Consequences of Climate Change on the Emergence of Pathogenic, Environmentally Acquired Nontuberculous Mycobacteria

Abstract

Background

Climate change, manifested by global warming, unpredictable precipitation, and increased frequency and severity of catastrophic weather, is a growing health threat. However, the impact that climate changes pose to environmental bacteria is not fully recognized.

Methods

To understand pathogen response to climate change, we interrogated nontuberculous mycobacteria (NTM) on a continental scale using open-source data products including Surface water microbe community composition data, soil microbe community composition data, and 16S ribosomal RNA (rRNA) gene sequences provided by the National Ecological Observatory Network (NEON) between 2015 and 2018.

Results

Of 6343 soil and water samples, 81.8% were positive for mycobacteria; soil samples had a higher positivity rate. NTM were also identified among a subset of 31 archived DNA samples, albeit in low proportion (6.5% [n = 2]). Viable Mycobacterium chelonae and Mycobacterium arabiense were recovered from 3.7% (3 of 81) biobanked NEON soil and aquatic sediment samples. Finally, using geographic coordinates of NTM from work in Hawai’i (a geographic hot spot for NTM infections), we modeled habitat associations during current and future climates. We found that the potential ranges for NTM are forecast to increase under future climate conditions and are strongly associated with increases in temperature, with pathogenic species accounting for most of the predicted surge.

Conclusions

Very little is known of the possible negative climate impacts on the emergence of disease due to environmental microbes. These data support the notion that NTM prevalence may be heavily augmented by climate change resulting in expansion into new geographic niches and posing new clinical consequences for humans.

Climate change is known to alter the frequency of soilborne, airborne, and waterborne pathogens. Herein, we characterize the distribution of environmental nontuberculous mycobacterial (NTM) species across the United States and forecast increased habitat suitability for NTM under future climate conditions.

The negative effects of climate change on human health are of growing concern [1]; changes include rising average and peak global temperatures and the increasing frequency of natural disasters, which increase levels of airborne particulates and expand distribution of microbial agents. Airborne infectious diseases pose a mounting burden on public health systems [2], but the impact of climate change on environmental bacteria of respiratory significance remains understudied.

Nontuberculous mycobacteria (NTM) are common to soil, dust, freshwater, and biofilms [3]. When these NTM-containing fomites are aerosolized and subsequently inhaled, a chronic, difficult-to-treat lung disease (LD) may occur [4–8]. NTMLD is associated with high morbidity and mortality rates, continues to increase in incidence and prevalence worldwide, and is now more prevalent in the United States and several parts of the world than tuberculosis, caused by Mycobacterium tuberculosis [9–12]. “Serovar” regional variations of NTM infections were initially highlighted during the AIDS epidemic [9, 10]. Since, the “geocentric” nature of US NTMLD is widely recognized, with more cases noted in southern, hotter, and more humid states than in northern states [11] and Hawai’i showing the highest NTMLD prevalence [11–13]. Due to their association with both natural and built environments and the overlap of their niches with those of humans, anthropogenic driven climate events are likely to influence NTM dispersal and increase human NTM exposures [3].

NTM acquisition is likely influenced by large-scale movement of soil and water, such as those observed during hurricanes, tornadoes, floods, or other natural disasters [14, 15]. Previously, our group reported that aerosolized volcanic ash from Hawai’i carry respiratory-relevant NTM [16]. While NTM most likely cannot survive in the active volcanic eruption itself, contact of cooled ash with NTM-enriched areas in the surrounding environment, along with the ease with which airborne ash may be carried for thousands of miles, suggests that NTM-associated ash is a newly discovered mode of NTM transmission.

In the current study, we leveraged archival data and banked samples from the National Ecological Observatory Network (NEON) [17] to assess the relative prevalence of NTM on a continental scale. NEON is a terrestrial, aquatic, atmospheric, and remote sensing measurement infrastructure that delivers environmental samples and calibrated data in an openly accessible data portal [18]. Sites are geographically distributed across the United States and established to understand longitudinal changes in ecosystems. We also applied previously collected NTM results from Hawai’i to model the habitat associations of various NTM species, predict future NTM movements, and identify future NTM hot spots. Our overarching goal was to expand our understanding of the relationship between climate changes and environmental NTM and explore the potential for NTM to proliferate with climate change.

MATERIALS AND METHODS

Open-Source Data Products Were Used to Screen for Mycobacteria at NEON Locations

NEON comprises 81 aquatic and terrestrial sites spanning 20 ecoclimate domains distributed across various US geographies. Each domain includes a terrestrial and aquatic core site selected to represent wildlands or more pristine ecosystems and gradient sites that encompass a range of ecological change that can be quantified [19]. Operational since 2019, NEON's purpose is to generate data and environmental samples for ecological research over a 30-year period. NEON open-source data products, including the surface water microbe community composition product DP1.20141.001 and soil microbe community composition product DP1.10081.001, were analyzed for mycobacteria. Using published NEON 16S ribosomal RNA (rRNA) gene sequences and taxonomic information, we queried 6343 surface water and soil samples collected between 1 November 2015 and 31 October 2018. Across this 3-year period, there were 633 site-days on which environmental samples were collected for 16S rRNA gene sequencing, with 1–7 samples per year depending on the site. The “mycobacteria” detected are overwhelmingly likely to be NTM since M tuberculosis is not readily found in the environment.

NTM Identification From Archival NEON Environmental Samples

A total of 112 archived NEON samples (31 DNA, 81 environmental) were assessed for NTM. First, 31 archival DNA samples collected between May 2014 and September 2018 from 15 NEON sites across 9 states were screened for NTM by sequencing the RNA polymerase beta subunit (rpoB) gene, as published elsewhere [13, 20]. In addition, 81 banked environmental samples—comprising aquatic sediment (n = 10), soil (n = 46), PM₁₀ air filters (n = 6), surface water microbes (n = 20), and sediments from the bottom of bodies of water (ie, benthic microbe) (n = 30) collected between July 2017 and May 2022 from 21 NEON sites across 9 states—were cultured onto Middlebrook 7H9 broth and on 7H10 agar plates, as described elsewhere [13, 20] (Supplementary Table 1A).

DNA Extraction and Gene Sequencing

Genomic DNA was extracted from pure environmental isolate cell pellets, following a mycobacterial specific protocol [21]. DNA extracts from mycobacterial pellets and previously extracted DNA supplied by NEON were identified by Sanger sequencing a 711–base pair rpoB gene region (Quintara) [22, 23]. Results were trimmed for quality control and compared with rpoB reference sequences in the National Center for Biotechnology Information GenBank, using their BLAST algorithm. If an isolate had a >90% BLAST identity and coverage match to any Mycobacterium, but no BLAST identity >97%, it was categorized as a putatively “novel” NTM.

Characteristics of Data From Prior Environmental Work in Hawai’i

Before the current study, we performed the largest environmental sampling campaign for NTM in the geographic hot spot of Hawai’i (2017–2021). Samples collected (n = 3735) included household biofilms, dust, and soil, as well as natural samples (eg, from streams or forest soil). Of the 3735 samples, 512 were positive for NTM, as determined by Sanger sequencing of the rpoB gene. Species-specific occurrence records were used for the ecological niche modeling below.

Regression Modeling of North-South and East-West Trends

Geographic trends in published Mycobacterium 16S rRNA gene sequences at NEON sites across the coterminous United States were estimated by regressing sample detection/nondetection against latitude and longitude (longitudes west of the prime meridian are considered negative) separately, and jointly in quasi-binomial logistic models in R 4.3.1 software [24, 25]. Latitude and longitude values were centered before regression. Models with quadratic terms in latitude and longitude were also investigated. All models included a separate intercept for each of the 4 site types (core aquatic, core terrestrial, gradient aquatic, and gradient terrestrial) to control for differences in underlying positivity rates across site types irrespective of geography. Statistical significance was assessed using an α cutoff of .05.

Ecological Niche Modeling

The Supplementary Methods contain additional modeling details. Briefly, separate ecological niche models were developed using the software Maxent for 4 NTM selected from the Hawai’i dataset: Mycobacterium chelonae, Mycobacterium abscessus, Mycobacterium intracellulare subsp chimaera, and Mycobacterium gordonae. Maxent applies the maximum entropy method to estimate the suitability of a rasterized (pixelated) landscape for each species on a pixel-by-pixel basis [26, 27]. The output is visualized as a heat map that displays a cloglog transformation, relating environmental data to habitat suitability for every location [28]. Habitat suitability scores range from 0 (most unsuitable) to 1 (most suitable). Maxent compares conditions at locations marked by the presence of documented species to conditions across the broader landscape and grades the habitat suitability of every pixel based on its similarity to locations where the species has been observed. The resulting map highlights areas with environmental conditions like those at observed species occurrence locations to predict other potentially suitable habitats not directly sampled.

Species location data for modeling were derived from DNA sequences recovered during sampling in Hawai’i. Environmental data consisted of gridded climate variables obtained from WorldClim [29] rather than weather or climate data collected previously in Hawai’i due to Maxent's requirement for complete, high-resolution datasets covering every landscape pixel including remote and unsampled areas. Species location data were integrated with WorldClim data to generate habitat suitability heat maps, with scores ranging from 0 to 1, across the study area. Maxent was also use to generate maps of future habitat suitability. These projections were not separate models but represented extrapolations of the current models to future climatic conditions. Existing models, trained on present-day data, predicted future habitat suitability by evaluating each pixel of future data based on its resemblance to present-day conditions at observed species presence locations, allowing habitat suitability estimations across the landscape under future scenarios.

High-resolution bioclimatic variables from Worldclim were at a 30-second scale and based on 30–40-year averages [30]. Variables with correlation less than |0.7| were retained for modeling purposes to avoid overfitting and to simplify model interpretations [31] using the “raster.cor.matrix” function of the ENMTools package [32] in R 4.0.0 [24]. Modeling was validated using the area under the receiver operating characteristic curve (AUC), the probability that a randomly chosen presence site will be ranked above a randomly chosen background site [33]. Models with an AUC >0.7 were considered reliable [34]. Model fit was measured with the gain statistic. Gain is a likelihood (deviance) statistic that measures the model performance compared with a model that assigns all areas of the landscape as equally suitable. Taking the exponent of the final gain gives the mean probability of the occurrence records compared with the background points. To determine future habitat suitability across the landscape, ecological niche models were projected onto future climate for the years 2041–2070, based on the IPSL-CM6A-LR climate model [35] and a shared socioeconomic pathway (SSP) SSP3-7.0 using Worldclim [36].

For each NTM species, Maxent settings were chosen using the “ENMevaluate” function of the ENMeval package [37] in R [24, 25]. All combinations of feature classes were considered at a range of regularization multipliers [25] from 0 to 3 at intervals of 0.5. Settings that resulted in the lowest corrected Akaike information criterion were chosen [38].

Finally, we estimated the potential geographic range sizes of NTM [39]—areas where NTM species are predicted to be found and areas where they could potentially live. We used the Maxent predictions of habitat suitabilities across the entire landscape to calculate the median habitat suitability for each NTM, conforming to the gradational nature of species boundaries [40]. To understand the extent to which the potential range sizes of NTM are forecast to change, we used the predicted habitat suitabilities of each species under both current and future climatic conditions.

RESULTS

Higher Prevalence of Mycobacteria in Soil Than in Aquatic NEON Sites

From our analyses of the NEON 16S rRNA sequencing data available, 5189 of the 6343 surface water and soil samples (81%) were positive for mycobacteria. The detection rate for mycobacteria was greater at core terrestrial sites (2475 of 2782 [89%]) and gradient terrestrial sites (2554 of 2936 [87%]) than at core aquatic sites (102 of 444 [23%]) and gradient aquatic surface water sites (63 of 181 [35%]). These NEON locations where mycobacteria were detected were mapped for the contiguous United States (Figure 1). NEON's 81 field sites are scattered among 20 ecoclimatic domains (D01–D20), established by NEON through strategic partitioning of the United States based on distinguishable qualities of the physical environment and ecosystems. The NEON domains with the highest detection rates were D05 (Great Lakes; 91.8%), D17 (Pacific Southwest; 91.8%), and D10 (Central Plains; 82.6%). The ecological regions with the lowest detection rates were D18 (Tundra; 37.9%), D15 (Great Basin; (58.1%), D04 (Atlantic Neotropical; 58.9%), D08 (Ozarks Complex; 66.4%), and D16 (Pacific Northwest; 66.5%).

Figure 1.

Sequencing data for the 16S ribosomal RNA (rRNA) gene reveal mycobacteria at National Ecological Observatory Network (NEON) sites. Existing NEON 16S rRNA gene sequencing data from 1 November 2015 and 31 October 2018 were analyzed for mycobacteria. The percentage of samples in which mycobacteria were detected are shown. Sites with no sequencing data between 1 November 2015 and 31 October 2018 are denoted by gray symbols. NEON sites in Hawai’i, Alaska, and Puerto Rico are not displayed. NEON ecological domain boundaries are included.

In our regression models controlling for site type, mycobacteria were positively associated with latitude (odds ratio [OR], 1.038 [95% confidence interval [CI], 1.024–1.053]; P < .001) and longitude (1.015 [1.009–1.020]; P < .001). The quadratic model for latitude yielded a positive estimate for the linear term (OR, 1.035 [95% CI, 1.021–1.050]; P < .001) and a negative estimate for the squared term (0.997 [.994–.999]; P = .02). The quadratic model for longitude yielded positive estimates for the linear term (OR, 1.017 [95% CI, 1.012–1.023]; P < .001) and the squared term (1.0006 [1.00024–1.00098]; P = .001). The model with both latitude and longitude yielded positive estimates for both latitude (OR, 1.045 [95% CI, 1.031–1.061]; P < .001) and longitude (1.018 [1.012–1.023]; P < .001).

In addition to 16S rRNA sequencing data, we applied partial rpoB sequencing to 31 archival DNA samples from the NEON biorepository, of which 2 of 31 (6.5%) were NTM positive. The rpoB-positive samples included soil DNA from the Harvard Forest site and soil DNA from the Smithsonian Conservation Biology Institute site, from which Mycobacterium wuenschmannii and a novel NTM species were identified, respectively (Supplementary Table 1B). An additional 81 archived NEON environmental samples, comprising 5 sample types, were cultured for viable NTM. Most were NTM culture negative (Supplementary Table 1C). However, Mycobacterium arabiense was cultured from an aquatic sample from the Prairie Lake site, and M chelonae was recovered from soil collected at both the Rocky Mountain National Park and the Smithsonian Conservation Biology Institute sites (Supplementary Table 1B).

Likely Proliferation of NTM Under Future Climate Conditions

Finally, we applied existing distribution data of environmental NTM from Hawai’i [41, 42] to model habitat associations and distributions on the landscape. Average annual temperature and temperature isothermality (a measure of how consistent the temperature range is throughout the year compared to the temperature range within a single day) were the most important variables predicting NTM species occurrence under current and future conditions. Suitability increased with increasing temperatures and decreased with increasing isothermality, which approximates a sigmoidal relationship where habitat suitability levels off at the extremes (Supplementary Table 2). The average annual temperature in Hawai’i is expected to increase and temperature isothermality is expected to decrease in the future (Figure 2), both of which favor the NTM species studied here (Figure 3).

Figure 2.

Increases in average annual temperatures (B) and decreases in temperature isothermality (a measure of how consistent the temperature range is throughout the year compared with the temperature range within a single day) (D) under future climatic conditions, compared with current conditions (A and C, respectively), meaning that temperatures will generally be higher throughout the year and will fluctuate less during the course of the day.

Figure 3.

More suitable habitats for pathogenic NTM are predicted to emerge under future climate conditions. Maxent modeling predicts that multiple NTM—including Mycobacterium chimaera, Mycobacterium gordonae, Mycobacterium chelonae, and Mycobacterium abscessus—will thrive under hotter future climates (A) and less fluctuating future temperatures (B), especially M chimaera, and that habitat suitability will increase under future climate conditions for all species studied (C). Gray vertical dotted and dashed lines in A and B represent current (dashed) and future (dotted) average conditions in Hawai’i. As indicated, the average annual temperature is predicted to increase in the future, whereas temperature isothermality (a measure of how consistent the temperature range is throughout the year compared with the temperature range within a single day) is predicted to decrease. See Supplementary Materials and Methods for more description of these variables.

Habitats suitable for rapid-growing M chelonae and M abscessus were predicted to increase by 120% and 205%, respectively, in the future (2041–2070) (Figure 4A–4D). Habitats were predicted to become more suitable for slow-growing M chimaera and M gordonae, increasing by 311% and 48%, respectively, in the future (Figure 4E–4H). Models predicted future expanded potential geographic range sizes for all NTM species studied, largely due to the relationships between average annual temperature and habitat suitability (Figure 3). Relationships between the species and other variables (eg, precipitation), were generally less important in determining the species' current and future potential range sizes (Supplementary Table 2 and Supplementary Figures 1 and 2).

Figure 4.

More suitable habitats emerge for rapid- and slow-growing NTM species under future climate conditions. A–D, Habitat suitability for rapid-growing, pathogenic Mycobacterium chelonae and Mycobacterium abscessus under current (A, C) and future (B, D) climates in Hawai’i. Habitat suitability scores range from 0 (most unsuitable) to 1 (most suitable). E–H, Habitat suitability for slow-growing Mycobacterium chimaera (pathogenic) and Mycobacterium gordonae (nonpathogenic) under current (E, G) and future (F, H) climates in Hawai’i.

DISCUSSION

Understanding the climactic factors that favor growth and spread of pathogenic bacteria is critical to public health. However, the consequences that climate changes pose to environmental bacteria, particularly those with the potential to cause human LD, are still not fully appreciated. NTM infections are believed to be acquired through human-environmental interactions, as evinced by genotype matching between environmental and respiratory isolates [43]. Because NTM infections occur as a result of inhalation or aspiration from environmental sources, we predict that NTMLD association with soil and water will be influenced by the accelerating trajectory of climate change. Indeed, Mycobacterium 16S rRNA gene sequences have been detected in dust caused by natural desert storms [44, 45], and NTMLD cases spike after hurricanes [15]. Thus, while the literature speculates that increased temperatures due to climate change will cause NTM proliferation [46, 47], to our knowledge this has never been quantitatively modeled, despite calls for such research [15, 46, 48, 49]. To address this gap in knowledge, we investigated the distribution of environmental NTM on a continental scale using NEON data products. In addition, using existing data we previously generated, we demonstrate the expansion of NTM habitats under future climate conditions, likely due to elevated temperatures.

NEON was established to create an infrastructure for ecology to facilitate landscape-level studies on a continental scale [50] and applied to study ecological niches of mosquito pathogens and soil properties [51–53]. Brumfield et al [54] demonstrated the feasibility of applying existing metagenomic datasets from the NEON open-source data portal to elucidate microbial community structure and diversity nationally. Herein, our combined analyses of 16S rRNA gene sequencing data from NEON and microbiological culture for viable NTM suggest that there are geographic differences in the detection of mycobacteria. Analysis of national NEON 16S rRNA gene sequencing data found 23%–89% NTM positivity rates, higher than the 0.03%–2.9% rates detected in a global survey of mycobacterial diversity in soil using similar analyses of 16S rRNA sequencing data [55]. Mycobacteria were most frequently detected at terrestrial and aquatic sites in the D05 (Great Lakes), D17 (Pacific Southwest), and D10 (Central Plains) NEON domains, though all domains yielded detection rates >50%, except D18 (Tundra), which encompasses coastal northern and western Alaska.

In regression models covering NEON sites in the coterminous US controlling for site type, the odds of mycobacteria detection increased with both longitude (ie, increasing further east) and rising latitude, up to roughly 44 degrees. Trends predicted the highest and lowest detection rates at NEON regions due to the models' adjustment for site type, the distribution of which varied across regions. Yet, the domain furthest to the east, D01 (Northeast) had the sixth-highest detection rate (80.0%), D02 (Mid Atlantic) had the fourth highest, and D05 (Great Lakes) had the highest, suggesting that NTM may have a mild preference for coniferous and broadleaf forests and/or cool, wet climates. However, we recognize limitations such as the snapshot nature and narrow pocket of time of the data used. Furthermore, this analysis cannot distinguish pathogenic from nonpathogenic NTM species, which may have different geographic distributions.

From soil DNA extracted by NEON, we identified M wuenschmannii and a novel NTM from Harvard Forest and Smithsonian Conservation Biology Institute NEON sites, respectively. Located 65 miles west of Boston, Massachusetts, Harvard Forest is a cool and temperate area covered by hardwoods and coniferous forest within the Quabbin watershed, the largest US freshwater reservoir. Located in the foothills of the Blue Ridge Mountains in Front Royal, Virginia, the Smithsonian Conservation Biology Institute is a temperate and humid terrestrial field site.

In parallel, we recovered viable and clinically relevant M chelonae from Rocky Mountain National Park and Smithsonian Conservation Biology Institute soil. The Rocky Mountain site is located 75 miles northwest of Denver, Colorado, and is dominated by granite and mollisol soil within lower montane ecosystems of pine and firs. This concurs with our prior finding of M chelonae from Colorado but from water biofilms collected across the Denver-Rocky Mountain corridor [56]. M chelonae is also common to the NTMLD hot spot of Hawai’i [13], an area characterized by iron rich soil [57]. It's likely that M chelonae is an environmental generalist, capable of surviving in varied niches. Viable M arabiense was cultured from an aquatic sediment sample from the Prairie Lake, North Dakota, site comprising rolling hill grassland, agricultural, and wetland/aquatic areas. M arabiense is a rapid-growing NTM of ill-described pathogenesis, first identified from Dubai coastal sand [58] and India soil [59]. Recovery of M chelonae and M arabiense exemplifies the ability of NTM to adapt, survive, and thrive in varied environments spanning continental miles.

Our group previously recovered viable NTM known to cause LD from both built and natural environments in Hawai’i [13, 20]. We applied this knowledge to elucidate the potential impacts of climate change on the emergence of pathogenic NTM using modeling approaches. We found that bioclimatic rainfall patterns did not appreciably contribute to modeled NTM prevalence, indicating that NTM are suitable for survival in either dry or wet climates. More studies are needed, however, since other work suggests that floods, tsunamis, and hurricanes likely exacerbate NTM acquisition [14, 15]. A significant finding was that our models predicted that Hawai’i will be warmer and experience less cooling at night in the future, features associated with the proliferation of M abscessus and M chimaera. Thus, we quantitatively modeled climate changes to identify geographic areas of greatest concern for emerging NTM threats that can be targeted for further surveillance [60] and preemptive mitigations [61, 62]. Other climate-associated natural disaster events should be continually monitored including wildfires and volcanic eruptions. We have already reported respiratory relevant NTM from Kīlauea volcanic ash [16], suggesting links between pathogenic NTM and volcanic activity.

Study limitations include the low number of NEON DNA and environmental samples tested, thus, making generalizing our results difficult. While our focused work in Hawai’i is a clear strength, it is also a limitation and warrants further investigations to elucidate the association between NTM and climate changes on a broader scale. We previously reported that NTM recovery from soil is not pH driven and that soil kaolin and halloysite nutrients do not modulate the replication of environmental M abscessus and M chimaera isolates in vitro [63], but other variables, such as increased temperatures and humidity, should be assessed in future work. Finally, although we focused on prediction of NTM prevalence in the environment based on future climate changes, we recognize as a potential confounder that virulence may increase following infection in the host. One mechanism for this increase is antimicrobial resistance due to environmental or in vivo biofilm formation or selection pressures imposed by antibiotics. A separate mechanism is displayed by NTM that possess glycopeptidolipids in their cell envelope where, following infection in the host, glycopeptidolipids may be reversibly lost (as with M abscessus), with the transition from smooth to rough colony morphology having been shown to increase organism virulence [64] or irreversibly lost (in the case of Mycobacterium avium complex).

While increasing temperature was linked to higher recovery of pathogenic NTM in Hawai’i, in the coterminous United States, detection of mycobacteria at NEON sites was positively associated with latitude (up to 44 degrees). This potential contradiction could be caused by a number of factors. The positive trend in latitude at NEON sites may be driven by the pathogenic species not analyzed in the Hawai’i study, which may counteract a putative negative trend for species that were analyzed in that study. This could be due to the difference in soil or ecosystem composition in Hawai’i compared with the coterminous United States. However, despite the seeming inconsistency with respect to temperature, both studies found that NTM was more common in areas with higher humidity.

Climate change is now widely recognized as a serious contemporary challenge facing humanity [65]. For the thousands living at increased risk of NTM infections nationally, and even more globally, it is crucial to determine whether NTM are climate-sensitive opportunistic pathogens. Existing evidence using a Salmonella outbreak and meteorological data suggest that warmer temperatures are positively correlated with gastrointestinal infections and increased bacterial replication [66]. Warming climates are also linked to increasing antibiotic resistance in Europe, possibly related to increased bacterial mutation rates with more rapid replication [67]. Data from the Antimicrobial Surveillance Network and Statistical Yearbook database in China revealed an association between increased ambient temperatures and higher rates of antibiotic resistance [68]. In closing, this work provides justification for further studies to elucidate the consequential role of weather changes on respiratory pathogens related to public health globally.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.