Signal Detection of Depression and Suicidality Associated with Finasteride and Dutasteride: Updated Pharmacovigilance Evidence and Recommendations for Comprehensive Psychiatric Assessment

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Your manuscript addresses an important and timely pharmacovigilance issue regarding potential neuropsychiatric adverse events associated with 5α-reductase inhibitors (5-ARIs), particularly finasteride and dutasteride. The topic is clinically relevant and aligned with ongoing regulatory scrutiny. The comparative disproportionality analysis between the two agents is of interest, and the proposed structured psychiatric monitoring approach represents a practical contribution.

However, substantial revisions are required to improve methodological transparency, internal consistency, statistical clarity, and precision of language. In its current form, the manuscript requires major revision before it can be considered for publication.

Below are detailed, line-specific, and section-specific comments.

Line 23-31: Finasteride selectively inhibits type-2 5-androgen receptors. Correction: Replace with “type-2 5α-reductase isoenzyme.”

Line 31: Causality remains uncertain. -> Causality cannot be inferred from spontaneous reporting data.

Line 33-36: Lack of details (Period of FAERS data extraction, Which version of MedDra, It was EFDA duplication removed?) 

Line 43-46: “Baseline psychiatric assessment and regular monitoring… are recommended.” -> Considering the signal-detection architecture, the recommendation phrasing is robust.

Line 63: “side effects dealing with the sexual domain”-> “sexual adverse effects.”

Lines 71-73: "possible association of the 5-ARI intake and the onset…” -> "possible association between 5-ARI use and the onset…”

Lines 74-76: “hence exerting antidepressant- and neuroprotective-effects.” ->Remove hyphenation. Improve clarity: “which modulate GABAergic neurotransmission and exert antidepressant and neuroprotective effects.”

Lines: 86-88: “Cleaned case listings were used." Please specify: which quarterly files were uesd ()2004Q1-2025Q2), Duplication method.

Also here I'd like to clarify if there were concomitant drugs exculded

Lines 90-91: Specify the libraries used and offer proof of the availability of the code.

Lines 93-97: It must clearly reference category describe whether the comparison was just between finasteride and dutasteride or vs. entire FAERS database.

Line 107. It think that it's a mistake here because CIs are not negative libe here'When ROR >1but CIs were negative values.'

Line 115. Remove or replace: 'All codes were tested and met all parameters to a 100% accuracy level' -because it is not scientifically appropiate!

Lines 127-130: Please report also the percentages, not only counts (4132 vs 202)

Lines 136-141: Please may you clarify what is 'disabled'? It is part of FAD outcome field?

Table 2: it is difficult to interpret; there are too many categories,  the age is inconsistently formatted, and some country codes are unclear. Try to simplify it and add more to supplementart materials.

Lines 151-156 there are some inconstant values regarding in lines 182-184 about RORs (Suicide and Depression) - differenteces. Are these external study values? 

Line 165: Please notice that the absolute novelty should be softened -> “To our knowledge, few studies have comparatively examined…”

Lines 169-178: Are these FAERS study or a cohort study? 

Lines 245-258: I think that you may add: lack of dose information, treatment duration, indicate bias BPH vs alopecia, psychiatric history/??

Bibliography: Justify the use of refs 7 and 11; add a DOI to all the references, and check to see if all the 2025 references are in press or already published.

 

Suggestion of the statistics: Stratify y age: <45 years vs. >=45 years and use timetrend ROR every 5 years interval

Comments on the Quality of English Language

The paper effectively communicates its scientific point and is generally comprehensible. However, to enhance accuracy, fluency, and a formal scientific tone, little language modification is advised.

Particular problems include:

• sporadic errors in grammar (e.g., verb agreement, article use);
• excessive hyphenation;
• the phrase "side effects dealing with the sexual domain" is one example of such bad wording;
• little typographical mistakes;
• Line 71: “neurolopsychiatric” -> neuropsychiatric
• Line 245: “interpretated” -> interpreted
•  
• inconsistent use of technical terms.

Thorough expert language editing would improve the text's readability and ensure compliance with journal requirements.

Author Response

Ms. Sana Zhang
Assistant Editor
E-Mail：sana.zhang@mdpi.com

 

 

Dear Editor and Reviewers,

We would like to sincerely thank you for the time and effort you dedicated to reviewing our manuscript entitled “Signal Detection of Depression and Suicidality Associated with Finasteride and Dutasteride: Updated Pharmacovigilance Evidence and Recommendations for Comprehensive Psychiatric Assessment.” We are grateful for the insightful comments and constructive suggestions, which have significantly improved the quality and clarity of our work.

We have carefully considered all comments and revised the manuscript accordingly. Below, we provide a detailed, point-by-point response to each reviewer’s comments. All changes in the revised manuscript have been highlighted for ease of reference.

In addition to addressing the reviewers’ comments, we have made several minor revisions throughout the manuscript to improve clarity, readability, and consistency. These include grammatical corrections, improved figure/table descriptions, and refinement of terminology related to pharmacovigilance and psychiatric assessment.

We believe that the revised version of the manuscript has been substantially improved and hope that it is now suitable for publication in Brain Sciences. We sincerely appreciate your consideration and are happy to provide any further clarifications if needed.

Thank you again for your valuable feedback.

Prof. Fabrizio Schifano

 

 

 

Reviewer 1

Your manuscript addresses an important and timely pharmacovigilance issue regarding potential neuropsychiatric adverse events associated with 5α-reductase inhibitors (5-ARIs), particularly finasteride and dutasteride. The topic is clinically relevant and aligned with ongoing regulatory scrutiny. The comparative disproportionality analysis between the two agents is of interest, and the proposed structured psychiatric monitoring approach represents a practical contribution.

However, substantial revisions are required to improve methodological transparency, internal consistency, statistical clarity, and precision of language. In its current form, the manuscript requires major revision before it can be considered for publication.

Below are detailed, line-specific, and section-specific comments.

Line 23-31: Finasteride selectively inhibits type-2 5-androgen receptors. Correction: Replace with “type-2 5α-reductase isoenzyme.”

Line 31: Causality remains uncertain. -> Causality cannot be inferred from spontaneous reporting data.

Line 33-36: Lack of details (Period of FAERS data extraction, Which version of MedDra, It was EFDA duplication removed?)

Line 43-46: “Baseline psychiatric assessment and regular monitoring… are recommended.” -> Considering the signal-detection architecture, the recommendation phrasing is robust.

Line 63: “side effects dealing with the sexual domain”-> “sexual adverse effects.”

Lines 71-73: “possible association of the 5-ARI intake and the onset…” -> “possible association between 5-ARI use and the onset…”

Lines 74-76: “hence exerting antidepressant- and neuroprotective-effects.” ->Remove hyphenation. Improve clarity: “which modulate GABAergic neurotransmission and exert antidepressant and neuroprotective effects.”

Lines: 86-88: “Cleaned case listings were used.” Please specify: which quarterly files were uesd ()2004Q1-2025Q2), Duplication method.

Also here I’d like to clarify if there were concomitant drugs exculded

Lines 90-91: Specify the libraries used and offer proof of the availability of the code.

Lines 93-97: It must clearly reference category describe whether the comparison was just between finasteride and dutasteride or vs. Entire FAERS database.

Line 107. It think that it’s a mistake here because Cis are not negative libe here’When ROR >1but Cis were negative values.’

Line 115. Remove or replace: ‘All codes were tested and met all parameters to a 100% accuracy level’ -because it is not scientifically appropriate!

AUTHORS: Thank you for the above comments. Accordingly we have modified the manuscript.

 

Lines 127-130: Please report also the percentages, not only counts (4132 vs 202)

AUTHORS: It is not possible to report percentages in this context, as the present work is based on a pharmacovigilance study design. Pharmacovigilance data—typically derived from spontaneous reporting systems—do not provide a reliable denominator (i.e., the total number of exposed individuals). As a result, calculating incidence rates or meaningful percentages would be methodologically inappropriate and potentially misleading. Therefore, the findings are reported in terms of absolute numbers and descriptive analyses, in line with standard practices for pharmacovigilance research.

 

Lines 136-141: Please may you clarify what is ‘disabled’? It is part of FAD outcome field?

AUTHORS: Thank you for this note. We have further clarified that ‘disabled’ is reported in accordance with U.S. reporting regulations and referenced as such.

 

Table 2: it is difficult to interpret; there are too many categories,  the age is inconsistently formatted, and some country codes are unclear. Try to simplify it and add more to supplementart materials.

AUTHORS: Thank you for this valuable query. We have accordingly modified Table 2 and added two figures (Figure 1 and 2).

 

Lines 151-156 there are some inconstant values regarding in lines 182-184 about RORs (Suicide and Depression) – differenteces. Are these external study values?

AUTHORS: Thank you for this valuable query. We have reviewed the statistics and refined the comparison to relevant previous study findings.  

 

Line 165: Please notice that the absolute novelty should be softened -> “To our knowledge, few studies have comparatively examined…”

AUTHORS: Thank you for the above comments. Accordingly we have modified the manuscript.

 

 

Lines 169-178: Are these FAERS study or a cohort study?

Lines 245-258: I think that you may add: lack of dose information, treatment duration, indicate bias BPH vs alopecia, psychiatric history/??

AUTHORS: Thank you for this valuable query. We have clarified the Methodology and implemented the Limitations section as well. 

 

Bibliography: Justify the use of refs 7 and 11; add a DOI to all the references, and check to see if all the 2025 references are in press or already published.

AUTHORS: Thank you for this suggestion. We have completed a statistical analysis in this format and present the results within the relevant section.

 

Suggestion of the statistics: Stratify y age: <45 years vs. >=45 years and use timetrend ROR every 5 years interval

AUTHORS: Thank you for this interesting suggestion. We have completed a statistical analysis in this format and present the results within the relevant section.

 

Comments on the Quality of English Language

The paper effectively communicates its scientific point and is generally comprehensible. However, to enhance accuracy, fluency, and a formal scientific tone, little language modification is advised.

Particular problems include:

Sporadic errors in grammar (e.g., verb agreement, article use);

Excessive hyphenation;

The phrase “side effects dealing with the sexual domain” is one example of such bad wording;

Little typographical mistakes;

Line 71: “neurolopsychiatric” -> neuropsychiatric

Line 245: “interpretated” -> interpreted

Inconsistent use of technical terms.

Thorough expert language editing would improve the text’s readability and ensure compliance with journal requirements.

AUTHORS: Thank you for these important observations. We have revised the manuscript to clarify the points raised (see the manuscript throughout). Table 2 has also been revised. Finally, references have been modified where possible (not always) adding the DOI.

 

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript by Chiappinin et al. describes a pharmacovigilance study examining mental health–related adverse effects associated with finasteride and dutasteride. The authors use the FAERS database and established methodology.

The main findings show that adverse effects such as depression, anxiety, and suicide-related events were more commonly reported with finasteride among younger patients. The following are suggestions for minor revisions:

1. Expand the discussion of potential mechanistic pathways underlying depression and anxiety associated with finasteride and dutasteride use.
2. Consider discussing bidirectional associations between depression and sexual dysfunction (libido, erectile dysfunction), for example: Journal of Sexual Medicine, Vol. 9, Issue 6, 1497–1507. Such associations may help explain some reported adverse events due to the compounding effects of sexual dysfunction and dopaminergic changes.

 

Author Response

Ms. Sana Zhang
Assistant Editor
E-Mail：sana.zhang@mdpi.com

 

 

Dear Editor and Reviewers,

We would like to sincerely thank you for the time and effort you dedicated to reviewing our manuscript entitled “Signal Detection of Depression and Suicidality Associated with Finasteride and Dutasteride: Updated Pharmacovigilance Evidence and Recommendations for Comprehensive Psychiatric Assessment.” We are grateful for the insightful comments and constructive suggestions, which have significantly improved the quality and clarity of our work.

We have carefully considered all comments and revised the manuscript accordingly. Below, we provide a detailed, point-by-point response to each reviewer’s comments. All changes in the revised manuscript have been highlighted for ease of reference.

In addition to addressing the reviewers’ comments, we have made several minor revisions throughout the manuscript to improve clarity, readability, and consistency. These include grammatical corrections, improved figure/table descriptions, and refinement of terminology related to pharmacovigilance and psychiatric assessment.

We believe that the revised version of the manuscript has been substantially improved and hope that it is now suitable for publication in Brain Sciences. We sincerely appreciate your consideration and are happy to provide any further clarifications if needed.

Thank you again for your valuable feedback.

Prof. Fabrizio Schifano

 

Reviewer 2

The manuscript by Chiappini et al. Describes a pharmacovigilance study examining mental health–related adverse effects associated with finasteride and dutasteride. The authors use the FAERS database and established methodology.

The main findings show that adverse effects such as depression, anxiety, and suicide-related events were more commonly reported with finasteride among younger patients. The following are suggestions for minor revisions:

 

Expand the discussion of potential mechanistic pathways underlying depression and anxiety associated with finasteride and dutasteride use.

Consider discussing bidirectional associations between depression and sexual dysfunction (libido, erectile dysfunction), for example: Journal of Sexual Medicine, Vol. 9, Issue 6, 1497–1507. Such associations may help explain some reported adverse events due to the compounding effects of sexual dysfunction and dopaminergic changes.

 

AUTHORS: Thank you for this thoughtful comment. We have addressed it by almost entirely modifying the Discussion section of the manuscript.

Reviewer 3 Report

Comments and Suggestions for Authors

Dear authors,

• Despite an interesting initial concept, the study in its present form lacks the necessary consistency in order to be published in a journal with a 2.8 IF. The comparison between only two substances across four PTs, already well-documented in the literature and listed in the SmPC and is extremely limited in scope. Furthermore, the study fails to provide a comprehensive descriptive analysis or any sensitivity analyses. To ensure relevance and methodological rigour, I recommend expanding your research to include a broader spectrum of PTs and supplementary analysis (such as based on Evans criteria, EBGM, Time-to-Onset, etc.).
• The role of pharmacovigilance should be detailed in the Introduction section
• Why did you choose FAERS for this analysis?
• Please describe the cleaning, respectively, the extraction data process from FAERS.
• Also, the extraction date should be mentioned.
• What was the reporting period?
• Please update the 95% CI formula. The equation provided currently calculates the interval on the logarithmic scale. To obtain the CI values as typically reported in literature, the lower and upper bounds must be exponentiated.
• A flowchart with the extraction process could be helpful for the reader.
• Table 2. A comparative chart could be used for representing the yearly reporting
• Why do you mention that the “Confidence Interval has been calculated after being exponentiated (applying normalisation - no negative data”)?
• Table 3: Replace 'normalization' with 'log-transformation' (standard procedure for OR/ROR). Also, the phrase 'no negative data' is redundant, as ROR values are inherently non-negative
• Since the initial comparison of finasteride vs. dutasteride already demonstrated disproportionality, calculating the inverse ROR (dutasteride vs. finasteride) does not provide additional statistical value. Please justify the inclusion of both.
• The Reference section should be updated with supplementary titles.

Author Response

Ms. Sana Zhang
Assistant Editor
E-Mail：sana.zhang@mdpi.com

 

 

Dear Editor and Reviewers,

We would like to sincerely thank you for the time and effort you dedicated to reviewing our manuscript entitled “Signal Detection of Depression and Suicidality Associated with Finasteride and Dutasteride: Updated Pharmacovigilance Evidence and Recommendations for Comprehensive Psychiatric Assessment.” We are grateful for the insightful comments and constructive suggestions, which have significantly improved the quality and clarity of our work.

We have carefully considered all comments and revised the manuscript accordingly. Below, we provide a detailed, point-by-point response to each reviewer’s comments. All changes in the revised manuscript have been highlighted for ease of reference.

In addition to addressing the reviewers’ comments, we have made several minor revisions throughout the manuscript to improve clarity, readability, and consistency. These include grammatical corrections, improved figure/table descriptions, and refinement of terminology related to pharmacovigilance and psychiatric assessment.

We believe that the revised version of the manuscript has been substantially improved and hope that it is now suitable for publication in Brain Sciences. We sincerely appreciate your consideration and are happy to provide any further clarifications if needed.

Thank you again for your valuable feedback.

Prof. Fabrizio Schifano

 

 

Reviewer 3

 

Despite an interesting initial concept, the study in its present form lacks the necessary consistency in order to be published in a journal with a 2.8 IF. The comparison between only two substances across four PTs, already well-documented in the literature and listed in the SmPC and is extremely limited in scope. Furthermore, the study fails to provide a comprehensive descriptive analysis or any sensitivity analyses. To ensure relevance and methodological rigour, I recommend expanding your research to include a broader spectrum of PTs and supplementary analysis (such as based on Evans criteria, EBGM, Time-to-Onset, etc.).

AUTHORS: We thank the reviewer for this thoughtful and constructive feedback. We acknowledge that the scope of the present analysis is relatively focused. However, this was a deliberate methodological choice aimed at conducting a hypothesis-driven pharmacovigilance assessment targeting specific drug–event associations of clinical relevance. Regarding the limited number of Preferred Terms (PTs), these were selected a priori based on clinical plausibility and prior signals reported in the literature and product information (SmPC). Expanding the number of PTs indiscriminately may increase noise and reduce the interpretability of disproportionality analyses, particularly in the context of spontaneous reporting systems (See Supplemental materials).

 

The role of pharmacovigilance should be detailed in the Introduction section

AUTHORS: We appreciate this suggestion and have revised the Introduction accordingly. Specifically, we have expanded the section to better contextualize the role of pharmacovigilance in post-marketing drug safety surveillance.

 

Why did you choose FAERS for this analysis?

AUTHORS: We thank the reviewer for raising this important point. The choice of the FDA Adverse Event Reporting System was based on  FAERS accessibility, as it is a publicly available database, allowing reproducibility and transparency of analyses. Also it is one of the largest global pharmacovigilance databases, increasing the likelihood of capturing rare or emerging safety signals.  FAERS is widely used in pharmacovigilance research and by regulatory agencies, making findings more comparable with existing literature. Finally, the standardized coding using MedDRA terminology facilitates consistent identification of Preferred Terms (PTs) and enables robust disproportionality analyses. We have now clarified these points in the Methods section to better justify our methodological choice.

 

Please describe the cleaning, respectively, the extraction data process from FAERS.

AUTHORS: Thank you for this comment. We have clarified within the flow charts mentioned below, and implemented within the methods.

Also, the extraction date should be mentioned.

What was the reporting period?

AUTHORS: Thank you for this query. To clarify, all available cases were accessed. For dutasteride, this was Q3 2002, and finasteride, Q3 1992. Amendments have been made to the manuscript to clarify as such.

Please update the 95% CI formula. The equation provided currently calculates the interval on the logarithmic scale. To obtain the CI values as typically reported in literature, the lower and upper bounds must be exponentiated.

AUTHORS: Thank you to the reviewer for this note. This has now been amended in Equation 2.

 

A flowchart with the extraction process could be helpful for the reader.

AUTHORS: Thank you for this helpful suggestion. We have implemented in the manuscript.

Table 2. A comparative chart could be used for representing the yearly reporting

 

AUTHORS: Thank you for this suggestion. We have implemented 2 bar charts to aid understanding through visual representations in the Results section.

Why do you mention that the “Confidence Interval has been calculated after being exponentiated (applying normalisation - no negative data”)?

 

AUTHORS: Thank you for noting this oversight. This has been removed from the manuscript.

Table 3: Replace 'normalization' with 'log-transformation' (standard procedure for OR/ROR). Also, the phrase 'no negative data' is redundant, as ROR values are inherently non-negative.

 

AUTHORS: Thank you for this note. This has been removed from the manuscript.

Since the initial comparison of finasteride vs. dutasteride already demonstrated disproportionality, calculating the inverse ROR (dutasteride vs. finasteride) does not provide additional statistical value. Please justify the inclusion of both.

 

AUTHORS: Thank you for noting this oversight. This has been removed from the manuscript.

 

The Reference section should be updated with supplementary titles.

AUTHORS: Thank you for this suggestion. Accordingly, we expanded the Reference section.

 

 

Reviewer 4 Report

Comments and Suggestions for Authors

Overall evaluation

This manuscript addresses an important and contentious safety question: potential associations between finasteride/dutasteride and depression, suicidality, anxiety, and sexual dysfunction, using FAERS data and disproportionality analysis. The data extraction and basic ROR methodology are described clearly, and the topic is clinically relevant. However, in its current form the paper appears to over‑interpret comparative signals between finasteride and dutasteride, does not duplicate standard signal extraction methods, does not adequately integrate randomized trial evidence, and goes beyond its evidentiary base by proposing guideline‑like management algorithms.

Major issues

1. Although the authors quote the study of Lee et al. they do not adopt their more standard methods of signal detection.  I think the most informative strategy would be to report both: (a) standard drug‑versus‑database disproportionality measures (IC and/or ROR) for finasteride and dutasteride versus the full FAERS dataset, to align with current pharmacovigilance practice and with Lee et al., and (b) the finasteride–versus–dutasteride pairwise RORs as a secondary, exploratory comparison. Presenting both levels of analysis would make your findings more interpretable and comparable to the existing literature, while still addressing your specific interest in differences between the two 5‑ARIs.

The Discussion and Conclusions repeatedly state that “finasteride appears significantly more prone than dutasteride to generate such signals,” implying a higher intrinsic risk of depression, suicidality, and related ADRs with finasteride. However, the present FAERS two‑drug disproportionality analysis is not well-suited to establish robust conclusions about comparative risk between two drugs in the same class.

FAERS lacks denominators, has substantial under‑reporting and notoriety/media bias, and offers limited control for confounding by indication, age, baseline psychiatric risk, and co‑medications.  Finasteride and dutasteride are used in different populations (for example, younger men with alopecia vs older men with BPH), with different background risks of mood disorders and different likelihoods of reporting.  The comparison is restricted to two 5‑ARIs, without anchoring to other BPH drugs or to the background psychiatric ADR reporting level in FAERS.

2. The Discussion relies heavily on pharmacovigilance data and does not emphasize the data from RCTs which showed little difference between finasteride and dutasteride (Li et al.)

3. Comparator choice and lack of exposure/background context

The rationale for choosing dutasteride as the sole comparator is not fully justified. While both drugs are 5‑ARIs, their real‑world use differs in indication and patient demographics, and the manuscript does not explain clearly why a two‑drug comparison is preferable to other options such as pursued by Lee et al.

4. Scope of clinical recommendations and Figures 1–2

Figures 1 and 2 present detailed algorithms for psychiatric assessment and for deciding when to discontinue finasteride/dutasteride. These function as de facto clinical practice recommendations. No source, guideline, or consensus statement is cited for these specific algorithms.  They go well beyond what can be supported by a single pharmacovigilance disproportionality analysis with acknowledged limitations. No citations are offered to support these recommendations. Treatment and monitoring guidance is typically developed by professional societies or regulatory bodies based on comprehensive evidence synthesis, not de novo from FAERS signals.

Minor issues

Table 3 – redundant ratios: you report both “ROR dutasteride vs finasteride” and “ROR finasteride vs dutasteride,” which are exact reciprocals and thus redundant. I recommend keeping a single, clearly specified comparison direction (for example, finasteride vs dutasteride) and removing the inverse column to improve clarity.

Table 2 – readability: Table 2 is very dense, combining four reaction groups (depressive, suicide‑related, libido‑related, anxiety‑related) and two drugs with multiple age, outcome, year, and country strata in a vertically oriented layout that is hard to read. Please consider decomposing it into four subtables (for example, 2a–2d for each reaction group) with more horizontal structure so that finasteride and dutasteride can be compared side‑by‑side within each ADR category.

 

Author Response

 

 

 

 

Ms. Sana Zhang
Assistant Editor
E-Mail：sana.zhang@mdpi.com

 

 

Dear Editor and Reviewers,

We would like to sincerely thank you for the time and effort you dedicated to reviewing our manuscript entitled “Signal Detection of Depression and Suicidality Associated with Finasteride and Dutasteride: Updated Pharmacovigilance Evidence and Recommendations for Comprehensive Psychiatric Assessment.” We are grateful for the insightful comments and constructive suggestions, which have significantly improved the quality and clarity of our work.

We have carefully considered all comments and revised the manuscript accordingly. Below, we provide a detailed, point-by-point response to each reviewer’s comments. All changes in the revised manuscript have been highlighted for ease of reference.

In addition to addressing the reviewers’ comments, we have made several minor revisions throughout the manuscript to improve clarity, readability, and consistency. These include grammatical corrections, improved figure/table descriptions, and refinement of terminology related to pharmacovigilance and psychiatric assessment.

We believe that the revised version of the manuscript has been substantially improved and hope that it is now suitable for publication in Brain Sciences. We sincerely appreciate your consideration and are happy to provide any further clarifications if needed.

Thank you again for your valuable feedback.

Prof. Fabrizio Schifano

 

Reviewer 4

Overall evaluation

 

This manuscript addresses an important and contentious safety question: potential associations between finasteride/dutasteride and depression, suicidality, anxiety, and sexual dysfunction, using FAERS data and disproportionality analysis. The data extraction and basic ROR methodology are described clearly, and the topic is clinically relevant. However, in its current form the paper appears to over‑interpret comparative signals between finasteride and dutasteride, does not duplicate standard signal extraction methods, does not adequately integrate randomized trial evidence, and goes beyond its evidentiary base by proposing guideline‑like management algorithms.

 

Major issues

 

1. Although the authors quote the study of Lee et al. They do not adopt their more standard methods of signal detection. I think the most informative strategy would be to report both: (a) standard drug‑versus‑database disproportionality measures (IC and/or ROR) for finasteride and dutasteride versus the full FAERS dataset, to align with current pharmacovigilance practice and with Lee et al., and (b) the finasteride–versus–dutasteride pairwise RORs as a secondary, exploratory comparison. Presenting both levels of analysis would make your findings more interpretable and comparable to the existing literature, while still addressing your specific interest in differences between the two 5‑

 

AUTHORS: Thank you for this valuable suggestion. Given the recent update to FAERS, it is not possible for us to rerun this analysis with comparison to the full FAERS dataset. Regarding the inclusion of Lee et al, we have utilised this reference differently to highlight the consistency of signals found across both studies, despite different methodological decisions.

 

 

The Discussion and Conclusions repeatedly state that “finasteride appears significantly more prone than dutasteride to generate such signals,” implying a higher intrinsic risk of depression, suicidality, and related ADRs with finasteride. However, the present FAERS two‑drug disproportionality analysis is not well-suited to establish robust conclusions about comparative risk between two drugs in the same class.

AUTHORS: Dear Reviewer, thanks for this comment. Accordingly we have modified this sentence in the Conclusion.

 

FAERS lacks denominators, has substantial under‑reporting and notoriety/media bias, and offers limited control for confounding by indication, age, baseline psychiatric risk, and co‑medications.  Finasteride and dutasteride are used in different populations (for example, younger men with alopecia vs older men with BPH), with different background risks of mood disorders and different likelihoods of reporting.  The comparison is restricted to two 5‑ARIs, without anchoring to other BPH drugs or to the background psychiatric ADR reporting level in FAERS.

AUTHORS: Thanks for this comment. Accordingly, we have expanded the Limitations section and now recorded the ROR values considering the age subgroups (<45 years vs > 45 years).

 

2. The Discussion relies heavily on pharmacovigilance data and does not emphasize the data from RCTs which showed little difference between finasteride and dutasteride (Li et al.)

AUTHORS: Thank you, Reviewer, for this comment. Accordingly, we have now added a paragraph to the Discussion.

 

3. Comparator choice and lack of exposure/background context

The rationale for choosing dutasteride as the sole comparator is not fully justified. While both drugs are 5‑ARIs, their real‑world use differs in indication and patient demographics, and the manuscript does not explain clearly why a two‑drug comparison is preferable to other options such as pursued by Lee et al.

AUTHORS: Thank you for raising this important point regarding the choice of comparator. We agree that the rationale for focusing on dutasteride as the sole comparator required clearer justification, and we have revised the manuscript accordingly to better explain this decision. In particular, our choice was guided by both clinical and pharmacological considerations. Finasteride and dutasteride are the two most widely used 5-α-reductase inhibitors in clinical practice and, to our knowledge, represent the only agents within this class that are currently available at an international level. As such, they constitute the most relevant and homogeneous comparison when the objective is to explore whether signals potentially related to the pharmacological mechanism of 5-AR inhibition may emerge within this class. We acknowledge that their real-world use differs in terms of indication and patient demographics (e.g., androgenetic alopecia versus benign prostatic hyperplasia), and we now explicitly discuss how these differences may introduce confounding and affect reporting patterns. This limitation has been more clearly emphasized in the revised Discussion. Regarding the possibility of including additional comparators, such as other drugs used for benign prostatic hyperplasia, we agree that this could provide a broader contextual framework, as also pursued in previous studies (e.g., Lee et al.). However, the list of alternative treatments is extensive and pharmacologically heterogeneous, and incorporating all—or even a representative subset—would require substantially more complex analyses and time, potentially extending beyond the scope of the present, focused pharmacovigilance study. Importantly, given that the aim of this work was specifically to explore safety signals related to psychiatric outcomes in relation to 5-ARI pharmacology, we considered it methodologically coherent to restrict the analysis to finasteride and dutasteride. This approach allows for a more focused, hypothesis-generating assessment within a single drug class, while minimizing additional variability introduced by drugs with different mechanisms of action. We also now explicitly acknowledge that future studies including broader comparator groups would be valuable to better contextualize these findings.

 

 

 

4. Scope of clinical recommendations and Figures 1–2

 

Figures 1 and 2 present detailed algorithms for psychiatric assessment and for deciding when to discontinue finasteride/dutasteride. These function as de facto clinical practice recommendations. No source, guideline, or consensus statement is cited for these specific algorithms.  They go well beyond what can be supported by a single pharmacovigilance disproportionality analysis with acknowledged limitations. No citations are offered to support these recommendations. Treatment and monitoring guidance is typically developed by professional societies or regulatory bodies based on comprehensive evidence synthesis, not de novo from FAERS signals.

AUTHORS: We thank the reviewer for this important comment. At present, no specific clinical guidelines provide standardized recommendations for the assessment of depression in patients treated with 5-alpha reductase inhibitors. While psychiatric adverse events are increasingly recognized, no structured approaches (e.g., validated rating scales or defined screening intervals) are formally recommended in this setting. To ensure a comprehensive evaluation of the available literature, we performed a targeted search. Indeed we have already cited https://pubmed.ncbi.nlm.nih.gov/37182121/. In Italy the Agenzia Italiana del Farmaco released a recent Direct Healthcare Professional Communication (DHPC) highlighting the risk of depression and suicidal ideation associated with finasteride and, more cautiously, with dutasteride. This document recommends that patients be informed about potential mood changes and that treatment should be discontinued and medical advice sought if depressive symptoms or suicidal ideation occur.

 

 

Minor issues

Table 3 – redundant ratios: you report both “ROR dutasteride vs finasteride” and “ROR finasteride vs dutasteride,” which are exact reciprocals and thus redundant. I recommend keeping a single, clearly specified comparison direction (for example, finasteride vs dutasteride) and removing the inverse column to improve clarity.

Table 2 – readability: Table 2 is very dense, combining four reaction groups (depressive, suicide‑related, libido‑related, anxiety‑related) and two drugs with multiple age, outcome, year, and country strata in a vertically oriented layout that is hard to read. Please consider decomposing it into four subtables (for example, 2a–2d for each reaction group) with more horizontal structure so that finasteride and dutasteride can be compared side‑by‑side within each ADR category.

AUTHORS: Thank you for these comments. Accordingly, we have modified Table 2 and Table 3 and added two Figures (Figure 1 and 2).

 

Although the authors quote the study of Lee et al. they do not adopt their more standard methods of signal detection. I think the most informative strategy would be to report both: (a) standard drug‑versus‑database disproportionality measures (IC and/or ROR) for finasteride and dutasteride versus the full FAERS dataset, to align with current pharmacovigilance practice and with Lee et al., and (b) the finasteride–versus–dutasteride pairwise RORs as a secondary, exploratory comparison. Presenting both levels of analysis would make your findings more interpretable and comparable to the existing literature, while still addressing your specific interest in differences between the two 5‑ARIs.

AUTHORS: Thank you for this valuable suggestion. Given the recent update to FAERS, it is not possible for us to rerun this analysis with comparison to the full FAERS dataset. Regarding the inclusion of Lee et al, we have utilised this reference differently to highlight the consistency of signals found across both studies, despite different methodological decisions.

 

 

The Discussion and Conclusions repeatedly state that “finasteride appears significantly more prone than dutasteride to generate such signals,” implying a higher intrinsic risk of depression, suicidality, and related ADRs with finasteride. However, the present FAERS two‑drug disproportionality analysis is not well-suited to establish robust conclusions about comparative risk between two drugs in the same class. FAERS lacks denominators, has substantial under reporting and notoriety/media bias, and offers limited control for confounding by indication, age, baseline psychiatric risk, and co medications.

AUTHORS: Dear Reviewer, thank you for your thoughtful and constructive comments. We fully agree that the FAERS-based disproportionality approach adopted in our study has important limitations, particularly when attempting to draw comparative conclusions between two drugs within the same class. In response to your observation, we have revised the Discussion and Conclusions to more clearly reflect the exploratory nature of our findings and to avoid any implication of a definitive difference in intrinsic risk between finasteride and dutasteride. Specifically, statements suggesting that finasteride is “significantly more prone” have been softened to indicate that it “The present pharmacovigilance study suggests that the use of 5-ARIs may be associated with safety signals related to depression, anxiety, suicidality, and sexual dysfunction, with finasteride appearing somewhat more likely than dutasteride to be linked to these signals,” while explicitly acknowledging that this does not establish a causal or comparative risk. We have also expanded the limitations section to emphasize the well-known constraints of FAERS data, including the absence of exposure denominators, underreporting, and the potential for notoriety and media bias. In addition, we now more clearly discuss the limitations of this study.

 

Finasteride and dutasteride are used in different populations (for example, younger men with alopecia vs older men with BPH), with different background risks of mood disorders and different likelihoods of reporting. The comparison is restricted to two 5‑ARIs, without anchoring to other BPH drugs or to the background psychiatric ADR reporting level in FAERS.

AUTHORS: Thank you for this comment, Reviewer. We have highlighted that finasteride and dutasteride are often prescribed to distinct patient populations (e.g., younger individuals with androgenetic alopecia versus older patients with benign prostatic hyperplasia), which may differ substantially in their baseline risk of psychiatric outcomes and in reporting behaviour. We have now added new ROR values distinguished for the age of patients (<45 and > 45 years).

 

 

2. The Discussion relies heavily on pharmacovigilance data and does not emphasize the data from RCTs which showed little difference between finasteride and dutasteride (Li et al.)

AUTHORS: see above.

 

 

3. Comparator choice and lack of exposure/background context

The rationale for choosing dutasteride as the sole comparator is not fully justified. While both drugs are 5‑ARIs, their real‑world use differs in indication and patient demographics, and the manuscript does not explain clearly why a two‑drug comparison is preferable to other options such as pursued by Lee et al.

AUTHORS: As above. We agree that the rationale for focusing on dutasteride as the sole comparator required clearer justification, and we have revised the manuscript accordingly to better explain this decision.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors,

I am delighted to inform you that Brain Sciences has accepted your manuscript for publication.

Congratulations on your work and on successfully addressing the reviewers’ comments.

Kind regards,

Reviewer 4 Report

Comments and Suggestions for Authors

My concerns have been addressed.