Stage 4: Sleep as Cognitive Intervention
Concept 3 of 6
L4.3

Insomnia Across the Lifespan

Chronic insomnia is a cognitive risk factor responsive to treatment. CBT-I first; pharmacology in layered sequence; benzodiazepines and Z-drugs to be used sparingly.

The seasoned approach

Layered approach to chronic insomnia in the longevity-psychiatry frame. The discipline is to use the lowest-cognitive-cost intervention that produces consolidated restorative sleep, not the fastest-acting sedative.

  1. Layer 1 — Always first — CBT-I
    Cognitive Behavioral Therapy for Insomnia. In-person if available; digital CBT-I (Somryst, Sleepio, SHUTi) if not. Six to eight sessions. The behavioral foundation is what produces durable change. Most patients labeled "refractory" have not actually had CBT-I.
  2. Layer 2 — Trazodone — the workhorse
    25–100mg qhs. Cheap, generic, no respiratory depression, weight-neutral, minimal hangover, useful with comorbid depression. Watch for orthostasis in elderly, rare priapism. The default first pharmacological agent for most patients.
  3. Layer 3 — Mirtazapine — when appetite or anxiety is part of the picture
    7.5–15mg qhs. The dose paradox: lower is more sedating because H1 antagonism dominates until NE release overtakes it at higher doses. Useful in patients with poor appetite, weight loss, anxiety component, or post-stroke depression with weight loss. Watch weight gain (sometimes desired), morning hangover, RLS, rare neutropenia.
  4. Layer 4 — Low-dose doxepin — sleep maintenance, elderly-friendly
    3–6mg qhs. Highly H1-selective at low doses, minimal anticholinergic burden at these doses (a clinically meaningful distinction from doxepin at antidepressant doses). Approved for sleep maintenance. Brand Silenor is expensive; compounded generic is comparable and often covered. Useful in elderly patients where anticholinergic burden is a concern at higher doses of other agents.
  5. Layer 5 — DORAs — newer, especially elderly
    Suvorexant, lemborexant, daridorexant. Block orexin signaling at the wake-promoting nucleus rather than potentiating GABA. Better safety profile in older adults than benzodiazepines or Z-drugs — less fall risk, less cognitive impairment, less dependency. Expensive; insurance often refuses without step therapy. The most significant pharmacological advance in sleep medicine in two decades.
Special situations
  • PTSD with nightmares: Add prazosin titrated to BP tolerance (typically start 1mg qhs, titrate weekly to 5–15mg). Treats nightmares specifically; does not replace the sleep agent.
  • ADHD on stimulants: Ensure last stimulant dose is early enough. Consider switching to short-acting evening or eliminating afternoon dose. Insomnia in patients on long-acting evening stimulants is iatrogenic until proven otherwise.
  • Perimenopausal or postmenopausal women: Consider hormonal contribution. Hot flash-driven sleep disruption may respond to hormonal interventions (Stage 19). Sleep agents alone are incomplete treatment for hormonally-driven insomnia.
  • Comorbid depression: Treat depression aggressively first; insomnia often resolves with durable depression remission. The depression-insomnia loop frequently breaks at the depression end.
Generally avoid
  • Chronic benzodiazepines — cognitive cost, dependency, fall risk in elderly, associated with elevated dementia risk in observational studies.
  • Chronic Z-drugs (zolpidem, eszopiclone, zaleplon) — similar concerns, similar associations, often continued for years without reassessment.
  • Quetiapine 25–50mg for sleep without psychiatric indication — metabolic cost, anticholinergic burden, limited evidence for efficacy at sleep doses. Despite widespread use, the risk-benefit calculation is unfavorable.
  • Diphenhydramine and OTC anticholinergic sleep aids — chronic anticholinergic burden is itself associated with elevated dementia risk; the cost is real and accumulates.

The chief-resident note: Most patients labeled "refractory insomnia" have not actually completed CBT-I. Most patients on chronic quetiapine for sleep started it for reasons that no longer apply. Most patients on chronic Z-drugs in late middle age are accumulating cognitive cost they have not been informed about. Audit the medication list before adding to it. Slow tapers (months, not weeks) succeed where fast tapers fail.

Warm cream-tinted manuscript page, deep slate margin annotations, deep night-blue palette. The behavioral loop that maintains chronic insomnia — anticipatory anxiety about sleep, conditioned arousal in the bedroom, compensatory behaviors that worsen the cycle. Margin clusters on the CBT-I interventions that break each link.

Chronic insomnia — difficulty initiating sleep, maintaining sleep, or experiencing restorative sleep for three or more nights weekly across three months or more — affects approximately 10–15% of adults at any given time. The figure rises with age, reaches near-30% in adults over 65, and is substantially higher in psychiatric populations. Chronic insomnia is not merely a symptom of an underlying condition; it is independently associated with elevated risk of depression, anxiety, cardiovascular disease, and — relevant to this volume — cognitive decline and dementia. The clinical task is to treat insomnia as the cognitive risk factor that it is, not as a complaint to be silenced with a hypnotic.

Cognitive Behavioral Therapy for Insomnia (CBT-I) is the first-line treatment, supported by stronger evidence than any pharmacological intervention. CBT-I addresses the cognitive and behavioral factors that maintain insomnia: sleep restriction, stimulus control, cognitive restructuring of catastrophic thinking about sleep, sleep hygiene, and relaxation training. Effect sizes are comparable to or larger than those of hypnotic medications, and benefits persist after treatment ends — unlike pharmacological interventions, which generally require ongoing use. The clinical problem is access: CBT-I requires trained providers, takes six to eight sessions, and is unevenly covered by insurance. Digital CBT-I platforms (Somryst, Sleepio, SHUTi) have closed some of this access gap with FDA-cleared options that produce comparable outcomes for many patients.

The pharmacological landscape is layered and should be deployed accordingly. Trazodone (25–100mg qhs) is the workhorse first-line pharmacological agent — cheap, generic, no respiratory depression, weight-neutral, minimal hangover, and useful in patients with comorbid depression. Mirtazapine (7.5–15mg qhs) is useful when appetite stimulation or anxiolysis is also desired; its dose paradox (more sedating at lower doses) is clinically important. Low-dose doxepin (3–6mg) is selective for histamine-1 receptors at low doses and is approved specifically for sleep maintenance; the cost-versus-effect calculation depends on whether brand or compounded generic is used. The dual orexin receptor antagonists (DORAs) — suvorexant, lemborexant, daridorexant — are newer agents with favorable safety profiles, particularly in older adults, and represent the most significant pharmacological advance in sleep medicine in two decades.

Benzodiazepines and Z-drugs deserve specific caution. Both classes work — they reliably produce sleep — but the cognitive cost over chronic use is substantial. Benzodiazepines have been linked in observational studies to elevated dementia risk, particularly with chronic use, and the cognitive impact is mechanistically plausible (GABA-mediated effects, residual sedation, impaired memory consolidation). Z-drugs (zolpidem, eszopiclone, zaleplon) carry similar concerns. The clinical principle is that both classes are appropriate for short-term use in acute insomnia or transient situations; their use as chronic, indefinite sleep aids is not consistent with longevity-psychiatry principles. The patient on chronic zolpidem for years should be reevaluated for alternatives.

Quetiapine 25–50mg for sleep is widespread and problematic. Low-dose quetiapine is widely prescribed off-label for insomnia despite limited evidence for efficacy at sleep-dose ranges and meaningful concerns about metabolic effects (weight gain, glucose dysregulation, lipid effects) and anticholinergic burden (cognitive cost). For psychiatric indications that justify quetiapine, it has a clear role. As a sleep medication in patients without psychiatric indication for it, the risk-benefit calculation is unfavorable for longevity psychiatry. The patient on quetiapine 25mg for sleep — a common scenario — should be evaluated for alternatives.

Special populations deserve specific considerations. Elderly patients are particularly sensitive to anticholinergic burden, falls risk from sedatives, and the cognitive cost of chronic hypnotic use; DORAs and low-dose doxepin are typically preferred. Patients with PTSD often benefit from prazosin for nightmares (covered in detail in L11.3). Patients with comorbid depression may have insomnia that responds primarily to depression treatment; durable remission of depression often resolves the insomnia. Perimenopausal and postmenopausal women have insomnia patterns driven partly by hormonal transitions; the conversation may need to include hormonal evaluation (covered in Stage 19). ADHD patients on stimulants need attention to the timing of the last stimulant dose.

The longevity-psychiatry approach to chronic insomnia is layered, patient — many "refractory" patients have not actually had CBT-I — and conservative about chronic pharmacology. The aim is restorative sleep that supports cognitive trajectory across decades, not just nightly sedation.

Editorial illustration of the layered pharmacological approach — trazodone, mirtazapine, doxepin, DORAs — arranged by cognitive cost and clinical situation. Margin notes on which patients fit which layer.
The anchor

Chronic insomnia is a cognitive risk factor, not a complaint to be silenced. CBT-I is first-line; pharmacology is layered; chronic benzodiazepines, Z-drugs, and low-dose quetiapine for sleep are inconsistent with longevity-psychiatry principles.

Painterly editorial illustration of the agents to use sparingly or avoid — chronic benzodiazepines, chronic Z-drugs, low-dose quetiapine for sleep, OTC anticholinergics. The cognitive cost of each rendered visually, with the longevity-psychiatry alternative noted in the margin.
Prove it

A 67-year-old patient has been on zolpidem 10mg nightly for eight years. She sleeps adequately on it, has tried to stop unsuccessfully (rebound insomnia), and asks whether she should continue. From a cognitive-protection standpoint, what is the conversation and the plan?

This connects to

Locked concepts unlock as you reach them on the path.

Back