Co-Benefits of Antimicrobial Resistance Mitigation from China’s PM2.5 Air Pollution Reduction Between 2014–2020

Zhenchao Zhou; Zejun Lin; Xinyi Shuai; Xiaoliang Ba; Chioma Achi; Mark A. Holmes; Tong Xu; Yingru Lu; Yonghong Xiao; Jianming Xu; Baojing Gu; Hong Chen

doi:10.1016/j.eng.2024.09.013

Engineering ›› 2025, Vol. 45 ›› Issue (2) :257 -265. DOI: 10.1016/j.eng.2024.09.013

Research Air Pollution Control—Article

research-article

Co-Benefits of Antimicrobial Resistance Mitigation from China’s PM_2.5 Air Pollution Reduction Between 2014–2020

Author information +

^aCollege of Environmental and Resource Sciences, Zhejiang University, Hangzhou 310058, China

^bDepartment of Veterinary Medicine, University of Cambridge, Cambridge CB3 0ES, UK

^cIneos Oxford Institute of Antimicrobial Research, Department of Biology, University of Oxford, Oxford OX1 3RE, UK

^dCentre for Environment, Energy and Natural Resource Governance, Department of Land Economy, University of Cambridge, Cambridge CB3 0ES, UK

^eUK of Intensive Care Unit, The First Affiliated Hospital of Wenzhou Medical University, Wenzhou 325000, China

^fState Key Laboratory for Diagnosis and Treatment of Infectious Disease, The First Affiliated Hospital, Zhejiang University, Hangzhou 310003, China

^gInternational Cooperation Base of Environmental Pollution and Ecological Health, Science and Technology Agency of Zhejiang, Zhejiang University, Hangzhou 310058, China

* E-mail address: zhouzc@zju.edu.cn (Z. Zhou),

E-mail address: chen_hong@zju.edu.cn (H. Chen).

Show less

History +

PDF (1119KB)

Abstract

The One Health concept acknowledges the importance of multiple dimensions in controlling antimicrobial resistance (AMR). However, our understanding of how anthropological, socioeconomic, and environmental factors drive AMR at a national level remains limited. To explore associations between potential contributing factors and AMR, this study analyzed an extensive database comprising 13 major antibiotic-resistant bacteria and over 30 predictors (e.g., air pollution, antibiotic usage, economy, husbandry, public services, health services, education, diet, climate, and population) from 2014 to 2020 across China. The multivariate analysis results indicate that fine particulate matter with a diameter of 2.5 μm or less (PM_2.5) is associated with AMR, accounting for 12% of the variation, followed by residents’ income (10.3%) and antibiotic usage density (5.1%). A reduction in PM_2.5 of 1 µg·m⁻³ is linked to a 0.17% decrease in aggregate antibiotic resistance (p < 0.001, R² = 0.74). Under different scenarios of China’s PM_2.5 air-quality projections, we further estimated the premature death toll and economic burden derived from PM_2.5-related antibiotic resistance in China until 2060. PM_2.5-derived AMR is estimated to cause approximately 27 000 (95% confidence interval (CI): 646848 830) premature deaths and about 0.51 (95% CI; 0.12–0.92) million years of life lost annually in China, equivalent to an annual welfare loss of 8.4 (95%CI; 2.0–15.0) billion USD. Implementing the “Ambitious Pollution 1.5 °C Goals” scenario to reduce PM_2.5 concentrations could prevent roughly 14 000 (95% CI; 3324–26 320) premature deaths—with a potential monetary value of 9.8 (95% CI; 2.2–17.6) billion USD—from AMR by 2060. These results suggest that reducing air pollution may offer co-benefits in the health and economic sectors by mitigating AMR.

Graphical abstract

Keywords

Antimicrobial resistance / Air pollution / Antibiotic usage / Multivariable analysis / Premature deaths / Nationwide

Cite this article

Download citation ▾

Zhenchao Zhou, Zejun Lin, Xinyi Shuai, Xiaoliang Ba, Chioma Achi, Mark A. Holmes, Tong Xu, Yingru Lu, Yonghong Xiao, Jianming Xu, Baojing Gu, Hong Chen. Co-Benefits of Antimicrobial Resistance Mitigation from China’s PM_2.5 Air Pollution Reduction Between 2014–2020. Engineering, 2025, 45(2): 257-265 DOI:10.1016/j.eng.2024.09.013

登录浏览全文

4963

注册一个新账户忘记密码

1. Introduction

Antimicrobial resistance (AMR) is a growing threat to human health and social well-being. It is projected to cause up to 10 million deaths per year by 2050, and its cumulative cost will reach 100 trillion USD in a “do-nothing” scenario [1]. Despite ongoing efforts, AMR remains uncontrolled, contributing to an estimated 1.27 million deaths globally in 2019—a substantial increase from the 0.7 million deaths in 2016 [1], [2]. While the overuse and misuse of antibiotics in humans and animals are widely recognized as key contributors to this crisis, other factors such as socioeconomic conditions, climate change, sanitation, and environmental transmission through the air, water, and soil may also play crucial roles, particularly in developing countries [3], [4], [5]. Factors such as inadequate sewage treatment, overcrowdedness, poor sanitation, and high doses of antibiotic consumption intensify the transmission and evolution of antibiotic resistance in these regions [6], [7]. Notably, some antibiotic-resistance genes (ARGs), such as the carbapenem resistance gene (bla_NDM-1) in patients from New Delhi, India, and the plasmid-mediated colistin resistance gene (mcr-1) in animals from Shanghai, China, have emerged from these areas, spreading rapidly worldwide [8], [9], [10] and jeopardizing our capacity to manage infectious diseases. Therefore, a more in-depth and systematic understanding of the factors driving AMR in these regions is urgently required.

Many developing countries lack publicly available and comprehensive data or systems for reporting antibiotic resistance, antibiotic usage, and socioeconomic factors, making it difficult to evaluate national progress in addressing AMR. The multifaceted nature of AMR within the One Health context underscores the importance of understanding the drivers of AMR [11], [12], [13]. Previous studies [3], [4] have demonstrated significant correlations between AMR and various factors, including local temperature, gross domestic product (GDP) per capita, education, basic infrastructure, public healthcare spending, and antibiotic consumption. As a developing country, China is confronting challenges related to environmental pollution, economic development inequality, and antibiotic resistance. Since the 2010s, the country has diligently established a comprehensive nationwide antibiotic-resistance surveillance system and collected substantial amounts of data, providing a representative region for understanding the drivers of antibiotic resistance in developing countries.

Environmental pollution as a significant factor in AMR transmission, or AMR levels, has not been comprehensively analyzed alongside other possible factors. Air pollutants have recently been recognized as important vectors for AMR transmission [14]. Fine particulate matter with a diameter of 2.5 μm or less (PM_2.5) air pollution can carry diverse and abundant antibiotic-resistance elements, leading to the long-range aerial transmission of AMR via dust suspension and dry or wet deposition [15]. More than 80% of the materials in PM_2.5 air pollution can be deposited in the human body and penetrate via the lungs [16], and long-term exposure to antibiotic resistance from the environment has been shown to affect antibiotic resistance in humans [17]. In addition, there is evidence that ARGs and microbes are more abundant in contaminated air than in non-contaminated air in China [18]. Our recent research also demonstrates that air pollution significantly impacts antibiotic resistance worldwide [19]. However, there are too many factors that vary from country to country, making it unclear how this phenomenon plays out within countries.

Here, we present a comprehensive analysis of the relationship between AMR and potential contributing factors, including air pollution, antibiotic usage, economic measures, livestock agriculture, public services, health services, education, diet, climate, and population across China from 2014 to 2020 using univariate, multivariable, and structural equation model (SEM) analyses. Against a background of clean-air and carbon-neutrality policies, China has substantially reduced air contamination in recent years and is committed to continuing this trend [20]. Employing six scenarios that combine climate mitigation pathways, we further estimated and projected the burden of premature deaths and welfare loss attributable to AMR caused by PM_2.5 in China until 2060. These findings highlight the importance of environmental factors in understanding AMR prevalence and the potential of improved air quality in controlling AMR in developing countries.

2. Methods

2.1. Study design and data sources

We compiled a dataset that includes antibiotic resistance, antimicrobial use density (defined daily doses (DDD) per 100 people), and over 30 other factors in mainland China (31 provinces, autonomous regions, and municipalities) from 2014 to 2020 (Table S1 in Appendix A). Data on antibiotic resistance and antimicrobial use density were sourced from the Status Report on Antimicrobial Administration and Antimicrobial Resistance by the National Health Commission of the People’s Republic of China [21]. The present study utilized data on 13 antibiotic-resistant bacteria, as detailed in Table S1. These bacteria include carbapenem-resistant Escherichia coli (CREC), carbapenem-resistant Pseudomonas aeruginosa (CRPA), carbapenem-resistant Acinetobacter baumannii (CRAB), carbapenem-resistant Klebsiella pneumoniae (CRKP), erythromycin-resistant Streptococcus pneumoniae (ERSP), methicillin-resistant Staphylococcus aureus (MRSA), methicillin-resistant coagulase-negative staphylococci (MRCNS), penicillin-resistant Streptococcus pneumoniae (PRSP), quinolone-resistant Escherichia coli (QREC), third-generation cephalosporin-resistant Escherichia coli (3GCREC), third-generation cephalosporin-resistant Klebsiella pneumoniae (3GCRKP), vancomycin-resistant Enterococcus faecalis (VREFS), and vancomycin-resistant Enterococcus faecium (VREFM). The AMR in these bacteria was normalized and collectively termed “aggregate antibiotic resistance.” Datasets on sulfur dioxide (SO₂), nitrogen dioxide (NO₂), particulate matter with a diameter of 10 μm or under (PM₁₀), carbon monoxide (CO), and ozone (O₃) were collected from the National Bureau of Statistics of China. PM_2.5 data for China were collected from an online dataset of annual regional-level mean population-weighted PM_2.5^† [22]. Additional data on various potential factors such as the economy (all residents’ income and GDP), livestock agriculture (pork, beef, and lamb production), public services (urban water penetration rate, daily treatment capacity of urban sewage, public toilets per 10 000 people, harmless domestic waste treatment, daily domestic water consumption, and green park space), health services (number of beds in medical institutions, health technicians per 10 000 people, urban basic medical insurance, village clinics, medical, and health institutions), education (students in higher education institutions, high school, junior high school, primary school, and kindergarten), diet (meat consumption), climate (temperature, sunshine, and humidity), and population (resident population and urban population density) were sourced from the National Bureau of Statistics of China.

2.2. Statistical analysis

Multivariate correlations between aggregate antibiotic resistance and predictors were examined using a panel fixed-effects analysis in Stata 16 (StataCorp LP, USA) and a multivariate forward stepwise regression analysis in SPSS V25.0, IBM, USA. The quality of the results from the multivariable analysis was assured through multicollinearity analysis (variation inflation factors < 10), residual analysis, and White’s heteroskedasticity test. Unlike previous global and multivariable research [3], [19], we further employed a non-recursive SEM to reveal both the direct and indirect influences of the predictors on aggregate antibiotic resistance and the reciprocal effect of aggregate antibiotic resistance on antibiotic usage. This approach allowed for the derivation of comparable effects by integrating the multiple relationships between variables using standard deviation coefficients [23]. For example, we discovered through the SEM that antibiotic usage density directly influences aggregate antibiotic resistance, whereas temperature indirectly impacts aggregate antibiotic resistance. In addition, the non-recursive SEM addressed the reciprocal effects between antibiotic resistance and antibiotic usage, facilitating empirical studies of this feedback relationship on a national scale. Relationships among predictors with partial correlation, such as PM_2.5’s correlation with antibiotic usage density, were explored using the SEM. The equations for the SEM analysis are presented in the following forms:

(1)

A R = α_{A R} + β_{1} \times A U + β_{2} \times {P M}_{2.5} + β_{3} \times P T + β_{4} \times H T + β_{5} \times V C + β_{6} \times R I + β_{7} \times U P D + β_{8} \times L P + e . A R

(2)

A U = α_{A U} + γ_{1} \times A R + γ_{2} \times {P M}_{2.5} + γ_{3} \times T + γ_{4} \times V C + γ_{5} \times M C + γ_{6} \times U P D + γ_{7} \times P P + γ_{8} \times E C + e . A U

(3)

{D E}_{i} = β_{i}

(4)

{I E}_{i} = γ_{i} \times β_{1}

(5)

{T E}_{i} = β_{i} + γ_{i} \times β_{1}

where AR is antibiotic resistance, AU is antibiotic usage density, PT is public toilets, HT is health technicians, VC is village clinics, RI is residents income, T is temperature, EC is education, UPD is urban population density, LP is lamb production, MC is meat consumption, PP is pork production, DE is direct effect, IE is indirect effect, TE is total effect, α_AR and α_AU represents the constant, β₁–β₈ and γ₁–γ₈ represent the path coefficients that need to be calculated, e represents an error term, and the subscript i denotes an exogenous variable. The goodness-of-fit statistics for the model are within the following expected ranges: goodness of fit index (GFI) = 0.96 > 0.90, Tucker–Lewis index (TLI) = 0.95 > 0.90, comparative fit index (CFI) = 0.99 > 0.90, root mean square error (RMSE) = 0.05 < 0.08, and standardized root mean square residual (SRMR) = 0.012 < 0.05. The non-recursive SEM satisfies the stability index = 0.32 < 1. All the eigenvalues are positioned inside the unit circle.

2.3. Impact of PM_2.5-derived AMR on premature death

Death tolls involving infections were obtained from the global burden of disease studies [24]. Annual premature death tolls attributable to AMR were estimated using the population-attributable fraction (PAF) of the infection-based deaths. The formula for PAF is as follows [2]:

(6)

P A F = p \times (R R - 1) / (1 + p \times (R R - 1))

where RR refers to the relative risk of antibiotic-resistant pathogens summarized by a previous study [2], and p is the prevalence of AMR. Changes in aggregate resistance and PAF with shifts in PM_2.5 levels are shown in Fig. S1 in Appendix A.

Years of life lost (YLLs) for the 31 provinces, autonomous regions, and municipalities were estimated based on the Global Burden of Disease studies using an average YLL per premature death of 18.5 (18.0–19.2) years. The formula of

{YLL}_{r, y}

is as follows [25]:

(7)

{Y L L}_{r, y} = {Y L L}_{E U} \times {(G_{r, y} / G_{E U})}^{s}

where s is the income elasticity. The average base welfare loss of one year of life lost in EU countries (

{Y L L}_{E U}

) and the average GDP per capita of EU countries (

G_{E U}

) were used to estimate the welfare loss of one year of life lost (

{Y L L}_{r, y}

) in regions (r) in the year (y) based on the regional GDP per capita (

G_{r, y}

2.4. Scenario analysis

To evaluate pathways of PM_2.5 air quality until 2060, we utilized six scenarios: baseline, neutral goals (NGs), ambitious pollution 1.5 °C goals (AP1.5Gs), ambitious pollution 2 °C goals (AP2Gs), national determined contributions goals (NDCGs), and current goals (CGs) [26]. These scenarios combine representative concentration pathways (RCPs), shared socioeconomic pathways (SSPs), and air contamination control policies; definitions have been detailed in a previous study [26]. The baseline scenario integrated the climate constraints of RCP6.0 and the socioeconomic factors of SSP4 and controlled the same end-of-pipe pollution as observed at the 2015 level. The CGs scenario integrated RCP4.5, SSP2, and the recent policies in China. The NDCGs scenario combined RCP4.5 and SSP2 and used the best end-of-pipe contamination–mitigation strategies. The NGs scenario incorporated the climate constraints of net-zero carbon dioxide emissions in China by 2060 and SSP1 and controlled the same end-of-pipe contamination as the NDCGs. The AP2Gs scenario integrated RCP2.6 and SSP1 and used the same end-of-pipe pollution control as NDCGs. The AP1.5Gs scenario integrated RCP1.9 and SSP1 and controlled the same end-of-pipe contamination as NDCGs. Changes in antibiotic resistance were calculated based on the estimated coefficients of the equation [19].

Our focus was on the impact of future changes in air pollution, and other indicators remained unchanged as control variables. We projected the premature death and welfare loss attributable to AMR derived from PM_2.5 through 2060 using PM_2.5 concentrations, GDP (at constant USD prices from 2011) and population data from these six scenarios via a generalized additive model [25], [27], [28]. Distribution maps of China, linear regression plots, circus-bar plots, and projection plots were generated using R (Version 4.0.5) with maptools, ggplot2, tidyverse, and cowplot.

3. Results

3.1. Antibiotic resistance decreases with decreasing air pollution

Annual antibiotic resistance data and most potential factors were available from 2014 to 2020, except for antibiotic usage density and meat consumption, which were available from 2015 to 2020 (Table S1 in Appendix A). The average antibiotic resistance rate for aggregate resistance was 41.3% (standard deviation (SD) = 11.3%). For specific types of antibiotic resistance, the rates were as follows: 93.4% for ERSP, 76.0% for MRCNS, 55.1% for 3GCREC, 54.8% for CRAB, 52.8% for QREC, 32.9% for 3GCRKP, 31.9% for MRSA, 19.8% for CRPA, 8.1% for CRKP, 3.0% for PRSP, 1.6% for CREC, 1.5% for VREFS, and 0.5% for VREFM. From 2015 to 2020, PM_2.5 concentrations, aggregate antibiotic resistance, and antibiotic usage density decreased by 16.7 µg·m⁻³, 11.2%, and 3.7 DDDs per 100 inhabitants per day, respectively (Fig. S2 in Appendix A). Univariate analysis showed that aggregate antibiotic resistance was significantly associated with PM_2.5 (p < 0.001, R² = 0.97) (Figs. S3 and S4 in Appendix A). The correlations between the predictors are displayed in Fig. S5 in Appendix A. Linear regressions between aggregate antibiotic resistance and PM_2.5 were stratified by the years 2014–2015, 2016–2017, and 2018–2020 (Fig. 1(a)). Notably, changes in PM_2.5 levels were found to correlate with more significant changes in aggregate antibiotic resistance in subsequent years. The slope values (percentage resistance per microgram per cubic meter of PM_2.5(%·(µg·m⁻³)⁻¹PM_2.5)) of the linear regression for the years 2014–2015, 2016–2017, and 2018–2020 were 0.34, 0.39, and 0.44, respectively (p < 0.001, R² = 0.19–0.23). Fig. 1(b) illustrates the slopes from the relationship between PM_2.5 and the 13 antibiotic-resistant pathogens (p < 0.001, R² = 0.12–0.24). Changes in PM_2.5 were found to be associated with greater changes in CRAB (slope value of 0.3), 3GCRKP (0.26), CRPAg (0.23), and 3GCREC (0.2).

The multivariable regression analysis for the panel fixed-effects analysis was limited to data from 2015 to 2020, as antibiotic usage density data were available for that period. The multivariable analysis of panel fixed effects and the stepwise correlation analysis demonstrated that PM_2.5 and antibiotic usage density were significantly associated with antibiotic resistance (p < 0.001) (Table 1 and Table S2 in Appendix A). In the multivariable regression analysis, a reduction of 1 µg·m⁻³ in PM_2.5 and one DDD per 100 inhabitants per day in antibiotic usage density resulted in significant declines in aggregate resistance by 0.17% and 0.31% (p < 0.001, R² = 0.737), respectively. Among the factors with strong positive correlations, PM_2.5 contributed 12.0% to AMR, followed by residents’ income (10.3%), antibiotic usage density (5.1%), pork production (4.7%), public toilets per 10 000 people (4.0%), number of village clinics (3.8%), and education (2.3%). Some factors, such as health technicians, meat consumption, lamb production, urban population density, and annual temperature, were inversely correlated with aggregate antibiotic resistance, contributing between 0.9% and 2.3% to antibiotic resistance. The multivariable analysis confirmed that the contribution of PM_2.5 was higher to CRAB, 3GCRKP, CRPA, and 3GCREC than to other pathogens, ranging from 6.2% to 9.8% for these four pathogens (Table S3 in Appendix A).

To reveal the direct or indirect influences of multiple factors on aggregate antibiotic resistance, we introduced the SEM to reveal the pathways of influence (Fig. 2). The goodness-of-fit statistics, including the GFI, TLI, CFI, RMSE, SRMR, and non-recursive SEM, satisfied the stability index of the model and were within the expected ranges, indicating good reliability. Antibiotic usage density and public toilets exhibited positive direct effects of 0.34 and 0.47, respectively, on aggregate antibiotic resistance. Conversely, health technicians and lamb production demonstrated negative direct effects of –0.28 and –0.26, respectively, on aggregate antibiotic resistance. Temperature, pork production, meat consumption, and education influenced aggregate antibiotic resistance indirectly, with respective effects of 0.16, 0.16, –0.09, and –0.12 mediated through antibiotic usage density. Notably, aggregate antibiotic resistance negatively impacted antibiotic usage density by –0.3. PM_2.5 exerted a positive direct effect on aggregate antibiotic resistance with a standardized path coefficient of 0.38, alongside a positive indirect effect of 0.13 via antibiotic usage density. This resulted in a total effect of 0.51 for PM_2.5 on aggregate AMR, where the direct effect constituted 75% and the indirect effect via antibiotic usage density comprised 25%. Similarly, village clinics and urban population density had respective total effects of 0.28 and –0.19 on aggregate antibiotic resistance.

3.2. Premature mortality burden attributable to AMR derived from PM_2.5

Our analysis revealed that AMR derived from PM_2.5 caused approximately 27 000 (95% confidence interval (CI): 6468–48 830) premature deaths and about 0.51 (95% CI; 0.12–0.92) million years of life lost in China annually, equivalent to a welfare loss of 8.4 (95% CI; 2.0–15.0) billion USD (Fig. S6 (a) and (b) in Appendix A). From 2014 to 2020, the death tolls attributable to PM_2.5-derived AMR decreased by approximately 8000 (95% CI; 1842–14 607) nationwide (Fig. S6(c)). Hebei, Hunan, and Jiangsu provinces saw the greatest reductions in premature deaths, ranging from 600 to 900. The changes in welfare loss caused by premature death between 2014 and 2020 varied regionally. Hebei province experienced the greatest decrease in welfare loss at 91.7 (95% CI; 20.7–164.4) million USD, while Henan Province saw the greatest increase in welfare loss at 43.3 (95% CI; 9.8–77.7) million USD (Fig. S6(d)). In 2020, the marginal premature mortality caused by AMR derived from PM_2.5 in China was estimated at 640 (95% CI; 144–1142) deaths per microgram per cubic meter of PM_2.5 (Fig. S6(e)). Guangdong, Shandong, and Hunan Provinces saw the greatest increases in marginal premature mortality during this period, which increased by approximately four deaths per microgram per cubic meter of PM_2.5. The marginal welfare loss of premature mortality was estimated at 0.23 (95% CI; 0.05–4.08) billion USD per microgram per cubic meter of PM_2.5 in 2020 (Fig. S6(f)). Changes in marginal premature mortality between 2014 and 2020 increased by 60 (95% CI; 13–105) deaths per microgram per cubic meter of PM_2.5 (Fig. S6(g)). The changes in marginal welfare loss between 2014 and 2020 increased by 83.8 (95% CI; 1.9–150.3) million USD per microgram per cubic meter of PM_2.5 nationwide (Fig. S6(h)). Guangdong, Sichuan, and Jiangsu provinces saw the largest increases in marginal welfare loss, which increased by approximately seven million USD per microgram per cubic meter of PM_2.5.

3.3. Projection of premature death and welfare loss until 2060

We projected the premature death and welfare loss caused by AMR derived from PM_2.5 until 2060 under six scenarios. The AP1.5Gs scenario showed the most significant decrease in premature death and welfare loss in 2060, followed by NGs, AP2Gs, NDCGs, and CGs (Fig. 3). Under the baseline scenario, national premature death attributable to PM_2.5-derived AMR is projected to increase until 2040, followed by a decrease until 2060. Under the CGs and NDCGs scenarios, premature death is expected to remain steady until 2040 and then decline (Fig. 3(a)). The welfare loss from premature death is projected to increase under the baseline and CGs scenario and decrease under NDCGs, AP2Gs, NGs, and AP1.5Gs scenario until 2060 (Fig. 3(b)). AP1.5Gs, NGs, and AP2Gs scenario are projected to reduce the premature deaths caused by AMR derived from PM_2.5 to 4200–6000 in 2060, equivalent to a reduction in welfare loss of 3.2–4.5 billion USD. The benefits of reducing PM_2.5 under the AP1.5Gs scenario could prevent approximately 14 000 (95% CI; 3324–26 320) premature deaths—equivalent to 9.8 (95% CI; 2.2–17.6) billion USD—attributable to AMR in 2060 compared with the baseline scenario.

4. Discussion

This multivariate analysis of data from China in recent years provides important insights into the prevalence of AMR. This is the first study to examine nationally how multivariable factors influence AMR in China. Antibiotic resistance can spread globally, affecting individuals of any age, in any country; however, the environmental drivers of antibiotic resistance remain unclear. There is an urgent need for multivariable analyses to uncover the global drivers of antibiotic resistance. Driven by the One Health concept, the surrounding environment is recognized as a significant vector or source of AMR determinants [29]. Among the multiple factors explored in this study, all three multivariate analyses—the panel fixed-effects analysis, the multivariate forward stepwise regression analysis, and the non-recursive SEM—identified PM_2.5 as making a significant contribution to antibiotic resistance in China. Moreover, the number of confounders in the current study was much greater than that in the literature, improving the reliability of the analysis and reducing the disturbance of unknown factors in the results [3], [19].

The consistent results of the fixed-effects and SEM analyses suggest that the correlation between PM_2.5 air pollution and AMR can be considered robust. Regions with higher air pollution tended to have higher AMR. The SEM revealed that PM_2.5 had direct and indirect effects on antibiotic resistance. Regarding the direct pathway, PM_2.5 is the main vector of AMR, exposing humans through inhalation. Previous study have revealed that antibiotic resistance in various settings (e.g., livestock farming, hospitals, and sewage) can be transmitted to PM_2.5 through aerosol diffusion and dust suspension [14]. The concentration of ARGs in PM_2.5 can reach hundreds of copies per cubic meter, with adults inhaling an estimated 10²–10⁴ copies per day [16]. Concentrations of culturable multidrug-resistant bacteria are reported to be about three times higher in polluted air than in non-polluted air [18]. Highly polluted air contains higher concentrations of ARGs and pathogens, possibly increasing the risk of antibiotic-resistant bacterial infections. Research has indicated that ARGs in environments ranging from fresh snow in pristine ecosystems to heavily human-influenced ecosystems may be intensified by air pollution, significantly raising the health threats associated with AMR [14]. As for the indirect pathway, severe PM_2.5 air pollution was associated with increased incidence of respiratory diseases, indirectly contributing to rising levels of antibiotic resistance, as more drugs are utilized to cure infections caused by air pollution [30].

Air pollution is a critical health issue currently challenging the world. Over 90% of the global population is breathing air that does not fall under healthy World Health Organisation (WHO) guidelines, with developing countries experiencing the highest exposures [31]. Intense human activities exacerbate the environmental transmission of antibiotic resistance—a phenomenon frequently observed in developing countries due to economic growth and population expansion [32]. Our global analysis [19] and this intra-country analysis consistently confirm the link between air pollution and antibiotic resistance. Thus, strategies and tactics to reduce PM_2.5 air pollution are promising for mitigating antibiotic resistance in the future.

Changes in AMR are not related to just one or two factors [5]. Examining multiple factors—including anthropological, socioeconomic, and environmental factors, as well as climatic elements affecting antibiotic resistance—provides a comprehensive view of the drivers of antibiotic resistance. This study demonstrates that PM_2.5, antibiotic usage density, pork production, public toilets per 10 000 people, number of village clinics, and residual income are dominant contributors to antibiotic resistance across China. Antibiotic usage is widely recognized as a promoting factor in antibiotic resistance; the consumption of antibiotics in developing countries is expected to continue increasing in future, likely making these regions hotspots for AMR [33]. Thus, the rational use of antibiotics is still advocated as a long-term effective measure to combat AMR in developing countries. As the demand for pork increases, higher pork production may involve higher antibiotic use, especially as growth promoters or for prophylaxis in pig farming. Animal waste containing higher levels of antibiotics would then be released into the agricultural environment through manure application, which could in turn relate to higher antibiotic resistance [34]. Fortunately, China stopped using antibiotics as growth stimulants in animal husbandry in 2020 [35], [36].

Public services in China, such as public toilets, were established to provide convenient social services; however, inadequate disinfection measures or management of these facilities may exacerbate the spread of antibiotic resistance [37]. Unlike previous global studies [3], [19], the effect of the urban water penetration rate on antibiotic resistance was not found to be significant. The potential reason may be that Chinese people usually drink boiled water, which has a remarkable limiting effect on antibiotic-resistant bacteria transmission from drinking water. However, it should be noted that the urban water penetration rate does not represent the quality of drinking water, and the impact of drinking water quality on antibiotic resistance still requires further exploration.

Higher income was significantly associated with increased AMR, suggesting that wealthier populations may have greater access to antibiotics and, therefore, higher usage [3], [38]. Surprisingly, a greater number of village clinics was associated with increased AMR in China. According to the SEM results, village clinics could help regulate the use of antibiotics in suburban areas, but insufficient sanitation and infection–prevention control may exacerbate the transmission and infection of antibiotic resistance within these clinics, indicating that village clinics could become blind spots for antibiotic resistance control, as discussed in a previous study [39]. Conversely, increasing the proportion of health technicians could help reduce AMR, enhancing the provision of adequate healthcare as more individuals receive attention.

We also found that diet (e.g., meat consumption) affected antibiotic resistance. Previous studies [40], [41] have reported a high-fat diet to be associated with antibiotic tolerance and that more fiber in the diet was linked to lower ARGs in individuals due to the different microbial communities associated with various diets. Using non-recursive SEM, we determined that antibiotic resistance had negative reciprocal effects on antibiotic usage density. The underlying reason is that, once bacteria exhibit high antibiotic resistance, clinicians will no longer prescribe unrestricted antibiotics but will directly use restricted antibiotics or special-use-grade antibiotics [42]. Moreover, compared with previous research [19], the present study further utilized a non-recursive SEM to explore the influences of factors on antibiotic resistance and introduced more confounding predictors to comprehensively analyze the interaction between various factors within the developing country under study. Overall, the drivers of antibiotic resistance are multifactorial; therefore, controlling antibiotic use alone may not be sufficient to control antibiotic resistance, making multiple initiatives necessary to combat this issue. These may include the rational use of antibiotics, reducing environmental pollution, improving sanitation, and enhancing health services.

To address issues of antibiotic resistance and environmental pollution, China has implemented several strategies, such as the 2016–2020 one health national action plan to contain antimicrobial resistance and the air pollution action plan, to limit AMR and air pollution [43], [44]. These measures have proven successful in gradually reducing antibiotic usage, AMR, and air pollution in recent years. Premature death has decreased between 2014 and 2020 across China (Fig. S6(c)). However, the same cannot be said for the welfare impact (Fig. S6(d)), given that rapid economic development and China’s poverty alleviation policies have remarkably affected the GDP per capita in some provinces or regions, particularly in the western part of China [45]. As a result, while premature deaths have decreased in some provinces, there is also a corresponding increase in welfare impacts. The greater correlation effects between PM_2.5 and AMR indicate that controlling PM_2.5 could be one of the most effective ways of reducing antibiotic resistance. Both marginal premature death and welfare loss increased between 2014 and 2020, suggesting that a reduction in PM_2.5 per unit could yield considerable health benefits. From a policy perspective, the effects of marginal premature death and welfare loss are more important than the aggregate effect resulting from air pollution over time.

Air pollution is closely related to the global climate and ecosystem. Many factors, such as burning fossil fuels, cause air pollution in the form of greenhouse gas emissions [46]. Therefore, policies limiting air pollution provide a doubly beneficial solution for reducing the disease burden caused by air pollution and supporting climate change mitigation. In this work, six scenarios based on socioeconomic effects and policies to mitigate air contamination and climate change revealed different paths for the future consequences of antibiotic resistance in China [26], all of which may be essential references for developing countries aiming to control antibiotic resistance. The PM_2.5 concentration under CGs scenario exceeds twice the guideline (10 µg·m⁻³), indicating that the antibiotic resistance burden will continue to increase until 2060 and suggesting that current climate and air pollution measures are insufficient to limit antibiotic resistance—hence the need for stricter policies. The NDCGs scenarios share the same energy and social-economic developments with the CGs scenario but apply stricter clean air policies, with a PM_2.5 concentration of 14.6 µg·m⁻³ projected for 2060. This presents a middle pathway for air pollution and antibiotic resistance control. Moreover, the NGs and AP1.5Gs scenarios offer promising pathways to limit air pollution and antibiotic resistance. While the NGs scenario presents a pathway for carbon-neutral commitment and air-contamination mitigation by 2060, significantly reducing the antibiotic resistance burden, the AP1.5Gs scenario provides the strictest strategies to mitigate climate change and air pollution. These strategies drive the PM_2.5 concentration below the WHO guideline, resulting in the lowest number of premature deaths and amount of welfare loss caused by antibiotic resistance derived from PM_2.5. This study used scenarios applicable to China, which has important guiding significance for future policymaking.

Antibiotic resistance at the animal, human, and environmental interface, driven by multiple factors, is a growing global issue. Our report on antibiotic resistance and its contributors in China over a period of seven years provides an up-to-date comprehensive analysis. These results have significant policy implications for low- and middle-income countries. However, we acknowledge several limitations. First, the data were not uniformly available for all provinces or regions; therefore, it is necessary to further systematize and improve the national microbial resistance monitoring network, and the relative risk should be further measured at various locations and levels of antibiotic resistance. Second, other contributors may also affect antibiotic resistance (e.g., antimicrobial use in animals, drinking water quality, other forms of environmental pollution, culture, and religion) [47]; thus, these data need to be publicly stored in the corresponding database, such as the resistance-monitoring network. Third, research on more antibiotic-resistant pathogens is needed in order to comprehensively present AMR patterns. Fourth, we are aware that correlation does not imply causation. Although there may be confounders that explain the correlations we have observed, biologically plausible mechanisms still need to be identified.

5. Conclusions

This study is the first to comprehensively reveal the relationships between multiple predictors and clinical AMR nationwide in a developing country. Our study contributes to the emerging evidence that PM_2.5 is robustly associated with AMR. We demonstrate that failing to control air pollution may limit efforts to reduce AMR in China. Strategies addressing antibiotic resistance, air pollution, and carbon neutralization may jointly tackle the problem of antibiotic resistance. Although controlling antibiotic consumption is a primary measure to combat antibiotic resistance, simultaneous measures to improve air quality and sanitation, and to regulate access to antibiotics in village clinics and for farm animals must also be implemented. These findings emphasize the need from an environmental perspective to support policies that reduce the antibiotic-resistance burden.

Acknowledgments

This research was funded by the National Natural Science Foundation of China (22406168, W2411031, and 52270201), the China Postdoctoral Science Foundation (2023M733061), and the Zhejiang University Global Partnership Fund (100000-11320/198).

Compliance with ethics guidelines

Zhenchao Zhou, Zejun Lin, Xinyi Shuai, Xiaoliang Ba, Chimo Achi, Mark A. Holmes, Tong Xu, Yingru Lu, Yonghong Xiao, Jianming Xu, Baojing Gu, and Hong Chen declare that they have no conflict of interest or financial conflits to disclose.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.eng.2024.09.013.

References

Publishing order | Descend order by publishing year | Descend order by cited within

[1]	O’Neill J.Tackling drug-resistant infections globally: final report and recommendations.Report. London: Government of the United Kingdom; 2016.

[2]	Murray CJL, Ikuta KS, Sharara F, Swetschinski L, Robles Aguilar G, Gray A, et al.Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis.Lancet 2022; 399(10325):629-655.

[3]	Collignon P, Beggs JJ, Walsh TR, Gandra S, Laxminarayan R.Anthropological and socioeconomic factors contributing to global antimicrobial resistance: a univariate and multivariable analysis.Lancet Planet Health 2018; 2(9):398-405.

[4]	MacFadden DR, McGough SF, Fisman D, Santillana M, Brownstein JS.Antibiotic resistance increases with local temperature.Nat Clim Chang 2018; 8(6):510-514.

[5]	Allel K, Day L, Hamilton A, Lin L, Furuya-Kanamori L, Moore CE, et al.Global antimicrobial-resistance drivers: an ecological country-level study at the human–animal interface.Lancet Planet Health 2023; 7(4):291-303.

[6]	Ikhimiukor OO, Odih EE, Donado-Godoy P, Okeke IN.A bottom-up view of antimicrobial resistance transmission in developing countries.Nat Microbiol 2022; 7(6):757-765.

[7]	Nadimpalli ML, Marks SJ, Montealegre MC, Gilman RH, Pajuelo MJ, Saito M, et al.Urban informal settlements as hotspots of antimicrobial resistance and the need to curb environmental transmission.Nat Microbiol 2020; 5(6):787-795.

[8]	Liu YY, Wang Y, Walsh TR, Yi LX, Zhang R, Spencer J, et al.Emergence of plasmid-mediated colistin resistance mechanism MCR-1 in animals and human beings in China: a microbiological and molecular biological study.Lancet Infect Dis 2016; 16(2):161-168.

[9]	Walsh TR, Weeks J, Livermore DM, Toleman MA.Dissemination of NDM-1 positive bacteria in the New Delhi environment and its implications for human health: an environmental point prevalence study.Lancet Infect Dis 2011; 11(5):355-362.

[10]	Hendriksen RS, Munk P, Njage P, van Bunnik B, McNally L, Lukjancenko O, et al.Global monitoring of antimicrobial resistance based on metagenomics analyses of urban sewage.Nat Commun 2019; 10(1):1124.

[11]	Kahn LH.Perspective: the one-health way.Nature 2017; 543(7647):S47.

[12]	Hernando-Amado S, Coque TM, Baquero F, Martínez JL.Defining and combating antibiotic resistance from One Health and Global Health perspectives.Nat Microbiol 2019; 4(9):1432-1442.

[13]	Zhu YG, Gillings M, Simonet P, Stekel D, Banwart S, Penuelas J.Human dissemination of genes and microorganisms in Earth’s critical zone.Glob Change Biol 2018; 24(4):1488-1499.

[14]	Zhu G, Wang X, Yang T, Su J, Qin Y, Wang S, et al.Air pollution could drive global dissemination of antibiotic resistance genes.ISME J 2021; 15(1):270-281.

[15]	Wu D, Jin L, Xie J, Liu H, Zhao J, Ye D, et al.Inhalable antibiotic resistomes emitted from hospitals: metagenomic insights into bacterial hosts, clinical relevance, and environmental risks.Microbiome 2022; 10(1):19.

[16]	Jin L, Xie J, He T, Wu D, Li X.Airborne transmission as an integral environmental dimension of antimicrobial resistance through the “One Health” lens.Crit Rev Environ Sci Technol 2022; 52(23):4172-4193.

[17]	Sun J, Liao XP, D AW’Souza, Boolchandani M, Li SH, Cheng K, et al.Environmental remodeling of human gut microbiota and antibiotic resistome in livestock farms.Nat Commun 2020; 11(1):1427.

[18]	Mao Y, Ding P, Wang Y, Ding C, Wu L, Zheng P, et al.Comparison of culturable antibiotic-resistant bacteria in polluted and non-polluted air in Beijing, China.Environ Int 2019; 131:104936.

[19]	Zhou Z, Shuai X, Lin Z, Yu X, Ba X, Holmes MA, et al.Association between particulate matter (PM)_2.5 air pollution and clinical antibiotic resistance: a global analysis.Lancet Planet Health 2023; 7(8):649-659.

[20]	Qian H, Xu S, Cao J, Ren F, Wei W, Meng J, et al.Air pollution reduction and climate co-benefits in China’s industries.Nat Sustain 2021; 4(5):417-425.

[21]	National Health Commission of the People’s Republic of China.Status report on antimicrobial administration and antimicrobial resistance in China. Report. Beijing: Peking Union Medical College Press; 2021.

[22]	Southerland VA, Brauer M, Mohegh A, Hammer MS, van Donkelaar A, Martin RV, et al.Global urban temporal trends in fine particulate matter (PM_2.5) and attributable health burdens: estimates from global datasets.Lancet Planet Health 2022; 6(2):139-146.

[23]	Ren C, Zhou X, Wang C, Guo Y, Diao Y, Shen S, et al.Ageing threatens sustainability of smallholder farming in China.Nature 2023; 616(7955):96-103.

[24]

Stanaway JD, Afshin A, Gakidou E, Lim SS, Abate D, Abate KH, et al.Global, regional, and national comparative risk assessment of 84 behavioural, environmental and occupational, and metabolic risks or clusters of risks for 195 countries and territories, 1990–2017: a systematic analysis for the global burden of disease study 2017.Lancet 2018; 392(10159):1923-1994.

[25]	Gu B, Zhang L, Van Dingenen R, Vieno M, Van Grinsven HJ, Zhang X, et al.Abating ammonia is more cost-effective than nitrogen oxides for mitigating PM_2.5 air pollution.Science 2021; 374(6568):758-762.

[26]	Cheng J, Tong D, Zhang Q, Liu Y, Lei Y, Yan G, et al.Pathways of China’s PM_2.5 air quality 2015–2060 in the context of carbon neutrality.Nat Sci Rev 2021; 8(12):nwab078.

[27]	Liu C, Chen R, Sera F, Vicedo-Cabrera AM, Guo Y, Tong S, et al.Ambient particulate air pollution and daily mortality in 652 cities.N Engl J Med 2019; 381(8):705-715.

[28]	Chen J, Liu YJ, Zhang ER, Pan T, Liu YH.Estimating China’s population over 21st century: spatially explicit scenarios consistent with the shared socioeconomic pathways (SSPs).Sustainability 2022; 14(4):2442.

[29]	Shen LY, Wei XL, Yin J, Haley DR, Sun Q, Lundborg CS.Interventions to optimize the use of antibiotics in China: a scoping review of evidence from humans, animals, and the environment from a One Health perspective.One Health 2022; 14:100388.

[30]	Shen Y, Wu Y, Chen G, Van Grinsven HJM, Wang X, Gu B, et al.Non-linear increase of respiratory diseases and their costs under severe air pollution.Environ Pollut 2017; 224:631-637.

[31]	Kirby T.WHO: 92% of the world’s population breathe polluted air.Lancet Respir Med 2016; 4(11):862.

[32]	Larsson DGJ, Flach CF.Antibiotic resistance in the environment.Nat Rev Microbiol 2022; 20(5):257-269.

[33]	Friedrich MJ.Antibiotic consumption increasing globally.JAMA 2018; 319(19):1973.

[34]	Van Boeckel TP, Brower C, Gilbert M, Grenfell BT, Levin SA, Robinson TP, et al.Global trends in antimicrobial use in food animals.Proc Natl Acad Sci USA 2015; 112(18):5649-5654.

[35]	Wang Y, Xu C, Zhang R, Chen Y, Shen Y, Hu F, et al.Changes in colistin resistance and mcr-1 abundance in Escherichia coli of animal and human origins following the ban of colistin-positive additives in China: an epidemiological comparative study.Lancet Infect Dis 2020; 20(10):1161-1171.

[36]	Schoenmakers K.How China is getting its farmers to kick their antibiotics habit.Nature 2020; 586(7830):S60-S62.

[37]	Cave R, Cole J, Mkrtchyan HV.Surveillance and prevalence of antimicrobial resistant bacteria from public settings within urban built environments: challenges and opportunities for hygiene and infection control.Environ Int 2021; 157:106836.

[38]	Barriere SL.Clinical, economic and societal impact of antibiotic resistance.Expert Opin Pharmacother 2015; 16(2):151-153.

[39]	Dyar OJ, Yang D, Yin J, Sun Q, St LCålsby.Variations in antibiotic prescribing among village doctors in a rural region of Shandong Province, China: a cross-sectional analysis of prescriptions.BMJ Open 2020; 10(6):036703.

[40]	Oliver A, Xue Z, Villanueva YT, Durbin-Johnson B, Alkan Z, Taft DH, et al.Association of diet and antimicrobial resistance in healthy U.S. adults.MBio 2022; 13(3):0010122.

[41]	Liu Y, Yang K, Jia Y, Shi J, Tong Z, Fang D, et al.Gut microbiome alterations in high-fat–diet-fed mice are associated with antibiotic tolerance.Nat Microbiol 2021; 6(7):874-884.

[42]	Zhang J, Ma X, Tang L, Tian D, Lin L, Li Y, et al.Pattern of antibiotic prescriptions in Chinese children, a cross-sectional survey from 17 hospitals located across 10 provinces of China.Front Pediatr 2022; 10:857945.

[43]	Xiao Y, Li L.China’s national plan to combat antimicrobial resistance.Lancet Infect Dis 2016; 16(11):1216-1218.

[44]	Yue HB, He CY, Huang QX, Yin D, Bryan BA.Stronger policy required to substantially reduce deaths from PM_2.5 pollution in China.Nat Commun 2020; 11(1):1462.

[45]	Liu MY, Feng XL, Wang SG, Qiu HG.China’s poverty alleviation over the last 40 years: successes and challenges.Aust J Agric Resour Econ 2020; 64(1):209-228.

[46]	Huang RJ, Zhang Y, Bozzetti C, Ho KF, Cao JJ, Han Y, et al.High secondary aerosol contribution to particulate pollution during haze events in China.Nature 2014; 514(7521):218-222.

[47]	Paul RJ, Varghese D.AMR in animal health: issues and one health solutions for LMICs.S. Thomas (Ed.), Antimicrobial resistance: global challenges and future interventions, Springer, Singapore 2020; 135-149.

RIGHTS & PERMISSIONS

THE AUTHOR

PDF (1119KB)

Part of a collection:

Supplementary files

4596

Accesses

Citation

Detail

Sections

Recommended

Browse

Authors & reviewers

About the journal