Association Between Frailty Scoring and Cardiopulmonary Exercise Testing: A Retrospective Cohort Study
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis is a well-conducted retrospective study examining an important clinical question regarding the relationship between Clinical Frailty Scale (CFS) scores and cardiopulmonary exercise testing (CPET) parameters. The manuscript is generally well-written and structured, though there are some areas that could be strengthened.
Major Strengths:
- Addresses a clinically relevant question with potential impact on resource utilization
- Clear methodology and statistical analysis
- Appropriate acknowledgment of study limitations
- Good presentation of results with supporting visual data
Major Weaknesses:
- Relatively small sample size
- Limited CFS score range (1-5 only)
- Single-center study design
- Incomplete outcome data for the full cohort
Specific Comments:
Introduction:
- Well-structured and provides appropriate context
- Clear rationale for the study
- "Current UK guidelines(5) recommend screening all patients with validated tools, such as the Duke Activity Status Index, Godin-Shepard Leisure-Time Exercise Questionnaire, or the International Physical Activity Questionnaire." Authors should also report that European Society of Anaesthesiology and Intensive Care guidelines also recommend to screen for these tools. Please discuss doi: 10.1097/EJA.0000000000002069
Methods:
- Statistical methods are appropriate
- Recommend including power calculation or sample size justification
- Clarify exclusion criteria more explicitly
- Consider explaining why only host site patients were included in length-of-stay analysis
Results:
- Clear presentation of findings
- Tables are informative and well-organized
- Suggest including baseline characteristics comparison between different CFS groups
- Consider adding multivariate analysis adjusting for potential confounders
Discussion:
- Good interpretation of findings in context of existing literature
- Appropriate acknowledgment of limitations
- Could elaborate more on clinical implications of findings
- Consider discussing potential future research directions more specifically
Figures/Tables:
- Well-presented and relevant
- Regression plots in Appendix A are helpful
- Consider adding a flow diagram of patient selection
Technical Corrections:
- Line 46: "Shuttle" is misspelled as "shuĴle"
- Line 98-99: Consider rephrasing the ethical approval statement for clarity
- Several instances of inconsistent spacing after periods
- Some references lack complete formatting
Author Response
Reviewer 1
Thank you for reviewing our work, we are most grateful.
Comment 1: Please discuss doi: 10.1097/EJA.0000000000002069
Response 1: Thank you for this addition. This is now included, line 57.
Comment 2: Recommend including power calculation or sample size justification
Response 2: Added line 78.
Comment 3: Clarify exclusion criteria more explicitly
Response 3: updated line 76.
Comment 4: Consider explaining why only host site patients were included in length-of-stay analysis
Response 4: The authors had approved access to length of stay data at the host site only. Added Line 108.
Comment 5: Suggest including baseline characteristics comparison between different CFS groups
Response 5: Not included in this revision due to time constraints for resubmission.
Comment 6: Consider adding multivariate analysis adjusting for potential confounders
Response 7: This will be completed within subsequent work by the authors.
Comment 7: Could elaborate more on clinical implications of findings
Response 7: Included within paragraph from line 203.
Comment 8: Consider discussing potential future research directions more specifically
Response 8: Included within paragraph from line 203.
Comment 9: Consider adding a flow diagram of patient selection
Response 9: Not included in this revision due to time constraints for resubmission.
Comment 10: Line 46: "Shuttle" is misspelled as "shuĴle"
Response: Amended previously
Comment 11: Line 98-99: Consider rephrasing the ethical approval statement for clarity
Response 11: Amended now line 111.
Comment 12: Several instances of inconsistent spacing after periods
Response 12: Amened and will be amended further as required.
Comment 13: Some references lack complete formatting
Response 13: Amended and will be amended further as required.
Reviewer 2 Report
Comments and Suggestions for AuthorsThe study included 174 patients. However, a larger and more diverse population would have increased the generalizability of the results.
The exclusion of patients with a frailty score above 5 means that the frailest group of patients was excluded from the analysis. This limits the applicability of the results to the frailest groups.
The exclusion of 5 patients due to missing data indicates that a missing data analysis is required. This is not reported in the study.
A prospective study could more reliably assess the effects of frailty score on CPET performance.
The fact that data were collected from only one hospital limits the generalizability of the results to other centers.
The assessment of frailty scores by the clinical team can be subjective. Standardization of the assessment is an important factor.
A low correlation coefficient indicates that the relationship between frailty score and CPET performance is limited. However, the effects of other variables were not taken into account.
Low R² values in linear regression models indicate that the explanatory power of the results is weak.
The lack of association between CFS and length of stay requires a subgroup analysis with a larger number of patients.
The study is insufficient to support the use of frailty scores instead of CPET. Alternative analysis methods (e.g., machine learning) may provide stronger results in this regard.
A prospective design with a larger sample size should be suggested. Additionally, the independent effect of frailty scores on postoperative complications should be investigated.
More details on data analysis (e.g., which R packages were used) should be provided.
Better visualization of the data may facilitate interpretation of the results.
Comments on the Quality of English Languageminor editing
Author Response
Reviewer 2
Thank you for reviewing our work, we are most grateful.
Comment 1: The study included 174 patients. However, a larger and more diverse population would have increased the generalizability of the results.
Response 1: This will be completed in future work.
Comment 2: The exclusion of patients with a frailty score above 5 means that the frailest group of patients was excluded from the analysis. This limits the applicability of the results to the frailest groups.
Response 2: Patients with CFS > 5 were not excluded however no patients included had a CFS> 5. This in our experience is typical for these cohorts as the frailest patients are not put forward for many major elective operations. However, we do appreciate that this limits the applicability and this is now included in the discussion.
Comment 3: The exclusion of 5 patients due to missing data indicates that a missing data analysis is required. This is not reported in the study.
Response 3: This represents 2.7% of the study population and was deemed low risk to introduce significant bias to the analysis.
Comment 4: A prospective study could more reliably assess the effects of frailty score on CPET performance.
Response 4: We agree with this statement.
Comment 5: The fact that data were collected from only one hospital limits the generalizability of the results to other centers.
Response 5: Data were collected from all 4 centres however CPET testing was done at a single site as per methods. We would argue that this is an advantage of this study as it improved reliability of CPET values.
Comment 6: The assessment of frailty scores by the clinical team can be subjective. Standardization of the assessment is an important factor.
Response 6: We fully agree. Added to discussion line 186.
Comment 7: A low correlation coefficient indicates that the relationship between frailty score and CPET performance is limited. However, the effects of other variables were not taken into account.
Comment 7: This will be completed within subsequent work by the authors.
Comment 8: Low R² values in linear regression models indicate that the explanatory power of the results is weak.
Response 8: We agree. This is reflected in the discussion.
Comment 9: The lack of association between CFS and length of stay requires a subgroup analysis with a larger number of patients.
Response 9: We agree. This is reflected in the discussion, line 189.
Comment 10: The study is insufficient to support the use of frailty scores instead of CPET. Alternative analysis methods (e.g., machine learning) may provide stronger results in this regard.
Response 10: We fully agree with the first statement and hope this is communicated in the discussion. The authors have no experience and therefore cannot comment on ML algorithm use in this setting.
Comment 11: A prospective design with a larger sample size should be suggested. Additionally, the independent effect of frailty scores on postoperative complications should be investigated.
Response 11: Included in the discussion, line 205.
Comment 12: More details on data analysis (e.g., which R packages were used) should be provided.
Response 12: Added line 110.
Comment 13: Better visualization of the data may facilitate interpretation of the results.
Response 13: We accept this.
Reviewer 3 Report
Comments and Suggestions for AuthorsThe purpose of this study is to evaluate local cohort data and look at the relationship between length of hospital stay, CPET results, and clinical frailty grading. However, there are certain queries/ suggestions need to be addressed before considering it for publication
1. Provide more details on the patient inclusion/exclusion criteria, any stratification methods, and the rationale for selecting the study period or hospitals.
2. The subgroup analysis (n=59) is mentioned but lacks an explanation of why this subgroup was selected and its significance.
3. Include a line outlining the study's limitations (such as its small sample size, retrospective design, or selection bias) and how the results could affect future research or clinical practice.
4. Provide a succinct description of the models' construction. Indicate which CPET variables were used, how confounders were taken into account, and why ordinal logistic regression was selected for this research.
5. Give more information about the data anonymization process and explain why ethical approval was not necessary. If applicable, make reference to institutional or local service evaluation requirements.
6. Link each statistical method directly to the research objectives. For instance, explain how the logistic regression results address the predictive utility of the CFS score for 1-year mortality.
7. Discuss the potential unmeasured variables (e.g., comorbidities, socioeconomic factors) that might account for the unexplained variance and how they could be incorporated in future studies.
8. While the limitations of low adjusted R^2 values are acknowledged, the broader limitations of using retrospective data or the potential for bias in clinician-assessed CFS scoring are not discussed.
9. For better readability, simplify difficult words and think about presenting correlations between CFS scores, CPET variables, and mortality statistics using visual aids (such as graphs or tables).
10. The statement "Our data suggests a weak signal between CFS score and CPET results" is clear but could benefit from a more detailed synthesis of the findings. Summarize key statistical results (e.g., correlation coefficients, significance levels) to reinforce the evidence for the "weak signal." This adds specificity and credibility to the conclusion.
11. Include a concise acknowledgment of major limitations, such as the retrospective design, small sample size, or variability in CFS scoring in conclusion section.
Author Response
Reviewer 3
Thank you for reviewing our work, we are most grateful.
Comment 1: Provide more details on the patient inclusion/exclusion criteria, any stratification methods, and the rationale for selecting the study period or hospitals.
Response 1: Included lines 71-77. No straitifciation methods were used.
Comment 2: The subgroup analysis (n=59) is mentioned but lacks an explanation of why this subgroup was selected and its significance.
Response 2: Included line 108.
Comment 3: Include a line outlining the study's limitations (such as its small sample size, retrospective design, or selection bias) and how the results could affect future research or clinical practice.
Response 3: Included within the discussion from line 185.
Comment 4: Provide a succinct description of the models' construction. Indicate which CPET variables were used, how confounders were taken into account, and why ordinal logistic regression was selected for this research.
Response 4: Now included; lines 94 onwards and 101.
Comment 5: Give more information about the data anonymization process and explain why ethical approval was not necessary. If applicable, make reference to institutional or local service evaluation requirements.
Response 5: Updated line 111.
Comment 6: Link each statistical method directly to the research objectives. For instance, explain how the logistic regression results address the predictive utility of the CFS score for 1-year mortality.
Response 6: Now included.
Comment 7: Discuss the potential unmeasured variables (e.g., comorbidities, socioeconomic factors) that might account for the unexplained variance and how they could be incorporated in future studies.
Response 7: Now included in the discussion. Line 172.
Comment 8: While the limitations of low adjusted R^2 values are acknowledged, the broader limitations of using retrospective data or the potential for bias in clinician-assessed CFS scoring are not discussed.
Response 8: Now included in the discussion line 186-191.
Comment 9: For better readability, simplify difficult words and think about presenting correlations between CFS scores, CPET variables, and mortality statistics using visual aids (such as graphs or tables).
Response 9: Not included in this revision due to time constraints for submission.
Comment 10: The statement "Our data suggests a weak signal between CFS score and CPET results" is clear but could benefit from a more detailed synthesis of the findings. Summarize key statistical results (e.g., correlation coefficients, significance levels) to reinforce the evidence for the "weak signal." This adds specificity and credibility to the conclusion.
Response 10: Included in conclusion.
Comment 11: Include a concise acknowledgment of major limitations, such as the retrospective design, small sample size, or variability in CFS scoring in conclusion section.
Response 11: Included in conclusion.
Reviewer 4 Report
Comments and Suggestions for AuthorsThe paper "Association between frailty scoring and cardiopulmonary exercise testing. A retrospective cohort study" is carefully read and reviewed. Cardiopulmonary exercise testing (CPET) is widely recognized as the gold standard for assessing functional capacity and predicting postoperative outcomes, particularly in high-risk surgeries. However, its availability is limited due to its resource-intensive nature. Simpler tools like the Clinical Frailty Scale (CFS) have been proposed as alternative predictors for perioperative outcomes, but their correlation with CPET metrics and clinical utility remains unclear. Authors aimed to study whether CFS could replace CPET.
Authors reported that CFS showed a weak relationship with functional capacity as measured by CPET, but it lacked sufficient predictive power to replace CPET or reliably estimate perioperative outcomes.
Several issues must be considered by the authors.
1. Preoperative period is associated with some degree of inflammation (International Journal of Colorectal Disease, 2016, 31: 1611-1617.). Frailty is also characterized with increased inflammatory burden (Bratislava Medical Journal/Bratislavské Lekárske Listy, 2021, 122.2.). Hence, CFS could assess the inflammation in patients undergoing surgery.
2. The cohort's mean CFS score (mean=3\text{mean} = 3) indicates a relatively low level of frailty, limiting the generalizability to populations with higher frailty scores. Discuss please.
3. Authors focused on CPET metrics, length of hospital stay, and 1-year mortality but did not assess other critical postoperative outcomes, such as complications or quality of recovery.
4. The retrospective nature of the study may introduce selection bias and confounding factors. Acknowledge please.
5. The cohort size (n=174) limits the statistical power of the findings, particularly for subgroup analyses (n=59n = 59). Larger datasets are necessary to validate these results. Acknowledge please.
Author Response
Reviewer 4
Thank you for reviewing our work, we are most grateful.
Comment 1: Preoperative period is associated with some degree of inflammation (International Journal of Colorectal Disease, 2016, 31: 1611-1617.). Frailty is also characterized with increased inflammatory burden (Bratislava Medical Journal/Bratislavské Lekárske Listy, 2021, 122.2.). Hence, CFS could assess the inflammation in patients undergoing surgery.
Response 1: Thank you for highlighting this. After discussion, the authors feel they have little knowledge of this subject matter and hence have not commented upon it within the manuscript.
Comment 2: The cohort's mean CFS score indicates a relatively low level of frailty, limiting the generalizability to populations with higher frailty scores. Discuss please.
Response 2: This is a recurring issue and similar to the points raised by other reviewers whereby the maximum CFS in this cohort =5. However this finding is in keeping with the surgical population referred in the authors experience. In that surgeons, especially for major procedures, rarely refer patients with CFS > 5. Discussed lines 189 onwards.
Comment 3: Authors focused on CPET metrics, length of hospital stay, and 1-year mortality but did not assess other critical postoperative outcomes, such as complications or quality of recovery.
Response 3: Assessing these was not possible with the dataset, budget and time restraints of the authors. However, we acknowledge this in the discussion.
Comment 4: The retrospective nature of the study may introduce selection bias and confounding factors. Acknowledge please.
Response 4: Added to the discussion, line 190.
Comment 5:The cohort size (n=174) limits the statistical power of the findings, particularly for subgroup analyses (n=59n = 59). Larger datasets are necessary to validate these results. Acknowledge please.
Response 5: Acknowledged within the discussion and conclusion.
Reviewer 5 Report
Comments and Suggestions for AuthorsThe article focuses on an important clinical problem by examining the relationship between cardiopulmonary exercise testing (CPET) and frailty scoring. The study was based on a retrospective analysis and many variables related to CPET were evaluated. The suggested corrections for the article are listed below:
- The use of a period in the title was not deemed appropriate. It is recommended to use ":" instead. For example: "Association between frailty scoring and cardiopulmonary exercise testing: A retrospective cohort study."
- The expression "UK 1;" in line 5 should be corrected.
- The references should be checked throughout the article. For example, in line 54, a period should come after the reference, and appropriate space should be left before the parentheses (for example: lines 36, 43, 63...).
- Subscripts should be used in some places. For example, lines 80, 81 and 124.
- The title in line 131 should start with a capital letter.
- It is recommended to support the article results with visuals to make them more fluent. For example, graphs presented in the Appendix can be placed within the article.
- Although the findings are related to the existing literature, only one reference study (O’Mahoney et al.) is cited. Comparison with more studies is needed.
- The conclusion section of the article should be better structured. The implications of the findings for clinical practice should be emphasized and suggestions for future studies should be provided.
Author Response
Thank you for reviewing our work, we are most grateful.
Comment 1: The use of a period in the title was not deemed appropriate. It is recommended to use ":" instead. For example: "Association between frailty scoring and cardiopulmonary exercise testing: A retrospective cohort study."
Response 1: Amended. Thank you.
Comment 2: The expression "UK 1;" in line 5 should be corrected.
Response 2: Amended.
Comment 3: The references should be checked throughout the article. For example, in line 54, a period should come after the reference, and appropriate space should be left before the parentheses (for example: lines 36, 43, 63...).
Response 3: Amended.
Comment 4: Subscripts should be used in some places. For example, lines 80, 81 and 124.
Response 4: Amended.
Comment 5: The title in line 131 should start with a capital letter.
Response 5: Amended.
Comment 6: It is recommended to support the article results with visuals to make them more fluent. For example, graphs presented in the Appendix can be placed within the article.
Response 6: Thank you. These have now been included in the article.
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsIt can be accepted
Reviewer 3 Report
Comments and Suggestions for AuthorsAll comments were addressed
