Low Serum Total Calcium: Influencing Factors of Heart Failure with Preserved Ejection Fraction (HFpEF)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Comments to the Authors

The authors report an association between hypocalcemia and lower rates of HF readmission, all-cause mortality and cardiovascular events in 9,322 patients classified as having HFpEF. The cohort is large and follow-up extends to 1,000 days. However, the exposure is uncorrected total calcium, which cannot be assumed to represent biologically meaningful hypocalcemia in a hospitalized heart failure population without albumin correction or measurement of the ionized fraction. The principal conclusion is also inconsistent with the all-cause mortality panel of Figure 2, and numerical discrepancies run through the abstract, text and tables. These are not editorial problems. They require verification of the exposure and outcome definitions against the source data and a full reanalysis.

(1) Calcium is reported as total calcium only, with no albumin in Table 1 or Table 2 and no mention of correction, although hypoproteinemia is listed as a covariate in Section 3.2. Hypoalbuminemia is common in hospitalized heart failure and lowers total calcium without lowering the ionized fraction. A prevalence of 36.6% raises the possibility that the exposure is substantially capturing hypoalbuminemia or acute illness rather than reduced ionized calcium. Parathyroid hormone, 25(OH) vitamin D, phosphate, magnesium, acid-base status and calcium or vitamin D supplementation are also absent; deferring supplementation to future work understates this. Please repeat the analysis with albumin-corrected calcium, and ionized calcium where available, and report albumin at baseline.

(2) The hypocalcemia group is older with higher creatinine and NT-proBNP and more COPD, coronary disease, pleural effusion and hypokalemia, yet has fewer events (mortality 885/3408, 26.0%, vs 2254/5914, 38.1%; readmission 1059/3408, 31.1%, vs 2884/5914, 48.8%). An effect of this size against a generally higher-risk baseline profile raises concern about outcome ascertainment or residual bias rather than establishing a protective effect.

The follow-up data add to this concern. The final three rows of Table 1 appear to be times in days, but neither their definition nor their units is given. For all-cause mortality the reported value is shorter in the hypocalcemia group (743.7 ± 313.4 vs 845.0 ± 286.3 days) despite a substantially lower crude mortality proportion. If these values represent overall observed follow-up, that combination could reflect earlier deaths among the smaller number of decedents, differential censoring, or both, and the present reporting does not allow these to be distinguished. If they instead represent time to death among decedents, they should be labelled accordingly and should not appear among baseline characteristics.

Follow-up was by telephone and community visit. Please define these rows and report loss to follow-up by group, median follow-up by reverse Kaplan-Meier with IQR, and whether vital status was verified against a registry independent of patient contact. If death was ascertained only by contacting the patient or family, differential loss to follow-up could materially bias the mortality estimate and, at the extreme, account for the reported direction.

(3) In the lower panel of Figure 2 the hypocalcemia curve appears to lie below the comparator over most of the follow-up, indicating worse survival, which is the opposite of the HR of 0.85 (0.78–0.92) in Table 4. The other two panels show the reverse. Please establish which is correct and reconcile figure, table and abstract.

Table 3 is directionally consistent with the dichotomous analysis, since hazard ratios above 1.0 below the inflection points imply rising risk with rising calcium. The estimates are nevertheless difficult to evaluate. Values of 10.75 and 16.93 are extreme, the scaling is unclear, and it is not stated whether they are crude or adjusted or how the inflection points (2.30, 2.17, 2.29 mmol/L) were derived. Please explain how these estimates relate to the dichotomous models and confirm that the same population, outcome coding and covariate structure were used. Section 3.2 cites spline plots as Figure 2, but Figure 2 is the Kaplan-Meier panel and no spline figure was submitted. These are needed to judge the shape of the association, particularly as patients with calcium >2.6 mmol/L were excluded and the upper range is therefore truncated.

The Table 3 footnote gives the SD as 1 mmol/L, whereas Table 2 reports 0.16 mmol/L. Raising the Table 2 hazard ratios to the power 0.16 returns 1.22, 1.14 and 1.26, which are the per-SD values reported in Table 3, so the increment used in the analysis appears to have been 0.16 mmol/L and the footnote to be in error. Please confirm, express the estimates per 0.1 mmol/L or per SD, and state whether they are crude or adjusted.

(4) The following require regenerated output rather than correction of the text.

Abstract and Section 3.2 give adjusted HRs of 0.69 (0.61–0.77), 0.85 (0.78–0.92) and 0.65 (0.61–0.70); Table 4 Model 2 gives 0.71 (0.66–0.76), 0.86 (0.80–0.94) and 0.66 (0.61–0.70).

The Model 2 covariates in Section 3.2 do not match the Table 4 footnote. Four of the five footnote variables appear in the text, but hemoglobin appears only in the footnote and twelve variables listed in the text are absent from it.

Section 2.1 states that no patients in the cohort had calcium >2.6 mmol/L, while Figure 1 excludes 256 such patients. Please confirm that this refers to the post-exclusion analytic cohort, and explain why hypercalcemic patients were excluded rather than analyzed as a separate stratum.

Section 3.2 cites Table 3 for the multivariable models, which are in Table 4.

The denominator for nitrendipine, given as 10 (20.00%), is unclear; 10 of 9,322 is 0.11%. A drug used in 10 patients should not enter variable screening in any case.

Section 3.1 lists torasemide, nitroglycerin and ezetimibe twice.

(5) Univariate screening at P<0.05 followed by backward elimination gives unstable models and understates uncertainty; established confounders should be retained on clinical grounds. Please report one final model consistently across text and tables, and test and report the proportional hazards assumption for calcium. All-cause death should be treated as a competing risk for HF readmission; for the cardiovascular composite, non-cardiovascular death is the competing event, assuming the analysis is based on time to first event. The manuscript should also state that HF readmission is a component of the cardiovascular composite, so the two should not be presented as independent confirmatory endpoints.

(6) Figure 1 excludes 12,189 patients for absent or incomplete echocardiography, more than the number analyzed, with no screened total given. Please report the total screened and how the excluded patients differed from those retained.

(7) The proposed mechanism is not supported by the present data. Lower extracellular calcium would be expected to reduce trigger calcium entry and SR loading, with potential adverse consequences for contractility, and it does not correct impaired SR reuptake or the other determinants of diastolic dysfunction invoked in the Introduction. Ionized calcium, intracellular calcium handling, SERCA2a activity, diastolic function and filling pressures were all unmeasured, so improved myocardial relaxation cannot be inferred from a lower total serum calcium concentration. Hypocalcemia is clinically associated with QT prolongation, arrhythmia and hypocalcemic cardiomyopathy. Cohorts reporting worse outcomes with low corrected calcium in heart failure should be cited and reconciled; the contrary literature is currently dismissed as an HFrEF phenomenon on a single review (ref. 14). The claim that HFpEF is a syndrome of systemic inflammation and microvascular endothelial dysfunction is attributed to ref. 7, which does not support it. Describing hypocalcemia as protective, and as a Knight, invites the inference that it need not be corrected.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,
I have read this study with care and interest. While I would like to emphasize that the study is interesting, there are some methodological issues that need correction and clarification. These are:
1. In the study, hypocalcemia is defined only through total serum calcium (<2.2 mmol/L). However, it is not specified whether albumin-corrected calcium or ionized calcium was used. This may affect the results, especially since hypoalbuminemia is frequently seen in heart failure patients. The calcium measurement method used and, if necessary, the correction formula should be detailed in the methods section.
2. The finding that hypocalcemia is an independent protective factor for heart failure re-admission, cardiovascular events, and mortality is not an expected result in the literature. Therefore, not only biological mechanisms but also residual confounding variables, selection bias, and alternative explanations that may arise from the observational study design should be discussed more comprehensively.
3. The details of which variables were included in the multivariate model and the criteria used, and the details of the backward elimination process are not clear. Furthermore, it is not specified whether the proportional hazards assumption was tested. This information should be added to the methods section to increase the reliability of the model.
4. The study includes important clinical features that can affect serum calcium, such as chronic kidney disease, coronary artery disease, and high NT-proBNP levels. If possible, subgroup analyses according to kidney function, gender, or age groups, or at least interaction analyses, would increase the clinical interpretability of the findings. These subgroup analyses would enhance the value of the study.

5. Please include a message for clinicians in the conclusion section of your article, avoiding generalizations and definitive statements.

Sincerely

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The authors report a single-center, retrospective cohort study of 9,322 patients hospitalized with heart failure with preserved ejection fraction (HFpEF) at Xiangtan Central Hospital between January 2014 and January 2024, of whom 3,408 (36.6%) had hypocalcemia (total serum calcium below 2.2 mmol/L) at first admission. Patients were followed for up to 1,000 days for three endpoints: heart failure readmission, all-cause mortality, and a composite cardiovascular-events outcome. Using Cox proportional-hazards models adjusted stepwise for demographic and clinical covariates, the authors report that hypocalcemia was independently associated with a lower hazard of all three endpoints (adjusted HR 0.69, 0.85, and 0.65, respectively). Kaplan-Meier curves and a piecewise linear threshold analysis are used to characterize the calcium-outcome relationship as non-linear. The authors position the finding as the identification of hypocalcemia as a novel protective prognostic marker in HFpEF, a direction they describe as opposite to calcium's expected effect in heart failure with reduced ejection fraction, and propose that lower extracellular calcium may ease cardiomyocyte calcium overload and reduce myocardial stiffness in this population.

After careful reading of the manuscript, I have the following comments/questions to the authors:

1. The manuscript's central claim, that hypocalcemia is an independent protective marker in HFpEF, runs opposite to a directly relevant prior study: Atluri et al. (2021, Journal of the American College of Cardiology, National Inpatient Sample analysis) reported hypocalcemia as an adverse prognostic marker in hospitalized heart failure patients. I suggest to cite and discuss this paper.

2. 

The Abstract and Section 3.2 report Model 2 adjusted HRs of 0.69 (95% CI 0.61-0.77) for HF readmission, 0.85 (0.78-0.92) for all-cause mortality, and 0.65 (0.61-0.70) for CV events. Table 4's "Model 2" column reports different values for the same three endpoints: 0.71 (0.66-0.76), 0.86 (0.80-0.94), and 0.66 (0.61-0.70). The discrepancy appears to trace to two different covariate lists: the Results paragraph describes a roughly sixteen-variable adjustment (systolic blood pressure, D-dimer, HbA1c, HDL-C, NT-proBNP, respiratory failure, hypoproteinemia, pleural effusion, CHD, pulmonary arterial hypertension, renal insufficiency, hypoxemia, paroxysmal supraventricular tachycardia, dexamethasone, prednisone), while the Table 4 footnote describes a five-variable set (hemoglobin, CHD, pulmonary arterial hypertension, renal insufficiency, hypoxemia) chosen by backward elimination. Both cannot be "Model 2."

3. The manuscript states that covariates with P < 0.05 in univariate analysis were carried into the multivariate model. Table 1 reports significant between-group differences for age, BMI, creatinine, LDL-C, Ln NT-proBNP, gender, hypertension, hydrothorax, diabetes, COPD, hypokalemia, and hypochloridemia, but Table 2, which is supposed to report the univariate Cox screen feeding that selection rule, shows only a subset of variables. Diabetes, BMI, creatinine, hydrothorax, hypokalemia, and hypochloridemia do not appear in Table 2 at all.

4. The Methods do not state whether serum calcium was measured as total or ionized calcium, or whether total calcium was corrected for albumin. This matters here specifically: total calcium falls when albumin falls (roughly 0.8 mg/dL of calcium per 1 g/dL drop in albumin), and HFpEF patients with congestive hepatopathy, malnutrition, or systemic inflammation are prone to hypoalbuminemia. Albumin is itself a recognized independent predictor of heart failure outcomes. If "hypocalcemia" in this cohort partly reflects lower albumin rather than a true ionized-calcium derangement, the reported protective association could be confounded or even reversed after correction.

5. Table 3 reports, for HF readmission, a per-SD hazard ratio of 10.75 (95% CI 7.24-15.98) below the 2.30 mmol/L breakpoint, and 16.93 (11.66-24.60) below 2.29 mmol/L for CV events. These values are several-fold larger than the overall continuous-calcium HRs in Table 2 (3.46 and 4.20 for the same two endpoints) and are larger than nearly any hazard ratio reported for a single continuous biomarker in the cardiovascular literature. No confidence intervals are given for the breakpoints themselves (2.30, 2.17, 2.29 mmol/L), and the narrow calcium range in this cohort (Table 1 standard deviations of roughly 0.1-0.16 mmol/L) makes a stable below-threshold slope estimate difficult to obtain from a piecewise linear Cox model.

6. The hypocalcemia group has meaningfully worse renal function (creatinine 117 vs. 95 umol/L), more pleural effusion, more COPD, and more coronary heart disease than the non-hypocalcemia group (Table 1), yet experiences better outcomes. My suggestion is to add a paragraph addressing reverse causation and confounding by indication as alternative explanations, and note in the Limitations that calcium was measured only once at admission.

7. Table 2 row label; Table 4 footnote: "Hyoxemia" is misspelled and should read "Hypoxemia"; the error recurs in more than one place.

8. Figure 2, all three panels: The x-axis of each Kaplan-Meier panel is labeled with the endpoint name ("CV events," "HF readmission," "All-cause mortality") rather than the variable actually plotted, which is follow-up time in days. Figure 2, all three panels: None of the three panels report numbers-at-risk under the curves. This is standard practice for Kaplan-Meier plots and would help readers judge how much of the late divergence between curves is driven by a shrinking risk set.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

1. The limitation of using total serum calcium is now acknowledged. However, please report how many patients had albumin or ionized calcium available and, if feasible, provide a sensitivity analysis using corrected or ionized calcium.
2. Adjustment for baseline covariates does not resolve potential differential follow-up or outcome ascertainment bias. Please report loss to follow-up by group, median follow-up, and how vital status was confirmed in patients who could not be contacted.
3. The explanation based on confounding does not fully reconcile the discrepancy between crude event rates, Kaplan–Meier findings, and adjusted Cox estimates. Please verify outcome coding, censoring, and group assignment and provide a consistent interpretation.
4. Several previously identified inconsistencies remain, including duplicated medications, the handling of hypercalcemic patients, and inconsistent model descriptions. These issues should be corrected individually and consistently throughout the manuscript.
5. The response states that univariate screening was abandoned and competing-risk analyses were performed, but the revised Methods still describe P<0.05 screening and Model 1/Model 2 analyses. Please ensure that the manuscript accurately reflects the final statistical analysis and report the proportional hazards assessment.
6. Please report the total number of patients screened and compare patients excluded because of missing echocardiographic data with those included in the analysis, as originally requested.
7. The manuscript still describes hypocalcemia as “protective” and retains the “Knight” terminology despite acknowledging that these expressions may be misleading. The unsupported mechanistic speculation should also be substantially reduced, and the conclusions should be framed as an observational association with low total serum calcium.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

I congratulate you on your efforts in revising the article text. I find the revisions valuable. I believe the quality of the article has improved and it is publishable.

Best Regards

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

I would like to thank the authors for responding in my comments/questions to a reconstructive way and this was improved the clarity of their manuscript. In the new table 5 in Heart failure readmission and Cardiovascular events columns the 95% upper and lower CI are identical with HR and should be corrected. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 3

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript has improved substantially after revision, particularly in the more cautious interpretation of total serum calcium, the clarification of outcome definitions and follow-up, and the acknowledgment of important limitations. However, several major methodological concerns raised previously remain unresolved. In particular, inconsistencies across the statistical methods, results, tables, and figures make the main findings difficult to interpret reliably.

1. The discrepancy between Figure 2 and the unadjusted mortality analysis remains unresolved. Figure 2 appears to show worse overall survival in the hypocalcemia group, whereas the unadjusted Cox model reports a lower mortality hazard (HR 0.85). Because both analyses are unadjusted, this cannot be explained by differences in covariate adjustment. The group labels, event coding, censoring, and generation of the survival curves should be carefully re-checked.

2. The final statistical approach remains unclear. The Methods describe Fine-Gray competing-risk models with prespecified covariates, whereas the Results still refer to Cox models and Model 1/Model 2 analyses, and Table 3 continues to mention backward elimination. The Abstract also refers to multivariable Cox analysis. These descriptions should be made consistent throughout the manuscript, and the analytical approach should be clearly specified for each outcome. All-cause mortality, in particular, should be analyzed separately because competing death does not apply to this endpoint in the same way as it does to nonfatal outcomes. This is especially important because the response letter states that the previous Model 1/Model 2 and variable-selection approaches were removed, yet they remain in the revised manuscript.

3. The clarification that the exposure was total serum calcium is helpful. However, it is still unclear whether serum albumin and ionized calcium were entirely unavailable or available in only a subset of patients. This should be stated explicitly. Without albumin-corrected or ionized calcium measurements, the findings should be interpreted as associations with total serum calcium rather than with hypocalcemia itself.

4. Several other issues from the previous review remain only partially addressed. The manuscript should explicitly state whether there was any loss to follow-up and clarify the follow-up/event-time variables currently shown in Table 1. Given the large number of patients excluded because of missing echocardiographic data, a comparison of included and excluded patients would be useful. If this is not feasible, this should be stated. In addition, although the causal language has been appropriately softened, the mechanistic discussion regarding extracellular calcium, intracellular calcium overload, myocardial relaxation, and filling pressure remains more definitive than the data support. This section should be shortened or clearly presented as hypothesis-generating.

Overall, these are not simply isolated reporting issues. The statistical analysis and presentation require comprehensive re-evaluation. A full re-analysis, with regeneration and careful cross-checking of the corresponding tables and figures, would be necessary before the findings can be reliably assessed.

Author Response

Comments 1: The discrepancy between Figure 2 and the unadjusted mortality analysis remains unresolved. Figure 2 appears to show worse overall survival in the hypocalcemia group, whereas the unadjusted Cox model reports a lower mortality hazard (HR 0.85). Because both analyses are unadjusted, this cannot be explained by differences in covariate adjustment. The group labels, event coding, censoring, and generation of the survival curves should be carefully re-checked.

 

Response 1: We sincerely apologize for the confusion caused by inconsistent statistical descriptions across the manuscript. We appreciate the opportunity to clarify our analytical framework. To ensure absolute methodological rigor and appropriateness for each specific type of clinical endpoint, we have systematically organized our statistical approach throughout the manuscript, Abstract, and tables as follows:

1. Standard Cox Proportional Hazards Models (Table 3):We utilized multivariable standard Cox proportional hazards models to analyze endpoints where competing mortality does not apply in the same manner, specifically for [All-cause mortality, heart failure readmission, cardiovascular events] (presented in Table 3).
2. Fine-Gray Competing-Risk Models (Table 5):For non-fatal clinical events where competing events can significantly impact the cumulative incidence (specifically heart failure readmission), we applied multivariable Fine-Gray competing-risk regression models, treating non-cardiovascular death as a competing risk and reporting subdistribution hazard ratios (SHRs) (presented in Table 5).
3. Manuscript-Wide Harmonization:We have conducted a thorough review of the entire manuscript. All outdated references (such as "Model 1/Model 2" and "backward elimination" in Table 3 or text) have been completely removed, and the descriptions in the Abstract, Methods, Results, and Table footnotes have been fully harmonized to reflect this clear two-tier analytical strategy.

Thank you for your expert guidance, which has greatly improved the methodological clarity of our study.

We added the Events-free percent of HF-readmission and CV-event as well as the probability of survival of all-cause mortality in Figure 2 now.

 

Comments 2: The final statistical approach remains unclear. The Methods describe Fine-Gray competing-risk models with prespecified covariates, whereas the Results still refer to Cox models and Model 1/Model 2 analyses, and Table 3 continues to mention backward elimination. The Abstract also refers to multivariable Cox analysis. These descriptions should be made consistent throughout the manuscript, and the analytical approach should be clearly specified for each outcome. All-cause mortality, in particular, should be analyzed separately because competing death does not apply to this endpoint in the same way as it does to nonfatal outcomes. This is especially important because the response letter states that the previous Model 1/Model 2 and variable-selection approaches were removed, yet they remain in the revised manuscript. The clarification that the exposure

Response 2: We sincerely apologize for our oversight and the confusion caused by inconsistent statistical descriptions across the manuscript. We deeply regret that despite our previous intention, outdated terms (such as "Model 1/Model 2", "backward elimination", and uncorrected Cox models for all endpoints) inadvertently remained in the Abstract, Results, and Table footnotes.To resolve this completely, we have conducted a thorough, manuscript-wide review and standardized our statistical descriptions. Specifically, we have implemented the following corrections: 1. Clear Distinction of Statistical Methods by Outcome (Methods Section):Table 3 shows Cox regression analysis, and Table 5 shows Fine-Gray competing-risk models.2. Elimination of Outdated erminology Across the Manuscript: We have removed all references to "Model 1 / Model 2" and "backward elimination" across the entire text, including the footnotes of Table 3. We have updated the Abstract and Results sections to consistently and accurately reflect our final, predefined multivariable modeling strategy without any contradiction.

 

 

Comments 3: The clarification that the exposure was total serum calcium is helpful. However, it is still unclear whether serum albumin and ionized calcium were entirely unavailable or available in only a subset of patients. This should be stated explicitly. Without albumin-corrected or ionized calcium measurements, the findings should be interpreted as associations with total serum calcium rather than with hypocalcemia itself.

 

Response 3: We sincerely thank the reviewer for this insightful comment. We completely agree that total serum calcium is influenced by serum albumin levels and that ionized calcium represents the physiologically active form. We have now explicitly clarified the availability of these laboratory measurements in the Methods and Limitations sections of the revised manuscript.

Specifically, we have clarified the data availability as follows:

Serum Albumin: We did not include serum albumin data. Ionized Calcium: Ionized calcium measurements were largely unavailable due to the retrospective, real-world nature of this study, as ionized calcium is not routinely recorded in standard clinical documentation for all hospitalized HFpEF patients. Reframing the Interpretation: In accordance with the reviewer's expert suggestion, we have carefully revised the manuscript text to strictly interpret our findings as associations with total serum calcium rather than definitive clinical hypocalcemia per se. We have also explicitly listed the lack of routine ionized calcium data as a study limitation. Thank you for helping us refine the precision and scientific rigor of our interpretation.

 

 

 

Comments 4: Several other issues from the previous review remain only partially addressed. The manuscript should explicitly state whether there was any loss to follow-up and clarify the follow-up/event-time variables currently shown in Table 1. Given the large number of patients excluded because of missing echocardiographic data, a comparison of included and excluded patients would be useful. If this is not feasible, this should be stated. In addition, although the causal language has been appropriately softened, the mechanistic discussion regarding extracellular calcium, intracellular calcium overload, myocardial relaxation, and filling pressure remains more definitive than the data support. This section should be shortened or clearly presented as hypothesis-generating.

 

Response 4: We sincerely thank the reviewer for these constructive and thorough comments. We have now fully addressed each of these remaining issues in the revised manuscript:

1. Loss to Follow-up and Table 1 Variables:

Loss to Follow-up: Because this is a retrospective cohort study based on electronic health records, traditional active "loss to follow-up" (such as patients dropping out of a prospective trial) does not apply. Instead, our study evaluated clinical outcomes through comprehensive electronic medical record tracking within our healthcare system.Specifically, patients who lacked subsequent clinical records or sufficient follow-up data were not included in the final analytical cohort during the initial data-cleaning phase. Therefore, all patients included in the final analysis had complete tracking of their clinical status through our system, and no active follow-up attrition occurred post-baseline. We have clarified this distinction in the Methods section of the revised manuscript to prevent any potential misunderstanding.Table 1 Variables: We thank the reviewer for pointing this out. We have updated and clarified the definitions of the follow-up and event-time variables in Table 1. Specifically, we have explicitly stated in the table footnote that follow-up time is presented as median with interquartile range (Median [Q1–Q3]), along with clear definitions for all time-to-event metrics.

2. Comparison of Included and Excluded Patients:

Conducting a comprehensive baseline comparison was not feasible because the 12,189 patients excluded due to missing echocardiographic data lacked routine echocardiographic evaluations and extensive clinical documentation in our electronic health system, preventing the extraction of a matched set of covariates for comparison. To transparently address this limitation, we have explicitly stated this constraint in the Discussion section of the revised manuscript. We clarified that this patient attrition is an inherent limitation of real-world retrospective studies and acknowledged its potential impact on the generalizability of our findings.

3. Shortening and Re-framing the Mechanistic Discussion:

In accordance with the reviewer's advice, we have substantially shortened the mechanistic discussion regarding extracellular/intracellular calcium handling, myocardial relaxation, and filling pressures. Furthermore, we have explicitly reframed these brief physiological considerations as strictly hypothesis-generating, emphasizing that our study is observational and that future translational and experimental studies are warranted to explore the underlying molecular pathways.

We deeply appreciate the reviewer's patience and rigorous guidance, which have ensured the scientific balance and clarity of our manuscript.