A recent thread on the Women’s Mental Health listserve prompted the creation of this summary of information about medications used to improve sleep. It is a work in progress and I should note that the text in italics is from the Minkel and Krystal article. I will complete the article soon…
Before discussing medications, I want to make three important points:
- People with insomnia should be evaluated to rule out specific disorders of sleep (such as periodic leg movements during sleep (PLMS), sleep apnea, and narcolepsy) which need other types of treatments and should receive a general health screen to rule out other medical conditions that may cause disrupted sleep.
- Cognitive Behavioral Therapy for Insomnia (CBTi), which we have discussed elsewhere on this website, is generally more effective and certainly safer than medications for the treatment of most types of chronic insomnia.
- People with sleep disorders secondary to another psychiatric condition (major depression, PTSD, generalized anxiety, psychosis, etcetera) need to receive treatment for that condition or they will not get better.
This review attempts to focus on evidence based treatments. There are many, many medications and non-prescribed supplements that are felt to improve sleep quality, but have limited, if any, evidence to support their use. We have not included those in this list.
If you are interested in a more detailed discussion of the psychopharmacology of sleep and wakefulness (but one which does not focus on strength of the data supporting efficacy) I recommend this summary by Flavio Guzman.
In addition, where available, I have tried to identify sleep studies that include sleep studies. This seriously limits the list of studies since most research uses patient reports of sleep duration and quality. This type of research is much less costly to perform but much less useful. Among other problems with patient self-reported sleep quality is the fact that many of the medications used to treat insomnia also interfere with memory. This may explain why patients feel that benzodiazepines and related medications are much more effective than sleep study data would suggest – patients no longer remember transient awakenings, nor how long it takes to get to sleep.
This page uses the traditional scheme for classifying psychotropic medications. For information about the Neuroscience Based Nomenclature, an alternative system that is supported by a number of prestigious organizations click here.
The Evidence Base
This chart, from the Wilt, et al, Annals of Internal Medicine review shows how many studies have been done on various treatments of insomnia.
Not surprisingly, the bulk of the studies have been done on prescription medications and were sponsored by the manufacturer (zolpidem, zaleplon, eszopiclone, ramelteon, suvorexant, temazepam). The other treatments that have good evidence for efficacy, include CBTi and low dose doxepin. Melatonin does not have consistent data and has been studied at many different doses and formulations. Other medications that we will mention have even less data supporting their use.
Benzodiazepine and Nonbenzodiazepine GABA-A Enhancing Medications
The most widely prescribed medications for insomnia enhance the activity of GABA-A receptors. Since GABA-A receptors are distributed throughout the brain, these medications have many effects on brain activity.
These additional effects include reducing anxiety, stimulating the reward circuits in the brain (which is linked to abuse potential), anticonvulsant and muscle relaxation effects. Adverse effects may include sedation, cognitive impairment (anterograde short-term memory loss), motor impairment and the potential for abuse.
The nonbenzodiazepine agents were more recently introduced and were designed to avoid some of the adverse effects of benzodiazepines. It is not clear to what extent they succeed, and, in fact, there are only two studies that compared a benzodiazepine and nonbenzodiazepine hypnotic (zolpidem and temazepam) and these two agents are quite different in other important ways (temazepam has a much longer half life) which limits our ability to interpret the results.
A number of controlled trials have established the efficacy of benzodiazapines for the treatment of insomnia. Triazolam (Halcion), temazepam (Restoril), flurazepam, quazepam, and estazolam (ProSom) have been found to have beneficial effects on sleep onset and maintenance for insomnia patients aged 18–65 (Krystal, 2009).
These medications are more similar than different and the primary factors that guide the selection of medications are related to speed of onset (which is related to lipid solubility) and duration of action (which is related to the half-life of the medication).
Of the benzodiazapines commonly prescribed for insomnia, triazolam has the shortest half-life and is least likely to produce next-day effects. Some benzodiazepines have half-lives that exceed 24 hours, including flurazepam, quazepam and clonazepam (Klonopin). These medications are relatively likely to lead to next-day effects, such as daytime sedation.
Of note, these medications have not been extensively studied in chronic insomnia, mostly because the FDA did not encourage such studies when they were brought to market.
This summary from Minkel and Krystal (2013) summarizes the differences between these medications and the benzodiazepines.
Non-benzodiazepines include some of the most commonly prescribed sleep-promoting medications – zolpidem (Ambien), zaleplon (Sonata), and eszopiclone (Lunesta).
These medications are chemically unrelated to benzodiazepines but have similar pharmacologic effects. Although they bind to the same site on the GABA-A receptor complex as do the benzodiazepines, they bind more specifically to subtypes of GABA-A receptors. In contrast to benzodiazepines, which bind non-specifically to GABA-A receptors containing the α1, α2, α3, and α5 subunits, the non-benzodiazepines tend to selectively bind to a subset of GABA-A receptors.
Zolpidem and zaleplon, which bind relatively preferentially to α1 containing GABA receptors, have effects limited to sleep, anti-seizure, memory, and motor impairing effects.
Eszopiclone, on the other hand, has anti-anxiety and muscle relaxing effects, in addition to increasing sleep, because it bind also to GABA-A receptors containing α2 and α3 subunits.
Non-benzodiazepines have been evaluated for longer term efficacy and safety. Eszopiclone has been evaluated through 6 months in a two placebo-controlled trials and up to 12 months in an open-label extension. Zaleplon has been shown to be safe in 6–12 month open label studies of older adults.
The side effect profile of non-benzodiazepines has been found to be similar to the benzodiazepines and includes sedation, dizziness and psychomotor impairment. Non-benzodiazepines appear to have lower abuse potential at recommended dosages, but may still have a significant risk at higher doses.
Melatonin Receptor Agents
Melatonin itself, while widely recommended and taken for treatment of insomnia and chronic insomnia, has not been shown to be effective. It may be more effective as an agent to enhance normal circadian rhythms and may thus be more useful for people with insomnia due to jet lag, or sleep phase delays.
Ramelteon (Rozerem) is a prescription medication that that selectively binds to melatonin 1 and melatonin 2 receptors in the suprachiasmatic nucleus. Ramelteon has been shown to be efficacious in several randomized, placebo-controlled trials, including one six month study. Most studies have evaluated a dose of 8 mg administered 30 minutes prior to bedtime. Studies have been done in older adults (age 65+) as well as in those between the ages of 18 and 65. The effects are limited to reductions in how long it takes to get to sleep, and this effect is larger in sleep studies than by patient report (perhaps because patients compare ramelteon’s effects to the effects they experience with benzodiazepines which also impair memory of sleep delays). The most commonly reported side effects are headache, somnolence and sore throat, but these are not significantly greater than with placebo.
Suvorexant (Belsomra) has been recently approved for the treatment of insomnia on the basis of two medium duration (3 month) and one long term (one year) study showing modest improvements in total sleep time compared with placebo at doses of 15 or 20 mg (the long term study used a higher dose that was not approved by the FDA of 30 to 40 mg).
Response to treatment was modestly higher with suvorexant than with placebo (55% versus 42%) and the average increase in total sleep time was also a modest average of 16 minutes.
The FDA labeling information for suvorexant warns of cognitive and behavioral changes, such as amnesia, anxiety, hallucinations, and other neuropsychiatric symptoms; complex behaviors, such as sleep driving; worsening of depression, including suicidal thinking in individuals with depression; daytime impairments; sleep paralysis; and hypnagogic/hypnopompic hallucinations. However the evidence from the clinical trials is much more reassuring…
In the two 3-month trials, the only adverse effect clearly due to suvorexant was somnolence (rates of somnolence were 7% for suvorexant and 3% for placebo). Discontinuation rates due to adverse effects did not significantly differ between suvorexant, 15 to 20 mg, and placebo and no difference between groups in the proportions of participants reporting at least 1 adverse effect was seen.
Medications that pass through the blood brain barrier and block histamine 1 receptors are very sedating. Histamine (along with orexin) are the primary wakefulness promoting neurotransmitters.
Diphenhydramine (Benadryl) and doxylamine (Unisom) are the antihistamines most commonly used for insomnia. Both are available “over-the-counter” alone and in combination with non-prescription analgesics (e.g., Tylenol PM). Both medications have similar properties and the usual dose for each is 25–50 mg. Both have significant anticholinergic effects that are the primary source of side effects. Anticholinergic medications have recently been found to be associated with an increased risk of dementia in the elderly and are also associated with delirium, dry mouth, and constipation.
The only selective H1 antagonist that has been studied for the treatment of insomnia is doxepin (Sinequan), a tricyclic antidepressant which has FDA approval for the treatment of depression in dosages from 75–150 mg and for the treatment of insomnia in dosages from 3–6 mg. In addition, doxepin has been studied at intermediate doses of 25 – 50 mg where it has been shown to have effects on improving sleep onset, sleep duration and quality.
Doxepin is a more potent and selective H1 antagonist than any of the medications we refer to as antihistamines. As with diphenhydramine, low-dose doxepin has stronger effects on sleep maintenance than it has on sleep onset. The therapeutic effect on sleep maintenance has been demonstrated for both younger and older adults. Of particular note, doxepin is the only agent available which has been demonstrated to have therapeutic effects in the last third of the night, including the final hour of an eight hour sleep period, without significant morning adverse effects.
Many antidepressants are also used to enhance sleep. The agents we discuss here will only be those felt to have additional sleep enhancing effects beyond their role in resolving sleep disruption caused by depression.
One of the most commonly prescribed sleep medications (trazodone) also has the smallest evidence base for its use. trazodone has only been evaluated in a single large-scale placebo-controlled trial in patients with insomnia and in that study it failed to have significant sustained therapeutic effects compared to placebo.
Trazodone is metabolized to the wake-promoting mCPP molecule to a highly variable degree due to a genetic polymorphism that is not rare in the population.This may eliminate the therapeutic effects of trazodone for some patients and lead them to have distressing levels of anxiety.
Mirtazapine (Remeron) is believed to have sleep promoting effects related to its antagonism of serotonergic (5HT2 and 5HT3), adrenergic (α1), and histaminergic (H1) receptors. In addition, it is believed to antagonize adrenergic α2 receptors, which are presynaptic and inhibit the release of norepinephrine. As a result of this property, the sleep enhancing effects of mirtazapine are thought to decrease as the dose increases. While the range for antidepressant dosing is 7.5–45 mg, dosages below 30 mg are generally used to promote sleep.
There have been no placebo-controlled trials of mirtazapine for the treatment of insomnia. The evidence that mirtazapine may benefit sleep comes from a double-blind evaluation of mirtazapine vs. fluoxetine in patients with major depression, an open-label study of healthy volunteers without sleep complaints, preoperative patients at risk for insomnia due to surgery, and a pilot study of depressed patients.
Similar to the antidepressants used to enhance sleep, some antipsychotics are used “off-label” to treat insomnia. The antipsychotics most commonly used to treat insomnia are quetiapine (Seroquel) (dosed at 25–250 mg) and olanzapine (Zyprexa) (dosed at 2.5–20 mg).
Olanzapine has a tmax of 4–6 hours, making it better suited for the treatment of sleep maintenance problems than for sleep onset problems. Quetiapine on the other hand has a tmax of 1–2 hours and a half-life of 7 hours making it well suited for both sleep onset problems and sleep maintenance problems.
No placebo-controlled trials have been completed with these agents for the treatment of insomnia specifically, therefore the risk-benefit profile is difficult to assess. Evidence for a sleep enhancing effect of quetiapine (25–75 mg) was reported from open label studies with primary insomnia patients and with healthy volunteers while olanzapine has been noted to enhance sleep in an open-label study of healthy volunteers only.
Anticonvulsants such as gabapentin (Neurontin), pregabalin (Lyrica) and tiagabine (Gabitril), are sometimes used in the treatment of insomnia. Gabapentin and pregabalin bind to the alpha-2-delta subunit of N-type voltage-gated calcium channels, which decreases the activity of wake promoting glutamate and norepinephrine systems. Tiagabine enhances sleep by inhibiting the reuptake of GABA. Gabapentin has a relatively long tmax of 3 – 3.5 hours, making it relatively unlikely to facilitate sleep onset.
Gabapentin and pregabalin have been demonstrated to have sleep enhancing effects in a variety of populations including healthy volunteers, patients with restless legs syndrome, chronic pain patients and patients with partial seizures.
In placebo-controlled trials with primary insomnia patients, tiagabine has been found to increase slow-wave sleep, but has not been shown to improve sleep onset or sleep maintenance consistently.
The fact that all of these agents promote slow-wave sleep and have minimal adverse effects on REM sleep make them unusual and may represent a significant benefit to their use in chronic sleep disorders.
Gabapentin and pregabalin may be particularly helpful in treating sleep disruption associated with pain.
The most common side effects associated with gabapentin are ataxia and diplopia, while pregabalin is associated with dry mouth, cognitive impairment, peripheral edema and increased appetite. Tiagabine is most commonly associated with nausea. Of these medications, only pregabalin is associated with abuse potential and should be used with caution in patients who are prone to substance abuse.
Prazosin (Minipress) is an antihypertensive medication with relatively recently discovered benefits for sleep, primarily in those who experience frequent nightmares and sleep disturbance associated with post-traumatic stress disorder (PTSD). It is an α1 adrenoreceptor antagonist. Studies to date have generally prescribed 2–6 mg for most patients with an upper limit of 15–20 mg. The recommended starting dose is 1 mg to prevent hypotension, then the dose is slowly titrated upward until a therapeutic effect is achieved.105 In addition, it is important to warn patients that orthostatic hypotension is most likely to occur the morning after dose increases. Placebo-controlled trials have reported benefits of prazosin in the treatment of sleep disturbance and trauma-related nightmares in military veterans and civilians with PTSD.
Medications and Deep or Slow Wave Sleep
Although most people might assume there medication for insomnia will improve the quality of sleep and increase the amount of deep or slow wave sleep, most of the medications we discussed actually reduce slow-wave sleep. The significance of this is somewhat unclear but we do know that slow-wave sleep is associated with memory: reduced slow-wave sleep can lead to impaired memory. We also know that it is during slow-wave sleep that the brain releases the greatest amount of growth hormone which may play a role in repairing damage. Finally, we know that slow-wave sleep drops relatively dramatically beginning in midlife, particularly for men. All of this leads to the question of which medications actually increase slow-wave sleep. The table below shows that it’s a relatively select group of medications many of which are not traditionally used for the treatment of insomnia.
Tiagabine, gabapentin and pregabalin are discussed above, as are mirtazapine, olanzapine and trazodone.
One interesting medication that has not been discussed is gamma hydroxy-butyrate (GHB) or the related sodium oxybate (Xyrem). GHB became widely know as “the date rape drug” in the 80’s and 90’s. However, despite this negative publicity, it came to market as a medication for the treatment of narcolepsy because of its ability to treat this relatively rare sleep disorder.
It is a complicated medication to prescribe and to take because it is only distributed through one company which monitors its use extremely carefully to avoid diversion, however I have seen a few patients with severe insomnia not caused by narcolepsy who have had a positive response to the medication.
For More Information
Asnis GM, Thomas M, Henderson MA. Pharmacotherapy Treatment Options for Insomnia: A Primer for Clinicians. Hashimoto K, ed. International Journal of Molecular Sciences. 2016;17(1):50. doi:10.3390/ijms17010050.
de Almondes KM, Costa MV, Malloy-Diniz LF, Diniz BS. Insomnia and risk of dementia in older adults: Systematic review and meta-analysis. J Psychiatr Res. 2016 Jun;77:109-15. doi: 10.1016/j.jpsychires.2016.02.021. Epub 2016 Mar 8. Review. PubMed PMID: 27017287.
Krystal AD. A compendium of placebo-controlled trials of the risks/benefits of pharmacological treatments for insomnia: the empirical basis for US clinical practice. Sleep Med Rev. 2009; 13(4): 265. [PubMed: 19153052]
Lo HS, Yang CM, Lo HG, Lee CY, Ting H, Tzang BS. Treatment effects of gabapentin for primary insomnia. Clin Neuropharmacol. 2010 Mar-Apr;33(2):84-90. doi: 10.1097/WNF.0b013e3181cda242. PubMed PMID: 20124884.
McCall C, McCall WV. What is the role of sedating antidepressants, antipsychotics, and anticonvulsants in the management of insomnia? Curr Psychiatry Rep. 2012 Oct;14(5):494-502. doi: 10.1007/s11920-012-0302-y. Review. PubMed PMID: 22923053.
Minkel J, Krystal AD. Optimizing the Pharmacologic Treatment of Insomnia: Current Status and Future Horizons. Sleep medicine clinics. 2013;8(3):333-350. doi:10.1016/j.jsmc.2013.06.002.
Norman JL, Anderson SL. Novel class of medications, orexin receptor antagonists, in the treatment of insomnia – critical appraisal of suvorexant. Nat Sci Sleep. 2016 Jul 14;8:239-47. doi: 10.2147/NSS.S76910. eCollection 2016. Review. PubMed PMID: 27471419; PubMed Central PMCID: PMC4948724.
Roehrs T, Roth T. Drug-related Sleep Stage Changes: Functional Significance and Clinical Relevance. Sleep Med Clin. 2010 Dec 1;5(4):559-570. PubMed PMID: 21344068; PubMed Central PMCID: PMC3041980.
Wilt TJ, MacDonald R, Brasure M, Olson CM, Carlyle M, Fuchs E, Khawaja IS, Diem S, Koffel E, Ouellette J, Butler M, Kane RL. Pharmacologic Treatment of Insomnia Disorder: An Evidence Report for a Clinical Practice Guideline by the American College of Physicians. Ann Intern Med. 2016 Jul 19;165(2):103-12. doi: 10.7326/M15-1781. Epub 2016 May 3. PubMed PMID: 27136278.