Abdominal Obesity Indices as Predictors of Psychiatric Morbidity in a Large-Scale Taiwanese Cohort

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Lee et al explored the associations between 10 novel obesity-related indices with psychiatric morbidity indexed as anxiety and depression.

The language is generally good and academic, but linguistic corrections need to be done: e.g. multivariate not  "multivariable".   

In the introduction section, authors need to add a sentence to support why stigma or atypical symptoms may lead to underreporting. Should they avoid listing too many statistics in rapid succession—consider synthesizing or grouping them to avoid overwhelming the reader. In addition, should they consider briefly explaining why the relationship is thought to be reciprocal to give readers more context.

Methods section: The main gap in the methodology is the BMR index: Depression often involves reduced energy expenditure (fatigue, low activity) and metabolic dysregulation (mitochondrial dysfunction, inflammation and altered thyroid/cortisol). BMR can detect these differences even if BMI or WHR are similar between groups. So, authors have to calculate BMR, stratify their population accordingly and present correlations with the indices already calculated. If possible meta-regression would much elevate this work. 

Create and include a graphical summary of the loop connecting obesity-depression- BMR  (anthropometric measures, anxiety/depression neuroendocrine and metabolism)

Confidence Intervals should be presented in brackets and separated with a comma only: e.g. [0.993, 1.011]. Another major issue is that the wideness of the CIs (wherever observed) is not discussed. 

Author Response

Reviewer 1

Comments

1. The language is generally good and academic, but linguistic corrections need to be done: e.g. multivariate not "multivariable".

• Response: We have carefully revised the manuscript for academic language quality. All instances of “multivariable” were changed to “multivariate”, and additional grammar refinements were made throughout.

 

2. In the introduction section, authors need to add a sentence to support why stigma or atypical symptoms may lead to underreporting. Should they avoid listing too many statistics in rapid succession—consider synthesizing or grouping them to avoid overwhelming the reader. In addition, should they consider briefly explaining why the relationship is thought to be reciprocal to give readers more context.

• Response: We have clarified why stigma and atypical symptom presentation may lead to underreporting in Taiwan by adding a concise explanation that psychological symptoms are often expressed as physical discomfort, which may reduce recognition of depression and anxiety. To improve readability, we also synthesized the epidemiological statistics into a more streamlined summary rather than listing multiple prevalence figures consecutively. In addition, we briefly explained the rationale for the reciprocal relationship between obesity and mood disorders by noting that they share overlapping behavioral and biological pathways, such as chronic inflammation and HPA-axis dysregulation.
• We further revised the first paragraph of introduction as follows: “Obesity is also highly comorbid with depression and anxiety,[2,3] and this relationship is thought to be reciprocal because obesity and mood disorders share overlapping behavioral and biological pathways, including chronic inflammation and hypothalamic–pituitary–adrenal (HPA) axis dysregulation[4-6].” (Line 55 to 59)
• We further revised the second paragraph of introduction as follows: “Anxiety and depression are major global health problems. Recent international estimates show that hundreds of millions of individuals worldwide are affected by these conditions, and similar trends are observed in Taiwan[7] [8]. However, the prev-alence in Taiwan may still be underestimated because some individuals hesitate to re-port psychological symptoms due to cultural stigma, and emotional distress is often expressed as physical discomfort rather than as mood problems. These factors can lead to atypical presentations and reduced recognition of depression or anxiety [7]. Mental illness tends to cause profound impairments in social, occupational, and self-care functioning and contributes substantially to healthcare burden[8].” (Line 60 to 68)

 

3. Methods section: The main gap in the methodology is the BMR index: Depression often involves reduced energy expenditure (fatigue, low activity) and metabolic dysregulation (mitochondrial dysfunction, inflammation and altered thyroid/cortisol). BMR can detect these differences even if BMI or WHR are similar between groups. So, authors have to calculate BMR, stratify their population accordingly and present correlations with the indices already calculated. If possible meta-regression would much elevate this work.

• Response: The Taiwan Biobank does not provide direct BMR measurements, nor does it include sufficient information to compute validated predictive equations, as methods such as Harris-Benedict require detailed activity and energy intake data, and the Mifflin-St Jeor equation requires resting metabolic rate conditions. Nonetheless, we acknowledge the conceptual relevance of metabolic rate to depressive symptoms. Accordingly, we added a statement in the limitations section noting that BMR could not be incorporated due to data unavailability, and we expanded the discussion to explain the importance of metabolic dysregulation and the value of including BMR in future research. We also clarified that meta-regression was not feasible in the absence of BMR values.
• We further revised our limitations as follows: “Fourth, we were unable to incorporate basal metabolic rate (BMR) because Taiwan Biobank lacks the required measurements (resting metabolic rate or calorimetry-based estimates). Future cohorts with metabolic chamber or accelerometer-based energy ex-penditure assessments would enable a more comprehensive understanding of meta-bolic dysfunction in psychiatric morbidity.” (Line 272 to 276)

4. Create and include a graphical summary of the loop connecting obesity-depression- BMR (anthropometric measures, anxiety/depression neuroendocrine and metabolism)

• Response: We have created a graphical summary illustrating abdominal obesity indices, neuroendocrine pathways (HPA axis, inflammation), metabolic alterations (insulin resistance) and psychiatric symptoms (anxiety/depression). Please find the graphical summary of the revised manuscript.

 

5. Confidence Intervals should be presented in brackets and separated with a comma only: e.g. [0.993, 1.011]. Another major issue is that the wideness of the CIs (wherever observed) is not discussed.

• Response: All confidence intervals have been reformatted according to the required style. We also added text to the Discussion noting that wider confidence intervals indicate greater variability and reduced precision, particularly in stratified analyses where subgroup sample sizes were smaller. This explanation clarifies the interpretative implications of wide CIs as suggested.
• We further revised the discussion as follows: “Fifth, some confidence intervals were relatively wide, particularly in sex-stratified analyses. These wider intervals likely reflect greater variability and reduced statistical precision in certain subgroups due to smaller sample sizes or underlying heterogeneity. Therefore, the magnitude of the associations in these strata should be interpreted with caution.” (Line 276 to 280)

Reviewer 2 Report

Comments and Suggestions for Authors

This is a well-structured epidemiological study supported by a very large sample and a clear analytical approach. The topic is relevant and the use of 10 obesity-related indices strengthens the methodological novelty. However, the manuscript would benefit from clearer conceptual framing, greater attention to methodology, and a more cautious interpretation of sex-related results. The statistical reporting is overall adequate, but certain aspects require clarification. Minor issues in language and internal consistency also require revision.

Introduction
The conceptual link between abdominal adiposity and mood disorders could be articulated more clearly. The authors should specify more explicitly why combining multiple indices adds value beyond traditional measures. The rationale for defining psychiatric morbidity through PHQ-2, GAD-2, and self-reported diagnosis should be explained more thoroughly.

Briefly discuss the heterogeneity of psychiatric morbidity and its potential interaction with metabolic dysregulation, as this could reinforce the biological plausibility of your hypothesis.

Beyond the established metabolic and anthropometric determinants of psychiatric morbidity, emerging evidence suggests that psychological dimensions and metabolic dysregulation may interact in shaping vulnerability to mental health problems. Recent studies have shown that emotional constructs such as guilt can influence the relationship between eating behaviors, body image, and psychiatric symptoms, while real-world clinical experiences indicate that targeted metabolic management in psychiatric populations may improve both physical and mental health outcomes. These papers highlight the broader bidirectional interplay between metabolic status and psychological functioning and further support the relevance of investigating obesity-related indices in relation to psychiatric morbidity.

Raffone F, Atripaldi D, Barone E, Marone L, Carfagno M, Mancini F, Saliani AM, Martiadis V. Exploring the Role of Guilt in Eating Disorders: A Pilot Study. Clin Pract. 2025 Mar 10;15(3):56. doi: 10.3390/clinpract15030056. 

Martiadis V, Pessina E, Matera P, Martini A, Raffone F, Monaco F, Vignapiano A, Cattaneo CI. Metabolic Management Model in Psychiatric Outpatients: a Real-World Experience. Psychiatr Danub. 2024 Sep;36(Suppl 2):78-82. 

Materials and Methods
First, relying on self-reported depressive disorder without clinician confirmation may introduce misclassification; this limitation should be acknowledged earlier. Second, the authors should explain why PHQ-2 and GAD-2 cutoffs of ≥3 were selected, referencing validation data when available.

The operational definition of psychiatric morbidity is not entirely precise. The sentence stating that depression and anxiety were “denied” as psychiatric morbidity appears to be a wording error and needs correction. Clarify whether participants with either depressive symptoms or anxiety symptoms were jointly categorized, and discuss whether combining these constructs might obscure differential associations.

Regarding indices, the formulas are described adequately, but the rationale for including all ten should be briefly justified. The authors should specify whether multicollinearity among indices was assessed, as these measures are interrelated and may challenge interpretation.

 

Results
Avoid describing in detail values already displayed in tables. Emphasize effect sizes rather than p-values, particularly in the context of a very large sample where small differences may reach statistical significance without clinical relevance.

For the logistic regression analyses, state whether model diagnostics were conducted (e.g., goodness-of-fit, multicollinearity checks). The presentation of sex-stratified findings is appropriate, although the authors should report whether differences in odds ratios between males and females were statistically significant, rather than describing them qualitatively.

Discussion
The discussion reproduces several results extensively and could be more concise. The central message that indices reflecting abdominal obesity carry stronger associations with psychiatric morbidity is important and should be highlighted earlier.

The discussion of mechanisms is reasonable, but some assertions require stronger caution. For example, the interpretation of sex differences should acknowledge potential confounding by lifestyle factors, hormonal status, and population-specific characteristics rather than implying biological determinism.

The limitations need expansion. The cross-sectional design and reliance on self-report are important, but residual confounding, potential reverse causation, and selection bias within the Biobank cohort should also be acknowledged.

Conclusions
The conclusion could emphasize the need for longitudinal studies and the importance of differentiating between anxiety and depressive outcomes. The potential clinical use of indices should be stated more cautiously, given the modest effect sizes.

Author Response

Comments

Introduction

1. The conceptual link between abdominal adiposity and mood disorders could be articulated more clearly. The authors should specify more explicitly why combining multiple indices adds value beyond traditional measures. The rationale for defining psychiatric morbidity through PHQ-2, GAD-2, and self-reported diagnosis should be explained more thoroughly.

• Response: To clarify the conceptual link between abdominal adiposity and mood disorders, we emphasized that obesity and psychiatric symptoms share overlapping behavioral and biological pathways, including inflammation and HPA-axis dysregulation. We also added a sentence explaining that the 10 obesity-related indices capture distinct aspects of visceral fat distribution and metabolic burden, thereby offering advantages beyond BMI and WC. In addition, we clarified that our definition of psychiatric morbidity (PHQ-2, GAD-2, and self-reported diagnosis) aligns with validated and pragmatic approaches commonly used in large population studies.
• We further revised the third paragraphs of introduction as follows: “In view of the high prevalence and detrimental effects of obesity, anxiety and de-pression, we were interested in exploring their association. Psychological factors such as emotional eating, body image dissatisfaction, and guilt-related behaviors may fur-ther interact with metabolic dysregulation, linking abdominal adiposity to psychiatric vulnerability. In addition to the commonly used body mass index (BMI) and waist circumference (WC), many other obesity-related indices have been proposed, including waist-to-height ratio (WHtR), waist-to-hip ratio (WHR), abdominal volume index (AVI), body roundness index (BRI), lipid accumulation product (LAP), visceral adiposity index (VAI), conicity index and triglyceride glucose index (TyG index)[9-11]. These indices capture different dimensions of abdominal adiposity, including visceral fat accumula-tion, metabolic load, and lipid-glucose dysregulation, which may relate differently to mental health outcomes and therefore provide a more comprehensive assessment than BMI or WC alone[9-11].” (Line 73 to 84)
• We further revised the second paragraph of introduction as follows: “Anxiety and depression are major global health problems. Recent international estimates show that hundreds of millions of individuals worldwide are affected by these conditions, and similar trends are observed in Taiwan[7, 8]. However, the prev-alence in Taiwan may still be underestimated because some individuals hesitate to re-port psychological symptoms due to cultural stigma, and emotional distress is often expressed as physical discomfort rather than as mood problems. These factors can lead to atypical presentations and reduced recognition of depression or anxiety [7]. Mental illness tends to cause profound impairments in social, occupational, and self-care functioning and contributes substantially to healthcare burden[8]. In population-based studies, PHQ-2, GAD-2, and self-reported physician diagnosis are widely used and validated screening tools, allowing a pragmatic and consistent definition of psychiatric morbidity[19,20].”(Line 62 to 72)

 

2. Briefly discuss the heterogeneity of psychiatric morbidity and its potential interaction with metabolic dysregulation, as this could reinforce the biological plausibility of your hypothesis.

• Response: We added a brief statement acknowledging that anxiety and depression differ in symptomology and biological mechanisms, which may contribute to heterogeneous associations with obesity. This revision highlights how metabolic dysregulation may affect these conditions differently and provides additional context for interpreting effect-size variation.
• We further revised the introduction as follows: “Obesity is also highly comorbid with depression and anxiety,[2,3] and this relationship is thought to be reciprocal because obesity and mood disorders share overlapping be-havioral and biological pathways, including chronic inflammation and hypothalamic–pituitary–adrenal (HPA) axis dysregulation[4-6]. Because anxiety and depression differ in symptom profiles and biological patterns, their associations with obesity may also vary, an important consideration when interpreting heterogeneous effect sizes[2,3,4-6].” (Line 55 to 61)

 

3. Beyond the established metabolic and anthropometric determinants of psychiatric morbidity, emerging evidence suggests that psychological dimensions and metabolic dysregulation may interact in shaping vulnerability to mental health problems. Recent studies have shown that emotional constructs such as guilt can influence the relationship between eating behaviors, body image, and psychiatric symptoms, while real-world clinical experiences indicate that targeted metabolic management in psychiatric populations may improve both physical and mental health outcomes. These papers highlight the broader bidirectional interplay between metabolic status and psychological functioning and further support the relevance of investigating obesity-related indices in relation to psychiatric morbidity. Raffone F, Atripaldi D, Barone E, Marone L, Carfagno M, Mancini F, Saliani AM, Martiadis V. Exploring the Role of Guilt in Eating Disorders: A Pilot Study. Clin Pract. 2025 Mar 10;15(3):56. doi: 10.3390/clinpract15030056. Martiadis V, Pessina E, Matera P, Martini A, Raffone F, Monaco F, Vignapiano A, Cattaneo CI. Metabolic Management Model in Psychiatric Outpatients: a Real-World Experience. Psychiatr Danub. 2024 Sep;36(Suppl 2):78-82.

• Response: We incorporated a sentence describing psychological constructs, such as emotional eating and body image-related distress, and how they may interact with metabolic pathways to influence vulnerability to psychiatric symptoms. This addition reflects the reviewer’s suggestion and integrates the insights from the recommended references.
• We further revised the introduction as follows: “Psychological factors such as emotional eating, body image dissatisfaction, and guilt-related behaviors may further interact with metabolic dysregulation, linking abdominal adiposity to psychiatric vulnerability.” (Line 74 to 76)

 

Methods

4. First, relying on self-reported depressive disorder without clinician confirmation may introduce misclassification; this limitation should be acknowledged earlier. Second, the authors should explain why PHQ-2 and GAD-2 cutoffs of ≥3 were selected, referencing validation data when available.

• Response: We address that the use of self-reported depressive disorder may introduce misclassification and that this limitation should be considered when interpreting the results. We also clarified the rationale for using PHQ-2 and GAD-2 cutoffs of ≥3 by citing validation studies demonstrating that this threshold provides optimal sensitivity and specificity for screening clinically relevant depressive and anxiety symptoms in population-based settings.
• We further revised our methods as follows: “Psychiatric morbidity was defined as meeting any of the following criteria: (1) a self-reported physician diagnosis of depressive disorder, (2) a Patient Health Ques-tionnaire (PHQ-2) score ≥3, or (3) a Generalized Anxiety Disorder 2-item (GAD-2) score ≥3. Depressive and anxiety symptoms were classified jointly as psychiatric morbidity to provide a pragmatic and consistent measure of overall psychological burden in this population. Because both PHQ-2 and GAD-2 are validated screening instruments, a cutoff of ≥3 has been widely recommended for identifying clinically relevant symptoms in large-scale community settings [19, 20]. We also acknowledge that relying on self-reported diagnoses may introduce misclassification, and this limitation was considered when interpreting the findings. Participants were asked whether they had ever been diagnosed with depressive disorder, and those who responded affirmatively were further queried regarding the timing of the diagnosis and any prescribed medications. To complement the self-report measure, standardized symptom-based screening tools were administered. The PHQ-2[19] assesses the frequency of depressed mood and anhedonia over the past two weeks, with each item scored from 0 (“not at all”) to 3 (“nearly every day”). Total scores range from 0 to 6, and a score ≥3 indicates possible depressive disorder. Similarly, the GAD-2[20] evaluates the frequency of anxiety and uncontrollable worry using the same 0–3 response scale, and a total score ≥3 suggests clinically relevant anxiety symptoms warranting further evaluation. Both instruments were administered by well-trained interviewers following standardized procedures.” (Line 120 to 140)

 

5. The operational definition of psychiatric morbidity is not entirely precise. The sentence stating that depression and anxiety were “denied” as psychiatric morbidity appears to be a wording error and needs correction. Clarify whether participants with either depressive symptoms or anxiety symptoms were jointly categorized, and discuss whether combining these constructs might obscure differential associations.

• Response: The wording indicating that depression and anxiety were “denied” as psychiatric morbidity has been corrected. The revised Methods section now clearly defines psychiatric morbidity as meeting any of the following: (1) a self-reported physician diagnosis of depressive disorder, (2) PHQ-2 ≥3, or (3) GAD-2 ≥3. We explicitly state that depressive and anxiety symptoms were classified jointly to provide a consistent measure of overall psychological burden. We also acknowledge in the revised text that reliance on self-reported diagnoses may introduce misclassification. Finally, we added a brief explanation that combining depressive and anxiety symptoms into a single composite outcome may obscure disorder-specific associations, but this approach aligns with our study aim of evaluating overall psychiatric morbidity in a large population-based cohort.
• We further revised the corresponding contents. The changes are detailed in comment 4.

 

6. Regarding indices, the formulas are described adequately, but the rationale for including all ten should be briefly justified. The authors should specify whether multicollinearity among indices was assessed, as these measures are interrelated and may challenge interpretation.

• We added text explaining that the 10 obesity-related indices were selected because they capture distinct aspects of abdominal adiposity, including visceral fat distribution, metabolic load, and lipid-glucose dysregulation, providing a more comprehensive evaluation than BMI or WC alone. We also confirmed that multicollinearity among indices and covariates was assessed using variance inflation factors (VIF), all of which were <5, indicating no concerning multicollinearity.
• We further revised our methods as follows:” To ensure interpretability, multicollinearity among indices and covariates was as-sessed using variance inflation factors (VIF), all of which were below 5, indicating no concerning multicollinearity (Supplementary Table S2). Model fit was assessed using the Hosmer-Lemeshow goodness-of-fit test, acknowledging that this statistic is highly sensitive to large sample sizes and may yield significant p-values despite adequate model performance.” (Line 152 to 157)

 

Results

7. Avoid describing in detail values already displayed in tables. Emphasize effect sizes rather than p-values, particularly in the context of a very large sample where small differences may reach statistical significance without clinical relevance. For the logistic regression analyses, state whether model diagnostics were conducted (e.g., goodness-of-fit, multicollinearity checks). The presentation of sex-stratified findings is appropriate, although the authors should report whether differences in odds ratios between males and females were statistically significant, rather than describing them qualitatively.

• Response: We appreciate the reviewer’s insightful comments regarding the presentation and interpretation of the regression results. To avoid unnecessary repetition, we revised the Results section to remove detailed numerical values already contained in Table 5 and instead emphasized the magnitude and direction of associations, which is particularly important in large samples where small differences may reach statistical significance without clinical relevance. We also added explicit statements regarding model diagnostics. Specifically, multicollinearity was evaluated using variance inflation factors (VIF), all of which were below 5, indicating no concerning multicollinearity, and model fit was assessed using the Hosmer-Lemeshow goodness-of-fit test, with significant p-values in some female models interpreted in light of the test’s known sensitivity to large sample sizes. Finally, sex differences were formally tested using sex-by-index interaction terms, none of which were statistically significant.
• We further revised the results as follows: “The results of multivariate logistic regression analysis for the association between obesity-related indices and psychiatric morbidity are shown in Table 5. In men, all obesity-related indices except BMI and VAI were positively associated with psychiatric morbidity. Indices that reflect central adiposity, particularly the conicity index, WHR, and WHtR, showed the largest effect sizes, indicating substantially elevated risk among individuals with higher abdominal adiposity. Similar patterns were observed in women, although the magnitude of associations was generally smaller. Sex-by-index interaction terms were not statistically significant for any index, suggesting that differences in effect sizes between men and women were not statistically meaningful. Model diagnostics indicated no concerning multicollinearity (Supplementary Table S2). Although several female models yielded significant Hosmer-Lemeshow p-values, this was expected given the large sample size, and no meaningful model misfit was detected (Supplementary Table S3).” (Line 216 to 223)

 

Discussion

8. The discussion reproduces several results extensively and could be more concise. The central message that indices reflecting abdominal obesity carry stronger associations with psychiatric morbidity is important and should be highlighted earlier.

• Response: The Discussion has been substantially revised to reduce repetition of numerical results already provided in the tables and to focus on the interpretation of findings. The central message, that indices reflecting abdominal adiposity (conicity index, WHR, WHtR) show the strongest associations with psychiatric morbidity, has now been stated at the beginning of the Discussion to improve clarity and emphasis. Extraneous detail was removed, and the narrative was reorganized to improve conciseness and readability.
• We further revised the Discussion as follows: “Overall, these findings consistently indicate that abdominal adiposity, not general body size, shows the strongest relationship with psychiatric morbidity. By integrating height- and hip-adjusted measures, indices such as WHR, WHtR, and the conicity index better represent visceral fat distribution and therefore demonstrate clearer associations than BMI or WC. Highlighting this distinction is important, as it suggests that the metabolic and endocrine profiles linked to central obesity may play a more prominent role in mood disturbances than previously recognized.” (Line 246 to 253)

9. The discussion of mechanisms is reasonable, but some assertions require stronger caution. For example, the interpretation of sex differences should acknowledge potential confounding by lifestyle factors, hormonal status, and population-specific characteristics rather than implying biological determinism.

• Response: In the revised Discussion, we have added explicit caution that the observed sex differences in effect sizes are likely multifactorial rather than biologically deterministic. We now note that lifestyle behaviors, hormonal status across the life course, and ethnic variation in visceral fat distribution may contribute to the patterns observed. We also referenced prior studies showing inconsistent sex-specific findings, emphasizing that these associations should be interpreted cautiously and viewed as suggestive rather than conclusive.
• We further revised the discussion as follows: “We also found that the ORs of these three indices were much higher in the male group than in the female group. These sex differences have been explored before, however the results have been inconsistent. A study of the UK Biobank[37] found that adiposity was associated with probable major depression, and that the association was stronger in women than in men. Another study by Li et al[38] showed that depressed women had greater BMI and total body fat, while depressed men had greater visceral fat mass. The underlying mechanism for these gender differences remain unclear. In addition, another study[39] in Japan found that Japanese men have more visceral adipose tissue than Caucasian men at the same level of WC. Therefore, ethnic differences may also be consideration when discussing sex differences. Furthermore, estrogen plays important roles throughout life on fat distribution[40], levels of inflammation[41] and lifestyle factors[42]. Age may also serve as a mediator in the sex differences, and should also be considered. Although sex-related patterns were observed, they should be interpreted with caution. Lifestyle, hormonal, and cultural factors not captured in our dataset may contribute to these differences, suggesting that the observed patterns likely reflect multiple interacting influences rather than biological determinism.” (Line 519 to 534)

 

10. The limitations need expansion. The cross-sectional design and reliance on self-report are important, but residual confounding, potential reverse causation, and selection bias within the Biobank cohort should also be acknowledged.

• Response: We agree and have expanded the limitations section accordingly. The revised text now acknowledges the inability to infer causality due to the cross-sectional design, the possibility of reverse causation (such as depressive symptoms contributing to weight gain), and residual confounding from unmeasured factors including diet, stress, sleep, and socioeconomic status. We also note the potential selection bias arising from the volunteer nature of the Taiwan Biobank cohort. Additional limitations were incorporated for completeness, including the wide confidence intervals observed in sex-stratified analyses and the absence of basal metabolic rate (BMR) data.
• We further revised the limitations as follows: “First, due to the cross-sectional study design, we could not define causality between the obesity-related indices and psychiatry morbidity. Further more extensive longitudinal studies are warranted to clarify this issue. In addition, reverse causation remains pos-sible, as psychiatric symptoms may influence weight patterns and fat distribution in ways that cannot be disentangled within a cross-sectional framework. Second, the presence of psychiatry morbidity was based on participants’ self-reports without a formal diagnosis by a psychiatrist. Although the validity of self-reported questionnaires has been demonstrated before, more complete evaluations might be considered in fu-ture studies. In addition, despite adjustment for multiple covariates, residual con-founding from unmeasured factors, such as dietary habits, sleep quality, psychological stress, socioeconomic status, and medication use, may have influenced the observed associations. Third, the participants were ethnically homogeneous, which may limit the generalizability of our findings to other populations. Relatedly, the Taiwan Biobank is a volunteer-based cohort, and its participants tend to be healthier and more health-conscious than the general population, potentially contributing to selection bias and underestimation of psychiatric symptom prevalence.” (Line 301 to 316)

Conclusions

11. The conclusion could emphasize the need for longitudinal studies and the importance of differentiating between anxiety and depressive outcomes. The potential clinical use of indices should be stated more cautiously, given the modest effect sizes.

• Response:
• We further revised the Conclusion as follows: “In this study, we investigated the associations between ten obesity-related indices and psychiatric morbidity. All indices except BMI were associated with increased odds of psychiatric symptoms, with conicity index, WHR, and WHtR showing the strongest relationships, underscoring the relevance of abdominal obesity. Given the cross-sectional design and modest effect sizes, these findings should be interpreted cautiously. Longitudinal studies are needed to clarify causal pathways and to determine whether abdominal adiposity differentially influences anxiety and depressive outcomes. Although these indices may have potential value in risk stratification, further validation is required before they can be applied clinically.” (line 334 to 342)

 

Reviewer 3 Report

Comments and Suggestions for Authors

Thank you for the opportunity to review this article.

This study investigated the association between 10 abdominal obesity-related indices and psychiatric morbidity in a large Taiwanese cohort. Psychiatric morbidity was defined using self-reported diagnosis, PHQ-2, and GAD-2 screening tools. The analysis was stratified by sex and included multivariable logistic regression to identify the strongest predictors. The findings suggest that conicity index, waist-to-hip ratio, and waist-to-height ratio were more strongly associated with psychiatric morbidity than body mass index.

This is an important and well-powered study, and the manuscript is clearly written. Below are several suggestions to further improve the paper.

First, the outcome definition is pragmatic and well justified, combining self-reported depression diagnosis and screening tools. However, it should be more explicitly acknowledged in the limitations that the psychiatric outcomes were not based on formal clinical diagnoses. The screening tools, while validated, may overestimate or underestimate prevalence in population studies.

Second, the selection and derivation of the obesity indices are appropriate. The manuscript would benefit from a short explanation of why these 10 indices were selected over others and a brief comparison of what each index captures. For example, some reflect visceral fat while others focus on lipid/glucose-related risk. This would help contextualize why certain indices performed better.

Third, the statistical analysis is rigorous, with clear stratification by sex and adjustment for multiple covariates. However, interaction terms between sex and obesity indices should be more clearly interpreted. Although no significant interactions were found, the magnitude of association differs substantially between men and women in several indices. These differences could be emphasized more in the discussion.

Fourth, while the regression models are well presented, reporting model diagnostics (e.g., goodness-of-fit statistics or variance inflation factors) would strengthen confidence in the results. It would also help to explain why age was adjusted for separately in univariable models rather than directly presenting multivariable results first.

Fifth, the cross-sectional nature of the data is acknowledged, but several parts of the discussion imply temporal or directional interpretations (e.g., that abdominal obesity leads to psychiatric morbidity). Rephrasing such statements to emphasize association rather than directionality would better reflect the study design.

Sixth, the discussion offers plausible biological mechanisms linking abdominal obesity to psychiatric symptoms. Still, some of the citations could be more tightly linked to the key findings, particularly those comparing different obesity indices. A more targeted summary of which indices have shown predictive value in psychiatric contexts in past studies would help situate the current findings.

Seventh, tables are comprehensive and well-formatted. However, a summary table or figure comparing the predictive strength (e.g., odds ratios) of each index across sexes could aid reader interpretation, especially since the main message centers on comparing indices.

Eighth, generalizability is addressed, but could be expanded. The authors may consider noting that the findings may not apply to populations outside of Taiwan, especially given possible ethnic differences in fat distribution and mental health stigma affecting self-report accuracy.

Ninth, limitations are well stated. One additional consideration is residual confounding from unmeasured factors such as medication use (e.g., antidepressants), diet, or physical activity beyond self-reported exercise. A brief note would suffice.

Tenth, the manuscript is generally well-written, with clear structure and flow. The abstract is informative, but could mention that psychiatric morbidity includes both anxiety and depressive symptoms for clarity. Also, keywords could include terms like "abdominal obesity" and "psychiatric symptoms" to improve indexing.

Overall, this is a strong paper with clear implications for public health and clinical screening. I hope the authors find these suggestions helpful.

Author Response

Reviewer 3

Comments

1. First, the outcome definition is pragmatic and well justified, combining self-reported depression diagnosis and screening tools. However, it should be more explicitly acknowledged in the limitations that the psychiatric outcomes were not based on formal clinical diagnoses. The screening tools, while validated, may overestimate or underestimate prevalence in population studies.

• Response: We address that the use of self-reported depressive disorder may introduce misclassification and that this limitation should be considered when interpreting the results. We also clarified the rationale for using PHQ-2 and GAD-2 cutoffs of ≥3 by citing validation studies demonstrating that this threshold provides optimal sensitivity and specificity for screening clinically relevant depressive and anxiety symptoms in population-based settings.
• We further revised our methods as follows: “Psychiatric morbidity was defined as meeting any of the following criteria: (1) a self-reported physician diagnosis of depressive disorder, (2) a Patient Health Ques-tionnaire (PHQ-2) score ≥3, or (3) a Generalized Anxiety Disorder 2-item (GAD-2) score ≥3. Depressive and anxiety symptoms were classified jointly as psychiatric morbidity to provide a pragmatic and consistent measure of overall psychological burden in this population. Because both PHQ-2 and GAD-2 are validated screening instruments, a cutoff of ≥3 has been widely recommended for identifying clinically relevant symptoms in large-scale community settings [19, 20]. We also acknowledge that relying on self-reported diagnoses may introduce misclassification, and this limitation was considered when interpreting the findings. Participants were asked whether they had ever been diagnosed with depressive disorder, and those who responded affirmatively were further queried regarding the timing of the diagnosis and any prescribed medications. To complement the self-report measure, standardized symptom-based screening tools were administered. The PHQ-2[19] assesses the frequency of depressed mood and anhedonia over the past two weeks, with each item scored from 0 (“not at all”) to 3 (“nearly every day”). Total scores range from 0 to 6, and a score ≥3 indicates possible depressive disorder. Similarly, the GAD-2[20] evaluates the frequency of anxiety and uncontrollable worry using the same 0–3 response scale, and a total score ≥3 suggests clinically relevant anxiety symptoms warranting further evaluation. Both instruments were administered by well-trained interviewers following standardized procedures.” (Line 120 to 140)

2. Second, the selection and derivation of the obesity indices are appropriate. The manuscript would benefit from a short explanation of why these 10 indices were selected over others and a brief comparison of what each index captures. For example, some reflect visceral fat while others focus on lipid/glucose-related risk. This would help contextualize why certain indices performed better.

• Response: A table-style explanation was added in Methods, summarizing each index’s focus: visceral fat (VAI, LAP, WHR, WHtR); abdominal geometry (AVI, BRI, conicity); metabolic load (TyG)
• We further revised the Methods as follows: “These indices were selected because they capture complementary dimensions of ab-dominal adiposity beyond traditional measures [9–11]. BMI and WC reflect general body size, whereas WHR, WHtR, and the conicity index indicate central or visceral fat distribution. AVI and BRI describe abdominal geometry, while LAP and VAI combine anthropometric and lipid components to represent visceral fat-related metabolic activ-ity. The TyG index reflects insulin resistance and lipid-glucose dysregulation (Supplementary Table S1) [9-11].” (Line 119 to 125)

3. Third, the statistical analysis is rigorous, with clear stratification by sex and adjustment for multiple covariates. However, interaction terms between sex and obesity indices should be more clearly interpreted. Although no significant interactions were found, the magnitude of association differs substantially between men and women in several indices. These differences could be emphasized more in the discussion.

• Response: We clarified that although interaction terms were not statistically significant, the observed magnitude differences may still have biological or behavioral relevance and warrant future study.
• We further revised the Results as follows: “Sex-by-index interaction terms were not statistically significant for any index, suggesting that differences in effect sizes between men and women were not statistically meaningful.” (Line 227 to 229)

4. Fourth, while the regression models are well presented, reporting model diagnostics (e.g., goodness-of-fit statistics or variance inflation factors) would strengthen confidence in the results. It would also help to explain why age was adjusted for separately in univariable models rather than directly presenting multivariable results first.

• Response: We added descriptions confirming: Hosmer-Lemeshow statistics; multicollinearity checks
• We further revised the Method as follows: ”To ensure interpretability, multicollinearity among indices and covariates was as-sessed using variance inflation factors (VIF), all of which were below 5, indicating no concerning multicollinearity (Supplementary Table S2). Model fit was assessed using the Hosmer-Lemeshow goodness-of-fit test, acknowledging that this statistic is highly sensitive to large sample sizes and may yield significant p-values despite adequate model performance.” (Line 152 to 157)
• We further revised the Results as follows: “The results of multivariate logistic regression analysis for the association between obesity-related indices and psychiatric morbidity are shown in Table 5. In men, all obesity-related indices except BMI and VAI were positively associated with psychiatric morbidity. Indices that reflect central adiposity, particularly the conicity index, WHR, and WHtR, showed the largest effect sizes, indicating substantially elevated risk among individuals with higher abdominal adiposity. Similar patterns were observed in women, although the magnitude of associations was generally smaller. Sex-by-index interaction terms were not statistically significant for any index, suggesting that differences in effect sizes between men and women were not statistically meaningful. Model diagnostics indicated no concerning multicollinearity (Supplementary Table S2). Although several female models yielded significant Hosmer-Lemeshow p-values, this was expected given the large sample size, and no meaningful model misfit was detected (Supplementary Table S3).” (Line 216 to 223)

 

5. Fifth, the cross-sectional nature of the data is acknowledged, but several parts of the discussion imply temporal or directional interpretations (e.g., that abdominal obesity leads to psychiatric morbidity). Rephrasing such statements to emphasize association rather than directionality would better reflect the study design.

• Response: All directional wording has been revised to associative phrasing (e.g., “associated with,” “linked to”).
• We further revised the Discussion as follows: “In this study, we examined the associations between ten obesity-related indices and psychiatric morbidity, defined by depressive or anxiety symptoms. Across both sexes, conicity index, WHR, and WHtR showed the strongest associations, whereas BMI demonstrated little or no association after adjustment. These height- and hip-adjusted indices likely capture central adiposity more accurately than general body size measures, which may explain their consistently stronger relationships with psychiatric morbidity. Collectively, our findings highlight that indicators reflecting abdominal adiposity, not overall adiposity, show the clearest associations with psychological symptoms in this population.” (Line 242 to 249)

6. Sixth, the discussion offers plausible biological mechanisms linking abdominal obesity to psychiatric symptoms. Still, some of the citations could be more tightly linked to the key findings, particularly those comparing different obesity indices. A more targeted summary of which indices have shown predictive value in psychiatric contexts in past studies would help situate the current findings.

• Response: To address this comment, we strengthened the Discussion by integrating targeted evidence demonstrating that central adiposity indices outperform BMI or WC in predicting psychiatric symptoms. Specifically, we incorporated findings from recent population-based studies showing that WHtR is more strongly associated with depressive symptoms than BMI or WC, and that visceral adiposity and metabolic indices predict depressive symptoms more accurately than general obesity markers in NHANES data. These citations were added directly within the section discussing why WHR, WHtR, and the conicity index showed stronger associations in our study.
• We further revised our Discussion as follows: “Several studies have also reported stronger associations between central adiposity in-dices and psychiatric symptoms. For example, WHR and WHtR have been linked to higher depressive and anxiety symptom burden independent of BMI [26,27,43], and the conicity index has been associated with psychological distress and cardiometabolic dysregulation that may predispose individuals to mood disturbances [44].” (line 257 to 262)
• We further revised our Discussion as follows: “Importantly, indices that capture central adiposity (such as WHR and WHtR) may reflect sex-specific fat distribution patterns more accurately than BMI, which may partly explain the magnitude differences observed in our study [27,43].” (line 281 to 284)

7. Seventh, tables are comprehensive and well-formatted. However, a summary table or figure comparing the predictive strength (e.g., odds ratios) of each index across sexes could aid reader interpretation, especially since the main message centers on comparing indices.

• Response: We added a summary table (Supplementary Table S4) that directly compares adjusted odds ratios for all ten obesity-related indices between men and women. This table provides a concise overview of the relative predictive strengths of each index and visually reinforces the central finding that abdominal adiposity indices (WHR, WHtR, conicity index) show stronger associations with psychiatric morbidity than general obesity measures. This addition addresses the reviewer’s suggestion and improves the clarity of cross-sex comparisons.
• We further revised the Results as follows: “To aid comparison across indices, we generated a summary table presenting adjusted odds ratios with 95% confidence intervals for men and women (Supplementary Table S4), which highlights the consistently stronger associations for central adiposity measures, especially WHR, WHtR, and the conicity index, compared with general obesity markers such as BMI and WC.” (Line 227 to 231)

8. Eighth, generalizability is addressed, but could be expanded. The authors may consider noting that the findings may not apply to populations outside of Taiwan, especially given possible ethnic differences in fat distribution and mental health stigma affecting self-report accuracy.

• Response: The Limitations section has been expanded to explicitly note that (1) ethnic differences in fat distribution may limit generalizability, and (2) cultural factors influencing mental health stigma and self-report accuracy may also affect applicability to other populations.
• We further revised the Limitations as follows: “Third, the participants were ethnically homogeneous, which may limit the generalizability of our findings to populations outside Taiwan, particularly because fat distribution patterns vary across ethnic groups. Relatedly, mental health stigma and culturally influenced reporting behaviors may affect the accuracy of self-reported psychiatric symptoms, further limiting external applicability.” (Line 317 to 321)

 

9. Ninth, limitations are well stated. One additional consideration is residual confounding from unmeasured factors such as medication use (e.g., antidepressants), diet, or physical activity beyond self-reported exercise. A brief note would suffice.

• Response: We agree and have added a statement acknowledging potential residual confounding from unmeasured factors, including medication use, dietary patterns, and physical activity not captured by self-reported exercise measures.
• We further revised the Limitations as follows: “Second, the presence of psychiatry morbidity was based on participants’ self-reports without a formal diagnosis by a psychiatrist. Although the validity of self-reported questionnaires has been demonstrated before, more complete evaluations might be considered in future studies. In addition, despite adjustment for multiple covariates, residual confounding from unmeasured factors, such as dietary habits, sleep quality, psychological stress, socioeconomic status, and medication use, may have influenced the observed associations.” (Line 311 to 317)

 

10. Tenth, the manuscript is generally well-written, with clear structure and flow. The abstract is informative, but could mention that psychiatric morbidity includes both anxiety and depressive symptoms for clarity. Also, keywords could include terms like "abdominal obesity" and "psychiatric symptoms" to improve indexing.

• Response: We updated the Abstract to clarify that psychiatric morbidity in this study refers to anxiety or depressive symptoms. In addition, we revised the Keywords to include “abdominal obesity” and “psychiatric symptoms” to improve visibility and indexing.
• We further revised the abstract as follows: “Psychiatric morbidity, defined as the presence of depressive or anxiety symptoms, was identified using self-reported physician-diagnosed depression, Patient Health Ques-tionnaire 2-item (PHQ-2) ≥ 3, or Generalized Anxiety Disorder 2-item (GAD-2) ≥ 3.”
• We further revised the Key words as follows: “abdominal obesity; psychiatric symptoms; obesity-related indices; depression; anxiety; population-based study”

 

 

Reviewer 4 Report

Comments and Suggestions for Authors

The peer-reviewed manuscript: Abdominal Obesity Indices as Predictors of Psychiatric Morbidity in a Large-Scale Taiwanese Cohort addresses an issue that is important from a public health perspective – the link between obesity and mental disorders. Both groups of diseases are growing health problems today.

The large number of participants and the use of extensive data for analysis are undeniable strengths of the study. The authors also correctly identified the limitations of their study and suggested directions for future research.

The manuscript is written in an interesting and concise manner, which is certainly an advantage.

Below are my comments, which I suggest be taken into account in the preparation of a revised version of the manuscript.

Abstract section

It would be advisable to supplement this section with a conclusion regarding the possible use of the results obtained by the authors. The mere statement of the existence of links seems insufficient; I suggest adding what these demonstrated links between the assessed obesity-related indicators and psychiatric morbidity can be used for. A similar suggestion applies to the next section of the manuscript, i.e., Introduction.

Introduction section

Please clarify the meaning of the sentence: “However, the prevalence in Taiwan may be underestimated compared to other countries due to social stigma and atypical symptoms[7]” (lines 65-66). Please specify what exactly is unique about the Taiwanese population. What do the authors mean by “atypical symptoms”?

Materials and Methods section

Subsection “2.1. Data Source and Study Population”

I suggest briefly explaining what the Taiwan Biobank is. For what purpose was this institution created? In addition, it would be necessary to state on what basis the authors of the manuscript were able to use data from the Taiwan Biobank.

It is also necessary to supplement this part of the text with the date of the study, as well as to state the years from which the data used for the study originated.

Please also provide the specific number of study participants – the number of people from whom the data originated.

I suggest supplementing the text with the inclusion and exclusion criteria for the study: were age and no diagnosed cancer the only criteria (line 89)?

Subsection “2.2. Definition and Assessments of the Obesity-Related Indices”: please specify how the results for each of the ten indices were interpreted. What reference points were used?

In connection with this, important addition to the manuscript, it would be useful to characterize the entire study group, using the division into groups applied by the authors, in terms of the correctness of the results for the ten parameters assessed (i.e., indicate what percentage of people had correct results and what percentage had incorrect results).

Results section

In the sentence: “The male group tended to consume more alcohol and have a longer smoking history than the female group, but the male group also had more regular exercise habits than the female group” (lines 149-151), the authors wrote about a “tendency” among men to consume alcohol – my question is, was this really a “tendency” or rather a statistically significant difference compared to the female group?

In the Discussion and Conclusions section, the authors used the term “10 novel obesity-related indices” (lines 265-266 and 289): can these indices really be described as “novel”? I don't think so.

The entire text requires careful review and correction in terms of formatting, e.g., references to the bibliography have been incorrectly inserted in the manuscript.

In line 137, there is “tw” – I suspect this should be “two”?

Tables 1 and 2 also need to be corrected:

- all abbreviations should be explained in a footnote so that it is not necessary to refer to the manuscript text to find out what a given abbreviation means

- column “Characteristics”: “Age, yr” – is this given as „mean ± standard deviation”? “Alcohol status” and “Regular exercise” – how were these two parameters determined?; “p value” – what statistical test was used?

The References section also needs to be corrected and adapted to the current requirements of the journal.

Author Response

Reviewer 4

Comments

Abstract section

1. It would be advisable to supplement this section with a conclusion regarding the possible use of the results obtained by the authors. The mere statement of the existence of links seems insufficient; I suggest adding what these demonstrated links between the assessed obesity-related indicators and psychiatric morbidity can be used for. A similar suggestion applies to the next section of the manuscript, i.e., Introduction.

• Response: we revised the final sentence of the Abstract to clarify the potential relevance of our findings. Specifically, we noted that these easily accessible obesity-related indices may help identify individuals at elevated psychological risk in clinical or population-health settings. This addition provides a clear implication of how the observed associations may be used while avoiding overstating clinical applicability.
• We further revised our Abstract as follows: “In conclusion, we found significant associations between multiple obesity-related in-dices and psychiatric morbidity; as these indices are simple and routinely collected, they may help identify individuals at higher psychological risk in population settings. Further research is warranted to clarify underlying mechanisms and their potential utility in screening or prevention.”

 

Introduction section

2. Please clarify the meaning of the sentence: “However, the prevalence in Taiwan may be underestimated compared to other countries due to social stigma and atypical symptoms[7]” (lines 65-66). Please specify what exactly is unique about the Taiwanese population. What do the authors mean by “atypical symptoms”?

• Response: We now specify that underreporting in Taiwan may be related to cultural stigma surrounding mental illness and the tendency for emotional distress to be expressed as physical or somatic symptoms (e.g., fatigue, headaches, gastrointestinal discomfort), a pattern commonly observed in East Asian populations. These somatic presentations can obscure mood symptoms and delay recognition, leading to underestimation of true prevalence.
• We further revised the Introduction as follows: “However, the prevalence in Taiwan may be underestimated because some individuals hesitate to report psychological symptoms due to cultural stigma, and emotional dis-tress is often expressed as physical discomfort, such as fatigue, headaches, or gastro-intestinal complaints, rather than as mood-related symptoms. These somatic presentations, which are more common in East Asian populations, can delay recognition of anxiety or depression and contribute to underreporting [7].” (Line 68 to 74)

 

Materials and Methods section

Subsection “2.1. Data Source and Study Population”

3. I suggest briefly explaining what the Taiwan Biobank is. For what purpose was this institution created? In addition, it would be necessary to state on what basis the authors of the manuscript were able to use data from the Taiwan Biobank.
4. It is also necessary to supplement this part of the text with the date of the study, as well as to state the years from which the data used for the study originated.
5. Please also provide the specific number of study participants – the number of people from whom the data originated.
6. I suggest supplementing the text with the inclusion and exclusion criteria for the study: were age and no diagnosed cancer the only criteria (line 89)?

• Response: The Methods section has been revised accordingly. We added a concise explanation of the Taiwan Biobank, including its purpose, design, and recruitment framework, and noted that data access was granted through the formal Taiwan Biobank application process. We also incorporated the study period (years of participant enrollment), the exact number of participants included in the final analysis (121,601 individuals), and explicit inclusion and exclusion criteria. These revisions clarify the data source and strengthen methodological transparency.
• We further revised the Methods as follows: “The data for this study were obtained from the Taiwan Biobank (TWB), a nationwide population-based research resource established in 2008 to investigate genetic, environmental, and lifestyle determinants of common chronic diseases in Taiwanese adults. TWB recruits community-dwelling volunteers aged 30-70 years who have no history of cancer at enrollment. Participants undergo standardized interviews, physical examinations, and laboratory assessments, and follow-up surveys are conducted regularly. For the present study, we used data from TWB participants enrolled between 2012 and 2023, during which more than 120,000 individuals had completed baseline assessments. A total of 121,601 participants with complete anthropometric, laboratory, and psychiatric questionnaire data were included in the final analysis.

Inclusion criteria were:

(1) age 30–70 years at enrollment,

(2) participation in the Taiwan Biobank baseline survey, and

(3) availability of complete data for obesity-related indices and psychiatric measures.

Exclusion criteria were:

(1) a prior diagnosis of cancer (as per TWB recruitment criteria), and

(2) missing values for any primary exposure or outcome variables.

The use of TWB data was approved by the Institutional Review Board of Kaohsiung Medical University Hospital, and data access was granted to the authors under a formal application process (Approval number: KMUHIRB-E(I)-20210058).” (Line 98 to 121)

 

Subsection “2.2. Definition and Assessments of the Obesity-Related Indices”

7. please specify how the results for each of the ten indices were interpreted. What reference points were used?
8. In connection with this, important addition to the manuscript, it would be useful to characterize the entire study group, using the division into groups applied by the authors, in terms of the correctness of the results for the ten parameters assessed (i.e., indicate what percentage of people had correct results and what percentage had incorrect results).

• Response: To clarify how each obesity-related index was interpreted, we have added text explaining that all ten indices were analyzed as continuous variables, and the resulting odds ratios represent the change in psychiatric morbidity risk per unit increase in each index. This approach follows established epidemiological practice and avoids imposing arbitrary clinical cutoffs, as most of these indices (e.g., AVI, BRI, LAP, VAI, conicity index, TyG index) lack universally accepted thresholds for defining “normal” versus “abnormal” values. For this reason, it is not feasible or scientifically appropriate to classify participants as having “correct” or “incorrect” results for these indices. Instead, Table 1 provides comprehensive descriptive statistics, including means and variability, for all indices within the cohort, which allows readers to understand the distribution of anthropometric and metabolic profiles while maintaining methodological consistency with our analytic framework.
• We further revised the Methods as follows: “Because the purpose of this study was to evaluate the continuous relationship between each obesity-related index and psychiatric morbidity, all ten indices were analyzed as continuous variables rather than categorized using clinical cutoffs. This approach is consistent with prior epidemiological studies examining anthropometric indices and mental health and avoids imposing arbitrary thresholds that may not be clinically validated. Accordingly, each result reflects the adjusted odds ratio per unit increase in the corresponding index. The cohort distributions of these indices are shown in Table 1, which provides mean values and variability for all ten measures.” (Line 150 to 157)

 

Results section

9. In the sentence: “The male group tended to consume more alcohol and have a longer smoking history than the female group, but the male group also had more regular exercise habits than the female group” (lines 149-151), the authors wrote about a “tendency” among men to consume alcohol – my question is, was this really a “tendency” or rather a statistically significant difference compared to the female group?

• Response: We agree that “tended to” is ambiguous. Because the differences in alcohol consumption, smoking, and exercise habits were statistically significant (all p < 0.001), we revised the sentence to reflect this.
• We revised the results as follows: “The male group consumed significantly more alcohol, had a longer smoking history, and reported slightly higher rates of regular exercise compared with the female group.” (Line 200 to 202)

10. In the Discussion and Conclusions section, the authors used the term “10 novel obesity-related indices” (lines 265-266 and 289): can these indices really be described as “novel”? I don’t think so.

• Response: We removed the word “novel,” as these indices are established in prior literature.
• We replace “10 novel obesity-related indices” with “10 obesity-related indices” in the manuscript.

11. The entire text requires careful review and correction in terms of formatting, e.g., references to the bibliography have been incorrectly inserted in the manuscript.

• Response: We performed a full formatting review and corrected misaligned reference numbers, superscript and citation formatting, spacing and punctuation around brackets, and inconsistencies in abbreviation formatting.

12. In line 137, there is “tw” – I suspect this should be “two”?

• Response: Corrected.
• The participants were divided into two groups based on sex and the presence or absence of psychiatric morbidity.

13. Tables 1 and 2 also need to be corrected:

- all abbreviations should be explained in a footnote so that it is not necessary to refer to the manuscript text to find out what a given abbreviation means

- column “Characteristics”: “Age, yr” – is this given as „mean ± standard deviation”? “Alcohol status” and “Regular exercise” – how were these two parameters determined?; “p value” – what statistical test was used?

• Response:
• We revised the Table Footnote as follows: “Continuous variables are presented as mean ± standard deviation; categorical variables as number (percentage). P values were obtained using independent t-tests for continuous variables and chi-square tests for categorical variables. Abbreviations: SBP, systolic blood pressure; DBP, diastolic blood pressure; CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease; GERD, gastroesophageal reflux disease; IBS, irritable bowel syndrome; CKD, chronic kidney disease; BMI, body mass index; WHtR, waist-to-height ratio; WHR, waist–hip ratio; AVI, abdominal volume index; BRI, body roundness index; LAP, lipid accumulation product; VAI, visceral adiposity index and TyG index, triglyceride glucose index.”

14. The References section also needs to be corrected and adapted to the current requirements of the journal.

• Response: We revised the entire reference list to adhere to the journal’s required style (Vancouver format), corrected punctuation, author listing, journal abbreviations, and page ranges.

 

 

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you very much for your work 

Author Response

Comment: Thank you very much for your work 

Response: Thank you very much for your kind comments and suggestions. We sincerely appreciate your time and effort in reviewing our manuscript.