Association of Cardiovascular Disease and Long-Term Exposure to Fine Particulate Matter (PM_2.5) in the Southeastern United States

Round 1

Reviewer 1 Report

The manuscript I was sent to review has no line numbers which makes it difficult to refer precisely to locations within this manuscript.

This paper describes work that characterises the extent of association between self-reported measures of cardiovascular disease (CVD) and their risk factors in a large cohort, and the estimated exposure of the cohort individuals to annual mean PM2.5 in the year preceding their enrolment in the cohort.

The cohort is a very large one spanning 12 states in the south-eastern USA. This study uses data for 44,610 individuals. Appropriate multi-level models are applied.

The manuscript is well written and clearly presented. The findings are of interest. However, there are some shortcomings in the discussion that should be addressed.

(1) The abstract and final part of the discussion need to make clear that the ‘headline’ conclusions of this study are not necessarily reflective of the general population of this part of the USA since the cohort of individuals that underpins the data analysis in this work is substantially biased towards members of households with low incomes, low employment rates and high prevalences of self-reported morbidities.

(2) A critical data component in this study is the estimation of the PM2.5 exposure, yet whilst the paper contains a lot of information about the population enrolment and the health data, there is almost no information on the PM2.5 exposure, just one or two references. Although there does not need to be largescale repetition of data published elsewhere, the readers of this paper need to have some idea about estimated uncertainties in PM2.5 exposure assignment; both the uncertainties inherent in the model process used to derive a PM2.5 value, but also uncertainty in the use of 3 km grid averaging to assign an exposure to a given individual. What proportion of the individuals have a residential address in the same grid as another individual? (or perhaps where this occurs the date of enrolment was different such that a different PM2.5 value is assigned?)

(3) The Discussion needs to provide some discussion of the geographical settings and the population demographics contained therein for the cohort individuals that have exposures less than and greater than the binary threshold of 12 ug/m3. Presumably the former are dominated by rural communities and the latter by urban communities? This must be particularly the case given that the stated spatial resolution of the exposure assignment is as coarse as 3 km grids. What about differences in other confounder variables between these two communities that might partially or totally account for the CVD status, e.g. diet? Associated with this issue, the nature of the PM2.5 and its closely correlated air pollutants may be different between the sub-populations in these two different types of communities. For example, the chemical constituents of the PM2.5 may be different, and there may be greater annual mean ultrafine particle number concentration and/or NO2 in the urban populations, and/or greater annual mean ozone in the rural populations.

I also noted the following minor points.

Abstract L5: Start new sentence at “currently”

Abstract L5: The units given here for PM2.5 concentration are not correct.

Abstract L14: The terminology “unhealthful” is used here by the authors, which I find a poor choice of word compared with using “unhealthy”. However, the authors use this terminology throughout their manuscript so I leave it to the journal typesetters to accept this terminology or to make a change.

P1, l1: Insert “US” to read “the US Clean Air Act”

P1, first paragraph: PM pollution is not the largest CV health risk, but it is the largest environmental factor for CV health risk. Please ensure that the text is precise.

P1, first paragraph: why are transport and industrial combustion and industrial processes not listed as major sources of primary PM2.5 into the atmosphere? They should be: they are sources of primary BC, OC and metal-containing particles.

P1 penultimate paragraph: Why is access “only to ground level estimates of PM2.5 exposures” deemed to be a limitation of a study on association between PM2.5 and morbidity? Surely ground-level estimates of PM2.5 is exactly what is required, assuming the estimates are accurate of course.

P6, first paragraph of Results: Remove repetition of the statistic that only 36.3% of the cohort were employed.

P9, first paragraph. Put the phrase “of which PM2.5 is one” into a sub-clause bounded by commas.

Author Response

Please see attachment for responses to review comments and suggestions.

Author Response File: Author Response.docx

Reviewer 2 Report

Please check the attached document.

Comments for author File: Comments.pdf

Author Response

Please see the attachment for our responses to reviewers comments and suggestions.

Author Response File: Author Response.docx