Pericoronary, but Not Epicardial, Cardiac Fat Thickness Is Associated with Sarcopenia in Hospitalized Older Adults

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Several issues need clarification or correction:

1: the study design and outcome definition need sharper alignment. Sarcopenia was “already established clinically” and then used as a predefined outcome for imaging analyses. As far as I understand, the paper is not evaluating a contemporaneous, prospectively assessed imaging biomarker against a standardized index test performed under the same protocol, but rather testing association with an existing clinical classification assembled from retrospective data.

That distinction should be made explicit from the outset and in the title/abstract/discussion, because the current framing occasionally drifts toward diagnostic utility rather than association.

2: the timing window between CT and BIA is wide. Inclusion allowed CT and bioelectrical impedance analysis within 3 months, regardless of which came first. In hospitalized older adults, body composition and functional status can change materially over that interval, especially in the context of acute illness, immobilization, nutritional fluctuation, or deconditioning. This introduces temporal misclassification that directly affects the central exposure-outcome relationship and should be discussed more explicitly as a limitation.

3: the measurement approach for PCAT and EAT raises reproducibility concerns. A senior radiologist reviewed all datasets, but the manuscript does not report intraobserver or interobserver reproducibility, nor does it provide enough detail to assess measurement robustness in a non-ECG-gated thoracic CT setting. This is particularly important because PCAT was measured as a linear thickness across RCA, LAD, and LCx territories rather than by volumetric or attenuation-based techniques, while the discussion itself recognizes susceptibility to motion artifacts and segment-selection variability. Without reproducibility data, the reliability of the central exposure variable remains uncertain.

4: the choice of confounder adjustment appears incomplete for the main claim. The final main model adjusts for age, sex, and BMI, but baseline characteristics show that diabetes mellitus and body weight differed significantly between groups, and several other comorbidities are clinically plausible confounders in a study of adiposity and sarcopenia. The manuscript mentions a more fully adjusted Model D only as a sensitivity analysis because EPV dropped below 10. This is understandable statistically, but it also means the principal conclusion of “independent association” should be stated more cautiously, because residual confounding remains a real possibility.

5: the diagnostic language is too strong relative to the actual performance. The ROC results are only modest, with AUC 0.637 in the full cohort and 0.715 in the BMI ≥25 subgroup, and the gain in AUC after adding PCAT did not reach significance by DeLong testing, while IDI also did not cross the conventional significance threshold. These results support possible incremental information, but not a robust diagnostic tool. The manuscript should avoid implying that a clinically useful threshold has been established. At most, the data support exploratory signal detection.

6: the BMI-stratified interpretation should be toned down further. The subgroup with BMI ≥25 showed a significant association, but the interaction test was not significant. Therefore, claims that PCAT is especially informative in overweight or obese patients should remain clearly hypothesis-generating. At present, parts of the abstract, discussion, and conclusion still give this subgroup finding more weight than the formal statistics support.

7: the manuscript has substantial missingness in secondary geriatric variables, ranging from 28% to 49%, handled by complete-case analysis. Even if these variables were not included in the primary regression models, this weakens the descriptive geriatric profiling of the cohort and limits confidence in broader mechanistic interpretations linking PCAT to frailty-related constructs beyond the defined sarcopenia outcome.

8: there is some overinterpretation in the discussion. Several mechanistic statements about localized inflammatory signaling, vascular oxidative stress, and systemic muscle catabolism are biologically plausible, but they are not directly tested in this study. Since the paper measures only CT thickness and clinical/body composition variables, the mechanistic language should be framed more explicitly as hypothesis-based interpretation rather than inferred demonstration.

9: the conclusion section appears incomplete. The final sentence ends with “these subgroup findings require” and does not finish the thought. This must be corrected.

10: the manuscript would benefit from editorial cleanup. There are several punctuation and formatting inconsistencies, including stray apostrophes, spacing issues, and minor language problems. These do not invalidate the science, but they do affect readability and polish. Examples appear in the methods and results sections.

Author Response

Dear Editor

We would like to thank you and reviewers for taking your precious time to read our manuscript and to provide feedback.

The related changes and refinements have been made in the revised paper by taking into consideration your valuable comments. You can find the mentioned amendments below. I hope our revision has improved the paper to a level of your satisfaction.

Sincerely.

Corresponding author; Dr.Erdoğan

 

Comment 1: the study design and outcome definition need sharper alignment. Sarcopenia was “already established clinically” and then used as a predefined outcome for imaging analyses. As far as I understand, the paper is not evaluating a contemporaneous, prospectively assessed imaging biomarker against a standardized index test performed under the same protocol, but rather testing association with an existing clinical classification assembled from retrospective data.

That distinction should be made explicit from the outset and in the title/abstract/discussion, because the current framing occasionally drifts toward diagnostic utility rather than association.

Response 1: Thank you for this important explanation. We agree. Our study is a single center, retrospective observational association analysis: sarcopenia status was determined from routine comprehensive geriatric assessment data (EWGSOP2 components) and then used as a predefined outcome for epicardial and pericoronary fat thickness analyses derived from CT. The study was not designed as a prospective diagnostic validation of an imaging biomarker against a simultaneous standardized reference test under a combined protocol. Therefore, we revised the text to consistently reflect an association based framework and removed statements implying diagnostic validation or establishing a clinically applicable threshold. This explanation has been incorporated into the abstract,introduction, materials and methods, results, discussion ,conclusions, and sections of the revised text as follows:

‘’This study investigated the associations of EAT and PCAT thickness with sarcopenia’’in Abstract section- Background and Objectives page :1 line:23

‘’This is a retrospective observational study using CT data obtained for clinical purposes and routine geriatric assessment data’’in Abstract section- Materials and methods page :1 line:25

‘’PCAT thickness measured on routine non-contrast thoracic CT was associated with sarcopenia, whereas EAT thickness was not. ROC analyses showed modest discrimination; therefore, any proposed cut-offs should be considered exploratory and require prospective validation and external confirmation before clinical use.’’ in Abstract section- Conclusions page :2 line:41

‘’Notwithstanding this rationale, simultaneous evaluation of EAT and PCAT in relation to sarcopenia in geriatric inpatients remains limited. Therefore, we conducted a retrospective observational study to examine the association between CT-derived EAT/PCAT thickness and sarcopenia defined using EWGSOP2 components from routine comprehensive geriatric assessment. We also explored the discriminative performance of these measures within clinical models, without aiming to establish a definitive diagnostic threshold.’’ in İntroduction section page :3 line:92

‘’Muscle mass was classified as low at SMMI values below 9.2 kg/m² in men and below 7.4 kg/m² in women, following the Turkish-population reference values proposed by Bahat et al. [20]." in Materials and Methods section -2.4. Muscle Assessment  page :5 line:172

Grip weakness was defined as a peak value below 32 kg in men and below 22 kg in women, using the Turkish-population reference cutoffs of Bahat et al. [20]. in Materials and Methods section -2.4. Muscle Assessment )page :6 line:178

"Sarcopenia status had been established clinically according to the EWGSOP2 algorithm [1], applied with Turkish-population reference cutoffs from Bahat et al. [20]. Under EWGSOP2, isolated weakness indicates probable sarcopenia, combined low strength and reduced muscle mass yield confirmed sarcopenia, and the additional finding of reduced walking speed defines severe sarcopenia [1]. BIA-derived SMMI cut-points were used to characterize low muscle mass in accordance with these criteria. Muscle mass (BIA-derived SMMI) was available in all 101 patients. Handgrip dynamometry yielded numeric values in 89 of 101 patients (88.1%); in the remaining 12 patients (11.9%), grip dynamometry could not be performed because of severe cognitive impairment, ICU-acquired weakness, or hand pathology, and these patients were operationally classified as having low strength in the retrospective algorithm. This approach reflects clinically meaningful inability to complete strength assessment but may introduce outcome misclassification; concordance and sensitivity analyses were therefore performed and are reported in the Supplementary Materials. The five-times sit-to-stand test (n=44) served as a corroborative strength measure. Physical performance (4-m gait speed, SPPB) was used for severity grading when available but did not affect the binary sarcopenia outcome used in this study." "in Materials and Methods section -2.4. Muscle Assessment  page :6 line:184

"Among secondary variables used in comprehensive geriatric assessment (Katz, Lawton, MNA, and GDS), missing data ranged from 28% to 49% and were handled using complete-case analysis for each analysis. The availability and operational handling of EWGSOP2 muscle strength components (handgrip and 5×sit-to-stand) are described in Section 2.4."  in Materials and Methods section -2.5. Statistical Analysis page: 7    line: 241

" The algorithm reproduced the recorded clinical sarcopenia label in 99 of 101 patients (98.0%; Cohen's κ = 0.960, indicating very high agreement). Sensitivity analyses using the algorithm-concordant subset (n=99), an algorithm-defined outcome variable (n=101), additional adjustment for diabetes, and CT–BIA intervals restricted to ≤60 days (n=58) or ≤30 days (n=46) yielded materially similar PCAT effect estimates (adjusted ORs ranging from 1.36 to 1.52; all p<0.05; Supplementary Figure  4. in Results  section page: 14 line: 352

Supplementary Figure 4.  The relationship between CT and BIA intervals.

Because of the retrospective design and non-contemporaneous clinical assessments, our results should be interpreted as associations rather than as a formal diagnostic validation of an imaging biomarker. in discussions page :15 line 379

In hospitalized older adults, PCAT thickness measured on routine non-contrast thoracic CT was associated with sarcopenia after adjustment for age, sex, and BMI, whereas EAT thickness was not.”  In conclusions section  page : 20 line :572

"Nonetheless, the formal interaction test was not statistically significant, and these subgroup findings require confirmation in larger prospective cohorts with standardized sarcopenia ascertainment and direct measures of adiposity."  In conclusions section  page : 20 line :578

 

Comment 2: the timing window between CT and BIA is wide. Inclusion allowed CT and bioelectrical impedance analysis within 3 months, regardless of which came first. In hospitalized older adults, body composition and functional status can change materially over that interval, especially in the context of acute illness, immobilization, nutritional fluctuation, or deconditioning. This introduces temporal misclassification that directly affects the central exposure-outcome relationship and should be discussed more explicitly as a limitation.

Response 2: We thank the Reviewer for highlighting the importance of the CT–BIA interval. We agree that temporal misclassification is a legitimate concern in this population and have therefore (a) quantified the interval distribution in the analytic cohort, (b) tested the stability of the PCAT–sarcopenia association across progressively shorter interval windows, and (c) expanded the Limitations to acknowledge this issue. The findings are summarized below.

"Inclusion allowed up to 90 days between thoracic CT and BIA, and in hospitalized older adults body composition can change materially over that interval through acute illness, immobilization, nutritional fluctuation, or deconditioning. The median CT–BIA interval in this cohort was 36 days (IQR 10–81), and the interval did not differ by sarcopenia status. Sensitivity analyses restricted to ≤60 days (n=58) and ≤30 days (n=46) preserved the PCAT–sarcopenia association with materially similar effect estimates (adjusted OR 1.48 and 1.52 respectively, both p<0.05), arguing against attenuation by temporal misclassification. Nevertheless, prospective designs with simultaneous imaging and body-composition assessment would be required to fully eliminate this concern. in discussions page : 19 line :538

Comment 3 : the measurement approach for PCAT and EAT raises reproducibility concerns. A senior radiologist reviewed all datasets, but the manuscript does not report intraobserver or interobserver reproducibility, nor does it provide enough detail to assess measurement robustness in a non-ECG-gated thoracic CT setting. This is particularly important because PCAT was measured as a linear thickness across RCA, LAD, and LCx territories rather than by volumetric or attenuation-based techniques, while the discussion itself recognizes susceptibility to motion artifacts and segment-selection variability. Without reproducibility data, the reliability of the central exposure variable remains uncertain.

Response 3: Thank you for this important comment. We have expanded the CT measurement description to improve transparency (standardized landmarks, 0.625-mm thin-slice reconstructions, and exclusion of oblique/nonperpendicular planes on non–ECG-gated scans). We also performed an intraobserver reproducibility assessment by remeasuring a stratified random subset of 30 scans after ≥2 weeks, blinded to initial values. Reproducibility was excellent (ICC(2,1)=0.968 for PCAT and 0.965 for EAT; both 95% CI 0.93–0.98), and these results are now reported in the Materials and Methods .

‘’All thoracic CT measurements were performed by a single senior radiologist using a standardised protocol. PCAT thickness was measured on 1-1.5 mm thin slice reconstructions as the mean of linear measurements across the RCA, LAD, and LCx territories; EAT was measured as the maximal perpendicular thickness anterior to the right ventricular free wall. Oblique or nonperpendicular sections were excluded to limit motion related error in these non ECG gated examinations. To assess measurement robustness, the reader remeasured a stratified random subset of 30 scans after an interval of at least two weeks, blinded to the initial values; the intraobserver intraclass correlation coefficients (ICC[2,1], two-way random-effects, absolute agreement) were 0.968 (95% CI 0.93-0.98) for PCAT and 0.965 (95% CI 0.93-0.98) for EAT ‘’ in Materials and Methods section -2.3. Chest Computed Tomography page :4 line : 132

‘’Measurements were performed by a single radiologist; although intraobserver reproducibility was excellent (ICC 0.965-0.968), interobserver reproducibility could not be assessed in this retrospective cohort, and this remains a limitation. PCAT was assessed as a linear thickness on non ECG gated thoracic CT rather than by volumetric or attenuation based techniques, which are more susceptible to motion artefact and segment-selection variability. A standardised protocol with thin-slice reconstruction, multi territory averaging, and exclusion of oblique sections was used to mitigate these sources of error; nevertheless, prospective studies with standardised, ideally ECG-gated acquisition and full intra and interobserver reproducibility testing are warranted.’’ in discussion section page :18 line 511

Comment 4: the choice of confounder adjustment appears incomplete for the main claim. The final main model adjusts for age, sex, and BMI, but baseline characteristics show that diabetes mellitus and body weight differed significantly between groups, and several other comorbidities are clinically plausible confounders in a study of adiposity and sarcopenia. The manuscript mentions a more fully adjusted Model D only as a sensitivity analysis because EPV dropped below 10. This is understandable statistically, but it also means the principal conclusion of “independent association” should be stated more cautiously, because residual confounding remains a real possibility.

Response 4: Thank you for this important point. We agree that residual confounding is possible given clinically plausible covariates (e.g., diabetes and other comorbidities) that may influence both regional adiposity and sarcopenia. Our primary models were intentionally parsimonious (age, sex, BMI) to maintain adequate events-per-variable (EPV) and avoid overfitting in a cohort of 101 participants. As noted, a more comprehensively adjusted model (Model D) reduced EPV below the conventional threshold and was therefore presented as a sensitivity analysis. This explanation has been incorporated into  the revised text as follows:

In a diabetes-adjusted sensitivity model (Model E: age, sex, BMI, diabetes), the association was essentially unchanged (adjusted OR 1.39, 95% CI 1.12-1.71, p=0.003); diabetes was not an independent predictor (OR 0.47, 95% CI 0.18-1.18, p=0.107), and the likelihood-ratio test versus the primary model was non significant (chi-square=2.61, p=0.106). in Results section 3.1.4. Main Analysis: Hierarchical Regression Models,page :11 line :310

Comment 5: the diagnostic language is too strong relative to the actual performance. The ROC results are only modest, with AUC 0.637 in the full cohort and 0.715 in the BMI ≥25 subgroup, and the gain in AUC after adding PCAT did not reach significance by DeLong testing, while IDI also did not cross the conventional significance threshold. These results support possible incremental information, but not a robust diagnostic tool. The manuscript should avoid implying that a clinically useful threshold has been established. At most, the data support exploratory signal detection.

Response 5:

Thank you for this important comment. We agree that our ROC performance is modest and does not support a robust diagnostic tool or a clinically established threshold. We have revised the manuscript to (i) remove or temper diagnostic/predictive wording, (ii) explicitly report that the AUC improvement did not reach statistical significance by the DeLong test and that IDI did not meet conventional significance thresholds, and (iii) frame ROC-derived cut-offs as exploratory and hypothesis-generating, requiring external validation before any clinical application. This explanation has been incorporated into the revised text as follows:

‘’PCAT alone yielded an AUC value of 0.637. To formally evaluate the additional discriminative value of PCAT, the AUC of the baseline model including age, sex, and BMI (AUC=0.667) was compared with that of the PCAT-enhanced model (AUC=0.754). Although adding PCAT to the baseline model increased the AUC from 0.667 to 0.754, this increase was not statistically significant according to the DeLong test (ΔAUC=0.087, p=0.083). The likelihood ratio test confirmed that PCAT significantly improved model fit (χ²=11.68, p<0.001). However, the fact that this significant improvement in model fit did not translate into a statistically significant gain in AUC-level discrimination suggests that, although PCAT statistically contributes to the prediction equation, its ability to provide clinically meaningful reclassification at the individual patient level remains modest in the current sample size. Accordingly, the incremental discriminative value of PCAT should be interpreted not as a definitive finding, but rather as a supportive signal. The integrated discrimination improvement was 0.104 (95% CI −0.012 to 0.220; p=0.080; bootstrap SE=0.059, 2000 resamples), indicating a trend toward improved discrimination that did not reach the conventional threshold for statistical significance’’ in discussions section page : 17 line : 456

Comment 6: the BMI-stratified interpretation should be toned down further. The subgroup with BMI ≥25 showed a significant association, but the interaction test was not significant. Therefore, claims that PCAT is especially informative in overweight or obese patients should remain clearly hypothesis-generating. At present, parts of the abstract, discussion, and conclusion still give this subgroup finding more weight than the formal statistics support.

Response 6: Thank you for this important point. We agree. Although the association between PCAT thickness and sarcopenia was statistically significant in the BMI ≥25 kg/m² group, the formal BMI × PCAT interaction test was not statistically significant; this indicates that no effect modification based on BMI can be concluded from our data. Therefore, to ensure that the findings stratified by BMI are presented as exploratory/hypothesis-forming and to avoid overemphasizing the overweight/obese subgroup according to the overall results, it has been edited in the revised text as follows:

‘’ In an exploratory BMI-stratified analysis, the PCAT-sarcopenia association appeared numerically larger in participants with BMI >=25 kg/m2 (OR 1.48, p=0.008). However, the formal BMI-by-PCAT interaction was not statistically significant (p=0.290); this subgroup observation is hypothesis-generating only and requires confirmation in adequately powered prospective studies. in discussions section ; page: 16 line :418

Comment 7: the manuscript has substantial missingness in secondary geriatric variables, ranging from 28% to 49%, handled by complete case analysis. Even if these variables were not included in the primary regression models, this weakens the descriptive geriatric profiling of the cohort and limits confidence in broader mechanistic interpretations linking PCAT to frailty-related constructs beyond the defined sarcopenia outcome.

Response 7:

Thank you for this important observation. We agree that the degree of missingness in several secondary geriatric assessment variables limits the completeness of cohort profiling and constrains interpretation of relationships between PCAT and broader frailty-related constructs beyond the operational sarcopenia outcome. In the current dataset, missingness was primarily confined to secondary comprehensive geriatric assessment measures; variables required for the prespecified primary models (age, sex, BMI, EAT/PCAT, and muscle mass indices used for sarcopenia classification) had negligible missingness. Accordingly, we used complete-case analysis for descriptive secondary variables and explicitly reported the available sample size for each analysis. We have now strengthened the manuscript by (i) clearly stating the extent of missingness, (ii) clarifying that these measures were not included in the primary regression models, and (iii) adding a dedicated limitation noting that missingness reduces confidence in broader mechanistic/frailty extrapolations.

‘’Among secondary variables used in comprehensive geriatric assessment (Katz, Lawton, MNA, and GDS), missing data ranged from 28% to 49% and were handled using complete-case analysis for each analysis. The availability and operational handling of EWGSOP2 muscle strength components (handgrip and 5×sit-to-stand) are described in Section 2.4." in Materials and Methods section 2.5. Statistical Analysis ; page :7 line : 241

Comment 8: there is some overinterpretation in the discussion. Several mechanistic statements about localized inflammatory signaling, vascular oxidative stress, and systemic muscle catabolism are biologically plausible, but they are not directly tested in this study. Since the paper measures only CT thickness and clinical/body composition variables, the mechanistic language should be framed more explicitly as hypothesis-based interpretation rather than inferred demonstration.

 

Response 8: Thank you for this helpful comment. We agree that our study did not directly assess biological mechanisms (e.g., inflammatory signaling, oxidative stress, or causal pathways of muscle catabolism) and that mechanistic language should therefore be presented as biologically plausible hypotheses rather than demonstrated effects. We revised the Discussion to (i) remove causal phrasing, (ii) explicitly state that mechanistic pathways were not measured, and (iii) frame these interpretations as hypothesis-generating contextualization supported by prior literature.

‘’Importantly, this study was not designed to test mechanistic pathways. We did not measure inflammatory biomarkers, oxidative stress markers, or adipokine profiles, and therefore any mechanistic interpretation should be considered hypothesis-based rather than inferred evidence of causality.’’ in discussions section ; page : 16 line : 392

‘’Given its anatomic proximity to the coronary arterial wall, pericoronary adiposity has been proposed to be linked to local paracrine and inflammatory processes in prior imaging studies. In our cohort, the observed association between PCAT thickness and sarcopenia is consistent with a model in which regionally distributed cardiac adiposity may co-occur with systemic metabolic dysregulation that also relates to muscle decline.”Future studies incorporating contemporaneous measures of systemic inflammation, cardiometabolic biomarkers, and CT-based tissue characterization will be required to evaluate whether such biological pathways explain the observed association.” in discussions section ; page : 16 line : 403

Comment 9: the conclusion section appears incomplete. The final sentence ends with “these subgroup findings require” and does not finish the thought. This must be corrected.

Response 9: Thank you for pointing this out. We apologize for the incomplete sentence in the Conclusions. We have corrected and completed the statement to clearly communicate that the subgroup observations are exploratory and require confirmation in larger, prospectively designed studies with standardized imaging/assessment protocols and external validation.

‘’Nonetheless, the formal interaction test was not statistically significant, and these subgroup findings require confirmation in larger prospective cohorts with standardized sarcopenia ascertainment and direct measures of adiposity." In Conclusion section, page :20 line :578

Comment 10: the manuscript would benefit from editorial cleanup. There are several punctuation and formatting inconsistencies, including stray apostrophes, spacing issues, and minor language problems. These do not invalidate the science, but they do affect readability and polish. Examples appear in the methods and results sections.

Response 10:

Thank you for your comment. We agree that readability and presentation are important. The entire article has been edited for punctuation, spacing, and formatting.

 

 

Reviewer 2 Report

Comments and Suggestions for Authors

1. The manuscript addresses a clinically relevant and novel question by examining whether CT-derived epicardial adipose tissue and pericoronary adipose tissue are associated with sarcopenia in hospitalized older adults.
2. The title is clear, but it should be refined to emphasize that the main significant finding relates specifically to pericoronary adipose tissue thickness rather than cardiac fat thickness in general.
3. The abstract is well structured, but the diagnostic implications of PCAT should be softened because the reported AUC values suggest only modest discrimination.
4. The introduction provides a strong rationale linking sarcopenia, adipose tissue inflammation, cardiometabolic risk, and ageing, but it should better justify why linear PCAT thickness on non-contrast thoracic CT was selected.
5. The retrospective single-centre design is acceptable for an exploratory study, although it limits causal interpretation and should be described consistently as showing association rather than causation.
6. The study population is relevant, but the 3-month interval between CT and bioelectrical impedance analysis may introduce misclassification in hospitalized older adults.
7. The authors should report the median time between CT and BIA and consider a sensitivity analysis restricted to patients with a shorter interval.
8. The use of EWGSOP2 criteria for sarcopenia is appropriate, but the manuscript should clarify whether all diagnostic components were available for every participant.
9. Because several geriatric assessment variables had substantial missingness, the authors should explain whether missing data affected sarcopenia classification or only secondary analyses.
10. The CT measurement protocol is described in detail, but interobserver and intraobserver reproducibility for EAT and PCAT measurements should be reported.
11. The use of BIA for skeletal muscle mass is practical, but its limitations in hospitalized older adults with possible hydration changes should be discussed more explicitly.
12. The statistical analysis is comprehensive, but the small sample size and low events-per-variable ratio in extended models raise concerns about overfitting.
13. Diabetes mellitus differed significantly between groups and should be considered as a potential confounder in additional sensitivity analyses.
14. The results support an independent association between PCAT thickness and sarcopenia, whereas EAT was not associated, but subgroup findings by BMI should remain clearly labelled as exploratory because the interaction test was not significant.

Author Response

Dear Editor

We would like to thank you and reviewers for taking your precious time to read our manuscript and to provide feedback.

The related changes and refinements have been made in the revised paper by taking into consideration your valuable comments. You can find the mentioned amendments below. I hope our revision has improved the paper to a level of your satisfaction.

Sincerely.

Corresponding author; Dr.Erdoğan

 

Comment 1: The manuscript addresses a clinically relevant and novel question by examining whether CT-derived epicardial adipose tissue and pericoronary adipose tissue are associated with sarcopenia in hospitalized older adults.

Response 1: We sincerely thank the reviewer for this positive evaluation. We are pleased that the clinical relevance and novelty of examining CT-derived epicardial and pericoronary adipose tissue in relation to sarcopenia in hospitalized older adults were recognized. In the revised manuscript, we further strengthened the framing of the study’s contribution by clarifying the association-based design and highlighting its clinical context and limitations.

Comment 2: The title is clear, but it should be refined to emphasize that the main significant finding relates specifically to pericoronary adipose tissue thickness rather than cardiac fat thickness in general.

Response 2: Thank you for the reviewer's suggestion. However, we did not want to change the title as it reflects the integrity of the article. We hope this is not a problem.

Comment 3: The abstract is well structured, but the diagnostic implications of PCAT should be softened because the reported AUC values suggest only modest discrimination.

Response 3:

Thank you for this important suggestion. We agree that the AUC values indicate modest discrimination, and therefore the abstract should not imply that PCAT provides a robust diagnostic tool or a clinically actionable threshold. We revised the Abstract to (i) remove diagnostic/predictive wording, (ii) describe ROC findings as exploratory, and (iii) explicitly state that discriminative performance is modest and requires prospective validation and external replication.

‘’This study investigated the associations of EAT and PCAT thickness with sarcopenia.’’ in Abstract section Background and Objectives; page 1 line 23

‘’With the addition of the PCAT, the discriminatory power was modest (AUC 0.637 overall and 0.715 for BMI ≥25 kg/m²).” in Abstract section Background and Objectives; page 1 line 37

PCAT thickness measured on routine non-contrast thoracic CT was associated with sarcopenia, whereas EAT thickness was not. ROC analyses showed modest discrimination; therefore, any proposed cut-offs should be considered exploratory and require prospective validation and external confirmation before clinical use.  in Abstract section Conclusions; page 1  line 41

Comment 4: The introduction provides a strong rationale linking sarcopenia, adipose tissue inflammation, cardiometabolic risk, and ageing, but it should better justify why linear PCAT thickness on non-contrast thoracic CT was selected.

Response 4: Thank you for this helpful suggestion. We agree that the article should provide a clearer justification for why we selected linear thickness-based PCAT measurement in clinically acquired non-contrast chest CT. The reasons for choosing this approach are as follows: (i) the study utilised opportunistic imaging obtained from routine inpatient care rather than dedicated cardiac CT/coronary CT angiography; (ii) thickness measurements are a practical and widely applicable method in non-ECG-triggered chest CT when volumetric segmentation or attenuation-based metrics are not feasible; (iii) our aim is to test the association using a method that reflects real-world accessibility in a geriatric inpatient setting. We have revised the introduction to clearly state this rationale.

‘’In this study, we prioritized linear thickness measurements of PCAT on routine, unenhanced thoracic CT to emphasize the feasibility and 'opportunistic' utility of this approach in geriatric clinical practice. While volumetric or attenuation-based analyses are often preferred in dedicated cardiac research, they typically require specialized software and ECG-gated protocols that are not part of standard care for most hospitalized older adults. By utilizing linear measurements on non-contrast scans—which are frequently performed for diverse clinical indications in this population—we aim to demonstrate a pragmatic method for integrating cardiometabolic biomarkers into multidimensional geriatric assessment without additional radiation exposure or specialized post-processing requirements’’ in introduction section page :2 line :74

Comment 5: The retrospective single-centre design is acceptable for an exploratory study, although it limits causal interpretation and should be described consistently as showing association rather than causation.

Response 5: Thank you for your comment. We agree that, due to the retrospective single-center observational design, our findings should be interpreted as evidence of association, not causality. We have made the terminology consistent throughout the paper; replacing causal or predictive statements with relationship-focused statements such as “was related,” “showed a relationship,” and “hypothesis-forming.” We also reinforced the limitations in the conclusion section on line 572, explicitly stating that causal inference is not possible and that reverse causality and confounding may occur.

Comment 6:The study population is relevant, but the 3-month interval between CT and bioelectrical impedance analysis may introduce misclassification in hospitalized older adults.

Response 6: We thank the Reviewer for highlighting the importance of the CT–BIA interval. We agree that temporal misclassification is a legitimate concern in this population and have therefore (a) quantified the interval distribution in the analytic cohort, (b) tested the stability of the PCAT–sarcopenia association across progressively shorter interval windows, and (c) expanded the Limitations to acknowledge this issue. The findings are summarized below.

"Inclusion allowed up to 90 days between thoracic CT and BIA, and in hospitalized older adults body composition can change materially over that interval through acute illness, immobilization, nutritional fluctuation, or deconditioning. The median CT–BIA interval in this cohort was 36 days (IQR 10–81), and the interval did not differ by sarcopenia status. Sensitivity analyses restricted to ≤60 days (n=58) and ≤30 days (n=46) preserved the PCAT–sarcopenia association with materially similar effect estimates (adjusted OR 1.48 and 1.52 respectively, both p<0.05), arguing against attenuation by temporal misclassification. Nevertheless, prospective designs with simultaneous imaging and body-composition assessment would be required to fully eliminate this concern’’ in discussion section ; page : 19 line : 538

Comment 7: The authors should report the median time between CT and BIA and consider a sensitivity analysis restricted to patients with a shorter interval.

Response 7: We thank the Reviewer for highlighting the importance of the CT–BIA interval. We agree that temporal misclassification is a legitimate concern in this population and have therefore (a) quantified the interval distribution in the analytic cohort, (b) tested the stability of the PCAT–sarcopenia association across progressively shorter interval windows, and (c) expanded the Limitations to acknowledge this issue. The findings are summarized below.

"Inclusion allowed up to 90 days between thoracic CT and BIA, and in hospitalized older adults body composition can change materially over that interval through acute illness, immobilization, nutritional fluctuation, or deconditioning. The median CT–BIA interval in this cohort was 36 days (IQR 10–81), and the interval did not differ by sarcopenia status. Sensitivity analyses restricted to ≤60 days (n=58) and ≤30 days (n=46) preserved the PCAT–sarcopenia association with materially similar effect estimates (adjusted OR 1.48 and 1.52 respectively, both p<0.05), arguing against attenuation by temporal misclassification. Nevertheless, prospective designs with simultaneous imaging and body-composition assessment would be required to fully eliminate this concern’’ in discussion section ; page : 19 line: 538

Comment 8: The use of EWGSOP2 criteria for sarcopenia is appropriate, but the manuscript should clarify whether all diagnostic components were available for every participant.

Response 8:

Thank you for this important clarification request. We agree that the manuscript should explicitly state the completeness of EWGSOP2 diagnostic components. We have revised the Methods and Results to clarify the availability of the sarcopenia-defining measures (muscle strength, muscle mass, and physical performance) for the analytic cohort. We also specified how missingness was handled if any EWGSOP2 component was unavailable, and we reported the number of participants contributing to each component.

Muscle mass was classified as low at SMMI values below 9.2 kg/m² in men and below 7.4 kg/m² in women, following the Turkish-population reference values proposed by Bahat et al. [20]. in Materials and Methods section 2.4 Muscle Assessment; page :5 line :172

Grip weakness was defined as a peak value below 32 kg in men and below 22 kg in women, using the Turkish-population reference cutoffs of Bahat et al. [20]." in Materials and Methods section 2.4 Muscle Assessment; page :6 line :178

Sarcopenia status had been established clinically according to the EWGSOP2 algorithm [1], applied with Turkish-population reference cutoffs from Bahat et al. [20]. Under EWGSOP2, isolated weakness indicates probable sarcopenia, combined low strength and reduced muscle mass yield confirmed sarcopenia, and the additional finding of reduced walking speed defines severe sarcopenia [1]. BIA-derived SMMI cut-points were used to characterize low muscle mass in accordance with these criteria. Muscle mass (BIA-derived SMMI) was available in all 101 patients. Handgrip dynamometry yielded numeric values in 89 of 101 patients (88.1%); in the remaining 12 patients (11.9%), grip dynamometry could not be performed because of severe cognitive impairment, ICU-acquired weakness, or hand pathology, and these patients were operationally classified as having low strength in the retrospective algorithm. This approach reflects clinically meaningful inability to complete strength assessment but may introduce outcome misclassification; concordance and sensitivity analyses were therefore performed and are reported in the Supplementary Materials. The five-times sit-to-stand test (n=44) served as a corroborative strength measure. Physical performance (4-m gait speed, SPPB) was used for severity grading when available but did not affect the binary sarcopenia outcome used in this study." in Materials and Methods section 2.4 Muscle Assessment; page :6 line :184

 

Comment 9: Because several geriatric assessment variables had substantial missingness, the authors should explain whether missing data affected sarcopenia classification or only secondary analyses.

Response 9: We thank the Reviewer for this helpful comment. Missing data did not affect sarcopenia classification, as all variables required for the EWGSOP2-based sarcopenia definition were complete or fully resolved for the entire analytic cohort.

Specifically, the final sarcopenia status, BIA-derived skeletal muscle mass and SMMI, age, sex, BMI, and the main exposure variables, PCAT and EAT, were available for all 101 patients. The observed missingness was limited to secondary comprehensive geriatric assessment variables, including Katz, Lawton, MNA, and GDS, which were used only for descriptive cohort characterization and were not included in the sarcopenia definition or in the primary regression models.

Regarding muscle strength, handgrip dynamometry provided numeric values in 89 patients. In the remaining 12 patients, strength testing could not be completed; therefore, these patients were operationally classified as having low strength, as described in our response to the EWGSOP2 component comment. An internal consistency audit reapplying the formal EWGSOP2 algorithm reproduced the recorded sarcopenia label in 99 of 101 patients, showing excellent agreement (98.0%; Cohen’s kappa = 0.960).

Thus, missing data were confined to secondary descriptive variables and did not influence the primary exposure variables, the sarcopenia outcome, or any variable included in the primary regression analyses. We have clarified this point in the Materials and Methods section.

BIA-derived SMMI cut-points were used to characterize low muscle mass in accordance with these criteria. Muscle mass (BIA-derived SMMI) was available in all 101 patients. Handgrip dynamometry yielded numeric values in 89 of 101 patients (88.1%); in the remaining 12 patients (11.9%), grip dynamometry could not be performed because of severe cognitive impairment, ICU-acquired weakness, or hand pathology, and these patients were operationally classified as having low strength in the retrospective algorithm. This approach reflects clinically meaningful inability to complete strength assessment but may introduce outcome misclassification; concordance and sensitivity analyses were therefore performed and are reported in the Supplementary Materials. The five-times sit-to-stand test (n=44) served as a corroborative strength measure. Physical performance (4-m gait speed, SPPB) was used for severity grading when available but did not affect the binary sarcopenia outcome used in this study." in Materials and Methods section 2.4 Muscle Assessment; page :6 line :188

‘’Among secondary variables used in comprehensive geriatric assessment (Katz, Lawton, MNA, and GDS), missing data ranged from 28% to 49% and were handled using complete-case analysis for each analysis. The availability and operational handling of EWGSOP2 muscle strength components (handgrip and 5×sit-to-stand) are described in Section 2.4."Missing data were confined to secondary comprehensive geriatric assessment instruments and did not affect the sarcopenia classification or any variable used in the primary regression models; all components required for the EWGSOP2 outcome, together with PCAT and EAT, were available for the full analytic cohort (n=101)’’ in Materials and Methods section 2.5 -2.5. Statistical Analysis page :7 line: 241

 

Comment 10: The CT measurement protocol is described in detail, but interobserver and intraobserver reproducibility for EAT and PCAT measurements should be reported.

Response 10: Thank you for this important comment. We have expanded the CT measurement description to improve transparency (standardized landmarks, 0.625-mm thin-slice reconstructions, and exclusion of oblique/nonperpendicular planes on non–ECG-gated scans). We also performed an intraobserver reproducibility assessment by remeasuring a stratified random subset of 30 scans after ≥2 weeks, blinded to initial values. Reproducibility was excellent (ICC(2,1)=0.968 for PCAT and 0.965 for EAT; both 95% CI 0.93–0.98), and these results are now reported in the Materials and Methods.

‘’All thoracic CT measurements were performed by a single senior radiologist using a standardised protocol. PCAT thickness was measured on 0.625 mm thin slice reconstructions as the mean of linear measurements across the RCA, LAD, and LCx territories; EAT was measured as the maximal perpendicular thickness anterior to the right ventricular free wall. Oblique or nonperpendicular sections were excluded to limit motion related error in these non ECG gated examinations. To assess measurement robustness, the reader remeasured a stratified random subset of 30 scans after an interval of at least two weeks, blinded to the initial values; the intraobserver intraclass correlation coefficients (ICC[2,1], two-way random-effects, absolute agreement) were 0.968 (95% CI 0.93-0.98) for PCAT and 0.965 (95% CI 0.93-0.98) for EAT.’’ in Materials and Methods section -2.3. Chest Computed Tomography page :4 line :132

‘’Measurements were performed by a single radiologist; although intraobserver reproducibility was excellent (ICC 0.965-0.968), interobserver reproducibility could not be assessed in this retrospective cohort, and this remains a limitation. PCAT was assessed as a linear thickness on non ECG gated thoracic CT rather than by volumetric or attenuation based techniques, which are more susceptible to motion artefact and segment-selection variability. A standardised protocol with thin-slice reconstruction, multi territory averaging, and exclusion of oblique sections was used to mitigate these sources of error; nevertheless, prospective studies with standardised, ideally ECG-gated acquisition and full intra and interobserver reproducibility testing are warranted.’’ in discussion section page :18 line 511

Comment 11: The use of BIA for skeletal muscle mass is practical, but its limitations in hospitalized older adults with possible hydration changes should be discussed more explicitly.

Response 11: Thank you for this important point. We agree that BIA is a pragmatic method in clinical geriatric practice, but its estimates of skeletal muscle mass can be influenced by hydration status particularly in hospitalized older adults who may experience fluid shifts, edema, diuretic use, acute illness, or electrolyte disturbances. We have therefore expanded the Discussion section to explicitly discuss the potential impact of hydration-related measurement error on BIA-derived muscle mass indices and, consequently, on sarcopenia classification. We also clarified in the Methods that BIA was not performed in patients with severe edema or marked electrolyte imbalance to reduce the risk of inaccurate measurements.

‘’Skeletal muscle mass was estimated using BIA, which is practical and widely used but can be sensitive to hydration status. In hospitalized older adults, acute illness, diuretic therapy, fluid shifts, and subclinical edema may affect impedance-derived estimates and could lead to misclassification of muscle mass and sarcopenia status. Although we attempted to standardize measurements and excluded patients with severe edema or major electrolyte imbalance, residual measurement error related to hydration cannot be fully excluded.’’ in Discussion section ,page :19 line :547

‘’BIA was not performed in patients with cardiac pacemakers or other implanted electronic devices, those exhibiting clinically relevant peripheral edema, or those with pronounced electrolyte imbalances.’’ in Materials and Methods section 2.4 Muscle Assessment; page :5 line :165

 

Comment 12: The statistical analysis is comprehensive, but the small sample size and low events-per-variable ratio in extended models raise concerns about overfitting.

Response 12:

We are grateful for this comment and fully agree that overfitting is a legitimate concern in models with a low events-per-variable (EPV) ratio. We would respectfully clarify the EPV structure of our models. With 55 sarcopenia events, the primary model (PCAT, age, sex, BMI; 4 covariates) has an EPV of 13.8, and the diabetes-adjusted sensitivity model (Model E; 5 covariates) has an EPV of 11.0; both meet the conventional threshold of 10. The more fully adjusted model incorporating additional comorbidities (Model D) does fall below this threshold, and for that reason it is presented only as a sensitivity analysis and is not used to support the principal conclusion. We have stated this limitation explicitly and have tempered the strength of our conclusions accordingly. We have also added the EPV values to the Methods so that readers can assess the robustness of each model directly.

Comment 13: Diabetes mellitus differed significantly between groups and should be considered as a potential confounder in additional sensitivity analyses.

Response 13: We thank the Reviewer for this suggestion, with which we agree. Because diabetes mellitus differed significantly between groups at baseline (chi-square p=0.016), we added a diabetes-adjusted sensitivity model (Model E: PCAT, age, sex, BMI, diabetes; EPV=11.0). The PCAT-sarcopenia association remained essentially unchanged (adjusted OR 1.39, 95% CI 1.12-1.71, p=0.003), and diabetes was not an independent predictor in the multivariable model (OR 0.47, 95% CI 0.18-1.18, p=0.107). The likelihood-ratio test comparing the primary model with Model E was non-significant (chi-square=2.61, p=0.106), indicating that the addition of diabetes did not materially change the model. These results have been added to the manuscript.

Comment 14: The results support an independent association between PCAT thickness and sarcopenia, whereas EAT was not associated, but subgroup findings by BMI should remain clearly labelled as exploratory because the interaction test was not significant.

Response 14: We thank the Reviewer and agree. The BMI-by-PCAT multiplicative interaction was not statistically significant (p=0.291); accordingly, the BMI-stratified findings are now labelled as hypothesis-generating only in the Abstract, Results, Discussion, and Conclusion, and the BMI-stratified figures have been moved to the Supplementary Materials so that they do not receive undue prominence in the main text. We have also moderated the wording describing the principal PCAT finding, in line with our response to the comments on confounder adjustment and diagnostic language, so that the association is presented as robust but not definitively independent.

Reviewer 3 Report

Comments and Suggestions for Authors

I recommend that the authors pay attention to this publication in the introduction section DOI: 10.2174/0113816128284695240219093612
Is it possible to add histological and/or immunohistochemical indicators of EAT and PCAT, especially in the discussion section, because it is necessary to clearly understand what the difference is at the cellular level.
Point 2.3, the principle of measuring EAT and PCAT_ look, I am an operating cardiac surgeon, I see the heart in the open every day, and naturally, the anatomy of each patient, both in the area of the free wall of the right ventricle and in the area of the main coronary arteries, is constantly different, and in half of the cases fat inclusions may be absent, the arteries pass intramyocardially, or vice versa, in the thickness of adipose tissue, and since you write "approximate point" on line 127, it means that you are actually blindly determining the place of the study, in this regard, the question is, where is the reliability of the result? Either make it a limitation of the study, or refine the diagnostic search.
Line 129 is the standard correction factor of 1mm. Does it often happen that the EAT and PCAT layer itself can reach this value, so if you subtract it, you get 0? Is it not there, even if it is present?
Point 3.1, table 1, is there a history of myocardial infarction? Because it also contributes to changes in the epicardium in the acute phase, further leading to inflammation and fibrosis. Perhaps this should also be taken out as a limitation of the study.
Lines 358-366 - you have begun to discuss your work as a new special marker, but in my opinion you have not finished your thought, expand it in practical significance.
The "conclusions" section has no practical significance, despite the large number of correlations in the work.
In general, the authors have done a lot of work that is at the junction of several specialties, which may be of interest to many medical professionals. The introduction is described concisely and clearly, the design of the study is worked out clearly and consistently, the statistics are very painstaking and understandable. The results are also presented consistently and clearly. If the authors listen to the comments, the work will become more detailed and understandable.

Author Response

Dear Editor

We would like to thank you and reviewers for taking your precious time to read our manuscript and to provide feedback.

The related changes and refinements have been made in the revised paper by taking into consideration your valuable comments. You can find the mentioned amendments below. I hope our revision has improved the paper to a level of your satisfaction.

Sincerely.

Corresponding author; Dr.Erdoğan

 

Comment 1: I recommend that the authors pay attention to this publication in the introduction section DOI: 10.2174/0113816128284695240219093612

Response 1: Thank you for your suggestion. We have read the article you suggested and mentioned it in the introduction, page 2 ,line 59, and added it to the references. Alterations in lipid metabolism, such as cholesteryl ester transfer protein (CETP), have been associated with an increased risk of cardiometabolic disease, providing additional context for the systemic metabolic environment in which regional adiposity and sarcopenia may coexist.

Comment 2: Is it possible to add histological and/or immunohistochemical indicators of EAT and PCAT, especially in the discussion section, because it is necessary to clearly understand what the difference is at the cellular level.

Response 2:Thank you for your valuable suggestion. We agree that histological and immunohistochemical data will provide valuable cellular-level information regarding differences between epicardial and pericoronary fat deposits. However, our study is retrospective and based on clinically obtained non-contrast thoracic CT and routine geriatric assessments. Histological and immunohistochemical analyses could not be performed due to the lack of available tissue samples. This is noted as a limitation in the discussion section.

’’Although tissue sampling was not performed in our study, previous histological and immunohistochemical studies support biological heterogeneity between cardiac and perivascular fat deposits [29]. However, these mechanisms were not directly evaluated in our cohort. To validate the cellular mechanisms underlying the observed associations, future prospective studies integrating CT-based measurements  with circulating and tissue-level histological and immunohistochemical markers are needed.’’ in Discussion section ; page : 20 line : 554

Comment 3: Point 2.3, the principle of measuring EAT and PCAT_ look, I am an operating cardiac surgeon, I see the heart in the open every day, and naturally, the anatomy of each patient, both in the area of the free wall of the right ventricle and in the area of the main coronary arteries, is constantly different, and in half of the cases fat inclusions may be absent, the arteries pass intramyocardially, or vice versa, in the thickness of adipose tissue, and since you write "approximate point" on line 127, it means that you are actually blindly determining the place of the study, in this regard, the question is, where is the reliability of the result? Either make it a limitation of the study, or refine the diagnostic search.

Response 3:" Dear Reviewer, thank you for your invaluable contribution from your daily surgical practice and assessment of clinical reality. You rightly acknowledge that anatomical anomalies such as the non-homogeneous course of the intramyocardial arteries or the epicardial adipose tissue present with significant medical changes in the patient. Because our study is retrospective and conducted using routine non-contrast thoracic CT scans, it was not possible to replace the diagnostic method with a volutric or 3D scan at this stage. The 25%, 50%, and 75% reference points in the free wall of the right ventricle were not random but standardized according to previously validated radiological measurement principles in the literature (References 16 and 17). However, as your impeccable observations confirm, these 2D linear measurements may be insufficient to reflect 3D anatomical reality. Your suggestion that these anatomical constraints (absence of fat inclusions, intramyocardial course, etc.) would enhance the robustness of 2D measurements has been added to the 'Discussion' section as a significant limitation of our study.

‘’Anatomical variations, such as the intramyocardial course of coronary arteries or the non-homogeneous distribution of epicardial fat tissue, show significant differences from patient to patient. The absence of fat inclusions and the fact that anatomical variations such as intramyocardial course can affect the reliability of 2D measurements is a significant limitation of our study.’’ in Discussion section ; page : 19 line : 519

 

‘’Moreover, linear measurements cannot fully account for physiological anatomical variations, such as the intramyocardial course of coronary arteries or the heterogeneous, focal absence of epicardial fat deposits, which may affect the reliability of cross-sectional measurements.’’ in Discussion section ; page : 19 line : 524

Comment 4: Line 129 is the standard correction factor of 1mm. Does it often happen that the EAT and PCAT layer itself can reach this value, so if you subtract it, you get 0? Is it not there, even if it is present?

Response 4:"Dear Reviewer, thank you for your detailed evaluation. First, we would like to point out that the 1 mm standard correction factor included in our methodology was applied only to Epicardial Fat Tissue (EAT) measurements to avoid the inclusion of adjacent vascular structures in the measurement, and this deduction was not made for Pericoronary Fat (PCAT) measurements. While it is theoretically possible for healthy and very thin young individuals to have a fat layer of 1 mm or less, our study group consisted of elderly hospitalized individuals with a mean age of 78.5 years. In our dataset, the median EAT thickness was measured as 6.0 mm (Interquartile Range: 5.0–8.0 mm). Even the lowest raw EAT measurement detected in our data was well above 1 mm (approximately 2–3 mm), therefore we had no cases where '0 mm' was obtained or where the absence of fat tissue was assumed after deducting the 1 mm correction factor. This is a natural consequence of our study with an elderly and specific patient group. The materials and  methods, and discussion sections have been revised to avoid misunderstandings in the text.

‘’The 1 mm correction applies only to EAT and not to PCAT.’’  in Materials and Methods section -2.3. Chest Computed Tomography page :5 line: 150

‘’Moreover, linear measurements cannot fully account for physiological anatomical variations, such as the intramyocardial course of coronary arteries or the heterogeneous, focal absence of epicardial fat deposits, which may affect the reliability of cross-sectional measurements.’’ in Discussion section; page: 19 line :524

Comment 5: Point 3.1, table 1, is there a history of myocardial infarction? Because it also contributes to changes in the epicardium in the acute phase, further leading to inflammation and fibrosis. Perhaps this should also be taken out as a limitation of the study

Response 5: Thank you for this clinically important point. We agree that a prior history of myocardial infarction (MI) and recent acute coronary syndromes could influence epicardial and pericoronary tissue characteristics through inflammatory and fibrotic remodeling,and could act as a potential confounder in analyses linking cardiac adiposity to sarcopenia. In our dataset, comorbidities were primarily recorded as broader cardiovascular categories . A dedicated variable capturing MI history or recency was not consistently available in the retrospective records, and therefore it could not be included as a covariate in the main models. We have now added this explicitly as a limitation and clarified that future studies should incorporate detailed cardiovascular phenotyping, including MI history and timing, and ideally exclude or separately analyze patients with recent acute coronary events.

‘’Data regarding the timing of previous myocardial infarction, and especially recent acute coronary syndrome, were not consistently available. Since post-infarction inflammatory and fibrotic remodeling can alter epicardial and pericoronary tissue characteristics, unmeasured MI history and timing may contribute to residual confounding in the observed relationships.’’ in Discussion section; page: 19   line: 528

Comment 6:Lines 358-366 - you have begun to discuss your work as a new special marker, but in my opinion you have not finished your thought, expand it in practical significance. The "conclusions" section has no practical significance, despite the large number of correlations in the work.

Response 6: Thank you for your constructive suggestion. We agree that the Discussion and Conclusion sections should more clearly articulate the practical and clinical significance of our findings, while remaining consistent with the modest discrimination and retrospective association design. In the revised text, we have added a realistic clinical use case to the Discussion section regarding the use of opportunistic PCAT measurement in routine non-contrast thoracic CT as an adjunct risk indicator rather than an independent diagnostic test. We have also revised the Conclusion section to provide a concise, practical summary message and to clearly state that prospective validation and external replication are necessary before clinical practice.

‘’In an exploratory BMI-stratified analysis, the PCAT-sarcopenia association appeared numerically larger in participants with BMI >=25 kg/m2 (OR 1.48, p=0.008). However, the formal BMI-by-PCAT interaction was not statistically significant (p=0.290); this subgroup observation is hypothesis-generating only and requires confirmation in adequately powered prospective studies." in Discussion section; page: 16 line: 418

‘’Nonetheless, the formal interaction test was not statistically significant, and these subgroup findings require confirmation in larger prospective cohorts with standardized sarcopenia ascertainment and direct measures of adiposity." in conclusions section ; page : 20 line :578

Comment 7: In general, the authors have done a lot of work that is at the junction of several specialties, which may be of interest to many medical professionals. The introduction is described concisely and clearly, the design of the study is worked out clearly and consistently, the statistics are very painstaking and understandable. The results are also presented consistently and clearly. If the authors listen to the comments, the work will become more detailed and understandable.

Response 7: We sincerely thank the reviewer for this positive and encouraging evaluation. We appreciate the recognition of the interdisciplinary relevance, clarity of the study design, and the transparency of the statistical approach. We have carefully addressed all reviewer comments to further improve precision, interpretability, and clinical relevance, and we believe the revised version is clearer and more detailed as a result.

 

 

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript now better frames the study as a retrospective observational analysis rather than as a diagnostic validation study. The authors have also appropriately acknowledged the exploratory nature of the BMI-stratified findings, especially since the formal BMI-by-PCAT interaction was not statistically significant. The addition of intraobserver reproducibility data for PCAT and EAT measurements, the diabetes-adjusted model, the sensitivity analyses according to CT–BIA interval, and the discussion of the modest ROC performance all strengthen the manuscript.

In its current form, the study presents an interesting and potentially relevant observation: PCAT thickness measured opportunistically on routine non-contrast thoracic CT was associated with sarcopenia in hospitalized older adults, whereas EAT thickness was not. This is a clinically attractive hypothesis-generating finding, particularly in the context of geriatric risk stratification and opportunistic imaging.

However, I still recommend minor revision before acceptance, mainly for editorial and structural reasons.

Remaining comments

The Discussion remains somewhat repetitive and should be shortened.

Several limitations and interpretative cautions are repeated in multiple places, including the exploratory nature of the BMI-stratified analyses, the absence of significant interaction, the limitations of linear PCAT measurement, and the lack of mechanistic biomarkers. The Discussion would benefit from being more concise and better structured.

The manuscript still contains several editorial artifacts and formatting issues.

There are residual quotation marks, spacing problems, and typographical inconsistencies throughout the text. For example, “3.1.1Characteristics” should be corrected, and the manuscript contains occasional stray quotation marks. The section “6. Patents” is inappropriate unless there are actual patents to declare and should be corrected according to journal formatting.

The clinical interpretation of ROC-derived thresholds should remain cautious.

Although the authors now appropriately state that PCAT alone has modest discriminatory performance, this point should be consistently maintained throughout the manuscript. The proposed cut-offs should not be presented as clinically actionable thresholds, but only as exploratory estimates requiring external validation.

The limitations related to CT acquisition and measurement methodology should be kept prominent.

The use of non-ECG-gated thoracic CT, linear rather than volumetric or attenuation-based PCAT assessment, and the absence of interobserver reproducibility remain important limitations. These do not invalidate the study, but they restrict the strength of the conclusions.

Author Response

Dear Editor

We would like to thank you and the reviewers for taking your precious time to read our manuscript and to provide feedback.

The related changes and refinements have been made in the revised paper by taking into consideration your valuable comments. You can find the mentioned amendments below. I hope our revision has improved the paper to a level of your satisfaction.

Sincerely.

Corresponding author: Dr.Erdoğan

 

Comment 1: The Discussion remains somewhat repetitive and should be shortened.

Response 1 : We thank the reviewer for the careful reevaluation and for recommending minor revision. We agree and have substantially condensed the Discussion.

Two paragraphs starting on line 523 were deleted,

‘’Anatomical variations, such as the intramyocardial course of coronary arteries or the non-homogeneous distribution of epicardial fat tissue, show significant differences from patient to patient. The absence of fat inclusions and the fact that anatomical variations such as intramyocardial course can affect the reliability of 2D measurements is a significant limitation of our study.”

 “Moreover, linear measurements cannot fully account for physiological anatomical variations, such as the intramyocardial course of coronary arteries or the heterogeneous, focal absence of epicardial fat deposits, which may affect the reliability of cross-sectional measurements.”

and a new paragraph was added. (Line 547, page 19) The sentence we added is ‘’Linear, cross-sectional measurements cannot fully capture physiological anatomical variation — such as the intramyocardial course of coronary arteries or the focal, heterogeneous absence of epicardial fat — which may reduce measurement reliability; this remains an inherent limitation of the 2D approach’’

The sentence on line 421 of the "Discussion" section has been deleted because it was a repetition.’’ Accordingly, these subgroup findings should be interpreted as hypothesis-generating rather than as definitive evidence of effect modification.’’

The sentence on line 556 has been deleted because it was a repetition.’’Nonetheless, the formal interaction test was not statistically significant, and these subgroup findings require confirmation in larger prospective cohorts with standardized sarcopenia ascertainment and direct measures of adiposity.’’

 

Comment 2: There are residual quotation marks, spacing problems, and typographical inconsistencies throughout the text. For example, “3.1.1Characteristics” should be corrected, and the manuscript contains occasional stray quotation marks. The section “6. Patents” is inappropriate unless there are actual patents to declare and should be corrected according to journal formatting.

Response 2: Thank you for your suggestions. We inserted the missing space in “3.1. Characteristics” and removed three stray quotation marks (in the Abstract, in Section 2.5, and in the Discussion). We additionally identified and fixed further artifacts during a full re-proofread: the misplaced space in “(BIA) preceded” (Section 2.1), an unclosed parenthesis in Section 3.5. (“…Supplementary Figure 4”), and a broken inline character/formatting around the word “causality” in the Discussion. The empty “6. Patents” section has been removed, as there are no patents to declare. The whole manuscript was re-read for residual typographical inconsistencies.

 

Comment 3 : The clinical interpretation of ROC derived thresholds should remain cautious.

Although the authors now appropriately state that PCAT alone has modest discriminatory performance, this point should be consistently maintained throughout the manuscript. The proposed cut-offs should not be presented as clinically actionable thresholds, but only as exploratory estimates requiring external validation.

Response 3 : We agree and have made the wording consistent throughout. The exploratory, non-actionable nature of the cut-offs is now stated explicitly at first mention in the Results (Section 3.6.) and in the footnote to Table 5, in addition to the Abstract and Conclusions where this was already noted. In the Discussion, the 11.44 mm value is now framed strictly as an exploratory estimate requiring external validation rather than as a clinically actionable threshold. We made no change to the underlying analyses or reported values.

Comment 4 : The limitations related to CT acquisition and measurement methodology should be kept prominent.

Response 4 : We agree with your comment. As you said, we clearly stated this as a limitation on line 538 of the article.

‘’Measurements were performed by a single radiologist; although intraobserver reproducibility was excellent (ICC 0.965-0.968), interobserver reproducibility could not be assessed in this retrospective cohort. This is a limitation of the article.’’

Reviewer 2 Report

Comments and Suggestions for Authors

1. Thank you for the revised manuscript and detailed responses. The manuscript has improved substantially, and most of the previous concerns have been addressed satisfactorily.

2. The authors have appropriately clarified the retrospective and association-based nature of the study, softened the diagnostic interpretation of PCAT, added CT–BIA interval sensitivity analyses, clarified EWGSOP2 component availability, discussed missing data and BIA-related limitations, addressed overfitting concerns, added diabetes-adjusted sensitivity analysis, and labelled BMI-stratified findings as exploratory. Only a few minor issues remain before the manuscript can be considered fully acceptable.

• First, the title should be revised to better reflect the main finding, which relates specifically to pericoronary adipose tissue rather than cardiac fat thickness in general. Because epicardial adipose tissue was not associated with sarcopenia, the current title may overstate the broader relevance of cardiac fat thickness. I recommend revising the title to: “Pericoronary, but Not Epicardial, Cardiac Fat Thickness Is Associated with Sarcopenia in Hospitalized Older Adults.”
• Second, the authors have added useful intraobserver reproducibility data; however, interobserver reproducibility was not assessed. If this was not feasible because measurements were performed by a single radiologist in this retrospective cohort, this should be stated explicitly in the response letter and clearly retained as a limitation in the manuscript.

3.Finally, the manuscript would benefit from careful formatting and language editing before publication. Minor issues with spacing, punctuation, quotation marks, and typographical consistency should be corrected. The authors should also ensure that causal, predictive, or diagnostic wording remains appropriately qualified throughout the manuscript. Overall, the manuscript is substantially improved and would be suitable for publication after these minor revisions.

Author Response

Dear Editor

We would like to thank you and the reviewers for taking your precious time to read our manuscript and to provide feedback.

The related changes and refinements have been made in the revised paper by taking into consideration your valuable comments. You can find the mentioned amendments below. I hope our revision has improved the paper to a level of your satisfaction.

Sincerely.

Corresponding author: Dr.Erdoğan

 

Comment 1: Thank you for the revised manuscript and detailed responses. The manuscript has improved substantially, and most of the previous concerns have been addressed satisfactorily.

Response 1: Thank you for your valuable comment. We have taken your suggestions into consideration and tried our best to make improvements.

Comment 2:  The authors have appropriately clarified the retrospective and association-based nature of the study, softened the diagnostic interpretation of PCAT, added CT–BIA interval sensitivity analyses, clarified EWGSOP2 component availability, discussed missing data and BIA-related limitations, addressed overfitting concerns, added diabetes-adjusted sensitivity analysis, and labelled BMI-stratified findings as exploratory. Only a few minor issues remain before the manuscript can be considered fully acceptable.

• First, the title should be revised to better reflect the main finding, which relates specifically to pericoronary adipose tissue rather than cardiac fat thickness in general. Because epicardial adipose tissue was not associated with sarcopenia, the current title may overstate the broader relevance of cardiac fat thickness. I recommend revising the title to: “Pericoronary, but Not Epicardial, Cardiac Fat Thickness Is Associated with Sarcopenia in Hospitalized Older Adults.”
• Second, the authors have added useful intraobserver reproducibility data; however, interobserver reproducibility was not assessed. If this was not feasible because measurements were performed by a single radiologist in this retrospective cohort, this should be stated explicitly in the response letter and clearly retained as a limitation in the manuscript.

Response

2.1: Thank you so much for your feedback. As you suggested, we have changed the title to “Pericoronary, but Not Epicardial, Cardiac Fat Thickness Is Associated with Sarcopenia in Hospitalized Older Adults.”

2.2: Thank you for your comment. In the retrospective cohort, it was not possible to assess inter-observer reproducibility because the measurements were performed by a single radiologist. This was noted as a limitation in the discussion section. “This is a limitation of the article.” Page : 19 Line : 540

Comment 3: Finally, the manuscript would benefit from careful formatting and language editing before publication. Minor issues with spacing, punctuation, quotation marks, and typographical consistency should be corrected. The authors should also ensure that causal, predictive, or diagnostic wording remains appropriately qualified throughout the manuscript. Overall, the manuscript is substantially improved and would be suitable for publication after these minor revisions.

Response 3: Thank you for your suggestions. As you requested, unnecessary quotation marks have been removed, and punctuation and spacing have been corrected.

We systematically audited the manuscript for causal, predictive, and diagnostic wording and ensured each instance is appropriately qualified. Given the cross-sectional, retrospective design, the central finding is consistently described as an association  rather than a causal or predictive effect, and the absence of mechanistic data is stated explicitly. Where residual wording could imply prediction or diagnosis, we revised it: the Figure 3 legend now refers to discrimination rather than “predicting” sarcopenia, the Table 5 heading now reads “classification  performance” rather than “diagnostic performance,” and the ROC derived cut offs are framed throughout as exploratory estimates requiring external validation rather than clinically actionable thresholds.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors took into account all the remarks, the publication is recommended for adoption

Author Response

Dear Editor

We would like to thank you and the reviewers for taking your precious time to read our manuscript and to provide feedback.

The related changes and refinements have been made in the revised paper by taking into consideration your valuable comments. You can find the mentioned amendments below. I hope our revision has improved the paper to a level of your satisfaction.

Sincerely.

Corresponding author: Dr.Erdoğan

Comment 1: The authors took into account all the remarks, the publication is recommended for adoption

Response 1 : We have benefited from your constructive and helpful criticism. Thank you. Sincerely