Environmental predictors of pulmonary nontuberculous mycobacteria (NTM) sputum positivity among persons with cystic fibrosis in the state of Florida.

Nontuberculous mycobacteria (NTM) are opportunistic human pathogens that are commonly found in soil and water, and exposure to these organisms may cause pulmonary nontuberculous mycobacterial disease. Persons with cystic fibrosis (CF) are at high risk for developing pulmonary NTM infections, and studies have shown that prolonged exposure to certain environments can increase the risk of pulmonary NTM. It is therefore important to determine the risk associated with different geographic areas. Using annualized registry data obtained from the Cystic Fibrosis Foundation Patient Registry for 2010 through 2017, we conducted a geospatial analysis of NTM infections among persons with CF in Florida. A Bernoulli model in SaTScan was used to identify clustering of ZIP codes with higher than expected numbers of NTM culture positive individuals. Generalized linear mixed models with a binomial distribution were used to test the association of environmental variables and NTM culture positivity. We identified a significant cluster of M. abscessus and predictors of NTM sputum positivity, including annual precipitation and soil mineral levels.

Introduction

Nontuberculous mycobacteria (NTM) are opportunistic human pathogens that reside in the environment and are commonly found in soil and water [1]. Exposure to these organisms may cause pulmonary nontuberculous mycobacterial (PNTM) disease, which poses a threat to high-risk groups including older adults and individuals with chronic lung conditions; persons with CF are particularly vulnerable [2]. Infection likely results from a combination of behavioral and environmental exposure which jointly increase the risk of NTM infection [3]. Studies have tried to evaluate common exposure sources including household plumbing such as showerheads [4,5], water heating units, and dust, as well as external sources such as soil, watersheds, and climatic factors [6-8]. These factors play a role in determining environmental suitability for the pathogen, with higher abundance of NTM associated with increased rainfall, humidity, certain watersheds, and soil composed of particular elements. Studies have also shown that prolonged exposure to certain environments can increase the risk of PNTM [9]. It is therefore important to determine the risk associated with different geographic areas, so high risk individuals can either avoid an area or focus prevention efforts to reduce their risk of exposure. Previous studies have shown significant clustering of NTM in multiple regions of the United States, including Florida, with geographical heterogeneity in overall and species-specific risk of pulmonary disease. In Florida, the 5-year NTM sputum positivity prevalence was 31% from 2010 through 2014 [2], and this state has the highest prevalence of NTM in the contiguous US [10,11]. Identifying risk factors for exposure and subsequent infection with NTM are central to prevention efforts in the CF community [12]. In this study we describe spatial clusters of NTM infections in Florida and identify environmental predictors of NTM sputum positivity.

Materials and methods

Data sources

We conducted a nested case-control study, using annualized registry data obtained from the Cystic Fibrosis Foundation Patient Registry (CFFPR) for 2010 through 2017 [13]. Our study population comprised patients aged ≥ 12 years residing ≥ 2 consecutive years in Florida. Patients with a history of lung transplant or Mycobacterium tuberculosis infection were excluded. Incident NTM cases were persons with a positive pulmonary culture after ≥ 1 negative culture(s) and with residence in Florida the year before and the year of their first positive NTM culture. Controls were defined as persons with ≥1 negative NTM culture(s) during the study period, residence in Florida the year of and the year before a negative culture, and no positive cultures during Florida residency. Because the majority of individuals had multiple cultures, we used the first culture and residential ZIP code associated with the year of first culture for each person meeting these criteria for analysis.

We selected environmental variables for analysis based on prior findings. Variables that have been previously found to be predictive of sputum positivity include evapotranspiration [10], saturated vapor pressure [14], vapor pressure [15], temperature [6], and rainfall [6], as well as soil or water mineral concentration including copper [10], sodium [10], manganese, [8,10], calcium [7], and molybdenum [7]. Environmental data sources used in this study are described in Table 1. Soil geochemistry collected from 2007 through 2010 included data on calcium, copper, molybdenum, manganese, and sodium content from samples measured in the top 5 cm of soil. Annual temperature and rainfall, and eight-day evapotranspiration data were extracted for the years 2010 through 2017. Evapotranspiration was averaged to create annual estimates. Ordinary kriging was performed to estimate the broader spatial distribution over Florida from the original sampling sites or weather stations for all soil geochemistry variables, temperature, and rainfall. To adapt county-level census data to the ZIP-code level, we determined the county each ZIP code primarily fell within by overlaying ZIP code polygons on a map of Florida counties using the R packages “rgdal” and “sp” [16-18]. The ZIP code-level mean of each environmental variable was calculated, scaled, and mean-centered by year, when data were available, for analysis. Driving distance from ZIP code centroids to CF clinics in Florida were calculated using a Distance Matrix API via Google Cloud Services and the R package “gmapsdistance” [19] to control for potential spatial clustering near CF clinics. The closest clinic for individuals was assumed to be the pediatric or adult clinic with the shortest driving distance for all patients in each ZIP code under or at least 18 years of age, respectively.

Table 1

Data sources for variables used in analysis.

Variable	Source	Type and frequency	Period	Citation
Soil minerals	United States Geological Services’ (USGS) Geochemical and Mineralogical Data for Soils of the Conterminous United States study	Survey sites, collected once	2007–2010	https://pubs.usgs.gov/ds/801/
Temperature	National Oceanic and Atmospheric Administration (NOAA) Global Summary of the Year (GSOY) database	Weather station, annual	2010–2017	https://data.nodc.noaa.gov/cgi-bin/iso?id=gov.noaa.ncdc:C00947
Rainfall	National Oceanic and Atmospheric Administration (NOAA) Global Summary of the Year (GSOY) database	Weather station, annual	2010–2017	https://data.nodc.noaa.gov/cgi-bin/iso?id=gov.noaa.ncdc:C00947
Evapotranspiration	MOD16A2 Version 6 Evapotranspiration/Latent Heat Flux dataset by the National Aeronautics and Space Administration (NASA) in collaboration with the USGS	Remotely sensed, 8-day	2010–2017	https://lpdaac.usgs.gov/products/mod16a2v006/
CFF centers	Cystic Fibrosis Foundation website	Not applicable		https://www.cff.org/ccd/CareCenters?State=FL&Zip=&Distance=100
Median income	United States Census Bureau American Community Survey (ACS) data	Survey, annual	2010–2017	https://www.census.gov/programs-surveys/acs/data.html
Metropolitan centers	Centers for Disease Control and Prevention (CDC) 2013 National Center for Health Statistics (NCHS) Urban-Rural Classification Scheme for Counties	Census-based data, collected once	2013	https://www.cdc.gov/nchs/data_access/urban_rural.htm

Spatial and statistical analysis

We used a Bernoulli model comparing individuals with CF who were NTM culture positive (cases) to individuals with CF who were NTM culture negative (controls) to identify clustering of ZIP codes with higher than expected numbers of NTM culture positive individuals from 2011 through 2017, using SaTScan version 9.6 [20] with default setting, limiting cluster radius to 50km. This was done as a way to assess the presence of geographic locations in Florida where people with CF might be at increased risk of NTM. We repeated the analysis for overall NTM as well as each species of NTM separately. To test the association of environmental variables and NTM culture positivity, we used generalized linear mixed models with a binomial distribution. The variance inflation factor was used to assess collinearity and model fit was assessed via Akaike Information Criterion. The final environmental variables included were copper, manganese, molybdenum, sodium, annual precipitation, and evapotranspiration from the year in which an individual had their first NTM culture. Additionally, we included patient gender, age, number of years receiving chronic macrolides through the year of culture, the closest CF clinic in miles, whether the ZIP code was considered metropolitan, ZIP code median household income, and a random intercept for ZIP code. Statistical analyses were conducted using R versions 3.6.1–4.0.2 [21]. The study was determined to be not Human Subjects Research by the NIH Office of Human Subjects Research Protection.

Results

Of the 1293 patients in the CFFPR residing ≥ 2 consecutive years in Florida from 2010 through 2017, 979 patients met inclusion criteria; 261 (26.7%) were classified as cases and 718 (73.3%) as controls (Fig 1). Species identified were distributed as follows: 109 (41.8%) M. avium complex (MAC), 118 (45.2%) M. abscessus and its subspecies, and 62 (23.8%) other species. The proportions of sex, age, and years of chronic macrolide use were similar between cases and controls, while years of Florida residency from 2010 to the year of the first culture were generally greater for cases (Table 2). We found one statistically significant high-risk cluster in southeast Florida (p-value: 0.035, radius: 48.8 km), which included Broward, Miami-Dade, and Palm Beach counties (Fig 2). This high risk-cluster was associated with M. abscessus; no significant clustering was observed for other species.

Fig 1

Study population for incident NTM cases and negative controls in the CFFPR from 2010 through 2017.

Table 2

Baseline demographics of the study population by NTM culture result, Florida, 2011–2017.

	NTM cases* (n = 261)
	All NTM (n = 261)	MAC* (n = 109)	Mabs* (n = 118)	Controls (n = 718)
Sex
Female	119 (45.6)	51 (46.8)	52 (44.1)	361 (50.3)
Male	142 (54.4)	58 (53.2)	66 (55.9)	357 (49.7)
Age group, yr†
12 to <18	80 (30.7)	38 (34.9)	44 (37.3)	251 (35.0)
18 to <60	178 (68.2)	69 (63.3)	73 (61.9)	462 (64.3)
≥ 60	3 (1.1)	2 (1.8)	1 (0.8)	5 (0.7)
Chronic macrolide use‡
0 yr	68 (26.1)	25 (22.9)	40 (33.9)	195 (27.2)
1–2 yr	88 (33.7)	43 (39.4)	49 (41.5)	332 (46.2)
3–4 yr	66 (25.3)	24 (22.0)	21 (17.8)	122 (17.0)
5+ yr	39 (14.9)	17 (15.6)	8 (6.8)	69 (9.6)
Florida residency‡
2–3 yr	129 (49.4)	58 (53.2)	69 (58.5)	620 (86.4)
4–5 yr	80 (30.7)	27 (24.8)	31 (26.3)	71 (9.9)
6+ yr	52 (19.9)	24 (22.0)	18 (15.3)	27 (3.8)

*NTM = nontuberculous mycobacteria; MAC = Mycobacterium avium complex; Mabs = M. abscessus.

†Age groups are determined by year of NTM culture.

‡From 2010 to year of NTM culture.

Fig 2

Distributions of persons residing in Florida for ≥ 2 years from 2011 through 2017.

A. Persons with cystic fibrosis (CF), n = 979. B. Nontuberculous mycobacteria (NTM) cases among persons with CF, n = 261. C. ZIP codes within a high-risk cluster of NTM cases among persons with CF.

Sputum-positive patients were more likely to live in a ZIP code with higher average yearly precipitation, with a 34% increase in the odds of a positive culture for each standard deviation (SD) increase in average annual precipitation (adjusted odds ratio [aOR]: 1.34, 95% confidence interval [95% CI]: 1.13–1.58). Soil geochemistry was also associated with NTM positivity; a one SD increase in levels of sodium in the soil was associated with a 92% increased risk of culture positivity (aOR: 1.92, 95% CI: 1.46–2.52), and a one SD increase in soil manganese was associated with a 40.7% decreased risk (aOR: 0.59, 95% CI, 0.46–0.77). Species-specific analysis showed the same associations for M. abscessus (Table 3).

Table 3

Generalized linear mixed models with a binomial distribution and a random intercept for zip code of demographic and environmental risk factors for NTM culture positivity in persons with cystic fibrosis, Florida, 2011–2017.

			All NTM*	MAC*	Mabs*
	Fixed effects	SD*	aOR (95% CI)*	aOR (95% CI)	aOR (95% CI)
Individual level	Female	-	0.81 (0.60, 1.09)	0.91 (0.59, 1.41)	0.75 (0.50, 1.12)
	Age, per year	-	1.01 (1.00, 1.03)	1.00 (0.98, 1.02)	1.01 (0.99, 1.03)
	Years on chronic macrolide	-	1.10 (1.02, 1.19)	1.10 (0.98, 1.23)	0.88 (0.79, 0.99)
Zip code level	Mean precipitation†	7.4 in	1.34 (1.13, 1.58)	1.18 (0.92, 1.51)	1.25 (1.00, 1.57)
	Mean yearly evapotranspiration†	895 mm	0.85 (0.70, 1.04)	0.75 (0.54, 1.03)	1.03 (0.78, 1.36)
	Mean soil manganese†	16.8 ppm	0.59 (0.46, 0.77)	0.82 (0.58, 1.17)	0.53 (0.34, 0.82)
	Mean soil sodium†	46.2 ppm	1.92 (1.46, 2.52)	1.44 (0.98, 2.13)	2.27 (1.45, 3.54)
	Closest CF clinic*†	32.5 miles	0.90 (0.76, 1.07)	0.76 (0.57, 1.02)	0.97 (0.76, 1.25)
	Median household income, per $1000†	$17.10	0.98 (0.84, 1.15)	1.01 (0.80, 1.28)	1.17 (0.96, 1.42)
	In a metropolitan area	-	0.76 (0.37, 1.56)	0.72 (0.25, 2.07)	0.68 (0.24, 1.93)

*NTM = nontuberculous mycobacteria; MAC = Mycobacterium avium complex; Mabs = M. abscessus; CF = cystic fibrosis; SD = standard deviation; aOR (95% CI) = adjusted odds ratio (95% confidence interval).

†Variables were scaled and centered to zero; interpretation is by standard deviation.

Discussion

Because the prevalence of PNTM in Florida is so high [10,11], understanding whether there are environmental predictors or geospatial clustering of NTM is of interest to the CF community and public health. We found a significant cluster of NTM culture positivity, specifically M. abscessus, among persons with CF living in Florida in the southeast part of the state. M. abscessus is one of the most commonly isolated species of NTM from persons with CF, found in up to 16–68% of NTM-positive sputum cultures [22,23] and is considered difficult to treat due to high levels of inducible resistance to macrolides, the typical first-line therapy for infections with other NTM species [24]. It is therefore important to understand geographical areas of higher risk to acquiring this pathogen.

In addition to spatial clustering, we also found an association between sputum positivity and annual precipitation, soil sodium levels, and levels of soil manganese. The risk associated with higher sodium and lower manganese levels in soil is consistent with two other studies in the US, however these studies were not able to examine species-specific associations [8,10]. An association with rainfall has been recently identified for the province of Queensland, Australia; this relationship varied by species and geographic region [6]. One of the challenges of studying the environmental risks for PNTM disease is that the incubation period for PNTM is unknown, which creates uncertainty about the appropriate timescale for measuring exposure. While we studied incident infections to limit this potential bias, it will be important to conduct further analyses varying the timescale of possible exposure to environmental variables. Additionally, longitudinal, granular data on soil geochemistry is currently unavailable; the soil mineral concentrations throughout Florida likely vary more than represented in our study, limiting our ability to adjust these variables based on time and location. These are particularly important considerations as studies have shown that prolonged exposure to certain environments increases the risk of PNTM [9], so cumulative exposures to a variety of high risk sources may increase the risk of infection.

Climatic and environmental factors that contribute to increased mycobacterial abundance likely vary by region, making identification of a uniform set of determinants contributing to PNTM disease across national and global geographic areas challenging. Factors related to the built environment also likely interact with soil and water sources to affect the presence of mycobacteria in the environment; a recent study quantifying mycobacteria in showerheads found that both type of chlorination and showerhead material influenced their abundance [5]. Future studies which estimate risk related to mycobacterial abundance as well as components of the natural and built environmental will allow more complete and precise elucidation of these factors. Because persons with CF are at increased risk for NTM infection, continued studies to determine high-risk geographical areas and specific predictors of disease are critical so that precautions can be taken to reduce risk of exposure.

20 Apr 2021

PONE-D-21-06131

Environmental predictors of pulmonary nontuberculous mycobacteria (NTM) sputum positivity among persons with cystic fibrosis in the state of Florida

PLOS ONE

Dear Dr. Ricotta,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Both reviewers found your study to be a potentially important contribution to the epidemiology of NTM infections in CF patients based on geographic location. If you choose to submit a revised manuscript, Reivew 2 has a number of comments all of which need to be addressed in your response, along with revision of the manuiscript.

Please submit your revised manuscript by May 17, 2021. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Thomas Byrd

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

and

2. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

3. Your ethics statement should only appear in the Methods section of your manuscript. If your ethics statement is written in any section besides the Methods, please move it to the Methods section and delete it from any other section. Please ensure that your ethics statement is included in your manuscript, as the ethics statement entered into the online submission form will not be published alongside your manuscript.

Additional Editor Comments:

Both reviewers found your manuscript interesting and a potentially important contribution to the role that geographic location plays in the incidence of NTM lung infections in CF patients. Reivewer 2 raises a number of questions that need to be addressed. If you choose to submit a revised version of your manuscript, please respond to the issues raised by Reviewer 2 and modify the manuscript accordingly.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: No

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: An important contribution toward understanding the environmental characteristics that are determinants of environmental mycobacterial geographical distribution.

I read the article closely, but could find no instances where I thought changes or additions/deletions were needed.

Reviewer #2: Review: Foote et al. Environmental predictors of pulmonary nontuberculous mycobacteria (NTM) sputum positivity among persons with cystic fibrosis in the state of Florida

Manuscript (MS) in review. Interesting and much needed analysis of environmental risk factors for NTM sputum culture positivity. The use of CF patients is a novel and useful approach to identifying those with substantial exposure to NTM. The use of cluster analysis with environmental analysis at the zip code level is a specific and granular approach to identifying potentially important exposures.

General comment. This reader had to reread the manuscript multiple times to understand the role of the SE cluster of NTM cases. Were they modelled separately? It does not appear so, the model encompasses the state. I have come to the conclusion that you called them out because you identify them as a cluster and that is all. This is confusing. Maybe there should be less emphasis on this discovery (no inclusion in Figure 2) or somehow otherwise alert the reader that this was an incidental finding. Did you run a cluster analysis on controls? Maybe looking at case and control clusters is another environmental analysis report (suggestion). But in this report, the cluster is a distraction as currently presented.

Methods

Q- How was zipcode assigned to patient and control? Did you use residential or hospital zipcode or something else? Was the same method used for both patient and control?

Q- You analyze sputum positive CF patients (cases) and compare to controls: What is the rate of false negatives among sputum samples for NTM? Did you evaluate later cultures of controls to see if they became cases later during your six year study period?

Q-Line 64- you analyzed weather, soil or water mineral concentrations. I did not see the resulting values for these environmental exposures (maybe in a supplement)? Nor do I see a data source for water minerals in Table 1, although soil is listed.

Q-How were zip code and census polygons (income, metropolitan centers) reconciled?

The reader needs more information about the temporal and spatial variability of covariates to be comfortable with the presence of environmental exposures included in the final model. Because variables such as temperature, precipitation, and evapotranspiration (“weather”) vary greatly month to month, year to year. Sea surface temperatures influence air temperature and precipitation both influence evapotranspiration. Early in your study period, there were anomalous lows, then anomalous highs, then lows again, ending in anomalous highs in 2017 in sea surface temperatures. See: https://twitter.com/MichaelRLowry/status/1249061412896464896/photo/1

(Questions to authors in bold text) This reader is concerned that there is a lack of information about this variability and how “weather” data are related to NTM culture positivity. What is the typical lag between NTM colonization and detection vs sputum sampling? How often are CF patients cultured? Please cite sources. This reader is seeing that analyses were conducted over a six year period (2011-2017). All subjects appear to be lumped into that time period. Mean “weather” variables are defined by year. What is the distribution of NTM culture positivity by month? Please include table. How is weather associated with sputum positivity over time, given potential lags? Was there enough power to perform a temporal analysis between weather and sputum positivity?

In short, there is a major mismatch between exposure and outcome. If you wish to include exposure variables that can change so much year to year, as I reader I expect to see time in the model. It is Florida. What about outlier precipitation years due to hurricanes? Effect on sputum positivity? I also expect to see the outcome analyzed by time. But NTM is considered a chronic condition. How do you reconcile this analysis of highly temporally variable “weather” with an outcome (sputum positivity) with no granularity or knowledge of time?

Soil is the also problematic. In the absence of additional information in the MS, these are the issues.

1-The reference for soil mineral data: https://pubs.usgs.gov/ds/801/ leads one to a summary report: “U.S. Department of the Interior, U.S. Geological Survey, Data Series 801, Geochemical and Mineralogical Data for Soils of the Conterminous United States” (“Data for Soils”). The report describes sampling methods which include a single point sample collection from “target sites that represented a density of approximately 1 site per 1,600 km2”.

Spatial variability of soil minerals is a large concern for this report. Lines 70-72 state ” Where necessary, ordinary kriging was performed to estimate the broader spatial distribution over Florida from the original sampling sites or weather stations. “ This statement comes after “weather” variables were discussed. Is this kriging approach also true for point estimates of minerals in soil by zipcode?

Please include citations for why a soil mineral analysis of this type -using kriging among spatial points 1,600 km2 apart, sampled once near the time of the study period, represents the exposure of people living among the points during the study period.

When one looks at variability in soil characteristics, this reader was surprised by the extreme temporal and spatial variability in concentrations of minerals. Let’s examine the two that made it to the final model, sodium and manganese.

Sodium: the exposure analysis using grid spacing for sodium samples of 1 site per 1,600 km2 appears to be too coarse to estimate your subjects exposures at the zipcode level. See: Trangmar et al. Application of Geostatistics to Spatial Studies of Soil Properties. Advances in Agronomy 1986. https://doi.org/10.1016/S0065-2113(08)60673-2

Manganese: in a study of agricultural fields, manganese concentrations differed among months of the year and over the extent of the field. Areas with higher concentrations did not keep rank order of concentrations over the year. See: Hoskinson, et al. Temporal Changes in the Spatial Variability of Soil Nutrients. INEEL/CON-99-00290. 1999.

https://www.osti.gov/servlets/purl/9823-CNgTWm/webviewable/

It appears that soil minerals are variable spatially and temporally due to multiple factors such as land use, vegetative cover, soil microbial communities, and precipitation.

Therefore, I fear that your assignment of soil mineral exposure was too simplistic. Please prove me wrong, because we need to better understand environmental risk factors for NTM colonization.

Note- attachment preserves figure and bold text features of review.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

20 Jul 2021

PONE-D-21-06131 - Environmental predictors of pulmonary nontuberculous mycobacteria (NTM) sputum positivity among persons with cystic fibrosis in the state of Florida

Response to reviewers

Reviewer #2: Review: Foote et al. Environmental predictors of pulmonary nontuberculous mycobacteria (NTM) sputum positivity among persons with cystic fibrosis in the state of Florida

Manuscript (MS) in review. Interesting and much needed analysis of environmental risk factors for NTM sputum culture positivity. The use of CF patients is a novel and useful approach to identifying those with substantial exposure to NTM. The use of cluster analysis with environmental analysis at the zip code level is a specific and granular approach to identifying potentially important exposures.

QUESTION 1: General comment. This reader had to reread the manuscript multiple times to understand the role of the SE cluster of NTM cases. Were they modelled separately? It does not appear so, the model encompasses the state. I have come to the conclusion that you called them out because you identify them as a cluster and that is all. This is confusing. Maybe there should be less emphasis on this discovery (no inclusion in Figure 2) or somehow otherwise alert the reader that this was an incidental finding. Did you run a cluster analysis on controls? Maybe looking at case and control clusters is another environmental analysis report (suggestion). But in this report, the cluster is a distraction as currently presented.

RESPONSE: For clustering analysis, individuals with CF who were NTM culture positive (cases) were compared to individuals with CF who were NTM culture negative (controls) using a Bernoulli model. This analysis was repeated for all NTM cases as well as each species of NTM separately, resulting in finding the cluster of M. abscessus. We performed this analysis to highlight a specific geographical area of Florida where there may be a higher risk of acquiring NTM, important knowledge for persons with CF in addition to the environmental findings of our study. For this type of analysis, controls serve to determine the expected rate of cases in a given area, therefore we would not do a separate analysis on controls because they serve as our comparison group.

We have clarified the methods and purpose of this type of analysis in the text, which now reads: “We used a Bernoulli model comparing individuals with CF who were NTM culture positive (cases) to individuals with CF who were NTM culture negative (controls) to identify clustering of ZIP codes with higher than expected numbers of NTM culture positive individuals from 2011 through 2017, using SaTScan version 9.6 [17] with default setting, limiting cluster radius to 50km. This was done as a way to assess the presence of geographic locations in Florida where people with CF might be at increased risk of NTM. We repeated the analysis for overall NTM as well as each species of NTM separately."

Methods

QUESTION 2: How was zipcode assigned to patient and control? Did you use residential or hospital zipcode or something else? Was the same method used for both patient and control?

RESPONSE: Residential ZIP codes were obtained from the Cystic Fibrosis Foundation Patient Registry (annualized registry data) for both for both cases and controls. The ZIP codes used were those associated with the year the patients had their first positive culture and for controls, the year they had their first negative culture with Florida residency (i.e., at least two years living in Florida)

This approach has been clarified in the manuscript: “The first culture and residential ZIP code associated with the year of first culture for each person meeting these criteria was used for analysis.”

QUESTION 3: You analyze sputum positive CF patients (cases) and compare to controls: What is the rate of false negatives among sputum samples for NTM? Did you evaluate later cultures of controls to see if they became cases later during your six year study period?

RESPONSE: The choice of a control group was based on the approach that controls had to be at risk of infection in Florida – in other words, persons with CF that lived in the same state and during the same time period as cases. If any individuals were misclassified as controls but were sputum positive, this would only dilute the observed association between the exposures and the outcome of sputum positivity. For this reason, any disease misclassification resulting from a false negative within our control group would be biasing our findings towards a null association rather than a bias in the opposite direction.

We are not aware of systematic evaluation of rates of false negatives for nontuberculous mycobacterial cultures among persons with CF. Even though different laboratory techniques could have variable sensitivities, the laboratory methods for NTM species identification are relatively standard. We are not sure if the concern of the reviewer is false negatives related to laboratory techniques or sputum sampling techniques, or other sources of variability in mycobacterial detection. However, there is no reason to believe that the rate of misclassification would vary by zip code and therefore the cohort should be affected homogeneously.

To address the last point, it is possible that individuals may become reinfected or relapse, or that individuals who are negative in one year could become positive later, and individuals who are positive could clear infection. For this reason, we evaluated culture status of controls at follow up. Among controls who remained resident in Florida, none became cases during the six-year follow-up period. Four controls had a positive culture after moving out of Florida; we kept these four controls in our analysis to give our findings more power.

QUESTION 4: Line 64- you analyzed weather, soil or water mineral concentrations. I did not see the resulting values for these environmental exposures (maybe in a supplement)? Nor do I see a data source for water minerals in Table 1, although soil is listed.

RESPONSE: The variables listed on line 64 are examples of environmental variables previously found to be predictive of sputum positivity. The resulting values for environmental exposures used in analysis can be found in Table 3. We did not analyze water mineral variables in this study. We recognize that the wording of this section may be misleading, so have revised it to:

“We selected environmental variables for analysis based on prior findings. Variables that have been previously found to be predictive of sputum positivity include evapotranspiration [10], saturated vapor pressure [14], vapor pressure [15], temperature [6], and rainfall [6], as well as soil or water mineral concentration including copper [10], sodium [10], manganese, [8,10], calcium [7], and molybdenum [7]. Environmental data sources used in this study are described in Table 1.”

QUESTION 5: How were zip code and census polygons (income, metropolitan centers) reconciled?

RESPONSE: ZIP code and census polygons were reconciled by mapping the census data by county, then overlaying the ZIP code polygons to determine the counties each ZIP codes primarily fell within. We then joined the relating census data to the ZIP codes within each county for further analysis.

This methodology has been clarified in the manuscript: “To adapt county-level census data to the ZIP-code level, we determined the county each ZIP code primarily fell within by overlaying ZIP code polygons on a map of Florida counties using the R packages “rgdal” and “sp”.”

The reader needs more information about the temporal and spatial variability of covariates to be comfortable with the presence of environmental exposures included in the final model. Because variables such as temperature, precipitation, and evapotranspiration (“weather”) vary greatly month to month, year to year. Sea surface temperatures influence air temperature and precipitation both influence evapotranspiration. Early in your study period, there were anomalous lows, then anomalous highs, then lows again, ending in anomalous highs in 2017 in sea surface temperatures. See: https://twitter.com/MichaelRLowry/status/1249061412896464896/photo/1

RESPONSE: The authors recognize that there is temporal variability of the covariates that is not addressed in this analysis, however this is for two reasons: first, we only have annualized data on culture positivity for this analysis and second, the incubation period for NTM is not known but could range from months to years. For these reasons, we are unable to match more granular temporal data to positivity data.

QUESTION 6: This reader is concerned that there is a lack of information about this variability and how “weather” data are related to NTM culture positivity. What is the typical lag between NTM colonization and detection vs sputum sampling? How often are CF patients cultured? Please cite sources. This reader is seeing that analyses were conducted over a six year period (2011-2017). All subjects appear to be lumped into that time period. Mean “weather” variables are defined by year. What is the distribution of NTM culture positivity by month? Please include table. How is weather associated with sputum positivity over time, given potential lags? Was there enough power to perform a temporal analysis between weather and sputum positivity?

RESPONSE: As stated above, the incubation period for NTM is unknown, but is thought to be on the order of month to years, although likely shorter for persons with CF. In persons without CF, the time between symptom onset and diagnosis has been reported to range from months to years (Ratnatunga et al., 2020). Given the unknown incubation period and the delays in diagnosis, it seems reasonable to study associations on the order of years. While it is true that weather is highly variable, on average there is likely to be more variability across regions than within a given region. Although we recognize the limitations of these kinds of analyses, especially given the uncertain incubation period, the fact that our findings are in line with other studies evaluating the association of NTM culture positivity and environmental variables would suggest that these factors do play a role in NTM infection.

We did not assess the study as a six-year time period. To account for time and the temporal association of positivity with environmental variables, for each individual we used the environmental data associated with the year of the person’s first culture (first positive for cases, first negative for controls). So for example if someone was positive in 2013, we would use the annual mean precipitation, soil variables, and evapotranspiration from 2013. This has been clarified in the text, which now reads: “The final environmental variables included were copper, manganese, molybdenum, sodium, annual precipitation, and evapotranspiration from the year in which an individual had their first NTM culture.”

We did try lagging the environmental variables (evapotranspiration and temperature, specifically) by a year, but there were no significant effects in this analysis, so they were not included in any of the final models.

Ratnatunga, C. N., Lutzky, V. P., Kupz, A., Doolan, D. L., Reid, D. W., Field, M., Bell, S. C., Thomson, R. M., & Miles, J. J. (2020). The Rise of Non-Tuberculosis Mycobacterial Lung Disease. In Frontiers in Immunology (Vol. 11, p. 303). Frontiers Media S.A. https://doi.org/10.3389/fimmu.2020.00303

QUESTION 7: In short, there is a major mismatch between exposure and outcome. If you wish to include exposure variables that can change so much year to year, as I reader I expect to see time in the model. It is Florida. What about outlier precipitation years due to hurricanes? Effect on sputum positivity? I also expect to see the outcome analyzed by time. But NTM is considered a chronic condition. How do you reconcile this analysis of highly temporally variable “weather” with an outcome (sputum positivity) with no granularity or knowledge of time?

RESPONSE: We agree with the reviewer that weather conditions may be highly variable by time, including by month and year, and that the incubation period, the interval between exposure and infection is unknown. For these reasons, it is challenging to precisely associate exposure with outcome. This limitation is inherent to the multiple “ecologic” environmental studies that have been conducted to date for NTM where we are associating climatic variables, such as precipitation and humidity, at a broad scale with individual disease/infection or clusters of disease without knowing the exact period when a person was exposed. These types of biases would likely lead to misclassification of exposure resulting in associations closer to the null value. The finding in prior studies of associations between climatic factors and sputum positivity supports that these are real associations; for example, a prior national study of sputum positivity among persons with CF used climate data averaged over an 11-year period and found similar associations. This averaging likely smooths out the extreme values [see papers by Adjemian et al in main text references], although it is beyond the scope of the current analysis to be able to say whether an extreme event is a risk factor. Moreover, there is likely more variability across geographic areas within Florida than within areas, such that even an extreme event would likely not change the association a great deal.

In these particular models, the most precise approach we could take was to associate the temperature, precipitation, and evapotranspiration for each case and control from the year of the first culture that met the study definition. Thus, the model does account for outlier years, such as precipitation due to hurricanes. The relation of time in our model has been clarified throughout the methods.

We reconciled the analysis of climatic variables with sputum positivity by including only incident cases that had at least one negative culture before the first positive culture. However, since the incubation period of pulmonary NTM is unknown, there are unresolvable uncertainties around the appropriate timescales for measuring environmental exposures. This aspect is emphasized as a limitation of our study.

Soil is the also problematic. In the absence of additional information in the MS, these are the issues.

1-The reference for soil mineral data: https://pubs.usgs.gov/ds/801/ leads one to a summary report: “U.S. Department of the Interior, U.S. Geological Survey, Data Series 801, Geochemical and Mineralogical Data for Soils of the Conterminous United States” (“Data for Soils”). The report describes sampling methods which include a single point sample collection from “target sites that represented a density of approximately 1 site per 1,600 km2”.

QUESTION 8: Spatial variability of soil minerals is a large concern for this report. Lines 70-72 state ” Where necessary, ordinary kriging was performed to estimate the broader spatial distribution over Florida from the original sampling sites or weather stations. “ This statement comes after “weather” variables were discussed. Is this kriging approach also true for point estimates of minerals in soil by zipcode?

Please include citations for why a soil mineral analysis of this type -using kriging among spatial points 1,600 km2 apart, sampled once near the time of the study period, represents the exposure of people living among the points during the study period.

RESPONSE: We appreciate the concerns of the reviewer and recognize that the lack of precision in soil measurements is a limitation of this study. More precise measurements are not available and we felt that even though these measures were imperfect approximations of soil exposure, they could be indicative of risk. This approach was supported by similar methodologists that USGS has taken to produce continuous estimates of soil mineral concentrations throughout the contiguous United States [https://pubs.usgs.gov/of/2014/1082/pdf/ofr2014-1082.pdf].

Again, imprecision in measurement of exposure will likely lead to misclassification toward the null value, such that finding an association is likely indicative of a true effect. In two prior studies of NTM environmental risk, similarly imprecise measures of soil exposure were used but consistent associations were identified. For example, manganese was identified as a protective factor in two studies using various approaches in measurement (see Adjemian et al 2012, Lipner et al 2017 in main text references). Thus, although these studies and measurements have their limitations, they can indicate factors that may be causal and warrant further study.

Regarding the kriging, we performed ordinary kriging on all soil mineral variables as well as each year for precipitation and temperature to obtain continuous estimates of the variables across the state of Florida. The average of each variable within each ZIP code was then used for analysis.

This has been clarified in the manuscript: “Ordinary kriging was performed to estimate the broader spatial distribution over Florida from the original sampling sites or weather stations for all soil geochemistry variables, temperature, and rainfall.”

When one looks at variability in soil characteristics, this reader was surprised by the extreme temporal and spatial variability in concentrations of minerals. Let’s examine the two that made it to the final model, sodium and manganese.

Sodium: the exposure analysis using grid spacing for sodium samples of 1 site per 1,600 km2 appears to be too coarse to estimate your subjects exposures at the zipcode level. See: Trangmar et al. Application of Geostatistics to Spatial Studies of Soil Properties. Advances in Agronomy 1986. https://doi.org/10.1016/S0065-2113(08)60673-2

Manganese: in a study of agricultural fields, manganese concentrations differed among months of the year and over the extent of the field. Areas with higher concentrations did not keep rank order of concentrations over the year. See: Hoskinson, et al. Temporal Changes in the Spatial Variability of Soil Nutrients. INEEL/CON-99-00290. 1999.

https://www.osti.gov/servlets/purl/9823-CNgTWm/webviewable/

QUESTION 9: It appears that soil minerals are variable spatially and temporally due to multiple factors such as land use, vegetative cover, soil microbial communities, and precipitation.

Therefore, I fear that your assignment of soil mineral exposure was too simplistic. Please prove me wrong, because we need to better understand environmental risk factors for NTM colonization.

RESPONSE: We recognize that soil mineral concentrations likely vary more than represented in our model and have clarified this as a limitation in the manuscript:

“Additionally, longitudinal, granular data on soil geochemistry is currently unavailable; the soil mineral concentrations throughout Florida likely vary more than represented in our study, limiting our ability to adjust these variables based on time and location.”

When we average over time this may reduce some of the variability; it could be spurious, and that is why we feel that further investigation of these factors is warranted – we hope these data can serve as a call for further studies with more precise data. In addition, since many people are exposed to water, and soil mineral content is likely eventually reflected in the water, we feel that further studies which look at the water mineral content would be important.

Submitted filename: PONE_Reviewer Response_FINAL.docx

Click here for additional data file.

2 Nov 2021

Environmental predictors of pulmonary nontuberculous mycobacteria (NTM) sputum positivity among persons with cystic fibrosis in the state of Florida

PONE-D-21-06131R1

Dear Dr. Ricotta,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Thomas Byrd

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

1 Dec 2021

PONE-D-21-06131R1

Environmental predictors of pulmonary nontuberculous mycobacteria (NTM) sputum positivity among persons with cystic fibrosis in the state of Florida

Dear Dr. Ricotta:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Thomas Byrd

Academic Editor

PLOS ONE