Association Between Dyspeptic Symptoms and Eating Habits in the Colombian Population

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

First of all, I would like to congratulate the authors on this article. The paper addresses a relevant and up-to-date topic: the association between FD and eating habits in the Latin American population.

The manuscript provides useful information for the clinician regarding the relationship between certain foods and the severity of FD symptoms, as well as the possible protective effects of foods rich in phenolic compounds.

However, I consider that some clarifications and methodological corrections are necessary to improve the rigor of the study:

1. The severity index of FD: the items included don't correspond to the validated domains of FD symptoms (Rome IV) and combine symptoms, behaviors and diagnostic details, which affect construct validity. Clarifications and methodological adjustments are needed so that the index reflects FD severity.
2. Separation into MFD and LFD groups: a clinical or statistical justification of the threshold used for the separation of the groups is necessary. The manuscript doesn't explain the reason for this division of the classification.
3. Additional table regarding foods associated with symptoms: it would be useful to include a table that highlights the foods positively/negatively associated with FD symptomatology, in order to facilitate the clinical interpretation and applicability of the results.
4. The limitations of the study need to be extended: it is recommended to explicitly mention the limitations related to the self-reporting of symptoms by the participants, the subjective variability of psychological perceptions (anxiety, stress) and the potential biases inherent to the cross-sectional design.
5. The references require a complete check according to the MDPI Reference Guide: the reference list shows numerous inconsistencies that need to be corrected. Examples:

Ref. 1 – pages or article number are missing.

Ref. 3 – pages / article number are missing.

Refs. 4 and 7 – incomplete, without article number.

Refs. 8 and 13 – duplicates, require unification.

Ref. 12 – critically incomplete (“Talley 2017”), without bibliographic data; it must be completed or removed.

Ref. 14 – requires article number.

Several references (20–23 etc.) – incomplete (volume, issue, pages, DOI formatting) ….

I consider that systematic verification according to the MDPI guide is necessary for accuracy and consistency.

Author Response

1. The severity index of FD: the items included don't correspond to the validated domains of FD symptoms (Rome IV) and combine symptoms, behaviors and diagnostic details, which affect construct validity. Clarifications and methodological adjustments are needed so that the index reflects FD severity. R = Thanks for your observations. We included the main symptoms of functional dyspepsia in the calculation of the FD index; additionally, this index was positively correlated with the Bristol Scale, which was missed from the previous version of the manuscript and might support the external validity of our index. We are exploring how these FD symptoms are associated with the participants´ demographic data and health habits. In this study, we are unable to assess FD severity. Therefore, we will include this as a limitation in our exploratory study.
2. Separation into MFD and LFD groups: a clinical or statistical justification of the threshold used for the separation of the groups is necessary. The manuscript doesn't explain the reason for this division of the classification. R= Thanks for your suggestion, we modified the introduction section to make clearer our first exploratory aim “characterize differences between participants with more and less FD symptoms in sociodemographic, environmental, and dietary factors, and gastrointestinal and non-communicable diseases (such as diabetes and hypertension)”. In addition, we included a statistical description for the cluster calculation to explain the classification.
3. Additional table regarding foods associated with symptoms: it would be useful to include a table that highlights the foods positively/negatively associated with FD symptomatology, in order to facilitate the clinical interpretation and applicability of the results. R= Thanks for your suggestion, we included the table as you suggested.
4. The limitations of the study need to be extended: it is recommended to explicitly mention the limitations related to the self-reporting of symptoms by the participants, the subjective variability of psychological perceptions (anxiety, stress) and the potential biases inherent to the cross-sectional design. R= We agreed to your suggestions; we included this in the limitation section.
5. The references require a complete check according to the MDPI Reference Guide: the reference list shows numerous inconsistencies that need to be corrected. Examples: Ref. 1 – pages or article number are missing., Ref. 3 – pages / article number are missing., Refs. 4 and 7 – incomplete, without article number., Refs. 8 and 13 – duplicates, require unification., Ref. 12 – critically incomplete (“Talley 2017”), without bibliographic data; it must be completed or removed., Ref. 14 – requires article number., Several references (20–23 etc.) – incomplete (volume, issue, pages, DOI formatting) ….I consider that systematic verification according to the MDPI guide is necessary for accuracy and consistency. R = Done

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript investigates a clinically relevant topic: the link between dietary habits and dyspeptic symptoms in a Latin American population, where evidence is still sparse. The introduction does a solid job framing the issue, referencing global prevalence, Rome IV criteria, and key pathophysiological elements like diet, inflammation, and the gut-brain axis. I also appreciate the validation of the dietary assessment tool, with a Cronbach’s alpha of 0.857 signaling good internal consistency, and the thoughtful effort to tie epidemiological findings to plausible biological mechanisms (from CCK and acrolein to capsaicin, TRPV1, probiotics, and polyphenols). These strengths make the paper stand out beyond basic associations.

However, the manuscript needs substantial major revisions before it can be accepted. Several methodological and interpretive issues undermine its validity and generalizability.

The sample is modest (n=102 Colombians recruited by convenience, skewed young with mean ages of 23.6 years in the MFD group and 28.3 in LFD), so claims about Latin America as a whole don’t hold up.

The title and abstract should be dialed back to “Colombian population” or “Colombian sample,” and the age imbalance needs explicit discussion as a confounder.

Functional dyspepsia (FD) diagnosis relies on self-reported items like “Do you suffer from gastritis or dyspepsia?” or “postprandial nausea?” without structured Rome IV evaluation or ruling out organic causes (e.g., endoscopy), making “FD” a misnomer; reframe it as “dyspeptic symptoms” or “uninvestigated dyspepsia” throughout, with a clear limitation statement.

The LFD/MFD groups come from K-means clustering on the FD index, not clinical thresholds or Rome IV subtypes (PDS, EPS, overlap), so they’re arbitrary rather than true diagnostic entities. Therefore, the authors should position this firmly as exploratory work on a continuous symptom score.

The regression analysis feels murky and underdescribed: it’s unclear if it’s simple or multivariable, what covariates were included (age, sex, BMI, comorbidities), whether collinearity among foods (fast food, fried items, red meat) was checked, or if model assumptions, residuals, and overfitting were validated. A full table with standardized betas, SEs, p-values, and R² is missing, and Figure 2’s scatter plot lacks clear labeling (x-axis, “predicted value”), and a coefficient plot with confidence intervals would be sharper.

Diving into sections, the title and abstract overreach: specify if the questionnaire is original or an adapted FFQ (items, scoring, psychometrics), and tweak “phenolic-rich foods linked to better GI health” to “lower FD index” to avoid post-hoc overinterpretation.

The introduction could temper claims like “certain foods alleviate FD symptoms” to “may alleviate” or “associated with,” distinguishing trials from observational data, and should flag eating disorders or behaviors (chewing speed, portion sizes, skipped meals) as potential confounders.

Methods need more information on recruitment (online? clinic?), inclusion/exclusion (PPI use, H. pylori), response rates, administration mode (completion time, pilot testing), and a full statistical rewrite with test rationales, multiple-comparison corrections, and a summary table.

In the results, the authors should add p-values for continuous variables (age, BMI), normality tests (Kolmogorov-Smirnov), and justification for t-tests vs. Mann-Whitney and clarify units (daily/weekly) in Figure 1.

The discussion offers plausible speculation (e.g., students’ symptoms from stress) but marks it as hypothetical without data on anxiety, depression, or sleep; chamomile, peppers, and sesame as “worsening” factors clash with literature (chamomile is typically anti-dyspeptic), so double-check the data and explain. Increase limitations to cover no clinical FD diagnosis, missing eating disorders/anxiety/stress/physical activity data, cross-sectional limits (no causality), and convenience sampling.

Conclusions are too bold (“dietary changes improve symptoms”); stick to associations and predictors.

In summary, tone down title/abstract claims, overhaul methods (recruitment, stats), sharpen results/figures, separate data from speculation, and bolster cautious limitations/conclusions. This has real promise for shedding light on diet and dyspepsia in Colombia, but it needs tighter stats and diagnostic rigor to fit Nutrients.

I recommend major revision and look forward to seeing the updated version.

Author Response

1. The sample is modest (n=102 Colombians recruited by convenience, skewed young with mean ages of 23.6 years in the MFD group and 28.3 in LFD), so claims about Latin America as a whole don’t hold up. R = We modified the title and whole manuscript to describe that our study was performed in the Colombian population. Additionally, we included a map of Colombia in Appendix A that describes the sample distribution across different regions.
2. The title and abstract should be dialed back to “Colombian population” or “Colombian sample,” and the age imbalance needs explicit discussion as a confounder. R = We modified the title and abstract as you suggested. Additionally, we argued differences in age between LFD and MFD in the discussion section.
3. Functional dyspepsia (FD) diagnosis relies on self-reported items like “Do you suffer from gastritis or dyspepsia?” or “postprandial nausea?” without structured Rome IV evaluation or ruling out organic causes (e.g., endoscopy), making “FD” a misnomer; reframe it as “dyspeptic symptoms” or “uninvestigated dyspepsia” throughout, with a clear limitation statement. R = You are right. We modified the whole manuscript to indicate that we are assessing dyspeptic symptoms rather than functional dyspepsia.
4. The LFD/MFD groups come from K-means clustering on the FD index, not clinical thresholds or Rome IV subtypes (PDS, EPS, overlap), so they’re arbitrary rather than true diagnostic entities. Therefore, the authors should position this firmly as exploratory work on a continuous symptom score. R = Given that we are assessing dyspeptic symptoms (DS) rather than functional dyspepsia, as you suggested in a prior request, our index seems to be reflecting the symptoms instead of diagnostic entities. On the other hand, we modified our manuscript to clarify that our first aim is to explore DS only.
5. The regression analysis feels murky and underdescribed: it’s unclear if it’s simple or multivariable, what covariates were included (age, sex, BMI, comorbidities), whether collinearity among foods (fast food, fried items, red meat) was checked, or if model assumptions, residuals, and overfitting were validated. A full table with standardized betas, SEs, p-values, and R² is missing, and Figure 2’s scatter plot lacks clear labeling (x-axis, “predicted value”), and a coefficient plot with confidence intervals would be sharper. R= We performed a multicollinearity analysis before including predictor variables in the regression model, following the next tests: correlation analyses, tolerance values, variance inflation factors, and condition indices. Variables with values outside the parameter range were excluded from the regression analysis. We included this information in the methods section.
6. Diving into sections, the title and abstract overreach: specify if the questionnaire is original or an adapted FFQ (items, scoring, psychometrics), and tweak “phenolic-rich foods linked to better GI health” to “lower FD index” to avoid post-hoc overinterpretation. R= We modified the title and abstract as you suggested.
7. The introduction could temper claims like “certain foods alleviate FD symptoms” to “may alleviate” or “associated with,” distinguishing trials from observational data, and should flag eating disorders or behaviors (chewing speed, portion sizes, skipped meals) as potential confounders. R= We agreed with your comment, so we added eating behaviors as a study limitation.
8. More information on recruitment (online? clinic?), inclusion/exclusion (PPI use, H. pylori), response rates, administration mode (completion time, pilot testing), and a full statistical rewrite with test rationales, multiple-comparison corrections, and a summary table. R= We included the information requested in the participants and data analysis methods sections.
9. In the results, the authors should add p-values for continuous variables (age, BMI), normality tests (Kolmogorov-Smirnov), and justification for t-tests vs. T and clarify units (daily/weekly) in Figure 1. R= Thanks for your suggestion. We included the Kolmogorov-Smirnov test in the manuscript, and we replaced Figure 1 with Table 2, which includes the requested units.
10. The discussion offers plausible speculation (e.g., students’ symptoms from stress) but marks it as hypothetical without data on anxiety, depression, or sleep; chamomile, peppers, and sesame as “worsening” factors clash with literature (chamomile is typically anti-dyspeptic), so double-check the data and explain. Increase limitations to cover no clinical FD diagnosis, missing eating disorders/anxiety/stress/physical activity data, cross-sectional limits (no causality), and convenience sampling. R= We agreed with your observation, we included these data, such as stress, anxiety, or physical activity, as a limitation for our manuscript. Additionally, we ran again; therefore, our predictors changed.
11. Conclusions are too bold (“dietary changes improve symptoms”); stick to associations and predictors. R= Thanks for your suggestion, we modified the conclusions as you suggested.
12. In summary, tone down title/abstract claims, overhaul methods (recruitment, stats), sharpen results/figures, separate data from speculation, and bolster cautious limitations/conclusions. This has real promise for shedding light on diet and dyspepsia in Colombia, but it needs tighter stats and diagnostic rigor to fit Nutrients. I recommend major revision and look forward to seeing the updated version. R = We modified the manuscript as you suggested, thanks for your observations.

Reviewer 3 Report

Comments and Suggestions for Authors

The article addresses the important and topical issue of the relationship between eating habits and the severity of functional dyspepsia symptoms in the Latin American population. Interest in this condition is justified by its high prevalence, significant public health burden, and diverse aetiology. The authors attempt not only to analyse dietary factors but also to validate a research tool in Spanish, which adds value.

However, the manuscript in its current form does not fully meet the standards required by Nutrients. The statistical analyses are insufficiently described, the dyspepsia assessment tool lacks methodological grounding, and the construction of the FD index raises serious psychometric concerns. In interpreting the results, the authors draw conclusions that exceed the scope of the data. The study is based on a small, unrepresentative sample, which significantly limits the external validity of the generalisations.

The authors declare the validation of the questionnaire (Cronbach α = 0.857); however:
- there is no information on the structural validity of the tool (e.g., CFA, EFA with clear justification),
- there is no assessment of external validity, such as comparison with Rome IV, GSRS, or other standardised tools,
- the design of the FD Index raises doubts:
 – it consists of simple questions about symptoms,
 – there is no described scoring scale,
 – it is not known whether it is linear,
 – there are no clearly defined cut-off points between LFD and MFD.

Furthermore, the use of K-means clustering to define LFD and MFD groups is problematic:
- clustering is not theoretically justified,
- it may be unstable at n = 102,
- it does not reflect clinical categories of dyspepsia.

I recommend redesigning the FD index with reference to Rome IV and conducting a full psychometric validation.

The study was based on non-representative convenience sampling, which limits the possibility of generalisation. The sample consists mainly of young people, with an average age of 23–30 years. Fifty-three per cent of the MFD group are students, indicating a likely recruitment bias. There is no information on refusals to participate, making it difficult to assess selection bias. The authors should place greater emphasis on the limitations resulting from the sample characteristics.

The authors mainly use χ² tests, t-tests, and multiple linear regression. There are several key problems:

a) Regression with R² = 0.80
Obtaining such a high R² with a small sample (n = 102) and a large number of predictors indicates:
- risk of overfitting the model,
- lack of control for collinearity,
- likely misestimation of effects,
- no reporting of VIF,
- no checking of the normality of residuals,
- no theoretical justification for introducing so many predictors.

The results of the regression model should be interpreted with great caution.

b) Lack of control for confounding variables
In the literature, FD is closely related to stress, psychosocial status, lifestyle, smoking, and H. pylori infection. The failure to consider these variables limits the internal validity of the study.

c) Problems in comparing MFD vs LFD groups
Since the clusters were identified by an algorithm, it is not known:
- what range of FD Index scores differentiated the groups,
- whether the differences between the groups are clinically significant, not just statistically significant.

The authors indicate that fatty foods, fast food, sesame, chamomile, and paprika increase symptoms,while lemon, pear, chia, and garlic alleviate symptoms. These interpretations are too elaborate for observational data because:
- correlation does not imply causation,
- there is no quantitative analysis of consumption, only frequency,
- some of the biological explanations (especially the long paragraphs on the effects of phenolic compounds, CCK, TRPV1) are not based on data from this study but on experimental literature.

The discussion should be shortened and more closely linked to the authors' own results.

The following are missing: the exact response scale in the questionnaire (frequency of consumption, intensity of symptoms), a clear justification for the choice of predictors in the regression, the procedure for dealing with missing data, information on the size of the effects (OR, β with standardised values), and supplementary tables.

The authors suggest that ‘Dietary changes can alleviate FD symptoms.’ While this statement is generally correct, it does not follow directly from their study, which is cross-sectional, does not evaluate interventions, is based on self-reporting, and lacks clinical measurements. The conclusions need to be significantly tempered.

The study concerns dyspepsia but does not use the Rome IV diagnostic standard, which is a serious methodological shortcoming.

Figure 1 and Figure 2 are difficult to read and lack complete axis descriptions. There is no legend explaining the FD Index values. The tables need improvement (e.g., Table 3 contains an imprecise description).
The article presents an interesting topic and provides necessary data on FD in the Colombian population. However, in its current form, the manuscript requires major revisions before it can be considered for pu

Author Response

1. The authors declare the validation of the questionnaire (Cronbach α = 0.857); however: - there is no information on the structural validity of the tool (e.g., CFA, EFA with clear justification),
   - there is no assessment of external validity, such as comparison with Rome IV, GSRS, or other standardised tools, - the design of the FD Index raises doubts:  – it consists of simple questions about symptoms, – there is no described scoring scale, – it is not known whether it is linear,  – there are no clearly defined cut-off points between LFD and MFD. R= Thanks for your observations; these improve our manuscript. In the new version of the manuscript, we included the information as you suggested.
2. Furthermore, the use of K-means clustering to define LFD and MFD groups is problematic:
   - clustering is not theoretically justified, - it may be unstable at n = 102, - it does not reflect clinical categories of dyspepsia. R= We modified the manuscript as suggested by another reviewer, who proposed that we are not assessing clinical categories; instead, in our manuscript, we are addressing dyspeptic symptoms. Additionally, we have included a paragraph in the manuscript explaining the k-means clustering.
3. I recommend redesigning the FD index with reference to Rome IV and conducting a full psychometric validation. R= We have included the information as you suggested.
4. The study was based on non-representative convenience sampling, which limits the possibility of generalisation. The sample consists mainly of young people, with an average age of 23–30 years. Fifty-three per cent of the MFD group are students, indicating a likely recruitment bias. There is no information on refusals to participate, making it difficult to assess selection bias. The authors should place greater emphasis on the limitations resulting from the sample characteristics. R= Thanks for your suggestions, we added this observation to the limitation section.
5. The authors mainly use χ² tests, t-tests, and multiple linear regression. There are several key problems:
6. a) Regression with R² = 0.80
   Obtaining such a high R² with a small sample (n = 102) and a large number of predictors indicates:
   - risk of overfitting the model, - lack of control for collinearity, - likely misestimation of effects, - no reporting of VIF,- no checking of the normality of residuals, no theoretical justification for introducing so many predictors.

R= We appreciate your recommendation. We wrote a justification for this analysis throughout the study. Additionally, we conducted a collinearity analysis prior to the regression model and reported the VIF values as you suggested.

The results of the regression model should be interpreted with great caution.

1. b) Lack of control for confounding variables
   In the literature, FD is closely related to stress, psychosocial status, lifestyle, smoking, and H. pylori infection. The failure to consider these variables limits the internal validity of the study. R We make our inclusion criteria more explicit, indicating that participants with pylori infection and cancer were excluded from the study. In addition, we included the confounding variables as limitations in this study.
2. c) Problems in comparing MFD vs LFD groups
   Since the clusters were identified by an algorithm, it is not known: - what range of FD Index scores differentiated the groups, - whether the differences between the groups are clinically significant, not just statistically significant. R = We modified the manuscript to indicate that this study is not clinical; instead, it is exploratory, and we added a section explaining the range of dyspeptic symptoms (DS) taken into account for this analysis.
3. The authors indicate that fatty foods, fast food, sesame, chamomile, and paprika increase symptoms, while lemon, pear, chia, and garlic alleviate symptoms. These interpretations are too elaborate for observational data because:- correlation does not imply causation,- there is no quantitative analysis of consumption, only frequency,- some of the biological explanations (especially the long paragraphs on the effects of phenolic compounds, CCK, TRPV1) are not based on data from this study but on experimental literature. The discussion should be shortened and more closely linked to the authors' own results.

R = We substantially modified the analysis and discussion section, excluding causation arguments as you proposed.

7. The following are missing: the exact response scale in the questionnaire (frequency of consumption, intensity of symptoms), a clear justification for the choice of predictors in the regression, the procedure for dealing with missing data, information on the size of the effects (OR, β with standardised values), and supplementary tables. R = We added supplementary tables that include the requested information.
8. The authors suggest that ‘Dietary changes can alleviate FD symptoms.’ While this statement is generally correct, it does not follow directly from their study, which is cross-sectional, does not evaluate interventions, is based on self-reporting, and lacks clinical measurements. The conclusions need to be significantly tempered. R= We modified the conclusion section as you suggested.
9. The study concerns dyspepsia but does not use the Rome IV diagnostic standard, which is a serious methodological shortcoming. Figure 1 and Figure 2 are difficult to read and lack complete axis descriptions. There is no legend explaining the FD Index values. The tables need improvement (e.g., Table 3 contains an imprecise description). R = Done
   The article presents an interesting topic and provides necessary data on FD in the Colombian population. However, in its current form, the manuscript requires major revisions before it can be considered for publication.

 

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Dear authors,

Congratulations on the revisions made and on the clarity with which you addressed the reviewers’ comments. The manuscript has been significantly improved, and the results are now easier to interpret.

Author Response

Thanks for your suggestions, now our results are now easier to interpret.

Reviewer 2 Report

Comments and Suggestions for Authors

The revised manuscript has improved substantially, but several points still require clarification to strengthen methodological rigor and temper causal interpretations. In particular, the construction and validation of the dyspeptic symptom index, the clustering procedure, and the regression model would benefit from more detailed reporting. It would also be important to more clearly distinguish between exploratory associations and true predictors, given the cross‑sectional design and limited, convenience-based sample predominantly composed of students. Furthermore, the Discussion and Limitations sections should emphasize potential selection and information biases, the absence of clinical assessment, and the exploratory nature of the findings regarding specific foods and probiotics.

Author Response

1. In particular, the construction and validation of the dyspeptic symptom index, the clustering procedure, and the regression model would benefit from more detailed reporting. R Thanks for your suggestion, we included this infotmation in section 2.3.1.
2.  It would also be important to more clearly distinguish between exploratory associations and true predictors, given the cross‑sectional design and limited, convenience-based sample predominantly composed of students. R We distiguished that our results are exploratory associations through the new version of the manuscript.
3.  Furthermore, the Discussion and Limitations sections should emphasize potential selection and information biases, the absence of clinical assessment, and the exploratory nature of the findings regarding specific foods and probiotics. R We added this information as you suggested in limitation section.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have substantially revised the manuscript in response to the previous review. Several important methodological clarifications and corrections have been introduced, and the overall tone of the manuscript has become more cautious and appropriate for an exploratory, cross-sectional study. In particular, the authors have improved the description of the regression analysis, expanded the limitations section, and moderated causal interpretations in the discussion and conclusions.

The authors present an assessment of the internal consistency of the tool (Cronbach’s α = 0.857) and provide additional information on factor analysis and correlations with the Rome IV criteria. This represents an improvement over the previous version of the manuscript.

Nevertheless, several significant problems remain:

The description of the factor analysis is insufficient. Key information such as the KMO statistic, Bartlett’s test of sphericity, the number of factors extracted, and factor loadings should be clearly presented.

The correlation with the Rome IV criteria appears weak and cannot be considered full external validation. At best, it only supports preliminary convergent validity.

The DS index scoring system remains unclear. The manuscript does not clearly describe:
the response scale used for individual questions,
the method of summing or weighting items,
whether the index is linear,
how to interpret the total scores obtained.

The authors should clearly state that the DS index is not a validated diagnostic tool, but an exploratory scale of symptom severity. The scoring procedure must also be described in a more transparent and detailed manner to ensure the reproducibility of the study.

The authors have modified the interpretation of the results and now state that the study does not address clinical categories of functional dyspepsia, but focuses on clusters of dyspeptic symptoms. This clarification is justified and improves the conceptual consistency of the work.

However, methodological concerns remain:

The theoretical justification for the use of k-means clustering is limited.
With a relatively small sample size (n = 102), the stability of the clusters obtained is uncertain.
No sensitivity or stability analyses (e.g. bootstrapping or alternative clustering solutions) were performed.
The manuscript does not clearly specify the DS index score ranges that differentiate between groups with low and moderate symptom severity.

The authors should clearly state that the clustering results are hypothetical and that the identified clusters should not be treated as clinically significant categories. Providing the point ranges corresponding to each cluster would increase the clarity and interpretability of the results.

The authors addressed several earlier comments by reporting variance inflation factors (VIF), justifying the selection of predictors, and presenting standardised regression coefficients (β). These changes significantly improve the quality of the analysis.

The R² value (~0.80) remains unusually high for a cross-sectional questionnaire-based study conducted on a relatively small sample.

There is a lack of information on the verification of the regression model assumptions (normality and homoscedasticity of residuals).

No cross-validation or sensitivity analysis of the model was performed.

The authors should interpret the results of the regression analysis even more cautiously and clearly indicate that the results obtained may reflect overfitting of the model. At least a brief comment on the diagnosis of residuals should be added.

The authors clarified the exclusion criteria (e.g. H. pylori infection, cancer) and appropriately discussed uncontrolled confounding factors such as stress, lifestyle, and psychosocial factors in the section on study limitations.

Although this does not fully resolve the problem, such transparency is appreciated and acceptable in an exploratory study.

The discussion has been significantly revised, and causal language has been largely removed, which is a clear improvement.

However, some mechanistic explanations remain relatively extensive given the observational nature of the data and the lack of quantitative assessment of food intake. Further shortening and clarification of the discussion would improve the consistency between the data and the interpretation.

The response scales used in the questionnaire (frequency of consumption of products and severity of symptoms) should be clearly described in the Methods section.

Figures 1 and 2, despite improvements, require clearer axis descriptions and legends explaining the DS index values.

The tables should be carefully checked for accuracy and clarity of descriptions (e.g. the description of Table 3).

The supplementary materials are useful but should be clearly referenced in the main text.

The revised version of the manuscript is a significant step forward compared to the original version and addresses many of the reviewer’s comments. Nevertheless, key issues concerning the accuracy of measurement, clustering methodology, and the reliability of the statistical model remain only partially resolved. The study should be clearly presented as exploratory, and conclusions should be formulated with appropriate caution.

Author Response

Dear Reviewer,

Thanks for your observations; these have improved our manuscript. We have modified our manuscript as you suggested and look forward to your final decision. 

Best Regards, 

PhD. Catalina Alatorre

Author Response File: Author Response.pdf