospinal fluid (CSF) to enter the interstitial space and clear away metabolic waste-products, including amyloid-beta. Could this be our mechanism? If it could be shown that reduced sleep duration or quality causes a build-up of amyloid-beta, we could make a strong case that poor sleep causes AD.

And in fact this experiment has been done, in a randomised controlled trial published in 2014 by Ooms et al. Researchers took 26 healthy middle-aged men and randomly assigned them to either one night of normal sleep or one night of total sleep deprivation. They measured cerebro-spinal fluid (CSF) levels of amyloid-beta 42 at multiple points in the evening and again in the morning. They found that in the well-slept group, CSF levels of amyloid-beta 42 decreased by 6% over night while in the sleep-deprived group, CSF levels of amyloid-beta 42 were unchanged.

The study also found a correlation between hours slept and reductions of CSF amyloid-beta 42 overnight. The more participants slept, the greater was their clearance of amyloid-beta.


6% may sound like a small difference, but this study demonstrates that sleep is required for the clearance of amyloid-beta. Without it, the body cannot shuttle away this toxic compound. If one night of sleep deprivation prevents amyloid-beta clearance, what does this look like over months and years of poor and reduced sleep? Note also that these subjects were 50 years old on average, so we can’t make the claim that poor sleep is a risk factor for older individuals only. If sleep disrupts the glymphatic system in 50 year-olds, presumably it has the same effect in 40 and 30 year-olds too. How confident are you now that poor sleep is not putting you at risk?

Another observational study looked at the link between self-reported sleep duration and brain scans of amyloid-beta deposits in a group of elderly people. Although fairly small (76 subjects), the results of this study serve as another piece of supporting evidence showing that, not only does poor sleep cause acute build-up of amyloid-beta in CSF, it is also associated with a chronic build-up of the protein in the brain itself. Researchers found a negative (albeit small: -0.36-0.38) correlation between sleep duration and amyloid-beta deposits. The images of brain-scans of representative subjects from different sleep cohorts shows a visible relationship which is hard to ignore (see below).

You could argue that the case is not closed here. Since there is controversy over whether amyloid-beta plaques are a causal factor in the development of AD, evidence of a causal relationship between poor sleep and decreased amyloid-beta clearance cannot be put forward as conclusive evidence that poor sleep causes AD. But with strong observational data as well as a possible mechanism supported by randomised controlled trial research, why take the risk? It looks very likely that there is a strong link between poor sleep and AD. If you are in the 90th percentile of sleep fragmentation, you may be increasing your risk of AD by as much as 50% and every time you have a bad night’s sleep, your clearance of amyloid-beta is compromised. As far as I’m concerned, that’s enough to take the matter of sleep exceptionally seriously. Of course, sleep is likely protective against many diseases, not just AD and it can increase performance in a number of ways. So this is just one angle on the importance of sleep; but hopefully an angle that may scare you into re-thinking a ‘sleep when I die’ philosophy.

As ever, if you have an questions or comments, or if you think I’ve got anything wrong, I’d be more than happy to hear from you. Just comment below or email me at

Leave a Reply

Your email address will not be published. Required fields are marked *