Antipsychotic Medications in Parkinson’s Disease Psychosis; A Systematic Review of Double-Blind, Randomised, Placebo-Controlled Trials

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript addresses a clinically very relevant question: the comparative efficacy and tolerability of antipsychotic agents for Parkinson’s disease psychosis (PDP), focusing on double-blind, randomized, placebo-controlled trials. The topic is highly important for clinicians and fits well with the scope of Healthcare. The review is generally well written, the structure is clear, and the main findings (superiority of clozapine and pimavanserin vs placebo; lack of convincing efficacy for quetiapine; unfavourable motor profile for olanzapine) are clinically meaningful.

However, several aspects need to be strengthened before the study can reach its full potential and align more closely with Healthcare’s expectations for systematic reviews. In particular, (1) the introduction and discussion should position the work more clearly in relation to recent meta-analyses and umbrella reviews, including those authored by the present team; (2) the methods need more detailed reporting and, ideally, methodological reinforcement; and (3) some journal-specific structural elements (highlights format and back matter sections) should be adjusted.

Overall, I consider the scientific basis of the work sound, but substantial revisions are required, mainly on methodological reporting and contextualization in the existing literature.

Major comments

1. Introduction and positioning in the literature

• The introduction provides a good general overview of PDP and the clinical dilemma facing clinicians. However, the state of the art is not yet sufficiently detailed given the amount of high-quality evidence accumulated in recent years.

• Several recent meta-analyses, network meta-analyses and umbrella reviews (2019–2024) are directly relevant and, in some cases, reach very similar conclusions to those presented here. Some of these are already cited, but others are only briefly mentioned or not fully integrated into the narrative.

• Importantly, there appears to be previous work by the same group on antipsychotics in PDP, including a systematic review and meta-analysis. The relationship between that work and the present manuscript is not entirely clear to the reader. At present, it risks appearing as a partial overlap or “re-packaging” of similar data.

Suggestions:

• Expand the introduction with a more detailed and explicit review of existing systematic reviews/meta-analyses/umbrella reviews on antipsychotic treatment in PDP (e.g., those published between 2019 and 2024).

• Clearly articulate what is genuinely new in the present manuscript. For example, you might emphasise:

  • the strict restriction to double-blind, randomized, placebo-controlled trials;

  • a more granular focus on dose/titration and motor tolerability;

  • the translation of evidence into practical implications for real-world prescribing and service organisation.

• Explicitly clarify how this paper differs from and complements your own previous work (same or overlapping author team) so that readers and editors can appreciate the added value and avoid concerns about duplication.

2. Methods: reporting and design issues

The overall design (systematic review of double-blind, randomized, placebo-controlled RCTs) is appropriate and methodologically robust. However, the reporting of the methods is not yet sufficient for full transparency and reproducibility.

Key issues:

1. Protocol registration

   • It is not stated whether the review was preregistered (e.g., in PROSPERO). Current standards (and MDPI’s expectations for systematic reviews) strongly recommend preregistration.

   • If no prior registration was made, this should be clearly acknowledged as a limitation, with a brief rationale.

2. Search strategy and databases

   • The manuscript lists several major databases (e.g., MEDLINE, Embase, PsycINFO, PubMed, Google Scholar), which is commendable.

   • However, for a review of RCTs, one would also expect Cochrane CENTRAL and clinical trial registries (e.g., ClinicalTrials.gov, WHO ICTRP) to be considered or at least discussed. If they were not searched, please explain why and discuss the potential impact on the completeness of evidence.

   • To comply with PRISMA standards, the full electronic search strategy for each database (or at least for the main ones) should be presented in an appendix, including all search terms, limits, and filters.

3. Study selection and data extraction

   • It is mentioned that two reviewers were involved, but the process should be described more explicitly:

     • Were titles/abstracts and full texts screened independently by two reviewers?

     • How were discrepancies resolved (discussion, third reviewer)?

   • Similarly, data extraction procedures should specify how disagreements were handled and whether a standardised extraction form was used.

4. Risk of bias assessment

   • You correctly employed a Cochrane risk-of-bias tool, but please specify whether you used RoB 1.0 or RoB 2.0 and clarify how domains were judged and synthesised.

   • If any domain was frequently at “unclear” or “high” risk, it would be useful to highlight this more explicitly in the Results and discuss its implications.

5. Choice not to perform a quantitative meta-analysis

   • The decision not to perform a pooled meta-analysis, despite the presence of multiple RCTs with similar design and outcomes, is a pivotal methodological choice.

   • At present, the justification based on heterogeneity and limited number of trials is not fully convincing, given that other groups have conducted meta-analyses with comparable datasets.

   • While it is legitimate to opt for a more qualitative or “trial-by-trial” synthesis, it is important to:

     • state clearly a priori why a formal meta-analysis was not performed;

     • provide a more detailed explanation of the specific sources of heterogeneity that prevented pooling (e.g., different outcome scales, follow-up durations, dosing strategies, concomitant medications).

   • Alternatively, you could consider performing at least limited meta-analyses for agents with more homogeneous data (e.g., clozapine, pimavanserin), which would greatly strengthen the quantitative backbone of the paper.

6. Use of GenAI tools

   • In line with MDPI’s current policy, please add a brief statement on whether any generative AI tools were used in the manuscript preparation. If not, a simple statement such as “No generative AI tools were used in the writing of this manuscript” would suffice.

Overall, I would recommend substantially expanding and clarifying the Methods section to fully meet PRISMA and journal standards.

3. Results: clarity and structure

The results section is generally clear and logically organised by agent, with informative tables. The PRISMA flow diagram is appropriate.

Suggestions for improvement:

• In some instances, the textual description is overly detailed and repeats numerical information already presented in the tables. To improve readability, consider:

  • summarising key patterns in the text (e.g., “all quetiapine trials failed to show significant improvement in psychotic symptoms versus placebo”) while leaving detailed numbers to the tables;

  • highlighting effect directions, magnitude, and consistency across trials, rather than restating each value.

• If no meta-analysis is performed, you might still consider adding simple visual summaries (e.g., forest-like plots with individual study effects for each drug) to make the pattern of results more immediately evident.

4. Discussion and conclusions

The discussion is clinically thoughtful and makes important points about the misalignment between evidence and common prescribing practices. However, several refinements would enhance its robustness:

• Provide a more systematic comparison with previous reviews and meta-analyses, explicitly stating where your findings agree or disagree, and why.

• The tone of some statements is relatively strong and prescriptive (e.g., “X should be first-line”). Given that this is a review, not a guideline, I recommend slightly tempering the language:

  • prefer formulations such as “Our findings support the consideration of clozapine as a preferred option…” or “These data suggest that quetiapine should not be considered evidence-based first-line therapy…”.

• The limitations section should be expanded to include:

  • absence of protocol registration;

  • restriction to English-language trials;

  • exclusion of non–placebo-controlled studies and registries, and the impact this may have on generalizability and publication bias;

  • the limited sample sizes and number of available RCTs for some agents.

• Consider elaborating a brief paragraph on future research directions, such as:

  • head-to-head trials comparing clozapine and pimavanserin;

  • longer-term safety studies (hematologic, cardiovascular, cognitive);

  • integration of clinical trials with real-world data and digital monitoring of psychotic and motor symptoms.

5. Journal-specific and formal aspects (Healthcare / MDPI)

A few structural and formal issues should be corrected to conform fully to Healthcare’s guidelines:

1. Highlights

   • The journal asks for up to two bullet points for “main findings” and two for “implications”. At present, your Highlights section exceeds these limits.

   • Please condense the main findings into two concise bullets, and similarly keep implications to two bullets.

2. Back matter sections

   • For Healthcare articles, the following sections are required, even for systematic reviews using only published data:

     • Author Contributions (CRediT);

     • Funding;

     • Institutional Review Board Statement;

     • Informed Consent Statement;

     • Data Availability Statement;

     • Acknowledgments (if applicable);

     • Conflicts of Interest.

   • Some of these are currently missing or incomplete. Please add them in the standard MDPI format (e.g., “Ethical review and approval were waived…”, “No new data were created or analysed…”, etc.).

3. Figures and tables

   • Ensure that all figures and tables have clear titles and legends, including full definition of all abbreviations at first use.

   • In the final version, figures should appear as close as possible to their first citation in the text.

4. References

   • The references currently cited are appropriate and relevant. Nonetheless, I strongly encourage the inclusion of the most recent key systematic reviews, meta-analyses, and umbrella reviews, which will reinforce the discussion and clearly show how your work builds upon and differentiates itself from them.

   • Please double-check all references for consistency of year, volume, pages, and DOIs.

Minor comments

• Ensure consistent use of terminology and spelling throughout (e.g., “randomized” vs “randomised”; “Parkinson’s disease psychosis” vs “PD psychosis”).

• Check that all abbreviations are defined at first appearance in the main text and abbreviations list.

• A small number of sentences could be slightly rephrased for clarity, but overall the quality of English is good and does not require major editing.

In summary, this is a clinically important and potentially valuable systematic review. With a clearer positioning in relation to existing literature, more detailed and transparent methods, and alignment with Healthcare’s structural requirements, the manuscript could make a strong contribution to the field.

Comments on the Quality of English Language

The overall quality of English in this manuscript is good, and the text is generally clear, precise, and appropriate for an international medical audience. Sentences are well structured, the terminology is technically correct, and the argument flows logically across sections. In its current form, the manuscript does not require major language editing, and it is certainly understandable for Healthcare’s readership.

A few minor refinements could further enhance readability and stylistic consistency:

• Ensure consistent use of US vs UK spelling (e.g., “randomized” vs “randomised”) throughout the manuscript, including the title, abstract, and tables, in line with the journal’s preferred style.

• Harmonize terminology so that key constructs are referred to in a uniform way (e.g., consistently using “Parkinson’s disease psychosis (PDP)” rather than alternating with “psychosis in PD” or “PD psychosis” unless context specifically requires differentiation).

• Check that all abbreviations are defined at first use in the main text and in tables/figures, and avoid introducing abbreviations that are used only once.

• In the Results section, you may consider lightly streamlining some long descriptive sentences that repeat numerical details already reported in the tables, to make the narrative more concise and easier to follow.

These are minor, mostly stylistic adjustments. From a linguistic standpoint, the manuscript is well written and fits the standard of Healthcare; any additional polishing would be optional and aimed only at improving flow and internal consistency.

Author Response

Please see attachment 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This is an interesting review which emphasize double blind trial investigating efficacy and tolerability of antipsychotic drugs in parkinson disease psychosis

The paper is well written and very clear. The message is important for neurologists and psychiatrists who may be confronted with Parkisnon disease psychosis
I have only a few remarks

• the surveillance rules for clozapine will probably be less severe in the future (since agranulocytosis is rare after the first year of treatment). This point could be evoked
• recent meta analysis are absent in the references:
  • Srisurapamont et al Gen Hosp Psychiatry 2024
  • Iketani et al PRD 2020
  • others?

Author Response

Please see attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This manuscript discusses antipsychotic treatments PD psychosis.

I wish to offer the following observations for the authors:

 

Major points -

• The authors can consider adding controlled vocabulary (like MeSH terms) to their search strategy to improve the sensitivity and avoid missing articles
• Line 120 – I am not sure why comparator studies are mentioned in exclusion criteria, when RCTs are being included
• Line 142 – The authors may specify the tool used instead of software, so they may mention RoB v1.0 or the other tool
• Even when not performing a meta-analysis, a systematic review finding may be synthesised formally using recommendations like SWiM (Synthesis Without Meta-analysis) guideline - https://pubmed.ncbi.nlm.nih.gov/31948937/
• Line 160 - Applying search limits of “abstract” and “randomised controlled trial” post-search is not recommended, and this should not be done, and the results (especially when so low results) should be screened
• Figure 1 – The PRISMA flowchart should report numbers separately for each database, not clubbing them
• The reasons in the list of manually excluded reasons add up to 108, the authors may find the missing five studies
• Line 43 in Discussion – Effect size should usually be interpreted in the light of minimally important difference, and not just using effect size
• Table 2 seems more apt for results section than the discussion section
• Line 159 in Conclusions – LOCF is not usually recommended as a strategy because of a pre-decision, and multiple imputation should be preferred

 

Relatively minor point -

• Line 130 – The authors may change it to adverse events

 

Author Response

Please see attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The revised manuscript shows a substantial and appreciable improvement compared with the initial version. You have clearly taken the previous review comments seriously and implemented changes at multiple levels: conceptual framing, methodological transparency, integration with the existing literature, and formal alignment with MDPI requirements.

The topic remains highly relevant clinically: the management of Parkinson’s disease psychosis (PDP) with antipsychotic agents, and in particular the question of how to weigh efficacy on psychotic symptoms against motor tolerability and safety, is central to everyday practice in neurology, geriatrics, and psychiatry. Your focus on double-blind, randomized, placebo-controlled trials ensures that the review synthesizes the highest level of evidence available, which is entirely appropriate for Healthcare.

Overall, the revised manuscript is much stronger and, in my view, now close to being suitable for publication, provided a few remaining points are addressed.

 

1. Introduction and positioning in the literature

You have substantially strengthened the Introduction:

• The background on PDP and the clinical dilemma around antipsychotic use is now clear and concise.
• The state of the art is far better detailed. The addition of the table summarizing previous systematic reviews, meta-analyses and network meta-analyses on antipsychotics in PDP is particularly helpful. It allows the reader to immediately understand the different scopes, inclusion criteria, and time windows of prior work.
• Importantly, you now situate your review as a methodologically stringent update of earlier meta-analyses that used similar eligibility criteria (double-blind, randomized, placebo-controlled trials only), while also contrasting it with broader meta-analyses and network meta-analyses that included open-label and active-comparator trials.

This revision directly addresses earlier concerns about novelty and potential redundancy. The added clarity about where your work stands in relation to the existing literature is a major improvement.

Minor suggestion:
You might consider adding one or two sentences in the Introduction that anticipate more explicitly the clinical take-home message (e.g., regarding the relative roles of clozapine, pimavanserin, and quetiapine in practice), so that readers immediately grasp why this update matters for everyday decision-making.

 

2. Methods: transparency and rigor

The Methods section has become considerably more transparent and aligned with PRISMA and Healthcare expectations.

Positive changes include:

• You now state clearly that the review was conducted with reference to PRISMA 2020, and that the protocol was developed a priori but not preregistered. This explicit acknowledgment of non-registration is important and honest.
• The search strategy is more clearly described, including:
• databases used,
• time-frame covered,
• restriction to fully published, peer-reviewed trials,
• and the use of supplemental searches.
• The decision to focus exclusively on double-blind, randomized, placebo-controlled trials is now consistently linked to the review’s aim of synthesizing the highest level of efficacy evidence, and the trade-offs of this choice are appropriately discussed in the limitations.
• The risk-of-bias assessment is much more detailed. You discuss power calculations, incomplete outcome data, attrition, and the implications of LOCF and other analytical choices in the primary trials, not merely as a checklist but as factors that genuinely shape the interpretation of your findings.
• Crucially, you have introduced a limited pooled meta-analysis where data permit, rather than relying solely on narrative summaries of individual trial effect sizes.

These are all significant and commendable improvements.

Suggestions for further refinement:

1. Statistical details of the pooled analyses
• Since you now perform limited meta-analyses, it would be valuable to specify a few additional technical details, for example:
• whether a fixed-effect or random-effects model was used (and why);
• how heterogeneity was assessed (e.g., I² statistics) and interpreted;
• how you handled trials with multiple arms (e.g., different doses of the same drug vs placebo).
• If feasible, consider providing simple forest plots (even in the Appendix) to visually summarize the pooled effects for clozapine and pimavanserin. This would reinforce the quantitative backbone of your conclusions.
2. Trial registries and unpublished data
• You now explain that you limited inclusion to fully published, peer-reviewed RCTs. This is methodologically defensible, and you do discuss the possible impact on publication bias in the Limitations.
• It may still be helpful to explicitly state in the Methods that trial registries were not systematically searched and that unpublished or non–peer-reviewed data were deliberately excluded, so that this choice is transparent already at the design stage (not only in the Limitations section).

With these additions, the methodological reporting would be fully in line with the level of detail expected for systematic reviews in Healthcare.

 

3. Results: clarity, structure, and interpretability

The Results section remains well structured, but in the revised version it is more interpretive and better integrated with both the risk-of-bias considerations and the limited meta-analysis:

• Presentation by drug (clozapine, quetiapine, olanzapine, pimavanserin) is clear and logical.
• You now more clearly distinguish between statistical significance and clinical relevance, particularly for pimavanserin, where small-to-moderate effect sizes and mixed findings across scales raise legitimate questions about the practical meaning of some improvements.
• The discussion of heterogeneity across trials (duration, dosing schemes, diagnostic criteria, scales used, baseline severity) is tighter and more consistently linked to how confidently we can generalise your findings.

Suggestion:
In some passages, especially where individual trials are described in detail, you could still slightly streamline the text by focusing on the most clinically salient results and directing the reader to the tables and appendices for detailed numerical values. This would improve readability without sacrificing information.

 

4. Discussion and conclusions

The Discussion has been substantially enriched and is now one of the strengths of the paper.

You:

• Provide a much more systematic comparison with previous meta-analyses and network meta-analyses, explaining where your estimates converge or diverge and why (e.g., exclusion of non-published pimavanserin trials, exclusion of comparator and open-label studies).
• Elaborate on how study-level limitations (small sample sizes, incomplete outcome data, attrition patterns, heterogeneity) influence your confidence in the results for each agent.
• Offer a nuanced and clinically relevant reading of the pimavanserin evidence, particularly the tension between modest mean improvements on psychosis scales and the more robust proportions of responders on global impression scales.

The Limitations section is also much stronger: you now explicitly mention non-registration of the protocol, language restrictions, exclusion of non–peer-reviewed/non-published data, and the impact of heterogeneity and small sample sizes on effect estimates.

Regarding the Conclusions and clinical implications:

• Your statement that quetiapine has not demonstrated benefit over placebo in DBRPCTs and should not be considered first-line antipsychotic treatment for PDP is clearly grounded in the evidence you review.
• The updated discussion of clozapine monitoring requirements and the recent relaxation of long-term blood-test schedules is particularly useful. It meaningfully shifts the risk–benefit balance for clozapine and is exactly the sort of practice-oriented information that Healthcare readers will appreciate.
• Your formulation of pimavanserin as a promising alternative, but with limitations in the current evidence base, is well balanced.

Minor suggestion on tone:
Some sentences could still be made slightly more cautious in wording, not to dilute the clinical message, but to maintain a clear distinction between evidence synthesis and prescriptive guideline. For example, phrases like “should not be considered first-line” could be rephrased as “current evidence does not support quetiapine as first-line antipsychotic treatment for PDP, particularly when agents with demonstrated superiority over placebo are available.” This maintains the strength of the message but keeps it within the natural boundaries of a review article rather than a guideline.

 

5. Formal aspects, structure, and language

You have also addressed most of the formal and structural issues raised in the previous round:

• The inclusion of the full set of back-matter sections (Author Contributions, Funding, IRB Statement, Informed Consent, Data Availability, Acknowledgments, Conflicts of Interest) now aligns the manuscript with MDPI’s standard format for research and review articles.
• The Highlights have been appropriately revised in length and structure (though please ensure they fully match the journal’s current template—two bullet points for main findings and two for implications).
• The English language has improved noticeably. The text is now fluent, technically correct and coherent, with only minor stylistic polishing potentially needed at copy-editing stage. Your explicit acknowledgment of the use of an AI tool for language refinement is transparent and consistent with MDPI’s policy.

 

6. Summary

In summary, the progression from the initial to the revised version of your manuscript is clearly and substantially positive:

• The scientific content is more robust, better positioned within the existing literature, and supported by clearer methodological reporting and limited meta-analytic synthesis.
• The clinical implications are more nuanced and up to date, especially regarding clozapine monitoring and the practical role of pimavanserin.
• The manuscript is now largely aligned with Healthcare’s expectations for systematic reviews in terms of structure, transparency, and style.

The remaining suggestions are mainly refinements: modest clarifications about the statistical methods used in the pooled analyses, possible slight softening of the prescriptive tone of certain clinical statements, and minor streamlining of some descriptive passages in the Results. Addressing these points would further strengthen an already solid and clinically important paper.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

The overall quality of English in this manuscript is good, and the text is generally clear, precise, and appropriate for an international medical audience. Sentences are well structured, the terminology is technically correct, and the argument flows logically across sections. In its current form, the manuscript does not require major language editing, and it is certainly understandable for Healthcare’s readership.

A few minor refinements could further enhance readability and stylistic consistency:

• Ensure consistent use of US vs UK spelling (e.g., “randomized” vs “randomised”) throughout the manuscript, including the title, abstract, and tables, in line with the journal’s preferred style.

• Harmonize terminology so that key constructs are referred to in a uniform way (e.g., consistently using “Parkinson’s disease psychosis (PDP)” rather than alternating with “psychosis in PD” or “PD psychosis” unless context specifically requires differentiation).

• Check that all abbreviations are defined at first use in the main text and in tables/figures, and avoid introducing abbreviations that are used only once.

• In the Results section, you may consider lightly streamlining some long descriptive sentences that repeat numerical details already reported in the tables, to make the narrative more concise and easier to follow.

These are minor, mostly stylistic adjustments. From a linguistic standpoint, the manuscript is well written and fits the standard of Healthcare; any additional polishing would be optional and aimed only at improving flow and internal consistency.

Author Response

Please see the attachment

Author Response File: Author Response.pdf