Sleep and aging: what the evidence shows about non-drug approaches that work

Sleep changes with age. That much is settled science. What is less settled, and where a lot of bad advice creeps in, is what those changes mean and what an older adult can do about them without reaching for a pill. This piece reports what randomized trials, systematic reviews, and clinical guidelines have found. It is evidence reporting, not medical advice.

How sleep changes with age

The clearest map of normal aging sleep comes from a meta-analysis of 65 studies and 3,577 healthy people from childhood to age 102 (Ohayon, Sleep, 2004). Across adulthood, total sleep time, sleep efficiency, the share of deep slow-wave sleep, and the share of REM sleep all decline, while the time it takes to fall asleep and the time spent awake after falling asleep both rise. In plain terms, sleep gets lighter, shorter, and more broken up.

The size of these shifts is modest and gradual. A review of the aging literature reports that men lose roughly 1.7% of slow-wave sleep per decade, while women show little age-related change in deep sleep. Time awake during the night rises about 10 minutes per decade between ages 30 and 60, then tends to plateau (Sleep in Normal Aging, 2017). The body clock also shifts earlier with age, an advance of around an hour, which is why many older adults feel sleepy earlier in the evening and wake before dawn. One practical point: waking more often is common with age, but loud snoring, gasping, and heavy daytime sleepiness are not part of normal aging and can signal sleep apnea, which is worth raising with a clinician.

The behavioral approaches with the strongest evidence

When insomnia becomes chronic, the first-line treatment in the guidelines is not a drug. The American College of Physicians gives cognitive behavioral therapy for insomnia (CBT-I) a strong recommendation as initial treatment for all adults with chronic insomnia (ACP, Annals of Internal Medicine, 2016). The American Academy of Sleep Medicine reached the same conclusion, with a strong recommendation for CBT-I and conditional support for its component parts, such as stimulus control and sleep restriction (AASM, JCSM, 2021). CBT-I bundles education about sleep, a consistent schedule, and techniques to break the anxious association between bed and wakefulness. It usually runs four to eight sessions.

Several everyday habits inside CBT-I have their own trial support:

Consistent sleep and wake times. A steady schedule anchors the body clock. This is a core component of CBT-I rather than a stand-alone tested therapy, so the evidence is strongest as part of the package.

Morning bright light. Light is the dominant signal for the circadian clock. Randomized trials in older adults, including nursing-home residents and people with cognitive impairment, show variable but generally positive effects on sleep and rhythm stability, though results are mixed and the highest-quality evidence is limited (JAMDA systematic review, 2022).

Physical activity. This has some of the better randomized evidence. A meta-analysis of randomized controlled trials found exercise improved self-reported sleep quality by about 2.2 points on the Pittsburgh Sleep Quality Index versus control (14 trials, 1,358 people; PMC8215288). Effects were seen with both conventional and mind-body exercise such as tai chi, though heterogeneity between trials was high.

Alcohol, caffeine, and naps

Alcohol is a common self-prescribed sleep aid, and the evidence explains why it backfires. A 2024 systematic review of 27 controlled studies found alcohol raises deep sleep early in the night, then suppresses REM sleep and fragments the second half of the night (Sleep Medicine Reviews, 2024). The net effect is worse, not better, sleep.

Caffeine is dose- and timing-dependent. Pooled across controlled trials, caffeine cut total sleep by about 45 minutes and lowered sleep efficiency by about 7% (Sleep Medicine Reviews, 2023). A randomized crossover trial found little effect from 100 mg taken up to four hours before bed, but 400 mg disrupted sleep even when taken many hours earlier (Sleep, 2025). Caffeine’s half-life varies widely between people, from roughly 2 to 10 hours, so individual tolerance differs.

Naps are more nuanced. Short naps appear low-risk, but long habitual naps are associated with worse outcomes in cohort studies. A meta-analysis linked naps of 60 minutes or more to higher cardiovascular disease and all-cause mortality risk (PLOS One, 2024). This is an association, not proof of harm. Long napping may be a marker of poor nighttime sleep or existing illness rather than a cause.

Where sleep meets health: what is causal and what is not

Poor sleep travels with worse health, but the direction of cause is often unclear. A dose-response meta-analysis of prospective cohorts found both short and long sleep linked to higher mortality and cardiovascular events, with the lowest risk near 7 hours (JAHA, 2017). Because these are observational data, they show correlation, not a proven cause.

For mood, the evidence is a little stronger. In a meta-analysis of longitudinal studies, insomnia roughly doubled the later risk of depression (odds ratio about 2.60; Baglioni, 2011). The relationship runs both ways, so depression also predicts later sleep problems.

On dementia, caution is warranted. Meta-analyses of cohort studies report that insomnia, excessive daytime sleepiness, and both short and long sleep are associated with higher risk of later cognitive decline and dementia (GeroScience, 2025). But sleep changes can be an early sign of neurodegeneration already underway, so this link is associative and cannot yet be read as proof that treating sleep prevents dementia.

A note on sleep medications

Guidelines put drugs second for a reason. The American Geriatrics Society Beers Criteria list benzodiazepines and the “Z-drugs” (such as zolpidem) as medications older adults should generally avoid for insomnia, citing raised risks of confusion, falls, and fractures (AGS Beers Criteria, 2023). That is why the guideline bodies place behavioral treatment first.

This article is evidence reporting, not medical advice. Sleep problems can have many causes, and medication decisions in particular carry real trade-offs. Talk with a qualified clinician about your own situation before making changes.

Common questions

Do older adults need less sleep?

Not really. Research on healthy aging finds that the ability to produce deep, consolidated sleep declines, so older adults often get less sleep, but the underlying need stays broadly similar to other adults. Most cohort data point to lowest health risk around 7 hours. Waking more during the night is common, but heavy daytime sleepiness is not a normal part of aging and is worth discussing with a clinician.

What is the most evidence-based non-drug approach for insomnia after 50?

Cognitive behavioral therapy for insomnia (CBT-I). Both the American College of Physicians (2016) and the American Academy of Sleep Medicine (2021) give it their strongest recommendation as first-line treatment, ahead of sleep medications. It combines a consistent schedule, stimulus control, and techniques to reduce sleep-related anxiety, usually over four to eight sessions.

Does exercise genuinely improve sleep, or is that just a common belief?

There is randomized-trial support. A meta-analysis of controlled trials found exercise improved self-reported sleep quality by about 2.2 points on the Pittsburgh Sleep Quality Index compared with control groups. Both conventional exercise and mind-body forms such as tai chi showed benefit, though results varied between studies.

Is a nightcap a reasonable way to fall asleep?

The evidence says no. Controlled studies show alcohol may help you fall asleep faster and deepen sleep early in the night, but it then suppresses REM sleep and fragments the second half of the night, leaving sleep worse overall.

Does poor sleep cause dementia?

The answer from the data is that we cannot say it causes dementia. Cohort studies link insomnia, excessive daytime sleepiness, and both short and long sleep to higher later risk of cognitive decline. But sleep changes can also be an early sign of brain disease already developing, so the relationship is associative, not proven cause and effect.