Insomnia in older adults — three things confused into one word

What changes with age, and what does not

Sleep changes structurally across decades. Not all of what changes is disease. Worth being specific.

Sleep architecture changes — normal biology

Total sleep time decreases roughly ten minutes per decade after age thirty, stabilizing around age sixty-five. Sleep efficiency drops — the ratio of time-asleep to time-in-bed shrinks. Slow-wave sleep, the deepest stage, declines substantially across decades; at age seventy, it is roughly forty percent of what it was at twenty. REM sleep decreases more modestly. Sleep becomes more fragmented — more arousals, more brief awakenings, more difficulty returning to sleep after a wake. The circadian phase advances. Body temperature rhythm flattens, and melatonin secretion declines with age.

What this means in practice

An eighty-year-old who falls asleep at nine-thirty pm, wakes at five am, sleeps a total of six and a half hours with three to four brief awakenings, and feels reasonably rested during the day is having normal old-age sleep. This is not insomnia. It is not pathology. It is what age does to the sleep machine, and treating it as disease creates iatrogenic harm — usually in the form of pharmacology with a worse risk-benefit profile in this demographic than any other.

When it is disorder, not architecture

Clinical insomnia disorder applies the same DSM-5 criteria as in younger adults: sleep difficulty at least three nights per week for at least three months, with daytime impairment. Roughly fifteen percent of adults over sixty-five meet these criteria. Those patients deserve treatment. The remaining fifteen percent who report sleep problems but function adequately during the day need calibration of expectations more than they need a prescription.

An eighty-year-old waking at five am with six and a half hours of sleep is not sick. Calling it sick is how the pharmacology starts.

Comorbidity is the rule, not the exception

The defining feature of older-adult insomnia is that it is almost never primary. Four categories drive the majority of cases.

Medical

Chronic pain — arthritis, neuropathy, late-stage cardiovascular discomfort — is the most common upstream driver after age sixty-five. Cardiovascular disease (heart failure, atrial fibrillation) produces arousals. Obstructive sleep apnea is present in roughly thirty percent of adults over sixty-five and frequently undiagnosed. Nocturia — driven by BPH in men, overactive bladder in women, late-day diuretic dosing in both — disrupts continuous sleep. Restless legs syndrome rises in prevalence with age and is often missed without iron studies. GERD and thyroid dysfunction complete the medical picture.

Psychiatric

Depression is both cause and symptom of insomnia, and prevalence rises in later decades particularly around bereavement and chronic illness. Anxiety contributes similarly. Grief is a separate clinical entity from depression — grief responds to time and support rather than to antidepressants in most cases, and the sleep disturbance of grief is typically self-limiting over months.

Neurological

Pre-dementia sleep changes can precede cognitive symptoms by years (covered separately below). Parkinson's disease often presents with sleep disturbance as an early sign. REM sleep behavior disorder — a specific dream-enactment pattern — is a strong predictor of future Parkinson's or Lewy body dementia and is the differential most often missed in primary care.

Iatrogenic

Polypharmacy is the under-discussed engine. The average adult over sixty-five takes four to seven prescription medications. Several common classes disrupt sleep: diuretics taken too late in the day produce nocturia; beta-blockers can suppress endogenous melatonin; SSRIs commonly cause insomnia in the first six weeks; anticholinergic medications fragment sleep; steroids elevate cortisol. A complete medication review by a pharmacist or geriatrician familiar with sleep effects is often the highest-leverage single intervention available.

Older-adult insomnia is almost never primary insomnia. It is the symptom of something the medical workup has not yet found — a comorbidity, a medication side effect, sleep apnea masquerading as aging, or normal sleep architecture misread as disease. Treating it with a sleeping pill without workup is treating a check-engine light by putting tape over the dashboard.

The most common driver of older-adult insomnia is the medication for the other condition. Audit before adding.

The medication trap

Older adults are systematically over-medicated for sleep. The data is unambiguous and the pharmacy aisle has not adjusted.

Z-drugs and benzodiazepines

The Beers Criteria — geriatric medicine's standard reference for medications inappropriate in adults over sixty-five — explicitly recommends against chronic use of Z-drugs (zolpidem, eszopiclone, zaleplon) and benzodiazepines for sleep. Fall risk is elevated fifty to one hundred percent in users; hip fracture risk follows from falls, with associated mortality. Cognitive impairment and daytime confusion are common. Tolerance and dependence develop on routine prescribing schedules. Full comparison in our CBT-I versus sleeping pills piece.

Diphenhydramine and first-generation antihistamines

ZzzQuil, Tylenol PM, Benadryl-as-sleep-aid. Beers Criteria: avoid. Anticholinergic burden from long-term use is associated with elevated dementia risk in longitudinal data — Gray's 2015 JAMA Internal Medicine analysis found a hazard ratio around 1.54 in the highest exposure tertile. Falls, daytime confusion, urinary retention, dry mouth all rise in older users. Full ranking with safer substitutions in OTC sleep aids ranked.

Better options when pharmacology is genuinely needed

Low-dose doxepin (3 to 6 milligrams) is FDA-approved for insomnia and has the safer profile of any prescription option in this demographic. Cleanly leaves sleep architecture intact. First-choice when pharmacology is part of the plan.

Suvorexant and the orexin antagonists have less abuse liability and tolerable side-effect profiles. Reasonable second-line. More expensive and often not covered.

Trazodone at twenty-five to one hundred milligrams is widely used in geriatric practice off-label. Acceptable when other options are unavailable, with monitoring for orthostatic hypotension.

Low-dose melatonin (0.3 to 0.5 milligrams) is the right tool for circadian-aligned use cases — early-evening dosing for shifted phase. Not for general sleep onset. Full version in our melatonin deep-dive.

The Beers Criteria has explicitly warned against most over-the-counter sleep aids in older adults for over a decade. The pharmacy aisle has not received the memo.

CBT-I works in older adults

The most under-prescribed intervention in this demographic. Worth saying clearly.

Multiple randomized controlled trials and meta-analyses establish that CBT-I works as well in adults over sixty-five as in younger adults. Effect sizes for sleep efficiency improvement run d = 0.6 to 0.8, which is large in the clinical literature. The barrier is access, not efficacy — the population of clinicians trained in CBT-I is small, the subset who work routinely with older adults is smaller, and primary care physicians often default to prescriptions because pharmacology is faster to deliver than a referral that takes weeks to schedule.

Adaptations exist for the demographic. Stimulus control is fully applicable and arguably more important — the bed-arousal pairing is easier to consolidate in patients with smaller social and work schedules. Sleep restriction may need calibration in older adults, particularly those with fatigue-related fall risk; some clinicians use a less aggressive protocol with the same eventual outcome. Cognitive restructuring — challenging catastrophic thoughts about sleep — works well, particularly in the demographic that has internalized the cultural expectation of sleeping like a thirty-year-old.

Group CBT-I has reasonable evidence in this population and is often the most accessible format, since group programs scale better than one-on-one. Digital CBT-I (Sleepio, Somryst) is also evidence-based and is the cheapest path when geographic access to a clinician is limited.

CBT-I works as well at seventy as at thirty. The barrier is access, not biology. Most over-sixty-five patients have never been offered it.

The morning circadian advance — feature, not disorder

One of the most-misread aspects of older-adult sleep is the natural phase advance. Worth reframing because it produces a meaningful share of the unnecessary insomnia diagnoses in this demographic.

The circadian phase advances with age. By the seventies and eighties, the natural sleep window often runs nine pm to five am rather than eleven pm to seven am. This is normal biology, driven by changes in the suprachiasmatic nucleus and in melatonin secretion. Going to bed at nine and waking at five is biologically appropriate at seventy. Trying to sleep until seven by going to bed at midnight is fighting the clock — and losing.

The cultural script is that seven am is a healthy wake time. The script is age-inappropriate. An adult child who pushes a parent to sleep later because five am wakes seem too early is asking the parent to suppress a normal circadian shift. The right adjustment is the social schedule — earlier dinner, earlier evening wind-down, earlier bedtime — not the wake time.

Morning light exposure remains useful, but older adults tend to get it already by virtue of being up early; the dose is usually not the issue. Evening light avoidance becomes more important — paradoxically, older eyes are more sensitive to the phase-shifting effects of evening light, even though they transmit less light overall. For patterns where the timing problem is more pronounced, the relevant cluster reference is the circadian hub.

Sleep apnea — the under-diagnosed driver in this demographic

Worth its own section. Roughly thirty percent of adults over sixty-five have obstructive sleep apnea. Most are undiagnosed. Many were told they have insomnia.

OSA in older adults frequently presents as insomnia rather than as the classic snoring-plus-daytime-sleepiness picture more common in younger patients. The pattern: frequent awakenings, fragmented sleep, daytime fatigue, often without the dramatic gasping a partner would notice. In older women specifically, the diagnostic mismatch is large — OSA in this group presents as insomnia three times as often as it presents as snoring, and most clinicians do not have that pattern in their default differential.

Signs that warrant escalation to a sleep apnea workup: witnessed apneas reported by a partner (gasping or pauses in breathing), morning headaches, refractory hypertension despite multiple medications, atrial fibrillation, persistent daytime fatigue despite adequate time in bed. Any of these alongside sleep complaints justifies a home sleep apnea test or in-lab polysomnography.

CPAP, when prescribed appropriately, is well-tolerated in older adults — modern devices are quieter and mask options have expanded. The treatment effect when OSA is the dominant problem is large enough that the insomnia complaint often resolves entirely after several weeks of therapy. The barrier to treatment is usually identification, not adherence.

Sleep apnea in older women presents as insomnia three times as often as it presents as snoring. The diagnostic pattern is decades behind the data.

When sleep change signals neurological concern

Calibrated section, not alarmist. Some sleep changes precede neurological diagnoses by years. Knowing the pattern is useful; panic is not warranted by any single occurrence.

REM sleep behavior disorder

RBD is dream enactment — punching, kicking, vocalizing, sometimes injuring self or bed partner during REM sleep. The body is acting out what the brain is dreaming, because the normal REM-stage muscle paralysis has failed. RBD most often appears in adults over fifty and is a strong predictor of future neurodegenerative disease: sixty to eighty percent of patients with idiopathic RBD develop Parkinson's disease, dementia with Lewy bodies, or related conditions within ten to fifteen years of RBD onset. This is not occasional sleep talking or restless movement; RBD is dramatic and recognizable. A patient with this pattern warrants neurology evaluation, not for alarm but for monitoring and for the access it provides to early intervention as new treatments emerge.

Pre-dementia sleep changes

Increasingly fragmented sleep over two to five years, escalating daytime napping, sundowning (confusion or agitation in the evening hours), and progressive sleep efficiency decline can precede the cognitive symptoms of dementia. These are statistical signals at the population level rather than diagnostic at the individual level — a single year of fragmented sleep at seventy-eight is not a dementia diagnosis. The pattern across several years is the signal worth bringing to a clinician.

REM sleep behavior disorder is a punching, kicking, dream-enactment of the body. It is also a clinical signal the brain is decades earlier in a process the patient does not yet know is happening.

What to do this week

Three reader profiles, three protocols.

If you are an older adult experiencing sleep changes

First, calibrate. Sleep at seventy-five is not sleep at thirty. Some changes are normal and not pathology. Then audit medications with a pharmacist or physician — focus on diuretic timing, beta-blockers, anticholinergic drugs, and any over-the-counter sleep aids. Track sleep for fourteen days: total time-in-bed, total time asleep, awakenings, daytime impact. If daytime impact is minimal, you may not have insomnia disorder. Bring the tracking and the medication list to your next physician visit. Our treatments and substances hub is worth reading before any pharmacology conversation.

If you are an adult child concerned about a parent's sleep

Help with the tracking. Bring it to the physician with the medication list. Specifically ask about a sleep apnea screen, REM sleep behavior disorder, and a polypharmacy review. Do not push the seven-am wake — five am may be biologically appropriate at their age. If they are functioning well during the day and waking at five does not bother them, it is not a problem that needs solving. Regular moderate exercise is one of the most impactful single interventions for sleep in this demographic — the relevant article is exercise and sleep.

If you are a clinician

The differential is wide. Sleep apnea, depression, polypharmacy, RBD, pain, hyperthyroidism. Workup before pharmacology. CBT-I referral is appropriate when primary insomnia is identified; stimulus control and sleep restriction protocols apply with minor adaptations. The acute vs chronic insomnia piece is the cornerstone for staging and protocol selection.

Frequently asked questions