Microbiota Composition and Probiotics Supplementations on Sleep Quality—A Systematic Review and Meta-Analysis

The gut microbiota (GM) plays a crucial role in human health. The bidirectional interaction between GM and the central nervous system may occur via the microbiota–gut–brain axis, possibly regulating the sleep/wake cycle. Recent reports highlight associations between intestinal dysbiosis and sleep disorders, suggesting that probiotics could ameliorate this condition. However, data are poor and inconsistent. The aim of this quantitative metanalytic study is to assess the GM composition in sleep disturbances and evaluate probiotics’ effectiveness for managing sleep disorders. A systematic review was carried out until July 2022 in online databases, limiting the literature research to human studies and English language articles. No significant GM diversity between patients with sleep disturbances versus healthy controls was found, revealed by α-diversity, while β-diversity is missing due to lack of proper reporting. However, probiotics supplementation significantly reduced the self-assessed parameter of sleep quality and disturbances Pittsburgh Sleep Quality Index (PSQI) score compared with the placebo. No difference in the Epworth Sleepiness Scale (ESS) score was found. While available data suggest that GM diversity is not related to sleep disturbances, probiotics administration strongly improves sleep quality as a subjective perception. However, heterogeneity of data reporting in the scientific literature should be considered as a limitation.


Introduction
The gut microbiota (GM) is a community of intestinal microorganisms, including bacteria, archaea, and eukarya, constituting the intestinal flora [1].Physiologically, GM shows (i) metabolic properties, being able to produce essential nutrients, (ii) protective functions through the regulation of mucus production, (iii) structural actions, mediating the expression of tight junction proteins, and (iv) neurological properties, interacting with the peripheral and the central nervous system [2].It is accepted that GM has broad impacts on human health, impacting the colonization and the resistance to pathogens, maintaining the intestinal epithelium, metabolizing dietary and pharmaceutical compounds, and controlling immune function [3].The GM composition is heterogeneous and represents an individual signature, reflecting dietary habits [4].In humans, more than two thousand prokaryotic species distributed in 11 different phyla have been recognized [5,6], among which the large dominant phyla are Firmicutes and Bacteroidetes [7,8].Derangement of microbiota may impact its metabolic activities, leading to dysbiosis, which, in turn, could lead to several dysfunctions.Therefore, the evaluation of GM composition gathered increasing relevance [9][10][11][12][13].GM heterogeneity is described by the αand β-diversity parameters as measures of species diversity within a community at a local scale and between different communities, respectively [14].In particular, α-diversity is a measure of microbiome diversity applicable to a single sample, whereas β-diversity quantifies the similarity or dissimilarity between two communities.
In the last decade, the link between GM composition and pathological conditions has gathered increasing interest, suggesting possible connections with infectious and autoimmune diseases, metabolic dysfunctions, neurocognitive disorders, and cancer [15][16][17][18][19][20][21].The mutual relationship between GM and the host is expected to modulate additional physiological processes, such as sleep, which could be perturbed in the case of dysbiosis [22].Sleep disturbances are classified into major categories, such as insomnia, sleep-related breathing disorders, central disorders of hypersomnolence, circadian rhythm sleep-wake disorders, sleep-related movement disorders, and parasomnias [23].Sleep could be described both quantitatively and qualitatively using multiple tools [24].Polysomnography (PSG) remains the gold standard approach to quality and quantitatively assess sleep, providing detailed information about brain activity, sleep stages, patterns, oxygen saturation, and eye and leg movements [25].However, PSG is scarcely used in large-scale investigations due to its cost and invasiveness.Alternatively, actigraphy is based on small wrist-watch digital devices monitoring motor activity data for extended periods [26] and is frequently employed in population-based epidemiologic studies [27].However, it is not recommended for the routine management of sleep disturbances.Thus, several alternative subjective tools, such as direct observations, sleep diaries, and questionnaires, are largely used as well [28].In particular, validated questionnaires are common tools because of the advantageous cost/benefit ratio despite the less accurate data collection [29].Among these, the Pittsburgh Sleep Quality Index (PSQI) is a 19-items questionnaire assessing sleep quality [30], while the Epworth Sleepiness Scale (ESS) is an 8-item self-administered questionnaire evaluating the perceived propensity to fall asleep [31,32].Finally, the nature and severity of insomnia may be assessed by the Insomnia Severity Index (ISI), determined by a 7-item questionnaire [29].
Several pieces of evidence suggest that GM is essential for the maintenance of normal sleep physiology, modulating the production of metabolites involved in sleep homeostasis, such as interleukin (IL)-1β, short-chain fatty acids (SCFAs), serotonin (5-HT), γ-aminobutyric acid (GABA), and melatonin [1,[33][34][35].Experiments in mouse models demonstrated that the SCFA butyric acid modulates the expression of the clock genes Per2 and Bmal1, potentially disrupting circadian rhythms [36,37].A recent study identified the GM signature of good sleepers, which would be abundant in Firmicutes, low in Prevotella, and with a high α-diversity index [38].These data suggest that sleep disorders may be treated by acting on the GM composition using probiotics.They are defined as "live microorganisms which, when administered in adequate amounts, confer a health benefit on the host" [39][40][41] and may be used to treat dysbiosis [39].Probiotic preparations are largely heterogeneous, and their effects depend, for instance, on dosage, supplementation duration, and bacterial strains [42][43][44].However, the possible beneficial impact of probiotics on sleep disorders is a matter of debate.In particular, few pieces of evidence are available so far to apply these compounds in clinical practice.
This systematic review and meta-analytic study has two aims: (i) to evaluate differences in GM composition between subjects with sleep disturbances versus healthy controls and (ii) to investigate the effectiveness of probiotics in treating sleep disorders.

Materials and Methods
The systematic review and meta-analysis were performed according to criteria described in the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) 2015 Statement.Two separate literature searches were carried out from the first available study on probiotics until July 2022, querying the online databases PubMed (MEDLINE), Embase, the Cochrane Library, and Scopus.Then, two different meta-analyses were performed in sequence.

Study Selection and Inclusion Criteria
The first literature search evaluated all published articles concerning the relationship between sleep disorders and GM composition, using the following keywords: microbiota OR microbiome OR (gut microbiota) OR (gut microbiome) OR (gut microflora) OR dysbiosis AND sleep OR (sleep disorder) OR (sleep disturbance) OR (sleep problem).Studies were considered eligible when investigating patients with sleep disorders and reporting GM composition.No restrictions on age, sex, study design (i.e., observational or interventional), year of publication, or tool used to evaluate sleep dysfunctions were applied.
The second literature search was designed to identify all studies in which oral daily supplements containing probiotics (live bacteria) or para-probiotics (heat-inactivated/killed probiotic bacteria) were administered to human subjects with sleep disorders.The following search strategy was applied: probiotic AND sleep OR (sleep disorder) OR (sleep disturbance) OR (sleep problem).Interventional studies in which probiotics or para-probiotics were administered together with other substances were excluded to avoid potential biases.A first screening collected all the interventional studies available, but subsequently, only placebo-controlled studies (probiotic treatment versus placebo) were selected.The randomization was not considered an inclusion criterion.
For both literature searches, two authors (FC and VD) independently collected the eligible articles, and conflicts were resolved by a third investigator (DS).A preliminary screening identified potentially relevant studies based on their titles and/or abstracts.The selected studies were systematically screened for inclusion by full text, according to the aforementioned criteria.

Data Extraction and Analysis
In the first literature search, all extracted data were included in a single dataset, considering authors, year of publication, journal, study design, subjects' age, patient's number, microbiota profiling method, sleep assessment methods, criteria for group subdivision according to sleep disorder, α-diversity, β-diversity, and Firmicutes to Bacteroidetes ratio, representing a clinical biomarker of dysbiosis [45].The α-diversity index was considered as the primary endpoint.The main outcome was the comparison of the α-diversity index between subjects with vs. without sleep disorders, as it was the parameter reported across almost all studies, while it was not for β-diversity.When studies reported the median and the interquartile range (IQR) of the index, the corresponding mean ± standard deviation (SD) was calculated [46,47].
In the second literature search, the following data types were extracted: authors, year of publication, journal, study design, subjects' age and body mass index (BMI), patients' number, microbiota profiling method, sleep assessment methods, intervention(s), probiotic bacteria used and colony forming unit (CFU), placebo treatment, treatment duration, and sleep data (both pre-and post-treatment).Since the most used tool to assess sleep quality was PSQI, it was considered the primary endpoint.The PSQI mean ± SD before and after treatment was used to quantify the probiotics' effect on sleep quality.When articles reported the standard error of mean (σ), SD was calculated using the following equation: where n indicates the number of subjects.Meta-analyses were performed only when more than three studies were available.The parameter indicating heterogeneity among studies (I 2 ) was considered as "low," "moderate," and "high" for values of 25, 50, and 75%, respectively [48].Considering the high heterogeneity expected for the outcomes selected, the random effect model was applied to evaluate the mean difference (MD) among continuous data when available.The Review Manager (RevMan) 5.3 software (Version 5.3.1 Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014) was used to perform meta-analyses.When a significant difference was detected by meta-analysis, meta-regression analyses were performed using "Statistical Package for the Social Science" software for Windows (version 28.0; SPSS Inc., Chicago, IL, USA).Statistical significance was considered for p values < 0.05.

Overview of Included Studies
The first literature search evaluated the relationship between sleep disorders and GM composition.Among 16846 studies detected, fifty-eight potentially relevant studies were extracted, and eighteen articles were finally included (Figure 1).Table 1 summarizes data from studies analyzed during the first literature search.
high heterogeneity expected for the outcomes selected, the random effect model was applied to evaluate the mean difference (MD) among continuous data when available.The Review Manager (RevMan) 5.3 software (Version 5.3.1 Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014) was used to perform meta-analyses.When a significant difference was detected by meta-analysis, meta-regression analyses were performed using "Statistical Package for the Social Science" software for Windows (version 28.0; SPSS Inc., Chicago, IL, USA).Statistical significance was considered for p values < 0.05.

Overview of Included Studies
The first literature search evaluated the relationship between sleep disorders and GM composition.Among 16846 studies detected, fifty-eight potentially relevant studies were extracted, and eighteen articles were finally included (Figure 1).Table 1 summarizes data from studies analyzed during the first literature search.While not statistically quantifiable, visual assessment of the number of GM components reported in studies shows high variability across studies.The mean age of patients considered was 35.7 ± 22.1 years, ranging from a minimum of 2-year-old children [54] to adults of 68.0 years [62] (Table 1).Similarly, high variability in sleep disorders was detected among studies.Indeed, sleep disturbance pattern includes several conditions of different nature, such as insomnia, reduced sleep quantity, obstructive sleep hypopnea/apnoea syndrome (OSHAS) [54,62], and narcolepsy type 1 [55], not allowing clear conclusions about associations between GM and sleep disturbances (Table 1).Relatively high study heterogeneity is also characterized by the tool used to evaluate sleep quality, which includes PSQI [65], sleep clinical record (SCR) [54], PSG [54,55], ESS [55], and multiple sleep latency test (MSLT) [55], while all studies considered the 16S ribosomal subunit mRNA to profile the microbiota (Table 1).
The second literature collection included studies describing oral daily supplements of probiotics or para-probiotics administered to individuals with sleep disorders.Among 5684 studies, thirty were evaluated and twenty-four finally included  (Figure 1).Finally, eight studies were eligible to be quantitatively analyzed for PSQI score [68,71,80,87,88,90,91] and three for ESS [87][88][89].Table 2 summarizes data from studies extracted during the second literature search.Overall, 274 subjects treated with probiotics were compared with 261 subjects receiving a placebo.Only one study showed a cross-over design, whereas others employed a randomized approach.Five studies recruited healthy individuals [68,71,78,82], one study [87] enrolled cirrhosis patients recovered from hepatic encephalopathy (HE), one employed [88] patients with irritable bowel syndrome (IBS) and sleep complaints, and one [90] recruited post-menopausal women.Moreover, three studies [71,87,88] reported baseline BMI, and one study enrolled patients with overweight (BMI > 25 kg/m 2 ).Regarding the use of probiotics, three studies employed single strains of bacteria [68,78,82] and five multiple strains [71,87,88,90,91] (Table 2).Typically, probiotics were administered within the 108-1011 CFU dose range for a variable duration of 3 [71] to 24 weeks [68,87].Different media were used to convey the treatment, such as capsules [78,88,91], tablets [68], powder sachets [71,87], fermented milk [82], or yogurt [90], revealing a certain grade of variability among studies for probiotic administration.Compliance with supplementation protocols, if shown, was higher than 95% [82,87,91], except for one study [68], reporting 92%.Thus, as occurred in the first literature search, a high heterogeneity among studies was expected for the aforementioned reasons.Among these, the patient's compliance with the probiotic consumption protocol should be considered as a further source of heterogeneity.However, none of the studies reported side effects or adverse events, suggesting general good tolerance to the treatment.

Meta-Analyses
Meta-analysis to compare GM composition between sleep disorders vs. healthy individuals was performed using α-diversity data (Figure 2).α-diversity was not different between patients with and without sleep disorders, considering the Shannon (p = 0.580, Figure 2A), Simpson (p = 0.160, Figure 2B), and Chao (p = 0.410, Figure 2C) indexes.Instead, β-diversity was not considered due to lack of reporting in more than three (i.e., 13) studies.
Overall, 274 subjects treated with probiotics were compared with 261 subjects receiving a placebo.Only one study showed a cross-over design, whereas others employed a randomized approach.Five studies recruited healthy individuals [68,71,78,82], one study [87] enrolled cirrhosis patients recovered from hepatic encephalopathy (HE), one employed [88] patients with irritable bowel syndrome (IBS) and sleep complaints, and one [90] recruited post-menopausal women.Moreover, three studies [71,87,88] reported baseline BMI, and one study enrolled patients with overweight (BMI > 25 kg/m 2 ).Regarding the use of probiotics, three studies employed single strains of bacteria [68,78,82] and five multiple strains [71,87,88,90,91] (Table 2).Typically, probiotics were administered within the 108-1011 CFU dose range for a variable duration of 3 [71] to 24 weeks [68,87].Different media were used to convey the treatment, such as capsules [78,88,91], tablets [68], powder sachets [71,87], fermented milk [82], or yogurt [90], revealing a certain grade of variability among studies for probiotic administration.Compliance with supplementation protocols, if shown, was higher than 95% [82,87,91], except for one study [68], reporting 92%.Thus, as occurred in the first literature search, a high heterogeneity among studies was expected for the aforementioned reasons.Among these, the patient's compliance with the probiotic consumption protocol should be considered as a further source of heterogeneity.However, none of the studies reported side effects or adverse events, suggesting general good tolerance to the treatment.

Meta-Analyses
Meta-analysis to compare GM composition between sleep disorders vs. healthy individuals was performed using α diversity data (Figure 2).α diversity was not different between patients with and without sleep disorders, considering the Shannon (p = 0.580, Figure 2A), Simpson (p = 0.160, Figure 2B), and Chao (p = 0.410, Figure 2C) indexes.Instead, β diversity was not considered due to lack of reporting in more than three (i.e., 13) studies.For the meta-analytic investigation of probiotics effectiveness in sleep disorders and health, only PSQI and ESS data were accessible in more than three articles.Other sleep measures were not suitable for meta-analysis.Probiotic consumption significantly reduced PSQI score compared with placebo administration (p = 0.04, Figure 3).For the meta-analytic investigation of probiotics effectiveness in sleep disorders and health, only PSQI and ESS data were accessible in more than three articles.Other sleep measures were not suitable for meta-analysis.Probiotic consumption significantly reduced PSQI score compared with placebo administration (p = 0.04, Figure 3).Meta-regression analysis was performed using patients' age as a covariate and th PSQI as the dependent variable.The difference between the study and control group wa significantly related to patients' age (Chi-squared 10.9, p = 0.012).
The use of probiotics was linked to ESS score reduction vs. placebo, although it wa not statistically significant (p = 0.070, Figure 4).Finally, the mean age of subjects did no differ between the study and control groups (−0.06; 95%CI: −0.05, 0.54 years, p = 0.850) excluding the age as a potential source for differences in PSQI score.

Discussion
This meta-analysis demonstrates that the use of probiotics improves sleep quality, a established by self-assessment questionnaires.However, the self-assessed propensity to fall asleep does not change after probiotics consumption, suggesting that the GM-sleep quality relationship could be only slightly influenced by this therapeutic intervention Indeed, here we demonstrate that the GM diversity is not different between patients with vs. without sleep disorders, suggesting that there is not a stable substrate on which probiotic could act to improve sleep quality.Results were obtained using α diversity data, which doe not provide a measure of the specific abundance of taxa or species in the gut.Moreover β diversity is missing due to a lack of reporting in more than three studies.This is limitation depending on the heterogeneity of the current literature and precludes th meta-analysis of individual microbiome taxa and/or species.Further, we have considered GM modulation by therapeutic measures on sleep quality as a topic recently explored in a meta analytics study [92], where authors evaluated the efficacy of probiotic or prebioti intervention on sleep characteristics.In this study, people with type 2 diabetes mellitus dementia, and bone fracture were included, and no improvements in sleep quality Meta-regression analysis was performed using patients' age as a covariate and the PSQI as the dependent variable.The difference between the study and control group was significantly related to patients' age (Chi-squared 10.9, p = 0.012).
The use of probiotics was linked to ESS score reduction vs. placebo, although it was not statistically significant (p = 0.070, Figure 4).Finally, the mean age of subjects did not differ between the study and control groups (−0.06; 95%CI: −0.05, 0.54 years, p = 0.850), excluding the age as a potential source for differences in PSQI score.
ks&Sleep 2023, 5, FOR PEER REVIEW For the meta-analytic investigation of probiotics effectiveness in sleep disorders a health, only PSQI and ESS data were accessible in more than three articles.Other sle measures were not suitable for meta-analysis.Probiotic consumption significan reduced PSQI score compared with placebo administration (p = 0.04, Figure 3).Meta-regression analysis was performed using patients' age as a covariate and PSQI as the dependent variable.The difference between the study and control group w significantly related to patients' age (Chi-squared 10.9, p = 0.012).
The use of probiotics was linked to ESS score reduction vs. placebo, although it w not statistically significant (p = 0.070, Figure 4).Finally, the mean age of subjects did n differ between the study and control groups (−0.06; 95%CI: −0.05, 0.54 years, p = 0.85 excluding the age as a potential source for differences in PSQI score.

Discussion
This meta-analysis demonstrates that the use of probiotics improves sleep quality, established by self-assessment questionnaires.However, the self-assessed propensity fall asleep does not change after probiotics consumption, suggesting that the GM-sle quality relationship could be only slightly influenced by this therapeutic interventi Indeed, here we demonstrate that the GM diversity is not different between patients w vs. without sleep disorders, suggesting that there is not a stable substrate on which probio could act to improve sleep quality.Results were obtained using α diversity data, which do not provide a measure of the specific abundance of taxa or species in the gut.Moreov β diversity is missing due to a lack of reporting in more than three studies.This i limitation depending on the heterogeneity of the current literature and precludes meta-analysis of individual microbiome taxa and/or species.Further, we have consider GM modulation by therapeutic measures on sleep quality as a topic recently explored in a me analytics study [92], where authors evaluated the efficacy of probiotic or prebio intervention on sleep characteristics.In this study, people with type 2 diabetes mellit dementia, and bone fracture were included, and no improvements in sleep qual

Discussion
This meta-analysis demonstrates that the use of probiotics improves sleep quality, as established by self-assessment questionnaires.However, the self-assessed propensity to fall asleep does not change after probiotics consumption, suggesting that the GM-sleep quality relationship could be only slightly influenced by this therapeutic intervention.Indeed, here we demonstrate that the GM diversity is not different between patients with vs. without sleep disorders, suggesting that there is not a stable substrate on which probiotics could act to improve sleep quality.Results were obtained using α-diversity data, which does not provide a measure of the specific abundance of taxa or species in the gut.Moreover, βdiversity is missing due to a lack of reporting in more than three studies.This is a limitation depending on the heterogeneity of the current literature and precludes the meta-analysis of individual microbiome taxa and/or species.Further, we have considered GM modulation by therapeutic measures on sleep quality as a topic recently explored in a meta-analytics study [92], where authors evaluated the efficacy of probiotic or prebiotic intervention on sleep characteristics.In this study, people with type 2 diabetes mellitus, dementia, and bone fracture were included, and no improvements in sleep quality (eighteen trials analyzed) and sleep duration (five trials analyzed) were found [92].These results contrast with our results since we found improvement in sleep quality parameters with probiotics administration versus placebo.This discrepancy could be due to the subject analyzed since only healthy subjects or subjects with specific diseases, such as inflammatory bowel disease and fibromyalgia, were included in the present study.Moreover, the authors considered sleep quality parameters using both standardized questionnaires and self-reported information.In the present work, only PSQI reported significant improvement of sleep disturbances compared with the placebo.ESS score seems to be not significantly different, although is shows a trend similar to what observed by PSQI analysis.Unfortunately, heterogeneity of patient's characteristics with sleep disorders, and wide variability in study experimental settings, make it difficult to generalize further about these results.The role of GM in sleep disorder was recently discussed and converging findings indicated that microbiota could be targeted by interventions aimed at improving sleep [93].However, we found that sleep disorders could not be associated with GM composition, irrespective to the index used to measure its diversity.On the contrary, PSQI offers the subjective perspective that probiotics administration, obviously influencing GM, improves sleep quality.We may assume that GM could impact sleep behavior, even if possibly weakly, and that this issue requires specific clinical studies with well-selected populations to be addressed.
The scientific interest in the relationship between GM and human functions' homeostasis has been increasing in recent decades [94,95].In fact, several lines of evidence indicate the existence of a connection between GM and the central nervous system, leading to the definition of the gut-brain axis (GBA) [96].The bidirectional flow of information between the GM and the central nervous system is poorly understood and probably works through different mechanisms, such as the hypothalamic pituitary adrenal (HPA) axis [2], the immune system [33,97,98], the intestinal neuroendocrine cells [99], the vagus nerve [100] and the enteric nervous system [2].The GBA influences sleep status in close interaction with emotions, physiological stress, and circadian rhythms [96].While the existence of these connections is overall clear, the potential impact of GM on sleep physiology has not been completely unraveled.Sleep patterns could be modulated by changes in intestinal permeability, inflammation, immune system activation, energy harvest, and bacterial diversity [101,102].However, the relationship between either GM composition or the effect of GM changes after probiotics consumption and sleep patterns remains questionable [103].Our meta-analysis was not able to quantify the relationship between GM composition and sleep disturbance pattern since no differences in α-diversity were found by any indexes used, such as Shannon, Simpson, and Chao.However, taken individually, authors found reduced α-and β-diversity of GM in specific groups with sleep disturbances [38,51,53,54,57,61].For example, total GM diversity is positively correlated with increased sleep efficiency and total sleep time when otherwise healthy subjects were considered [104].Thus, we could speculate that the relationship between GM composition and sleep must be considered, evaluating each group of sleep disturbances separately.
The beneficial effect of probiotics supplementation on sleep disturbance seems to be weak but relatively well established, taking the present study into account with previous analyses [105].In fact, our findings confirmed that probiotics consumption improves the subjective measure of sleep quality in patients with sleep disturbances when compared with placebo groups.Also, in this case, both sleep disturbance patterns and probiotic composition are extremely variable among studies.Despite this heterogeneity, PSQI is statistically different between the two groups considered, suggesting that the beneficial effect of probiotics could be stronger overall than expected, overcoming potential biases, such as selection and detection biases.Moreover, the mean age of participants is similar between the two groups, excluding the role of age in the observed differential composition of microbial community under probiotics administration.Other subjective sleep questionnaires could be less indicative than PSQI of effects obtained upon probiotics administration.For instance, ESS investigates general parameters such as day time sleepiness, instead of straightly the sleep quality [106].These considerations might explain why we found no significant improvement in ESS after probiotics administration.However, it is worthy of note that only three studies evaluated ESS [87][88][89], limiting the statistical power of this subgroup meta-analysis.This is a drawback that may be improved by increasing the number of clinical trials on this specific topic.Interestingly, one study was performed on subjects with bipolar disorder, reporting a negative correlation between Lactobacillus count and sleep [61].This finding, obtained in a specific and limited set of subjects, is suggestive of a possible effect of this bacterial strain on sleep health.
It is intriguing that major benefits on sleep were observed among university students exposed to stressful conditions (e.g., academic examinations or courses), as indicated by some of the studies evaluated [68,71,82].We may speculate that, in these groups, probiotics mitigate markedly the physiological response to stress exposure, via modulation of glucocorticoid action.A perceived environmental threat, such as a stressful event, triggers the HPA axis to secrete glucocorticoids and the sympathetic nervous system, leading to catecholamine release [107].Moreover, increasing glucocorticoid levels were detected even in short sleep conditions [103].Some studies indicated that probiotics administration attenuates the physiological increase in salivary glucocorticoids [82,106], potentially promoting sleep enhancement.This hypothesis suggests a beneficial role of probiotics for the general management of sleep disturbances, although subgroup analyses would be useful to establish firm conclusions.
Gender should be considered among potential determinants for high heterogeneity.Women were enrolled in only one study of those considered [90], whereas other trials included mixed groups of both males and females.Since sex-related differences in the effect of probiotics on sleep were suggested [74,84], we may assume that the heterogeneous gender composition of studies meta-analyzed could be a limitation.According to these studies, we could speculate that better outcomes occurred in men than women after Lactobacillus gasseri supplementation, although the limited availability of data prevented more accurate evaluations.Second, the participant's health status varied among trials, including either healthy subjects or individuals affected by different pathological conditions [87,88].Last, the experimental intervention differed for the type of probiotic bacteria, i.e., species, single vs. multiple strains, dosage, duration of the supplementation, and delivery method.In conclusion, methodological biases should be carefully considered by other studies on this topic.
The present study shows several limitations.The number of trials using the same methodological approach to evaluate sleep disturbance, GM composition, and probiotics type is low.Moreover, studies are highly heterogeneous, including participants with different ages and health statuses and assessment methods to evaluate sleep quality (i.e., subjective or objective measures).In this context, we were not able to collect data useful to adjust the analyses performed.The only available parameter was the patient's age, which, as demonstrated by meta-regression analysis, potentially influenced the final result of the meta-analysis.Moreover, we could collect only α-diversity data, while β-diversity is missing due to lack of proper reporting, limiting the analysis of how individual microbiome taxa or species may contribute to mitigate sleep disturbances.In particular, the variability in Systematic Reviewpatients' ages should be carefully considered since it is well-known that sleep patterns undergo physiological changes with aging [108,109].In our study, the limited number of studies detected does not allow subgroup analyses based on age and gender.Third, the evaluation of sleep quality by questionnaires is obviously influenced by the subjectivity of these instruments.Although self-reported questionnaires are broadly used for clinical and research purposes, they do not provide objective parameters.Thus, data collected could be less accurate than those obtained by objective sleep assessment tools, such as PSG or actigraphy.Finally, only gut microbial genera/species significantly associated with the outcome are likely to be reported in the literature.This would introduce publication bias and limit the number of included studies.Further clinical trials with randomized placebo-controlled designs and objective measurements are required to achieve solid conclusions.
In conclusion, the connection between sleep disturbances and GM homeostasis remains to be established since major drawbacks prevent definitive conclusions about this issue.Promising results based on subjective assessment of sleep quality suggested that probiotics could represent an effective clinical intervention to manage sleep disorders targeting GM.The understanding of variables impacting GM, such as diet, physical activity, stress, and pathological conditions, will be crucial to clarify the relationship between intestinal bacteria and sleep.

Figure 1 .
Figure 1.Study flow chart.The relationship between GM composition and sleep disorders (metanalysis 1) and the effect of probiotics on sleep disturbance (metanalysis 2) were assessed.

Figure 1 .
Figure 1.Study flow chart.The relationship between GM composition and sleep disorders (metanalysis 1) and the effect of probiotics on sleep disturbance (metanalysis 2) were assessed.

Table 1 .
Studies detected during the first literature search.

Table 2 .
Studies detected during the second literature search.