Suvorexant

Suvorexant is a dual orexin receptor antagonist (DORA) developed by Merck [3]. It was the first drug of its kind approved for insomnia [4], in the United States in 2014 [1] and also in Japan [5], and it is licensed for difficulty falling asleep, staying asleep or both [1].

Insomnia. In a phase 2 crossover trial, 10–80 mg produced dose-related gains in polysomnographic sleep efficiency on the first night and after 4 weeks, with dose-related effects on sleep onset and maintenance [6]; 10 and 20 mg were each better than placebo on both [1]. Two 3-month phase 3 trials randomized 2,040 adults, testing 40 mg (30 mg in people aged 65 and over) and a smaller arm of 20 mg (15 mg) [7]. At the licensed 15–20 mg, suvorexant beat placebo by 11–22 minutes of self-reported sleep and 17–31 minutes less wake after sleep onset; the effect on polysomnographic sleep onset was 8–10 minutes at month 1 and had disappeared in one trial by month 3 [1]. In the year-long safety trial at 30–40 mg, people reported 22.7 more minutes of sleep and fell asleep 9.5 minutes sooner than on placebo at month 1, and 62% completed the year [3]. In pooled data from people aged 65 and over, 30 mg and 15 mg both improved sleep maintenance and onset, although the onset effect of 15 mg was less evident at later time points [8]. Calculated from the phase 3 data, about 8 people need to be treated for one more to improve by at least 6 points on the Insomnia Severity Index at 3 months [4].

How it compares. A meta-analysis of four trials (3,076 patients) found better subjective sleep onset, total sleep and sleep quality at 1 and 3 months [5]. In a 2025 network meta-analysis of the three licensed DORAs, every dose beat placebo, but suvorexant had the smallest effect on time to fall asleep (standardized mean difference −0.16), and lemborexant 5 and 10 mg did better than it on that measure [9]. The 2022 Lancet network meta-analysis of 154 trials grouped suvorexant with drugs that work in the short term but have poor tolerability or no long-term data [10], and the American Academy of Sleep Medicine's 2017 guideline weakly suggests it for sleep-maintenance insomnia [11].

Dementia and delirium. In 285 people with mild-to-moderate Alzheimer's disease and insomnia, 4 weeks of 10–20 mg added 28 minutes of total sleep time over placebo (73 against 45 minutes) and cut wake after sleep onset by 16 minutes, most markedly in the last third of the night, with no change in cognitive test scores [12]; the Cochrane review of sleep drugs in dementia rated this moderate-certainty evidence [13]. For delirium, a rater-blinded trial of 72 older emergency admissions found no cases on 15 mg against 6 of 36 on placebo [14], but a larger double-blind phase 3 trial in 203 high-risk inpatients found 16.8% against 26.5%, a difference that was not statistically significant [15]. A 2025 meta-analysis pooling suvorexant, lemborexant and ramelteon found fewer cases of delirium, on low-certainty evidence and without clear differences between the drugs [16].

Other groups. In a pilot trial in 19 night-shift workers, 20 mg before daytime sleep added about 2 hours of sleep by actigraphy [17]. In 56 midlife women whose insomnia came with night-time hot flushes, insomnia severity fell 8.1 points against 5.6 on placebo and night-time flushes became less frequent [18]. In trauma-related insomnia it did no better than placebo [19]. During a buprenorphine taper in 38 people with opioid use disorder, suvorexant increased sleep during the taper and reduced withdrawal symptoms afterward, with no sign of abuse potential [20].

Alzheimer's biomarkers. In 38 cognitively normal adults with spinal catheters, 20 mg lowered the phosphorylated fraction of tau at threonine 181 by about 10–15% and amyloid-β by 10–20% for some hours after each of two doses, without increasing sleep time; other tau sites did not change and 10 mg had no significant effect [21]. The authors say chronic studies are needed before it can be considered for preventing Alzheimer's disease [21].

The orexin system. Orexin A and B (also called hypocretins) are made by neurons found only in the lateral hypothalamus, which project to the brain's arousal centers [22]. These neurons fire during wakefulness and fall silent during sleep; their loss causes narcolepsy in people, a mutation of the OX2 receptor causes it in dogs, and mice lacking orexin develop a narcolepsy-like disorder [22][23]. The OX2 receptor carries orexin's effect on histamine-mediated arousal, and both receptors take part in the brainstem circuits that gate wake and sleep [23].

Blocking the wake signal. Suvorexant binds OX1 and OX2 receptors with Ki values of 0.55 and 0.35 nM [1]. The rationale is to lower the threshold for falling asleep and dampen orexin-driven arousal, rather than suppressing brain activity broadly as GABA-A receptor modulators do [2]. Taken by mouth it reduced activity and promoted sleep in rats, dogs and rhesus monkeys, with consistent changes in sleep architecture across species, and occupied over 90% of human OX2 receptors in transgenic rat brain at plasma levels of 1.1 µM [24]. Its half-life was tuned to cover the night at low-to-moderate doses [2].

Effect on sleep stages. Reviews of its development report that DORAs promote both non-REM and REM sleep without altering the stage-specific EEG spectra, and that people stay able to wake to salient stimuli [2]. In the Alzheimer's trial the proportion of time in each sleep stage did not change significantly compared with placebo [12].

Side effects from the mechanism. Because losing orexin causes narcolepsy, blocking it can produce narcolepsy-like symptoms: sleep paralysis, hallucinations at sleep onset or waking and cataplexy-like weakness [1]. Dogs given 5 and 30 mg/kg a day showed cataplexy-like limb buckling when offered food [1], and across DORA trials excessive daytime sleepiness and sleep paralysis were about two and three times as common as on placebo [25].

Tablets and absorption. Suvorexant is a white powder that does not dissolve in water, sold as film-coated tablets of 5–20 mg [1]. Exposure rises less than in proportion to dose because absorption falls at higher doses, and a high-fat meal delays the peak by about 1.5 hours without changing total exposure [1].

Who reaches higher levels. Women have 17% higher exposure than men at 40 mg, people with obesity 31% higher, and obese women 46% higher than non-obese women [1]. Moderate liver impairment did not change exposure but lengthened the half-life, and severe kidney impairment did not change exposure; the drug has not been studied in severe liver impairment [1].

Common effects. In the 3-month trials, somnolence affected 7% on 15–20 mg against 3% on placebo, and 3% stopped because of side effects against 5% on placebo [1]. Somnolence rose with dose, from 2% at 10 mg to 5% at 20 mg and 11–12% at 40–80 mg, and was more common in women (8%) than men (3%) [1]. Over a year at 30–40 mg, 69% had any adverse event against 64% on placebo, 13% had somnolence against 3%, and serious events occurred in 5% against 7% [3]. Across all three DORAs, treatment-related adverse events were about 1.7 times as common as on placebo [25].

Next-day function and driving. In 28 healthy adults under 65, 20 mg raised next-morning lane weaving by 1.01 cm, statistically significant but below the 2.4 cm threshold taken as clinically meaningful; more drivers were impaired than improved, and four women stopped five test drives because they felt too drowsy, all after suvorexant [28]. In 24 healthy adults aged 65–80, 15 and 30 mg did not impair next-morning driving [30]. A network meta-analysis found suvorexant 15–20 mg caused less impairment than zopiclone [31]. The label cautions people taking 20 mg against next-day driving and notes that its depressant effects can last several days after stopping in some people [1]. With alcohol, 40 mg had additive effects on attention, memory and balance, though the combination did not change suvorexant's pharmacokinetics [32].

Breathing. In 26 people with mild-to-moderate obstructive sleep apnea, four nights of 40 mg raised the apnea–hypopnea index by 2.66 events an hour; the upper confidence bound of 5.09 just missed the trial's prespecified non-inferiority limit of 5, oxygen saturation did not change, and responses varied widely within and between people [33]. In 25 people with mild-to-moderate COPD, 30–40 mg did not lower oxygen saturation [34]. A meta-analysis of four DORA trials in sleep apnea found more sleep and no rise in the apnea–hypopnea index or in oxygen desaturation [35]. Suvorexant has not been studied in severe sleep apnea or severe COPD [1].

Mood and complex behaviors. Questionnaires picked up a dose-dependent increase in suicidal ideation in clinical studies, and the label warns that hypnotics can worsen depression [1]. In the phase 3 program suicidal ideation was reported by 0.2% on 15–20 mg, 0.7% on 30–40 mg and 0.1% on placebo, none with a formal plan or suicidal behavior [4]. In US adverse-event reports, DORAs were not linked to more suicidal ideation or behavior than trazodone, and completed suicides were reported less often [36]. Sleep-walking, sleep-driving and other complex sleep behaviors have been reported with hypnotics including suvorexant, and the label says to stop it if they occur [1]; abnormal sleep-related events were its strongest signal in a 2025 analysis of US reports [37].

Falls. In the Alzheimer's trial three people on suvorexant fell, one breaking an ankle, against none on placebo; none of the falls was judged to be cataplexy [12]. In hospitalized older adults, 6.9% fell against 2.9% on placebo [15].

Abuse and dependence. Suvorexant is in Schedule IV [1]. In 36 recreational drug users, single 40–150 mg doses were liked more than placebo and about as much as zolpidem, but the effect did not grow with dose, "high" ratings were lower than with zolpidem 30 mg and euphoria-type events were numerically fewer [38]. No withdrawal or clear rebound insomnia appeared when it was stopped after 3 or 12 months [1][4][7].

Other findings. Doses up to 240 mg did not prolong the QTc interval to a clinically relevant extent [1]. Rats dosed for two years developed retinal atrophy at all doses, and thyroid and liver tumors at high doses that were attributed to rodent-specific mechanisms [1]. Human pregnancy data are insufficient; in breast milk the infant dose is under 1% of the mother's [1].

Suvorexant is cleared mainly by CYP3A, so ketoconazole raised its exposure 2.79-fold and diltiazem 2.05-fold, and rifampin cut it by 88% [29]. The label advises against strong CYP3A inhibitors such as itraconazole, clarithromycin and ritonavir, recommends 5 mg with moderate inhibitors such as fluconazole, erythromycin, verapamil and grapefruit juice, and warns that strong inducers such as carbamazepine and phenytoin may make it less effective [1]. It adds to the effects of alcohol and other central depressants, including benzodiazepines, opioids and tricyclics; the label advises against alcohol and against combining it with other insomnia drugs, and asks for digoxin to be monitored [1]. Paroxetine did not interact with it [1].