Small-Scale Farming, Pesticide Exposure, and Respiratory Health: A Cross-Sectional Study in Bolivia

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

General comment:

In general, the manuscript is well written – logical layout, concise background, clear aim, descriptive methodology (including statistical analysis of data, appropriate use of figures and tables in the results and elaborate discussion of findings (including  study limitations).

Major issue/concern:

p.3 line 108: Data collection took place between October 2016 and January 2017 and raises the question on whether the findings reported are still valid eight years later. It is not clear why there is such a long delay between data collection and reporting (manuscript).

Other comments:

p.2 line 75: The authors mention that few studies evaluated potential effects of pesticide exposure on respiratory health in small-scale farming in Bolivia without referencing these studies.

p.3 line 89: Consider using the word ‘conditions’ and not weather ‘characteristics’.

p.3 line 113 and p.4  Section 2.3.2: What competency did the field workers have in assessing respiratory health signs and symptoms? As examples, if they are not clinically trained, how did they know the difference between chronic cough and phlegm or able to distinguish nasal allergy from other (respiratory) allergies?

p.3 line 113: Additionally, spirometry …. (delete ‘a’)

Table 1: For height, farmers – please check the reported value 158.2?

Table 1: Intoxication, consider the word use (could emit the word) as signs/symptoms after exposure will suffice. Intoxication more associated with alcohol consumption.

Table 1: footnote – What is meant by ‘mask and two others’? Please specify/clarify.

p.8 line 256: What contributed to 13.8% (60) spirometry tests being excluded (as inferior quality)?

p.8 line 258: FEV1/FVC is a ratio and not an index.

p.8 line 270: Was the fuel type used for cooking considered as a confounder?

p.8 Table 4: Multiple linear regression models, should be named Table 5 (not Table 5) see also in text line 274.

p.9 line 294: Is it possible to elaborate on the type of PPE used?

Author Response

Comments 1: General comment:

In general, the manuscript is well written – logical layout, concise background, clear aim, descriptive methodology (including statistical analysis of data, appropriate use of figures and tables in the results and elaborate discussion of findings (including  study limitations).

Response 1: Thank you very much for the review. It helped us improve the presentation of the document.

Major issue/concern:

Comments 2: p.3 line 108: Data collection took place between October 2016 and January 2017 and raises the question on whether the findings reported are still valid eight years later. It is not clear why there is such a long delay between data collection and reporting (manuscript).

Response 2: We appreciate the clarification on this point. The fieldwork was completed in 2017, and the report with the general results (mainly descriptive) was sent to local authorities that same year. However, we were unable to work on the scientific publication until several years later. This was due to the team's workload and personal issues, such as the change of country of residence of some of the authors.

While we recognize that several years have passed since the fieldwork was completed, the situation in this region regarding pesticide management has remained relatively stable. No additional programs have been implemented to reduce pesticide use or promote more responsible use, nor have any regulations or legislation on the subject been proposed or approved. On the contrary, during the last year, the introduction of some mining activities near the study area has been debated (https://correodelsur.com/local/20250327/sedes-chuquisaca-rechaza-explotacion-minera-en-el-rio-chico-es-un-atentado-contra-la-salud.html), which could further endanger the health of the communities. With this background, we believe the results are still relevant. On the other hand, no new research projects on this topic have been implemented at the local level, and very few have addressed the issue of pesticides and health at the national level. Therefore, we believe it is important to publish these results to provide scientific evidence from low- and middle-income countries. Having this information published and openly accessible will also provide a baseline for assessing changes in pesticide use practices for future studies.

Other comments:

Comments 3: p.2 line 75: The authors mention that few studies evaluated potential effects of pesticide exposure on respiratory health in small-scale farming in Bolivia without referencing these studies.

Response 3: That's correct, thank you very much for the observation. We've added a reference (now Reference 21) to a scoping review article from 2000 to 2022 exploring human pesticide exposure in Bolivia. They identified 39 peer-reviewed articles, 27 reports/documents on Bolivian regulations, and 12 other documents. Of these, the majority focused on acute symptoms of poisoning, and none of the studies were conducted in our study area.

Comments 4: p.3 line 89: Consider using the word 'conditions' and not weather 'characteristics'.

Response 4: Thank you very much for the suggestion. We have incorporated it.

Comment 5: p.3 line 113 and p.4  Section 2.3.2: What competency did the field workers have in assessing respiratory health signs and symptoms? As examples, if they are not clinically trained, how did they know the difference between chronic cough and phlegm or able to distinguish nasal allergy from other (respiratory) allergies?

Response 5: Thank you very much for the observation. Very important!. The lung function spirometry test was performed by a final-year medical school graduate who received prior training and instruction. All recommended protocols for this test were also followed. We added this information: “The spirometer was performed by a health professional previously trained in lung function, spirometry, and equipment management.” (Lines 180-182).

Concerning assessing respiratory symptoms, we used the Spanish version of the European Community Respiratory Health Survey (ECRHS). This questionnaire, being a standardized instrument, should be administered in the same way in all studies that use it to facilitate subsequent comparisons. In this regard, the protocol for administering the questionnaire states: “The interviewers will ask the questions exactly as they are, using the exact wording and order, as written on the survey questionnaire, avoiding any hints or verbal clues. The ideal interviewer, either clinical or non-clinical staff, should be neutral and non-judgemental!. Never surprised or disappointed, never approving or disapproving, never feeling embarrassed about personal questions, never asking leading questions but sticking to the exact wording and order as on the survey questionnaire.”

Considering this, the interviewers were trained to administer the questionnaire to avoid biasing the responses and to collect the information as people reported it.

We have clarified this information in the manuscript: “All recommendations for administering the questionnaire were followed”.  (Line 169)

Comment 6: p.3 line 113: Additionally, spirometry …. (delete 'a')

Response 6: We have corrected the error, thank you very much!

Comment 7:Table 1: For height, farmers – please check the reported value 158.2?

Response 7: We have corrected it in order to be uniform in the report. thank you very much!

Comment 8: Table 1: Intoxication, consider the word use (could emit the word) as signs/symptoms after exposure will suffice. Intoxication more associated with alcohol consumption.

Response 8: Thank you very much! We've corrected the term “intoxication” in the document everywhere it was mentioned.

Comment 9: Table 1: footnote – What is meant by 'mask and two others'? Please specify/clarify.

Response 9: Thank you for your feedback. We have clarified the categorization we used to describe personal protective equipment. For the general description, we consider the use of at least three personal protective equipment, one of which is a face mask (respiratory protection), as this can mitigate the entry of pesticides through the respiratory tract and their potential direct effects on respiratory health.

Comment 10: p.8 line 256: What contributed to 13.8% (60) spirometry tests being excluded (as inferior quality)?

Response 10: The factors we identified as contributing to poor spirometry quality were varied. Some included: Difficulty understanding the test, especially in older individuals or those with language difficulties. In some cases, individuals exceeded the number of attempts, failing to exhale with the necessary force. It also happened in some cases (younger individuals) that the test was a little embarrassing because it was the first time they had seen such a test. However, we did not identify a pattern that should be taken into account in subsequent analyses; we believe it was due to the more or less random distribution.

We clarified this in lines 278-281:

“Among the main reasons attributed to low-quality spirometries were difficulties in understanding the test, especially in older people or those with language difficulties, as well as some cases where the recommended number of attempts was exceeded without being able to exhale with the necessary force.”

Comment 11: p.8 line 258: FEV1/FVC is a ratio and not an index.

Response 11: Thank you very much! We've corrected the term

Comment 12: p.8 line 270: Was the fuel type used for cooking considered as a confounder?

Response 12: Thank you for the comment. Initially, exposure to cooking or heating fuels was not incorporated as a confounding variable in the linear regression model. Thank you for the observation. In the new version of the manuscript, I have reanalyzed the data, incorporating this variable as a confounding factor, as we did in the logistic regression model (Table 4).

Comment 13: p.8 Table 4: Multiple linear regression models, should be named Table 5 (not Table 5) see also in text line 274.

Response 13: Yes! Thanks for your observation! We corrected the mistake.

Comment 14: p.9 line 294: Is it possible to elaborate on the type of PPE used?

Response 14: Thank you for your comment. In the new version of the manuscript, we've described the frequency of the personal protective equipment explored in the study (Lines 233-235).

 

 

Reviewer 2 Report

Comments and Suggestions for Authors

The introduction provides a solid foundation on pesticide exposure and respiratory health, but it could better highlight the study's novelty in the Bolivian context, where data are scarce. The abstract summarizes key findings but fails to emphasize the dual nature of farming's effects (protective for nasal allergies but harmful for lung function). The authors should ensure all in-text citations (e.g., [1-18]) match the reference list, and update older references (e.g., [7], [18]) with recent literature (2023–2025) to reflect current evidence. The abstract could also briefly mention the study's policy implications, such as the need for pesticide regulation and farmer education.

The methods are well-detailed, but the exposure assessment could be strengthened by incorporating biomonitoring (e.g., urinary pesticide metabolites) to complement the IEB score. The spirometry protocol is robust, but post-bronchodilator tests would enhance obstructive pattern detection. For statistical analysis, consider advanced techniques like:

Machine learning (e.g., random forests to identify non-linear relationships between exposure and outcomes).

Structural equation modeling to explore mediating factors (e.g., PPE use mitigating effects).

Propensity score matching to reduce confounding by balancing farmer/non-farmer characteristics.
Clarify how missing data (e.g., 9 missing IEB scores) were handled, and justify the IEB cutoff (median = 2) with sensitivity analyses.

 

The results are clear, but I believe the discussion could delve deeper into mechanisms. For example:

1. Please link the lower FEV1/FVC in farmers to chronic airway inflammation from pesticide exposure, citing oxidative stress pathways [8,30].
2. Expand on the "farm effect" (reduced nasal allergies) by discussing microbiome diversity [33,34] or endotoxin exposure.
3. Address limitations more explicitly: recall bias in self-reported symptoms, cross-sectional design (cannot infer causality), and potential exposure misclassification (non-farmers near crops).

The authors should compare findings with similar Latin American studies [11,27] to contextualize results. The discussion of pesticide types (e.g., herbicides vs. insecticides) and their specific effects would add depth.

Although the conclusion reiterates findings well but it should propose actionable steps (e.g., "Implement PPE training programs" or "Policy reforms to restrict highly toxic pesticides"). The authors should nsure that all in-text citations are listed (e.g., check [19] in-text vs. references). Adhere to journal guidelines. Remove duplicates (e.g., Table 4 appears twice in the PDF). Not compulsory, but it will be good if the authors use reference managers (e.g., EndNote) to automate formatting.

Author Response

Comment 1: The introduction provides a solid foundation on pesticide exposure and respiratory health, but it could better highlight the study's novelty in the Bolivian context, where data are scarce. The abstract summarizes key findings but fails to emphasize the dual nature of farming's effects (protective for nasal allergies but harmful for lung function). The authors should ensure all in-text citations (e.g., [1-18]) match the reference list, and update older references (e.g., [7], [18]) with recent literature (2023–2025) to reflect current evidence. The abstract could also briefly mention the study's policy implications, such as the need for pesticide regulation and farmer education.

Response 1: Thank you very much for your comments and suggestions. It helped us improve the presentation of the document.

We have strengthened the introduction, emphasizing the contribution of the study, especially in the Bolivian context.

We review the abstract to highlight the results found for farmers in relation to nasal allergies and lung function. We also incorporate, in a more explicit way, the implications of public policies.

We have also reviewed the use of bibliographic references and updated older references whenever possible. We have retained the reference to Jors's study (Reference 19 - previously 18) since we used the instrument associated with that publication.

Comment 2: The methods are well-detailed, but the exposure assessment could be strengthened by incorporating biomonitoring (e.g., urinary pesticide metabolites) to complement the IEB score. The spirometry protocol is robust, but post-bronchodilator tests would enhance obstructive pattern detection. For statistical analysis, consider advanced techniques like:

Machine learning (e.g., random forests to identify non-linear relationships between exposure and outcomes).

Structural equation modeling to explore mediating factors (e.g., PPE use mitigating effects).

Propensity score matching to reduce confounding by balancing farmer/non-farmer characteristics.

Response 2: I greatly appreciate your comments and suggestions. Implementing this study has not been easy, as it required significant fieldwork, a limited budget, and research equipment, as well as some somewhat substandard conditions that prevented further measurement. I want to take this opportunity to mention that the initial project idea included taking urine samples (a less invasive sample, and more accepted by the population) to identify pesticide metabolites. Unfortunately, the logistics from the rural area to the city, while maintaining optimal conditions for sample handling, were quite challenging. It was also not possible to conduct the analyses in another town (Santa Cruz), which had the necessary laboratory equipment, due to limited staff training. Due to these issues, as we mentioned, it is somewhat difficult to use other exposure measurements. However, some progress has been made, and my university is currently strengthening some laboratory techniques to allow for further measurement. Therefore, we will consider strengthening pesticide measurement, but also outcome measurement in a future study. In that sense, this study will serve as an essential background (and we will also incorporate all the lessons learned from this project) to advance this important topic.

We reviewed the study's limitations section to make these aspects explicit:

“Future studies can complement exposure measurement by evaluating the IEB and some biological measurements that will strengthen a more comprehensive exposure assessment.” (lines 398-400)

Regarding advanced statistical techniques, we find the proposal very interesting. We have reviewed their applications, uses, and assumptions, and believe that applying these could help deepen this analysis, as well as future ones. However, considering the objective of this manuscript and the limitations of some resources, we believe that the linear regression used, although a very basic analysis, is sufficient to achieve the proposed objective. We have also verified that the assumptions for this analysis are met.

However, we have reflected this potential limitation in the analysis in the limitations section of our study, so that it can be considered from the outset in future studies.

“Future studies could also consider more advanced statistical analyses that would allow for a more precise analysis of the relationship between exposure and outcome, taking into account potentially mediating factors and other confounding factors.” (Lines 406-409)

Comment 3: Clarify how missing data (e.g., 9 missing IEB scores) were handled, and justify the IEB cutoff (median = 2) with sensitivity analyses.

Response 3: Thank you for your feedback. Missing values were excluded from the analysis. In this new version of the manuscript, we have clarified this information and revised the sample sizes in the analyses.

“Missing values were excluded from further analysis”. (Line 193), and changes were made in Table 2 and Table 3.

 

Since no standard cutoff point for pesticide exposure was reported, the IEB score was used to categorize individuals with high or low pesticide exposure based on the distribution of the data (median). This was clarified in the text:

“For further analysis, high and low exposure to pesticides were categorized using the distribution of the data as a criterion, defining individuals with high pesticide exposure as those with an IEB score greater than the median (2 points).” (Lines 148-150)

Now, we performed a sensitivity analysis considering cut-off points 1 and 3. The results maintained the same trend. They are now included as supplementary material. This was clarified in the manuscript (methods and results sections).

“A sensitivity analysis was performed considering different cut-off points for IEB (more than one and more than three points in the IEB score) to analyze the consistency of the results.” (Lines 206-208)

“Sensitivity analysis with IEB cut-off points of 1 and 3 maintained the same trend (Supplementary Material).” (Lines 287-289)

Comment 4: The results are clear, but I believe the discussion could delve deeper into mechanisms. For example:

1. Please link the lower FEV1/FVC in farmers to chronic airway inflammation from pesticide exposure, citing oxidative stress pathways [8,30].
2. Expand on the "farm effect" (reduced nasal allergies) by discussing microbiome diversity [33,34] or endotoxin exposure.
3. Address limitations more explicitly: recall bias in self-reported symptoms, cross-sectional design (cannot infer causality), and potential exposure misclassification (non-farmers near crops).

Response 4: Thank you very much for the suggestions. You bring up very relevant issues for the discussion, which were incorporated into this section. (1: lines 344-348; 2:lines 365-374; 3:lines 386-392)

Comment 5: The authors should compare findings with similar Latin American studies [11,27] to contextualize results. The discussion of pesticide types (e.g., herbicides vs. insecticides) and their specific effects would add depth.

Response 5: Thank you very much for the suggestions. We have incorporated the points mentioned into the discussion and have related our results to other Latin American studies, especially regarding working conditions and pesticide handling, which are common in Latin America. (lines 349-355)

Comment 6: Although the conclusion reiterates findings well but it should propose actionable steps (e.g., "Implement PPE training programs" or "Policy reforms to restrict highly toxic pesticides"). The authors should nsure that all in-text citations are listed (e.g., check [19] in-text vs. references). Adhere to journal guidelines. Remove duplicates (e.g., Table 4 appears twice in the PDF). Not compulsory, but it will be good if the authors use reference managers (e.g., EndNote) to automate formatting.

Response 6: Thank you very much for your comments and suggestions. We have modified the conclusions to be more specific in the study's final recommendations (424-427).

We have also reviewed the use of the bibliography. Thank you for pointing out the inconsistencies and errors. It was modified in the new version of the manuscript.

 

Reviewer 3 Report

Comments and Suggestions for Authors

This study analyzed the relationship between pesticide exposure with respiratory symptoms, and lung function, among small-scale farmers compared to non-farmers in rural communities of Bolivia.  The authors provide a solid justification for the research.  The methods are fairly well described.  The results are presented clearly.  The discussion is appropriate for the results.

 

The authors could improve their presentation by addressing several concerns.

 

1. Section 2.2 (line 97 ff):  The authors do not provide any information about participation / refusal numbers or rates.  (Finally mentioned on Line 333 of the Discussion.)

 

2. Line 101: “vicinal joins” meaning?

 

3. Line 200: “authorization of the local authorities in the study communities was requested.”  Was it obtained.

 

4. The authors include the use of personal protective equipment in their constructed measure of pesticide exposure – the Individual Exposure Burden (IEB). They further discuss the use of PPE (line 294) to explain reports of pesticide poisoning symptoms.  This might be a bit circular.  More importantly, the authors should detail the types of PPE that were used.  It would be helpful to know whether the PPE was in good repair and being used properly.

 

5. Line 209: “Four hundred ninety-one individuals participated in the study comprising 277 farmers (56.4%) and 214 non-farmers (43.6%).” These 491 individuals resided in 215 households (line 103).   How was clustering accounted for in the analysis?

 

6. Line 256: “A total of 375 of 435 spirometry (86.2 %) were considered of sufficient quality (Quality A) and included in the subsequent analyses. Of this, 399 (94%) and 244 (57%) participants reported an FEV1/FVC% index above 70 and 80% respectively.” This is confusing – how did 375 valid spirometry tests result in 399 results?

Author Response

This study analyzed the relationship between pesticide exposure with respiratory symptoms, and lung function, among small-scale farmers compared to non-farmers in rural communities of Bolivia.  The authors provide a solid justification for the research.  The methods are fairly well described.  The results are presented clearly.  The discussion is appropriate for the results.

The authors could improve their presentation by addressing several concerns.

Comment 1: Section 2.2 (line 97 ff):  The authors do not provide any information about participation / refusal numbers or rates.  (Finally mentioned on Line 333 of the Discussion.)

Response 1: Thank you very much for all the comments and suggestions. It helped us improve the presentation of the document.

Concerning Comment 1, we've included the information on the number of people invited to participate and the response rate in section 2.2. 

“Approximately 645 persons were invited, resulting in a participation of 76%.” (Lines 110-111)

Comment 2: Line 101: "vicinal joins" meaning?

Response 2: Thanks for the comment. I think the best term for this would be “neighborhood councils” . We're referring to the organization of a group of homes in a specific area. We corrected this mistake.

Comment 3: Line 200: "authorization of the local authorities in the study communities was requested."  Was it obtained.

 Response 3: That's correct. We corrected the term.

Comment 4: The authors include the use of personal protective equipment in their constructed measure of pesticide exposure – the Individual Exposure Burden (IEB). They further discuss the use of PPE (line 294) to explain reports of pesticide poisoning symptoms.  This might be a bit circular.  More importantly, the authors should detail the types of PPE that were used.  It would be helpful to know whether the PPE was in good repair and being used properly.

 Response 4: You are absolutely right. In the revised version of the manuscript, we have described the percentage of use of each of the personal protective equipment items asked about (lines 233-235) In that sense, we have also modified the discussion to relate the high percentage of signs and symptoms after handling pesticides with the use of PPE, specifically low use of face masks (38%) and gloves (30%) (lines 321-323).

Although we were unable to verify the correct use of PPE, in our study, we explored the reasons for not using it, indicating that in most cases, they do not use it because it is uncomfortable. However, not all people answered these last questions.

Comment 5: Line 209: "Four hundred ninety-one individuals participated in the study comprising 277 farmers (56.4%) and 214 non-farmers (43.6%)." These 491 individuals resided in 215 households (line 103).   How was clustering accounted for in the analysis?

 Response 5: Thank you for your observation. The Intraclass Correlation Coefficient (ICC) was explored for the main outcomes. The ICC ranged from 0.005 to 0.016. Given the low value, which was expected due to the number of households, the results were not adjusted for the ICC, as it is suggested that low ICC values do not significantly impact the results (Bangdiwala, 2011)

We have incorporated the ICC values calculated in the study:

“Intraclass correlation coefficient (ICC) was computed to analyze the relationship of cases within a cluster (household) for each of the main outcomes. ICC ranged from 0.005 for FEV1 and 0.016 for FEV1/FVC%). Considering the low correlation between observations, no adjustments by cluster sampling were performed.”(25) (Lines 195-198)

 

25. Bangdiwala, S.I. Are injuries correlated? Interpreting the intraclass correlation coefficient. International journal of injury control and safety promotion 2011, 18, 169-171.

Comment 6: Line 256: "A total of 375 of 435 spirometry (86.2 %) were considered of sufficient quality (Quality A) and included in the subsequent analyses. Of this, 399 (94%) and 244 (57%) participants reported an FEV1/FVC% index above 70 and 80% respectively." This is confusing – how did 375 valid spirometry tests result in 399 results?

Response 6: We really appreciate your feedback! We have corrected the error.

“Considering high-quality spirometry, 348 (93%) and 213 (57%) participants reported an FEV1/FVC% ratio above 70 and 80%, respectively.” (line 282)

 

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have done the required correction

Author Response

Comment 1: The authors have done the required correction

Response 1: Thank you very much for the feedback. This has helped us significantly improve our research report.

Reviewer 3 Report

Comments and Suggestions for Authors

Thank you for the revisions.  An excellent study.

Author Response

Comment 1: Thank you for the revisions.  An excellent study.

Response 1: Thank you very much for the feedback. This has helped us significantly improve our research report.