Ultra-Processed Food Intake Is Not Associated with Systemic Inflammation in People with HIV

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study investigates the relationship between the intake of ultra-processed foods (NOVA4) and systemic inflammation, intestinal barrier, cardiovascular metabolic risks, and body composition in HIV-infected individuals with viral suppression. The topic is of clinical value, with comprehensive detection indicators. The core negative results form an important contrast with studies on the general population, and have potential for publication. However, some issues need to be further improved. Specific suggestions are as follows:
1. This study only uses a single 24-hour dietary recall to assess the proportion of ultra-processed food (NOVA4) intake. This method can only reflect the daily dietary situation and cannot represent the habitual intake level over a long period. It also has obvious diurnal variation, underreporting, and social expectation bias.
2. In the generalized additive model (GAM), the authors only corrected for age, gender, race, and CD4 count, but omitted variables such as physical activity, smoking, drinking, and the use of lipid-lowering/antihypertensive/antidiabetic drugs that have a strong impact on inflammation, metabolism, and body composition.
3. The study used GAM to analyze linear and non-linear relationships. It is recommended to supplement complete statistics, parameter basis, model diagnostic results, and use FDR and other methods to correct multiple tests.
4. The title, abstract, and conclusion of the article use strong causal expressions such as "will not lead to" and "no contribution", but this study is a cross-sectional design, which can only indicate "no association", and cannot infer causal relationships, nor can it rule out reverse causality (such as the influence of inflammatory status on dietary choices) or temporal sequence issues.
5. This study only included HIV-infected individuals in a single center in the United States with sustained viral suppression and stable antiviral treatment, and 66% were non-white, with an average BMI at the level of obesity. The population characteristics are highly concentrated. The research results are difficult to generalize to HIV-infected individuals with uncontrolled viral status, unstable treatment, different races/regions, or those with normal weight.
6. The study uses NOVA4 to define ultra-processed foods, but this classification does not distinguish the nutritional density differences within the same category (such as nutritionally fortified ultra-processed foods and high-sugar, high-fat, low-nutrition ultra-processed foods), nor does it conduct sensitivity analyses for continuous variables and different cut-off points for the NOVA4 proportion. It is impossible to determine whether the results are affected by the classification method.
7. The mechanism explanation lacks data support and is overly speculative. It is recommended to revise the mechanism discussion based on the statistical results and avoid unsupported inferences.

Author Response

This study investigates the relationship between the intake of ultra-processed foods (NOVA4) and systemic inflammation, intestinal barrier, cardiovascular metabolic risks, and body composition in HIV-infected individuals with viral suppression. The topic is of clinical value, with comprehensive detection indicators. The core negative results form an important contrast with studies on the general population and have potential for publication. However, some issues need to be further improved. Specific suggestions are as follows:

General response: We sincerely thank the reviewer for the thoughtful evaluation of our manuscript and for the constructive and encouraging comments. We greatly appreciate the recognition of the clinical relevance of our study and the comprehensive nature of the analyses. The reviewer’s insightful suggestions have helped us improve the clarity, rigor, and overall quality of the manuscript. We have carefully addressed each comment in detail below and revised the manuscript accordingly.

 

Comment 1: This study only uses a single 24-hour dietary recall to assess the proportion of ultra-processed food (NOVA4) intake. This method can only reflect the daily dietary situation and cannot represent the habitual intake level over a long period. It also has obvious diurnal variation, underreporting, and social expectation bias.

Response 1: This is an important comment. We acknowledge that the use of a single 24-hour dietary recall may not fully capture habitual dietary intake and we reported in our limitations section potential underreporting, and social desirability bias (line 333-334) and the need for future studies using repeated dietary assessments or longitudinal designs to better capture habitual intake and temporal relationships. The 24-hour recall method remains a widely used and validated approach in nutritional epidemiology, particularly in clinical studies, due to its feasibility and ability to provide detailed quantitative dietary data (https://epi.grants.cancer.gov/dietary-assessment-primer/profiles/recall/). In our study, recalls were conducted by trained dietitian using standardized protocols and analyzed with the NDSR system, which helps improve data quality and consistency.

 

Comment 2: In the generalized additive model (GAM), the authors only corrected for age, gender, race, and CD4 count, but omitted variables such as physical activity, smoking, drinking, and the use of lipid-lowering/antihypertensive/antidiabetic drugs that have a strong impact on inflammation, metabolism, and body composition.

Response 2: We thank the reviewer for this important comment. We agree that lifestyle factors and medication use may influence inflammation, metabolism, and body composition and could act as potential confounders. Our models included key established and known covariates (age, sex, race, and CD4 count). In our sample, smoking status and past medical history were similar between groups, but alcohol use differed between groups and may contribute to residual confounding. We also acknowledge that other variables, including physical activity and cardiometabolic medication use may impact our outcomes. These limitations are now explicitly addressed in the manuscript (line 331-333).

 

Comment 3: The study used GAM to analyze linear and non-linear relationships. It is recommended to supplement complete statistics, parameter basis, model diagnostic results, and use FDR and other methods to correct multiple tests.

Response 3: We thank the reviewer for this helpful comment. In response, we revised the Methods and Results sections to provide more complete reporting of the generalized additive models. Specifically, we now report estimated degrees of freedom (EDF) for nonlinear smooth terms to summarize model complexity, along with regression coefficients for linear terms. In addition, given the number of outcomes examined, we implemented correction for multiple comparisons using the Benjamini–Hochberg false discovery rate (FDR) procedure, applied to the primary dietary exposure variables across outcomes. Results tables have been updated to report both nominal p‑values and FDR‑adjusted q‑values. Some associations were attenuated after correction, and these changes are reflected in the revised Results and figures.

 

Comment 4: The title, abstract, and conclusion of the article use strong causal expressions such as "will not lead to" and "no contribution", but this study is a cross-sectional design, which can only indicate "no association", and cannot infer causal relationships, nor can it rule out reverse causality (such as the influence of inflammatory status on dietary choices) or temporal sequence issues.

Response 4: Thank you for this comment. We agree that causal language should be avoided given the cross-sectional design of our study. We have revised the title to replace causal wording (“does not contribute”) with more appropriate terminology (“is not associated with”). We also revised the conclusion of our abstract. In addition, we carefully reviewed the manuscript to ensure that the findings are consistently presented using non-causal language appropriate for an observational cross-sectional study design. We mentioned in the limitations that the cross-sectional nature of the study precludes causal inference.

 

Comment 5: This study only included HIV-infected individuals in a single center in the United States with sustained viral suppression and stable antiviral treatment, and 66% were non-white, with an average BMI at the level of obesity. The population characteristics are highly concentrated. The research results are difficult to generalize to HIV-infected individuals with uncontrolled viral status, unstable treatment, different races/regions, or those with normal weight.

Response 5: We thank the reviewer for this insightful observation. We agree that the characteristics of our study population may limit the generalizability of our findings. However, in our clinic population (and nationally) most people living with HIV who are in care have undetectable viral load and are on stable ART except for switches due to tolerability concerns.  We have revised the limitations to explicitly acknowledge that.

 

Comment 6: The study uses NOVA4 to define ultra-processed foods, but this classification does not distinguish the nutritional density differences within the same category (such as nutritionally fortified ultra-processed foods and high-sugar, high-fat, low-nutrition ultra-processed foods), nor does it conduct sensitivity analyses for continuous variables and different cut-off points for the NOVA4 proportion. It is impossible to determine whether the results are affected by the classification method.

Response 6: We thank the reviewer for this thoughtful comment. We acknowledge that the NOVA classification system does not account for heterogeneity in nutritional quality within ultra-processed foods. This is a limitation explicitly acknowledged in our manuscript. Regarding the use of cut-off points, we would like to clarify that our primary analyses did not rely on categorical classification. While participants were stratified by median %NOVA4 intake for descriptive purposes (Table 1-2), %NOVA4 was modeled as a continuous variable in GAMs, allowing us to evaluate both linear and nonlinear associations. This approach reduces dependence on arbitrary thresholds and serves as an implicit sensitivity analysis to different cut-offs.

 

Comment 7: The mechanism explanation lacks data support and is overly speculative. It is recommended to revise the mechanism discussion based on the statistical results and avoid unsupported inferences.

Response 7: We thank the reviewer for this comment. After carefully reviewing the manuscript, we found that most of the Discussion is grounded in the statistical results and comparison with prior literature. We revised statements to more clearly present them as possible interpretations rather than data-driven mechanistic conclusions. We agree that mechanistic interpretations should be made cautiously and supported by available evidence.  However, we would like to highlight that data examining the relationship between food processing and comprehensive inflammatory and gut integrity biomarkers in people with HIV are currently limited. Accordingly, our findings contribute to an emerging area of research.

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript addresses a relevant and topical issue, namely the role of ultra-processed foods in modulating inflammation and cardiometabolic risk in people living with HIV, an area that remains largely unexplored compared to the general population. The rationale is well constructed and the introduction adequately frames the problem, highlighting the knowledge gap and justifying the use of the NOVA classification as a complementary approach to traditional nutritional indices.

The study design is appropriate for an exploratory analysis and the cohort size is adequate in comparison with the available literature in this field. Particularly commendable is the multidimensional characterisation of the participants, which includes detailed measures of body composition, inflammatory biomarkers, cardiometabolic parameters and indicators of gut integrity. The use of generalised additive models also represents a methodological strength, allowing for the exploration of non-linear relationships between dietary exposure and outcomes.

The results are presented clearly and coherently. The absence of associations between consumption of ultra-processed foods and markers of inflammation, metabolism and body composition is reported transparently and supported by statistical analyses. This finding, which contrasts with that observed in the general population, is plausibly discussed by the authors, who suggest a possible predominant effect of HIV-related chronic inflammation capable of masking the impact of diet.

However, there are some aspects that require clarification or further investigation. The cross-sectional nature of the study severely limits the causal interpretation of the results, and this point, although mentioned, would merit a more critical discussion, particularly in light of the authors’ initial hypothesis. Furthermore, dietary assessment based on 24-hour recall may introduce misclassification and underestimation biases, particularly for ultra-processed foods, and this may have attenuated any associations. The NOVA classification, whilst widely used, contains elements of subjectivity that could influence the reproducibility of the results.

A further limitation concerns the possible presence of residual confounding factors. Although the models are adjusted for relevant variables such as age, sex, race and CD4 count, other behavioural and socio-economic determinants of diet and inflammatory status may not have been adequately accounted for. Furthermore, the lack of stratified analyses, for example by sex or duration of infection, reduces the ability to identify subgroups in which the effect of diet might be more pronounced.

The discussion is generally balanced, but could be strengthened by a more in-depth comparison with studies using different dietary approaches, such as overall dietary patterns or diet quality, to better contextualise the results. The high average calorie intake reported in the cohort also warrants more critical consideration, as it could indicate problems with data collection or specific characteristics of the sample.

In conclusion, the study is well conducted and makes an original contribution to the literature, suggesting that the role of ultra-processed foods in people with HIV may differ from that in the general population. However, certain methodological and interpretative limitations need to be discussed in greater detail to strengthen the robustness of the conclusions.

Author Response

The manuscript addresses a relevant and topical issue, namely the role of ultra-processed foods in modulating inflammation and cardiometabolic risk in people living with HIV, an area that remains largely unexplored compared to the general population. The rationale is well constructed and the introduction adequately frames the problem, highlighting the knowledge gap and justifying the use of the NOVA classification as a complementary approach to traditional nutritional indices.

 

The study design is appropriate for an exploratory analysis and the cohort size is adequate in comparison with the available literature in this field. Particularly commendable is the multidimensional characterisation of the participants, which includes detailed measures of body composition, inflammatory biomarkers, cardiometabolic parameters and indicators of gut integrity. The use of generalised additive models also represents a methodological strength, allowing for the exploration of non-linear relationships between dietary exposure and outcomes.

 

The results are presented clearly and coherently. The absence of associations between consumption of ultra-processed foods and markers of inflammation, metabolism and body composition is reported transparently and supported by statistical analyses. This finding, which contrasts with that observed in the general population, is plausibly discussed by the authors, who suggest a possible predominant effect of HIV-related chronic inflammation capable of masking the impact of diet.

General response: We sincerely thank the reviewer for the thorough and insightful evaluation of our manuscript, as well as for the constructive and encouraging comments. We greatly appreciate the recognition of the relevance and novelty of our study, particularly in addressing an understudied area in people living with HIV, and for highlighting the strengths of our multidimensional approach and analytical methods.

 

The reviewer’s thoughtful suggestions have helped us further improve the clarity, interpretation, and contextualization of our findings. We have carefully considered each point raised and revised the manuscript accordingly, as detailed in our responses below.

Comment #1: However, there are some aspects that require clarification or further investigation. The cross-sectional nature of the study severely limits the causal interpretation of the results, and this point, although mentioned, would merit a more critical discussion, particularly in light of the authors’ initial hypothesis. Furthermore, dietary assessment based on 24-hour recall may introduce misclassification and underestimation biases, particularly for ultra-processed foods, and this may have attenuated any associations. The NOVA classification, whilst widely used, contains elements of subjectivity that could influence the reproducibility of the results.

Response 1: We thank the reviewer for the positive and constructive feedback. We agree that the cross-sectional design limits causal inference and have clarified this in our limitations. We also acknowledge the limitations of 24-hour dietary recall and NOVA classification subjectivity, including potential misclassification and underestimation of ultra-processed food intake, and have also explicitly acknowledged this in our limitation.

 

Comment 2: A further limitation concerns the possible presence of residual confounding factors. Although the models are adjusted for relevant variables such as age, sex, race and CD4 count, other behavioural and socio-economic determinants of diet and inflammatory status may not have been adequately accounted for. Furthermore, the lack of stratified analyses, for example by sex or duration of infection, reduces the ability to identify subgroups in which the effect of diet might be more pronounced.

Response 2: We thank the reviewer for this important comment. We agree that lifestyle factors and socioeconomic factors use may influence inflammation, metabolism, and body composition and could act as potential confounders. Our models included key established and known covariates (age, sex, race, and CD4 count). In our sample, smoking status and past medical history were similar between groups, but alcohol use differed between groups and may contribute to residual confounding. We also acknowledge that other variables may impact our outcomes. These limitations are now explicitly addressed in the manuscript (line 331-333).

 

Comment 3: The discussion is generally balanced, but could be strengthened by a more in-depth comparison with studies using different dietary approaches, such as overall dietary patterns or diet quality, to better contextualise the results. The high average calorie intake reported in the cohort also warrants more critical consideration, as it could indicate problems with data collection or specific characteristics of the sample.

Response 3: We thank the reviewer for this valuable suggestion. We had included a comparison with studies using alternative dietary assessment approaches. Regarding high reported caloric intake in our cohort, this point has now been incorporated into the limitations section.

 

In conclusion, the study is well conducted and makes an original contribution to the literature, suggesting that the role of ultra-processed foods in people with HIV may differ from that in the general population. However, certain methodological and interpretative limitations need to be discussed in greater detail to strengthen the robustness of the conclusions.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript has been revised in accordance with the reviewers' comments and has now met the acceptance criteria.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have addressed all my concerns