Microbiome-gut-brain signalling pathways encompass a wide variety of the physiological systems implicated in anxiety disorders (Fig. 1). A large body of evidence exists supporting a role for the microbiome in the early development and ongoing regulation of stress responsivity. Preclinical studies indicate that disruption in microbiome composition early in life using antibiotics, bacterial infections, Caesarean-section births, various acute and chronic stress exposures and other environmental influences can result in significant, enduring alterations in HPA axis activity and stress response [19]. Similar preclinical methods have been used to demonstrate the substantial immunomodulatory properties of the gut microbiome at birth and throughout the lifespan [20]. Although the stress response system and immune function are perhaps the most studied gut-brain signalling pathways, numerous other gut-brain cross-talk mechanisms are at play. Microbial regulation of the metabolism of tryptophan is important in stress-related psychiatric disorders, both due to the role of tryptophan as a precursor to serotonin and its metabolism via the neuroactive kynurenine pathway [21]. We have previously demonstrated an alteration in kynurenine pathway metabolites in social anxiety disorder [22]. Microbial metabolites are important mediators of gut-brain communication. Short-chain-fatty-acids (SCFA) are key metabolites produced by bacterial fermentation of fibre and exert both direct and indirect effects in the brain. They influence intestinal barrier integrity [23], which we have recently found to be disrupted in people with social anxiety disorder (unpublished data). SCFAs augment blood brain barrier tight junction expression and protect against central neurotoxin infiltration in addition to regulating host GI immunity and peripheral immunity, modulating HPA axis response and directly impacting the concentrations of neurotransmitters and neurotrophic factors in the gut lumen [23]. Moreover, exciting work has demonstrated that the SCFA, acetate, is one of the mechanisms underlying the known impact of the gut microbiota on the maturation and function of brain microglia [24, 25]. The endocannabinoid system, which is thought to buffer against many of the effects of stress [26], is also influenced by gut microbiota composition [27]. It has been demonstrated that the impact of the gut microbiota on depressive-like behaviours in mice is mediated by the endocannabinoid system [28]. An additional mechanism of gut-brain signalling is via hippocampal neurogenesis. Faecal microbiota transplant and probiotic supplementation in animal models has demonstrated that microbial signalling can alter levels of hippocampal BDNF and influence adult hippocampal neurogenesis, with subsequent consequences for cognition, stress and emotional regulation [29]. The myriad pathways through which the gut microbiota impact brain function and behaviour highlight the complexity of gut-brain interplay and the many potential avenues for the development of anxiety disorders.

The question is raised as to which level of taxonomic classification is most useful when exploring disease-associated microbiome differences [31,32,33]. Some studies report differences across all taxonomic levels from phylum down to species-level. However, others only report differences in genus and species level. A recent study that explored the classification of six diseases using a machine learning algorithm and gut microbiome data reported that the performance of classification is improved by using a lower taxonomy level; the highest performance was observed at the genus level [34]. This may be because lower taxonomic ranks show greater correlation with the faecal metabolome than higher order taxonomic groups and thus provide greater insights regarding crosstalk between the intestinal microbiome and the host [35]. We will, therefore, focus on genus- and species-level findings here.

Generalised Anxiety Disorder

Gut microbiome richness (observed number of species, Operational Taxonomic Units (OTUs) or Amplicon Sequence Variants (ASVs)) is reduced in GAD [36,37,38]. However, alpha diversity measures that also consider taxa evenness, e.g., Shannon and Simpson indices, do not appear to be different to healthy controls [36, 37].

Coprococcus and Faecalibacterium, other prominent butyrate-producers in the human gut, also appear important in anxiety disorders. Chen et al. [37] reported depleted levels of Coprococcus in GAD patients, with a similar finding seen in perimenopausal PD [42]. Such outcomes are consistent with a large cross-sectional study which used data from 7,656 participants of the Dutch Lifelines Microbiome Project (DMP) cohort [43]. This study collected metagenomic sequencing data along with a wide range of biomedical, socio-demographic, behavioural, physical and psychological metadata. They assessed for the presence of a range of psychiatric disorders based on Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR) criteria using a standardized diagnostic interview. Their analyses pertaining to anxiety disorders involved those with ‘any anxiety disorder’ (which they defined as GAD, SAD or PD) or those with GAD only. They did not analyse SAD or PD separately given the small numbers. The study aimed to explore the associations of the gut microbiome with anxiety and depressive disorders while adjusting for the use of psychotropic medications. This was an important study, given the potentially confounding impact of psychotropic drug use in many smaller cross-sectional studies. The study reported that any anxiety disorder (defined as GAD, SAD or PD), and GAD analysed individually, were significantly associated with a decreased relative abundance of Coprococcus eutactus, even after adjusting for psychotropic drug use. The overall conclusion from this study was that mood and anxiety disorders rather than psychotropic drugs are associated with compositional gut microbiome differences relative to controls. Faecalibacterium is another bacterial group that repeatedly emerges in the literature about psychiatric disorders and the gut microbiome. One GAD study found it to be significantly depleted in the patient group [36] and another study in GAD patients reported that Faecalibacterium negatively correlated with total plasma cortisol [38]. Additionally, Faecalibacterium was more abundant is people with GAD in remission compared with the active state [36]. An important metagenomics study using a large microbiome population cohort (Flemish Gut Flora Project, n = 1,054) with validation in independent datasets (n = 1,070) has previously found Faecalibacterium and Coprococcus bacteria to be consistently associated with higher quality of life indicators [44].

It is interesting that certain compositional findings have been replicated in a number of GAD studies, however it is unlikely that such changes are highly specific to GAD. A recent meta-analysis of gut microbiome alterations across a wide variety of mental disorders found a transdiagnostic pattern of microbiota signatures as opposed to any evidence of disorder specificity [30]. Depleted levels of Faecalibacterium and Coprococcus and enriched levels of Eggerthella were consistently shared between major depressive disorder, bipolar affective disorder, psychosis (undefined) and schizophrenia, and anxiety.

While Eubacterium, Faecalibacterium and Copcococcus are depleted in GAD, several bacterial groups may be more abundant in these individuals. Genera associated with GAD included Ruminococcus gnavus and Fusobacterium [36]. In a subgroup analysis of treatment-naïve patients, Escherichia–Shigella and Bacteroides were also enriched [36]. These taxa were also elevated in a subsequent cross-sectional study where their abundance was positively correlated with GAD symptom severity [37]. Additionally, a high abundance of Bacteroides eggerthii immediately after a two-month frontline work period during the Covid-19 pandemic was associated with future PTSD symptoms [41]. Bacteroides are a complex group of bacteria. While they represent a significant proportion of human gut commensals, these gram-negative obligate anaerobes can be highly pathogenic [45]. Similarly, Escherichia–Shigella is another pathogen which is associated with several human diseases [46]. Stress-induced proliferation of Escherichia coli in mice is associated with increased anxiety-like behaviours, decreased hippocampal BDNF expression and elevated gastrointestinal and hippocampal inflammation [47]. No such bacterial taxa were enriched in GAD by the larger Bruschett et al. (2023) study, which used the Dutch Lifelines cohort data. This may be because psychotropic drugs were accounted for as well as differences in study design (such as small sample sizes, differences in microbiome preparation and analysis, etc.).

Data in relation to functional microbiome differences in GAD is limited. No association was found between any functional gut-brain modules (GBMs) and anxiety disorders in the Dutch Lifelines cohort study [43]. GBMs represent a database of manually-curated microbial pathways known to impact brain function, based on extensive literature review. Each GBM corresponds to a single neuroactive compound production or degradation process [44]. A small study exploring microbiome differences between GAD, MDD and a control group reported differences in 69 Kegg Orthologues between the three groups, thus suggesting some differences in predicted microbiome function [38]. However, the numbers in each group were small and this observation requires further replication.

Panic Disorder and Agoraphobia

Studies exploring microbiome composition are limited in PD and, to our knowledge, have not been conducted in agoraphobia. A small cross-sectional Chinese study has investigated the oral microbiome in PD [48]. The oral microbiome was significantly more diverse in PD patients, and many taxa differences were observed between the patients and controls. The relative abundances of Prevotella and Veillonella were higher in the PD group. Authors reported a predominance of these taxa in periodontal disease, which is more likely in PD. Another small study has explored the gut microbiota in perimenopausal PD [42]. They reported reduced alpha diversity in perimenopausal PD patients. Similar to findings in GAD, butyrate-producing groups, including Faecalibacterium, Copcococcus and Roseburia were depleted in relative abundance in perimenopausal PD, while the genus Bacteroides was elevated.

Social Anxiety Disorder

Our research group has recently reported the first findings on the gut microbiota in social anxiety disorder [49]. While there had long been interest in the gut microbiota in anxiety and stress regulation, a growing appreciation for the role of the microbiome in social development and behaviour has developed in recent years [50]. We investigated the composition and function of the gut microbiome in 32 patients with social anxiety disorder in comparison to a healthy control group. No differences were seen in alpha diversity. However, we found that overall microbiota composition, as measured by beta-diversity, differed between the SAD and control groups. Several taxonomic differences were seen at a genus- and species-level: the relative abundance of the genera Anaeromassillibacillus and Gordonibacter were elevated in SAD, while Parasuterella was enriched in healthy controls. Anaeromassilibacillus is a member of the Clostridiales order of bacteria, a group which appears to show altered abundance in many psychiatric disorders and may represent disease-shared microbial responses [51]. In relation to functional differences, the gut metabolic module ‘aspartate degradation I’ was elevated in SAD patients. This functional pathway is associated with tryptophan-kynurenine metabolism, which we have previously demonstrated to be altered in SAD [22].

In order to test the hypothesis that the microbiota plays a causal role in SAD, we subsequently used faecal microbiota transplantation (FMT), a method used to assess potential causality and mechanisms [52, 53]. This involved the transfer of the microbiota from patients with SAD to antibiotic-depleted mice recipients and assessment of the behavioural and biological impact of such microbiota alteration [54]. Interestingly, the mice who received the SAD microbiome demonstrated a specific heightened social fear response, a validated mouse model of SAD [55]. They performed normally across other tests evaluating general anxiety-like and depression-like behaviours, an important feature of the study highlighting specificity for social fear responses. Additionally, changes in central and peripheral immune function and oxytocin expression in the bed nucleus of the stria terminalis were evident in the SAD-FMT-recipient mice.

Probiotics, Synbiotics and Prebiotics

Numerous systematic reviews and meta-analyses have explored the impact of probiotics and prebiotics on stress and anxiety symptoms over the past five years [56,57,58,59,60,61]. For the most part, probiotics appear to be a promising intervention for reducing stress and anxiety symptoms, although results can vary depending on study inclusion criteria and the type of probiotic used, since strain-specific effects and different mechanisms of action are likely. The vast majority of studies included in these meta-analyses were not patients with a formal clinical diagnosis of an anxiety disorder. Rather, they include various combinations of healthy volunteers, patients with IBS (a disorder of gut-brain interactions with significant psychiatric comorbidity including anxiety) or other medical conditions, subjects under stress or people with depression. A meta-analysis of 1146 healthy subjects found that probiotics reduced subjective stress and improved stress‐related subthreshold anxiety/depression levels, although no impact on cortisol levels was observed [58]. A larger meta-analysis of 29 randomised controlled trials (RCTs) (n = 2035 participants) found that probiotics and synbiotics were effective in reducing anxiety symptoms, but prebiotics had no effect [59]. This was similar to an earlier meta-analysis, which included 34 RCTs involving healthy subjects, medical patients (with a range of medical problems including IBS, multiple sclerosis, obesity, fibromyalgia, rheumatoid arthritis and laryngeal cancer) and subjects with MDD, which reported a small anxiolytic effect of probiotics but not prebiotics [56]. A recent meta-analysis included only patients who had a clinical diagnosis (MDD: n = 4 studies, GAD: n = 1 study) or healthy subjects who were under stress (academic stress: n = 4 studies, socially-evaluated cold pressor test (SCEPT) condition: n = 1 study) [57]. They reported that a probiotic reduced depression scores but not anxiety scores. A meta-analysis involving pregnant (n = 946) or lactating (n = 524) women reported that probiotics were effective in reducing both anxiety and depressive symptoms. This may be a particularly important group when it comes to increasing therapeutic options for anxiety and depression, given uncertainty about the potential impact of antidepressant exposure in pregnancy [62].

Only one randomised controlled trial to date has investigated the impact of a probiotic in GAD [63]. This Iranian study randomised 48 medication-free patients with GAD to receive either a placebo or multispecies probiotic (18*10⁹ CFU Bifidobacterium longom, Bifidobacterium bifidum, Bifidobacterium lactis, Lactobacillus acidophilus) in addition to 25 mg of Sertraline for eight weeks. The group receiving the adjunctive probiotic had significantly greater reductions in the clinician-rated Hamilton-Anxiety Rating Scale, although no differences were seen in Beck Anxiety Inventory or the State Trait Anxiety Inventory.

To our knowledge, there have been no clinical trials using microbiota-based therapies in SAD, PD or agoraphobia. However, a cross-sectional study of over 1000 university students found that higher intake of fermented foods appeared to be protective against developing SAD in those at higher genetic risk, as measured by trait neuroticism [64]. High intake of fermented foods may also be protective against general anxiety symptoms [65]. Fermented foods are an important source of potentially beneficial bacteria, generally containing various strains of lactic acid bacteria [66]. When consumed in high amounts by humans, certain fermented foods have anti-inflammatory effects [67] which may, in part account for the benefit in mental health.

Dietary Interventions

Diet is a major determinant of microbiome composition [68] and a promising intervention for psychiatric disorders, recently reframed under the banner of Nutritional Psychiatry [69]. It is well recognised from population studies that a Mediterranean diet is protective against depression [70]. More recently, there is evidence that high adherence to a Mediterranean-type diet may also be protective against anxiety symptoms in both adults [71, 72] and older people [73], as well as being associated with lower odds and severity of anxiety disorders [74]. Conversely, a diet characterised by high-fat, high-sugar and low fruit and vegetable intake, characteristic of the ‘Western style’ pattern of eating, is associated with elevated anxiety symptoms [75].

The mental health benefits of a Mediterranean diet extend beyond prevention. In 2017, the SMILES trial demonstrated for the first time that a Mediterranean diet intervention could improve depressive symptoms in Australian patients with MDD alongside standard treatment including psychotherapy and/or pharmacotherapy. [76]. Several subsequent clinical trials, also in Australia, demonstrated similar findings in depressed patients [77,78,79]. To date, no clinical trials have specifically explored the Mediterranean diet as a therapeutic intervention in anxiety disorders. However, in the aforementioned SMILES trial, a reduction in the Hospital Anxiety and Depression Scale (HADS)-anxiety subscale score was reported as a secondary outcome.

Interestingly, a Mediterranean diet intervention results in an increased abundance of Faecalibacterium and Roseburia [80, 81] taxa, which are depleted in GAD and PD [36, 42]. A study from our research group recently investigated the impact of a ‘psychobiotic diet’ in healthy volunteers [82]. The ‘psychobiotic diet’ included aspects of the Mediterranean diet like fruits, vegetables, whole grains, legumes, and seeds, as well as fermented foods. After four weeks of adhering to this diet, subjects reported reductions in perceived stress which were greatest in those with high adherence. Dietary intervention remains a promising therapeutic strategy for anxiety disorders and needs to be explored further.

It is an exciting time in neuroscience and psychiatry. The exponential increase of microbiome-gut-brain axis research over the past two decades has led to hope of new approaches for the treatment of anxiety. Given the growing burden of anxiety and stress-related disorders, along with the significant number of patients who do not respond fully to conventional treatments, alternative options and the availability of adjunctive approaches are vital. However, much work remains to be done. A key priority now must be extending the evidence base for microbiome interventions from studies in healthy, non-psychiatric populations to people with clinically diagnosed anxiety disorders, and with an increased focus on function over form. This applies to the spectrum of microbiome-based interventions, including probiotics, prebiotics, synbiotics, whole-diet interventions and individual dietary components such as fermented foods. There is reasonably robust evidence for using specific adjunctive probiotics in patients with MDD [56, 83, 84]. Additionally, a Mediterranean diet intervention can also be recommended to depressed patients [85]. However, evidence is lacking in patients with clinical anxiety disorders and although similar interventions do hold promise, they cannot confidently be recommended by psychiatrists at present.

Adequately powered clinical trials in well-characterised groups of people with GAD, PD, agoraphobia and SAD are required to investigate the therapeutic potential of microbiome-based interventions. The many confounding factors which influence the human gut microbiome must be accounted for including diet, psychotropic and other medications, smoking, alcohol use and body mass index. An additional avenue for future research is clarifying the mechanisms underlying the mental health benefit of such interventions. Further exploration of the impact of probiotics, dietary components such as fermented foods and bacterial metabolites such as short-chain fatty acids on HPA axis function [86, 87], immune response, tryptophan metabolism, vagal nerve communication, BDNF expression, blood–brain-barrier integrity and other physiological processes involved in brain function and mental health is needed. Moreover, while much attention has focused on gut bacteria, the gut virome is also an important component of the human microbiome. It was recently demonstrated in rodents that the virome plays a role in the modulation of the microbiota–gut–brain axis during stress [88] indicating that viral populations should be considered when designing future microbiome-directed therapies.

The microbiome-gut-brain axis may be a promising new therapeutic target for the millions of people worldwide suffering from anxiety disorders. However, it has been a neglected topic of research in clinical anxiety disorder cohorts despite the promising preclinical signals, which were among the first to be noted in the field. It will be important to parse the common or distinct roles of the microbiome in clinically-diagnosed anxiety disorders as well as in those with high trait anxiety as a risk factor in otherwise healthy individuals and in people with comorbid anxiety in psychiatric and other disorders. The limitation of a symptomatic ‘floor effect’ when exploring the anxiolytic properties of microbiome interventions in healthy non-clinical populations may have underestimated the potential of this option and strain-specific effects also need to be taken into account. It is time for microbiome researchers to turn their attention towards people suffering with GAD, PD, agoraphobia and SAD, as well as those with sub-threshold anxiety symptoms, in order to fully elucidate the potential of the microbiome-gut-brain axis in such conditions.