Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

​1. Core Content​

This study investigates the relationship between allergic asthma episodes and concentrations of six tree pollen types (Cupressaceae, Olea, Platanus, Pinus, Ulmus, Populus) and six atmospheric pollutants (O₃, PM₁₀, PM₂.₅, NO₂, CO, SO₂) across urban municipalities in Madrid (2014–2017). Using multiple linear regression, the authors report statistically significant associations (adjusted R² >30%, p<0.0001), identifying Olea pollen and O₃ as the most frequent contributors to asthma exacerbations, with stronger correlations in urban areas. The work highlights the dual role of urban trees as both public health assets and sources of allergenic pollen interacting with pollutants.

 

​2. Major Limitations​

While the study addresses an important public health issue, several methodological and contextual limitations reduce the robustness and generalizability of its conclusions:

 

​A. Methodological Concerns​

​Multicollinearity and Variable Selection:

 

Pollen and pollutant concentrations often correlate with meteorological factors (e.g., temperature, humidity), which were not included. This omission risks confounding results, as weather influences both pollen dispersal and pollutant distribution.

No justification is provided for the backward stepwise regression method, which can overfit models. A sensitivity analysis (e.g., variance inflation factors) is needed to confirm the absence of multicollinearity among predictors.

​Data Limitations:

 

Asthma episodes were recorded only on weekdays, ignoring weekend trends. This introduces temporal bias, as pollen/pollutant exposures and healthcare-seeking behaviors may differ on weekends.

Pollen and pollutant data were aggregated at the municipal level, potentially masking micro-scale variations in exposure. Spatial interpolation methods (e.g., kriging) could improve granularity.

​Thresholds for Significance:

 

The use of adjusted R² >30% as a benchmark for "medium-high correlation" is arbitrary and inadequately justified. In ecological studies, low R² values (even <30%) can still indicate meaningful associations if effect sizes are consistent.

​B. Contextual and Analytical Gaps​

​Urban-Rural Dichotomy:

 

All municipalities studied are classified as urban (Table 1), yet the conclusion emphasizes rural-urban differences. This misalignment undermines claims about rurality’s role. A reanalysis using continuous metrics (e.g., population density gradients) would strengthen spatial inferences.

​Pollen-Pollutant Interactions:

 

The mechanistic link between pollen and pollutants (e.g., pollutant-induced allergenicity changes) is discussed but not quantified. Including allergen potency data (e.g., Bet v 1 levels for Olea) would clarify biological relevance.

​Temporal Resolution:

 

The analysis uses daily data but does not explore lag effects (e.g., delayed asthma responses to pollen exposure). Time-series analyses with distributed lag models could better capture dynamic relationships.

​C. Presentation and Reporting Issues​

​Missing Information:

 

Appendix Tables A/B (model equations) are referenced but not provided, hindering reproducibility.

The exclusion criteria for outliers (studentized residuals >3 SD) are mentioned but lack documentation (e.g., how many data points were removed?).

​Inconsistent Terminology:

 

PM₂.₅ is inconsistently formatted (e.g., “PM2,5” in Tables). Standardize notation to align with environmental science conventions.

References to “clinical prevalence” (Introduction) are misleading, as the study uses healthcare visits, not population-level prevalence data.

​Discussion of Conflicting Evidence:

 

The manuscript cites studies linking PM₂.₅ and NO₂ to asthma but does not reconcile why these pollutants were less prominent in their models. Acknowledging data gaps (e.g., missing PM₂.₅ measurements at some stations) would improve transparency.

​Recommendations for Revision​

Address multicollinearity and refine regression methods.

Incorporate meteorological variables and lag effects.

Reassess urban-rural comparisons using continuous spatial metrics.

Provide missing appendix tables and outlier removal details.

Clarify limitations related to weekday-only data and PM₂.₅ availability.

This study offers valuable insights into pollen-pollutant interactions but requires methodological strengthening and clearer contextual framing to fully support its conclusions.

Author Response

Reviewer 1

Comments and Suggestions for Authors

​1. Core Content​

This study investigates the relationship between allergic asthma episodes and concentrations of six tree pollen types (Cupressaceae, Olea, Platanus, Pinus, Ulmus, Populus) and six atmospheric pollutants (O₃, PM₁₀, PM₂.₅, NO₂, CO, SO₂) across urban municipalities in Madrid (2014–2017). Using multiple linear regression, the authors report statistically significant associations (adjusted R² >30%, p<0.0001), identifying Olea pollen and O₃ as the most frequent contributors to asthma exacerbations, with stronger correlations in urban areas. The work highlights the dual role of urban trees as both public health assets and sources of allergenic pollen interacting with pollutants.

​2. Major Limitations​

While the study addresses an important public health issue, several methodological and contextual limitations reduce the robustness and generalizability of its conclusions:

​A. Methodological Concerns​

​Multicollinearity and Variable Selection: 

Pollen and pollutant concentrations often correlate with meteorological factors (e.g., temperature, humidity), which were not included. This omission risks confounding results, as weather influences both pollen dispersal and pollutant distribution.

-Thank you very much for your suggestions, and for contributing to the improvement of this article.

 

-Another study by the same authors, published very recently, deals with the correlation between pollen concentration and various meteorological factors in the city of Madrid. In this article, which is cited with reference [22] in the present study on asthma, the relationship with atmospheric pollutants is also discussed.

No justification is provided for the backward stepwise regression method, which can overfit models.

-All right. Thank you. The explanation on the selection of the regression method used in lines 506-510 and 513-517 has been expanded.

A sensitivity analysis (e.g., variance inflation factors) is needed to confirm the absence of multicollinearity among predictors.

-As this is a descriptive study, and therefore not predictive, it is not considered necessary to carry out a sensitivity analysis. Furthermore, as can be seen in the study carried out by the same authors, and mentioned in this article on asthma, with reference [2], there are significant correlations between pollen and air pollutant concentrations. However, this fact does not affect the robustness of the conclusions of this study, as it is limited to describing, over a given period of time, the interrelationships of the independent variables with asthma episodes of care, and does not focus on making future predictions about these interrelationships.

​Data Limitations:

 Asthma episodes were recorded only on weekdays, ignoring weekend trends. This introduces temporal bias, as pollen/pollutant exposures and healthcare-seeking behaviors may differ on weekends.

-Given that the consultations of the Health Centres are only open on working days, it is not possible to carry out the study on the dates on which data are not available (lines 376-377). Therefore, this is a limitation beyond the researchers' intentions.

Pollen and pollutant data were aggregated at the municipal level, potentially masking micro-scale variations in exposure. Spatial interpolation methods (e.g., kriging) could improve granularity.

-Thank you for your proposal. However, this is not the subject of this study. It is a descriptive study, limited to specific geographical areas. The factors under study are subject to the action of other factors, such as meteorological factors, which imply a variability of their concentrations.

 

​Thresholds for Significance:

 The use of adjusted R² >30% as a benchmark for "medium-high correlation" is arbitrary and inadequately justified. In ecological studies, low R² values (even <30%) can still indicate meaningful associations if effect sizes are consistent.

- In this connection, an explanation of the rationale for the criterion followed with respect to the value of the coefficient of determination has been included (lines 493-498).

​B. Contextual and Analytical Gaps​

​Urban-Rural Dichotomy:

 All municipalities studied are classified as urban (Table 1), yet the conclusion emphasizes rural-urban differences. This misalignment undermines claims about rurality’s role. A reanalysis using continuous metrics (e.g., population density gradients) would strengthen spatial inferences.

-Thank you. The conclusion you refer to has been qualified and corrected (lines 843-844). On the other hand, in the Materials section it is clearly explained that there is an urban gradient in the study municipalities.

​Pollen-Pollutant Interactions:

The mechanistic link between pollen and pollutants (e.g., pollutant-induced allergenicity changes) is discussed but not quantified. Including allergen potency data (e.g., Bet v 1 levels for Olea) would clarify biological relevance.

-This is a very interesting aspect that could be addressed in a future study, where information on the atmospheric levels of pollen allergens, such as Ole e 1, could be gathered.

​Temporal Resolution:

The analysis uses daily data but does not explore lag effects (e.g., delayed asthma responses to pollen exposure). Time-series analyses with distributed lag models could better capture dynamic relationships.

-If it were to do so, it would be a predictive study, and therefore outside the scope of the present study, which is descriptive in nature.

​C. Presentation and Reporting Issues​

​Missing Information:

 Appendix Tables A/B (model equations) are referenced but not provided, hindering reproducibility.

-Thank you. Tables A and B are contained in the Appendix, starting on line 855. The specification of the results in Table A has been included at the bottom of the table.

The exclusion criteria for outliers (studentized residuals >3 SD) are mentioned but lack documentation (e.g., how many data points were removed?).

-In Table A of the appendix you can check the information on the number of eliminated lines, relative to the initial data lines. That is, “No. observations without outliers/No. total observations”.

A reference to this information in Table A has been included in the text of the Methods section (lines 520-522).

​Inconsistent Terminology:

 PM₂.₅ is inconsistently formatted (e.g., “PM2,5” in Tables). Standardize notation to align with environmental science conventions.

-All right. Thank you. It has been corrected.

References to “clinical prevalence” (Introduction) are misleading, as the study uses healthcare visits, not population-level prevalence data.

-As explained in the Materials section, although it has not been possible to make calculations with pure clinical prevalence data, an approximation to these calculations has been made with the only data available in this respect from the Regional Ministry of Health of the Community of Madrid. Therefore, it is possible to speak, in the Introduction, of asthma prevalence, since in this section we are not talking about the calculations made with the characteristic biases of a scientific study.

​Discussion of Conflicting Evidence:

 The manuscript cites studies linking PM₂.₅ and NO₂ to asthma but does not reconcile why these pollutants were less prominent in their models. Acknowledging data gaps (e.g., missing PM₂.₅ measurements at some stations) would improve transparency.

- Thanks. The details of these data gaps are specified in the manuscript (líneas 476-487).

​Recommendations for Revision​

Address multicollinearity and refine regression methods.

Incorporate meteorological variables and lag effects.

Reassess urban-rural comparisons using continuous spatial metrics.

Provide missing appendix tables and outlier removal details.

Clarify limitations related to weekday-only data and PM₂.₅ availability.

This study offers valuable insights into pollen-pollutant interactions but requires methodological strengthening and clearer contextual framing to fully support its conclusions.

 

 

 

Submission Date

24 February 2025

Date of this review

05 Mar 2025 04:31:39

 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The article by Chico-Fernández and Ayuga-Téllez needs to be improved in many aspects. In some cases the text is too long and the results are confusing. More graphics should be presented, so that the results are easier to understand.

The article needs major revisions.

 

The introduction is quite complete. However, the author needs to include many more references. The state of the art is not well developed. I strongly recommend including these articles:
Ortega-Rosas et al., https://doi.org/10.1007/s10653-020-00752-7 
Senechal et al., https://doi.org/10.1155/2015/940243 
de Lira-quezada et al., https://doi.org/10.1016/j.waojou.2023.100867 
Capone et al., https://doi.org/10.3390/atmos14101544

 

Line 51: The acronym CAM must be specified again. Currently it is present in the summary.

 

In line 156, the author specifies that the year 2013 was not taken into consideration for the article. It does not make much sense, in line 154, the author had already specified the years considered in the study.

 

In the map in figure 1, in addition to including the site of the Palinocam network stations, you must insert the position of the contamination monitoring stations. Where are installed? How many stations? Are located near the pollen collection sites?

 

Line 181-186: Please include the web page of the statistical institute where the data was taken.

 

Line 196: What is the OECD? Please explain.

 

In the materials and methods section the author makes no reference to the pollen sampling technique. How were collected? How were analyzed? The same goes for the contaminants.

 

Line 296-307: In addition, the author spends a whole paragraph explaining the cases of asthma in the year 2013. But this year is not included in the study, as mentioned in line 154. In case the author wants to compare the years taken into consideration with other years, he should include a comparison table and discuss the differences in the results section. In the materials and methods, he should only include how the data were collected and the analyses performed for the study and the years considered.

 

In the results section I strongly recommend including graphs. To make this part easier to understand. Only tables are present. Furthermore, some statistical results are presented but in the methodology it is not explained which methods were applied. This needs to be explained in detail. A graphical and statistical comparison between air pollutants and asthma incidences could be included. And also a graph comparing pollen and asthma.

Author Response

Reviewer 2

 

Comments and Suggestions for Authors

The article by Chico-Fernández and Ayuga-Téllez needs to be improved in many aspects. In some cases the text is too long and the results are confusing. More graphics should be presented, so that the results are easier to understand.

The article needs major revisions.

The introduction is quite complete. However, the author needs to include many more references. The state of the art is not well developed. I strongly recommend including these articles:

Ortega-Rosas et al., https://doi.org/10.1007/s10653-020-00752-7 
Senechal et al., https://doi.org/10.1155/2015/940243 
de Lira-quezada et al., https://doi.org/10.1016/j.waojou.2023.100867 
Capone et al., https://doi.org/10.3390/atmos14101544

-Thank you very much for your suggestions, and for contributing to the improvement of this article.

-The references you have suggested have already been included in the Introduction section, plus one more.

Line 51: The acronym CAM must be specified again. Currently it is present in the summary.

-Very well, thank you. It has already been specified in the introduction (line 49).

In line 156, the author specifies that the year 2013 was not taken into consideration for the article. It does not make much sense, in line 154, the author had already specified the years considered in the study.

-All right. The reference to 2013 has been deleted.

In the map in figure 1, in addition to including the site of the Palinocam network stations, you must insert the position of the contamination monitoring stations. Where are installed? How many stations? Are located near the pollen collection sites?

-Thank you. Two additional maps (which becomes Figure 1) have been included in addition to the existing map (which becomes Figure 2), together with an explanation of the number of stations, their location and the proximity between the two types of survey stations (lines 206-226).

Line 181-186: Please include the web page of the statistical institute where the data was taken.

-All right. Thank you. The reference ([26]) has been included (line 313).

Line 196: What is the OECD? Please explain.

 -Thanks. OECD stands for Organisation for Economic Co-operation and Development. It has already been implemented in the article (line 316).

In the materials and methods section the author makes no reference to the pollen sampling technique. How were collected? How were analyzed? The same goes for the contaminants.

-An explanation has been implemented (lines 227-300).

Line 296-307: In addition, the author spends a whole paragraph explaining the cases of asthma in the year 2013. But this year is not included in the study, as mentioned in line 154. In case the author wants to compare the years taken into consideration with other years, he should include a comparison table and discuss the differences in the results section. In the materials and methods, he should only include how the data were collected and the analyses performed for the study and the years considered.

-This paragraph alludes to the study by López Pereira et al., carried out in the period 1996-2013 to show the evolution of the prevalence of asthma in that period, which may be as interesting as the allusions made in the text (lines 440-443) to the different periods studied by the two cited editions of the SEAIC Allergological work (Alergológica 2005 and 2015).

In the results section I strongly recommend including graphs. To make this part easier to understand. Only tables are present. Furthermore, some statistical results are presented but in the methodology it is not explained which methods were applied. This needs to be explained in detail. A graphical and statistical comparison between air pollutants and asthma incidences could be included. And also a graph comparing pollen and asthma.

-Thank you. Two graphs have been included to make it easier to visualise the results obtained.

 

Submission Date

24 February 2025

Date of this review

03 Mar 2025 19:11:52

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The article has been improved according to my comments.

In my opinion, the article is suitable for publication.

Author Response

Thank you very much for your great contribution to the improvement of this manuscript.

Mr Javier Chico Fernández

Author Response File: Author Response.docx