Sleep Disruption and Alzheimer Disease: A Bidirectional, Treatable Relationship

Winnie Pao, MD, described sleep disruption as a treatable contributor to cognitive decline, even though sleep has not been included among the 14 modifiable dementia risk factors identified by the Lancet Commission.¹ Pao outlined a bidirectional relationship in which sleep apnea and other sources of sleep fragmentation promote amyloid accumulation, hypoxia, inflammation, and vascular risk, which in turn further degrade sleep architecture. Sleep fragmentation, reduced sleep efficiency, reduced rapid eye movement sleep, increased light sleep, and sleep-disordered breathing have been identified as predictors of cognitive decline in mild cognitive impairment and Alzheimer disease.² Pao emphasized that the therapeutic goal is consolidated, high-quality sleep rather than adherence to continuous positive airway pressure alone, and that clinicians should review medications, pain, and environmental contributors. Pao also noted that future studies require better characterization of sleep apnea endotypes and of baseline neurodegeneration among long sleepers, and encouraged psychiatric clinicians to partner with sleep specialists on pharmacologic management.

Psychiatric Times: How do you see the relationship between sleep disorders and Alzheimer disease?

Winnie Pao, MD: I like to tell my patients that dementia is very hard to treat. There's the aging process, and we don't have very good pharmacologic treatments. We're trying to do better, but it's still not as good as we would like.

Lancet has put out a list of 14 modifiable risk factors for dementia, starting around 2017, and every few years they add more. Each time I wait for sleep to be included, but it isn't yet. We're hoping that in the next edition sleep will be right at the top.

So when a patient has sleep apnea, I say, "Congratulations, because now we have something we can treat." Many people believe the relationship is bidirectional. Sleep interruption, which can be caused by sleep apnea, leads to accumulation of abnormal proteins such as amyloid. These deposit in the breathing center, the circadian rhythm center, and the wakefulness centers, and they affect sleep structure. Worse sleep structure then causes more interruption, which causes more deposition of these harmful proteins, and you get a vicious cycle. You want to keep that cycle from occurring, which is why treating sleep is so important.

Beyond sleep structure, the drop in oxygen and the inflammation also increase vascular risk factors, which can cause vessel disease and ischemia and worsen aging. So sleep apnea isn't just changing the brain through oxygen drops and disrupted sleep, and the brain's inability to clear out these proteins. It also causes comorbidities such as cardiovascular disease, high blood pressure, and diabetes. It becomes a whole-system problem.

Patients are usually surprised when I congratulate them. Then comes the CPAP issue: treating sleep apnea doesn't always mean using CPAP, and using CPAP doesn't always treat sleep apnea. Patients tend to think of CPAP right away and are taken aback. They don't like the mask on their face and don't think they can handle it, so they have no vested interest in treating the sleep apnea. The key is not about using CPAP. It's about how we get good, undisrupted sleep and how we maximize the benefit of sleep. We all sleep, and it's very important. That's what I try to convey to my patients: it's not "I want to treat your sleep apnea," it's "I want you to sleep better."

PT: Besides sleep apnea, are there other sleep disorders linked to cognitive decline or dementia?

Pao: Anything that disrupts your sleep is going to cause a problem. Restless legs syndrome is one example. These patients may have a hard time falling asleep to begin with, and it can be very painful, so they're up walking around. Many of them also have periodic limb movements of sleep, which can wake them, so their sleep isn't consolidated. Insomnia is another. As people get older, there are aches and pains, medications that affect sleep, worries, noise, and the environment.

It all comes back to disruption of the actual sleep. You can look at pain, heart disease, breathing disorders, restless legs syndrome, peripheral neuropathy, and situational factors too: a loud neighborhood, a dog that won't let you sleep, a bed partner who snores. If any of these disrupt your sleep, you're at risk. It's about sleep health and the quality of sleep.

PT: If a psychiatric clinician hears a patient's sleep quality is declining, how can they assess for suspected cognitive decline or Alzheimer disease?

Pao: With Alzheimer disease, it's more complicated and depends on severity. Early on, the circadian rhythm is affected. As we get older, our sleep is less consolidated, and with changes in sleep spindles we have more difficulty with memory consolidation. We also have more medical conditions, aches, and pains.

On top of that, the rhythm can be worsened by neurodegenerative processes, such as accumulation of amyloid in the suprachiasmatic nucleus, the master clock. We don't sleep as deeply or as well. Body temperature drops later than it should, and since that drop helps us fall asleep, sleep is pushed later. Patients become more active toward nighttime, and at night they may get more confused. That's sundowning, which comes from a circadian rhythm problem. They become agitated, suspicious, and start pacing. They don't sleep at night and sleep during the day. The sleep-wake pattern becomes disrupted, and we call this irregular sleep-wake rhythm disorder. It's the dreaded thing we see in advanced Alzheimer disease and dementia.

I recently looked at an article in the Journal of Alzheimer's Disease (2024) that reviewed 17 longitudinal studies. It identified five sleep predictors of cognitive decline in mild cognitive impairment or Alzheimer disease: sleep fragmentation, reduced sleep efficiency, reduced REM sleep, increased light sleep, and sleep-disordered breathing such as sleep apnea. Essentially, it's all about poor sleep. Whatever the reason a person can't get consolidated sleep, whether it's abnormal aging, a medical condition, or the environment, it raises the risk.

I have an interesting patient case. A woman in a nursing home came to us because she was very sleepy during the day. She told the staff that her roommate was very noisy at night and wouldn't quiet down, so by the time the roommate settled, it was early morning before she could finally sleep. She had asked them to change something, but nothing happened. So she continued to not sleep. This wasn't insomnia. It was her roommate, who had a behavioral problem.

When we ask why a patient is having trouble sleeping, we need a good history first. What medical conditions do they have? What's the environment like? What medications are they taking? A water pill or too much water may be waking them at night, or a beta blocker may be suppressing melatonin secretion. So we look at whether we can adjust or reduce medications, improve the environment, or increase daytime activity so they have enough sleep pressure at night. We should also address pain.

Parkinson disease is another example. Patients take medication during the day for movement, but sometimes we skip it at night because they "don't need to move." Then they get stiff, can't turn over, and are uncomfortable and restless. We may need to adjust the dosing or use a longer-acting formulation to carry them through the night. So knowing the person's history, situation, and medications will help, and everyone may be quite unique.

PT: If clinicians start treating these sleep disorders, how can they evaluate whether it's helping cognition?

Pao: They need to define what "cognitive aspect" they mean and what the baseline is. They might say confusion, behavioral changes, argumentativeness, or agitation, or they might mean memory problems. We need to understand what they mean. We also need to ask whether the patient is improving because we treated the sleep, since psychiatric problems can also affect cognition.

I recently saw a patient with many medical problems, including chronic pain. He was fighting to sleep. He has sleep apnea, depression, and anxiety, and he's on high doses of psychotropic medication and pain medication. He has excessive daytime sleepiness that we can't fully manage by treating the sleep apnea alone, and the sleep doctor can't do much given his medication burden. The psychiatrist is best positioned for the pharmacology.

I may have to give him a little stimulant to see whether it works, because I can't test his cognition at all while he's fighting to sleep. I can't tell whether his test performance reflects his cognition or just his sleepiness, and I'm worried about giving him more medication. So people need to come together. So many people are prescribing that we need some consolidation, and pharmacy consults can help, but someone has to take charge and see whether the medications can be cleaned up. More often than not, I've seen people, usually younger people, say their memory is much better after a psychiatrist cleans up some of their medications.

So sometimes medication makes sleep worse. In addition to treating sleep, we have to look at whether patients are taking too many medications. Providers are often at different institutions, and even under the same roof we don't always talk to each other. When we don't, the medication picture gets messy.

PT: What are the biggest gaps in understanding how sleep and Alzheimer disease interact? Where should more research be done?

Pao: There's a lot of research going on right now. One interesting point is that not all sleep apnea is the same, and when we treat it, we need to ask whether we see improvement. Long retrospective studies may show that people without sleep apnea are less likely to develop Alzheimer disease. But how much of that is truly sleep apnea alone? And CPAP studies of a few months aren't long enough, while long-term studies need a lot of funding. Some studies use registry data and assume that if Medicare is paying for CPAP, the patient must be using it. But you don't know how often or how much they're using it, or whether the therapy is adequately controlled.

There are also different endotypes, or endophenotypes. Two people can have the same apnea-hypopnea index but differ: one may drop oxygen more, another may have more arousals. Those differences generally aren't captured. So what would help is longitudinal studies with better characterization of the patient population. Some treatments work better in some patients than others, and if you lump everyone together, you dilute the effect, if there is one, and you may not be tracking the right thing.

Here's an example. Studies consistently show that short sleep time increases the risk of cognitive decline and dementia, and people also describe a U-shaped curve in which long sleep time increases risk as well. But those studies are less than 10 years long. So the question, which someone else raised, is whether we're sampling patients correctly. A neurodegenerative process may already be underway and making these people sleep longer, so when you recruit long sleepers, they aren't cognitively normal. Of course they develop dementia later, because the seed of the disorder was there when they were recruited. That's what I mean by selecting the right population. If we don't, we may not answer the right question.

We also need to define what we're testing. Are we testing CPAP, or are we testing treatment of sleep apnea? Or are we talking about consolidating sleep and improving sleep quality? How we ask the question affects how we conduct the study and what answer we get. If different endotypes respond differently to CPAP and you combine them, you won't get the right answer. Sometimes CPAP even hurts a person, and if you don't capture that, you say "CPAP is good" and push CPAP. Let's step back from CPAP and ask whether we're treating sleep apnea, or improving sleep.

PT: That's an interesting distinction. It isn't "everyone gets CPAP and that fixes it."

Pao: Exactly, it doesn't. There are people who hate it and don't find it helpful, and they aren't lying. Then we scratch our heads: it should work, so what are we missing? We tell them, "You're fine, go home." But they're not fine. They're still sleepy, they still can't function, and they still have memory problems. It may be that there's still a sleep problem.

PT: What one piece of advice would you give practicing psychiatric clinicians about sleep issues and dementia?

Pao: If you run into difficulty, there's always a friendly neighborhood sleep physician who can help. Reach out to us. If we can't answer the question, we might be able to brainstorm with you. Two heads can see things that one person who is deeply involved may not notice.

We also need psychiatry. A lot of patients with insomnia have psychiatric issues going on, and there's the medication side. Psychiatrists are the experts in pharmacology, definitely better than the rest of us. Knowing which medications are right to use and what to be careful about is very important. We'd love to have a sleep psychiatrist, but they're hard to find, which shows how important psychiatry is to sleep. If you don't have a sleep psychiatrist, then you have a sleep doctor and a psychiatrist, and you can work together.

Dr Pao is a neurologist and expert in sleep, memory, and cognitive disorders at the Cleveland Clinic.

References

1. Livingston G, Huntley J, Liu KY, et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. Lancet. 2024;404(10452):572-628.

2. Carpi M, Fernandes M, Mercuri NB, et al. Sleep biomarkers for predicting cognitive decline and Alzheimer's disease: a systematic review of longitudinal studies. J Alzheimers Dis. 2024;97(1):121-143.