Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript is focused on a rapid review of human data on Trifluoroacetic Acid (TFA) exposure across medical, occupational, and accident scenarios. It identifies 17 primary studies including case reports, observational work, pharmacokinetic studies,  exposure assessments, and concludes that i) acute systemic toxicity is low; ii) corrosive injury can occur through direct skin contact; and iii) evidence is insufficient on chronic , low-level exposure. The topic is timely given the persistence of these solvents and need for regulatory attention.

Some major revisions are indicated:

• The title and text contain a recurring misspelling: “Trifluoractetic.” Please correct to Trifluoroaceti. Also ensure consistent use of “trifluoroacetate (TFA). 
• In the abstract, add dates of search (April 28–29, 2025) and databases and report n by design consistently
• The manuscript cites 17 studies, but the categorization varies across sections (e.g., 5 vs  3 PK + 2 on methodological exposure). Please harmonize counts.
• The study adheres to Cochrane for rapid review guidance and PRISMA, and reference an a priori OSF protocol. However, the study selection and evaluation were largely based on single reviewer. Please clarify exactly which stages were dual vs single, and discuss the potential bias this may introduce. In addition, for pharmacokinetic studies, you rely on “expert assessment” - provide a transparent definition
• The review excludes biomonitoring studies reporting TFA in human matrices without a specified source. This may narrow the scope and miss clinically relevant associations with liver, renal, or inflammatory biomarkers. Please justify the exclusion.
• Several assumptions on exposure reconstruction are insufficiently justified. For anesthetized patients you used “sleeping” inhalation rates from EPA, but under general anesthesia, ventilation is typically controlled and differs from sleeping physiology. Please either use anesthesiology-appropriate ventilation volumes for adults/children or present this as a scenario analysis. For fraction absorbed and metabolized (halothane → TFA), you use fixed 27%/50% absorption and 12%/20% metabolism as “most-probable”/“worst-case.” Please provide a justification.
• Desflurane to TFA (Table 3): the worst-case totals (e.g., 891 mg/kg) exceed expectations given the reported low metabolism of desflurane to trifluoroacetylated metabolites. Please re-check the assumed metabolized fraction, and whether you inadvertently treated fluoride or TFA-adduct formation as if fully converting to free TFA.
• Consider adding a summary of findings by exposure context (dermal accident, occupational inhalation, anesthetic-related, experimental inhalation) listing key clinical outcomes.
• In the discussion consider providing a clear distinction between TFA as a strong acid causing local injury vs systemic effects of circulating TFA and trifluoroacetylation pathways linked to halothane hepatitis.
• Several references appear incomplete or oddly formatted. Please standardize to the Life (MDPI) reference style

Author Response

Reviewer 1

This manuscript is focused on a rapid review of human data on Trifluoroacetic Acid (TFA) exposure across medical, occupational, and accident scenarios. It identifies 17 primary studies including case reports, observational work, pharmacokinetic studies, exposure assessments, and concludes that i) acute systemic toxicity is low; ii) corrosive injury can occur through direct skin contact; and iii) evidence is insufficient on chronic, low-level exposure. The topic is timely given the persistence of these solvents and need for regulatory attention.

Some major revisions are indicated:

Comment 1: The title and text contain a recurring misspelling: “Trifluoractetic.” Please correct to Trifluoroaceti. Also ensure consistent use of “trifluoroacetate (TFA). 

Response 1: Thank you for pointing this out. We corrected this mistake in the title [page 1, line 2], added the spelling “Trifluoroacetic Acid” to the Keywords [page 1, line 21]. Furthermore, we adapted it for a consistent use on [page 1, paragraph 1, line 24], [page 1, paragraph 3, line 41], [page 1, paragraph 4, line 43], [table 2, page 9, row 6].

Comment 2: In the abstract, add dates of search (April 28–29, 2025) and databases and report n by design consistently

Response 2: We added the databases and the dates of the search to the abstract, see [page 1, paragraph 1, lines 13 and 14]. We applied changes to have a consistent report of n by design [page 7, paragraph 1, lines 250 and 251]

Comment 3: The manuscript cites 17 studies, but the categorization varies across sections (e.g., 5 vs  3 PK + 2 on methodological exposure). Please harmonize counts.

Response 3: We harmonized the counts [page 7, paragraph 1, line 251].

Comment 4: The study adheres to Cochrane for rapid review guidance and PRISMA, and reference an a priori OSF protocol. However, the study selection and evaluation were largely based on single reviewer. Please clarify exactly which stages were dual vs single, and discuss the potential bias this may introduce. In addition, for pharmacokinetic studies, you rely on “expert assessment” - provide a transparent definition.

Response 4: Thank you for your comment. As stated in the manuscript, the study selection was (at least) a dual stage: the title/abstract screening was performed by 10 reviewers [page 5, paragraph 4, line 205] and the full-text screening was performed by two reviewers [page 5, paragraph 5, line 210]. The quality assessment was performed by one person and reviewed by a second one [page 3, paragraph 6, lines 124 and 125], except for the pharmacological studies. We did not assess the quality of the evidence and did not synthesize multiple studies, therefore we believe that this approach sufficiently minimizes the risk of systematic bias and is appropriate for the scope and design of our study.

We added an explanation for “expert assessment” [page 3, paragraph 5, line 119-121] also see our response to Reviewer 2, comment 3(b).

Comment 5: The review excludes biomonitoring studies reporting TFA in human matrices without a specified source. This may narrow the scope and miss clinically relevant associations with liver, renal, or inflammatory biomarkers. Please justify the exclusion.

Response 5: As already stated in the protocol, we excluded studies with unclear TFA exposure, as our aim was to associate a known or at least estimable exposure with its potential effects. Biomonitoring studies with unknown sources of exposure do not allow for such conclusions and, in our view, would have broadened the scope of the review too much. We added a sentence in the discussion [page 14, paragraph 6, line 355 and 356]   

Comment 6: Several assumptions on exposure reconstruction are insufficiently justified. For anesthetized patients you used “sleeping” inhalation rates from EPA, but under general anesthesia, ventilation is typically controlled and differs from sleeping physiology. Please either use anesthesiology-appropriate ventilation volumes for adults/children or present this as a scenario analysis. For fraction absorbed and metabolized (halothane → TFA), you use fixed 27%/50% absorption and 12%/20% metabolism as “most-probable”/“worst-case.” Please provide a justification.

 

Response 6:

Thank you for your comment and the constructive feedback. We have researched inhalation rates for anesthesia and updated the results in table 3 and figure 2. Furthermore, we have included a summary of the assumptions used for exposure calculation in the ANNEX.

 

Comment 7: Desflurane to TFA (Table 3): the worst-case totals (e.g., 891 mg/kg) exceed expectations given the reported low metabolism of desflurane to trifluoroacetylated metabolites. Please re-check the assumed metabolized fraction, and whether you inadvertently treated fluoride or TFA-adduct formation as if fully converting to free TFA.

 

Response 7:

Thank you for your valuable feedback. We did use total metabolization rates for desflurane, which range from 0.2% to 0.002% according to the literature, since we could not find specific rates for its metabolization to TFA. Because using the upper limit of 0.2% resulted in unrealistically high TFA exposure values, we have re-calculated this “worst-case” scenario for desflurane in Table 3 in the manuscript. Given that other products are also formed during desflurane’s metabolism, 0.002% may still overestimate the amount converted to TFA. Therefore, we used 0.002% as the metabolization rate for both the “most-probable” and “worst-case” scenarios, adjusting only the inhalation rate. This is described in more detail in the material and methods section [page 4, paragraph 2, line 158 and 159] and the ANNEX under Metabolism of Desflurane, Uncertainty Analysis)

Comment 8: Consider adding a summary of findings by exposure context (dermal accident, occupational inhalation, anesthetic-related, experimental inhalation) listing key clinical outcomes.

Response 8:

Thank you for the suggestion. We have added a table showing the clinical outcomes for all studies in which results other than laboratory parameters were available, broken down by exposure. [page 13, Table 4],

Comment 9: In the discussion consider providing a clear distinction between TFA as a strong acid causing local injury vs systemic effects of circulating TFA and trifluoroacetylation pathways linked to halothane hepatitis.

Response 9: Thank you for this helpful comment. To clarify this point, we have added a corresponding sentence to the discussion section. [page 14, paragraph 2, lines 322 – 325].

Comment 10: Several references appear incomplete or oddly formatted. Please standardize to the Life (MDPI) reference style

Response 10: Thank you for this comment. We revised the references and standardized them to the Life (MDPI) reference style [page 22-24].

Reviewer 2 Report

Comments and Suggestions for Authors

Major comments

1. Scope and contribution: the review addresses an important and timely issue given growing environmental TFA levels. The work is clearly structured and the PECO framework is appropriate. Nonetheless, the manuscript would be strengthened by a clearer description of how this rapid review adds to existing assessments (e.g., EFSA draft and recent reviews cited) and by explicitly stating the review’s unique contribution in the Introduction and Conclusion.
2. Search strategy and reproducibility: the PubMed/Embase search strategy is provided in Appendix A, but more transparency is required for reproducibility. Please provide (a) the full, exact search strings used for each database (including date stamps), (b) the deduplication method and software settings, and (c) the registration details for the protocol (the OSF link is noted — include the protocol ID and a direct reference). Also indicate whether grey literature or non-English studies beyond the stated languages were considered.
3. Study selection and quality assessment: the review includes a heterogeneous set of study designs ranging from case reports to pharmacokinetic studies. While the chosen appraisal tools are acceptable, the manuscript should provide (a) a detailed table listing individual quality assessment scores for each included study (not only aggregate judgments), (b) justification for treating pharmacokinetic studies via “expert opinion” rather than an established tool, and (c) sensitivity analyses showing how conclusions would change if lower-quality studies were excluded.
4. Exposure estimation and uncertainty: substantial effort was made to estimate “most-probable” and “worst-case” TFA doses, but the assumptions (absorption fractions, metabolism rates, inhalation rates) introduce considerable uncertainty. I recommend (a) adding a dedicated subsection that lists each assumption, its literature source, and a plausibility range; (b) conducting a simple uncertainty analysis (e.g., range reporting or tornado table) to show how sensitive total exposure estimates are to each assumption; and (c) explicitly flagging studies where exposure is likely strongly underestimated (short urine collection windows, incomplete monitoring).
5. Missing biomonitoring studies: the authors acknowledge excluding biomonitoring studies. Since biomonitoring could substantially inform exposure assessment and temporality, the exclusion should be justified more fully. Either (a) perform a complementary search for biomonitoring studies and summarize them briefly, or (b) outline clearly why such studies were outside scope and discuss limitations this imposes on conclusions about chronic exposure.
6. Temporal coverage and currency: many included observational studies predate 1991; the manuscript should make clearer how this affects relevance to current exposure scenarios (e.g., modern HFC/HFO replacements, PFAS sources). Explicitly compare historical exposure contexts (halothane era) to present environmental TFA sources and concentrations to avoid misleading inference.
7. Synthesis and interpretation: the conclusion that “no clinically relevant effects attributable to TFA were observed” is reasonable given the available evidence, but it risks downplaying possible long-term risks from chronic low-level exposure. Reframe conclusions to emphasize evidence gaps and recommend priority research (longitudinal biomonitoring, controlled clinical studies at environmentally relevant concentrations, mechanistic human toxicokinetics). Provide concrete recommendations for future research and regulatory monitoring.
8. Presentation of results: include a compact summary table that lists each included study with study design, population, exposure metric (most-probable and worst-case dose), outcomes measured, and quality rating. This will greatly improve usability for readers and policymakers. Additionally, ensure the PRISMA flow diagram is complete and legible (dates and reasons for exclusion clearly stated).

 

Minor comments

9. Clarify language: a few typographical and formatting issues are present (e.g., inconsistent use of “Trifluoractetic” vs. “Trifluoroacetic” in the title/abstract); correct spelling and standardize terminology.
10. Reference updates: include and discuss very recent regulatory and monitoring documents (EFSA PC-1508 consultation is cited — consider expanding discussion to its key points and how they relate to this review’s findings).
11. Limitations: expand the Limitations section to explicitly acknowledge potential publication bias, historical context bias (older studies), and the narrow inclusion criteria (exclusion of biomonitoring).
    12. Policy implications: since the topic has regulatory relevance, add a short paragraph on what the current evidence means for surveillance and risk communication (for example, whether existing drinking-water or food monitoring programs should be modified).

Author Response

Reviewer 2

Major comments

Comment 1: Scope and contribution: the review addresses an important and timely issue given growing environmental TFA levels. The work is clearly structured and the PECO framework is appropriate. Nonetheless, the manuscript would be strengthened by a clearer description of how this rapid review adds to existing assessments (e.g., EFSA draft and recent reviews cited) and by explicitly stating the review’s unique contribution in the Introduction and Conclusion.

Response 1:

Thanks for this idea. We have added to the introduction [page 1, paragraph 1, lines 37 – 38] and expanded the discussion to include context to existing regulations [page 14, paragraph 7, lines 360 – 387].

Comment 2: Search strategy and reproducibility: the PubMed/Embase search strategy is provided in Appendix A, but more transparency is required for reproducibility. Please provide (a) the full, exact search strings used for each database (including date stamps), (b) the deduplication method and software settings, and (c) the registration details for the protocol (the OSF link is noted — include the protocol ID and a direct reference). Also indicate whether grey literature or non-English studies beyond the stated languages were considered.

Response 2: Thank you for your comment. (a) We divided the Appendix A into A1 for the PubMed search string and A2 for the Scopus search string and added the dates [pages 16 and 17]; (b) We added details concerning the deduplication process [page 5, paragraph 1, lines 192 and 193]; (c) unfortunately we do not understand what you mean by “direct reference” concerning the protocol. Reference, DOI and Link are included [page 22, line 444 and 445]

Concerning the grey literature and the languages, we added some details [page 2, paragraph 9, lines 92-94]

Comment 3: Study selection and quality assessment: the review includes a heterogeneous set of study designs ranging from case reports to pharmacokinetic studies. While the chosen appraisal tools are acceptable, the manuscript should provide (a) a detailed table listing individual quality assessment scores for each included study (not only aggregate judgments), (b) justification for treating pharmacokinetic studies via “expert opinion” rather than an established tool, and (c) sensitivity analyses showing how conclusions would change if lower-quality studies were excluded.

Response 3:

Thank you very much for these suggestions, we have made some adjustments:

(a) Since the three tools use different scales—and not all of them are score-based—we only reported the aggregated results (Table 1). To ensure full transparency, we have now provided the individual quality assessment in the ANNEX.

(b) Given the considerable heterogeneity and the relatively old publication dates of the included studies, we considered it more appropriate to have them reviewed by an expert for potential errors and methodological shortcomings, rather than applying a generic Risk of Bias tool that may not be well-suited to each individual study. To our knowledge, no specific tool has been widely adopted for pharmacokinetic studies. Therefore, a flexible, case-by-case assessment by an experienced staff member of the Austrian medical market surveillance authority appeared to us to be a more reliable approach to evaluating study quality than rigid RoB tools [page 3, paragraph 5, lines 119 - 121]

(c) We do not summarize evidence or calculate risks; instead, we present the results exclusively in a narrative form. A weighting does not seem appropriate in this context. The quality assessment serves only as additional information, and we did not find any reason to exclude any of the included studies due to poor quality. In our view, a sensitivity analysis based on study quality does not make sense given the studies ultimately included.

 

Comment 4: Exposure estimation and uncertainty: substantial effort was made to estimate “most-probable” and “worst-case” TFA doses, but the assumptions (absorption fractions, metabolism rates, inhalation rates) introduce considerable uncertainty. I recommend (a) adding a dedicated subsection that lists each assumption, its literature source, and a plausibility range; (b) conducting a simple uncertainty analysis (e.g., range reporting or tornado table) to show how sensitive total exposure estimates are to each assumption; and (c) explicitly flagging studies where exposure is likely strongly underestimated (short urine collection windows, incomplete monitoring).

Response 4:

(a) We have described the assumptions underlying the TFA exposure estimation in detail in a dedicated section in the ANNEX indicating their literature source. A short summary can be found in Table 2 in the ANNEX.

(b) We have added a section in the ANNEX regarding uncertainty analysis. In Table 3, the difference that variation of one parameter makes to the TFA exposure estimation of the “most-probable” scenario is shown for several studies, which are representative for the different calculation methods.

(c) Sources for possible significant underestimation of TFA in certain studies were specifically mentioned in an additional paragraph in the ANNEX in the section Uncertainty Analysis.

Comment 5: Missing biomonitoring studies: the authors acknowledge excluding biomonitoring studies. Since biomonitoring could substantially inform exposure assessment and temporality, the exclusion should be justified more fully. Either (a) perform a complementary search for biomonitoring studies and summarize them briefly, or (b) outline clearly why such studies were outside scope and discuss limitations this imposes on conclusions about chronic exposure.

Response 5:

As already stated in the protocol, we excluded studies with unclear TFA exposure, as our aim was to associate a known or at least estimable exposure with its potential effects. Biomonitoring studies with unknown sources of exposure do not allow for such conclusions and, in our view, would have broadened the scope of the review too much. We agree that biomonitoring studies can provide valuable information but addressing them appropriately would require a separate review [page 14, paragraph 1, lines 355 and 356].

Comment 6: Temporal coverage and currency: many included observational studies predate 1991; the manuscript should make clearer how this affects relevance to current exposure scenarios (e.g., modern HFC/HFO replacements, PFAS sources). Explicitly compare historical exposure contexts (halothane era) to present environmental TFA sources and concentrations to avoid misleading inference.

Response 6:

Thank you for highlighting this important distinction. We have added two corresponding paragraphs in the Discussion [page 13, paragraph 3, lines 311 – 315; page 14, paragraph 6, lines 357 - 359]

Comment 7: Synthesis and interpretation: the conclusion that “no clinically relevant effects attributable to TFA were observed” is reasonable given the available evidence, but it risks downplaying possible long-term risks from chronic low-level exposure. Reframe conclusions to emphasize evidence gaps and recommend priority research (longitudinal biomonitoring, controlled clinical studies at environmentally relevant concentrations, mechanistic human toxicokinetics). Provide concrete recommendations for future research and regulatory monitoring.

Response 7:

Thank you for the comment. We do not intend to downplay any potential risks and have revised the Conclusions accordingly [page 15, paragraph 4, lines 396 – 399]

Comment 8: Presentation of results: include a compact summary table that lists each included study with study design, population, exposure metric (most-probable and worst-case dose), outcomes measured, and quality rating. This will greatly improve usability for readers and policymakers. Additionally, ensure the PRISMA flow diagram is complete and legible (dates and reasons for exclusion clearly stated).

Response 8: Thank you for the input. We created a summary table after your suggestion and added it in the Appendix B [pages 18 - 20]. We updated the PRISMA flow chart with exact data and exclusion criteria [page 6, lines 222 – 238]

 

Minor comments

Comment 9: Clarify language: a few typographical and formatting issues are present (e.g., inconsistent use of “Trifluoractetic” vs. “Trifluoroacetic” in the title/abstract); correct spelling and standardize terminology.

Response 9: Response 1: Thank you for pointing this out. We corrected this mistake in the title [page 1, line 2], added the spelling “Trifluoroacetic Acid” to the Keywords [page 1, line 21]. Furthermore, we adapted it for a consistent use on [page 1, paragraph 1, line 24], [page 1, paragraph 3, line 41], [page 1, paragraph 4, line 43], [table 2, page 9, row 6].

Comment 10: Reference updates: include and discuss very recent regulatory and monitoring documents (EFSA PC-1508 consultation is cited — consider expanding discussion to its key points and how they relate to this review’s findings).

Response 10:

We appreciate your suggestion regarding the inclusion of recent regulatory and monitoring developments. We have added a new paragraph that summarises the latest biomonitoring data in humans and provides an overview of both proposed and enacted drinking water limit values for TFA in various European countries [page 14, line 360-381].

Comment 11: Limitations: expand the Limitations section to explicitly acknowledge potential publication bias, historical context bias (older studies), and the narrow inclusion criteria (exclusion of biomonitoring).

Response 11:

Thank you for the suggestion. We have expanded the paragraph on limitations accordingly. (ad biomonitoring also see Response 5 and 8) [page 14, paragraph 6, line 355-359]

Comment 12:  Policy implications: since the topic has regulatory relevance, add a short paragraph on what the current evidence means for surveillance and risk communication (for example, whether existing drinking-water or food monitoring programs should be modified).

Response 12:

In our opinion our findings do not allow any policy implications. This would only apply if the studies had revealed new, previously unknown warning signals. The value of our work lies in the fact that, despite an extensive search, we did not identify any such signals. However, the reverse conclusion is equally inappropriate: these results do not allow for an all-clear. We have emphasized this point clearly in the Conclusions.

 

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors provided excellent and exhaustive responses to all comments