Association Between Remnant Cholesterol and Muscle Mass and Quality: Insights from Muscle Quality Mapping and Abdominal Computed Tomography

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

General evaluation

This manuscript investigates the association between remnant cholesterol and skeletal muscle mass/quality in a large health-check cohort using BIA-derived ASM/BMI and CT-based NAMA/TAMA analysis. The topic is clinically relevant, the sample size is large, and the inclusion of myosteatosis in addition to muscle mass adds value to the study. The sex-stratified analyses and multivariable models are also strengths. However, several methodological and interpretive issues should be addressed before the manuscript can be considered for publication. Therefore, I recommend some revision.

Major comments

1. The main limitation is that the study is cross-sectional, yet some parts of the discussion and conclusion are written in a way that may imply a predictive or clinically actionable role of remnant-C. The wording should be made more cautious throughout the manuscript, especially in the conclusion, because causality cannot be inferred from the present design. The authors themselves acknowledge this limitation, but the interpretation still goes slightly beyond what the data can support.

2. The manuscript refers to sarcopenia, but the study actually evaluates low muscle mass and myosteatosis rather than a full contemporary sarcopenia definition. Muscle strength and physical performance were not available, and this should be emphasized earlier and more clearly. The current framing risks overstating the clinical scope of the findings.

3. The definition of myosteatosis needs stronger justification. The NAMA/TAMA index T-score below −1 is used as the cutoff, but it is not entirely clear how clinically validated and generalizable this threshold is outside the source population. Since this is central to the study outcome, the rationale should be expanded and the limitations of this threshold should be discussed more explicitly.

4. CT attenuation-based muscle quality analysis may be influenced by scanner differences, reconstruction settings, and especially contrast enhancement. The study used multiple CT scanners and contrast-enhanced abdominal CT examinations, but the possible effect of these factors on attenuation-based NAMA/LAMA classification is not discussed sufficiently. This is an important imaging-methodology issue and deserves a dedicated paragraph in the limitations section.

5. Selection bias is likely substantial. The cohort comes from a health screening center and extensive exclusions were applied, resulting in a relatively healthy subset. This limits generalizability and may also affect the observed prevalence of low muscle mass and metabolic disorders. This point is mentioned, but it should be discussed in a more concrete way.

6. The female results should be interpreted more carefully. For low muscle mass, the overall models are reported as significant, but the fully adjusted OR for Q4 is not statistically significant and the number of affected female patients is low. The discussion should avoid overinterpretation of female-specific findings and should emphasize limited power for that subgroup.    

7. The statistical approach would be stronger if the authors also analyzed remnant-C as a continuous variable, ideally with a dose-response approach such as restricted cubic spline analysis, rather than relying only on quartiles. Quartiles are easy to read, but they may obscure the actual shape of the association.

Minor comments

-There appears to be a clear typographical/numerical error in Table 2, female section, where Q3 is written as N = 14834 instead of 1484. This must be corrected.  

- There are minor language issues and awkward expressions that require editorial polishing. Examples include phrases such as “has been associated to various diseases” and some grammatical inconsistencies in the discussion.

-The statistical reporting in tables could be clarified further. In particular, when overall model P-values are significant but individual quartile confidence intervals cross 1.0, the text should clearly distinguish overall association from pairwise quartile significance.

Author Response

Comments 1: The main limitation is that the study is cross-sectional, yet some parts of the discussion and conclusion are written in a way that may imply a predictive or clinically actionable role of remnant-C. The wording should be made more cautious throughout the manuscript, especially in the conclusion, because causality cannot be inferred from the present design. The authors themselves acknowledge this limitation, but the interpretation still goes slightly beyond what the data can support.

Response 1: We thank the reviewer for this constructive comment. We agree that, given the cross-sectional design of our study, causal inference is not warranted. In response, we have carefully revised the Discussion and Conclusion sections to ensure that the language consistently reflects the associative—rather than predictive or causal—nature of our results.

These results suggest a potential association that may warrant evaluation for muscle mass and quality among individuals with elevated remnant-C levels. (page 13, line 316–317)

Our results suggest that remnant-C may be associated with a broader state of metabolic dysregulation characterized by excess adiposity and adverse body composition. (page 13, line 324–326)

These distinct trajectories suggest that sex-specific patterns of remnant-C should be considered in future prospective studies. (page 14, line 354–355)

These mechanisms may provide a biological basis for the observed association between remnant-C and myosteatosis.(page 14, line 365–367)

Our study showed that Korean adults with elevated remnant-C levels exhibited a significantly higher prevalence and greater odds of low muscle mass and poor muscle quality. Our findings suggest that individuals with elevated remnant-C levels may benefit from further evaluation for muscle mass and muscle composition. Prospective studies are warranted to clarify whether elevated remnant-C plays a causal role in muscle deterioration. (Conclusion, page 15, line 395–400)

 

Comments 2: The manuscript refers to sarcopenia, but the study actually evaluates low muscle mass and myosteatosis rather than a full contemporary sarcopenia definition. Muscle strength and physical performance were not available, and this should be emphasized earlier and more clearly. The current framing risks overstating the clinical scope of the findings.

Response 2: We have revised the manuscript to clarify that the study evaluates two specific components—low muscle mass and myosteatosis—rather than sarcopenia as defined by current consensus criteria, which also require assessment of muscle strength and physical performance. We acknowledged this distinction in the Introduction section and have added a more prominent statement in the Limitations. The following changes were made:

Therefore, we aimed to identify the relationship between remnant-C and two key components of sarcopenia—muscle mass and myosteatosis—applying the conventional cutoffs for muscle mass and the novel NAMA/TAMA index from abdominal CT analysis and muscle quality mapping. (Introduction, page 2, line 79–82)

Third, our dataset lacks functional measures such as handgrip strength and gait speed required for sarcopenia diagnosis [2], therefore, the present study should be understood as examining sarcopenia-related muscle composition outcomes rather than sarcopenia as a clinical entity. (Discussion, page 14, line 375–378)

 

Comments 3: The definition of myosteatosis needs stronger justification. The NAMA/TAMA index T-score below −1 is used as the cutoff, but it is not entirely clear how clinically validated and generalizable this threshold is outside the source population. Since this is central to the study outcome, the rationale should be expanded and the limitations of this threshold should be discussed more explicitly.

Response 3: We thank the reviewer for this important methodological concern. The NAMA/TAMA T-score cutoff of −1 was not derived within the study but adopted from Kim et al. [Clin Nutr 40:4022–4028], who established sex- and age-specific reference values for this index in a large Korean health screening cohort, following the conceptual framework of the WHO T-score criterion for osteoporosis. We acknowledge, however, that this threshold has not yet been formally validated in external or non-Asian cohorts. To address this limitation transparently, we have (1) stated the rationale for this cutoff and in the Methods section and (2) expanded this as a limitation of the study in the Discussion section.

We defined patients with the NAMA/TAMA index T-score below −1 as having myosteatosis. T-score cutoff was adopted from Kim et al. [11], who established age- and sex-specific reference values for the NAMA/TAMA index in a large Korean health screening population (<73 in men and <72 in women). (Methods, page 4, line 152–155)

Regarding the myosteatosis threshold, the NAMA/TAMA index T-score below −1 was derived from a Korean health screening population and lacks external validation in diverse populations. While its application across multiple studies has demonstrated consistent associations with cardiometabolic conditions [13–17], population-specific cutoffs remain to be established, and future prospective studies are needed to improve generalizability. (Discussion, page 14–15, line 378–383)

 

Comments 4: CT attenuation-based muscle quality analysis may be influenced by scanner differences, reconstruction settings, and especially contrast enhancement. The study used multiple CT scanners and contrast-enhanced abdominal CT examinations, but the possible effect of these factors on attenuation-based NAMA/LAMA classification is not discussed sufficiently. This is an important imaging-methodology issue and deserves a dedicated paragraph in the limitations section.

Response 4: Thank you for this important comment. We agree that CT attenuation-based muscle quality assessment may be influenced by technical factors such as scanner differences, reconstruction parameters, and contrast enhancement. In response, we have addressed in the potential effects of these factors on attenuation-based NAMA/LAMA classification in the limitation section.

Finally, CT attenuation-based muscle quality assessment may be influenced by technical factors such as scanner differences and contrast enhancement. However, the NAMA/TAMA index—as a ratio derived from the same CT acquisition—may partially mitigate such variability compared with absolute attenuation values. (Discussion, page 15, line 389–393)

 

Comments 5: Selection bias is likely substantial. The cohort comes from a health screening center and extensive exclusions were applied, resulting in a relatively healthy subset. This limits generalizability and may also affect the observed prevalence of low muscle mass and metabolic disorders. This point is mentioned, but it should be discussed in a more concrete way.

Response 5: We agree that the use of a health screening cohort and the application of extensive exclusion criteria may have introduced selection bias and limited the generalizability of our findings. Therefore, we have revised the relevant limitations paragraph to provide a more concrete discussion of the potential impact of selection bias on our findings. The following change was made:

Second, as our cohort was drawn from a single health screening center with several exclusion criteria applied, the study population represents a relatively healthy subset. Compared with Korean nationwide data [47–49], our population showed a lower prevalence of diabetes and hypertension and a higher rate of regular physical activity. Although the median remnant-C level (15 mg/dL) was consistent with a nationwide estimate [50], our findings may not be generalizable to other demographic groups. (Discussion, page 14, line 369–374)

 

Comments 6: The female results should be interpreted more carefully. For low muscle mass, the overall models are reported as significant, but the fully adjusted OR for Q4 is not statistically significant and the number of affected female patients is low. The discussion should avoid overinterpretation of female-specific findings and should emphasize limited power for that subgroup.

Response 6: Thank you for pointing this out. We agree with this comment. Therefore, we have revised the relevant passage in the Discussion section and explicitly mentioned the limitation of statistical power to more carefully interpret the female-specific findings and to explicitly acknowledge the limited statistical power in this subgroup.

In female patients, although a statistically significant overall association for low muscle mass was observed in all models, individual quartile-specific ORs did not reach statistical significance, with confidence intervals crossing 1.0. This is likely attributable to the very low prevalence of low muscle mass among female participants (74 out of 5215; 1.4%), thereby limiting statistical power to detect quartile-level differences. (Discussion, page 13, line 342–346)

Fourth, very low prevalence of low muscle mass in female participants (74 out of 5215; 1.4%) likely resulted in insufficient statistical power to detect significant quartile-level differences, and female-specific findings should therefore be considered exploratory. (Discussion, page 15, line 383–386)

 

Comments 7: The statistical approach would be stronger if the authors also analyzed remnant-C as a continuous variable, ideally with a dose-response approach such as restricted cubic spline analysis, rather than relying only on quartiles. Quartiles are easy to read, but they may obscure the actual shape of the association.

Response 7: We appreciate this valuable suggestion and have addressed it in the revised manuscript. We have performed an additional linear regression analysis treating remnant-C as a continuous variable, with NAMA/TAMA index as the outcome, adjusting for the same covariates used in the primary analysis. The results demonstrated a statistically significant inverse association between remnant-C and NAMA/TAMA index in both male (β = −0.029, SE = 0.008, p < 0.001) and female patients (β = −0.059, SE = 0.014, p < 0.001), consistent with the quartile-based findings and supporting a dose-response relationship between remnant-C and muscle quality. These results have been added as a Supplementary Table in the revised manuscript.

In multivariable linear regression analysis, each 1 mg/dL increase in remnant-C was independently associated with a decrease in the NAMA/TAMA index in both male (β = −0.029, 95% CI: −0.045 to −0.013) and female patients (β = −0.059, 95% CI: −0.087 to −0.031), after adjustment for age, VFA/SFA, smoking status, alcohol consumption, regular exercise, hypertension, and diabetes (Supplementary Table 1). (Results, page 12, line 300–304)

 

Comments 8: There appears to be a clear typographical/numerical error in Table 2, female section, where Q3 is written as N = 14834 instead of 1484. This must be corrected.

Response 8: Thank you for pointing this out. We have revised the mentioned typographical error in Table 2.

 

Comments 9: There are minor language issues and awkward expressions that require editorial polishing. Examples include phrases such as “has been associated to various diseases” and some grammatical inconsistencies in the discussion.

Response 9: We appreciate for this comment. We have carefully reviewed the entire manuscript for language and grammatical issues. All identified awkward expressions have been corrected.

 

Comments 10: The statistical reporting in tables could be clarified further. In particular, when overall model P-values are significant but individual quartile confidence intervals cross 1.0, the text should clearly distinguish overall association from pairwise quartile significance.

Response 10: We agree with this comment. Therefore, we have revised the relevant Results section to explicitly distinguish overall trend significance from individual quartile-level significance.

In female patients, the overall association between remnant-C quartiles and low muscle mass remained significant in all models. However, the individual OR for the highest quartile compared with the lowest quartile was not significant in Model 3 (Table 3 and Fig. 4). (Results, page 9, line 263–266)

Reviewer 2 Report

Comments and Suggestions for Authors

 

The manuscript investigates the association between remnant cholesterol and both low muscle mass and myosteatosis using CT-derived muscle quality mapping in a large Korean cohort.

Major Comments

• The definition of myosteatosis using a NAMA/TAMA T-score below −1 appears internally derived from the study population. Has this threshold been externally validated?
• The prevalence of low muscle mass is very low, particularly in females (1.4%), raising concerns regarding statistical robustness and possible underpowering of subgroup analyses. This should be mentioned in limitations.
• The manuscript repeatedly refers to “sarcopenia,” although functional parameters such as handgrip strength or gait speed were not available. According to contemporary consensus definitions, low muscle mass alone is insufficient for diagnosing sarcopenia. The terminology should therefore be revised throughout the manuscript.
• Remnant cholesterol was calculated indirectly rather than directly measured. Given the known discordance between calculated and directly measured remnant cholesterol, the potential impact on classification accuracy and associations should be discussed.
• You need to add information regarding segmentation validation, interobserver agreement, or quality control would strengthen methodological reliability.

 

Minor Comments

• Figure legends could be shortened and simplified for readability.
• Several typographical and formatting inconsistencies remain (spacing, citation punctuation, capitalization).

Author Response

Comments 1: The definition of myosteatosis using a NAMA/TAMA T-score below −1 appears internally derived from the study population. Has this threshold been externally validated?

Response 1: Thank you for this important comment. We would like to clarify that the NAMA/TAMA T-score cutoff of −1 was not internally derived from the present study population. Rather, this threshold was adopted from the study by Kim et al. [11], which established age- and sex-specific reference values for the NAMA/TAMA index in a large Korean health screening cohort based on the T-score framework commonly used in osteoporosis research. We acknowledge, however, that this cutoff has not yet been formally validated in independent external or non-Asian populations. To address this point more clearly, we revised both the Methods and Discussion sections to explain the rationale for the threshold and to explicitly acknowledge the current limitation in external validation. We also noted that this threshold has been consistently applied in several subsequent studies demonstrating associations between myosteatosis and various cardiometabolic conditions, supporting its construct validity as an imaging biomarker of muscle quality.

We defined patients with the NAMA/TAMA index T-score below −1 as having myosteatosis. T-score cutoff was adopted from Kim et al. [11], who established age- and sex-specific reference values for the NAMA/TAMA index in a large Korean health screening population (<73 in men and <72 in women). (Methods, page 4, line 152–155)

Regarding the myosteatosis threshold, the NAMA/TAMA index T-score below −1 was derived from a Korean health screening population and lacks external validation in diverse populations. While its application across multiple studies has demonstrated consistent associations with cardiometabolic conditions [13–17], population-specific cutoffs remain to be established, and future prospective studies are needed to improve generalizability. (Discussion, page 14–15, line 378–383)

 

Comments 2: The prevalence of low muscle mass is very low, particularly in females (1.4%), raising concerns regarding statistical robustness and possible underpowering of subgroup analyses. This should be mentioned in limitations.

Response 2: Thank you for this important comment. We agree that the relatively low prevalence of low muscle mass, particularly among female participants, may limit the statistical robustness of subgroup analyses. In response, the following text was added to the Discussion section:

In female patients, although a statistically significant overall association for low muscle mass was observed in all models, individual quartile-specific ORs did not reach statistical significance, with confidence intervals crossing 1.0. This is likely attributable to the very low prevalence of low muscle mass among female participants (74 out of 5215; 1.4%), thereby limiting statistical power to detect quartile-level differences. (Discussion, page 14, line 342–346)

Fourth, very low prevalence of low muscle mass in female participants (74 out of 5215; 1.4%) likely resulted in insufficient statistical power to detect significant quartile-level differences, and female-specific findings should therefore be considered exploratory. (Discussion, page 15, line 383–386)

 

Comments 3: The manuscript repeatedly refers to “sarcopenia,” although functional parameters such as handgrip strength or gait speed were not available. According to contemporary consensus definitions, low muscle mass alone is insufficient for diagnosing sarcopenia. The terminology should therefore be revised throughout the manuscript.

Response 3: Thank you for this important comment. We agree that, according to contemporary consensus definitions, sarcopenia should not be diagnosed without functional assessments such as handgrip strength and gait speed. In response, we carefully revised the terminology throughout the manuscript, including the Abstract, Introduction, Discussion, and Limitations sections, to avoid using “sarcopenia” when referring only to muscle composition measures. Specifically, we replaced several instances of “sarcopenia” with more precise terms such as “low muscle mass,” “myosteatosis,” or “muscle composition outcomes,” as appropriate. We also added clarifying statements in the Introduction and Limitations sections to explicitly acknowledge that the present study evaluated sarcopenia-related muscle composition parameters rather than sarcopenia as a fully defined clinical entity.

Therefore, we aimed to identify the relationship between remnant-C and two key components of sarcopenia—muscle mass and myosteatosis—applying the conventional cutoffs for muscle mass and the novel NAMA/TAMA index from abdominal CT analysis and muscle quality mapping. (Introduction, page 2, line 79–82)

Third, our dataset lacks functional measures such as handgrip strength and gait speed required for sarcopeniadiagnosis [2], therefore, the present study should be understood as examining sarcopenia-related muscle composition outcomes rather than sarcopenia as a clinical entity. (Discussion, page 14, line 375–378)

 

Comments 4: Remnant cholesterol was calculated indirectly rather than directly measured. Given the known discordance between calculated and directly measured remnant cholesterol, the potential impact on classification accuracy and associations should be discussed.

Response 4: Thank you for this important comment. We agree that the use of calculated rather than directly measured remnant-C may affect classification accuracy and potentially influence the observed associations. In response, we have mentioned this point in the Limitations section.

Fifth, remnant-C was calculated as TC minus HDL-C and LDL-C because directly measured remnant-C levels were unavailable. Although previous studies have shown a strong correlation between calculated and directly measured remnant-C [51], potential misclassification of remnant-C levels cannot be excluded [52]. (Discussion, page 14, line 386–389)

 

Comments 5: You need to add information regarding segmentation validation, interobserver agreement, or quality control would strengthen methodological reliability.

Response 5: Thank you for this valuable comment. We have added methodological details regarding segmentation validation in the Methods section. The AI-based algorithm was validated on an internal dataset (426 CT scans from 308 subjects) and an external dataset (171 CT scans from 171 subjects from two other hospitals) using the Dice similarity coefficient, cross-sectional area error, and Bland-Altman analysis. All CT images were reviewed by an image analyst and a radiologist, both blinded to clinical information, to ensure segmentation quality. We acknowledge that formal interobserver agreement was not evaluated in the present study. We recognize this as a limitation and will consider incorporating interobserver agreement metrics in future studies.

Cross-sectional CT images were automatically interpreted using an AI-based program with the segmentation technique of a fully convolutional network [33,34], which was designed to select the inferior endplate level of the L3 vertebra and delineate the boundaries of TAMA, visceral fat area (VFA), and subcutaneous fat area (SFA). TAMA encompassed all muscles visible in the axial image, including the psoas, paraspinal, transversus abdominis, rectus abdominis, quadratus lumborum, and internal and external obliques. An image analyst and a radiologist, blinded to clinical information, reviewed and validated all selected CT images and segmented areas. The AI-based algorithm has been previously validated on an internal and an external dataset using the Dice similarity coefficient and cross-sectional area error [33].(Methods, page 3–4, line 134–143)

 

Comments 6: Figure legends could be shortened and simplified for readability.

Response 6: We have revised the figure legends for Figure 3 and Figure 4, removing detailed methodological information that is already provided in the Methods section.

 

Comments 7: Several typographical and formatting inconsistencies remain (spacing, citation punctuation, capitalization).

Response 7: Thank you for your careful review. We have thoroughly revised the manuscript to correct the remaining typographical and formatting inconsistencies, including spacing, citation punctuation, and capitalization, throughout the text.

Reviewer 3 Report

Comments and Suggestions for Authors

The research investigated the association between remnant cholesterol  and skeletal muscle mass and quality in a large Korean population using abdominal CT scans and Muscle Quality Mapping.The study is well-motivated, addressing a significant gap in metabolic health research. The finding that RC is independently associated with myosteatosis  even after adjusting for BMI and waist circumference is clinically relevant.

The manuscript is generally well-written, and I have only a few minor comments for the authors to address.

1.Including a representative CT image showing the muscle quality mapping (color-coded for HAMI/LAMI) would be very helpful for readers unfamiliar with this technique.

2.The 'Patents' section is currently misaligned.

 

 

Author Response

Comments 1: Including a representative CT image showing the muscle quality mapping (color-coded for HAMI/LAMI) would be very helpful for readers unfamiliar with this technique.

Response 1: Thank you for this thoughtful comment. We would like to respectfully clarify that the muscle quality parameters used in our study are NAMA (normal attenuation muscle area) and LAMA (low attenuation muscle area). In accordance with your suggestion, we have added a new figure (Figure 1) to the Methods section, which provides a schematic diagram illustrating abdominal muscle segmentation based on CT attenuation values, with color-coded depictions of each tissue compartment. We hope this addition improves clarity for readers unfamiliar with this technique.

Figure 1. Schematic diagram of abdominal muscle segmentation based on CT attenuation. Modified from Kim et al. [8]. Abbreviations: CT, computed tomography; HU, Hounsfield unit; IMAT, inter/intramuscular adipose tissue; LAMA, low attenuation muscle area; NAMA, normal attenuation muscle area; SMA, skeletal muscle area; TAMA, total abdominal muscle area.

 

Comments 2: The 'Patents' section is currently misaligned.

Response 2: Thank you for pointing this out. We deleted the misplaced “Patents” section in the revised manuscript and carefully reviewed the manuscript formatting to correct any remaining alignment issues.