What do Cochrane systematic reviews say about interventions for insomnia?

ABSTRACT CONTEXT AND OBJECTIVE: Insomnia is a frequent complaint that generates more than five million visits to doctors per year in the United States. This study summarizes all Cochrane systematic reviews (SRs) that evaluated interventions to treat insomnia. DESIGN AND SETTING: Review of SRs, conducted in the Discipline of Evidence-Based Medicine, Escola Paulista de Medicina (EPM), Universidade Federal de São Paulo (UNIFESP). METHODS: A sensitive search was carried out in the Cochrane Database of Systematic Reviews to identify Cochrane SRs that assessed the effects of any type of intervention for people with insomnia. The results, main characteristics of the SRs and the certainty of the evidence obtained from them were synthesized and discussed. RESULTS: Seven SRs were included. They addressed the benefits and harm of acupuncture (n = 1), behavioral interventions (n = 1), music (n = 1), pharmacotherapy (n = 2), phototherapy (n = 1) and physical exercise (n = 1). The certainty of the evidence ranged from moderate to very low. CONCLUSION: Acupuncture, music, physical exercise, paroxetine, doxepin, trimipramine and trazodone seem to present some benefit for patients with insomnia. However, the uncertainty around these results means that no robust and definitive recommendations for clinical practice can be made until the benefits and harms from each intervention for patients with insomnia have been confirmed through further studies.


INTRODUCTION
Insomnia is a frequent complaint that generates more than five million visits to doctors per year in the United States. 1 It is considered to be a subjective condition that affects sleep maintenance, onset or early waking. 2 It is also a public health issue because of its impact on people's wellbeing. 3 Patients with insomnia usually present difficulty in initiating or maintaining sleep and may wake up without the capacity to go back to sleep. This situation may induce development of symptoms during the day, such as sleepiness, mood disturbances and fatigue. 4,5 Insomnia may precede or appear along with other diagnoses and may occur as a result of different stressors. The following individual factors are commonly associated with the risk of insomnia: female sex, older age (for any sex), previous episodes of insomnia and family history. 6,7 It is classified as a disorder and may be divided into primary or secondary according to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5). Primary insomnia is more common and secondary insomnia is associated with another disorder. 4,5 Acute or short-term insomnia is considered to consist of at least three nights of difficulty in sleeping for no more than three months. If no treatment is provided, it may progress to chronic insomnia, which is a stage of the disorder that is more difficult to treat. 8 Insomnia may be a risk for developing hypertension, diabetes, obesity and cardiovascular diseases. [9][10][11][12] It may also increase the risk of psychiatric comorbidities and substance abuse and it reduces the quality of life. [13][14][15] The treatments available include pharmacological options, psychological interventions and alternative therapies, but all of these may have limitations. Physicians need to take into account the symptoms relating to the insomnia, treatment availability, effectiveness and safety.

OBJECTIVE
The aim of this study was to make a synthesis of the evidence from Cochrane systematic reviews that assessed different therapies for insomnia and to discuss this evidence.

Design
Review of Cochrane systematic reviews (SRs).

Types of studies
We considered the latest version of Cochrane SRs. We did not include any protocols or any systematic reviews (SRs) that had been withdrawn from the Cochrane Database of Systematic Reviews (CDSR).

Types of participants
We considered participants with insomnia. We excluded SRs in which was not possible to identify insomnia as the type of sleep disturbance presented by the participant. We did not impose any restriction based on age or sex.

Types of intervention
We considered any surgical, pharmacological or non-pharmacological intervention.

Types of outcomes
We considered any outcomes that had been evaluated by the authors of the SRs included. These outcomes included any clinical, social, laboratory or economic outcomes that had been reported.

Search for reviews
We conducted a broad and unrestricted systematic search in the Cochrane Database of Systematic Reviews (via Wiley) on August 1, 2018. The search strategy is presented in Table 1.

Selection of systematic reviews
The selection phase consisted of independent reading of all the abstracts retrieved, by two researchers, to check their eligibility in relation to the inclusion criteria. Any disagreement was resolved through reaching a consensus or by consulting a third author.

Presentation of the results
We made a synthesis of all the SRs included and presented the key results and methodological issues using a narrative approach (qualitative synthesis).
The SRs included were summarized based on the following characteristics: • Inclusion criteria/PICO (population, intervention, comparator and outcomes) • Methodological issues relating to searching for and coding of studies • Main results • Critical assessment of studies included (risk-of-bias assessment) • Analyses performed (including methods for pooling research through meta-analysis) • Assessment of the certainty of the evidence using the GRADE approach. 16 • Applicability When the SRs included considered multiple interventions, we presented only those that were relevant for this review.

Acupuncture
The aim of this review 17 was to determine the efficacy and safety of acupuncture for insomnia. The authors compared groups receiving the same interventions with or without associated acupuncture, against placebo or sham or no treatment. Randomized controlled trials (RCTs) that compared different acupuncture methods or acupuncture against other treatments were not considered. The review included 33 RCTs (2293 participants, aged 15 to 98 years) that had assessed needle acupuncture, electroacupuncture, acupressure or magnetic acupressure. The main results are presented below.

Acupressure versus no treatment/sham/placebo
Acupressure resulted in a benefit regarding sleep quality, compared with no treatment (odds ratio, OR = 13.08; 95% confidence interval, CI = 1.79 to 95.59; 2 RCTs; 280 participants) Low *GRADE (Grading of Recommendations Assessment, Development and Evaluation) has the aim of assessing the certainty of the evidence. From this, the results are classified as having high certainty of evidence (high confidence that the estimated effect is close to the true effect); moderate certainty of evidence (likely that the estimated effect is close to the real effect, but there is a possibility that it is not); low certainty of evidence (limited confidence in the effect estimate) or very low certainty of evidence (the true effect is likely to be substantially different from the estimate effect). **The certainty of the evidence was graded considering all studies on melatonin that were included in the review. However, only one of them was on insomnia (the other studies considered all types of sleep disturbance. Thus, the GRADE that was specific for this study was not assessed. or sham/placebo (OR = 6.62; 95% CI = 1.78 to 24.55; 2 RCTs; 112 participants). However, in a sensitivity analysis in which an assumption was made that dropouts had worse outcomes, acupuncture ceased to be conclusively beneficial.

Acupuncture as an adjunct to another treatment versus this other treatment alone
Acupuncture as an adjunct to another treatment was found to present the possibility of benefit regarding sleep quality (OR = 3.08; 95% CI = 1.93 to 4.90; 13 RCTs; 883 participants,).
In subgroup analyses, needle acupuncture alone was shown to be beneficial, but not electroacupuncture.

Antidepressants for treating insomnia in adults
The aim of this review 18 was to assess the effects of antidepressants for treating insomnia. Twenty-three RCTs were included (2806 participants), comparing any antidepressant with placebo, other medications for insomnia or a different antidepressant.
• Fluoxetine: no important difference between groups.
There was also no proper reporting of adverse events. The overall certainty of evidence for the subjective outcomes ranged from low to very low.

Other antidepressants versus placebo
Eight RCTs compared other antidepressants with placebo (one used mianserin and seven used trazodone). The main results are presented below. The purpose of this review 20 was to assess the efficacy and safety of cognitive-behavioral interventions (CBT) among older people (older than 60 years) with insomnia. All forms of CBT were considered, including sleep hygiene, stimulus control, muscle relaxation, sleep restriction and cognitive therapies alone. CBT was compared with no intervention, waiting-list control and/or placebo ("quasi-desensitization"). Six randomized clinical trials (RCTs) were included (282 participants). The main results are presented below. The benefits associated with CBT regarding waking after sleep onset were probably clinically modest. The data were based on a single study with a small-sized sample regarding total time awake.

Music for insomnia among adults
This review 21 assessed the effects of listening to music on insomnia among adults. Six RCTs and quasi-randomized controlled trials (qRCTs) were included, comprising 314 participants. The studies compared the effects of listening to music daily at one's own house with no treatment or treatment-as-usual, on sleep improvement among adults with insomnia. The main results are presented below.
• Sleep quality as assessed using the Pittsburgh Sleep Quality Index (scale from 0 to 21 scale, on which lower scores mean better sleep quality): benefit from listening to music (MD = -2.80; 95% CI = -3.42 to -2.17; 5 studies; 264 participants; moderate certainty of evidence).
• Sleep onset latency, total duration of sleep, sleep interruption and sleep efficiency assessed using a questionnaire in the morning, with evaluations on participants using polysomnography. One study (50 participants) reported results from these outcomes but did not provide sufficient numerical data for analysis. The authors of that study only reported that there was no evidence of any effect from the intervention. The certainty of evidence regarding these outcomes was downgraded due to risk of bias and imprecision.
• Adverse events: none of the studies included reported this outcome.

DISCUSSION
The present review included seven Cochrane systematic reviews (SRs) that assessed the effects of different interventions for insomnia in the general population or among groups with specific disorders (people with dementia and the elderly). Overall, the current evidence from Cochrane SRs shows that acupuncture, music, physical exercise, paroxetine, doxepin, trimipramine and trazodone seem to present some benefit for patients with insomnia. However, the certainty of the evidence provided by these SRs ranged from very low to moderate, which means that it is likely or very likely that further studies may change the current evidence.
Three of the SRs included are out of date, since they were published in 2002 and 2003. 19,20,23 Updates for these reviews are urgently needed, to aid in searching for new studies and to revise the analyses in line with newer recommendations.
There are seven ongoing Cochrane SRs (protocols) that will be published in the future. Their aims are to assess new non-benzodiazepine hypnotics (eszopiclone, zaleplon, zolpidem and zopiclone), 24-27 ramelteon (melatonin receptor agonist) 28 and pharmacological and non-pharmacological interventions for treating insomnia during pregnancy. 29 Additionally, a network meta-analysis is being conducting to compare the efficacy and acceptability of all pharmacological treatments for insomnia among adults. 30 These SRs will be useful both for healthcare professionals and for patients, to aid in decision-making.

CONCLUSION
This review identified seven Cochrane systematic reviews that assessed pharmacological or non-pharmacological interventions for treating insomnia. Based on their findings, acupuncture, music, physical exercise, paroxetine, doxepin, trimipramine and trazodone seem to present some benefit for patients with insomnia. However, the uncertainty relating to these results means that no robust and definitive recommendations for clinical practice can be made until the benefits and harm from of each intervention for patients with insomnia have been confirmed through further studies.