The effects of aerobic, resistance, and meditative movement exercise on sleep in individuals with depression: protocol for a systematic review and network meta-analysis

Worldwide lifetime prevalence of unipolar depression is estimated to be 10% [1]. Unipolar depression is the leading cause of burden of disease in middle- and high-income countries and is projected to become the leading cause worldwide by 2030 [2]. The economic burden of unipolar major depression alone was estimated to be $210 billion in the USA in 2010 [3]. In addition to the direct and often debilitating symptoms, unipolar depression also substantially increases the risk of all-cause mortality [4] and morbidity [5].

One of the diagnostic criteria for depression according to DSM-V [6] and ICD-10 [7] is disturbed sleep, i.e., insomnia or hypersomnia. This is reflected by the high prevalence (up to 90%) of co-occurring insomnia in individuals with depression [8]. Insomnia has a negative impact on health-related quality of life and daytime functioning [9, 10]. Moreover, sleep disturbances have been recognized as a mechanistic (i.e., causal or bidirectional) process in depression. Sleep problems are an independent risk factor for depression [11]. Depression and insomnia are linked in a bidirectional manner [12]. Comorbid sleep disorders have a negative influence on the disease trajectory in depressive disorders [13]. Some of the most prevalent residual symptoms after treatment response or remission in this clinical population are insomnia symptoms [14, 15]. This is pertinent because persistent sleep disorders increase the likelihood of relapse [13, 16, 17]. Sleep disorders are an independent risk factor for suicide [18, 19].

According to the British national (NICE) treatment guidelines, individuals with depression are to be treated with psychotherapy and pharmacotherapy [20]. However, antidepressants are of limited efficacy [21] and can cause considerable adverse effects. In a primary care setting the number needed to treat (NNT) for improvement of depressive symptoms by selective serotonin reuptake inhibitors is approximately seven whereas the number needed to harm (NNH) (withdrawal due to side effects) is estimated to be between 20 and 29 [22]. A combination of antidepressants and benzodiazepine receptor agonists (Z-drugs) in patients with depression, similarly, has been shown to be effective (NNT = 10) but causes considerable adverse events (NNH = 20) [23]. Due to the frequency of adverse side effects, non-adherence to antidepressants is high and associated with decreased remission rates, increased risk of relapse, and increased health care utilization [24]. Hypnotics are frequently prescribed for insomnia. However, they have a poor benefit-to-risk ratio with serious adverse effects including cognitive impairment, injury from falls and automobile accidents (including in younger individuals), cancer, suicide, and hypnotic withdrawal insomnia [25, 26]. In light of this evidence, interest in adjuvant and alternative therapies, especially exercise, has increased in the last decade.

The effects of exercise on depressive symptoms have been summarized in multiple meta-analyses [27‐38]. Systematic reviews found moderate-to-large effect sizes for aerobic [28], resistance [36] as well as yoga [35] exercises on depression. Moreover, no significant differences between these interventions and antidepressant medication were found [28, 35]. The effect of other meditative movement exercises such as qi gong and tai chi seems to be positive, albeit less pronounced [37, 38]. Aerobic exercise interventions in depressive patients have also been found to improve cardiorespiratory fitness [27]. This is relevant because depression is known to increase the risk of cardiovascular mortality and morbidity [29, 30].

Current data suggest that exercise might be a suitable therapeutic option to improve sleep quality. Aerobic exercise has been shown to have positive acute (during the night immediately following exercise) and chronic (over several weeks) effects on sleep in healthy individuals with small-to-moderate effect sizes [39, 40]. These findings have been replicated in populations with sleep complaints [39, 41‐43] and chronic disorders [44‐49] and confirmed by a meta-analysis of previous meta-analyses [50]. A recent meta-analysis also found moderate-to-large effect sizes for mainly chronic resistance training on sleep quality [51]. Lastly, numerous meta-analyses show a positive effect of meditative movement on sleep quality in a variety of patient [49, 52] and elderly [53] populations. However, to the knowledge of the authors, no systematic review concerning the effect of exercise on sleep in patients with depression has been performed.

Although the etiology of insomnia (with or without comorbidities) is not yet fully understood, hyperarousal is widely considered a causal and maintaining factor [54‐56].

Multiple mechanisms of action, including ones which involve hyperarousal, have been proposed to explain the effect exercise has on sleep (confer the reviews of Buman and King (2010) [57] and Uchida et al. 2012 [58] for aerobic exercise). Insomniacs have been shown to have impaired thermoregulation [59]. Chronic exercise, on the other hand, improves thermoregulation [60, 61]. Increased skin temperature, which occurs during and immediately after acute aerobic [62], resistance [63], and meditative movement [61, 64] exercise, seems to modulate neural circuits in a way which might be conducive to sleep [65]. Exercise causes changes in the levels of pro-inflammatory cytokines [66], growth hormone [67, 68], and brain-derived neurotrophic factor [69‐71] which seem to play a role in the regulation of sleep [72‐74]. Aerobic [28], resistance [51], and meditative movement [75] exercise have positive effects on anxiety as well as depression and might thereby reduce psychophysiological arousal. Although it is not fully understood why humans sleep, one hypothesis states that humans sleep to optimize restorative processes [76]. Aerobic and resistance exercise increase energy expenditure and require muscle repair, thus stimulating such restorative processes. Aerobic exercise has also been shown to consistently produce phase shifts (i.e., changes in circadian rhythm within the 24 h cycle) in individuals of different ages and fitness levels. This effect has been found in individuals irrespective of age and cardiorespiratory fitness as well as independent from the effect of light. [77]. Therefore, aerobic exercise may act as a so-called ‘zeitgeber’ positively affecting entrainment (i.e., the synchronization of the endogenous and exogenous rhythms). It should be noted that it is unclear whether the mechanisms of action differ between insomniacs with and without psychiatric comorbidity.

The rationale for this review can be summarized in four points. (1) Sleep disturbances are of high prognostic relevance for remission in depression [11]. (2) Current therapies have a dissatisfactory benefit-to-risk-ratio. (3) Exercise has been shown to have positive effects on depression [28, 36, 75] as well as sleep [39, 41, 49, 51, 53]. (4) To the best of our knowledge, no systematic review has been performed to ascertain the effects of aerobic, resistance, and meditative movement exercise on sleep in people with depression.

The main objective of this review is to assess the effects of aerobic, resistance, and meditative movement exercise on sleep quality in patients with depression. A secondary goal is to ascertain the effects of exercise on sleep duration, sleepiness, daytime functioning, use of hypnotics, and adverse events (e.g., injuries, cardiovascular incidences).

Multiple sources will be used in this systematic review. A systematic computerized search will be performed in the following online databases: PubMed (on PubMed.​gov), EMBASE (on Ovid), Cochrane Library (on cochranelibrary-wiley.​com), PsycINFO (on Ovid), SPORTDiscus (on EBSCOhost), and CINHAL (on EBSCOhost). OpenGrey (on opengrey.eu) and ProQuest Dissertations and Theses A&I (on proquest.​com) will be searched to include gray literature. Bibliographies of all included studies as well as any other relevant reviews identified via the search will be screened. Clinicaltrials.​gov and WHO International Clinical Trials Registry will be searched in order to identify ongoing as well as unpublished studies. Due to lack of controlled vocabulary and restricted length of search strings on these websites, a modified query will be used. Authors of included studies will be contacted via e-mail in order to inquire whether they know of any other relevant publications. All databases will be searched from their inception to the search date.

The search strategy will be constructed using the PICOS (patient, intervention, comparison, outcome, study design) framework. The search string will be comprised of controlled vocabulary whenever possible and free text. These terms (including appropriate truncation) will be selected in an iterative scoping search using the PICOS approach as well as backward and forward chaining. The study design component will be identified using the “Cochrane highly sensitive search strategies for identifying randomized trials” [89] and translated according to the database. Terms within each group will be combined with a Boolean “OR” and groups will be combined using a Boolean “AND” command. The PubMed search strategy (see Additional file 2) was adapted according to the controlled vocabulary in each database (see Additional file 2). The search strategy has been reviewed by an information scientist from the Basel Medical University Library using the Peer Review of Electronic Search Strategies (PRESS) guideline [90]. Test searches have been performed in order to ensure the validity of the search string.

For the calculation of relative treatment effects group means, corresponding standard deviations and group sizes will be extracted primarily. In case one of these values was missing, other statistical data that can be converted into means and standard deviations will be extracted. Conversions will be calculated according to formulas provided, e.g., [98, 99]. If standard deviations cannot be calculated from the available study information, we will impute them using the standard deviations reported in the other included studies [100]. We will conduct sensitivity analyses excluding studies in which standard deviations had to be imputed. If the N was missing in the table of analysis, we will use the N of the descriptive statistics. If studies report medians and interquartile ranges, a normal distribution will be assumed, if not indicated otherwise, to convert these values to means and standard deviations [98]. If studies only report adjusted outcome values, data will be extracted, but sensitivity analyses will be calculated without these studies to check for possible bias. We plan to extract the effect size provided by the study authors only if no other information was available for effect size calculation. If it is not possible to impute appropriate measures for the calculation of effect sizes, and no effect sizes are reported we will contact the authors.

Among others, the following information will be extracted from each study:

The primary outcome will be standardized mean differences (SMD) of sleep quality at post-exercise-intervention and at the last available follow-up assessment, measured by self-reports (e.g., PSQI [102], ISI [103]) or clinician ratings (sleep-related HAM-D items [104]).

Secondary outcomes will be:

The rationale for the selection of the primary outcome is that perceived sleep quality, i.e., difficulties initiating or maintaining sleep or early morning awakening is one of the main complaints in insomnia. Reduced sleep duration [108] and daytime impairments are a further important category of complaints, markedly increasing the perceived need for treatment [109]. Adverse events must be considered in order to inform decision-makers on the benefit-to-risk ratio of an exercise intervention.

The risk of bias will be evaluated independently by two reviewers (GB and TZS) at the study level. Disagreement will be resolved by consensus or if no consensus can be reached, disagreement will be resolved by adjudication of a designated third reviewer (AST). Bias will be assessed using the revised Cochrane risk of bias tool [101]. This tool assesses five domains: (1) randomization process, (2) deviations from intended interventions, (3) missing outcome data, (4) measurement of outcomes, and (5) selection of reported interventions. The three possible judgments are possible: low risk, some concerns, and high risk of bias. A summary table of bias assessment on study level will be included in the publication. These assessments will contribute to the evaluation of overall confidence in the findings of the network meta-analysis using the CINeMA framework [110].

Data will be synthesized descriptively. A summary table of included studies will entail information on the authors, population characteristics (diagnostic criteria, baseline severity of sleep quality, depression, age, and numbers), interventions (exposure in each group), outcomes measures used, and results (sleep quality, sleep duration). Network meta-analysis will be performed. Statistical (number of studies and heterogeneity of results), clinical (heterogeneous populations), and methodological (low quality of trials or follow-up duration) aspects will be considered to decide whether network meta-analysis is valid. If network meta-analysis results must be deemed methodologically inaccurate, a pairwise meta-analysis will be considered. Should a pairwise meta-analysis also not be possible, studies will be summarized narratively.

The package netmeta [111] for the open-source software environment R [112] will be used to calculate network meta-analyses within a frequentist framework.

A network will be created including all available jointly randomizable treatments. We assume that any patient that meets all inclusion criteria is likely, in principle, to be randomized to any of the interventions in the synthesis comparator set.

We will address the assumption of transitivity which underlies network meta-analysis [113], by (1) assessing whether the included interventions are similar across studies using a different design, and (2) checking whether the distribution of potential moderators is balanced across comparisons [114]. A priori we have defined depression severity, comorbidities, age, and gender as potential effect modifiers and will evaluate the comparability of the respective characteristics across comparisons qualitatively.

We expect considerable diversity of outcome measures and will, therefore, calculate standardized mean differences (SMD) using Hedge’s g with 95% confidence intervals [115]. SMD is the mean difference between groups divided by the pooled standard deviation. The effect size measure allows comparison of effect sizes across similar measurements of a single outcome. The conventional and somewhat arbitrary classification of SMD proposed by Cohen (1988) [116] has been expanded to include very small (.01), small (0.2), medium (0.5), large (0.8), very large (1.2), and huge (2.0) effect sizes [117]. Random-effects pairwise SMDs across studies will be calculated based on the available comparisons between treatment and comparator treatments [118]. Inverse variance weighting is used for pooling. In addition, indirect evidence will be estimated using the entire network of evidence. Random-effects netmeta accounts for dependencies between comparisons in case of multi-arm trials [119]. The command pairwise will be used in case of multi-arm trials, in order to transform the dataset to the comparison level, which is needed for conducting the network meta-analysis.

The primary outcome will be SMD of sleep quality assessed via self- or observer-reported measures. If more than one primary outcome is reported, the most frequently used scale will be included in the analysis to reduce between-study heterogeneity. If possible, we will assess the association between instruments and changes in sleep quality. Two individual analyses will be run for the outcome data at the end of treatment, and the last available follow-up. Separate network meta-analyses will be conducted for secondary outcomes if possible. Results from network meta-analysis will be presented as summary SMD for each possible pair of treatments. Whenever possible, measures of uncertainty will be reported in the form of the 95% confidence interval and 95% prediction interval.

To calculate statistical heterogeneity between studies on the pairwise level, the Q statistic will be used [89]. Further τ² will be analyzed to estimate the variance caused by the distribution of the true study means [120]. I² will be evaluated to indicate the amount of observed variance that can be attributed to between-study heterogeneity [121]. I² and the corresponding confidence interval can be interpreted as the percentage of overall heterogeneity that is due to variation of the true effects. An I² value of 0% to 40% might not be important, 30 to 60% may represent moderate heterogeneity, 50 to 90% may represent substantial heterogeneity, and 75 to 100% considerable heterogeneity [89]. In NMA, we will assume a common estimate for the heterogeneity variance across the different comparisons.

Local and global methods will be used to detect inconsistency [122]. The presence of inconsistency will be evaluated using the following approaches: (1) locally using the netsplit command (i.e., testing the difference between estimates derived from direct evidence and estimates derived from indirect estimates for statistical significance) and (2) globally using the decomp.design command (i.e., using the design-by-treatment interaction model). For this purpose, the total Q statistic (i.e., the measure of total heterogeneity/inconsistency in the network) will be decomposed to an inconsistency factor (between designs) and a heterogeneity factor (within designs). We will compare the magnitude of heterogeneity between consistency and inconsistency models to determine how much heterogeneity will be explained by inconsistency. We will do this by testing the residual inconsistency, which remains under the assumption of a full design by treatment interaction model for statistical significance.

In the case of statistical heterogeneity or inconsistency between results from individual studies, we will investigate the potential impact of the following trial-level effect modifiers: (1) year of publication, (2) study precision (i.e., sample size), (3) studies reporting non-adjusted vs. adjusted means, (4) studies with imputed standard deviations vs. studies which reported standard deviations. If the number of studies allows it, theoretically driven subgroup analyses will be done according to population (e.g., severity of depression), duration of intervention, duration of follow-up, outcome characteristics (i.e., self- vs. observer ratings, objective vs. subjective sleep duration), and methodological quality.

The confidence in the network meta-analyses will be estimated using the Confidence in Network Meta-Analysis (CINeMA) framework [110]. This includes study limitation, indirectness, inconsistency (heterogeneity, incoherence), imprecision, and publication bias. Publication bias will be assessed according to the GRADE guideline [123] and by comparing eligible trials identified in registries (e.g., clinicaltrials.​gov) with published data. Selective reporting bias will be assessed by comparing protocols (if available) and reports of trials.

The results will be published in a peer-reviewed journal and presented at conferences as well as invited talks.

PROSPERO stage 1, preliminary searches completed.