Why Lemborexant Matters
Lemborexant was the second dual orexin receptor antagonist (DORA) on the US market, approved in December 2019 for adults with trouble falling asleep, staying asleep, or both. Like suvorexant, it blocks both orexin receptors (OX1R and OX2R) and turns down the brain's wake drive instead of pushing GABA receptors. It has two pharmacologic features of its own: an active metabolite, M10, that binds orexin receptors about as tightly as the parent drug, and a long effective half-life of 17 hours at 5 mg and 19 hours at 10 mg.
Its trial record speaks directly to older patients. In SUNRISE-1 (1,006 women 55 and older and men 65 and older with sleep-maintenance insomnia), both doses beat placebo on polysomnographic sleep onset and maintenance and beat zolpidem ER 6.25 mg on wakefulness in the second half of the night. In healthy older volunteers, middle-of-the-night body sway was greater on zolpidem ER than on lemborexant. SUNRISE-2 showed the benefit held through 12 months, with no rebound insomnia or withdrawal on stopping.
Lemborexant has the strictest CYP3A rules of the class: avoid it with strong or moderate CYP3A inhibitors and with strong or moderate inducers, and cap it at 5 mg with weak inhibitors. Modafinil, fluconazole and fluvoxamine all fall into those groups. It is brand-only, with no US generic as of November 2026. It is Schedule IV; a DEA proposal from August 2026 to move all three DORAs to Schedule V has not been finalized.
Reach for lemborexant when an older adult, or anyone who should not take a benzodiazepine or a Z-drug, needs a hypnotic for both sleep onset and maintenance and may need it for months. Start at 5 mg, check the medication list for CYP3A drugs before you write it, and go to 10 mg only if 5 mg is tolerated and not enough, knowing that 10 mg adds somnolence and nightmares.
Part 1: Indications
FDA-Approved Use
- Insomnia in adults, characterized by difficulties with sleep onset and/or sleep maintenance
That is the whole label. Safety and effectiveness have not been established in patients under 18.
The Evidence
SUNRISE-1 (Rosenberg et al., JAMA Netw Open 2019). A 1-month polysomnography trial in women aged 55 and older and men aged 65 and older (median age 63, 86% women) with sleep-maintenance insomnia. Arms were placebo (n=208), zolpidem ER 6.25 mg (n=263), lemborexant 5 mg (n=266) and 10 mg (n=269). At nights 29 and 30, latency to persistent sleep fell to a geometric mean of about 15.5 minutes on 5 mg and 14.5 minutes on 10 mg versus 20 minutes on placebo, from a baseline near 33 minutes. Wake after sleep onset fell by 24 to 25 minutes more than placebo, and sleep efficiency rose by 7 to 8 percentage points more. Against zolpidem ER, lemborexant cut wakefulness in the second half of the night by a further 6.7 minutes (5 mg) and 8.0 minutes (10 mg). The label notes that the effects seen at the start of treatment were generally consistent with later time points. Keep the comparator in mind: zolpidem ER 6.25 mg is the low dose, so the trial says nothing about lemborexant against full-dose zolpidem.
SUNRISE-2 (Kärppä et al., Sleep 2020). A 12-month trial of 949 adults aged 18 to 88 (median 55, 28% aged 65 or older) using sleep diaries. Over the 6-month placebo-controlled period both doses beat placebo on subjective sleep onset latency, sleep efficiency and wake after sleep onset. At month 6, sleep onset latency was about 20 minutes on either dose versus 27 minutes on placebo, and subjective wake after sleep onset was 17.5 minutes (5 mg) and 12.7 minutes (10 mg) lower than placebo. In the second 6 months, patients kept on lemborexant held their gains through month 12 (Yardley et al., Sleep Med 2021).
The two doses land close together on the main endpoints. In both trials 5 mg and 10 mg produced similar results on the primary and key secondary measures, while somnolence, nightmares and discontinuations climbed at 10 mg (Part 6). The 2022 Lancet network meta-analysis (De Crescenzo et al.) found eszopiclone and lemborexant more effective than placebo in long-term treatment, at very low certainty, and it called lemborexant's safety data inconclusive.
Sleep architecture in older adults. A secondary analysis of SUNRISE-1 found lemborexant increased total sleep time and REM sleep, and shortened REM latency, more than placebo or zolpidem ER (Moline et al., J Clin Sleep Med 2021). The changes were small and their clinical meaning is unproven.
Off-Label / Investigational
- Irregular sleep-wake rhythm disorder in Alzheimer's dementia. One 4-week phase 2 trial (62 patients aged 60 to 90, doses 2.5 to 15 mg) found improvements in some actigraphy measures of nighttime restlessness and day-night distinction at some doses, with no worsening on the MMSE or ADAS-Cog (Moline et al., J Prev Alzheimers Dis 2021). These are proof-of-concept data from a small trial.
- Insomnia with depression. The registration trials excluded significant depression, anxiety and substance use. The main data in depression are a 12-week open-label study of 83 Japanese patients in a major or bipolar depressive episode: insomnia and depression scores improved, somnolence ran 10% to 25% in three of four groups and nightmares 4% to 12.5% (Takaesu et al., CNS Drugs 2026). With no control group, it cannot separate the drug from treatment of the depression. Remember the label's depression and suicidality warning.
- PTSD and substance use. There are no lemborexant trials in PTSD, and nightmares are one of its 10 mg side effects. In opioid use disorder, a safety study of 18 patients on buprenorphine-naloxone found no change in oxygen saturation or end-tidal CO2 with single daytime doses of 5 and 10 mg (Martin et al., Drug Alcohol Depend Rep 2024); it did not test whether sleep improved.
Who Is a Good Candidate?
- An older adult with sleep-maintenance insomnia, especially one with fall risk you would not hand a Z-drug or benzodiazepine
- A patient who may need a hypnotic for months (diary benefit held through 12 months of continuous use in SUNRISE-2)
- A patient with a substance use history for whom you want an option other than a GABAergic hypnotic (it is still Schedule IV; see Part 7)
- A patient whose Z-drug trial ended in a complex sleep behavior (switching class is reasonable, though DORA labels carry the same warning)
Who Is a Poor Candidate?
- Narcolepsy (contraindicated)
- Anyone on a strong or moderate CYP3A inhibitor or inducer that cannot be changed, including modafinil, fluconazole, fluvoxamine, carbamazepine or rifampin
- Severe hepatic impairment
- A patient who cannot reliably stay in bed 7 hours, or who drives for a living early in the morning
- A patient for whom cost or prior authorization will end treatment before it starts
Part 2: Mechanism
Orexin A and orexin B come from a small group of hypothalamic neurons and promote arousal by driving the wake-promoting transmitter systems: acetylcholine, dopamine, histamine, norepinephrine and serotonin. Lemborexant is a competitive antagonist at both receptors (IC50 6.1 nM at OX1R and 2.6 nM at OX2R), and blocking them is thought to suppress wake drive. Its major circulating metabolite, M10, binds the same receptors with similar affinity (IC50 4.2 and 2.9 nM), so the metabolite contributes to the effect.
The same mechanism explains the odd side effects. Orexin also keeps REM features out of wakefulness, and narcolepsy with cataplexy is a disease of orexin-neuron loss. Blocking orexin can produce sleep paralysis, hypnagogic or hypnopompic hallucinations and brief cataplexy-like leg weakness, and it is why narcolepsy is the one contraindication.
Pharmacokinetics. Peak levels come 1 to 3 hours after a dose. A high-fat, high-calorie meal delayed the peak by 2 hours, lowered Cmax by 23% and raised AUC by 18%, so a dose taken right after dinner starts working late. Exposure accumulates 1.5- to 3-fold at steady state. CYP3A4 does most of the metabolism, with a smaller CYP3A5 share. The effective half-life is 17 hours at 5 mg and 19 hours at 10 mg. Age, sex, race and body mass index made no clinically significant difference to exposure, unlike suvorexant, where obese women run higher levels.
In vitro, lemborexant and M10 can induce CYP3A, weakly inhibit CYP3A and induce CYP2B6. In people, midazolam exposure rose only about 1.1-fold, while bupropion exposure fell (Part 8). The QT interval was not prolonged to a clinically relevant degree at 5 times the maximum dose.
A 17- to 19-hour effective half-life is long for a hypnotic. It may be why lemborexant beat zolpidem ER 6.25 mg on second-half-of-the-night wakefulness in SUNRISE-1, though the trial did not test that explanation. It also fits the label warning that CNS depressant effects may persist for up to several days after stopping.
Part 3: Before You Start: Workup and Candidacy
The label requires no baseline labs. The workup is a history and a medication review, and the medication review matters more for this DORA than for the other two.
| Check | Why it matters |
|---|---|
| Narcolepsy, cataplexy, daytime sleep attacks | Narcolepsy is the only contraindication. Ask about sudden muscle weakness with laughter or surprise and about sleep paralysis. |
| Snoring, witnessed apneas, COPD | Studied in mild to severe OSA and moderate to severe COPD only in short trials; the label says to weigh respiratory effects. Untreated OSA is also a cause of insomnia that a hypnotic will not fix. |
| Depression and suicidality | Labeled warning. Suicidal ideation by questionnaire ran 0.3% to 0.4% on lemborexant vs 0.2% on placebo. Prescribe the smallest feasible quantity in high-risk patients. |
| Substance use history | Schedule IV. In recreational sedative users, drug liking matched zolpidem and suvorexant. Not a contraindication; follow these patients closely. |
| Full medication list | Strong or moderate CYP3A inhibitors or inducers mean avoid; weak inhibitors cap the dose at 5 mg; bupropion and methadone levels fall. See Part 8. |
| Liver disease | Moderate impairment (Child-Pugh B): 5 mg maximum. Severe: not recommended. Check LFTs if liver disease is suspected; this is practice, the label sets no lab schedule. |
| Morning duties and driving | Some subjects on 10 mg had impaired next-morning driving. A patient who must drive early or cannot give the drug 7 hours in bed is a poor fit. |
| Age and fall history | Drowsiness raises fall risk, particularly in older adults; the label asks for caution above 5 mg at 65 and older. |
| Pregnancy and breastfeeding | No human pregnancy data; low transfer into milk. See Part 9. |
Ask what the patient has already tried. CBT-I is the recommended first treatment for chronic insomnia (American College of Physicians guideline, Qaseem et al., Ann Intern Med 2016), and a patient who has never had it should be offered it alongside or before any hypnotic.
Part 4: How to Start and Dose
Adults 65 and olderFDA labelExercise caution when using doses higher than 5 mg in patients 65 and older.In practiceMany patients are best left at 5 mg; in older adults somnolence at 10 mg was 9.8% versus 4.9% at 5 mg.
With CYP3A inhibitors or inducersFDA labelWeak CYP3A inhibitors: 5 mg maximum. Strong or moderate inhibitors, and strong or moderate inducers: avoid.
Liver or kidney impairmentFDA labelModerate hepatic impairment: 5 mg maximum. Severe hepatic impairment: not recommended. Renal impairment: no dose adjustment.
Switching from a benzo or Z-drug1 row
Stopping1 row
Dosing at a glance
| Situation | Label dose | Notes |
|---|---|---|
| Start | 5 mg, no more than once per night | Immediately before going to bed, with at least 7 hours before planned waking |
| Increase | 10 mg, based on response and tolerability | Maximum recommended dose; caution against next-day driving at 10 mg |
| Age 65 and older | No separate label dose | Label: exercise caution above 5 mg |
| Weak CYP3A inhibitor | 5 mg maximum | Label examples: chlorzoxazone, ranitidine |
| Strong or moderate CYP3A inhibitor | Avoid | Itraconazole, clarithromycin, fluconazole, verapamil |
| Strong or moderate CYP3A inducer | Avoid | Rifampin, carbamazepine, St. John's wort, modafinil, efavirenz |
| Moderate hepatic impairment | 5 mg maximum | Child-Pugh B |
| Severe hepatic impairment | Not recommended | Not studied |
| Renal impairment | No adjustment | Severe impairment raises exposure; watch for somnolence |
| With food | No fixed rule | Sleep onset may be delayed if taken with or soon after a meal |
Formulations
Film-coated tablets, 5 mg (pale yellow) and 10 mg (orange), unscored, in bottles of 30 and 90. Brand only (Dayvigo, Eisai).
How to Take It
- Immediately before going to bed, once per night, and only with a full night (at least 7 hours) left before the patient has to be active.
- Not right after a meal. A high-fat meal pushes the peak back about 2 hours, so sleep onset comes late and the peak lands later in the night.
- No second dose after a middle-of-the-night awakening. The label allows one dose per night, and a 4-hour awakening on lemborexant already showed impaired balance and memory in older volunteers.
- No alcohol that night. Alcohol raised lemborexant exposure by about 70% and added to cognitive impairment.
Choosing Between 5 mg and 10 mg
Start at 5 mg, the label dose. The label allows 10 mg based on response and tolerability, which is reasonable if 5 mg is tolerated and not enough, but set expectations: in the trials the two doses performed about the same on the main sleep outcomes, and 10 mg roughly doubled discontinuation for adverse events over 6 months (8.3% vs 4.1%, placebo 3.8%). Somnolence, nightmares and sleep paralysis were all more frequent at 10 mg. In older adults, somnolence at 10 mg was 9.8% versus 4.9% at 5 mg. In practice, many patients are best left at 5 mg.
Switching From a Z-Drug, Benzodiazepine or Trazodone
The label says lemborexant should not be combined with other drugs for insomnia, and other CNS depressants may need dose reductions. No randomized trial has tested a switch to lemborexant, so this plan rests on a consensus statement, observational data and practice.
- After a week or two of a Z-drug or benzodiazepine: the 2026 German and Austrian sleep societies' consensus allows stopping abruptly after 1 to 2 weeks of use (Hajak et al., Dtsch Arztebl Int 2026); start lemborexant the next night (practice, not label).
- After months or years of nightly use: taper the old drug by 10% to 25% a week, more slowly after high doses or a year or more of use (same consensus). Those weeks of overlap are what the label advises against, so keep the old dose falling and warn about next-morning sedation and falls. Adding suvorexant to a benzodiazepine receptor agonist, rather than switching, caused more oversedation and dropouts in a retrospective study of 228 patients (Hatano et al., Clin Psychopharmacol Neurosci 2018).
- What the observational data show: of 605 long-term Japanese users who stayed on suvorexant or lemborexant for 6 months, 42% at least halved their bedtime benzodiazepine or Z-drug dose and 21% stopped it, with no comparison group (Matsui et al., J Clin Sleep Med 2026). Switches from short-acting agents failed less often than from longer-acting ones (Kanahara et al., Clin Psychopharmacol Neurosci 2025).
- Trazodone: the consensus tapers antidepressants used for sleep the same way after more than 2 weeks of use. Once lemborexant is working, do not keep trazodone as a second hypnotic (practice, not label).
Patients used to a GABAergic hypnotic sometimes find a DORA subtler and need a few weeks before judging it; that is a clinical impression, and there are no trial data on it.
Missed dose
- One dose: if it is forgotten at bedtime, take it only if at least 7 hours remain before the patient has to be up; otherwise skip that night. Never two doses in one night (label: once per night, with 7 hours left before waking).
- Several days: nothing to manage; abrupt stopping caused no withdrawal or rebound insomnia in the trials.
- Restarting after a gap: restart at the dose the patient was taking; there is no titration to repeat (practice, not label).
What to tell the patient at the first prescription
- What to expect and when: it usually helps from the first nights, and at the first follow-up we decide whether 5 mg is enough.
- Early side effects: morning grogginess, so do not drive until you feel fully awake, especially on 10 mg. Some people briefly cannot move or speak as they fall asleep or wake, or see things that are not there; it passes, but tell me.
- Call me about: walking, eating or driving while not fully awake (stop the drug), sudden leg weakness, or worsening mood or thoughts of suicide.
- How to take it: as you get into bed, only with 7 or more hours to sleep, not right after a big meal, never a second dose in the night, and no alcohol that night.
- Next visit: about 7 to 10 days in; if sleep is no better by then, we look for another cause (practice, not label).
Part 5: Monitoring
There are no drug levels and no routine labs. The label sets no follow-up schedule; the one below is ordinary practice for a hypnotic.
| Parameter | Baseline | Follow-up |
|---|---|---|
| Next-morning sleepiness, driving | ✓ | First follow-up, after any dose increase, and when a CYP3A drug is added |
| Response at 7 to 10 days | n/a | If insomnia has not remitted, look for a medical or psychiatric cause (label) |
| Complex sleep behaviors | ✓ | Ask at every visit; any episode means stop the drug |
| Sleep paralysis, hallucinations, leg weakness | ✓ | Ask at every visit, especially at 10 mg |
| Mood and suicidality | ✓ | Each visit in patients with depression |
| Falls | ✓ | Each visit in older adults |
| Breathing symptoms | ✓ | In OSA or COPD, ask about worsening snoring, apneas or morning headache |
| Clozapine level | If on clozapine | Recheck after starting lemborexant (one case report; see Part 8) |
| Ongoing need | n/a | Every few months; ask whether the patient still needs it |
Ask about the medication list at every visit. A new course of fluconazole or clarithromycin, or a new modafinil prescription from another clinician, changes whether lemborexant can be used at all.
If it isn't working
- First, check: that it is taken as the patient gets into bed, not right after dinner, with 7 hours in bed, and that no CYP3A inducer (carbamazepine, modafinil, rifampin, St. John's wort) has been added. Ask about morning grogginess and nightmares, which lead patients to skip doses.
- At 7 to 10 days: if the insomnia has not remitted, look for the psychiatric or medical illness driving it, including untreated sleep apnea, as the label asks. If 5 mg is tolerated, the label allows one step to 10 mg, though in the trials 10 mg added little on average.
- Do not: go above 10 mg, or add a second hypnotic on top; the label advises against combining insomnia drugs.
- Then: offer CBT-I if the patient has not had it, or change class: low-dose doxepin for maintenance or ramelteon for onset (not with fluvoxamine). There are no data on whether a patient who fails one DORA responds to another.
Part 6: Side Effects and How to Manage Them
Over the first 30 days of the pooled trials, discontinuation for adverse events was 1.4% on 5 mg and 2.6% on 10 mg, against 1.5% on placebo. Over 6 months of SUNRISE-2 it rose to 4.1% and 8.3%, against 3.8% on placebo, mostly for somnolence and nightmares. The rates below are from the label's pooled first 30 days.
Somnolence and Fatigue
Somnolence or fatigue occurred in 6.9% on 5 mg and 9.6% on 10 mg versus 1.3% on placebo, the only adverse reaction above 5% and at least twice placebo. It led to discontinuation in 2.9% on 10 mg over 6 months.
- Management: drop from 10 mg to 5 mg, enforce the 7-hour rule, move the dose away from dinner, check for a new CYP3A inhibitor or alcohol, and stop it if sleepiness persists at 5 mg.
Headache
Headache was 5.9% on 5 mg and 4.5% on 10 mg versus 3.4% on placebo, so unlike somnolence it did not rise with dose.
Nightmares and Abnormal Dreams
Reported in 2.2% on 10 mg versus 0.9% on 5 mg and on placebo. Nightmares caused discontinuation in 1.3% on 10 mg over 6 months. Going back to 5 mg is a reasonable response.
Sleep Paralysis, Hallucinations and Cataplexy-Like Symptoms
Sleep paralysis occurred in 1.3% on 5 mg and 1.6% on 10 mg, with none on placebo. Hypnagogic hallucinations were 0.1% and 0.7%, also none on placebo. The label warns that symptoms resembling mild cataplexy can occur: leg weakness lasting seconds to a few minutes, by day or night, sometimes without a trigger.
- Management: explain these before the first dose, as the label asks, so a frightening episode does not end treatment. Lower the dose or stop if they recur. New daytime weakness with laughter or surprise should prompt a narcolepsy evaluation.
Middle-of-the-Night and Next-Morning Function
- Middle of the night: when healthy older adults were woken 4 hours after dosing, both doses impaired balance compared with placebo, and attention and memory worsened in a dose-dependent way. Body sway was larger on zolpidem ER 6.25 mg than on either lemborexant dose, and the ability to wake to a sound was unaffected (Murphy et al., J Clin Sleep Med 2020). The label asks you to warn patients about unsteadiness and memory lapses if they get up during the night.
- Next morning: postural stability and memory did not differ from placebo. On an on-road driving test about 9 hours after dosing, neither dose caused statistically significant impairment in adults or older adults, but some individuals on 10 mg were impaired.
The label calls lemborexant a CNS depressant that can impair daytime wakefulness even when taken as prescribed, and says the effect may last up to several days after stopping. The risk rises with less than a full night in bed, with doses above the label, and with alcohol or other sedatives. Patients on 10 mg should be cautioned specifically about driving the next morning.
Complex Sleep Behaviors
Two complex sleep behavior events were reported in the trials, both on 10 mg. The label carries the class warning: sleep-walking, sleep-driving, eating, phone calls or sex without full awareness, after the first or any later dose, with or without alcohol. Stop lemborexant at the first episode. Unlike the Z-drugs, the DORAs carry no boxed warning or contraindication for this.
What You Don't See
- Withdrawal or rebound insomnia: none after stopping in the 1-month and 12-month trials (Part 10).
- Loss of effect: sleep diary benefits held through 12 months of continuous use.
Part 7: Toxicity, Overdose, and Warnings
Boxed Warning
None. The contraindication is narcolepsy. The warnings and precautions are CNS depressant effects and daytime impairment, sleep paralysis with hallucinations and cataplexy-like symptoms, complex sleep behaviors, compromised respiratory function, worsening depression and suicidal ideation, and the need to evaluate for comorbid diagnoses.
Depression and Suicidality
In the insomnia trials, suicidal ideation or behavior by questionnaire was 0.4% on 5 mg and 0.3% on 10 mg against 0.2% on placebo. In depressed patients, hypnotics have been associated with worsening depression and suicidal thoughts and acts, and intentional overdose is more common, so the label asks for the lowest feasible number of tablets per prescription. Evaluate any new behavioral change promptly.
Abuse and Scheduling
In a human abuse potential study in 29 recreational sedative users, lemborexant 10, 20 and 30 mg produced drug liking and "take drug again" ratings statistically similar to zolpidem 30 mg and suvorexant 40 mg, and higher than placebo. In animals and in the trials, stopping chronic dosing produced no withdrawal, and the label says this suggests lemborexant does not produce physical dependence. Follow patients with addiction histories carefully, as the label asks.
Lemborexant is Schedule IV. On August 11, 2026 the DEA published a proposed rule (91 FR 51606) to move lemborexant, suvorexant and daridorexant to Schedule V. HHS recommended the move. The proposal notes fewer abuse and misuse cases than with zolpidem or temazepam but a number similar to triazolam, possibly because the DORAs are used less, and says the available data suggest the DORAs do not produce physical or psychological dependence. Comments closed September 10, 2026, and no final rule had been published as of November 19, 2026. Prescribe it under Schedule IV rules until a final rule takes effect.
Overdose
Clinical overdose experience is limited. Healthy volunteers given repeated doses up to 75 mg (7.5 times the maximum) had dose-dependent increases in somnolence. There is no antidote. Give supportive care with close monitoring and consider other drugs, especially alcohol and other sedatives. Lemborexant is highly protein bound, so hemodialysis is not expected to help. Poison Control (1-800-222-1222) can advise. In one published case of about 720 mg taken with eszopiclone, the patient recovered with supportive care, but her plasma level kept rising until 9 hours after the ingestion (Suzuki et al., Toxicol Rep 2025), a reason not to judge the course from the first few hours.
Part 8: Drug Interactions
Lemborexant is a CYP3A4 substrate with tighter interaction rules than suvorexant or daridorexant. It also lowers levels of CYP2B6 substrates, which puts bupropion and methadone on the list.
| Interaction | Effect | Risk | Action |
|---|---|---|---|
| Strong CYP3A inhibitors (itraconazole, clarithromycin, nefazodone) | Itraconazole raised lemborexant AUC about 3.7-fold; Cmax rose 1.4- to 1.6-fold across itraconazole and fluconazole (Landry et al., Clin Pharmacol Drug Dev 2021; label Figure 2) | HIGH | Avoid the combination |
| Moderate CYP3A inhibitors (fluconazole, verapamil, erythromycin, fluvoxamine) | Fluconazole raised AUC about 4-fold (same study) | HIGH | Avoid. With fluvoxamine, suvorexant 5 mg or daridorexant 25 mg stay within their labels |
| Weak CYP3A inhibitors (chlorzoxazone, ranitidine) | Modeling predicts less than a 2-fold rise in exposure | MODERATE | 5 mg maximum |
| Strong CYP3A inducers (rifampin, carbamazepine, phenytoin, St. John's wort) | Rifampin cut AUC and Cmax by more than 90% | HIGH | Avoid; it will likely not work |
| Moderate CYP3A inducers (modafinil, efavirenz, etravirine, bosentan) | Lower exposure, which may reduce efficacy | HIGH | Avoid |
| Alcohol | AUC up about 70% and Cmax up 35%; added impairment of attention and memory (Landry et al., J Psychopharmacol 2022) | HIGH | No alcohol with it |
| Benzodiazepines, opioids, TCAs, other sedatives | Additive CNS depression and daytime impairment | MODERATE | Dose reductions may be needed; do not combine with other insomnia drugs |
| CYP2B6 substrates (bupropion, methadone) | Lemborexant 10 mg cut S-bupropion AUC by about 45% | MODERATE | Watch for loss of response; a higher dose of the substrate may be needed |
| Clozapine | One case of rising clozapine and norclozapine levels with oversedation that reversed when lemborexant stopped; in vitro, lemborexant was a time-dependent CYP3A4 inhibitor (Watanabe et al., Br J Clin Pharmacol 2024) | MODERATE | If combined, check a clozapine level and watch for sedation |
| Midazolam, oral contraceptives, famotidine | Midazolam AUC about 1.1-fold; no clinically relevant change with ethinyl estradiol and norethindrone; famotidine had no effect | NONE | No adjustment |
Inhibitor and inducer examples come from the label and from the FDA's table of clinical CYP3A inhibitors and inducers (fluvoxamine and erythromycin as moderate inhibitors; phenytoin as a strong inducer). Nefazodone is listed as a strong CYP3A inhibitor in the Belsomra label.
Three common psychiatric prescriptions rule lemborexant out: fluvoxamine (moderate CYP3A inhibitor), carbamazepine (strong inducer) and modafinil (moderate inducer). Do not swap in ramelteon for the fluvoxamine patient either; ramelteon's label says not to combine it with fluvoxamine.
Part 9: Special Populations
Elderly
- Of 1,418 patients in the phase 3 trials, 491 were 65 or older and 87 were 75 or older. Efficacy was similar to younger adults.
- Somnolence on 10 mg was 9.8% at 65 and older versus 7.7% under 65; on 5 mg it was about 5% in both groups. The label asks for caution above 5 mg in patients 65 and older and warns of fall risk.
- In analyses of the 65-and-older participants of three trials, lemborexant did not impair next-morning balance or memory, or next-day driving in healthy older volunteers, compared with placebo (Gotfried et al., Drugs Aging 2024). The label still notes impaired driving in some subjects on 10 mg, and the middle-of-the-night balance data in Part 6 still apply.
- In Japanese records of adults 65 and older starting a hypnotic, lemborexant was linked to fewer hip-fracture hospitalizations over a year than the Z-drugs (hazard ratio 0.31); the suvorexant comparison did not reach significance. The study was observational (Matsuo et al., J Am Geriatr Soc 2026).
- The 2023 AGS Beers Criteria list the Z-drugs and benzodiazepines as hypnotics to avoid in older adults. The orexin antagonists do not appear in them.
Pregnancy
There are no human pregnancy data. In rats and rabbits, malformations and fetal toxicity appeared only at very high multiples of the maximum human exposure (no-effect levels about 143 and 23 times the maximum human dose by AUC). CBT-I is the first option in pregnancy. If a hypnotic is needed, agents with more human pregnancy data are usually preferred; a patient already stable on lemborexant needs an individual decision. Encourage enrollment in the National Pregnancy Registry for Psychiatric Medications (1-866-961-2388).
Lactation
In 8 lactating women given a single 10 mg dose, the estimated infant dose was 0.0029 mg/kg/day, a relative infant dose under 2%, and about 70% of the drug in milk was excreted within 24 hours. There are no data on effects in the infant or on milk supply, or after repeated dosing. If used, watch the infant for excess sedation.
Renal Impairment
No dose adjustment for mild, moderate or severe impairment. Exposure was about 1.5-fold higher in severe impairment, so expect more somnolence.
Hepatic Impairment
- Mild (Child-Pugh A): no dose change, but exposure rises (half-life unchanged); watch for somnolence.
- Moderate (Child-Pugh B): exposure and half-life increase; start and stay at 5 mg.
- Severe (Child-Pugh C): not studied; not recommended.
Sleep Apnea and COPD
Eight nights of 10 mg did not change the apnea-hypopnea index in mild OSA (37 patients; difference -0.06 events per hour) or moderate to severe OSA (33 patients; -0.80), and did not lower oxygen saturation in moderate to severe COPD (30 patients; +0.47%). The label notes these were short studies and clinically meaningful respiratory effects cannot be excluded, including with long-term use, and COPD with FEV1 under 30% of predicted was not studied. Treat the apnea first; lemborexant is not a substitute for CPAP.
Pediatric
Safety and effectiveness have not been established under 18. Any use in this age group is off-label.
Part 10: Discontinuation
Lemborexant can be stopped without a taper. In both phase 3 trials, patients stopped abruptly with no withdrawal on a benzodiazepine withdrawal questionnaire and no rebound insomnia on sleep diaries over the next 2 weeks. The label sets no taper, and the absence of withdrawal is the basis for stopping it directly; that is practice drawn from the trial data.
- Expect a few nights of carryover. With a 17- to 19-hour effective half-life, CNS depressant effects may persist for several days after the last dose.
- Expect the insomnia to return if its cause is still there. No rebound means sleep goes back to baseline, and baseline may still be poor.
- Switching to another DORA: there are no data. Do not overlap them, since the label advises against combining insomnia drugs, and allow for the carryover above.
- Switching onto lemborexant from a Z-drug, benzodiazepine or trazodone is covered in Part 4.
Part 11: Lemborexant vs the Alternatives
The Three Orexin Antagonists
No head-to-head trials compare the three DORAs. The table sets their US labels side by side.
| Feature | Suvorexant (Belsomra) | Lemborexant (Dayvigo) | Daridorexant (Quviviq) |
|---|---|---|---|
| US approval | 2014 | 2019 | 2022 |
| Tablets | 5, 10, 15, 20 mg | 5, 10 mg | 25, 50 mg |
| Label dose | 10 mg; maximum 20 mg | 5 mg; maximum 10 mg | 25 to 50 mg; maximum 50 mg |
| When to take | Within 30 minutes of bed, 7 or more hours before waking | Immediately before bed, 7 or more hours before waking | Within 30 minutes of bed, 7 or more hours before waking |
| Half-life | About 12 hours | Effective 17 hours (5 mg) and 19 hours (10 mg); active metabolite M10 | About 8 hours |
| High-fat meal | Peak delayed about 1.5 hours; exposure unchanged | Peak delayed 2 hours; Cmax down 23%, AUC up 18% | Peak delayed 1.3 hours; Cmax down 16%; AUC unchanged |
| Strong CYP3A inhibitor | Not recommended | Avoid | Avoid |
| Moderate CYP3A inhibitor | 5 mg; may raise to 10 mg | Avoid | 25 mg maximum |
| Weak CYP3A inhibitor | No label limit | 5 mg maximum | No label limit |
| CYP3A inducers | Strong: efficacy may be reduced | Avoid strong or moderate | Avoid strong or moderate |
| Effect on other drugs | Digoxin rises slightly; monitor digoxin levels | Lowers CYP2B6 substrates (bupropion, methadone) | Raises CYP3A4 and P-gp substrates; caution with narrow therapeutic index drugs |
| Hepatic impairment | No change if mild or moderate; severe not recommended | Moderate: 5 mg maximum; severe not recommended | Moderate: 25 mg maximum; severe not recommended |
| Pivotal trials | Two 3-month placebo-controlled trials (20 mg, 15 mg if 65 or older, alongside higher doses) | SUNRISE-1: 1 month, women 55 and older and men 65 and older, beat zolpidem ER 6.25 mg on late-night waking. SUNRISE-2: 6 months placebo-controlled, 12 months total | Two 3-month trials plus a 40-week extension; 50 mg also improved daytime sleepiness scores |
| Next-morning driving (label study, about 9 hours after dosing) | Impaired in some non-elderly subjects after one 20 mg dose | No significant mean impairment; some subjects impaired on 10 mg | Significant impairment after the first 50 mg dose, not after 4 nights; some subjects impaired (driving simulator study) |
| DEA schedule (November 2026) | C-IV; DEA proposed C-V in August 2026, not final | C-IV; DEA proposed C-V in August 2026, not final | C-IV; DEA proposed C-V in August 2026, not final |
| US generic | None marketed | None | None |
Choosing Among the Three
- Sleep onset or maintenance: all three helped both. In network estimates, lemborexant 10 mg had the largest effect on subjective sleep onset and daridorexant 50 mg on subjective total sleep, and lemborexant 10 mg, daridorexant 25 mg and suvorexant were linked to more somnolence than placebo in a sensitivity analysis (Kishi et al., J Psychiatr Res 2020; Transl Psychiatry 2025). Cross-trial estimates like these do not pick the drug for a given patient.
- Next-morning risk: the half-lives differ (about 8 hours for daridorexant, 12 for suvorexant, 17 to 19 effective for lemborexant), but no head-to-head data show that this changes next-morning impairment; each label reports impairment in some subjects in its driving study.
- Older adults: lemborexant has the only active-comparator trial in older adults (SUNRISE-1) and an observational hip-fracture result in its favor against the Z-drugs; daridorexant's older-adult data are a subgroup analysis. Every DORA label warns about falls.
- Liver disease: in moderate impairment, suvorexant needs no dose change, lemborexant is capped at 5 mg and daridorexant at 25 mg; none is recommended in severe impairment.
- Interacting drugs: with a moderate CYP3A inhibitor such as fluvoxamine, diltiazem or fluconazole, use suvorexant 5 mg or daridorexant 25 mg, not lemborexant. Only lemborexant has a cap with weak inhibitors, and only its label warns that it lowers bupropion and methadone. Suvorexant calls for digoxin levels, and daridorexant for caution with narrow-index CYP3A4 and P-gp substrates. With carbamazepine or another strong inducer, none of the three is a good choice.
- Cost and coverage: all three are brand-only in the US, so the formulary often decides more than the pharmacology; check it before you choose (practice, not label).
Lemborexant vs Other Hypnotics
| Comparison | Lemborexant advantage | Alternative advantage |
|---|---|---|
| vs Z-drugs (zolpidem, eszopiclone, zaleplon) | Beat zolpidem ER 6.25 mg on late-night waking and caused less middle-of-the-night body sway; no withdrawal or rebound in trials; no boxed warning for complex sleep behaviors; not on the Beers avoid list | Cheap generics; zaleplon is cleared mainly by aldehyde oxidase, so CYP3A drugs matter less; eszopiclone also beat placebo long-term in the 2022 network meta-analysis |
| vs Low-dose doxepin (Silenor 3 to 6 mg) | Helps sleep onset as well as maintenance | Not a controlled substance; generic; well tolerated in the network meta-analysis, though efficacy data were scarce; Beers flags doxepin only above 6 mg/day and rates the low dose's safety close to placebo |
| vs Ramelteon | Treats sleep maintenance as well as onset; ramelteon showed no material benefit overall in the 2022 network meta-analysis | Not a controlled substance; generic; the cleaner choice when any misuse risk is unacceptable (but not with fluvoxamine) |
| vs Trazodone | FDA-approved for insomnia; long-term placebo-controlled data; no orthostatic hypotension or priapism warnings | Generic and cheap; useful when depression is also being treated; effective acutely in the network meta-analysis, which grouped it with drugs that were poorly tolerated or lacked long-term data |
Lemborexant fits the older adult with middle-of-the-night and early-morning waking whose medication list is free of CYP3A drugs. It loses to daridorexant 25 mg or suvorexant 5 mg when a moderate CYP3A inhibitor cannot be stopped, and to the cheaper generics when insurance will not cover it.
The Bedside Cheat Sheet
What it is
- Dual orexin receptor antagonist (OX1R and OX2R); active metabolite M10
- Effective half-life 17 h (5 mg) to 19 h (10 mg)
- Adult insomnia, onset and/or maintenance; Schedule IV (Schedule V proposed August 2026, not final)
- Brand only (Dayvigo); no US generic
Starting and dosing
- Tablets 5 mg and 10 mg
- 5 mg immediately before bed, only with 7 or more hours to sleep; maximum 10 mg
- Age 65 and older: caution above 5 mg
- Moderate liver impairment: 5 mg maximum; severe: not recommended
- Not right after a meal; no second dose in the night; no alcohol
Side effects that matter
- Somnolence 7% (5 mg) and 10% (10 mg) vs 1%
- Headache; nightmares at 10 mg
- Sleep paralysis 1.3% to 1.6%; hallucinations; cataplexy-like leg weakness
- Complex sleep behaviors: stop the drug at the first episode
- No withdrawal or rebound insomnia in trials
Don't forget
- Avoid with strong or moderate CYP3A inhibitors or inducers: fluconazole, clarithromycin, fluvoxamine, carbamazepine, rifampin, modafinil
- Weak CYP3A inhibitor: 5 mg maximum
- Lowers bupropion and methadone levels; one case of raised clozapine levels
- Coming off a nightly benzodiazepine or Z-drug: taper it 10% to 25% a week; do not leave both
- Narcolepsy is the contraindication
The patient who most needs a plan is the one already taking a benzodiazepine or Z-drug every night. Write the taper of the old drug into the first lemborexant visit, with dates, so the overlap the label warns against does not turn into two hypnotics for good.