Sleep
Everything in the library studied for it, medical and holistic together. Within each kind, what was shown to help comes first, strongest evidence first. What was tested and not shown to help stays on the list.
Strong several reviews agree · Moderate one review or consistent trials · Limited one or a few small studies · Untested no study in people, listed with what it rests on
Medication20Peptides and SARMs3Supplements19Therapies6Practices14Daily life9Food and drink1Devices and wearables2
Medication
Insomnia
Shown to lengthen sleep by about an hour and shorten the time to fall asleep by 4 to 20 minutes in short trials, with more daytime drowsiness.
The numbers and sources
- Effect
- Against placebo, benzodiazepines increased total sleep duration by 61.8 minutes (95% CI 37.4 to 86.2). Sleep latency on sleep records fell by 4.2 minutes, which was not significant, and by 14.3 minutes in patients' own estimates [9]. A later meta-analysis found sleep latency fell by 10.0 minutes on sleep-lab recordings (95% CI 3.4 to 16.6) and 19.6 minutes in sleep diaries, with a higher risk of harm than placebo (risk difference 0.15) [10]. A network meta-analysis found benzodiazepines among the drugs that beat placebo for acute insomnia, and among those with more side effects than placebo [11]. In people over 60, sedatives added 25.2 minutes of sleep and a small gain in sleep quality (effect size 0.14) [12].
- Population
- Randomized trials of benzodiazepines for insomnia [9]; 105 studies of drugs for chronic insomnia [10]; 44,089 participants in 154 trials of 30 drugs [11]; 2,417 people aged 60 and older in 24 studies [12].
- Sources
- [9], [10], [11], [12], [7]
- Caveats
- The pooled trials cover benzodiazepines as a class, including sleep drugs outside this page, and the abstracts do not break results down by drug. Of the four labels cited here, only lorazepam's mentions insomnia, as insomnia due to anxiety or passing stress [7]. Trials were short. See rebound insomnia under Risks.
Daytime sleepiness in narcolepsy
Shown to reduce daytime sleepiness in adults with narcolepsy, about 3 points on the Epworth scale and 3 minutes longer awake on a lab test, but not cataplexy.
Compared with placebo [1][2][3]; sodium oxybate in one comparison [1]
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- Against placebo, modafinil lowered Epworth Sleepiness Scale scores by 2.73 points (95% CI 2.08 to 3.39) and lengthened time to fall asleep on the Maintenance of Wakefulness Test by 2.82 minutes (95% CI 2.40 to 3.24). It did not change the number of cataplexy attacks a day. Its effect on sleepiness was similar to sodium oxybate's [1]. An updated meta-analysis found 3.56 more minutes on the Maintenance of Wakefulness Test (95% CI 2.25 to 4.86) and a 3.34-point drop on the Epworth scale (95% CI 2.56 to 4.13) [2]. In the two label trials, 58% to 72% on modafinil were rated improved against 37% to 38% on placebo [3].
- Population
- 1,054 patients in 9 randomized trials [1]; 997 adults in 5 trials pooled, out of 9 studies reviewed [2]; 558 adults in two 9-week US trials [3].
- Sources
- [1], [2], [3]
- Caveats
- The 2026 review notes the randomized trials date mostly from the 1990s and early 2000s, with no new eligible trials in the past decade, and that long-term efficacy is not established [2]. Patients on modafinil still averaged well below normal wakefulness on the lab test [3].
Sleepiness in sleep apnea despite CPAP
Shown to reduce residual sleepiness in sleep apnea, about 2 to 3 points on the Epworth scale, with about three times the adverse events of placebo.
Compared with placebo, on top of CPAP [3][4][5]
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- Modafinil or armodafinil improved Epworth Sleepiness Scale scores by 2.2 points (95% CI 1.5 to 2.9) and Maintenance of Wakefulness Test times by 3 minutes (95% CI 2.1 to 3.8) over placebo. Adverse events tripled and withdrawals for adverse events doubled, with no rise in hospitalizations or deaths [4]. A second meta-analysis found Epworth scores lower by 2.96 points with modafinil (95% CI 2.19 to 3.73) and 2.63 with armodafinil (95% CI 1.85 to 3.40) [5]. In the 12-week label trial, 61% on modafinil 200 mg and 68% on 400 mg were rated improved against 37% on placebo [3].
- Population
- 1,466 patients in 10 trials of at least 2 weeks [4]; 723 patients in 11 modafinil trials and 1,009 in 5 armodafinil trials [5]; 327 adults with sleep apnea, 80% fully using CPAP [3].
- Sources
- [3], [4], [5]
- Caveats
- The European Medicines Agency removed this indication in 2011 over an unfavorable balance of benefit and risk, which the first review sought to re-examine [4]. The drug treats sleepiness, not the airway obstruction [3].
Insomnia
Shown to improve self-rated sleep quality by a small to moderate amount in short trials of people with insomnia, with mixed results on measured sleep.
The numbers and sources
- Effect
- In seven placebo-controlled trials, people on trazodone rated their sleep quality higher (SMD -0.41, 95% CI -0.82 to -0.00) and woke less often (SMD -0.51, 95% CI -0.97 to -0.05). Sleep efficiency did not differ (SMD 0.09, 95% CI -0.19 to 0.38), nor did time to fall asleep or total sleep time [5]. A Cochrane review pooled three trazodone studies and found a similar gain in subjective sleep (SMD -0.34, 95% CI -0.66 to -0.02) and little or no difference in sleep efficiency on sleep-lab recordings (MD 1.38 percentage points, 95% CI -2.87 to 5.63) [6]. A meta-analysis of sleep-lab studies found longer total sleep time (39.88 minutes, 95% CI 14.44 to 65.32), a shorter time to persistent sleep (19.30 minutes) and fewer awakenings (SMD -0.67) [7]. A dose-response meta-analysis found improved subjective sleep quality (Cohen's d 0.67, 95% CI 0.36 to 0.98), with the benefit leveling off around 50 mg a day [8].
- Population
- 429 patients in 7 trials [5]; 370 participants in 3 pooled trazodone studies [6]; 466 participants with insomnia disorder in 11 RCTs [7]; 1,366 participants in 13 RCTs [8].
- Sources
- [5], [6], [7], [8]
- Caveats
- Not an approved use. Trials were small and short, and the Cochrane review rated the evidence low quality and found no evidence on long-term use [6]. Heterogeneity was high in the dose-response analysis (I² 81%) [8]. The analyses disagree on total sleep time [5][7].
Falling asleep faster
Shown to shorten the time to fall asleep by about 22 minutes on sleep-lab recordings in adults with insomnia, an effect the authors called small.
The numbers and sources
- Effect
- In 13 placebo-controlled trials submitted to FDA, eszopiclone, zaleplon and zolpidem together shortened sleep latency on polysomnography by 22 minutes (95% CI 11 to 33), with the 95% CI of the standardized mean difference running from -0.57 to -0.16. Self-rated sleep latency improved less (SMD -0.33, 95% CI -0.62 to -0.04). Latency fell more with larger doses, longer treatment, younger or female participants, and zolpidem [1]. A meta-analysis of 6 trials found one month of zolpidem reduced sleep latency against placebo [2]. In adults aged 60 and older, zaleplon reduced measured sleep onset latency by 21.63 minutes and self-rated latency by 15.86 minutes against control [3]. Across 75 sleep laboratory studies, zolpidem showed significant initial efficacy, with marginal tolerance over intermediate and long-term use [4].
- Population
- 4,378 participants in 13 trials [1]; 1,068 patients with insomnia disorder in 6 trials [2]; 5,917 older adults in 24 articles [3]; 1,276 people in 75 studies, 804 of them with insomnia [4].
- Sources
- [1], [2], [3], [4]
- Caveats
- The authors of [1] described the drug effect as small and of questionable clinical importance, and noted that the placebo response accounted for much of the total improvement. The abstract of [2] gives no effect sizes. Most trials lasted days to weeks.
Sleep, low-dose quetiapine
Shown to improve sleep quality and add about 48 minutes of sleep against placebo, mostly in people with anxiety or depression, with frequent side effects.
The numbers and sources
- Effect
- In a meta-analysis of 21 trials, quetiapine improved sleep quality against placebo (SMD -0.57, 95% CI -0.75 to -0.40), with significant effects in the generalized anxiety, major depression and healthy subgroups and at 50, 150 and 300 mg. Total sleep time rose by 47.91 minutes against placebo (95% CI 28.06 to 67.76) and did not differ from other psychiatric drugs (-4.19 minutes, 95% CI -19.43 to 11.05). Adverse events and discontinuation because of them were common [13].
- Population
- 21 clinical trials, mainly in people with generalized anxiety disorder or major depressive disorder and in healthy volunteers [13].
- Sources
- [13]
- Caveats
- Not an approved use. Most trials measured sleep as a secondary outcome in people treated for another condition, and results in older people were highly heterogeneous. See the block on low-dose quetiapine and cardiovascular events below.
Insomnia
Shown to lower insomnia severity by about 6 points more than placebo at 6 weeks on low-dose mirtazapine, with no difference from placebo from 12 weeks.
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- In adults with insomnia disorder in general practice, low-dose mirtazapine lowered Insomnia Severity Index scores by 6.0 points more than placebo at 6 weeks (95% CI 3.0 to 9.0). Improvement and recovery rates were 52% and 56% against 14% on placebo. From 12 weeks onward scores did not differ from placebo [3]. In adults aged 65 and over, 28 days of mirtazapine 7.5 mg lowered scores by 6.5 points against 2.9 on placebo, and improved self-rated time awake in the night, total sleep time and sleep efficiency. Six people on mirtazapine and one on placebo stopped because of adverse events [4].
- Population
- 80 adults with insomnia disorder and trouble staying asleep [3]; 60 adults aged 65 and over with chronic insomnia [4].
- Sources
- [3], [4]
- Caveats
- Two small trials. Not an approved use. The authors of [3] concluded the results do not support use for several months.
Sleepiness in shift work disorder
Shown to reduce sleepiness on night shifts in people diagnosed with shift work disorder, about 1 point on a 10-point scale. No trials tested shift workers without the disorder.
Compared with placebo [6][7][8]
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- Armodafinil taken before the night shift lowered Karolinska Sleepiness Scale scores by 0.99 points (95% CI 0.67 to 1.32, range 1 to 10) and sped simple reaction time by 50 milliseconds at 3 months. Modafinil lowered the same score by 0.90 points (95% CI 0.35 to 1.45) [7]. In the pivotal 3-month trial, modafinil 200 mg lengthened nighttime time to fall asleep by 1.7 minutes against 0.3 on placebo, and 74% against 36% were rated clinically improved [8]. The armodafinil label reports the same kind of trial results for 150 mg [6].
- Population
- 2 armodafinil trials with 572 participants and 1 modafinil trial with 183, within a Cochrane review of 15 trials and 718 shift workers [7]; 209 patients with shift work disorder [8].
- Sources
- [6], [7], [8]
- Caveats
- The trial authors noted that treated patients remained very sleepy at night [8]. The Cochrane review found no trials in shift workers without a diagnosed disorder, and rated the evidence moderate quality [7].
Fibromyalgia
Shown to improve overall fibromyalgia symptoms with cyclobenzaprine, with about 5 treated for one extra person improved, and no change in fatigue or tender points.
The numbers and sources
- Effect
- Across 5 placebo-controlled trials, people on cyclobenzaprine were more likely to report overall improvement (OR 3.0, 95% CI 1.6 to 5.6), a risk difference of 0.21 (95% CI 0.09 to 0.34) and a number needed to treat of 4.8 (95% CI 3.0 to 11). Pain improved early in treatment and sleep improved. Fatigue and tender points did not improve at any time [9].
- Population
- People with fibromyalgia in 5 randomized, placebo-controlled trials of cyclobenzaprine [9].
- Sources
- [9]
- Caveats
- One meta-analysis of 5 trials, of cyclobenzaprine only. Not an FDA-approved use [10].
Obstructive sleep apnea
Shown to cut sleep apnea events by about 13 to 25 per hour more than placebo in adults with obesity, with smaller drops in those with type 2 diabetes.
The numbers and sources
- Effect
- At 80 weeks, in people without diabetes, the apnea-hypopnea index fell by 13.2 (4 mg), 24.7 (9 mg) and 22.5 (12 mg) events per hour more than placebo by intention to treat. In people with type 2 diabetes the differences were 8.8, 13.9 and 15.0 events per hour [8].
- Population
- 243 adults without diabetes (baseline 58.6 events per hour) and 448 with type 2 diabetes (55.9 events per hour), with moderate to severe sleep apnea and BMI of 27 or more [8].
- Sources
- [8]
- Caveats
- Two trials nested inside the larger weight trials, both funded by Eli Lilly [8].
Self-rated sleep quality
Shown to improve self-rated sleep quality over placebo in short trials of eszopiclone and zolpidem, by more than zaleplon, melatonin or ramelteon.
The numbers and sources
- Effect
- In a network meta-analysis of acute treatment, benzodiazepines, doxylamine, eszopiclone, lemborexant, seltorexant, zolpidem and zopiclone each improved self-rated sleep quality more than placebo (SMD range 0.36 to 0.83, certainty high to moderate). Benzodiazepines, eszopiclone, zolpidem and zopiclone improved it more than melatonin, ramelteon and zaleplon (SMD 0.27 to 0.71, certainty moderate to very low). Zolpidem caused more dropouts due to adverse events than placebo (OR 1.79, 95% CI 1.25 to 2.50) [8]. One month of zolpidem improved sleep quality against placebo in a second meta-analysis [2]. In people aged 60 and over, sedative hypnotics of all kinds improved sleep quality with an effect size of 0.14 [5].
- Population
- 44,089 adults with insomnia disorder in 154 double-blind trials of 30 interventions [8]; 1,068 patients in 6 trials [2]; 2,417 participants aged 60 or over in 24 studies [5].
- Sources
- [8], [2], [5]
- Caveats
- The SMD ranges cover several drug classes, and the abstract does not give the figure for each Z-drug. Data on zaleplon were described as scarce [8]. The effect in older adults was small [5].
Insomnia treated for six months or longer
Shown to keep improving sleep against placebo across 6 to 8 months of use in four trials of eszopiclone and zolpidem in adults with chronic insomnia.
The numbers and sources
- How common
- Very common: more than 1 in 10 people.
- Effect
- Over 6 months of nightly eszopiclone 3 mg, sleep latency, wake time after sleep onset, total sleep time and sleep quality all stayed better than placebo at every month (P 0.003 or lower), with no evidence of tolerance [9]. In a second 6-month trial, Insomnia Severity Index scores fell below clinically meaningful levels in 50% on eszopiclone and 19% on placebo [10]. Extended-release zolpidem 12.5 mg taken 3 to 7 nights a week was rated as helping sleep by 89.8% at week 12 against 51.4% on placebo, and by 92.3% against 59.7% at week 24 [11]. Over 8 months of nightly zolpidem, total sleep time and sleep efficiency rose and sleep latency and wake after sleep onset fell against placebo on sleep-lab recordings at months 1 and 8 [12]. In a network meta-analysis of long-term treatment, eszopiclone improved sleep quality more than placebo (SMD 0.63, 95% CI 0.36 to 0.90) and more than zolpidem (SMD 0.60, 95% CI 0.00 to 1.20), both rated very low certainty [8].
- Population
- 788 adults aged 21 to 69 with primary insomnia, 593 on eszopiclone [9]; 830 adults with primary insomnia [10]; 1,018 adults aged 18 to 64, 669 on zolpidem [11]; 91 adults aged 23 to 70 with primary insomnia [12]; long-term trials within 154 pooled trials [8].
- Sources
- [9], [10], [11], [12], [8]
- Caveats
- In the 8-month trial, self-rated sleep did not differ between zolpidem and placebo [12]. In the 6-month eszopiclone trial, 12.8% on the drug and 7.2% on placebo left because of an adverse reaction [7]. The zolpidem label describes its supporting trials as 4 to 5 weeks long and its approved use as short-term [13].
Long COVID
Limited evidence: fatigue, brain fog, sleep and pain improved from baseline in uncontrolled studies of long COVID. No randomized trials were found.
The numbers and sources
- Effect
- A 2026 systematic review screened 397 records and found no randomized trials. In 4 pre-post studies (155 people) at 1 to 6 mg a day, fatigue (Hedges' g -0.74, 95% CI -1.11 to -0.37), brain fog (g -0.53), sleep (g -0.60), pain (g -0.93) and daily function (g -0.93) improved from baseline [12].
- Population
- 155 people with long COVID in the US and Ireland in uncontrolled studies [12].
- Sources
- [12]
- Caveats
- The review rated certainty as low, and its studies had no control group [12]. No peer-reviewed randomized trial was located.
Sleep apnea and snoring in children
Small trials found montelukast improved sleep study measures, snoring and mouth breathing in children with mild to moderate obstructive sleep apnea over short periods.
Compared with placebo [7]; nasal mometasone alone [7]
The numbers and sources
- Effect
- Against placebo, oral montelukast improved polysomnography measures, snoring, mouth breathing and adenoid size. Added to mometasone nasal spray, it improved sleep study measures and snoring more than the spray alone [7].
- Population
- 305 children with mild to moderate obstructive sleep apnea in 4 randomized trials [7].
- Sources
- [7]
- Caveats
- Four small short-term trials. The abstract gives no effect sizes [7].
Sleep in dementia
One trial found trazodone 50 mg added about 42 minutes of night sleep over two weeks in 30 people with moderate to severe Alzheimer's disease.
The numbers and sources
- Effect
- Total night-time sleep rose by 42.46 minutes against placebo (95% CI 0.9 to 84.0) and sleep efficiency by 8.53 percentage points (95% CI 1.9 to 15.1). Night-time awakenings and daytime sleep changed little. No serious adverse effects were reported [9].
- Population
- 30 people with moderate to severe Alzheimer's disease in one trial within a Cochrane review [9].
- Sources
- [9]
- Caveats
- One small trial of two weeks, rated low certainty [9].
Sleep
Mixed results: gabapentin improved sleep measures in people with medical illness in one meta-analysis, while a review of trials in insomnia called the results inconclusive.
The numbers and sources
- Effect
- In people with sleep disturbance alongside a medical illness, gabapentin beat placebo on five of six composite sleep endpoints, for example SMD 0.50 (95% CI 0.28 to 0.71) on one. People on gabapentin left treatment more often (RR 1.38, 95% CI 1.08 to 1.76), at an average dose of about 1,800 mg a day [14]. A systematic review of gabapentin and pregabalin for insomnia found inconclusive results and minimal evidence [13].
- Population
- 4,684 participants in 26 studies of people with medical illness [14]; trials of gabapentinoids in insomnia [13].
- Sources
- [14], [13]
- Caveats
- Not an approved use. The positive analysis was in people whose poor sleep came with another illness, and its authors asked for trials in primary sleep disorders [14].
Total sleep time and waking in the night
Mixed results: one meta-analysis found a month of zolpidem lengthened total sleep, while pooled FDA trial data found no significant change in sleep time or awakenings.
The numbers and sources
- Effect
- In the FDA trial data, waking after sleep onset, number of awakenings, total sleep time and sleep efficiency did not differ significantly from placebo, and the authors said too few studies reported these outcomes for firm conclusions [1]. One month of zolpidem increased total sleep time against placebo, with no significant difference in waking after sleep onset [2]. In adults aged 60 and older, eszopiclone added 26.69 minutes of measured total sleep time against placebo [3]. Across sedative hypnotics of all kinds in people aged 60 and over, total sleep time rose by a mean of 25.2 minutes and night awakenings fell by 0.63 [5]. The zaleplon label states it has not been shown to increase total sleep time or decrease the number of awakenings [6]. The eszopiclone label reports that 3 mg reduced waking after sleep onset against placebo in adults [7].
- Population
- 4,378 participants in 13 trials [1]; 1,068 patients in 6 trials [2]; 5,917 older adults in 24 articles [3]; 2,417 participants aged 60 or over in 24 studies [5].
- Sources
- [1], [2], [3], [5], [6], [7]
- Caveats
- [5] pooled benzodiazepines with zaleplon, zolpidem and zopiclone and does not report the newer drugs alone. Results differ by drug: the evidence for longer sleep is for eszopiclone and zolpidem, not zaleplon.
Sleep (diphenhydramine)
Small trials found diphenhydramine gave modest sleep gains in mild insomnia, such as better sleep efficiency or fewer awakenings, with no change on overnight sleep recordings.
The numbers and sources
- Effect
- In adults with mild insomnia, diphenhydramine 50 mg raised diary sleep efficiency more than placebo and showed a trend toward longer sleep in the first 14 days, with lower insomnia severity ratings. Polysomnography showed no difference in sleep continuity, and no rebound insomnia followed stopping [8]. In older adults, diphenhydramine 50 mg reduced only the number of awakenings (1.7 against 2.0 a night) compared with placebo [9]. The American Academy of Sleep Medicine suggested clinicians not use diphenhydramine for sleep onset or maintenance insomnia [10].
- Population
- 184 adults, mean age 44, over 14 days [8]; older adults with insomnia, mean age 73.9, in a 14-night crossover trial [9]; guideline review of trials [10].
- Sources
- [8], [9], [10]
- Caveats
- Few small trials. Many sleep-aid products combine diphenhydramine with other drugs.
Sleep
Small trials found mixed results for falling asleep, staying asleep and sleep quality on 25 to 100 mg of hydroxyzine at bedtime.
The numbers and sources
- Effect
- A systematic review found five studies, one open-label and four randomized, of hydroxyzine 25, 50 or 100 mg at bedtime. Results were mixed for sleep onset, sleep maintenance and sleep quality. Dry mouth was the most common adverse effect, and four of the five studies did not report safety outcomes [11].
- Population
- 207 adults in 5 studies [11].
- Sources
- [11]
- Caveats
- Not an approved use. Small, short studies with no pooled estimate and no long-term data.
Sleep and insomnia
Not studied in people with insomnia. A Cochrane review found no trials of amitriptyline for insomnia, though sedation is one of its side effects.
No study in people. This rests on how it acts in the body.
The numbers and sources
- Effect
- A Cochrane review of antidepressants for insomnia found no evidence for amitriptyline, despite its common use [12]. The label describes amitriptyline as an antidepressant with sedative effects [13], and depression trials found more sedation than on placebo [1].
- Population
- 23 trials with 2,806 participants of other antidepressants for insomnia [12].
- Sources
- [12], [13], [1]
- Caveats
- Sedation in depression trials does not show better sleep in people with insomnia.
Peptides and SARMs
Sleep
Small trials of the natural hormone GHRH, not sermorelin, found more deep sleep when it was given in pulses at night, but not as a continuous infusion.
Compared with placebo [7][8]; the same men's sleep without GHRH, as the abstract describes it [9]
The numbers and sources
- Effect
- In 7 healthy men, four 50 microgram IV boluses of GHRH through the night raised slow-wave sleep to 20.2% of the night, against 14.0% on placebo [7]. In healthy volunteers, the same pulsed schedule increased slow-wave sleep (P < 0.01) and REM sleep (P < 0.05) against placebo, while a continuous infusion of a similar total dose did not change sleep significantly [8]. In healthy young men, a single IV dose in the first deep-sleep period did not change slow-wave sleep but increased REM sleep. Given in the third REM period, it was followed by an almost 10-fold rise in slow-wave sleep and less wake time [9].
- Population
- 7 healthy men [7]. Healthy volunteers [8]. Healthy young men [9].
- Sources
- [7], [8], [9]
- Caveats
- These trials gave GHRH intravenously, not sermorelin, the 1-29 fragment of GHRH that the trials on this page gave under the skin [2][4]. They were small, lasted one night each, and the effect depended on timing and on pulsed rather than continuous dosing [8][9].
Sleep
We haven't sourced research on this yet.
Sleep and insomnia
We haven't sourced research on this yet.
Supplements
Sleep
Shown to improve overall sleep in adults, more in people with insomnia, at 600 mg a day or more, and after 8 weeks or longer.
The numbers and sources
- Effect
- In 5 RCTs (400 participants), ashwagandha extract improved overall sleep (SMD -0.59, 95% CI -0.75 to -0.42, I² 62%). Effects were larger in adults diagnosed with insomnia, at 600 mg a day or more, and at 8 weeks or longer. Mental alertness on rising and anxiety improved, and quality of life did not change [5]. In people with mental disorders, sleep quality improved (SMD -1.35, 95% CI -1.79 to -0.91) [2].
- Population
- Adults 18 and older in 5 RCTs [5]. People with mental disorders [2].
- Sources
- [5], [2]
- Caveats
- Five trials in [5], whose authors write that data on serious adverse effects are limited. The mental disorders review describes a role in insomnia as still to be clarified [2].
Restless legs
Shown to lower restless legs scores about 5 points and double the share who improved, pooling oral and intravenous iron.
The numbers and sources
- Effect
- In 12 RCTs (511 patients), iron lowered the International RLS score by 5.28 points against placebo (95% CI -7.66 to -2.90), doubled the number of patients whose score improved (RR 2.06, 95% CI 1.49 to 2.84), and improved RLS quality of life (MD 7.42) and a sleep scale (MD -24.83) [9].
- Population
- Patients with restless legs syndrome [9].
- Sources
- [9]
- Caveats
- Oral and intravenous iron were pooled, and the authors single out intravenous ferric carboxymaltose. Overall adverse events were twice as common on iron (RR 2.04, 95% CI 1.46 to 2.85), with no difference in serious events [9].
Sleep
Shown to improve self-rated sleep onset, daytime function and sleep quality modestly in people of all ages.
The numbers and sources
- Effect
- In 18 RCTs pooled (897 participants across 19 articles), theanine improved self-rated sleep onset latency (SMD 0.15, 95% CI 0.01 to 0.29), daytime dysfunction (SMD 0.33, 95% CI 0.16 to 0.49) and overall sleep quality score (SMD 0.43, 95% CI 0.04 to 0.83) [6]. In a 4-week crossover trial of 200 mg a day in 30 adults, PSQI scores for sleep latency, sleep disturbance and use of sleep medication fell compared with placebo [7].
- Population
- People of all ages and health status [6]. 30 adults without major psychiatric illness [7].
- Sources
- [6], [7]
- Caveats
- The review authors note a lack of studies on pure theanine [6]. The crossover trial was funded by Taiyo Kagaku, which supplied the theanine, and two authors were its employees [7].
Time to fall asleep and total sleep in sleep disorders
Shown to shorten time to fall asleep about 7 minutes and add about 8 minutes of sleep in people with primary sleep disorders.
The numbers and sources
- Effect
- Across 19 studies (1,683 subjects) with primary sleep disorders, melatonin reduced sleep onset latency by 7.06 minutes (95% CI 4.37 to 9.75), increased total sleep time by 8.25 minutes (95% CI 1.74 to 14.75), and improved sleep quality (SMD 0.22, 95% CI 0.12 to 0.32). Longer trials and higher doses showed larger effects on latency and total sleep, and the effects did not fade with continued use [1].
- Population
- Adults and children with primary sleep disorders [1].
- Sources
- [1]
- Caveats
- The authors note the absolute benefit is smaller than with other insomnia drugs [1].
Sleep quality in adults with other conditions
Shown to improve sleep quality scores about 1.2 points in adults with other conditions, with results varying by condition and no clear change in cancer.
The numbers and sources
- Effect
- Across 23 RCTs, Pittsburgh Sleep Quality Index scores improved by 1.24 points (95% CI -1.77 to -0.71, I² 80.7%). Improvements were 2.20 points in respiratory diseases, 2.74 in metabolic disorders and 0.67 in sleep disorders, with no effect found in mental disorders or neurodegenerative diseases [2]. Later reviews in neurodegenerative disease differ. In Parkinson's disease, PSQI scores improved by 1.75 points across 7 RCTs (409 patients, 95% CI -2.94 to -0.55) [3], and by 2.90 points at doses above 4 mg but not at 4 mg or less across 5 RCTs (206 patients) [4]. In 10 RCTs of adults 65 and over with mild cognitive impairment or dementia (516 people), total sleep time rose by 12.4 minutes (95% CI 8.1 to 16.7) [5]. In 7 pooled trials in people with cancer, sleep quality did not change significantly (SMD -0.85, 95% CI -2.46 to 0.76, I² 94.5%) [6].
- Population
- Adults with respiratory, metabolic, sleep and other disorders [2]; people with Parkinson's disease [3][4]; older adults with cognitive impairment [5]; people with cancer [6].
- Sources
- [2], [3], [4], [5], [6]
- Caveats
- Heterogeneity was high, and health status explained part of it [2]. The Parkinson's improvements in subjective sleep quality did not reach the minimal clinically important difference [3]. Heterogeneity was 72% for total sleep time in [5] and 94.5% in [6].
Jet lag
Shown to reduce jet lag after flights crossing five or more time zones when taken near bedtime at the destination. Doses above 5 mg did no better.
The numbers and sources
- Effect
- Nine of ten trials found melatonin taken near target bedtime at the destination (10 pm to midnight) reduced jet lag after flights crossing five or more time zones. Doses of 0.5 to 5 mg gave similar results, people fell asleep faster after 5 mg than 0.5 mg, and doses above 5 mg did no better. The number needed to treat was 2. Benefit was smaller for westward flights, and dosing early in the day was liable to cause sleepiness and delay adaptation [7].
- Population
- Airline passengers, airline staff and military personnel in 10 RCTs [7].
- Sources
- [7]
- Caveats
- The review dates from 2002. One of the ten trials had a design fault and could not be used [7].
Sleep in children
Shown to improve sleep onset or sleep time in children with autism, cerebral palsy, atopic dermatitis or other neurologic conditions. No data in typically developing young children.
The numbers and sources
- Effect
- In children with autism, melatonin improved sleep quality (Hedges' g 0.75) and total sleep time (g 0.58) [8]. In 3 RCTs in children with atopic dermatitis, 3 to 6 mg a day shortened sleep onset latency (SMD -0.63, 95% CI -1.00 to -0.26) without a significant change in total sleep time (18.29 minutes, 95% CI -10.31 to 46.88) [9]. In 120 children aged 3 to 12 with cerebral palsy, 3 to 10 mg a day for 12 weeks gave 1.31 hours more total sleep and 27.72 minutes faster sleep onset than placebo [10]. A systematic review of children aged 0 to 6 found 6 trials (167 children), all in children with neurologic conditions, that reported improved sleep onset with few adverse events [11].
- Population
- Children with autism [8], atopic dermatitis [9], cerebral palsy [10] or other neurologic conditions [11].
- Sources
- [8], [9], [10], [11]
- Caveats
- No efficacy data were found for young children with typical development, and long-term outcomes were absent [11]. The atopic dermatitis review pooled only 3 trials [9].
Self-reported sleep improvement
Shown to raise the share of people reporting better sleep, about 37% more than placebo, in adults with or without sleep problems. Most studies had method problems.
The numbers and sources
- Effect
- In 18 RCTs, the share of people reporting improved sleep was higher with valerian than placebo (RR 1.37, 95% CI 1.05 to 1.78) [1]. An earlier meta-analysis of 16 studies (1,093 patients) found a similar result (RR 1.8, 95% CI 1.2 to 2.9) with evidence of publication bias [2]. A 2020 meta-analysis of 10 studies (1,065 participants) on subjective sleep quality linked inconsistent results across trials to variable extract quality [3].
- Population
- Healthy adults and people with sleep disturbance or insomnia [1][2][3].
- Sources
- [1], [2], [3]
- Caveats
- Doses, preparations and durations varied widely, and most studies had methodological problems [2]. These are yes-or-no or self-rated outcomes.
Sleep
Small trials found 5-HTP improved sleep scores in older adults who slept poorly, raised REM sleep in Parkinson's disease, and cut sleep terrors in children.
The numbers and sources
- Effect
- In 30 older adults given 100 mg a day or nothing for 12 weeks, poor sleepers on 5-HTP improved their Pittsburgh Sleep Quality Index global score by 2.80 points (p = 0.005), while good sleepers did not change [11]. In 18 people with Parkinson's disease and REM sleep behavior disorder, 50 mg a day for 4 weeks raised the share of REM sleep over placebo, with no rise in episodes and a non-significant fall in arousals and waking after sleep onset [12]. In 45 children with sleep terrors, 29 of 31 (93.5%) given 2 mg/kg at bedtime for 20 days responded at 1 month, against 4 of 14 (28.6%) untreated, and 26 of 31 (83.9%) were free of sleep terrors at 6 months [13].
- Population
- 30 older adults in Singapore, mean age 66 [11]. 18 adults with Parkinson's disease [12]. 45 children aged 3 to 10 [13].
- Sources
- [11], [12], [13]
- Caveats
- The older adult trial was single-blinded with an untreated control, not placebo, and the benefit appeared only in the poor sleeper subgroup [11]. The Parkinson's trial was small and short [12]. The sleep terror trial was open label with no placebo [13]. No trial in adults with insomnia was located.
Sleep quality and next-day fatigue
Small trials found 3 g of glycine before bed improved self-rated sleep and next-day fatigue in people with poor or shortened sleep.
The numbers and sources
- Effect
- In people with ongoing unsatisfactory sleep, 3 g of glycine before bed improved self-rated sleep quality and sleep efficiency and shortened polysomnographic latency to sleep onset and to slow-wave sleep, without changing sleep architecture. Daytime sleepiness fell and memory recognition improved [1]. In healthy men whose sleep was cut by 25% for three nights, 3 g at bedtime reduced fatigue on a visual analog scale, showed a trend toward less sleepiness, and improved psychomotor vigilance [2]. A 2026 scoping review found 3 glycine sleep trials, all reporting a benefit on at least one sleep outcome [3].
- Population
- Volunteers with ongoing unsatisfactory sleep [1]. 10 healthy men aged 30 to 61, all employees of Ajinomoto Co. [2].
- Sources
- [1], [2], [3]
- Caveats
- Few, small trials. The lead author of the restriction trial was affiliated with Ajinomoto Co. [2]. The scoping review did not pool effect sizes [3].
Sleep
Small trials found sleep onset about 17 minutes faster in older adults with insomnia, with no significant change in total sleep. Quality was rated low.
The numbers and sources
- Effect
- In 3 RCTs (151 older adults), sleep onset was 17.36 minutes shorter on magnesium than placebo (95% CI -27.27 to -7.44). Total sleep time rose by 16.06 minutes, which was not statistically significant [7].
- Population
- Older adults with insomnia in three countries [7].
- Sources
- [7], [8]
- Caveats
- All three trials were at moderate to high risk of bias and the evidence was rated low to very low quality [7]. A 2026 systematic review of 12 RCTs in adults could not pool them. Findings were inconsistent across self-reported sleep, EEG, actigraphy and wearable measures, and certainty was low to very low [8].
Shift work
One trial found better sleep quality on days off in nurses on shift work who took 0.3 mg on nights off, with no change in sleep duration.
The numbers and sources
- Effect
- In 46 nurses on fixed shifts, 0.3 mg of melatonin taken on nights off improved self-reported sleep quality on days off by 35.33% (p < 0.001) and lowered climacteric symptoms by 15.8% against placebo (p = 0.01). Total sleep duration did not change [12].
- Population
- Climacteric women working morning, afternoon or night shifts in one hospital [12].
- Sources
- [12]
- Caveats
- One small trial, retrospectively registered. Melatonin was taken on nights off, so the trial does not show effects on sleep after night shifts.
Sleep
One trial found self-rated sleep improved on two forms of phenibut in older adults with insomnia, with no placebo comparison.
The numbers and sources
- Effect
- In a randomized trial in adults with insomnia and sensitivity to weather, phenibut taken as capsules and phenibut taken as powder emptied from the capsules for 4 weeks both improved total scores on a subjective sleep questionnaire (p = 0.00001 and p = 0.000001). The powder group scored higher on sleep quality (p = 0.00027) and quality of awakening (p = 0.00001) [4].
- Population
- 67 adults aged 60 to 70, 33 on capsules and 34 on powder [4].
- Sources
- [4]
- Caveats
- Both groups took phenibut, so the trial compared two forms of it. It did not compare phenibut with placebo or no treatment. Sleep was self-rated. Published in Russian with an English abstract [4].
Sleep
Limited evidence: a review found very limited support for sleep. One trial found faster sleep onset on GABA than placebo.
The numbers and sources
- Effect
- The same review found very limited evidence for sleep [1]. In a single-blind crossover trial, GABA shortened the time to fall asleep and increased non-REM sleep on EEG compared with placebo [4].
- Population
- Adults [1][4].
- Sources
- [1], [4]
- Caveats
- The abstract of [4] does not give participant numbers or effect sizes.
Sleep
Mixed results: one trial found better sleep quality on kava in people with anxiety-related sleep problems, while a larger internet trial found no relief of insomnia over placebo.
The numbers and sources
- Effect
- In 61 patients with sleep disturbances tied to anxiety, 200 mg a day of kava extract WS 1490 for 4 weeks improved self-rated sleep quality (p = 0.007) and feeling rested after sleep (p = 0.018) more than placebo, with no drug-related adverse events [8]. In 391 adults with anxiety and insomnia recruited online, neither kava nor valerian relieved anxiety or insomnia more than placebo over 28 days [2].
- Population
- 61 patients with anxiety-related sleep disturbances in a multicenter trial [8]. 391 adults with anxiety and insomnia in the US [2].
- Sources
- [8], [2]
- Caveats
- Both trials lasted 4 weeks and measured sleep by questionnaire. In the larger trial, the primary sleep comparison was for valerian, and the placebo group improved by 8.3 points on the insomnia index [2].
Sleep
Small trials found mixed results: one found longer sleep over 7 days, while two others found no change in sleep.
The numbers and sources
- Effect
- In 20 adults, 7 days of concentrate raised urinary melatonin and increased time in bed, total sleep time and sleep efficiency [7]. In 44 adults over 30 days, sleep time and quality did not differ from placebo [8]. In 34 adults with overweight or obesity, 500 mg capsules for 14 days changed no sleep or inflammation measure [9]. A 2025 review of 7 studies found 3 reporting improved sleep measures [10].
- Population
- Healthy adults [7][8]; adults with overweight or obesity and sleep complaints [9].
- Sources
- [7], [8], [9], [10]
- Caveats
- No meta-analysis. The two null trials were funded by the Cherry Marketing Institute [8][9].
Time to fall asleep and measured sleep
Not shown to shorten time to fall asleep or improve measured sleep in adults with or without sleep problems.
The numbers and sources
- Effect
- In 18 RCTs, time to fall asleep differed from placebo by 0.70 minutes (95% CI -3.44 to 4.83), and visual-scale sleep quality did not differ (SMD -0.02, 95% CI -0.35 to 0.31) [1]. A review of 29 controlled trials found most showed no difference from placebo, and none of the most recent, most rigorous trials found an effect on sleep [4].
- Population
- Healthy adults and people with sleep disturbance or insomnia [1][4].
- Sources
- [1], [4]
- Caveats
- The reviews disagree in their reading of the same body of trials [1][4][3].
Sleep
Small trials found CBD did not change total sleep time or significantly improve sleep disturbance, with a small gain in deep plus REM sleep in one.
The numbers and sources
- Effect
- In 125 people with insomnia, 300 mg of CBD with terpenes on at least 4 nights a week for 4 weeks raised the share of the night in slow-wave plus REM sleep by 1.3 percentage points over placebo (95% CI 0.1 to 2.5), measured by a wrist-worn tracker. Total sleep time did not change [9]. In 63 adults with high trait worry, 300 mg a day for 2 weeks did not improve sleep disturbance significantly more than placebo (p 0.10) [10].
- Population
- Adults with insomnia in a crossover trial [9]. Adults reporting elevated worry [10].
- Sources
- [9], [10]
- Caveats
- Two small trials. [9] tested a CBD and terpene formulation, not CBD alone, and used a consumer sleep tracker rather than a sleep lab.
Sleep
Not studied in people. Used in China for insomnia for hundreds of years. In rats, an extract lengthened total and non-REM sleep in one study but not another.
No study in people. This rests on animal studies, documented traditional use.
The numbers and sources
- Effect
- Reishi has been used in China as a tranquilizing agent for restlessness, insomnia and palpitations for hundreds of years [10]. In freely moving rats, 3 days of an oral fruiting body extract at 80 mg/kg increased total sleep time and non-REM sleep, an effect blocked by an antibody to TNF-α [10]. In another study, a water extract at 80 and 120 mg/kg did not change sleep architecture in normal rats, but shortened sleep onset and lengthened sleep in rats given pentobarbital, an effect partly blocked by the benzodiazepine antagonist flumazenil [11].
- Population
- Rats and mice [10][11].
- Sources
- [10], [11]
- Caveats
- Rodent studies with short dosing, and the two disagree on sleep in normal animals. No sleep trial in people was located.
Therapies
Insomnia
Shown to cut time to fall asleep by about 19 minutes and time awake at night by about 26 minutes against inactive comparisons, with little change in total sleep time.
Compared with inactive comparators [9]; no treatment [10]
The numbers and sources
- Effect
- After face-to-face CBT for insomnia, sleep onset latency improved by 19.03 minutes (95% CI 14.12 to 23.93), wake after sleep onset by 26.00 minutes, and sleep efficiency by 9.91%. Total sleep time rose 7.61 minutes, a change that was not significant (95% CI -0.51 to 15.74) [9]. Across 87 trials, insomnia severity improved (g 0.98), as did sleep efficiency (g 0.71) and sleep quality (g 0.40). The smallest effect was on total sleep time (g 0.16) [10].
- Population
- 1,162 adults with chronic insomnia, 64% women, mean age 56, in 20 trials [9]. 3,724 treated patients and 2,579 controls in 87 trials [10].
- Sources
- [9], [10]
- Caveats
- Sleep was mostly measured by diary. The first analysis excluded insomnia alongside other medical, sleep or psychiatric disorders [9].
Sleep quality
Shown to improve sleep quality scores in pooled trials, mostly against usual care or no intervention, with large effects.
Compared with control groups the source abstract does not describe [11]; nonactive controls: sham intervention, usual care, waiting list or psychoeducation [10]; control groups the source abstract does not describe [12]
The numbers and sources
- Effect
- Across 42 trials, sleep quality scores improved more than with controls (Hedges' g -1.37, 95% CI -1.81 to -0.94), and people given reflexology were less likely to still have sleep problems (OR 0.25, 95% CI 0.19 to 0.31) [11]. A second meta-analysis found improved sleep quality (g -1.665, 95% CI -2.361 to -0.970), larger with longer and more total reflexology time [10]. An earlier review found a weighted effect of 1.19 for sleep [12].
- Population
- 3,928 adults in 42 trials [11]. 2,366 adults in 26 trials [10]. 18 studies on sleep [12].
- Sources
- [11], [10], [12]
- Caveats
- Trials were heterogeneous and mostly unblinded. The 42-trial review included studies in English and Chinese only [11].
Insomnia, online CBT
Shown to improve insomnia severity, sleep efficiency and sleep onset when delivered over the internet, by amounts similar to those reported for face-to-face CBT.
Compared with control conditions the source abstract does not describe [12]
The numbers and sources
- Effect
- Internet-delivered CBT for insomnia improved insomnia severity, sleep efficiency, sleep quality, wake after sleep onset, sleep onset latency, total sleep time and night awakenings, with effect sizes from g 0.21 to 1.09, generally maintained at 4 to 48 weeks. Longer programs and more clinician support were associated with larger effects [12].
- Population
- 1,460 participants in 11 trials [12].
- Sources
- [12]
- Caveats
- The comparison with face-to-face CBT sets results from different trials side by side [12].
Sleep
Limited evidence: low-certainty trials in critically ill adults and cancer survivors found better self-reported sleep quality after massage.
Compared with control groups the source abstracts do not describe [16][17]
The numbers and sources
- Effect
- In critically ill adults, foot reflexology massage improved subjective sleep quality in subgroup analysis, and a two-night course helped in cardiac care patients [16]. In 4 trials in cancer survivors, massage improved self-reported sleep and accelerometer-recorded long sleep episodes [17].
- Population
- 569 participants in 10 trials in adult critical care [16]. Cancer survivors in 4 trials [17].
- Sources
- [16], [17]
- Caveats
- Overall certainty was rated low [16]. No trial in healthy adults or in people with insomnia alone was located.
Insomnia
Limited evidence: small, biased trials found more people sleeping better with acupressure or added acupuncture, and the effect became uncertain when dropouts were counted as failures.
Compared with no treatment [17]; sham or placebo [17]; other treatment alone [17]
The numbers and sources
- Effect
- Acupressure improved sleep quality in more people than no treatment (OR 13.08, 95% CI 1.79 to 95.59, 2 trials) or sham (OR 6.62, 95% CI 1.78 to 24.55, 2 trials). Acupuncture added to other treatment raised the share with better sleep (OR 3.08, 95% CI 1.93 to 4.90, 13 trials, 883 participants), with needle acupuncture and not electroacupuncture showing a benefit. When dropouts were assumed to do worse, the benefit was inconclusive [17].
- Population
- 33 trials with 2,293 people aged 15 to 98 with insomnia [17].
- Sources
- [17]
- Caveats
- All trials were at high risk of bias, and publication bias was likely. The reviewers wrote that the evidence could neither support nor refute acupuncture for insomnia [17].
Insomnia, compared with sleep medication
Limited evidence: five trials suggested CBT for insomnia outlasts sleep drugs in the long term, while benzodiazepines gave better short-term results.
Compared with benzodiazepine and non-benzodiazepine sleep drugs [11]
The numbers and sources
- Effect
- Low to moderate-grade evidence suggested CBT for insomnia gave better long-term results than benzodiazepine and non-benzodiazepine drugs. Very low-grade evidence suggested benzodiazepines gave better short-term results [11].
- Population
- Adults with primary or comorbid insomnia in 5 trials [11].
- Sources
- [11]
- Caveats
- Only 5 trials, with evidence graded low to very low [11].
Practices
Sleep quality
Shown to improve sleep quality slightly more than time-matched control programs in one meta-analysis, and no differently from established sleep treatments.
Compared with nonspecific active controls matched for time and attention [10]; evidence-based sleep treatments [10]
The numbers and sources
- Effect
- Against nonspecific active controls, sleep quality improved after the program (ES 0.33, 95% CI 0.17 to 0.48) and at follow-up (ES 0.54, 95% CI 0.24 to 0.84), at moderate strength of evidence. Against specific active controls such as evidence-based sleep treatments there was no difference after the program (ES 0.03, 95% CI -0.43 to 0.49) or at follow-up (ES -0.14) [10]. An earlier review found low or insufficient evidence for any effect on sleep [3].
- Population
- 1,654 participants in 18 trials [10]. 47 trials [3].
- Sources
- [10], [3]
- Caveats
- The reviewers call the findings preliminary [10]. Sleep was mostly self-rated.
Insomnia, against no program or sleep education
Shown in two trials of older adults with insomnia to improve sleep against no intervention or sleep education, by modest amounts on wrist-worn sleep measures.
Compared with no intervention [14]; a sleep seminar education control [15]
The numbers and sources
- Effect
- After 12 weeks, sleep efficiency rose 3.4% (95% CI 1.6 to 5.1), wake time after sleep onset fell 13.3 minutes and awakenings fell by 2.2 against no intervention, measured by actigraphy and maintained at 24 months [14]. Against a sleep seminar, Tai Chi Chih improved self-rated sleep quality, fatigue and depressive symptoms, but not insomnia remission, and polysomnography did not change [15].
- Population
- 320 adults aged 60 and over with chronic insomnia in Hong Kong, 80% women [14]. 123 older adults with chronic insomnia in Los Angeles [15].
- Sources
- [14], [15]
- Caveats
- The authors describe the absolute improvements as modest [14].
Sleep quality in older adults
Shown to improve self-rated sleep quality in older adults against inactive controls such as a waiting list, in one meta-analysis.
Compared with inactive controls such as a waiting list or education booklets [7]
The numbers and sources
- Effect
- Against inactive controls, sleep quality improved (effect size 0.65), along with vitality (0.31) and perceived mental and physical health. Against active controls, only depression improved significantly among quality of life outcomes [7].
- Population
- Older adults in 20 trials of quality of life outcomes [7].
- Sources
- [7]
- Caveats
- Sleep was self-rated.
Sleep and quality of life
Limited evidence: a systematic review described better sleep quality and quality of life in a few trials, without pooled estimates.
The numbers and sources
- Effect
- Narrative synthesis of randomized trials reported improvements in sleep quality and quality of life after cold water immersion [14]. In 5 patients with depression who swam regularly, Pittsburgh Sleep Quality Index scores went from 10.4 to 8.0 [21].
- Population
- Healthy adults in 11 randomized trials [14]. 5 patients [21].
- Sources
- [14], [21]
- Caveats
- The review abstract gives no effect sizes for sleep or quality of life. The patient study had no control group.
Sleep
Limited evidence: a sham-controlled pilot in 15 people with mild Alzheimer's disease and an uncontrolled study of 12 adults reported better sleep while grounded at night.
Compared with sham grounding [1]; before grounding, in the same people [2]
The numbers and sources
- Effect
- In patients with mild Alzheimer's disease, 12 weeks of grounding improved Pittsburgh Sleep Quality Index scores more than sham grounding (p = 0.006) [1]. Twelve adults with sleep problems, pain and stress who slept on a conductive mattress pad for 8 weeks reported sleep dysfunction reduced or eliminated in nearly all of them [2].
- Population
- 22 patients with mild Alzheimer's disease enrolled, 15 completing [1]. 12 adults with sleep dysfunction, pain and stress [2].
- Sources
- [1], [2]
- Caveats
- Seven of 22 patients did not finish the sham-controlled study [1]. The 12-person study had no control group and relied on self-report [2].
Snoring
Small uncontrolled studies found snoring fell by about half or more with mouth tape or an oral patch in people with mild sleep apnea.
Compared with before taping, in the same people [1][2]
The numbers and sources
- Effect
- In 20 mouth breathers, the median snoring index fell from 303.8 to 121.1 events an hour, and 13 of 20 (65%) cut their snoring index by at least half [1]. With a porous oral patch, the median snoring index fell from 146.7 to 40.0 an hour, snoring loudness from 49.1 to 41.1 dB, and self-rated snoring from 7.5 to 2.4 on a 10-point scale [2]. A scoping review notes one study in people without sleep apnea that reported less snoring with tape combined with other measures [5].
- Population
- 20 adults [1]. 30 patients [2]. 9 studies [5].
- Sources
- [1], [2], [5]
- Caveats
- No control group, and the responder share comes from 20 people. Participants were selected because they tolerated tape [1].
CPAP use in mouth breathers with sleep apnea
One crossover trial of 62 CPAP users who breathed through the mouth found tape added about 52 minutes of nightly CPAP use and eased daytime sleepiness and dryness.
Compared with CPAP without mouth tape for 30 days [6]
The numbers and sources
- Effect
- With silicone mouth tape, average CPAP use rose by 51.8 minutes a day (95% CI 33.1 to 70.4), nights with 4 or more hours of use rose by 17.6%, and the odds of meeting the trial's adherence target were 4.5 times higher. Epworth sleepiness scores, snoring, mouth and throat dryness and night awakenings improved [6]. A scoping review cites a study in which tape reduced mouth leak during bilevel ventilation [5].
- Population
- 62 adults with obstructive sleep apnea on CPAP, mean age 58, mean apnea-hypopnea index 45.8 [6].
- Sources
- [6], [5]
- Caveats
- One trial. Adverse effects were reported, and the abstract does not say which [6]. This is tape used with a CPAP machine, not tape alone.
Sleep apnea severity
Mixed results: two small uncontrolled studies found sleep apnea events fell by 35% to 47% with tape, while three other studies found no significant change.
Compared with before taping, in the same people [1][2]; sleeping without tape the night before, in the same people [3]; other treatments and placebo spray in studies summarized by a review [4]
The numbers and sources
- Effect
- In 20 mouth breathers with mild sleep apnea, median apnea-hypopnea index fell from 8.3 to 4.7 events an hour after a week of taping (p = 0.0002), and the oxygen desaturation index fell from 8.7 to 5.8 [1]. In 30 patients with mild sleep apnea, a porous oral patch lowered median apnea-hypopnea index from 12.0 to 7.8 an hour (P < 0.01) [2]. A systematic review of 10 studies found these two significant improvements, while three studies found no significant change in apnea-hypopnea index [4]. In 71 adults, oxygen desaturation did not improve with tape in about half of participants [3].
- Population
- 20 adults with mild sleep apnea who breathed through the mouth and tolerated tape [1]. 30 patients with mild sleep apnea [2]. 213 patients in 10 studies [4]. 71 adults with a two-night home sleep test [3].
- Sources
- [4], [1], [2], [3]
- Caveats
- No randomized trial against an untaped group or sham tape. Many studies excluded anyone with nasal obstruction [4]. The reviewers describe a potentially serious risk of harm in people taping indiscriminately [4].
Sleep
Mixed results: sleep scores improved in 10 athletes after 14 nights of whole-body red light, and pooled depression trials found no significant effect on sleep.
The numbers and sources
- Effect
- In 20 female basketball players, 30 minutes of whole-body red light every night for 14 days improved Pittsburgh Sleep Quality Index scores, serum melatonin and 12-minute run performance compared with no light (P < .05) [33]. In trials of photobiomodulation for depression, sleep outcomes did not differ significantly from control (SMD -0.82, 95% CI -2.41 to 0.77) [24].
- Population
- 20 athletes with a mean age of 18.6, 10 per group [33]. People with depression symptoms [24].
- Sources
- [33], [24]
- Caveats
- The athlete study had no sham light, so the comparison group knew it was untreated [33].
Sleep
Mixed results: 83.5% of surveyed sauna users reported sleep benefits, while an 8-week trial in people with chronic headache found no change in sleep disturbance.
The numbers and sources
- Effect
- In an online survey of 482 regular sauna users, 83.5% reported sleep benefits after sauna use [30]. In a randomized trial in chronic tension-type headache, sleep disturbance did not differ between the sauna and control groups after 8 weeks [20].
- Population
- 482 survey respondents who used a sauna about 1 to 2 times a week [30]. 37 people with chronic tension-type headache [20].
- Sources
- [30], [20]
- Caveats
- The survey is self-selected and self-reported, with no comparison group [30]. Sleep was a secondary outcome of the headache trial, and no trial measuring sleep with instruments was located.
Sleep
Mixed results: two small trials found better sleep with slow breathing before bed, while a month of 5-minute daily breathwork did not change sleep measured by a wearable.
Compared with social media use for the same time [13]; spontaneous breathing, before and after in the same people [14]; mindfulness meditation [4]
The numbers and sources
- Effect
- In 64 healthy young adults, 15 minutes of slow-paced breathing each evening for 30 days improved Pittsburgh Sleep Quality Index scores and raised overnight cardiac vagal activity, compared with 15 minutes of social media use [13]. In 14 people with self-reported insomnia, 20 minutes of paced breathing before bed shortened time to fall asleep, cut awakenings and raised sleep efficiency on polysomnography [14]. In the 1-month trial, hours of sleep, sleep efficiency and sleep score from a wearable did not change in any group or differ between groups [4].
- Population
- 64 healthy adults with a mean age of 22 [13]. 14 people with self-reported insomnia and 14 without sleep complaints [14]. 114 adults invited [4].
- Sources
- [13], [14], [4]
- Caveats
- Small samples, self-rated sleep in one trial [13], and no randomized control group in the insomnia study [14].
Sleep
Mixed results: one small trial found better rated sleep and fewer night movements after a session following evening training, and a depression trial found no change in sleep.
The numbers and sources
- Effect
- In 22 physically active men, 3 minutes at -40°C after evening training reduced movements during the night and improved subjective sleep quality compared with passive recovery (p < 0.05). Pain the next morning was lower (p < 0.01) [20]. In adults with depression, self-assessed sleep did not change with 10 sessions against sham (p > 0.05) [8].
- Population
- 22 physically active men [20]. 56 adults with a depressive episode who completed the trial [8].
- Sources
- [20], [8]
- Caveats
- The sleep trial had no sham, and its chamber ran at -40°C, warmer than the others on this page.
Insomnia, against exercise or CBT
Not shown to improve sleep more than conventional exercise. In one trial, CBT for insomnia produced more remission and better sleep than Tai Chi Chih.
Compared with conventional exercise [14]; cognitive behavioral therapy for insomnia [15]
The numbers and sources
- Effect
- Sleep measures by actigraphy did not differ between tai chi and conventional exercise [14]. CBT performed better than Tai Chi Chih for clinician-rated remission of insomnia and showed greater, more sustained improvement in sleep quality, fatigue and depressive symptoms (all P < 0.01) [15].
- Population
- 210 older adults in the tai chi and exercise arms [14]. 123 older adults [15].
- Sources
- [14], [15]
- Caveats
- Two trials of older adults.
Sleep
One small trial found no difference in sleep duration or quality the night after a post-exercise infrared sauna in 16 athletes.
The numbers and sources
- Effect
- Sleep duration (p = 0.377) and sleep quality rated 1 to 5 (p = 0.632) did not differ between the sauna and passive recovery nights. Night-time heart rate and heart rate variability did not differ either [20].
- Population
- 16 male basketball players, sleep recorded by diary [20].
- Sources
- [20]
- Caveats
- One night, self-reported by diary, in young athletes without sleep complaints.
Daily life
Eye masks for sleep in hospital
Shown to improve rated sleep quality in intensive care patients using eye masks, alone or with earplugs, in three reviews. Earplugs alone did not change sleep quality.
Compared with control groups in randomized trials the abstract does not describe [1]; control groups, including routine care, in randomized and nonrandomized studies [2]; comparison conditions the abstract does not describe [3]
The numbers and sources
- Effect
- Earplugs and eye masks together or apart improved self-reported sleep quality (SMD 1.44, 95% CI 0.80 to 2.09). Eye masks alone had a large effect (1.56, 95% CI 1.08 to 2.05), both together the largest (2.08, 95% CI 0.95 to 3.21), and earplugs alone none (0.07, 95% CI -0.50 to 0.64) [1]. Eye masks and earplugs improved sleep questionnaire scores, and polysomnography showed more total sleep time, higher sleep efficiency, more REM sleep and fewer awakenings and arousals [2]. Most of 19 studies showed better subjective sleep quality, while objective measures were not consistently improved [3].
- Population
- 797 adult ICU patients in 13 randomized trials [1]. 2,687 participants in 35 studies, 21 pooled [2]. 1,379 participants in 19 ICU and non-ICU studies [3].
- Sources
- [1], [2], [3]
- Caveats
- Patients in intensive care, where light and noise at night are far above home levels. Masks cannot be blinded, and sleep quality was mostly self-rated with tools that varied between studies [3].
Death from any cause
Linked to the lowest death rates at about 7 hours a night in pooled cohorts, a J or U shape, with a steeper rise for long sleep than for short.
Compared with people sleeping 7 to 8 hours a night, the reference in most studies [1]; people sleeping 7 hours a day [2]; normal sleepers [3]; people with normal sleep duration [4]
The numbers and sources
- Effect
- A dose-response meta-analysis found a U-shaped curve with the lowest risk at about 7 hours a day. Below 7 hours, risk rose 6% per hour less (RR 1.06, 95% CI 1.04 to 1.07). Above 7 hours, it rose 13% per hour more (RR 1.13, 95% CI 1.11 to 1.15). The shortest and longest groups had RRs of 1.13 and 1.35 [1]. A spline analysis of 40 cohorts found a J shape: against 7 hours, RR was 1.06 at 5 hours, 1.04 at 6 hours, 1.03 at 8 hours, 1.13 at 9 hours, 1.25 at 10 hours and 1.38 at 11 hours [2]. Short sleepers had a 12% higher risk (RR 1.12, 95% CI 1.08 to 1.16), with a linear rise below 6 hours [3]. Long sleepers had a 39% higher risk (RR 1.39, 95% CI 1.31 to 1.47) [4].
- Population
- 3,582,016 participants in 67 articles, follow-up 2.3 to 34 years [1]. 2,200,425 participants with 271,507 deaths in 40 cohorts [2]. 5,172,710 participants in 153 studies [3]. 5,134,036 participants in 137 cohorts [4].
- Sources
- [1], [2], [3], [4]
- Caveats
- Observational. Sleep was self-reported by questionnaire in most cohorts. Confounding is likely, and reverse causation is a strong candidate for long sleep: illness, depression and low fitness lengthen sleep, and some illnesses shorten it. The authors of the long sleep review call for work on whether the link is causal [4]. Heterogeneity was high for long sleep (I² 76.2%) [1]. In one review, short sleep was linked to higher risk in women and not in men [2].
Later school start times in teenagers
Shown to lengthen teenagers' sleep by about 25 to 77 minutes a night when school started later, in three reviews. Depressive symptoms fell slightly in pooled studies.
Compared with earlier school start times, as each study compared them [2]; the same students before the delay, and control schools in trials [3]; earlier school start times [4]
The numbers and sources
- Effect
- Six studies found later start times linked to more sleep. One study found fewer depressive symptoms, and results for alertness and absences were mixed. The evidence was rated very low in quality [2]. Start times delayed by 25 to 60 minutes were followed by 25 to 77 more minutes of sleep per weeknight, with some studies reporting less daytime sleepiness, depression and caffeine use [3]. Pooled, delayed start times added an average 69 minutes of sleep (SMD 0.63) and lowered depressive symptoms slightly (SMD -0.20) [4].
- Population
- 297,994 students aged 13 to 19 in 11 studies [2]. 6 studies, 2 of them randomized [3]. Middle and high school students in 6 studies [4].
- Sources
- [2], [3], [4]
- Caveats
- Most studies were not randomized and were at high or unclear risk of bias [2]. Heterogeneity was high [4]. Whether the extra sleep lasts was not shown. One study reported less parent and child time [2].
Activity and sleep after cutting family screen time
One trial found children were active about 46 minutes more a day after two weeks of strict family screen limits, with no change in sleep measured by EEG.
Compared with families told to carry on as usual [6]
The numbers and sources
- Effect
- Children's leisure non-sedentary activity rose by 45.8 minutes a day against controls (95% CI 27.9 to 63.6). Sleep measured by single-channel EEG did not differ between groups [6].
- Population
- 181 children and 164 adults in 89 Danish families whose parent used screens for leisure over 2.4 hours a day [6].
- Sources
- [6]
- Caveats
- Two weeks, with phones and tablets handed in, which is stricter than most everyday limits. 97% of children complied [6].
Redesigning shift schedules
Limited evidence: switching to faster or forward-rotating shifts, or letting workers choose their shifts, improved health or work-life balance in small, mostly uncontrolled studies.
Compared with the previous schedule, in experimental and quasi-experimental studies with or without a control group [1]
The numbers and sources
- Effect
- Across 26 studies, no type of schedule change was consistently harmful. Three types had beneficial effects on health, work-life balance or both: switching from slow to fast rotation, changing from backward to forward rotation (day, then evening, then night), and self-scheduling of shifts [1].
- Population
- Shift workers in 26 experimental and quasi-experimental studies from any country [1].
- Sources
- [1]
- Caveats
- The review gave a narrative synthesis with no pooled numbers, and the authors raised concerns about how far the evidence generalizes [1].
Sleep and body clock
Limited evidence: daylight was linked to fewer insomnia symptoms in one cohort, and a week camping in natural light shifted the body clock earlier in a small study.
Compared with each hour less of daylight outdoors a day [10]; the same people living with electric lighting [11]
The numbers and sources
- Effect
- Each extra daily hour outdoors was linked to fewer insomnia symptoms (95% CI for the odds ratio 0.94 to 0.97), less frequent tiredness (0.80 to 0.82), easier getting up (1.46 to 1.49) and an earlier chronotype [10]. After a week of natural light only, the internal biological night began at sunset and ended just after sunrise. Later chronotypes shifted the most [11].
- Population
- Over 400,000 UK Biobank adults [10]. 8 adults aged about 30, one week at home then one week camping in Colorado in July [11].
- Sources
- [10], [11]
- Caveats
- The cohort is observational [10]. The camping study was small and short, and measured clock timing, not sleep quality [11].
Blue-light filtering glasses in the evening
Mixed results: three reviews of blue-light filtering glasses found sleep improved in some trials and not others, with no change in objectively measured sleep against clear lenses.
Compared with non-blue-light filtering lenses [4]; visually identical clear-lens glasses worn at the same time [5]; control conditions the abstract does not describe [6]
The numbers and sources
- Effect
- Six trials gave inconsistent findings on sleep quality, three showing improvement and three no difference, with very low certainty of evidence [4]. In three double-blind crossover trials, the glasses did not change time to fall asleep (-4.86 minutes, 95% CI -20.23 to 10.52), total sleep time (8.75 minutes, 95% CI -35.31 to 52.82), sleep efficiency or wake after sleep onset, measured by wrist actigraphy [5]. A review of 16 randomized trials and other reports found substantial evidence of shorter time to fall asleep in people with sleep disorders, jet lag or variable shift schedules [6].
- Population
- 148 participants in 6 trials of sleep, within 17 trials overall [4]. 49 adults in 3 crossover trials [5]. 453 patients in 16 randomized trials, plus uncontrolled studies and case reports [6].
- Sources
- [4], [5], [6]
- Caveats
- Trials were small and short, and most were at high risk of bias [4]. The positive review was not a meta-analysis and included uncontrolled studies [6]. Populations ranged from healthy adults to people with insomnia, mania and depression.
Bright light, naps and other measures for sleepiness on shift
Mixed results: in a Cochrane review, bright light at night lowered sleepiness slightly in two studies, while naps, exercise and sleep education gave inconclusive results.
Compared with normal light [2]; control conditions the review abstract does not describe, for naps and other measures [2]
The numbers and sources
- Effect
- Bright light during the night shift lowered sleepiness by 0.83 points on the Stanford Sleepiness Scale (95% CI -1.3 to -0.36, very low-quality evidence). Another trial found no difference on another scale, and bright light during day shifts did not reduce sleepiness. A single nap opportunity changed reaction time by 11.87 ms, not significant. Exercise and sleep education showed promise without enough data to judge [2].
- Population
- 556 shift workers who also work nights, in 17 randomized trials [2].
- Sources
- [2]
- Caveats
- Most comparisons gave low to very low-quality evidence and studies were too varied to pool [2].
Blue-light filtering glasses
Not shown to ease eye strain from computer use against ordinary lenses in short trials, with effects on sleep uncertain, in a Cochrane review of 17 small trials.
Compared with non-blue-light filtering lenses [8]
The numbers and sources
- Effect
- Visual fatigue scores may not differ at under 1 week (mean difference 9.76, 95% CI -33.95 to 53.47, 120 participants, low certainty). Visual acuity probably did not differ (MD 0.00 logMAR). Of six trials of sleep, three found better sleep scores and three found no difference (very low certainty). Reported side effects included headache, discomfort and lower mood [8].
- Population
- Adults in 17 randomized trials of 5 to 156 participants, followed under 1 day to 5 weeks [8].
- Sources
- [8]
- Caveats
- No trial had a low risk of bias in every domain, and no meta-analysis was possible [8].
Food and drink
Falling asleep
Shown to shorten time to fall asleep only at about five standard drinks, in pooled studies of healthy adults. REM sleep fell even at two drinks.
Compared with a no-alcohol condition the abstract does not describe [10]
The numbers and sources
- Effect
- Shorter sleep onset and quicker deep sleep appeared only after a high dose of 0.85 g per kg or more, about five standard drinks. REM sleep was delayed and shortened from a low dose of 0.50 g per kg or less, about two drinks, worsening with dose. Effects on total sleep time, sleep efficiency and waking during the night were too uncertain to determine [10].
- Population
- Healthy adults in 27 studies [10].
- Sources
- [10]
- Caveats
- Laboratory nights after a set dose. The review notes faster sleep onset at high doses likely worsens later REM disruption [10].
Devices and wearables
Insomnia symptoms with tracker feedback and guidance
One trial found feedback on tracker data with guidance lowered insomnia severity more than sleep education in 113 adults with insomnia symptoms over 5 weeks.
The numbers and sources
- Effect
- Compared with sleep education and hygiene, the group given feedback on sensor-based sleep data with guidance on interpreting it had lower Insomnia Severity Index scores (p = .011, d = 0.51) and lower sleep disturbance (p = .036, d = 0.42) after the intervention. Sleep-related impairment, depression, anxiety and the gap between self-reported and sensor-measured sleep did not differ between groups (P > .40) [22].
- Population
- 113 adults with Insomnia Severity Index scores of 10 or higher (mean age 47.5, 64.9% female), 5 weeks. 103 completed [22].
- Sources
- [22]
- Caveats
- The intervention combined a Fitbit and an EEG headband with an individual session and two check-in calls, so the trial does not isolate the tracker from the guidance. Single-blind, with self-reported outcomes [22].
Sleep quality in healthy adults
One small trial found a week of wearing a tracker with feedback lowered self-reported sleep disturbance in 32 healthy adults, with total daily sleep unchanged.
The numbers and sources
- Effect
- In a randomized crossover trial, one week of wearing a WHOOP device that gave sleep feedback, plus sleep logs, was compared with one week of sleep logs alone. PROMIS sleep disturbance scores were lower with the device (B = -1.69, 95% CI -3.11 to -0.27, P = .021). Self-reported daytime naps rose slightly (B = 3.2, SE 1.4, P = .023), and total daily sleep did not change (P = .43) [14].
- Population
- 32 healthy adults (21 women, mean age 23.8) without medical conditions or sleep disorders [14].
- Sources
- [14]
- Caveats
- One week per condition, a self-reported outcome and no blinding. Young healthy volunteers at one center [14].