What it is:Repeated trouble falling asleep, staying asleep, or waking too early, with daytime impairment as the actual complaint that brings someone in. An occasional bad night is normal. Insomnia is when the pattern sticks around and starts wrecking function.
The core problem:Sleep is normally a switch: the brain's wake-promoting circuits shut off and GABA-driven inhibition takes over. In insomnia, that switch won't fully flip. The patient stays hyperaroused (physiologically, cognitively, or both) even when exhausted.
What you do about it:Find and fix secondary causes first, then start CBT-I regardless of cause, since it's the only intervention with durable benefit. Medication is a short-term bridge, not the plan.
Every drug class in this chapter works one of three ways: boost the brake(GABA, via benzodiazepines and Z-drugs), cut the gas(block orexin wake-signaling, via DORAs), or reset the clock(melatonin receptor agonism, via ramelteon). None of them "induce" sleep from nothing. Keep that framework and the whole drug list stops being an arbitrary memorization exercise.
Two questions matter before you do anything else: how long has this been going on, and is something else actually driving it.
| Type | Duration | Frequency | Typical driver |
|---|---|---|---|
| Short-term (acute) insomnia | <3 months | Variable | Identifiable stressor: a death, a new diagnosis, jet lag, a big deadline |
| Chronic insomnia disorder | ≥3 months | ≥3 nights/week | Often no single trigger; may have outlived the original stressor |
Chronic insomnia disorder = difficulty initiating or maintaining sleep, or early morning awakening, plus daytime impairment, occurring ≥3 nights per week for ≥3 months. Miss any one piece (the frequency, the duration, or the daytime impairment) and it's not chronic insomnia by definition, it's just poor sleep.
Only about 5%of the general population has primary insomnia (no identifiable secondary cause). Everyone else's insomnia is downstream of something: a psychiatric condition, another sleep disorder, a drug, or a habit. Comorbid psychiatric illness is the single biggest risk factor: roughly 40% of patients with recurrent insomniameet criteria for at least one mental disorder, versus about 15% of people without sleep complaints. Depression is the most common comorbidity, anxiety a close second. Older age and female sex also raise risk.
Your course material uses a mnemonic to force a differential before reaching for a hypnotic: Stress (anxiety disorder, PTSD, new meds), Obstructive sleep apnea (HTN, BMI, neck circumference, daytime fatigue), Urination (BPH, overactive bladder), Restless legs (check ferritin, ask about SSRIs or dopamine blockers), Drugs (caffeine, alcohol, stimulants), Routine (exercising right before bed), Environment (light, noise, screens), Anxiety, Mood (PHQ-9, undiagnosed bipolar), Subjective sleep-related anxiety. Treating the hypnotic before treating the cause is the classic miss.
Sleep and wakefulness are opposite ends of one continuum, controlled by an ascending arousal systemrunning from the brainstem up through the hypothalamus to the cortex. That system uses several neurotransmitters to keep you awake: norepinephrine, acetylcholine, serotonin, dopamine, histamine, and orexin/hypocretin(a wake-promoting peptide made in the hypothalamus). GABA does the opposite: it's the brain's main inhibitory transmitter, and it's what turns arousal down and lets sleep happen.
Independent of arousal chemistry, a biological clock decides whenyou should be sleepy. Light hits the retina, travels the retinohypothalamic tract, and syncs the suprachiasmatic nucleus (SCN)in the hypothalamus, the body's master clock. The SCN drives the pineal gland to release melatonin(synthesized from serotonin) primarily at night. Melatonin doesn't sedate directly, it signals "it's nighttime" to MT1/MT2 receptors and helps set the timing of sleep onset. Cortisol runs the opposite rhythm, rising toward morning to help you wake.
| Sleep stage | What's happening | Why it matters clinically |
|---|---|---|
| N1 (light) | Transition from wake, easily aroused, occasional muscle jerks | Fragmented by insomnia, OSA, RLS; increases daytime fatigue |
| N2 (light) | Sleep spindles and K-complexes on EEG, further drop in muscle tone | Memory consolidation happens here; disruption impairs cognition |
| N3 (deep, slow-wave) | Delta waves, hardest to wake someone from | The physically restorative stage; reduced by insomnia, OSA, benzodiazepines |
| REM | Brain activity resembles wakefulness, vivid dreaming, but the body is paralyzed except eyes and breathing muscles | Emotional regulation and learning happen here; most hypnotics suppress it to some degree |
Read the drug classes straight off the arousal-vs-clock model. Benzodiazepines and Z-drugspotentiate GABA, the brake. Suvorexant, lemborexant, and daridorexantblock orexin receptors, cutting the gas on wake-signaling instead of stepping on the brake. Ramelteondoesn't touch GABA or orexin at all, it just tells the clock it's nighttime via MT1/MT2 agonism. Doxepin, trazodone, and mirtazapineborrow sedating antihistamine or antiserotonergic side effects from drugs built for something else. Once you sort every agent into one of those four buckets, the chapter organizes itself.
Chronic insomnia is increasingly framed as a hyperarousal state: patients with chronic insomnia show higher metabolic rate, higher cortical activity, and higher cortisol even during attempted sleep, compared to good sleepers. That's why sleep hygiene alone rarely fixes chronic insomnia. You're not dealing with a bad habit, you're dealing with a nervous system that has learned to treat the bed as a place to be alert. That's exactly the mechanism CBT-I's stimulus control component targets.
The complaint is rarely just "I can't sleep." It's the daytime fallout that actually brings people to clinic.
| Category | Findings |
|---|---|
| Nighttime | Trouble falling asleep, frequent awakenings, waking earlier than desired, feeling that sleep isn't restorative |
| Daytime cognitive/mood | Fatigue, poor concentration, irritability, low mood |
| Psychological | Distress or worry specifically about the lack of sleep, which itself becomes a driver (sleep-related anxiety) |
A single bad night after a stressful day is nota diagnosis, it's normal physiology. What separates a diagnosable disorder from an ordinary bad week is the combination of frequency (≥3 nights/week), duration (≥3 months), and daytime impairment. All three, not just one.
Insomnia is a clinical diagnosis built from history, not a lab value. There's no confirmatory blood test.
Even though insomnia itself has no diagnostic lab, you still work up contributors: TSH (hyperthyroidism, or iatrogenic overreplacement with levothyroxine, both cause anxiety and insomnia), ferritin (low iron worsens restless legs, which fragments sleep), and a medication reconciliation for anything that raises arousal (stimulants, some SSRIs, corticosteroids, decongestants) or that's dosed at the wrong time of day.
Cognitive behavioral therapy for insomnia (CBT-I) is first-line for chronic insomnia, full stop, ahead of any drug. It has a large effect size (Cohen's d roughly 0.8 to 1.0 on sleep efficiency and ISI) and the benefit holds up for up to 2 yearsafter treatment ends. Medication's effect size is smaller (roughly 0.5 to 0.8), fades faster once stopped, and comes with dependence and rebound risk that CBT-I doesn't carry.
Limits time in bed to roughly match actual sleep time, building sleep drive so the bed becomes reliably associated with sleeping instead of lying awake.
Bed is for sleep only. Go to bed only when sleepy, get out of bed if you can't sleep, no screens or work in bed.
Identifies and challenges catastrophic beliefs about sleep loss ("I'll be useless tomorrow if I don't sleep 8 hours") that themselves fuel hyperarousal.
Regular schedule, no caffeine or alcohol near bedtime, plus breathing or progressive muscle relaxation to lower physical arousal.
Sleep hygiene by itself is not enoughfor chronic insomnia, it's only one component of CBT-I. This is a repeated trap: a question offering "sleep hygiene alone" as the answer for chronic or comorbid insomnia (including insomnia with alcohol use disorder) is testing whether you know the difference between the whole package and one piece of it. Digital or app-based CBT-I is validated and considered effective when in-person therapy isn't accessible.
Goals:improve sleep quality and continuity, improve next-day function, and do it without creating a new problem (dependence, falls, next-day sedation). Pharmacotherapy is reserved for when CBT-I is unavailable, insufficient, or the patient needs a short-term bridge while therapy takes effect.
| Scenario | Initial approach | If inadequate response |
|---|---|---|
| Short-term insomnia(identifiable stressor, <3 months) | Sleep hygiene plus a short-acting BZDRA or ramelteon | Reassess for an emerging chronic pattern or an untreated secondary cause |
| Chronic insomnia(≥3 nights/week, ≥3 months) | Sleep hygiene, CBT-I, and treat any underlying medical or psychiatric condition; add a BZDRA or DORA only if still needed | Trazodone, a different BZDRA, or a DORA |
Hypnotics are meant for lowest effective dose, shortest duration, generally ≤4 weeks. Long-term nightly use is not the goal for any BZDRA or Z-drug. If a patient has been on zolpidem nightly for years, that's a medication reconciliation flag, not a stable regimen to leave alone.
Grouped by mechanism, since that's how you should be memorizing this list anyway.
| Drug (Brand) | Usual Adult Dose | Half-Life | Key Notes |
|---|---|---|---|
| Benzodiazepine receptor agonists - traditional benzodiazepines | |||
| Triazolam (Halcion) | 0.125–0.25 mg | ~2 h | Highly lipophilic, fast in and out; CYP3A4 inhibitors (erythromycin, nefazodone, fluvoxamine, ketoconazole) raise levels significantly |
| Estazolam (ProSom) | 1–2 mg | 12–15 h | Intermediate acting |
| Temazepam (Restoril) | 15–30 mg | 10–15 h | Metabolized by conjugation, no active metabolites, one of the preferred BZDRAs in older adults |
| Flurazepam (Dalmane) | 15–30 mg | 8 h parent (active metabolite 47–100 h) | Long effective duration from the active metabolite; avoid in older adults |
| Quazepam (Doral) | 7.5–15 mg | 39 h | Long acting; avoid in older adults |
| Benzodiazepine receptor agonists - nonbenzodiazepine "Z-drugs" | |||
| Zolpidem (Ambien) | 5 mg (women, older adults, hepatic impairment) or 5–10 mg (men) | 2–2.6 h | Duration 6–8 h; take on an empty stomach; sleep eating reported; minimal effect on next-day psychomotor performance at labeled doses |
| Zaleplon (Sonata) | 10 mg (5 mg in older adults) | ~1 h | Fastest onset, shortest duration; helps sleep onsetonly, does not reduce nighttime awakenings or increase total sleep time |
| Eszopiclone (Lunesta) | 2–3 mg | ~6 h | Duration up to 6 h; the one Z-drug approved for nightly use up to 6 months; unpleasant taste and dry mouth are common |
| Dual orexin receptor antagonists (DORA) | |||
| Suvorexant (Belsomra) | 10–20 mg (max 20 mg/day) | ~12 h | Take within 30 min of bed with ≥7 h remaining before waking; caution in depression (dose-dependent worsening, suicidal ideation) |
| Lemborexant (Dayvigo) | 5–10 mg | ~17–19 h | Same timing rule and depression caution as suvorexant |
| Daridorexant (Quviviq) | 25–50 mg | ~8 h | Newest DORA (2022); same class precautions |
| Melatonin receptor agonist | |||
| Ramelteon (Rozerem) | 8 mg at bedtime | 1–2.6 h | MT1/MT2 selective; not a controlled substance; no acute drowsiness like the BZDRAs; safe option in COPD or sleep apnea since it doesn't suppress respiratory drive |
| Sedating antidepressants (off-label unless noted) | |||
| Doxepin, low dose (Silenor) | ≤6 mg | - | The one antidepressant with an FDA-approved insomnia indication(sleep maintenance); H1 antagonist mechanism; needs 7–8 h of remaining sleep opportunity |
| Trazodone | Off-label, dose individualized | - | Sedating via 5-HT2A antagonism and antihistamine effect; algorithm's fallback option; AEs include oversedation, orthostatic hypotension (alpha-adrenergic blockade), and rare priapism |
| Mirtazapine | Off-label, dose individualized | - | H1 antagonist mechanism; caution, it can worsen restless legs syndrome |
| OTC / herbal | |||
| Diphenhydramine / doxylamine | OTC labeling | Long | First-generation antihistamines; long half-life drives next-day sedation; anticholinergic; avoid in older adults |
| Valerian | 300–600 mg | - | Herbal, available without a prescription; efficacy data are lacking |
BZDRAs are the most commonly used drugs for insomnia and include both the traditional benzodiazepines and the newer Z-drugs, both acting at the GABAAreceptor. The FDA requires class-wide labeling for anaphylaxis, facial angioedema, and complex sleep behaviors(sleep driving, sleep eating, sleepwalking with no memory of the event afterward).
Benzodiazepines carry sedative, anxiolytic, muscle relaxant, and anticonvulsant properties because they're not selective, they hit GABAAreceptors broadly. They increase stage 2 sleep and decrease REM and deep (delta) sleep. Z-drugs are more selective for GABAAreceptors containing the alpha-1 subunit, which is why they carry less of the anxiolytic/muscle relaxant baggage and, per your course material, less REM suppression than traditional benzodiazepines.
Flurazepam and quazepamshould not be first-line because of their long effective half-lives (driven by active metabolites), which accumulate and raise the risk of falls and hip fractures, especially in older adults. If you must use a traditional benzodiazepine in an elderly patient, lorazepam, oxazepam, and temazepam("LOT") are preferred because they're cleared by conjugation with no active metabolites. Not all three are FDA-approved specifically for insomnia (temazepam is; lorazepam and oxazepam are used off-label), so know which one the question is actually asking about.
Rebound insomniaon discontinuation is minimized by using the lowest effective dose and tapering rather than stopping abruptly. Triazolamis worth remembering by name because of its drug interaction profile: it's cleared quickly due to high lipophilicity, but erythromycin, nefazodone, fluvoxamine, and ketoconazole all inhibit its clearance and can push levels up substantially.
Zolpidemis comparable in efficacy to the benzodiazepines with minimal effect on sleep stage architecture; take it on an empty stomach since food delays absorption. Zaleplon's ~1 hour half-life makes it useful purely for sleep-onset trouble, it will not help someone whose complaint is 3am awakenings. Eszopicloneis the one labeled for extended nightly use (up to 6 months), which distinguishes it from the "short-term only" framing of the rest of the class.
Suvorexant, lemborexant, and daridorexant block orexin A and orexin B receptors (OX1R/OX2R). Orexin is a wake-promoting neuropeptide from the hypothalamus; block it and the brain's "stay awake" signal turns off, rather than a sedative being layered on top. That mechanistic difference is why the counseling point is different too: patients should take a DORA within 30 minutes of going to bed, with at least 7 hours before they plan to wake, since these drugs are working by removing wakefulness drive for the hours they're active, not by knocking someone out immediately.
DORAs can worsen depression and trigger suicidal thinking in a dose-dependentmanner. This is not a class to reach for casually in a patient with underlying depression without weighing that risk.
They're Schedule IV, like the BZDRAs, so dependence potential and controlled-substance handling still apply. Common adverse effect is next-day sedation; rare narcolepsy-like symptoms (sudden muscle weakness) have been reported, which makes sense given the mechanism directly opposes the pathway that fails in narcolepsy.
Ramelteon is selective for MT1 and MT2 melatonin receptors, the same receptors endogenous melatonin uses to synchronize the SCN's clock signal. Because it's working through circadian timing rather than direct sedation, it doesn't cause the acute "hit by a truck" drowsiness some patients feel with a BZDRA, and it's not a controlled substance, no abuse liability, no DEA scheduling paperwork.
It's well tolerated; adverse effects are limited to headache, dizziness, and somnolence. Because it doesn't suppress respiratory drive the way GABAergic agents can, it's specifically called out as effective and appropriate in patients with COPD or sleep apnea, populations where you'd otherwise be nervous about a hypnotic.
These are antidepressants first, hypnotics second, borrowed for their off-target sedating properties. Low-dose doxepin (≤6 mg)is the exception, it's actually FDA-approved for sleep maintenanceinsomnia at this dose, working as a potent H1 (histamine) receptor antagonist without the anticholinergic burden seen at antidepressant doses.
Trazodoneworks through 5-HT2A antagonism and antihistamine activity, and it shows up explicitly in the treatment algorithm as an option when a first BZDRA or DORA hasn't worked. Its adverse effect profile is why it's not first choice: oversedation, orthostatic hypotension from alpha-adrenergic blockade, dizziness, and rare but notable priapism. It also carries serotonergic risk when stacked with other serotonergic drugs.
Mirtazapinesedates mainly through H1 antagonism as well. The high-yield interaction to remember: it can worsen restless legs syndrome, so before you blame "just more insomnia" in a patient on mirtazapine, ask about an urge to move the legs.
Diphenhydramine and doxylamineare first-generation antihistamines and the most common OTC "sleep aids." Their long half-lives are exactly the problem: patients get next-day grogginess, and the anticholinergic burden (dry mouth, constipation, urinary retention, confusion) is a real concern in older adults, where these agents are explicitly flagged to avoid.
Valerianis an herbal product taken at 300 to 600 mg. It's popular and unregulated, but the efficacy data simply aren't there. Don't oversell it to patients who ask.
This is the population where insomnia pharmacotherapy causes the most collateral damage. Benzodiazepines, Z-drugs, first-generation antihistamines, and TCAs are all flagged by Beers Criteriafor older adults because of increased falls, cognitive impairment, and (for the GABAergic agents) dependence risk.
Using alcohol to fall asleep is common and counterproductive, it fragments sleep architecture later in the night and raises relapse risk. Benzodiazepines should be avoidedhere because of dependence potential and dangerous additive CNS depression with alcohol. CBT-I is first-line with strong evidence for both insomnia relief and relapse prevention. If pharmacotherapy is needed, gabapentin, mirtazapine, or quetiapine have been used, though evidence is limited; melatonin has shown no benefit over placebo in this population at studied doses.
Two recurring teaching points from your practice cases: (1) an overreplaced levothyroxine dosecan cause iatrogenic hyperthyroidism presenting as anxiety and insomnia, and correcting the dose is the actual fix, not adding a sleep aid on top. (2) A diuretic dosed at night can cause nocturia that looks like "trouble staying asleep." Both are medication reconciliation catches, not pharmacotherapy decisions.
If a patient is already dependent, don't just stop the drug. Per ASAM guidance, for low-to-moderate withdrawal risk, taper by reducing the dose 10 to 25% every 1 to 2 weeks, adjusting pace to tolerance. Abrupt discontinuation risks a withdrawal syndrome that can include rebound anxiety and insomnia far worse than the original complaint.
| Parameter | When | Watching for |
|---|---|---|
| Effectiveness & adherence | After about 1 week of therapy | Improved sleep onset/maintenance, and whether the patient is actually doing the nonpharmacologic recommendations, not just taking the pill |
| Sleep diary | Ongoing, daily | Time to bed, awakenings, naps, medications taken, subjective sleep quality |
| ISI or similar validated tool | Baseline, then periodically | Trending severity objectively rather than relying on "I feel about the same" |
| Adverse effects | Every visit while on a BZDRA, Z-drug, or DORA | Next-day sedation, cognitive slowing, complex sleep behaviors, falls (especially older adults) |
| Ongoing need | Regularly, especially past 4 weeks | Whether the drug is still necessary, or has quietly become a long-term habit instead of a short-term bridge |
| Mood | Each visit if on a DORA | Dose-dependent worsening of depression or emergence of suicidal ideation |