Blood Pressure and Pleth Variability Index as Predictors of Tourniquet-Release Hypotension in Elderly Patients Undergoing Total Knee Arthroplasty: A Prospective Observational Study
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe authors should be congratulated for addressing a clinically relevant topic.
Introduction
The authors overstate the potential role of PVi despite known limitations of this parameter in elderly, vasoplegic, or vasopressor-treated patients.
The introduction does not clearly define what gap in the literature this study uniquely fills.
Materials
Multiple PVi time points were tested without correction for multiple comparisons.
Blood pressure monitoring every 1–5 minutes is inadequate for capturing abrupt tourniquet-release hypotension.
Vasopressor administration and fluid management were left to clinician discretion, creating substantial confounding.
The sample size calculation was based on correlation analysis, not the logistic regression/ROC.
The “clinically relevant post-deflation window” is vague and should be precisely defined.
Results
The main finding is partly circular: patients with lower pre-release MBP are naturally more likely to fall below the offMBP threshold.
AUC values are only moderate (“Good”), despite strong wording in parts of the manuscript.
Discussion
The authors appropriately acknowledge some limitations, but the implications of post hoc variable selection and overfitting should be emphasized more strongly.
Conclusions
The conclusion should be toned down. The study is hypothesis-generating rather than clinically practice-changing. Claims regarding “risk stratification” are premature without external validation and standardized hemodynamic management protocols.
Author Response
Reviewer 1 - Comment 1:
The authors overstate the potential role of PVi despite known limitations of this parameter in elderly, vasoplegic, or vasopressor-treated patients.
Response: We agree. We have toned down the clinical claims regarding PVi and added a dedicated discussion on its limitations in geriatric, vasoplegic, or vasopressor-treated patients in both the Introduction and Discussion sections.
Revisions can be found in: Introduction, page 2, paragraph 3 (highlighted in yellow).
Revised text in the manuscript:
“However, the clinical utility of PVi in geriatric cohorts remains controversial, as age-related vascular stiffness, baseline vasoplegia, and the frequent administration of vasoactive medications can significantly alter peripheral perfusion waveforms and compromise the reliability of dynamic preload indices [15,16].”
Reviewer 1 - Comment 2:
The introduction does not clearly define what gap in the literature this study uniquely fills.
Response: We thank the reviewer for this constructive feedback. We have revised the Introduction to clearly define our study's primary contribution—namely, addressing the clinical gap in dynamic, noninvasive predictors for tourniquet-release hypotension in geriatric patients where standard static parameters are often insufficient.
Revisions can be found in: Introduction, page 2, paragraph 4 (highlighted in yellow).
Revised text in the manuscript:
“Although pneumatic thigh tourniquets are widely used in TKA, abrupt hemodynamic depression after deflation is highly prevalent and hazardous in geriatric patients due to physiological aging and cardiovascular comorbidities [2,3]. Traditional static hemodynamic monitors are often insufficient for predicting this sudden vasodilation-driven hypotension, and there is a critical clinical gap in simple, noninvasive, and dynamic predictors to stratify patient risk before tourniquet deflation. Our study addresses this clinical gap by evaluating a combined pre-release perfusion pressure and early dynamic volume index (intuPVi) bedside risk assessment framework under real-world clinical constraints.”
Reviewer 1 - Comment 3:
Multiple PVi time points were tested without correction for multiple comparisons.
Response: We agree that testing multiple time points without multiple-comparison adjustments increases the risk of type I errors. We have explicitly labeled these analyses as exploratory and hypothesis-generating in the revised Methods and Discussion sections, adding a clear warning to this effect.
Revisions can be found in: Methods, page 4, Section 2.6, and Discussion, page 10, paragraph 9 (highlighted in yellow).
Revised text in the manuscript:
“We also acknowledge that multiple PVi time points were evaluated without formal statistical correction (e.g., Bonferroni correction) for multiple comparisons. Consequently, these findings must be interpreted as exploratory and hypothesis-generating.”
Reviewer 1 - Comment 4:
Blood pressure monitoring every 1–5 minutes is inadequate for capturing abrupt tourniquet-release hypotension.
Response: We agree that standard intermittent NIBP monitoring represents a clinical limitation in capturing transient post-deflation nadirs. We have added a dedicated section in the Discussion explaining that intermittent monitoring every 1–5 minutes lacks the temporal resolution to capture rapid, brief pressure drops immediately following tourniquet deflation, which is precisely why establishing pre-deflation bedside predictors is clinically valuable.
Revisions can be found in: Discussion, page 10, paragraph 9 (highlighted in yellow).
Revised text in the manuscript:
“In addition, we acknowledge that standard intermittent NIBP monitoring (every 1–5 minutes) rather than continuous invasive arterial pressure monitoring represents a clinical limitation because it lacks the temporal resolution to capture very brief or rapid hemodynamic nadirs immediately following tourniquet release. However, because standard TKA is routinely performed under noninvasive monitoring in real-world clinical practice, this constraint is precisely what motivated our study. Since NIBP cannot continuously monitor post-deflation trends, establishing pre-release predictors (such as early intuPVi and pre-release MBP) provides a pragmatic bedside risk-stratification framework to proactively identify patients at risk before tourniquet deflation occurs.”
Reviewer 1 - Comment 5:
Vasopressor administration and fluid management were left to clinician discretion, creating substantial confounding.
Response: We agree that clinician-guided fluid and vasopressor administration introduces confounding. We have explicitly discussed this in the revised Discussion, acknowledging that the absence of a standardized hemodynamic management protocol is a key limitation that may mask or attenuate post-deflation hypotension.
Revisions can be found in: Discussion, page 10, paragraph 9 (highlighted in yellow).
Revised text in the manuscript:
“Furthermore, intraoperative fluid management and vasopressor administration were left to the discretion of the attending anesthesiologists. This lack of a standardized hemodynamic protocol introduces confounding, as proactive vasoactive administration could mask or attenuate tourniquet-release hypotension.”
Reviewer 1 - Comment 6:
The sample size calculation was based on correlation analysis, not the logistic regression/ROC.
Response: We have clarified in the revised Methods section that our sample size was calculated a priori using the correlation coefficient from pilot data as a primary surrogate of association. We have also added a statement in the limitations acknowledging that the study was not powered specifically for higher-dimensional multivariable logistic regression or ROC analysis.
Revisions can be found in: Methods, page 4, Section 2.6 (highlighted in yellow).
Revised text in the manuscript:
“We acknowledge that while this sample size was calculated a priori based on the correlation coefficient as a primary surrogate of association, it was not specifically powered for higher-dimensional multivariable logistic regression or receiver operating characteristic (ROC) analysis, representing a limitation in statistical depth.”
Reviewer 1 - Comment 7:
The “clinically relevant post-deflation window” is vague and should be precisely defined.
Response: We have updated the Methods section to precisely define the post-deflation window as 'within 10 minutes following tourniquet deflation'.
Revisions can be found in: Methods, page 3, Section 2.4 (highlighted in yellow).
Revised text in the manuscript:
“and the minimum systolic and mean blood pressures after tourniquet deflation (offSBP and offMBP) within a precisely defined post-deflation window of 10 minutes following tourniquet deflation.”
Reviewer 1 - Comment 8:
The main finding is partly circular: patients with lower pre-release MBP are naturally more likely to fall below the offMBP threshold.
Response: We agree that lower pre-release blood pressure and post-deflation hypotension are mathematically related. We have added text in both the Results and Discussion acknowledging this expected relationship, while explaining that including pre-release MBP as a covariate was essential to demonstrate that the dynamic index (intuPVi) provides independent and additive predictive information beyond the absolute blood pressure level immediately before deflation.
Revisions can be found in: Results, page 5, Section 3.2, and Discussion, page 8, paragraph 2 (highlighted in yellow).
Revised text in the manuscript:
“We acknowledge that the association between lower pre-release MBP and post-deflation hypotension is partly expected and mathematically related. However, including pre-release MBP in our multivariable model was essential to demonstrate that the dynamic index (intuPVi) provides independent and additive predictive information beyond the absolute blood pressure level immediately before deflation.”
Reviewer 1 - Comment 9:
AUC values are only moderate (“Good”), despite strong wording in parts of the manuscript.
Response: We agree. We have revised the manuscript to replace overstating adjectives regarding the model's performance with 'moderate' (AUC = 0.71) in the Abstract, Results, and Discussion sections, removing the word 'acceptable' from the AUC expression to ensure a conservative, objective tone.
Revisions can be found in: Abstract (page 1), Results (page 6, Section 3.2), and Discussion (page 9, paragraph 5) (highlighted in yellow).
Revised text in the manuscript:
“The discriminatory performance of the combined model was moderate, with an AUC of 0.71 (Table 3 and Figure 1).”
Reviewer 1 - Comment 10:
The authors appropriately acknowledge some limitations, but the implications of post hoc variable selection and overfitting should be emphasized more strongly.
Response: We agree. We have expanded the Discussion section to emphasize the risks of post hoc variable selection and overfitting given our modest sample size (90 patients and 26 hypotension events).
Revisions can be found in: Discussion, page 10, paragraph 9 (highlighted in yellow).
Revised text in the manuscript:
“With 26 primary hypotension events, our multivariable model is susceptible to overfitting, which may inflate the predictive performance. We evaluated multiple PVi time points without formal statistical adjustment (e.g., Bonferroni correction) for multiple comparisons, and selected intuPVi as the primary predictor based on preliminary analyses; this exploratory approach carries an inherent risk of post hoc variable selection bias.”
Reviewer 1 - Comment 11:
The conclusion should be toned down. The study is hypothesis-generating rather than clinically practice-changing. Claims regarding
Response: We agree that these findings should be interpreted as hypothesis-generating. We have revised the manuscript (Abstract, Discussion, and Conclusions) to heavily tone down all claims regarding clinical risk stratification. We have added explicit statements emphasizing that our findings must be interpreted cautiously, and that translating these predictors into clinical practice is premature without prospective validation in larger, standardized cohorts.
Revisions can be found in: Abstract (page 1, Conclusions), Discussion (page 10, paragraph 10), and Section 5 (Conclusions, page 11) (highlighted in yellow).
Revised text in the manuscript (Discussion, page 5):
“Despite these limitations, our findings suggest that pre-release MBP and PVi could serve as preliminary indicators to help identify elderly TKA patients who may be vulnerable to tourniquet-release hypotension. In a research context, this underscores the potential value of examining pre-release hemodynamic states rather than relying solely on post-deflation reactive measures. However, translating these findings into clinical practice—such as implementing proactive hemodynamic protocols or preemptively adjusting vasopressor dosing before tourniquet deflation—would be premature without prospective validation in larger, standardized cohorts.”
Reviewer 2 Report
Comments and Suggestions for AuthorsThe present manuscript offers limited contributions to the current literature, as both mean blood pressure (MBP) and the pleth variability index (PVi) are well-established clinical metrics. The physiological sequelae of tourniquet release characterized by abrupt systemic vasodilation and the liberation of metabolic byproducts are also extensively documented. The primary strength of the study lies in its utilization of routinely monitored, noninvasive markers to facilitate risk stratification without requiring additional equipment. However, the internal validity of these findings is constrained by a modest sample size of 90 patients. To establish the robustness of the results, the authors should have provided a formal a priori power analysis to justify the statistical strength of the cohort. Furthermore, the inclusion of only 10 references is insufficient for a peer-reviewed article, failing to provide the necessary depth for a comprehensive introduction or a rigorous discussion of the findings. The results remain largely descriptive and, given the single-center design and limited enrollment, lack the generalizability required for broader clinical application. Both the introduction and discussion sections require significant expansion through the integration of more extensive relevant literature. While the authors are transparent regarding the lack of conceptual novelty—describing the pre-release MBP findings as "mechanistically expected" and the overall study as "hypothesis-generating"—the manuscript lacks academic rigor and statistical depth. The conclusions provide a succinct summary of the observations and rightly acknowledge the substantial limitations regarding the study's broader utility.
Author Response
Reviewer 2 - Comment 1:
To establish the robustness of the results, the authors should have provided a formal a priori power analysis to justify the statistical strength of the cohort.
Response: We appreciate the opportunity to clarify this point. A formal a priori power analysis was indeed conducted (Methods, Section 2.6) showing a target sample size of 85, which we met by enrolling 90 patients. However, we fully agree that this modest sample size represents a limitation for higher-dimensional multivariable modeling and generalizability. We have highlighted the power analysis details and expanded the study limitations in the Discussion to reflect this constraint.
Revisions can be found in: Methods, page 4, Section 2.6, and Discussion, page 10, paragraph 9 (highlighted in yellow).
Clarified text in the manuscript:
“Based on a pilot study conducted at our institution in patients undergoing TKA, the correlation coefficient between the degree of blood pressure decrease at tourniquet release and PVi was 0.493. A G*Power analysis (Exact test, Correlation: bivariate normal model, a priori computation of required sample size, two tailed, correlation ρH1 = 0.493, α error probability = 0.01, power = 0.99, correlation ρH0 = 0) yielded a required sample size of 85 patients. Assuming a 5% dropout rate, we set the target enrollment at 90 patients. We acknowledge that while this sample size was calculated a priori based on the correlation coefficient as a primary surrogate of association, it was not specifically powered for higher-dimensional multivariable logistic regression or receiver operating characteristic (ROC) analysis, representing a limitation in statistical depth.”
Reviewer 2 - Comment 2:
Furthermore, the inclusion of only 10 references is insufficient for a peer-reviewed article, failing to provide the necessary depth for a comprehensive introduction or a rigorous discussion... Both the introduction and discussion sections require significant expansion through the integration of more extensive relevant literature.
Response: We agree. The reference list has been expanded from 10 to 24 peer-reviewed publications, and the relevant literature on geriatric vascular physiology, tourniquet deflation hemodynamics, and PVi monitoring limitations has been integrated throughout the Introduction and Discussion sections to provide appropriate academic depth.
Revisions can be found in: References section, pages 13-14, and throughout the Introduction and Discussion (highlighted in yellow).
Reviewer 3 Report
Comments and Suggestions for AuthorsThe authors of the manuscript titled "Clinical and Radiographic Outcomes of Fibula-Preserving Supramalleolar Osteotomy Combined with Arthroscopic Modified Broström Operation in Varus Ankle Osteoarthritis" descibe the potential of pleth variable index and pre-release blood pressure as a predictor of tourniquet-release hypotension in elderly patients undergoing total knee arthroplasty. The manuscript is well written and the results are clearly presented. The study addresses an interesting clinical question and the findings are potentially useful. I have a few points that should be addressed before publication.
- Although propofol is a common anesthesia used in TKA, have the authors compared other types of anesthetics as a predictor in tourniquet-release hypotension? Please include response in the "Discussion" section.
- Sex may also be a risk factor in tourniquet-release hypotension. Have the authors considered using sex as predictor? Please include response in the "Discussion" section.
Author Response
Reviewer 3 - Comment 1:
The reviewer titled the response: 'Clinical and Radiographic Outcomes of Fibula-Preserving Supramalleolar Osteotomy Combined with Arthroscopic Modified Broström Operation in Varus Ankle Osteoarthritis' but described the paper correctly as using PVi and pre-release BP to predict tourniquet-release hypotension...
Response: We sincerely thank the reviewer for their encouraging evaluation and highly positive feedback. We are grateful for their insightful comments on the clinical utility of pre-release blood pressure and PVi as predictors of tourniquet-release hypotension, which have helped us improve the manuscript.
Reviewer 3 - Comment 2:
Although propofol is a common anesthesia used in TKA, have the authors compared other types of anesthetics as a predictor in tourniquet-release hypotension? Please include response in the 'Discussion' section.
Response: We thank the reviewer for this excellent suggestion. All patients in our study received a standardized general anesthetic with propofol induction and volatile maintenance. We did not compare other techniques (such as propofol-based TIVA vs. volatile anesthesia, or general vs. regional anesthesia). We have addressed this in the revised Discussion (page 5, paragraph 3) by comparing the hemodynamic impacts of these techniques based on the literature and suggesting this as an important direction for future research.
Revisions can be found in: Discussion, page 10, paragraph 7 (highlighted in yellow).
Revised text in the manuscript:
“We did not compare different anesthetic techniques, as all patients in our cohort received a standardized general anesthetic protocol consisting of propofol induction and volatile maintenance. In the literature, the choice between total intravenous anesthesia (TIVA) with propofol and volatile anesthesia has shown conflicting results regarding the severity of tourniquet-release hypotension, with some studies suggesting propofol preserves vascular tone slightly better, while others report no clinical difference [20,21]. Furthermore, regional anesthesia (e.g., spinal or epidural anesthesia) is also commonly used in TKA and is generally associated with lower baseline vascular resistance due to sympathetic blockade, which may alter the hemodynamic response and the predictive accuracy of dynamic indices like PVi [22].”
Reviewer 3 - Comment 3:
Sex may also be a risk factor in tourniquet-release hypotension. Have the authors considered using sex as predictor? Please include response in the 'Discussion' section.
Response: We agree that sex represents a potential predictor of interest. In our cohort, the sex distribution was identical between the hypotension and non-hypotension groups (p = 1.000, Table 1), indicating that sex was not a significant predictor in this specific sample. The high prevalence of female patients (86.7%) reflects the epidemiological characteristics of knee osteoarthritis in South Korea, which is highly prevalent in older women. As requested, we have added a dedicated paragraph in the Discussion section detailing these sex-based findings and epidemiological characteristics.
Revisions can be found in: Discussion, page 10, paragraph 8 (highlighted in yellow).
Revised text in the manuscript:
“We also evaluated the potential of sex as a predictor of tourniquet-release hypotension. In our cohort, sex distribution was identical between the hypotension and non-hypotension groups (p = 1.000, Table 1), indicating that sex was not a significant predictor in this cohort. The high prevalence of female patients (86.7%) in our study reflects the established epidemiological characteristics of knee osteoarthritis in South Korea, which is highly prevalent in older women [23,24]. Although sex-specific physiological differences in vascular compliance and autonomic responses could theoretically influence perioperative hemodynamic shifts, our study was not powered to detect subtle sex-based differences due to the small proportion of male participants.”
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsThe authors have substantially improved the quality of their manuscript. They have thoroughly acknowledged the study's inherent limitations and appropriately moderated their conclusions. Furthermore, the reference list has been significantly expanded from 10 to 24, incorporating a comprehensive selection of recent literature on the topic. Thus, I believe the manuscript is now suitable for publication.
