“Attacking” the Gut–Brain Axis with Psychobiotics: An Umbrella Review of Depressive and Anxiety Symptoms

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

 

• The authors provide an extensive overview of systematic reviews and meta-analysis on randomized controlled trials examining the effects of probiotic, prebiotic and synbiotic interventions in adults with depressive and/or anxiety symptoms or diagnoses.
• The reviewed articles showed consistent benefits of probiotics on MDD diagnoses, while benefits of prebiotics were limited. Results on anxiety and the use of synbiotics interventions were inconclusive.
• The authors provide a very thorough overview of the review setup, the geographical spread of the included clinical trials, and the connectivity analysis of the meta-analyses included (i.e. the shared references list which may reflect redundant results).

MAJOR COMMENTS

• Interestingly, the studies use rather different psychological assessments scales, which underscores a rather heterogeneous landscape amongst studies. I agree with the authors that this “challenges direct comparisons and underscores the need for analytical approaches that account for conceptual and sensitivity differences among instruments (Figure 5).”. However it is unclear for me how these different scales were homogenized in the review. P-values assess the statistical relevance, but do not tell how the scores connect. I suggest that they described this in their Summary and guidance of the results. If this results too complex, I then suggest they stratify the results based on scales, so separately for (at least some of) the scales, i.e. BDI, DASS, HADS. By doing this we also ensure that results are not driven by the characteristics of a i.e. specific scale.
• Extending to the previous point, I suggest the authors share their advice on a solution towards a more homogeneous landscape that allows for a better comparison (i.e. a common standard, or a consistent use of a multitude of assessments etc.? Please clarify your standpoint in the Discussion section.
• Regarding: “Here, “methodological robustness” refers to scales with stronger psychometric properties, broader validation across clinical populations, and clearer interpretability for anxiety severity”. Could they further specify the order of robustness of the different methods, and the criteria used?
• As mentioned, from the reviewed papers the authors found a consistent beneficial effect of probiotics on depressive symptoms. Practically, these are nearly all Lactobacillus and Bifidobacterium strains. While the authors (very) briefly touch up on a few functional characteristics of these strains, the  functional aspects should be elaborated more in detail. Is there a clear mechanistic insight, and also a proven link between the production of certain metabolites and the reduction of depressive symptoms - via gut-brain axis signaling? Or is this all still very speculative?
• Extending to the previous point: are the functional capacities i.e. to produce neurotransmitters unique to Lactobacillus and Bifidobacterium strains or could other genera potentially produce these as well? Or better: is the field focusing on two classical probiotics i.e. because of their safety and wide-spread use? Is this possibly not too reductive?
• I suggest the authors also elaborate on the effect of antidepressants and the combination of i.e. SSRIs with pre and probiotics.

MINOR COMMENTS

• “However, one of these trials, Kazemi et al., [57], published in Clinical Nutrition, appears with effect-size values that are statistically implausible.” I think that underlining the fact that the work was published in Clinical Nutrition can be removed as it suggests a judgement also about the journal.
• Please mark genus/species taxonomies in italic, i.e. L. Helveticus throughout the manuscript and tables (i.e. Table 1).
• In Table 1 be consistent to use the full genus name or the abbreviation. I.e. Lactobacillus acidophilus or L. Acidophilus.

Author Response

REVIEW PHARMACEUTICALS

REVIEWER 1

The authors provide an extensive overview of systematic reviews and meta-analysis on randomized controlled trials examining the effects of probiotic, prebiotic and synbiotic interventions in adults with depressive and/or anxiety symptoms or diagnoses. The reviewed articles showed consistent benefits of probiotics on MDD diagnoses, while benefits of prebiotics were limited. Results on anxiety and the use of synbiotics interventions were inconclusive. The authors provide a very thorough overview of the review setup, the geographical spread of the included clinical trials, and the connectivity analysis of the meta-analyses included (i.e. the shared references list which may reflect redundant results).

REPLY: We appreciate your dedication to this important review. With you, our paper becomes stronger!

MAJOR COMMENTS

Comment 1: Interestingly, the studies use rather different psychological assessments scales, which underscores a rather heterogeneous landscape amongst studies. I agree with the authors that this “challenges direct comparisons and underscores the need for analytical approaches that account for conceptual and sensitivity differences among instruments (Figure 5).”. However it is unclear for me how these different scales were homogenized in the review. P-values assess the statistical relevance, but do not tell how the scores connect. I suggest that they described this in their Summary and guidance of the results. If this results too complex, I then suggest they stratify the results based on scales, so separately for (at least some of) the scales, i.e. BDI, DASS, HADS. By doing this we also ensure that results are not driven by the characteristics of a i.e. specific scale.

REPLY: Dear reviewer, this is a very plausible and important question to answer. Regarding the issue of scales, perhaps I haven't been transparent or accurate enough. I think the crux of the matter lies not in the scales themselves, but in the subgroups selected for analysis. Therefore, to better clarify this issue, I will provide a supplementary list containing a plausible justification for their inclusion (global SMD vs. Subgroup analysis) in the final analysis. I believe this is transparent enough for all readers and critics. See our supplementary material (Table S4).

Additionally, we created a section in the results to properly explain this to readers and, especially, to be transparent about the matter.

Text added (results):

2.3.2 Synthesis Strategy and Selection of Effect Estimates

“The synthesis of results among the included meta-analyses required careful con-sideration of how SMDs were selected and extracted, given the substantial heterogeneity in study populations, symptom severity, and outcome assessment instruments. When available, subgroup-specific SMDs were preferentially extracted for analyses targeting depressive or anxiety symptoms, particularly when meta-analyses reported stratified estimates based on clinically diagnosed populations, symptom severity, or specific psychometric instruments. This approach aimed to enhance clinical interpret-ability and reduce dilution of effects that could arise from pooling heterogeneous populations, such as healthy individuals and participants without clinically relevant symptoms.

In contrast, global SMDs were retained when subgroup stratification was either not clearly defined, not methodologically justified, or when the study population was already homogeneous with respect to depressive or anxiety disorders. In several cases, global estimates were also preferred when further stratification would have substantially reduced the number of contributing studies, thereby compromising statistical stability and interpretability. For meta-analyses reporting multiple subgroup estimates based on different psychometric scales, the SMD derived from the most prevalent instrument was selected to ensure consistency and comparability among analyses.

This decision-making process was applied systematically and transparently across all included meta-analyses, with detailed justifications provided in Supplementary Table S4”.

Comment 2: Extending to the previous point, I suggest the authors share their advice on a solution towards a more homogeneous landscape that allows for a better comparison (i.e. a common standard, or a consistent use of a multitude of assessments etc.? Please clarify your standpoint in the Discussion section.

REPLY: a

Text added (3.5 Future directions):

“While the adoption of a single universal outcome measure could, in theory, en-hance comparability, such an approach is unlikely to be feasible or desirable given differences in study objectives, populations, symptom severity, and clinical versus sub-clinical contexts. From a practical and methodological standpoint, a more realistic solution lies in the standardization of analytical frameworks rather than in the exclusive use of a single instrument. In this review, comparability was addressed by prioritizing effect estimates derived from clinically relevant subgroups when available and by selecting SMDs based on instruments with stronger psychometric support, and greater prevalence among trials. Global SMDs were retained when populations were homogeneous or when further stratification would compromise statistical robustness.

Future research should move beyond highly generalized analyses that pool heterogeneous populations, symptom profiles, and probiotic formulations, as such approaches may obscure strain-specific effects and attenuate clinically meaningful signals. Accumulating evidence suggests that probiotic effects on mental health outcomes are not uniform [45,61,66,79,80], but instead depend on clearly defined strain characteristics, and well-characterized participant profiles. Accordingly, future trials should prioritize more phenotypically and biologically specific study designs, including the evaluation of single strains or well-justified.

In parallel, greater emphasis should be placed on recruiting clinically defined populations, such as individuals with established diagnoses of depressive or anxiety disorders, or on clearly stratifying participants by symptom severity and baseline microbiota characteristics. It is worth highlighting that, only a limited number of existing studies have adopted such stringent designs, focusing on specific strains, or clinically diagnosed populations (observed in more recent studies) [45,61,66,79,80], yet these studies tend to provide clearer and more interpretable effect estimates.

Finally, despite growing interest in microbiota-targeted interventions, the current evidence base remains heavily skewed toward prebiotic formulations, limiting the ability to differentiate the effects on depressive and anxiety symptoms. Future progress in this field will require methodologically rigorous primary studies specifically designed to evaluate prebiotics and synbiotics, with clearly defined compositions”.

Comment 3: Regarding: “Here, “methodological robustness” refers to scales with stronger psychometric properties, broader validation across clinical populations, and clearer interpretability for anxiety severity”. Could they further specify the order of robustness of the different methods, and the criteria used?

REPLY: Dear reviewer, significant heterogeneity is observed in the RCT papers available in the literature, and consequently, in the meta-analyses. Therefore, frankly, we had not initially planned for this issue. Upon encountering the results available in the literature, we sought to report our findings with due transparency and in the best possible way. In this context, methodological robustness was not intended to imply an absolute hierarchy of psychometric superiority, but rather a pragmatic and literature-supported prioritization of outcome measures based on their widespread use and interpretability across clinical and research settings. The criteria underlying this classification included: the extent of psychometric validation among diverse clinical populations; frequency of use in randomized controlled trials and meta-analyses; and clinical interpretability, including the availability of established severity thresholds and clinically meaningful change estimates. Based on these criteria, clinician-administered or extensively validated self-report instruments, such as BDI or STAI, was considered methodologically more robust due to their long-standing use, cross-cultural validation, and comparability among studies. Other instruments employed in the included trials (e.g., symptom-specific or mood-state questionnaires) are valid and sensitive to psychological change; however, they often present greater heterogeneity in scoring interpretation, population-specific validation, or limited cross-study comparability. Therefore, we improved the text to avoid any confusion. This approach is consistent with prior evidence-synthesis methodologies that prioritize outcome measures with greater standardization and cross-study interpretability.

Adjusted text (results):

2.3.2 Synthesis Strategy and Selection of Effect Estimates

“The synthesis of results among the included meta-analyses required careful con-sideration of how SMDs were selected and extracted, given the substantial heterogeneity in study populations, symptom severity, and outcome assessment instruments. When available, subgroup-specific SMDs were preferentially extracted for analyses targeting depressive or anxiety symptoms, particularly when meta-analyses reported stratified estimates based on clinically diagnosed populations, symptom severity, or specific psychometric instruments. This approach aimed to enhance clinical interpret-ability and reduce dilution of effects that could arise from pooling heterogeneous populations, such as healthy individuals and participants without clinically relevant symptoms.

In contrast, global SMDs were retained when subgroup stratification was either not clearly defined, not methodologically justified, or when the study population was already homogeneous with respect to depressive or anxiety disorders. In several cases, global estimates were also preferred when further stratification would have substantially reduced the number of contributing studies, thereby compromising statistical stability and interpretability. For meta-analyses reporting multiple subgroup estimates based on different psychometric scales, the SMD derived from the most prevalent instrument was selected to ensure consistency and comparability among analyses.

This decision-making process was applied systematically and transparently across all included meta-analyses, with detailed justifications provided in Supplementary Table S4”.

Comment 4: As mentioned, from the reviewed papers the authors found a consistent beneficial effect of probiotics on depressive symptoms. Practically, these are nearly all Lactobacillus and Bifidobacterium strains. While the authors (very) briefly touch up on a few functional characteristics of these strains, the functional aspects should be elaborated more in detail. Is there a clear mechanistic insight, and also a proven link between the production of certain metabolites and the reduction of depressive symptoms - via gut-brain axis signaling? Or is this all still very speculative?

REPLY: Dear reviewer, you are correct. Pharmaceuticals is a demanding journal, and I did not provide more in-depth explanations of the mechanisms, especially those that were cited.

Text added (discussion):

Strain-Specific Mechanistic Pathways Linking Lactobacillus and Bifidobacterium

It is important to highlight that a critical characteristic of the trials included in this review is the predominance of Lactobacillus and Bifidobacterium strains, admin-istered either as single strains or in multispecies formulations. This observation is rel-evant because the biological effects attributed to probiotics are not uniform, but rather depend on strain-specific and functional properties. Psychobiotic effects arise from specific microbial capacities related to metabolite production, immune modulation, and host–microbe signaling, rather than from probiotic exposure per se [6,47,59,79].

Experimental evidence provides robust support for neural communication as one of the pathways linking specific Lactobacillus strains to the regulation of stress and af-fective states. Chronic administration of Lactobacillus rhamnosus (JB-1) has been shown to induce region-specific alterations in GABA receptor expression and to at-tenuate stress-related behaviors; notably, these effects are abolished following vagot-omy, identifying the vagus nerve as a critical conduit for gut–brain signaling [78]. Cryan et al., [59] position this vagal pathway as one of the most biologically coherent routes of gut–brain communication. However, despite its strong biological plausibility, direct confirmation of vagal mediation in the improvement of depressive and anxiolyt-ic symptoms in humans remains limited, as few randomized controlled trials have simultaneously assessed vagal biomarkers.

In addition, the low-grade systemic inflammation commonly observed in patients with depressive disorders constitutes a biologically plausible target for microbio-ta-mediated interventions. Both Lactobacillus and Bifidobacterium strains have demonstrated anti-inflammatory properties, including reductions in pro-inflammatory cytokines and markers of oxidative stress. Clinical trials have reported concurrent im-provements in depressive symptom scores alongside reductions in C-reactive protein, IL-6 expression, and oxidative stress indices following probiotic supplementation [1,36]. Sulaiman et al., [47] identify immune modulation as one of the most consistently supported mechanisms across human studies, suggesting that attenuation of peripher-al inflammation may indirectly influence neuroinflammatory processes and HPA ac-tivity. Among the proposed pathways, immune–brain signaling currently represents one of the strongest translational links between probiotic exposure and reductions in depressive symptoms.

Another frequently cited mechanism involves the regulation of tryptophan avail-ability. Certain Bifidobacterium strains have been associated with alterations in tryp-tophan metabolism and increased peripheral serotonin availability in clinical trials [80]. This pathway provides a coherent intersection between immune activation, microbial function, and neurotransmission. However, it is now well recognized that depression and its comorbidities do not depend exclusively on monoaminergic pathways, and therefore this mechanism should be interpreted with caution [58,81].

Emerging neuroimaging data further suggest that probiotic interventions may in-fluence brain structure and functional connectivity within frontolimbic circuits impli-cated in mood regulation. Changes in hippocampal activation patterns, fractional ani-sotropy, and gray matter volume have been reported in small clinical samples follow-ing supplementation with multispecies probiotics [82,83]. Although these findings are intriguing, further studies are required to confirm and extend these observations.

Comment 5: Extending to the previous point: are the functional capacities i.e. to produce neurotransmitters unique to Lactobacillus and Bifidobacterium strains or could other genera potentially produce these as well? Or better: is the field focusing on two classical probiotics i.e. because of their safety and wide-spread use? Is this possibly not too reductive?

REPLY: The predominance of Lactobacillus and Bifidobacterium strains in psychobiotic research should not be interpreted as evidence that neuroactive or gut–brain signaling capacities are unique to these genera. A growing body of experimental and observational research indicates that multiple bacterial taxa possess the ability to produce or modulate neuroactive compounds, including gamma-aminobutyric acid, serotonin precursors, short-chain fatty acids, and other metabolites relevant to brain function. Genera such as Streptococcus, Enterococcus, Escherichia, Bacillus, and Akkermansia have also been implicated in microbiota–brain interactions, at least at the preclinical or associative level. The current emphasis on Lactobacillus and Bifidobacterium largely reflects pragmatic and regulatory considerations rather than exclusivity. These genera have a long history of safe use in humans, well-established manufacturing processes, and favorable regulatory status, which facilitate their inclusion in clinical trials. While this focus has enabled early translational progress, it may also contribute to a reductive view of psychobiotic potential.

Comment 6: I suggest the authors also elaborate on the effect of antidepressants and the combination of i.e. SSRIs with pre and probiotics.

REPLY: Dear reviewer, the suggested approach is possible when we discuss the results found; therefore, we will include a section in the discussion of our article. Consider it done. If the esteemed reviewer is suggesting including this comparison in our objectives, this becomes impossible, as we did not plan this comparison beforehand; we would be committing misconduct, and this is questionable in science and grounds for strong criticism.

Text added (discussion):

3.1 Interaction Between Antidepressants and Probiotics/Prebiotics

It is important to contextualize these findings within the broader pharmacological landscape of depression treatment, particularly regarding the widespread use of anti-depressants such as selective serotonin reuptake inhibitors (SSRIs). SSRIs remain first-line therapy for major depressive disorder; however, their clinical effectiveness is frequently limited by delayed onset of action, incomplete remission rates, substantial interindividual variability in response, and the occurrence of adverse effects, notably gastrointestinal symptoms [70,71].

Evidence indicates that antidepressants themselves may interact bidirectionally with the gut microbiota [71]. Several SSRIs exhibit intrinsic antimicrobial activity, capable of altering microbial composition and diversity, which may influence both therapeutic efficacy and tolerability [2,59,61,69,72]. These microbiota alterations may partially explain antidepressant-induced gastrointestinal side effects and, potentially, variability in clinical response [73]. Conversely, baseline dysbiosis has been associated with altered antidepressant metabolism and reduced treatment responsiveness, suggesting that the gut microbiome may act as a moderator of pharmacological efficacy [7,73,74].

Probiotics and prebiotics may complement antidepressant therapy through convergent modulation of the microbiota–gut–brain axis [1,10,21,61,69,75,76]. Proposed pathways include increased production of short-chain fatty acids, regulation of inflammatory and immune signaling, modulation of tryptophan metabolism and serotonergic neurotransmission, reinforcement of intestinal barrier integrity, and attenuation of HPA axis hyperactivity [59,77]. Several of these mechanisms overlap with bio-logical targets indirectly influenced by antidepressant drugs, supporting a biologically plausible rationale for combined or adjunctive approaches [59].

Confirming this, a recent high-quality systematic review and network me-ta-analysis by Zhao et al., [48] directly compared microbiota-targeted therapies with antidepressants in adults with MDD. Among 42 RCTs, probiotics demonstrated significant reductions in depressive symptom severity compared with placebo and were non-inferior to multiple antidepressants, including venlafaxine, vortioxetine, duloxetine, and citalopram. Notably, probiotics ranked second only to escitalopram in treatment hierarchy and showed comparable tolerability to antidepressants when administered for ≥ 8 weeks. Moreover, probiotic interventions used as adjuncts to antidepressant therapy were superior to several pharmacological agents alone.

So, preclinical e clinical studies provide robust support for this interaction, demonstrating that specific probiotic strains can exert antidepressant-like effects and modulate stress-related neuroendocrine and inflammatory responses, sometimes comparable to conventional antidepressants [1,26,36,39,44,46,50,57,65,78]”.

MINOR COMMENTS

Comment 7: “However, one of these trials, Kazemi et al., [57], published in Clinical Nutrition, appears with effect-size values that are statistically implausible.” I think that underlining the fact that the work was published in Clinical Nutrition can be removed as it suggests a judgement also about the journal.

REPLY: Dear reviewer, I understand your position. In fact, I had considered not including this evidence. However, the article has criteria aligned with our study and has not been retracted by science. Instead, to maintain accuracy and eliminate this bias, we recalculated the SMD to adequately represent the effect that the aforementioned author posits. For a better explanation, observe the data from Nikolova et al., 2019.

We reversed the positive sign, since the effect represents a reduction in symptoms. We recalculated the SMD from Kazemi et al., and compared it with other meta-analyses that cited the same study.

Comment 8: Please mark genus/species taxonomies in italic, i.e. L. Helveticus throughout the manuscript and tables (i.e. Table 1).

REPLY: OK, consider done! The table has been properly adjusted and is marked in yellow.

Comment 9: In Table 1 be consistent to use the full genus name or the abbreviation. I.e. Lactobacillus acidophilus or L. Acidophilus.

REPLY: OK, consider done! The table has been properly adjusted and is marked in yellow.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

Hi dear Editorial board and the respected authors

This article "Psychobiotics for Depression and Anxiety: An Umbrella Review and Critical Appraisal of Systematic Reviews and Meta-Analyses” was revised and has a novelty and I recommend it for publication after consideration of the following comments.

Title: If you can rewrite and make it more interesting for readers.

Abstract:

• Methodological Rigor: Given the high percentage (76.6%) that were rated as moderate, low, or critically low, what measures were taken to guarantee the methodological quality of the systematic reviews and meta-analyses included in the umbrella review?
• Consistency of Results: Considering the conflicting results regarding probiotic treatments for anxiety symptoms, what particular elements might be responsible for the variation in results among various studies?
• Evidence Gaps: How might future studies fill in the gaps in the literature regarding the effectiveness of prebiotics and synbiotics for treating anxiety and depression symptoms?
• Geographical Analysis: How do the studies' geographical origins affect the generalizability and applicability of the conclusions about psychobiotics for anxiety and depression symptoms?

Introduction:

• Definition Clarity: In the context of mental health interventions, how well-defined are the terms "psychobiotics," "probiotics," and "prebiotics"? How might different definitions affect how efficacy is interpreted across studies?
• Methodological Rigor: What particular methodological flaws in the psychobiotics systematic reviews and meta-analyses could affect the validity of their findings, and how might these be fixed in subsequent studies?
• Outcome Measures: How much of the observed heterogeneity in results across studies can be attributed to differences in clinical scales, self-reported symptoms, and biomarkers used to evaluate efficacy?
• Future Research Directions: In order to improve our comprehension of the effectiveness and applicability of psychobiotics, which particular gaps in the literature on the subject should be given priority for examination in upcoming clinical trials?
• You can refer to the following article in anywhere in your manuscript’s introduction:
• Nutrients2023, 15(17), 3715;https://doi.org/10.3390/nu15173715

Materials and methods

• Protocol Adherence: What particular changes could improve the methodological rigor of the systematic review process, and how well does the study follow PRISMA guidelines and the AMSTAR 2 tool requirements?
• Justification for Inclusion Criteria: What effects would it have on the generalizability of the results on psychobiotic interventions if studies involving other mental illnesses and unique circumstances, such as pregnancy, were excluded from the systematic review?
• Risk-of-Bias Evaluation: What steps can be taken to address identified weaknesses, and how does the risk-of-bias assessment using the AMSTAR 2 framework affect the interpretation of the results from the included systematic reviews?
• Statistical Analysis Consistency: How reliable are the statistical techniques used, specifically the application of random-effects models and sensitivity analyses, given the high variability anticipated in the pooled analyses?

“Results:

• Effect Size Interpretation: In order to improve the consistency and dependability of conclusions made regarding the effectiveness of probiotics, how should the notable variation in effect sizes among studies be addressed, especially with regard to the impact of outlier studies like Moshfeghinia et al.
• Population Stratification: When analyzing psychobiotic interventions, what methods can be used to guarantee that studies appropriately stratify populations by variables like age, gender, and baseline mental health status to minimize contamination of treatment effects?
• Methodological Restrictions: In light of the systematic reviews and meta-analyses' frequent methodological flaws, what particular changes should authors make to meet AMSTAR 2 requirements and strengthen the validity of their conclusions?

Discussion:

• Heterogeneity Assessment: How might the interpretation of the overall effectiveness of psychobiotic interventions in treating anxiety and depressive symptoms be affected by the significant statistical heterogeneity found in the meta-analyses?
• Methodological Improvements: What particular tactics might be used in upcoming systematic reviews to address the recurring methodological flaws found in the evaluation of recent literature, such as inadequate reporting of funding sources and lack of protocol registration?
• Population Specificity: How should future research prioritize and develop interventions to more accurately evaluate the effectiveness of psychobiotics in these underrepresented populations, given the gaps in studies involving older adults and those with particular underlying conditions?
• Discussion text must grammar improve and in some cases it is very weak and maybe there is no discussion at all.

Conclusions:

Conclusion is very general, try to make it more scientific, comprehensive and concise in detail, especially.

References: It is OK.

 

Author Response

REVIEW PHARMACEUTICALS

REVIEWER 2

Comment 1: This article "Psychobiotics for Depression and Anxiety: An Umbrella Review and Critical Appraisal of Systematic Reviews and Meta-Analyses” was revised and has a novelty and I recommend it for publication after consideration of the following comments.

REPLY: We appreciate your dedication to this important review. With you, our paper becomes stronger!

Title: If you can rewrite and make it more interesting for readers.

REPLY: Consider done!

Title modified:

Attacking the Gut–Brain Axis with Psychobiotics: An Umbrella Review of Depressive and Anxiety Symptoms

Abstract:

Comment 2: Methodological Rigor: Given the high percentage (76.6%) that were rated as moderate, low, or critically low, what measures were taken to guarantee the methodological quality of the systematic reviews and meta-analyses included in the umbrella review?

REPLY: Dear editor, we would very much like to improve the quality of the articles or intentionally select those with higher methodological quality. However, most of the time we "play" with the unexpected when conducting our search. And according to Cochrane's criteria, we cannot deliberately eliminate lower-quality studies. We kindly apologize, as such an action would not be feasible.

Comment 3: Consistency of Results: Considering the conflicting results regarding probiotic treatments for anxiety symptoms, what particular elements might be responsible for the variation in results among various studies?

REPLY: The inconsistency observed across studies evaluating psychobiotic effects on anxiety symptoms likely reflects a combination of methodological and conceptual factors. Unlike depressive symptoms, anxiety encompasses a heterogeneous set of constructs, including state anxiety, trait anxiety, stress-related symptoms, and clinically diagnosed anxiety disorders, which are often assessed interchangeably across trials. This conceptual variability is further compounded by the use of diverse psychometric instruments with differing sensitivities, thresholds, and clinical interpretability, increasing the likelihood of scale-driven heterogeneity. In addition, many trials assessing anxiety outcomes have enrolled heterogeneous or subclinical populations, frequently including healthy individuals or participants with mild stress-related symptoms rather than well-defined anxiety disorders.

We have created new sections to explain to the reader about future directions that help in understanding these nuances, whether for the analysis of depressive and/or anxiolytic symptoms. See below.

Text added (3.5 Future directions):

“Future research should move beyond highly generalized analyses that pool heterogeneous populations, symptom profiles, and probiotic formulations, as such approaches may obscure strain-specific effects and attenuate clinically meaningful signals. Accumulating evidence suggests that probiotic effects on mental health outcomes are not uniform [45,61,66,79,80], but instead depend on clearly defined strain characteristics, and well-characterized participant profiles. Accordingly, future trials should prioritize more phenotypically and biologically specific study designs, including the evaluation of single strains or well-justified.

In parallel, greater emphasis should be placed on recruiting clinically defined populations, such as individuals with established diagnoses of depressive or anxiety disorders, or on clearly stratifying participants by symptom severity and baseline microbiota characteristics. It is worth highlighting that, only a limited number of existing studies have adopted such stringent designs, focusing on specific strains, or clinically diagnosed populations (observed in more recent studies) [45,61,66,79,80], yet these studies tend to provide clearer and more interpretable effect estimates.

Finally, despite growing interest in microbiota-targeted interventions, the current evidence base remains heavily skewed toward prebiotic formulations, limiting the ability to differentiate the effects on depressive and anxiety symptoms. Future progress in this field will require methodologically rigorous primary studies specifically designed to evaluate prebiotics and synbiotics, with clearly defined compositions”.

Comment 4: Evidence Gaps: How might future studies fill in the gaps in the literature regarding the effectiveness of prebiotics and synbiotics for treating anxiety and depression symptoms?

REPLY: This question was already addressed in the previous resolution where we added:

Text added (3.5 Future directions):

“…despite growing interest in microbiota-targeted interventions, the current evidence base remains heavily skewed toward prebiotic formulations, limiting the ability to differentiate the effects on depressive and anxiety symptoms. Future progress in this field will require methodologically rigorous primary studies specifically designed to evaluate prebiotics and synbiotics, with clearly defined compositions”.

Comment 5: Geographical Analysis: How do the studies' geographical origins affect the generalizability and applicability of the conclusions about psychobiotics for anxiety and depression symptoms?

REPLY: The geographical distribution of the included studies has important implications for the generalizability and applicability of the present findings. Most trials evaluating psychobiotic interventions for depressive and anxiety symptoms have been conducted in specific regions, predominantly Europe, East Asia, and parts of the Middle East, with limited representation from low- and middle-income countries and other global populations. Given that gut microbiota composition is strongly influenced by geography-related factors such as habitual diet, lifestyle, environmental exposures, and cultural practices, baseline microbial profiles may differ substantially among regions, potentially modulating responsiveness to psychobiotic interventions.

However, our intention, as an Umbrella review, was not to address this issue, since it depends especially on primary studies (RCTs), which can indeed provide answers regarding behavior and outcomes related to geography. Our intention was merely to report where the science is being done, proposing a methodological rather than explanatory observation that would support the understanding of the outcomes, such as a scoping review.

Introduction:

Comment 6: Definition Clarity: In the context of mental health interventions, how well-defined are the terms "psychobiotics," "probiotics," and "prebiotics"? How might different definitions affect how efficacy is interpreted across studies?

REPLY: Within the context of microbiota-targeted mental health interventions, the terms probiotics, prebiotics, and psychobiotics are conceptually well defined in the scientific literature. Probiotics are generally defined as live microorganisms that, when administered in adequate amounts, confer a health benefit on the host; prebiotics as substrates selectively utilized by host microorganisms that confer a health benefit; and psychobiotics as a class of probiotics or prebiotics that exert mental health benefits through microbiota–gut–brain axis mechanisms. These definitions are supported by international consensus statements and have been widely adopted among systematic reviews and meta-analyses.

Hill et al., (2014). Expert consensus document: The International Scientific Association for Probiotics and Prebiotics consensus statement on the scope and appropriate use of the term probiotic. Nature Reviews Gastroenterology & Hepatology, 11(8), 506–514. https://doi.org/10.1038/nrgastro.2014.66

“An expert panel was convened in October 2013 by the International Scientific Association for Probiotics and Prebiotics (ISAPP) to discuss the field of probiotics. It is now 13 years since the definition of probiotics and 12 years after guidelines were published for regulators, scientists and industry by the Food and Agriculture Organization of the United Nations and the WHO (FAO/WHO). The FAO/WHO definition of a probiotic “live microorganisms which when administered in adequate amounts confer a health benefit on the host” was reinforced as relevant and sufficiently accommodating for current and anticipated applications. However, inconsistencies between the FAO/WHO Expert Consultation Report and the FAO/WHO Guidelines were clarified to take into account advances in science and applications. A more precise use of the term 'probiotic' will be useful to guide clinicians and consumers in differentiating the diverse products on the market. This document represents the conclusions of the ISAPP consensus meeting on the appropriate use and scope of the term probiotic”.

For the reader's better understanding, we have added this reference in the introduction.

Text added (introduction):

“Psychobiotics include probiotics (beneficial microbes), prebiotics (fibers that feed them), and synbiotics (their combined, mutually supportive pair) [12]”.

Comment 7: Methodological Rigor: What particular methodological flaws in the psychobiotics systematic reviews and meta-analyses could affect the validity of their findings, and how might these be fixed in subsequent studies?

REPLY: That's an excellent question, and one of our secondary purposes. I believe this has been answered in several points of the article (with the new updates). See below:

First point: “The assessment of heterogeneity among the primary studies within each me-ta-analysis revealed substantial variability, with I² values ranging from minimal (0.8%) to extremely high (99.7%). According to the interpretation framework recommended by the Cochrane Handbook, only five meta-analyses (16.1%) demonstrated low heterogeneity (I² ≤ 25%), indicating a high degree of consistency across their primary studies [15,30,39,40], whose findings are less influenced by methodological or clinical diversity”. “…a considerable number of meta-analyses (twelve studies; 38.7%) exhibited very high heterogeneity (I² > 75%), indicating pronounced disagreement among the included primary studies. This group includes [14,31-33,36,46,51], values exceeding 75% suggesting that methodological, clinical, or statistical factors strongly influence the consistency of aggregated effects”.

This condition hinders and influences the generalization of results. In our study, we attempted to homogenize the selection and inclusion of MDS, stratifying whenever possible by specific subgroups (such as patients actually diagnosed). We added a section to the results explaining and making transparent our actions to reduce variability. Basically, in almost all cases, we excluded healthy patients or those with systemic problems.

2.3.2 Synthesis Strategy and Selection of Effect Estimates

The synthesis of results among the included meta-analyses required careful con-sideration of how SMDs were selected and extracted, given the substantial heterogeneity in study populations, symptom severity, and outcome assessment instruments. When available, subgroup-specific SMDs were preferentially extracted for analyses targeting depressive or anxiety symptoms, particularly when meta-analyses reported stratified estimates based on clinically diagnosed populations, symptom severity, or specific psychometric instruments. This approach aimed to enhance clinical interpret-ability and reduce dilution of effects that could arise from pooling heterogeneous populations, such as healthy individuals and participants without clinically relevant symptoms.

In contrast, global SMDs were retained when subgroup stratification was either not clearly defined, not methodologically justified, or when the study population was already homogeneous with respect to depressive or anxiety disorders. In several cases, global estimates were also preferred when further stratification would have substantially reduced the number of contributing studies, thereby compromising statistical stability and interpretability. For meta-analyses reporting multiple subgroup estimates based on different psychometric scales, the SMD derived from the most prevalent instrument was selected to ensure consistency and comparability among analyses.

This decision-making process was applied systematically and transparently across all included meta-analyses, with detailed justifications provided in Supplementary Table S4.

We have updated the discussion by adding a section on how to proceed with future studies.

“While the adoption of a single universal outcome measure could, in theory, en-hance comparability, such an approach is unlikely to be feasible or desirable given differences in study objectives, populations, symptom severity, and clinical versus sub-clinical contexts. From a practical and methodological standpoint, a more realistic solution lies in the standardization of analytical frameworks rather than in the exclusive use of a single instrument. In this review, comparability was addressed by prioritizing effect estimates derived from clinically relevant subgroups when available and by selecting SMDs based on instruments with stronger psychometric support, and greater prevalence among trials. Global SMDs were retained when populations were homogeneous or when further stratification would compromise statistical robustness.

Future research should move beyond highly generalized analyses that pool heterogeneous populations, symptom profiles, and probiotic formulations, as such approaches may obscure strain-specific effects and attenuate clinically meaningful signals. Accumulating evidence suggests that probiotic effects on mental health outcomes are not uniform [45,61,66,79,80], but instead depend on clearly defined strain characteristics, and well-characterized participant profiles. Accordingly, future trials should prioritize more phenotypically and biologically specific study designs, including the evaluation of single strains or well-justified.

In parallel, greater emphasis should be placed on recruiting clinically defined populations, such as individuals with established diagnoses of depressive or anxiety disorders, or on clearly stratifying participants by symptom severity and baseline microbiota characteristics. It is worth highlighting that, only a limited number of existing studies have adopted such stringent designs, focusing on specific strains, or clinically diagnosed populations (observed in more recent studies) [45,61,66,79,80], yet these studies tend to provide clearer and more interpretable effect estimates.

Finally, despite growing interest in microbiota-targeted interventions, the current evidence base remains heavily skewed toward prebiotic formulations, limiting the ability to differentiate the effects on depressive and anxiety symptoms. Future progress in this field will require methodologically rigorous primary studies specifically designed to evaluate prebiotics and synbiotics, with clearly defined compositions”.

Comment 8: Outcome Measures: How much of the observed heterogeneity in results across studies can be attributed to differences in clinical scales, self-reported symptoms, and biomarkers used to evaluate efficacy?

REPLY: Interesting question. But I don't know if I really have the potential to answer it. Dear reviewer, we believe that the central point of heterogeneity is not actually the psychometric scales. but derived from the composition of existing probiotics and prebiotics. Consider that each primary study used a different composition and dose; therefore, in our view, that is the core of the heterogeneity. This is a problem that some studies have attempted to address by using standardized strains. In our study, we did not investigate specific strains, which is a limitation. You can see in Table 1 the wide variation of strains used. To improve this issue, we can only add a mention of the subject in the limitations section. I hope our efforts will be sufficient.

Text added (Future directions and Limitations):

“A major source of heterogeneity among the included studies likely arises from the substantial variability in probiotic and prebiotic formulations rather than from differences in outcome assessment alone. Across primary trials, interventions differed markedly with respect to microbial strains, combinations of species, dosages, intervention duration, and delivery formats. Such compositional heterogeneity may exert a stronger influence on observed effects than the choice of psychometric instruments used to assess depressive or anxiety symptoms.

Although some recent studies have attempted to address this issue by employing standardized or well-characterized strains, the majority of available trials continue to use heterogeneous formulations, limiting the ability to attribute effects to specific microbial profiles. The present review did not perform strain-specific or dose–response analyses, which represents an important limitation and reflects the current constraints of the evidence base. As illustrated in Table 1, the wide diversity of probiotic compositions precludes firm conclusions regarding the relative efficacy of individual strains or combinations. Future research should prioritize standardized formulations and strain-specific designs to reduce heterogeneity and enable more precise mechanistic and clinical inferences”.

Comment 9: Future Research Directions: In order to improve our comprehension of the effectiveness and applicability of psychobiotics, which particular gaps in the literature on the subject should be given priority for examination in upcoming clinical trials?

REPLY: Estimated reviewer, i believe we have already addressed this issue in a specific section of the discussion. Furthermore, we can see the resolution of this comment in comment 7.

Comment 10: You can refer to the following article in anywhere in your manuscript’s introduction: Nutrients2023, 15(17), 3715; https://doi.org/10.3390/nu15173715

REPLY: Consider done!

Materials and methods

Comment 11: Protocol Adherence: What particular changes could improve the methodological rigor of the systematic review process, and how well does the study follow PRISMA guidelines and the AMSTAR 2 tool requirements?

REPLY: Estimated reviewer, this is a crucial issue in this type of paper. Therefore, we have dedicated ourselves to exploring an exclusive session in the discussion on this subject. Please note:

3.4 Analysis of methodological quality

“The methodological assessment conducted in this umbrella review revealed a consistent pattern of substantial weaknesses among most of the included systematic reviews and meta-analyses, in line with the limitations previously identified by Morán et al., [23]. Using the AMSTAR 2 instrument, we found that the majority of reviews were classified as low or critically low in overall confidence. This distribution closely mirrors the findings of Morán et al., [23] who also reported a predominance of low-confidence ratings, highlighting Hofmeister et al., [39] as the only review meeting high-quality standards.

One of the most frequent shortcomings was the absence of a preregistered protocol (Item 2), which compromises transparency and increases the risk of post hoc decision-making. Recurrent issues also included inadequate reporting of funding sources for the included studies (Item 10) and the absence or insufficiency of publication bias assessments (Item 15). Additionally, as noted by Morán et al., [23] only a portion of the reviews appropriately incorporated risk-of-bias evaluations into the interpretation of their findings (Item 13). Even when such assessments were performed, their implications were seldom discussed with the necessary depth.

Another recurring problem was the high level of statistical heterogeneity, with I² values frequently exceeding 70%. Although some reviews conducted subgroup or sensitivity analyses to mitigate this issue, substantial heterogeneity persisted. Key contributors included notable differences in population characteristics (healthy individuals vs. clinically diagnosed patients), baseline symptom severity, intervention dosage and duration, and variability in psychometric instruments used to assess outcomes.

Despite the rapid expansion of the psychobiotics field, the methodological quality of existing systematic reviews remains inconsistent and often inadequate, thereby limiting the reliability and applicability of pooled effect estimates”.

Comment 12: Justification for Inclusion Criteria: What effects would it have on the generalizability of the results on psychobiotic interventions if studies involving other mental illnesses and unique circumstances, such as pregnancy, were excluded from the systematic review?

REPLY: The inclusion criteria adopted in this review were intentionally designed to balance generalizability with methodological coherence and clinical interpretability. Although psychobiotic interventions have been explored among a range of mental health conditions, the etiology and underlying biological mechanisms of many psychiatric disorders differ substantially. Including heterogeneous diagnostic categories beyond mood and anxiety disorders, such as neurodevelopmental, psychotic, or trauma-related conditions, would likely have introduced excessive biological and clinical heterogeneity, thereby obscuring disorder-specific effects and limiting the interpretability of pooled estimates. Restricting the analysis to depressive and anxiety-related outcomes therefore allowed for a more mechanistically coherent synthesis.

Similarly, studies conducted in unique physiological contexts, such as pregnancy, were excluded because pregnancy represents a distinct biological state characterized by profound hormonal, immunological, metabolic, and microbiota-related changes. Mental health symptoms during pregnancy, including perinatal depression and anxiety, are influenced by mechanisms that differ from those operating in non-pregnant populations. Including such studies would have introduced structural confounding that could not be adequately addressed without dedicated subgroup analyses. While these restrictions may limit the direct applicability of the findings to all populations, they enhance internal validity and support more precise conclusions regarding psychobiotic effects in adult populations with mood and anxiety disorders.

Comment 13: Risk-of-Bias Evaluation: What steps can be taken to address identified weaknesses, and how does the risk-of-bias assessment using the AMSTAR 2 framework affect the interpretation of the results from the included systematic reviews?

REPLY: The risk-of-bias assessment conducted using the AMSTAR 2 framework plays a critical role in contextualizing the interpretation of the findings derived from the included systematic reviews. Importantly, the methodological weaknesses identified, such as incomplete reporting, heterogeneity in primary study design, and limited consideration of bias in original trials, largely reflect inherent limitations of the available evidence base rather than deficiencies of the present synthesis.

For example, when positive effects of psychobiotic interventions are observed in reviews classified as having critically low confidence according to AMSTAR 2 criteria, despite a positive result, such findings should be interpreted with caution. In these cases, the presence of elevated risk of bias precludes definitive conclusions regarding efficacy, and the results should be viewed as indicative rather than confirmatory. Accordingly, the conclusions of this review were deliberately framed to reflect the strength and consistency of the underlying evidence, emphasizing uncertainty where methodological limitations predominate. This approach aligns with the intended purpose of the AMSTAR 2 tool, which is not to negate observed effects, but to guide appropriate confidence in the conclusions drawn from them.

Please note our conclusion:

“This umbrella review synthesized and critically appraised the highest level of evidence available on the effects of psychobiotic interventions on symptoms of depression and anxiety in patients with symptoms or a diagnosis consistent with a mood disorder. Despite the substantial growth of meta-analyses published in recent years, the overall certainty of the evidence remains constrained by methodological limitations, inconsistent reporting, and substantial heterogeneity among primary trials and reviews…” “Despite that, most reviews exhibited low to moderate confidence according to AMSTAR-2, primarily due to critical weaknesses in protocol registration, search completeness, selection bias and risk-of-bias integration…”

Comment 14: Statistical Analysis Consistency: How reliable are the statistical techniques used, specifically the application of random-effects models and sensitivity analyses, given the high variability anticipated in the pooled analyses?

REPLY: Given the substantial clinical and methodological heterogeneity anticipated across studies, the use of random-effects models was considered both appropriate and necessary. Unlike fixed-effect approaches, random-effects models explicitly account for between-study variability and assume that individual studies estimate different, yet related, true effects. This framework is particularly suitable for psychobiotic research, where variability in intervention composition, dosing, population characteristics, and outcome measures is expected.

Sensitivity analyses were conducted to assess the stability of pooled estimates and to evaluate whether overall conclusions were disproportionately influenced by individual studies or specific analytical choices. While such analyses cannot eliminate inherent heterogeneity, they provide important information regarding the consistency of observed effects. Accordingly, the statistical techniques applied in this review are reliable within the context of heterogeneous evidence, provided that pooled estimates are interpreted as average effects with associated uncertainty rather than as precise or universally generalizable effect sizes.

Thank you for your question!

Results:

Comment 15: Effect Size Interpretation: In order to improve the consistency and dependability of conclusions made regarding the effectiveness of probiotics, how should the notable variation in effect sizes among studies be addressed, especially with regard to the impact of outlier studies like Moshfeghinia et al.

REPLY: The notable variability in effect sizes observed across studies underscores the importance of cautious interpretation when synthesizing evidence on probiotic efficacy. Differences in intervention composition, dosage, study populations, and outcome measures contribute to dispersion in effect estimates and increase susceptibility to the influence of individual studies with comparatively large effects. To address this issue, sensitivity analyses were conducted to evaluate the pooled estimates in the presence of potential outliers.

Specifically, the exclusion of the study by Moshfeghinia et al., 2025 which reported a comparatively large effect size, resulted in attenuation of the pooled estimate and loss of statistical significance; however, the overall direction of the effect remained unchanged. This finding indicates that while individual outlier studies can meaningfully influence the magnitude and precision of pooled effects, the broader pattern of results does not rely on a single study.

Therefore, even excluding some of the studies, at least for this analysis, the conclusions will be the same.

Comment 16: Population Stratification: When analyzing psychobiotic interventions, what methods can be used to guarantee that studies appropriately stratify populations by variables like age, gender, and baseline mental health status to minimize contamination of treatment effects?

REPLY: Population stratification by age, sex, and baseline mental health status is largely determined at the level of primary study design. In this review, stratification was applied where data were available; however, incomplete reporting in many trials precluded systematic adjustment, representing an inherent limitation of the current evidence base rather than of the review methodology.

I'll give you an example. When we were extracting SMDs from meta-analyses, we considered collecting more specific information, such as sex or effects related to intervention time. However, most papers, despite performing subgroup analyses, present general analyses containing contamination from different populations; that is, when analyzing the effects on intervention time (<8 weeks or >8 weeks), this condition was often not stratified in relation to patients with mood disorders. Therefore, we observed that this would excessively complicate our work, in addition to deviating from the initially planned protocol. This condition is very poorly observed by the scientific community. Thus, we honestly established our purpose and followed the main extraction (which was already extremely difficult).

Comment 17: Methodological Restrictions: In light of the systematic reviews and meta-analyses' frequent methodological flaws, what particular changes should authors make to meet AMSTAR 2 requirements and strengthen the validity of their conclusions?

REPLY: The frequent methodological limitations identified among the included systematic reviews and meta-analyses reflect broader challenges within the current psychobiotics literature rather than deficiencies unique to individual studies. According to the AMSTAR 2 framework, critical weaknesses such as incomplete reporting, lack of protocol registration, limited assessment of risk of bias in primary trials, and insufficient consideration of heterogeneity reduce the overall confidence that can be placed in pooled estimates. Importantly, these limitations do not negate observed effects but constrain the strength and certainty of the conclusions that can be drawn.

In the context of the present review, no additional post hoc methodological adjustments could meaningfully correct these structural limitations without introducing further bias. Accordingly, the primary methodological response was to explicitly acknowledge these weaknesses and to frame conclusions with appropriate caution, avoiding definitive claims where the underlying evidence lacks robustness. Future systematic reviews seeking to meet AMSTAR 2 requirements more fully should prioritize prospective protocol registration, transparent outcome prioritization, comprehensive risk-of-bias integration, and predefined analytical strategies. Ultimately, strengthening the validity of conclusions in this field will depend on improvements in both primary study design and the methodological rigor of future evidence syntheses.

To better assist you, given your repeated concern about this issue, we will add this important question to our list of limitations and results. Thanks.

Text added (discussion - 3.4 Analysis of methodological quality):

“So, as indicated by the AMSTAR 2 assessment, several included studies exhibited critical or high-risk methodological weaknesses, which constrain the confidence that can be placed in pooled estimates. These limitations are largely inherent to the existing evidence base and could not be corrected retrospectively within the present synthesis. Consequently, while observed effects may indicate potential benefits of psychobiotic interventions, the findings should be interpreted with caution and should not be considered definitive evidence of efficacy”.

Discussion:

Comment 18: Heterogeneity Assessment: How might the interpretation of the overall effectiveness of psychobiotic interventions in treating anxiety and depressive symptoms be affected by the significant statistical heterogeneity found in the meta-analyses?

REPLY: Dear reviewer, I believe this question has already been addressed previously. Thank you for pointing it out (see Comment 8).

Comment 19: Methodological Improvements: What particular tactics might be used in upcoming systematic reviews to address the recurring methodological flaws found in the evaluation of recent literature, such as inadequate reporting of funding sources and lack of protocol registration?

REPLY: The answer to this question is very simple: analyze the checklists for each type of study design (https://www.equator-network.org/). For each type of design, we have a way to report and maintain transparency in our study. It's no different when we use the PRISMA checklist to conduct systematic reviews. Additionally, my group and I conducted a real-world test, asking other researchers (before publication) to complete the checklist point by point. Based on this, the chance of incurring a primary error, such as protocol documentation, is zero.

Comment 20: Population Specificity: How should future research prioritize and develop interventions to more accurately evaluate the effectiveness of psychobiotics in these underrepresented populations, given the gaps in studies involving older adults and those with particular underlying conditions?

REPLY: Future research should adopt a more population-specific approach to accurately evaluate the effectiveness of psychobiotic interventions in underrepresented groups, particularly older adults and segmented to sexes. We even made this suggestion in our discussion to encourage primary studies with this population. Note:

“It is important to highlight that specific population subgroups remain markedly underexplored in current systematic reviews with meta-analyses. Older adults, particularly those over 60 years of age, frequently exhibit age-related alterations in gut microbiota composition, compounded by multimorbidity and polypharmacy. These fac-tors may influence their responsiveness to probiotic supplementation; however, trials conducted in elderly populations are scarce. In the limited evidence available, Huang et al., [28] reported no significant effect of probiotics among participants older than 65 years (SMD = −0.18 [95%CI: −0.47 to 0.11]; p = 0.22), in contrast to younger adults, likely attributable to the small number of older individuals included. However, similar findings were observed in Zhu et al., [54] and Cheng et al., [52], which also failed to demonstrate significant probiotic effects compared with placebo or control groups (SMD = −0.13 [95%CI: −0.31 to −0.04]; p > 0.05; and SMD = −0.22 [95%CI: −0.57 to 0.13]; p = 0.12, respectively). These results highlight a persistent evidence gap concerning psychobiotic efficacy in older adults, underscoring the need for targeted, age-specific clinical research in this demographic”.

These populations present distinct biological, metabolic, and microbiota-related characteristics, such as age-related changes in immune function, polypharmacy, comorbidities, and altered gut microbiota composition, that may meaningfully influence responsiveness to microbiota-targeted interventions.

Comment 21: Discussion text must grammar improve and in some cases it is very weak and maybe there is no discussion at all.

REPLY: Thank you for pointing this out. Despite our efforts, there are always small errors that we don't notice properly. Please consider the review that has been done.

Conclusions:

Comment 22: Conclusion is very general, try to make it more scientific, comprehensive and concise in detail, especially.

REPLY: We will attempt to improve our conclusion by primarily highlighting our findings. Consider it done. See below:

“This umbrella review demonstrates that psychobiotic interventions are more consistently associated with reductions in depressive symptoms than in anxiety symptoms, for which findings remain heterogeneous and less stable. Probiotics emerged as the most frequently investigated intervention, demonstrating modest but generally favorable effects on depressive symptoms. Evidence for prebiotics and, particularly, synbiotics was limited in both contexts, with only a small number of reviews providing extractable estimates, and several failing to isolate intervention types or to stratify different formulations.

Despite the substantial growth in the number of meta-analyses published, the overall certainty of the evidence remains constrained by methodological limitations, inconsistent reporting, and substantial heterogeneity among reviews. Consistent with this, most reviews demonstrated low to moderate confidence according to the AMSTAR 2 framework, primarily due to critical weaknesses in protocol registration, search completeness, selection bias, and integration of risk-of-bias assessments.

Citation-based mapping and timeline analyses further revealed an uneven accumulation of evidence, characterized by clusters of redundant reviews and limited conceptual connectivity across the field”.

Comment 23: References: It is OK.

REPLY: Thanks!!!

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Dear authors,

Many thanks for your extensive answers to my comments and additions to the manuscript. I am satisfied with this.

I have some last remarks:

1. Please check the additions on grammar, i.e. in section 3.1: "preclinical e clinical" ... "preclinical and clinical";
2. In section 3.1 I suggest to somewhat attenuate the statements on the effects of probiotics as anti-depressants. I.e. "provide support" instead of "provide robust support", and "and possibly modulate" instead of "and modulate". Indeed there is evidence, but we are still far from substituting SSRI treatments with probiotic treatments, and the exact functional mechanisms are to be better investigated.
3. "Revise the search strategy specifications of Table S1: i.e. when I search in Pubmed on (psychobiotics OR probiotics OR prebiotics OR symbiotics) AND (depressive OR depressive disorder OR anxiety) AND Meta-Analysis", I get 98 results but when I click on the link below this I get around 600 results. For transparency the authors can best specify the keywords for each platform, the search date and the total number of hits.
4. For full transparency and completeness include in Table S1 a list of all the papers that were included in the review (including their source(s) and DOIs).
5. Add the DOIs of the different papers to Table S2.

Author Response

Rebuttal Letter

Comment 1: Please check the additions on grammar, i.e. in section 3.1: "preclinical e clinical" ... "preclinical and clinical";

REPLY: Sorry for our mistake. The term has been corrected.

Comment 2: In section 3.1 I suggest to somewhat attenuate the statements on the effects of probiotics as anti-depressants. I.e. "provide support" instead of "provide robust support", and "and possibly modulate" instead of "and modulate". Indeed, there is evidence, but we are still far from substituting SSRI treatments with probiotic treatments, and the exact functional mechanisms are to be better investigated.

REPLY: You are absolutely right. The terms have been adjusted and highlighted in yellow.

Comment 3: "Revise the search strategy specifications of Table S1: i.e. when I search in PubMed on (psychobiotics OR probiotics OR prebiotics OR symbiotics) AND (depressive OR depressive disorder OR anxiety) AND Meta-Analysis", I get 98 results but when I click on the link below this I get around 600 results. For transparency the authors can best specify the keywords for each platform, the search date and the total number of hits.

REPLY: Dear reviewer, we don't understand what's happening. When we paste the link below into our browser, we see these results (See the screenshot below):

https://pubmed.ncbi.nlm.nih.gov/?term=%28psychobiotics+OR+probiotics+OR+prebiotics+OR+symbiotics%29+AND+%28depressive+OR+depressive+disorder+OR+anxiety%29+AND+Meta-Analysis&filter=pubt.meta-analysis&filter=pubt.systematicreview

Please also note that the reported search engine appears in these terms: (psychobiotics OR probiotics OR prebiotics OR symbiotics) AND (depressive OR depressive disorder OR anxiety) AND Meta-Analysis.

Comment 4: For full transparency and completeness include in Table S1 a list of all the papers that were included in the review (including their source(s) and DOIs).

REPLY: Dear reviewer, to better present the paper, we have created a Table S2 containing only the requested information. For proper verification, please see our new version of the supplementary material. Thank you for your suggestion.

Comment 5: Add the DOIs of the different papers to Table S2.

REPLY: First, Table 2 was changed to Table 3. Second, the DOI registration number was duly added. See our supplementary material. All references to supplementary tables have been adjusted in the manuscript and are marked in yellow.

Thanks!!!

Author Response File: Author Response.pdf