The Gut Microbiome in Depression and Potential Benefit of Prebiotics, Probiotics and Synbiotics: A Systematic Review of Clinical Trials and Observational Studies

1Major depressive disorder (MDD) is a mood disorder characterized by discrete episodes of at least two weeks involving changes in affect, cognition, and neurovegetative functions, with inter-episode remissions. With an estimated lifetime prevalence of 10.8% in community samples globally, it is the most common psychiatric illness worldwide, and a leading cause of disability.

2One of the most widely accepted models of depression pathophysiology has been the monoamine hypothesis, which postulates that the underlying pathophysiology of depression is a depletion of the neurotransmitters serotonin, norepinephrine, and dopamine in the central nervous system. However, there are several limitations to this model, including the fact that up to 30% of patients with MDD do not respond to monoaminergic antidepressants, these medications are effective in psychiatric disorders with differing pathophysiologies, and disruptions of monoamine systems do not exacerbate existing depressive symptoms. Therefore, there likely exist other mechanisms by which depressive pathologies arise which might open new avenues for treatment.

3The gut–brain axis has been postulated to be involved in the onset of depression. The gut–brain axis is a bidirectional communication network between the gut and the brain that operates by neuroimmune and neuroendocrine processes. It is mediated by several molecules, including short chain fatty acids, secondary bile acids, GABA neurotransmitters, and tryptophan metabolites, which are derived from the microbiota.

4During dysbiosis, or a disruption to microbiota homeostasis, gut–brain pathways are dysregulated and associated with neuroinflammation and altered permeability of the blood–brain barrier. Microbiota alterations may produce changes in depression by directly affecting release of the neurotransmitters serotonin and dopamine, influencing the stress response and hypothalamus–pituitary–adrenal (HPA) axis, influencing levels of brain-derived neurotrophic factor (BDNF) and triggering the release of inflammatory cytokines. For example, depression is associated with the release of C-reactive protein (CRP) and cytokines such as IL-1, IL-2, IL-6, IFN-γ, and IL-1β. For an in-depth review of the gut brain axis, see this 2019 review by Cryan and colleagues.

5A study by Guida and colleagues (2017) revealed that antibiotic-induced dysbiosis in mice led to a general inflammatory state, and depressive-type behaviour, but was reversed with the probiotic Lactobacillus casei. Fecal microbial transplant from humans with MDD to microbiota-deficient rodents has also been shown to induce a depression-like phenotype. While human interventional trials are only beginning to gain traction, preliminary evidence from observational studies has shown that the gut microbiome (GMB) profiles of patients with MDD and depressive symptoms differ significantly from healthy controls (HC), as well as patients with other mood and anxiety disorders.

6Given this emerging evidence implicating the gut–brain axis in depression, there has been interest in developing treatments, namely probiotics, prebiotics, synbiotics, and microbiota restoration therapies that target the GMB. Probiotics are preparations of microorganisms that, when administered, improve gut microbial balance, while prebiotics are nondigestible compounds (e.g., fructooligosaccharides, galacto-oligosaccharides, and xylooligosaccharides) that are metabolized by gut microorganisms, modulating GMB composition to ultimately benefit the host. The scientific literature classifies prebiotics as functional foods, given their role in promoting health and preventing disease. When combined synergistically, probiotics and prebiotics are known as synbiotics. Newer treatment modalities such as microbiota therapeutics, have yet to be evaluated in MDD but based on proof-of-concept hold promise for the treatment of MDD. Microbiota therapeutics include whole fecal microbiota transplants (FMT), symbiotic microbial consortia, or engineered symbiotic microbes. The goal of microbiota therapeutics is to reconstitute a dysbiotic microbiota with a healthy microbiota.

7Probiotics most often consist of combinations of Lactobacillus and Bifidobacteria genera. These microorganisms have been shown to suppress inflammation and modulate the immune system by preventing the induction of the cytokine IL-8 in human colon epithelium, as well as reduce intestinal permeability, inhibiting endotoxemia. Several studies suggest that probiotic use confers physical and mental health benefits to the host, including as treatments for depression. In preclinical studies, administration of the probiotic Bifidobacterium infantis to rats has been shown to reverse experimentally-induced stress and depression, while supplementation with Lactobacillus rhamnosus for 28 days results in a decline in depressive symptom ratings. A more recent study by Li and colleagues (2018) showed that in a chronic mild stress mouse model of depression and anxiety, there was a reduction in Lactobacillus species, and an increase in the inflammatory markers IFN-γ, TNF-α, and indoleamine 2,3-dioxygenase-1 levels in the hippocampus. Probiotic supplementation attenuated anxiety and depressive-like behaviors, significantly increased Lactobacillus abundance, and reversed immune changes. These studies provide evidence that the antidepressant-like activity of probiotics may operate through a gut microbiota–inflammation–brain axis.

8Several reviews have also demonstrated a relationship between the GMB and major depressive disorder in human participants. However, these studies tend to be restricted to patients who meet strict criteria for MDD, at the expense of including more common, subclinical forms of depression that are prevalent in inpatient and healthy populations and may be more likely to respond to prebiotic, probiotic, and synbiotic supplementation. In a recent systematic review and meta-analysis, Hofmeister and colleagues (2021) did report a statistically significant benefit of probiotic, prebiotic, and synbiotic interventions in people experiencing depressive symptoms (irrespective of MDD diagnosis). It is important to note, however, that they included patients with comorbid medical and psychiatric disorders who may have differing gut microbiome compositions than patients with depression and HC, and did not review evidence from observational studies.

9The most recent systematic review of both observational studies and interventional trials was conducted by Sanada and colleagues in 2020 which included studies published until October 2019. The review synthesized evidence from ten observational studies that investigated differences in GMB diversity and taxa abundance in patients with MDD compared to HC, and six clinical trials that investigated changes in depressive symptom severity following probiotic or synbiotic administration. The authors report an overall effect of prebiotic and probiotic treatment on depressive symptoms, but inconsistent findings on GMB differences between MDD patients and healthy controls at the phylum level. Since then, an additional 13 interventional trials and 16 observational studies have been conducted which provide new evidence for our analysis. Thus, the present systematic review seeks to understand how the microbiota composition of patients with MDD or depressive symptoms differs from healthy controls, and the potential effects of prebiotic, probiotic and/or synbiotic treatment on depressive symptoms using an updated body of literature.

10We report the process used to select the observational studies and clinical trials in two PRISMA Flow Diagrams.

11In the search for observational studies (Figure 1), we identified 333 records through database searching. After removing 111 duplicates, we screened the title and abstracts of 222 articles, and assessed 37 full texts for eligibility. We also identified and screened three records from other reviews and websites. Ultimately, 24 observational studies (2817 participants) were included in our analysis.

12In the search for clinical trials (Figure 2), we identified 382 records through database searching. After removing 71 duplicates, we screened the title and abstracts of 311 articles, and 26 full texts based on our eligibility criteria. We also identified and screened three records from other reviews and websites. We ultimately selected 19 observational studies (1119 participants) for inclusion in our analysis.

13Characteristics of the 24 observational studies included in this analysis are reported in Table 1. The sample consisted mainly of patients with MDD, who were predominantly females, with ages ranging from 20 to 80 years. Diagnoses were confirmed by the Diagnostic and Statistical Manual of Mental Disorders (DSM), Mini International Neuropsychiatric Interview (MINI), Structured Clinical Interview for DSM-5 (SCID) (version IV or V) or International Classification of Diseases-10 (ICD-10). We also included six studies that involved patients with IBS, BD, and anxiety as they had subsamples of patients with MDD and HC alone. The majority of studies (19 out of 24) were recruited in Asia, but three studies were performed in North America, and two in Europe.

14All studies conducted genetic analyses on the fecal microbiota (Table 2). Most studies conducted 16S rRNA gene sequencing, examining the V3-V5 regions of the genome, however Rhee and colleagues (2020) as well as Caso and colleagues (2021) used 16s rDNA. Three studies reported shotgun metagenomic sequencing (SMS) with one study reporting metaproteomic analysis. The majority of studies reported no significant differences in the alpha diversity of patients with MDD and healthy controls (14 of 21). However, two thirds of the studies (12 of 18) showed that beta diversity significantly differed between patients with MDD and HC.

15The abundance of bacterial taxa in patients with MDD relative to HCs, based on the results of 23 studies, are reported in Table S1, and summarized in Table 3. The following taxa were increased in patients with MDD: the families Bifidobacteriaceae and Streptococcaceae (reported in four studies) as well as the genera Eggerthella (six studies) and Streptococcus (five studies). Conversely, there was a decrease in the phylum Bacteroidetes (four studies), family Sutterellaceae (four studies), genus Coprococcus (six studies), and genus Faecalibacterium (seven studies). We observed other changes in the relative abundance of bacterial species in three or fewer studies; however, the results are not described here.

16Characteristics of the 19 clinical trials included for synthesis are reported in Table 4, and major findings from these studies in Table 5. The sample consisted of patients with MDD, a depressive episode, or depressive symptoms, with more females than males. The mean age of participants ranged from 20 to 40 years. Diagnoses were ascertained by DSM-IV, DSM-V, and ICD-10. Five studies included subclinical forms of depression defined by various scales, including the Quick Inventory of Depressive Symptomatology (QIDS-SR16), the depression subscale of the Depression, Anxiety, and Stress Scale (DASS-42), the Hamilton Depression Rating Scale (HAM-D), and the Edinburgh Postnatal Depression Scale (EPDS). Studies involved populations from heterogeneous regions, including North America, Asia, Europe, and Australia.

17About half of the studies were double-blind, randomized controlled trials (RCT) (11 of 19 studies), but results were also published in open label trials (four studies), pilot trials (two studies) and triple blind RCTs (one study). Most often, the intervention used was a probiotic (17 studies), followed by prebiotic (three studies) and synbiotics (one study). Probiotics mainly consisted of a combination of Lactobacillus and Bifidobacterium species. One prebiotic study explored the benefits of Inulin 10 g per day relative to placebo. An additional two studies compared the benefits of the prebiotic galactooligosaccharide relative to the probiotic CEREBIOME® and placebo. Only one study compared the benefits of a synbiotic (Familact H, a combination of fructooligosaccharide and the species L. casaei, L. acidofilus, L. bulgarigus, L. rhamnosus, B. breve, B. longum, and S. thermophilus) relative to placebo.

18The majority of the trials were placebo-controlled (15 of 19), although some studies had no controls or used patients with antidepressant medication only as controls. The follow-up period for most studies was eight weeks, but ranged from four to nine weeks. Depressive symptoms were most often assessed by the Beck Depression Inventory (BDI) and HAM-D.

19More than half (10 out of 17) of the probiotic studies demonstrated a significant decrease in the depressive symptoms of patients treated with probiotics over time, while six reported no significant decrease. In addition, one study reported mixed results, where there was a significant decrease in MADRS score in the probiotic group between baseline and four weeks, but no significant decrease between four and eight weeks. Of the three studies examining the benefits of prebiotic treatment, none reported significant decreases in depressive symptom scores over an eight-week follow-up period. However, significant decreases in symptoms were observed following synbiotic treatment for eight weeks in the single synbiotic study by Ghorbani and colleagues.

20Five of the interventional trials also included microbiota analysis. The majority of these studies (four out of five) reported no significant differences in alpha and beta diversity following probiotic administration. Results from four studies conducting taxa abundance analyses revealed an increase in the proportion of Ruminococcus gauvreauii, Coprococcus 3, Desulfovibrio, Faecalibacterium, Bifidobacterium, Adlercreutzia, Megasphaera, and Veillonella, as well as a decrease in Rikenellaceae_RC9_gut_group, Sutterella, and Oscillibacter in patients treated with probiotics compared to healthy controls.

21Here we examined the gut microbiota in patients with MDD based on 24 observational studies and 19 interventional trials. Taken together, the observational studies demonstrated no significant differences in alpha diversity in patients with MDD compared to HC. As alpha diversity is a measure of species richness and evenness within a single population, these findings suggest that the diversity of the GMB is similar for patients with MDD and HC.

22Our findings are consistent with those of a recent systematic review of gut microbiome composition in patients with MDD, BD, and SZ compared to HC conducted by McGuinness and colleagues (2022). The authors observed no strong evidence for a difference in the alpha diversity of bacteria in patients with psychiatric disorders compared to HC. Previously, a decrease in alpha diversity was hypothesized to exist in patients with psychiatric disorders. This was in line with the assumption that greater species number and diversity contributed to metabolic functional redundancy and resistance to pathogenic colonization, preventing disease. More recently, however, human gut microbiome studies suggest that there is limited utility of alpha diversity metrics in measuring gut health and distinguishing disease cases and controls. The evidence for alpha diversity changes in patients with MDD relative to HC is in fact largely mixed. For example, while McGuinness and colleagues (2022) reported no significant differences in alpha diversity between patients with psychiatric illness and HC, in a recent systematic review and meta-analysis of a pooled sample of patients with MDD and comorbid mental illnesses, Nikolova and colleagues (2021) observed significant differences in alpha diversity relative to HC. Equivocal alpha diversity findings have also been reported in neuropsychiatric diseases with similar pathophysiologies to depression, including Parkinson’s disease, autism spectrum disorder, and anxiety.

23We observed a significant difference in beta diversity between patients with MDD and HC across most studies. Beta diversity is a measure of between-samples diversity, which in our case, is the similarity of microbial communities in patients with MDD compared to HC. Our observations are consistent with findings in patients with MDD, and psychiatric disorders more generally where changes in beta diversity have been noted, and aligns with the hypothesis that MDD involves a dysbiotic state.

24In patients with depression, relative to HC, we also observed several taxa abundance changes. Specifically, there was an increase in the abundance of Streptococcaceae and Bifidobacteriaceae and families as well as Eggerthella and Streptococcus genera. Cheung and colleagues (2019) also observed an increase in the genus Streptococcus in patients with MDD. Streptococcus is a high metabolizer of amino acids and proteins, which may divert essential host amino acids to the microbes in a process called putrefaction and result in toxic products such as ammonia, putrescine, and phenol. Dysbiosis from putrefaction has also been noted in colorectal cancer and autism spectrum disorder. An association of Eggerthella with increased gastrointestinal inflammation has also been noted in patients with MDD, and further supports the idea that depression involves inflammatory states.

25Bifidobacterium species have been found to have anti-inflammatory effects on stress and depression, and are expected to be decreased in MDD. However, our findings of increased abundance are consistent with those noted in a recent systematic review of human studies published from January 2000 to June 2019 on the GMB in depression by Barandouzi and colleagues (2020). Notably, there is considerable species heterogeneity within taxa, including at the family level, and changes in Bifidobacteriaceae levels may not necessarily reflect increases in Bifidobacterium species.

26We also observed a decrease in the relative abundance of Bacteroidetes phylum, Sutterellaceae family, as well as Coprococcus and Faecalibacterium genera. These results are consistent with the findings from three other reviews. Nikolova and colleagues (2021), observed a decrease in Coprococcus and Faecalibacterium across several psychiatric disorders. These genera have anti-inflammatory properties, and are involved in the production of butyrate, a short chain fatty acid involved in the maintenance of the gastrointestinal mucosa and reduction in pro-inflammatory cytokines. Mucosal integrity is important for preventing endotoxins from entering the circulation and reducing uncontrolled inflammation. Barandouzi and colleagues also reported similar reductions in the level of Sutterellaceae in patients with MDD, but inconsistent findings in the abundance of Bacteroidetes. This observation may be due to the fact that the latter taxon is at the phylum level and therefore encompasses a more heterogeneous set of species. Decreased levels of Bacteroidetes have also been observed in females compared to males, and females were overrepresented among patients with MDD relative to HC in the observational studies.

27When analyzed together, the interventional trials show a modest benefit of probiotic and synbiotic, but not prebiotic treatment in reducing depressive symptoms of patients with MDD over four to nine weeks relative to placebo. Ten probiotic studies with a combined sample size of 543 participants, five of which were double-blind RCTs and four of which were open-label trials, demonstrated a significant decrease in depressive symptoms over time relative to placebo or antidepressant medication. However, seven high quality studies (N = 462), which were either double-blind or triple-blind RCTs, demonstrated no significant changes in depressive symptoms over time compared to placebo. None of the prebiotic studies demonstrated significant changes in depressive symptoms following intervention. The evidence therefore supports some benefit of probiotic, and synbiotic treatment in patients with MDD relative to placebo, but is largely equivocal.

28Our observations parallel those of recent reviews on the gut microbiota in MDD. In a recent systematic review and meta-analysis, Hofmeister and colleagues (2021) synthesized evidence from 50 RCTs that evaluated probiotic, prebiotic, synbiotic, paraprobiotic, or fecal microbiota transplant interventions in an adult population. The authors reported statistically significant benefits of probiotic, prebiotic, and synbiotic interventions on depressive symptoms, as measured by the BDI and the depression subscale of the Hospital Anxiety and Depression Scale. These findings are supported by those of an earlier systematic review and meta-analysis of 16 RCTs by El Dib and colleagues (2021), which demonstrated significant improvement in depression and anxiety symptoms in patients treated with probiotics according to the BDI and State-Trait Anxiety Inventory (STAI). However, in another systematic review of nine randomized double and triple blind placebo-controlled clinical trials, Minayo and colleagues (2021) reported no definitive effect of probiotics on depression and anxiety. Therefore, while there appears to be some benefit of probiotic, prebiotic, and synbiotic treatment in reducing depressive symptoms at 4 to 9 weeks follow up, the evidence to support this observation is mixed.

29One possible explanation for these mixed findings, is that the majority of interventional trials published to date include patients who meet strict criteria for MDD. It is possible, however, that patients with mild depression may derive more benefit from probiotic and synbiotic treatment than those with chronic, treatment-resistant depression. Additional studies in subsamples of patients with depression would be helpful in elucidating the benefits of these treatments.

30Prebiotic and synbiotic treatments for depression are also largely understudied, and the evidence is less concrete. While some reviews conclude that prebiotic supplementation, either alone or with probiotics, can have beneficial effects on mental health disorders, others report that they do not improve depression or anxiety symptoms. Very few studies have been conducted on these supplements and there is a need for multiple studies on each compound to be performed and analyzed separately.

31Regarding potential mechanisms, two studies suggested that changes in depressive symptoms may be due to altered GMB composition as a result of supplementation. There was an increase in the relative abundance of Bifidobacterium in the fecal microbiota of patients supplemented with Bifidobacterium breve CCFM1025 over four weeks. Supplementation with OMNi-BiOTiC® Stress Repair and Bifidobacterium breve CCFM1025 over four weeks also resulted in increase in Coprococcus and Faecalibacterium genera, which are normally decreased in MDD. These findings suggest that supplementation may alter GMB pharmacokinetics and pharmacodynamics, and depressive symptoms in patients with depression.

32We did not observe any appreciable changes in the alpha or beta diversity of the GMB, following supplementation, however. Scholars have noted that changes in diversity are sometimes not observed despite the presence of a highly divergent community composition, and understanding how alterations in composition affect microbiota functioning may provide more insight into disease.

33We identify important limitations of the studies included in this synthesis. Studies did not consistently control for geographic region or diet, despite these being well-established confounders in the GMB literature. High fat diets can, for example, increase the concentration of lipopolysaccharides, and stimulate the immune system, affecting the gut–brain axis. Moreover, while antidepressant medications, including tricyclic antidepressants (TCAs), monoamine oxidase inhibitors (MAOIs), and selective serotonin reuptake inhibitors (SSRIs), have been found to have antimicrobial effects, contributing to GMB changes, including dysbiosis, few studies included drug-naïve patients during the time of the intervention. Indeed, one eight-week pilot trial of a probiotic in treatment-naïve patients with depression (which was included in this review) was associated with significant improvements in affective clinical symptoms at four and eight weeks follow-up.

34An additional limitation of this synthesis is that we analyzed the overall effect of various probiotics, prebiotics, and synbiotics. Ideally, multiple studies for each specific compound should be performed and analyzed separately. These studies would discern whether specific compounds influence depressive symptoms while others do not. It would also be valuable to understand the pharmacokinetics, pharmacodynamics, and mechanism of action of these supplements. For example, do the probiotics engraft in the gut (pharmacokinetics) and lead to a change in microbial composition and metabolites (pharmacodynamics) with a direct effect on mood? Additional studies are needed to determine these effects.

35Sex differences in the GMB may also affect the results of the observational studies. Relative to HC, the sample of patients with MDD was predominantly female, who have been shown to have different microbiome profiles than males. Our analysis also focused on gut microbial composition (pharmacokinetics) rather than function (pharmacodynamics). However, functional potentials, including changes in short-chain fatty acid synthesis, tryptophan metabolism, and neurotransmitter synthesis and degradation, provide essential insights into the mechanisms of psychiatric illness.

36There was considerable heterogeneity in the species and dosage administered in the prebiotic, probiotic, and synbiotic interventions. Additionally, the follow-up time for the majority of studies was short (four to nine weeks). One open-label trial by Bambling and colleagues (2017) included in our analysis, found that there was a significant decrease in BDI score in the probiotic group over eight weeks, but these results did not persist at 16 weeks follow-up. While antidepressants tend to see benefit after six weeks, the ideal duration for observing benefits of probiotic treatments is unknown. The benefits of these treatments relative to antidepressants have also not been extensively studied. Therefore, additional studies with continuous monitoring, longer follow-up times, and antidepressant controls are required to elucidate the optimal species, dosage, and treatment length necessary for prebiotic, probiotic and synbiotic interventions to be used in clinical practice.

37The systematic review was conducted according to the PRISMA 2020 Guidelines. We conducted two independent searches for observational studies and clinical trials in PubMed, EMBASE, and PsycINFO from 1 January 2016 up to 6 January 2022.

38The search strategy for observational studies was as follows:Depression.mp. or exp Depression/or exp Depressive Disorder, Major/or depressive.mp. or exp Depressive Disorder/or exp Depressive Disorder, Treatment-Resistant/Gastrointestinal microbiome.mp. or exp Gastrointestinal Microbiome/or gut.mp. or fecal.mp. or microbiota.mp. or exp Microbiota/or microbiome.mp.(healthy control or healthy or control).mp.(alpha diversity or beta diversity or abundance* or diversity or rRNA).mp.1 and 2 and 3 and 4Limit 6 to English language, human, 2016-current.

39Depression.mp. or exp Depression/or exp Depressive Disorder, Major/or depressive.mp. or exp Depressive Disorder/or exp Depressive Disorder, Treatment-Resistant/

40Gastrointestinal microbiome.mp. or exp Gastrointestinal Microbiome/or gut.mp. or fecal.mp. or microbiota.mp. or exp Microbiota/or microbiome.mp.

41(healthy control or healthy or control).mp.

42(alpha diversity or beta diversity or abundance* or diversity or rRNA).mp.

431 and 2 and 3 and 4

44Limit 6 to English language, human, 2016-current.

45The search strategy for clinical trials was as follows: Depression.mp. or exp Depression/or exp Depressive Disorder, Major/or depressive.mp. or exp Depressive Disorder/or exp Depressive Disorder, Treatment-Resistant/probiotic.mp. or exp Probiotics/or prebiotic.mp. or exp Prebiotics/or synbiotic.mp. or exp Synbiotics/or exp Lactobacillus/or Lactobacillus.mp. or exp Bifidobacterium/exp Clinical Trial/or trial.mp.1 and 2 and 3Limit 5 to English language, human, 2016–current.

46Depression.mp. or exp Depression/or exp Depressive Disorder, Major/or depressive.mp. or exp Depressive Disorder/or exp Depressive Disorder, Treatment-Resistant/

47probiotic.mp. or exp Probiotics/or prebiotic.mp. or exp Prebiotics/or synbiotic.mp. or exp Synbiotics/or exp Lactobacillus/or Lactobacillus.mp. or exp Bifidobacterium/

48exp Clinical Trial/or trial.mp.

491 and 2 and 3

50Limit 5 to English language, human, 2016–current.

51We also identified additional articles for inclusion by searching the full-texts of reviews conducted on the GMB and depression.

52Observational studies that met the following criteria were included: (1) they were conducted in patients with a diagnosis of MDD according to a validated scale, (2) they involved healthy controls, (3) they conducted GMB analysis (taxa abundance differences, alpha diversity or beta diversity), and (4) they were peer-reviewed and published as full-texts in English.

53Clinical trials that met the following criteria were included: (1) they were conducted in patients with a diagnosis of MDD, depressive episode or depressive symptoms according to a validated scale, (2) they involved a prebiotic, probiotic, or symbiotic intervention, and (3) they were peer-reviewed and published as full-texts in English.

54For both observational studies and clinical trials, we excluded (1) studies published before 2016, (2) reviews, case reports, conference abstracts, dissertations, or letters, (3) protocol descriptions of studies not yet conducted, and (4) studies reporting only pooled results in patients with comorbid psychiatric and medical conditions (e.g., bipolar disorder (BD), schizophrenia (SZ), anxiety, and irritable bowel syndrome (IBS)). Studies involving patients with comorbidities were included as long as they conducted subgroup analyses in patients with MDD alone.

55The first author (S.R.A.) screened all records (titles and abstracts) and reports (full-texts). In cases of uncertainty, articles were independently screened by D.J.M. and I.G. until consensus was reached. Original study investigators were not contacted. No automation tools were used.

56For the observational studies, the primary outcomes of interest were differences in GMB composition between patients with depression and healthy controls, as measured by alpha diversity, beta diversity, and taxa abundance changes.

57Alpha diversity is a measure of the richness (number of species) and evenness (distribution) of the microbial community within one sample. It is assessed by several measures, including the Shannon index, Simpson index, phylogenetic diversity, total observed species or operational taxonomic units (OTUs), Chao 1, Inverse Simpson index, Sobe index and Abundance-based Coverage Estimator (ACE).

58Beta diversity, on the other hand, is a measure of the similarity or dissimilarity between the microbiota communities of two samples. It can be measured by simple taxa overlap, the Bray–Curtis Dissimilarity index, and UniFrac distance, among other indices. A significant difference in beta diversity between two groups indicates that the communities have significantly different species composition. Therefore, alpha diversity measure represents a summary statistic of a single population, while beta diversity measure represents a similarity score between populations.

59Taxa abundance changes reflect the proportion of bacteria at domain, kingdom, phylum, class, order, family, and genus levels in one group relative to another (for our purposes, patients with MDD compared to HC). Given inconsistent reporting about species, we did not assess changes at the species level.

60For clinical trials, the primary outcome measures were rating scale scores for depressive symptoms at four to nine weeks follow-up. Where multiple measures of depressive symptoms were available, we reported the results of validated, commonly used rating scales. We also collected information on microbiota changes (alpha diversity, beta diversity, and taxa abundance) as a secondary outcome.

61Data items that were collected for all studies include study design, mean age and standard deviation, sex, country, population, sample size, diagnostic definition of depression, depression rating scale, and genetic analysis techniques. Specifically for the clinical trials, we also collected data on the intervention type, control groups, trial length, probiotic bacterial strains, and depressive symptom score changes (see Table 1 and Table 2).

62Taken together, the findings from observational studies conducted between January 2016–January 2022 provide evidence for a specific gut microbial profile of patients with MDD compared to HC. The interventional trials suggest that there is modest benefit of probiotic and synbiotic, but not prebiotic, supplementation in reducing the symptoms of depression relative to placebo, and that probiotic treatment may influence GMB composition. However, additional and more rigorous double-blind randomized-controlled trials, which consider confounding factors such as symptom severity, age, diet, and medication use, are needed. Critical questions about species administered, dosage, and length of treatment remain to be addressed before these therapies reach the implementation stage as treatments for depression.