You used to sleep eight hours and feel rested. Now you wake up at 4am, can’t fall back asleep, and feel slower all day. The body did change — but most of the complaint isn’t “just aging.” It’s three specific things, two of which are fixable in a single doctor visit, and one of which costs nothing at all.
If you’re over 50, you’ve probably noticed three things at once: it takes longer to fall asleep, you wake up earlier than you want to, and your sleep feels lighter when you do get it. All three are real and partly biological. But “partly biological” doesn’t mean “unfixable” — and a substantial chunk of what people call “just getting older” is actually one of three under-diagnosed conditions hiding in plain sight.
Two real biological shifts that affect almost everyone:
Both of these are normal aging. Neither, by itself, should make you feel exhausted, foggy, or unable to function. If you do, something else is going on — usually one of three things in the next section.
Senaratna et al., Sleep Medicine Reviews, 2017. Systematic review of OSA prevalence: roughly 30–40% of adults over 60 have obstructive sleep apnea (defined as an apnea-hypopnea index of 15 or more). The majority are undiagnosed. Marin et al., Lancet, 2005 showed untreated severe OSA approximately tripled the rate of fatal and non-fatal cardiovascular events compared to treated OSA over a 10-year follow-up.
If you snore loudly, wake unrefreshed, have hard-to-control blood pressure, or your bed partner says you stop breathing — ask your doctor about a sleep study. Home sleep tests now cost $200–$400, are covered by Medicare, and don’t require an overnight visit to a lab. If you do have OSA, CPAP treatment (or in some cases an oral appliance) remains one of the highest-leverage anti-aging interventions available to most adults over 50.
One important honest caveat: the cognitive case for CPAP is more nuanced than the cardiovascular case. A 2025 multicenter RCT (PubMed 39998447) found 12 months of CPAP did improve cortical thickness and default-mode-network connectivity on MRI — but did not improve MoCA cognitive scores. This is consistent with the older SAVE trial pattern: structural and physiological improvements show up before symptomatic cognitive ones do, and the “CPAP prevents dementia” framing is still hopeful rather than proven. A 2025 NHATS preprint (not yet peer-reviewed) suggests CPAP-adherent older adults show slower 10-year cognitive decline; that’s the kind of result we want, but it’s not yet locked in.
TSH levels creep up with age in many adults. Some end up in the “sub-clinical” range — high enough to flatten energy, slow metabolism, and disrupt sleep, but not high enough to trigger an automatic diagnosis. Most physicians don’t look unless you specifically ask, especially if your other thyroid numbers (free T4) look acceptable. Get TSH and free T4 checked on your next physical.
After 50, stomach acid production declines, which reduces B12 absorption from food. Sub-clinical B12 deficiency mimics “cognitive aging” — fatigue, mild brain fog, sometimes mood changes — and can disrupt sleep quality. The blood test is cheap, the fix is cheaper (sublingual B12 or injection if absorption is severely compromised), and it’s wildly under-diagnosed.
If you’re a woman over 45 whose sleep started getting worse around the time of cycle changes or hot flashes, the relevant question is whether vasomotor symptoms (night sweats, hot flashes) are driving the insomnia. The 2024-2025 Mayo Clinic/KEEPS-Continuation data and current Endocrine Society guidance show that hormone therapy in recently menopausal women improves both subjective and actigraphy-measured sleep when vasomotor symptoms are the driver. Progesterone may add a small additional non-REM benefit beyond estrogen alone. A four-arm RCT (NCT06306404, “Sleeping Through Menopause”) is the first head-to-head test of hormone therapy vs CBT-I vs both vs control — results are pending. Ask a clinician familiar with women’s health, not a generalist; this conversation is more sophisticated than the WHI-era 2002 talking points still circulating.
What this means for you: Before assuming poor sleep is “just aging,” ask your doctor for: a TSH and free T4, a B12 level, a home sleep study or referral if you snore or feel unrefreshed, and (for women) an honest conversation about whether vasomotor symptoms are part of the picture.
Windred et al., Sleep, 2024. Analysis of ~88,000 UK Biobank participants tracked with wrist-worn accelerometers showed sleep regularity was a stronger predictor of all-cause mortality than sleep duration. People with the most variable sleep timing had substantially higher mortality risk than those with the most consistent timing, even after adjusting for duration. The simplest interpretation: when you sleep matters as much as how long, possibly more.
Sansom et al., Journal of Sleep Research, 2026, extended the finding to a different outcome in the over-50 population specifically: in middle-aged and older community adults, higher Sleep Regularity Index scores predicted meaningfully better health-related quality of life. A separate 2025 stroke-survivor replication (Scientific Reports) found higher SRI also predicted better recovery and lower depression in clinical populations. The signal is not just UK Biobank-specific.
The translation: same bedtime, same wake time, every day. Including weekends. This is the single sleep habit with the strongest large-cohort mortality evidence and it costs nothing.
Within 30 minutes of waking, get outdoor light for 10–15 minutes (less on bright days, more if overcast). This anchors your circadian rhythm to the actual day-night cycle and helps both sleep onset that night and overall mood. The evidence isn’t RCT-grade for “morning light cures insomnia,” but the circadian-anchoring biology is solid (Mead, Environmental Health Perspectives, 2008, and many subsequent reviews) and the practical effect for most adults is meaningful within a week or two.
Trauer et al., Annals of Internal Medicine, 2015. Meta-analysis of CBT-I trials. CBT-I produced moderate-to-large improvements in sleep onset, sleep maintenance, and total sleep time that outperformed medication in long-term outcomes. The benefit is durable: most patients maintain gains years after treatment ends. It is the first-line treatment recommended by the American College of Physicians. The AASM’s 2024 Quality Measures Update reaffirmed CBT-I as the cornerstone of evidence-based insomnia care.
Ritterband et al., npj Digital Medicine, 2025. SHUTi OASIS — the strongest dedicated test of digital CBT-I in older adults to date. n=311, ages 55-95, 3-arm RCT (digital CBT-I, digital CBT-I plus stepped human support, patient education control). Both digital arms beat patient education on the Insomnia Severity Index at post-treatment, 6 months, and 12 months — with clinically meaningful remission rates. Older adults responded to a fully app-delivered intervention without needing in-person therapist contact. This is the trial that should retire the “CBT-I doesn’t work for old people who aren’t tech-comfortable” objection.
CBT-I is a structured 4–8 session program (in-person, via telehealth, or via well-validated apps like Sleepio, CBT-I Coach, or the newer SHUTi platform). The core moves are surprisingly simple: stimulus control (only use the bed for sleep), sleep restriction (consolidate your sleep window), and cognitive techniques for “but what if I don’t sleep” rumination. Apps now make it accessible without finding a therapist — and the 2025 SHUTi OASIS trial says they work in our age group specifically.
The least surprising piece of advice in this whole space, but the one most adults under-implement.
Neither of these is glamorous. Both work.
Two distinct categories worth understanding.
Melatonin helps with circadian-rhythm problems (jet lag, shift work, delayed sleep phase) more than it helps with general insomnia. For older adults whose endogenous melatonin production has declined, low-dose timed-release melatonin (0.3–1mg, 2–3 hours before bed) has modest evidence. The 3–10mg doses common in drugstore supplements often work less well, can produce next-day grogginess, and may disrupt sleep architecture. Lower is better in older adults.
Melatonin contamination flag. A CDC MMWR report (vol. 73, no. 9, 2024) documented over 260,000 pediatric melatonin ingestions tracked through US Poison Control Centers 2012-2021, with two deaths and a 530% increase — partly driven by the wild variability of actual melatonin content in supplements. A Canadian content-analysis study found supplements containing anywhere from -83% to +478% of the labeled dose. If you use melatonin, buy from a manufacturer with third-party testing (USP-verified or NSF Certified), not from the cheapest Amazon listing.
Sleeping pills — benzodiazepines (lorazepam, temazepam) and Z-drugs (zolpidem/Ambien, eszopiclone/Lunesta) — come with particular concerns after 50: next-morning fall risk, cognitive impairment, tolerance and dependence. The American Geriatrics Society’s Beers Criteria specifically flag these classes for caution in older adults. CBT-I, sleep regularity, and addressing the underlying conditions outperform these drugs in long-term outcomes.
The DORA class — the better pharmacologic option in 2026. Dual orexin receptor antagonists (suvorexant/Belsomra, lemborexant/Dayvigo, daridorexant/Quviviq) are now a meaningfully better option than Z-drugs for adults over 50 who need pharmacologic help.
Yeh et al., Translational Psychiatry, 2025. Network meta-analysis of 8 RCTs, n=5,198. Among DORAs, lemborexant 10 mg was best for sleep onset and daridorexant 50 mg was best for sleep maintenance. Critically: no tolerance, no rebound insomnia on discontinuation, no taper required (unlike Z-drugs and benzodiazepines which the 2023 Alliance for Sleep deprescribing guideline does require tapering for). The American Academy of Sleep Medicine’s 2024 Quality Measures Update added DORAs alongside CBT-I in the definition of evidence-based insomnia treatment.
This doesn’t make DORAs “safe” for casual use — they have side effects (next-day grogginess, abnormal dreams, rare REM-related effects) — but for an over-50 adult who genuinely needs medication, they are the class to ask your doctor about before Ambien or Lunesta.
The Over-50 Reverse-Aging Guide includes the complete sleep protocol — CBT-I sleep restriction calculator, the morning-light schedule, and the “what to ask your doctor” checklist for sleep apnea, B12, and thyroid.
See the Full GuideTwo things change. Sleep architecture shifts toward more light sleep and less deep sleep. Circadian rhythm tends to advance earlier. On top of that, three under-diagnosed conditions become much more common: obstructive sleep apnea, sub-clinical hypothyroidism, and B12 deficiency.
Keep a regular sleep schedule. Same bedtime, same wake time, every day including weekends. Sleep regularity has been shown to predict mortality even more strongly than sleep duration.
Yes if you snore loudly, your bed partner says you stop breathing, you wake unrefreshed, you have hard-to-control blood pressure, or you’re simply over 60. ~30–40% of older adults have OSA and most don’t know it. Home sleep studies are inexpensive and Medicare-covered.
It helps more with circadian-rhythm issues than general insomnia. Low-dose timed-release (0.3–1mg, 2–3 hours before bed) has modest evidence in older adults. Larger drugstore doses often work worse.
Benzodiazepines and Z-drugs (zolpidem, eszopiclone) carry particular risk over 50 — falls, cognitive impairment, dependence — and the AGS Beers Criteria flag them. The newer dual orexin receptor antagonists (DORAs) — suvorexant, lemborexant, daridorexant — are a meaningfully better class for over-50 use. A 2025 network meta-analysis (Yeh et al., Translational Psychiatry, n=5,198) identified lemborexant 10 mg as best for sleep onset and daridorexant 50 mg as best for maintenance, with no tolerance or rebound. CBT-I still wins on long-term outcomes.
Yes, including for older adults specifically. The 2025 SHUTi OASIS trial (n=311, ages 55-95) showed digital CBT-I beat patient education at post-treatment, 6 months, and 12 months — with or without stepped human support. App-delivered CBT-I works in this age group; the “older people can’t use the apps” objection no longer holds.
Most older adults do well on 6.5–7.5 hours. The 8-hour rule was a generalization that doesn’t apply uniformly. Consistency of timing matters more than hitting an exact duration.
— Scott Covert, 60, skeptic, not a physician. Sleep is the one anti-aging intervention where almost everyone has untapped upside. If you fix one thing on this page, fix sleep regularity. If you fix two, get the sleep apnea question off the table. Got a specific sleep question? Ask me.