Association between Long-Term Exposure to PM_2.5 and Lung Imaging Phenotype in CODA Cohort

Round 1

Reviewer 1 Report

I carefully examined the author's response that I was comment.

The response is well organized and fully satisfible.

My opinion is accept without further revision.

Author Response

Comment: I carefully examined the author's response that I was comment. The response is well organized and fully satisfible. My opinion is accept without further revision.

Response: Thank you for your detail comment in previous review. Thanks to your comments, our manuscript has become much better.

Author Response File: Author Response.docx

Reviewer 2 Report

The authors have presented a study investigating the association between two CT measures (emphysema index and man wall area percentage) and long term (1 – 5 year) exposure to PM2.5. They found a statistically significant relationship between PM2.5 and EI in participants with normal lung function. It is good that researchers do look for relationships with patho-physiological markers of disease as this can provide more information about the effects of air pollution exposure, particularly as clinical disease may not always manifest. One fundamental problem with the way the authors present the study and the results (in the abstract and discussion) is that they use the term ‘change’ in lung image phenotype. They also mention longitudinal follow-ups or observations. This study did not measure change, nor were there any longitudinal observations. It was a cross-sectional study. There was no follow-up of the participants and no repeat CT scans (from what I read). What the authors found was a significant association between EI, measured at a single time point, and estimates of long term exposure to PM2.5. The authors need to reword the manuscript to reflect that. Even the statistical analysis section states the authors were analysing the ‘Associations between air pollution and annual change of image measurements….’, which is not correct as there were no annual measurements. As a cross-sectional study there is no way to determine if exposure contributed to the differences. Therefore, in the Abstract the authors cannot state that the results suggest that long-term exposure to PM2.5 can cause emphysema in subjects with normal lung function. They can only state that there was a statistical association between the two that requires further investigation. I wasnt clear about the exposure assessment. In most LUR estimates, annual PM is estimated based on the participant’s home address, which seems case in this study (Methods, p3). It is not clear why PM2.5 concentrations were estimated using 1x1, 3x3, and 5x5 grids. Why use three these three? Why not just use the grid with the finest spatial resolution (1 x 1), or the one that fits the calibration model best? Were there any associations between long-term PM2.5 and lung function? Were there any associations between lung function and CT measures?

Author Response

Comment 1: The authors have presented a study investigating the association between two CT measures (emphysema index and man wall area percentage) and long term (1 – 5 year) exposure to PM2.5. They found a statistically significant relationship between PM2.5 and EI in participants with normal lung function. It is good that researchers do look for relationships with patho-physiological markers of disease as this can provide more information about the effects of air pollution exposure, particularly as clinical disease may not always manifest.

 

Response: Thank you for your valuable comments. Regarding the concerns raised by you, we provided a point-by-point response as below.

 

Comment 2: One fundamental problem with the way the authors present the study and the results (in the abstract and discussion) is that they use the term ‘change’ in lung image phenotype. They also mention longitudinal follow-ups or observations. This study did not measure change, nor were there any longitudinal observations. It was a cross-sectional study. There was no follow-up of the participants and no repeat CT scans (from what I read). What the authors found was a significant association between EI, measured at a single time point, and estimates of long term exposure to PM2.5. The authors need to reword the manuscript to reflect that. Even the statistical analysis section states the authors were analysing the ‘Associations between air pollution and annual change of image measurements….’, which is not correct as there were no annual measurements. As a cross-sectional study there is no way to determine if exposure contributed to the differences. Therefore, in the Abstract the authors cannot state that the results suggest that long-term exposure to PM2.5 can cause emphysema in subjects with normal lung function. They can only state that there was a statistical association between the two that requires further investigation. I wasnt clear about the exposure assessment.

 

Response: We fully agree with your concerns. As you commented, we have corrected several midleading expressions in the title, abstract, introduction, method, results and discussion. The modified parts are marked in red.  

 

Title: Association between long-term exposure to PM_2.5 and lung imaging phenotype IN CODA COHORT

Abstract line 58: emphysema index measurement

Abstract line 60: mean wall area measurement

Introduction line 102-104: the objective of this study was to examine the association between long-term exposure to air pollution (PM_2.5) and lung imaging phenotype in dust exposed Korean adults living near cement factories.

Method line 167-169: Associations between air pollution and image measurements (emphysema index and mean wall area percentage) were examined using a multiple linear regression model after adjusting for potential confounders.   

Result line 197: in longitudinal follow-ups -> even for measurement in different conditions

Result line 204: in various conditions

Discussion line 278-283: we investigated the lung imaging phenotype following prolonged PM_2.5 exposure. Our study had some remarkable findings. First, the specific findings in lung imaging phenotype were observed after long-term exposure to ambient PM_2.5. Second, there was no significant association in emphysema index or mean wall area in COPD patients who already had clinical manifestations. However, significant difference in emphysema index was confirmed in the normal lung function group following the long-term exposure of PM_2.5.

Conclusion line 332-334: phenotypic imaging finding presented as emphysema index was found in the normal lung function group, not in the COPD group.

 

Comment 3: In most LUR estimates, annual PM is estimated based on the participant’s home address, which seems case in this study (Methods, p3). It is not clear why PM2.5 concentrations were estimated using 1x1, 3x3, and 5x5 grids. Why use three these three? Why not just use the grid with the finest spatial resolution (1 x 1), or the one that fits the calibration model best?

 

Response: Thank you for your helpful comment. In order to see if the three grids (1x1, 3x3, and 5x5) show the same trend, all three cases were measured and presented.

 

Comment 4: Were there any associations between long-term PM2.5 and lung function? Were there any associations between lung function and CT measures? 

 

Response: Thank you for your valuable comments. As you pointed out, we think the subject is also important, and as a matter of fact, we are doing another study on that topic. Thank you very much.

 

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

The authors have adequately addressed most of my comments. There are a few issues, however, that could be amended and/or explained further. These are;

1. The conclusion of the Abstract needs to be reworded. As a cross-sectional study, the results do not suggest causality.
2. The reason for the 3 exposure models (1x1, 3x3, 5x5 grids) needs to be provided. It is not clear why three exposure models were chosen or how these improve exposure assessment. Why was it necessary to calculate exposure using different spatial resolution? (This needs to be included in the Methods section)
3. Change the term 'insignificant' (line 162) to 'not or non significant'

Author Response

Responses to the Reviewer 2’s Comments

 

Comment: The authors have adequately addressed most of my comments. There are a few issues, however, that could be amended and/or explained further.

 

Response: Thank you for your valuable comments. Regarding the concerns raised by you, we provided a point-by-point response as below.

 

Comment 1: The conclusion of the Abstract needs to be reworded. As a cross-sectional study, the results do not suggest causality.

 

Response: We fully agree with your concerns. As you commented, we have corrected the ambiguous expression in the conclusion of the Abstract. The modified parts are marked in red (page 4, line 62-63)

 

Conclusions: This study suggests that long-term exposure of PM_2.5 may affect the emphysematous change in patients with normal lung function.

 

Comment 2: The reason for the 3 exposure models (1x1, 3x3, 5x5 grids) needs to be provided. It is not clear why three exposure models were chosen or how these improve exposure assessment. Why was it necessary to calculate exposure using different spatial resolution? (This needs to be included in the Methods section)

 

Response: Thank you for your helpful comment. Looking at other papers that have studied the ambient air pollution such as PM_2.5 or VOCs with land-use regression, it can be seen that researchers have divided tha space to study on their own to confirm the on the exposure of air pollution^1-3. Such grids and would vary from study to study. In our study, we set the three grids (1x1, 3x3, and 5x5) to show the same trend about long-term exposure of PM_2.5. We added this description into the Methods section (page 7, line 146-147).

 

These grids (1x1km, 3x3km, and 5x5km) were determined to show the same trend about the long-term exposure of PM_2.5.

 

Comment 3: Change the term 'insignificant' (line 162) to 'not or non significant'

 

Response: Thank you for your valuable comments. As you pointed out, we have corrected that.  Thank you very much.

 

 

1. Liu T, Xiao J, Zeng W, Hu J, Liu X, Dong M, et al. A spatiotemporal land-use-regression model to assess individual level long-term exposure to ambient fine particulate matters. MethodsX 2019;6:2101-5.
2. Lu TJ, Lansing J, Zhang WW, Bechle MJ, Hankey S. Land Use Regression models for 60 volatile organic compounds: Comparing Google Point of Interest (POI) and city permit data. Science of the Total Environment 2019;677:131-41.
3. Chang TY, Liang CH, Wu CF, Chang LT. Application of land-use regression models to estimate sound pressure levels and frequency components of road traffic noise in Taichung, Taiwan. Environment International 2019;131.

Author Response File: Author Response.docx

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

This is an interesting paper on the association between estimated PM2.5 and higher emphysema index in a Korean population aged around 72 years. A number of methodological aspects are described with insufficient detail, making it difficult to interpret the results.

 

1) cohort inclusion criteria: line 74-75 state that the cohort included participants who had airflow limitation and healthy volunteers. No further detail on cohort inclusion criteria is provided. What were the inclusion criteria for the cohort? Was it aimed at including a large number of participants with airflow limitation? How was that determined (prior to study participation)? How were participants with airflow limitation recruited? How were ‘healthy’ participants recruited? Based on what criteria?

 

2) baseline measurements: it would be good to clearly state in the methods that baseline measurements (spirometry and lung imaging) took place over 2012-2017. Is it correct to say that the baseline measurements were taken at enrolment? This is important information in relation to how baseline data fit in relation to the time-period for which PM2.5 data are available. It would be helpful if descriptive data on year of baseline measurements were included in Table 1.

 

3) PM2.5 using satellite data: PM2.5 was estimated based on satellite-measured optical depth with 3 different resolutions (1 m2, 3 m2, 5 m2). Annual average PM2.5 were calculated for the years 2012-2017. The administrative area was 40-80 km2. No information on the estimated PM2.5 levels are provided. Could some data on the annual average PM2.5 estimated for this population be provided? This is important for the interpretation of results. Was there large variability in annual average PM2.5 among the study population? How was this geographically distributed? Was the presence of a cement factory associated with high PM2.5? Or is it chiefly associated with urban areas? Are there changes in mean PM2.5 over time (years 2012-2017)?

 

4) line 104-106: “We… observed their long-term exposure to annual average PM2.5 concentrations after 1 year, 3 year, and after 5 years before baseline examination.” Stating that these data are available “before baseline examination” does not seem to be correct. If a participant was recruited in 2012 and had their baseline assessments in 2012, the available PM2.5 data before baseline is for only one year. Only for the participants who had their baseline assessment in 2017 would a 5-year period of PM2.5 data be available. A clearer description of how the 1-year, 3-year and 5-year averages were calculated in relation to the year of baseline assessment is needed.

 

5) The presence of cement factories in the study area is presented as a relevant source of PM2.5 exposure in this population, and the study population seems to have been specifically chosen based on their proximity to cement factories. This raises the question of significant occupational exposure to dust at the cement factories for those who worked there. It is therefore surprising that the analyses were not adjusted for previous employment at a cement factory. Is information on previous employment at a cement factory available?

 

6) line 124 states that 76% of the population were ever/current smokers. Should this be 66%? Table 1: should ‘ever smoker’ be ‘ex smoker’? Normally ever smokers are considered those who ever smoked, including the current smokers.

 

7) line 161-162: “Changes of imaging phenotype were observed after long-term exposure to ambient PM2.5”. I think this is misleading: it suggests that actual changes in imaging phenotype were observed over time, suggesting that the study participants had undergone multiple medical assessments over time. From the methods, this does not seem to be the case: participants were only assessed at baseline (unless I totally misunderstood the methods). More careful wording should be used to accurately reflect the study findings.

Reviewer 2 Report

The manuscript by Kim et al. is interesting in that it shows PM 2.5 levels are associated with changes in lung image phenotype, especially emphysema in normal lung function subjects in korea cohort. This data seems interesting and provide some valuable information regarding the effect of PM 2.5 in lung health. However, the studies have limitations which need to more explanation. These concerns are outlined below. 

Comments

1. The data is from subjects living in six cities, which is cement plants is located. This may affect different result from other previous reports. In addition, is there different levels in PM 2.5 between cities? The authors should describe these in the discussion.
2. Regarding data, the authors just show association between emphysema and PM.2.5 levels. To more explain the characteristics of study subjects, the association between emphysema percent and personal factors ( 6 city area, age, sex, smoking) should be addressed as a new figure or table.
3. The demographic data of normal subjects (in Table 3), especially smoking, FVC, FEV1 and FEV1/FVC should be addressed in the supplement.