Metagenomic Meta-Analysis of Antibiotic-Resistance Genes in Wastewater: A Perspective from the COVID-19 Pandemic

Abstract

Antibiotic resistance is a silent global crisis intensified by the recent pandemic of coronavirus disease 2019 (COVID-19). To address this growing threat, wastewater-based surveillance (WBS) is emerging as a promising public health tool for monitoring antibiotic resistance within communities. Our meta-analysis aims to reveal the landscape of antibiotic-resistance genes (ARGs) in global wastewater during and after the COVID-19 pandemic. The analysis included wastewater samples collected between 2020 and 2024 from five countries across three continents: Asia (China), Europe (United Kingdom and Russia), and North America (United States and Canada). Our findings showed higher observed ARGs in Russia and China despite their small sample size, while the USA showed more diverse ARGs. Distinct patterns of ARGs were observed in European and North American wastewater samples (p-value < 0.001). We identified 2483 ARGs, with multidrug-resistant (MDR) genes dominating most regions and accounting for almost 45% of all ARGs detected in Europe. Country-specific indicator ARGs showed 22 unique ARGs for Russia, 3 for each of the UK and Canada, and 2 were specific for China. Continentally, 100 indicator ARGs were specific to Asia, 38 to Europe, and 18 to North America. These findings highlight the regional variations in ARG profiles, emphasizing the urgent need for region-specific strategies to combat antibiotic-resistance threat. Additionally, our study further supports the value of WBS as a valuable public health tool for monitoring antibiotic resistance.

1. Introduction

Antibiotic resistance is a silent global crisis, with an estimated 1.27 million deaths globally in 2019 alone, and is expected to reach 10 million deaths by 2050 [1,2]. In response to this growing concern, the World Health Organization (WHO) established the Global Antimicrobial Resistance Surveillance System (GLASS) in 2015 to facilitate the global sharing of surveillance data [3]. However, the WHO in its report for 2017–2018 highlighted the limitations in GLASS outcomes, including the lack of standardized sampling and reliance on data from individuals seeking medical care [4]. This report further emphasized the need to incorporate epidemiological and population data alongside laboratory information.

Wastewater-based surveillance (WBS) provides a promising approach for monitoring antibiotic resistance across the entire population [1,5]. It dynamically analyzes chemical and biological biomarkers in wastewater to provide real-time monitoring of public health threats [6,7,8]. Recently, in the pandemic of coronavirus disease 2019 (COVID-19), WBS demonstrated a remarkable ability to monitor the presence of the virus and its emerging variants within communities [6,9]. Similarly, WBS offers a valuable approach to understanding the circulation of antibiotic-resistance genes (ARGs) within communities by examining the collective genetic material discharged into wastewater.

Recent meta-analyses have highlighted the prevalence of ARGs in various aquatic environments. Kang et al. revealed a diverse antibiotic resistome in global hospital wastewater, dominated by multidrug, beta-lactam, and aminoglycoside-resistance genes [10]. Liu et al. investigated the contamination of groundwater with antibiotics and ARGs, identifying a significant presence of ARGs in various groundwater sources [11]. This alarming prevalence of ARGs in aquatic environments poses a substantial threat to public health, necessitating urgent measures to prevent the dissemination of antibiotic resistance and protect water resources.

The recent pandemic of COVID-19 has disrupted healthcare systems globally, leading to significant alterations in antibiotic-prescribing practices and healthcare-seeking behaviors [12]. In July 2024, the Centers for Disease Control and Prevention (CDC) reported an increase of 20% in antimicrobial-resistant pathogens related to healthcare settings in the United States during the COVID-19 pandemic compared to the pre-pandemic period [13]. These changes have likely influenced the dynamics of ARG dissemination, requiring an urgent assessment to understand the impact of the COVID-19 pandemic on the ARG landscape.

Therefore, this study aims to conduct a metagenomic meta-analysis of global wastewater samples collected during and after the COVID-19 pandemic to comprehensively characterize the dissemination, diversity, and trends of ARGs. By focusing on untreated wastewater, we aim to capture a better representation of the ARG profile in the community before potential alterations during the wastewater-treatment process. Moreover, this study assesses the utility of wastewater-based metagenomics as a public health surveillance tool for combating the ARG threat.

2. Materials and Methods

2.1. Systematic Review and Metagenomic Data Retrieval

A systematic literature search was conducted using PubMed and Google Scholar databases to identify relevant studies published between January 2020 and August 2024. The search terms included combinations of “metagenome”, “metagenomic”, “wastewater”, “sewage”, “antibiotic resistance”, “antimicrobial resistance”, and “resistome”. Inclusion criteria were limited to English peer-reviewed studies that employed DNA-based metagenomic approaches using the Illumina platform to characterize antimicrobial-resistance genes in untreated, raw wastewater of municipal, urban, and domestic origin collected between 2020 and 2024, with publicly available sequence data. Studies on treated wastewater, sludge, or other wastewater types (e.g., rural, hospital, industrial, pharmaceutical, aquaculture, farm, slaughterhouse, etc.) were excluded.

Metagenomic sequencing data for all wastewater samples included in this study were retrieved from the Sequence Read Archive (SRA) of the National Center for Biotechnology Information (NCBI) using the NCBI SRA Toolkit (3.1.1) [14] and the Genome Sequence Archive (GSA) of the China National Center for Bioinformation.

2.2. Metagenomic Sequences Pre-Processing

Raw metagenomic reads were initially assessed for quality using the Fast Quality Control (FastQC) tool (0.12.0) [15]. Subsequently, quality trimming and filtering were performed using FastQ Quality Control Software (FaQCs) (2.10) [16]. This process involved removing adapter sequences, which are short DNA sequences added to the ends of sequencing reads during library preparation, and filtering out low-quality bases with a quality score below 30.

2.3. Identification and Abundance Estimation of ARGs

For the ARG identification and abundance estimation, the high-quality reads were aligned against the Comprehensive Antibiotic Resistance Database (CARD) (3.2.9) [17] using the BBMap tool (35.85) [18]. The reads were filtered based on alignment length ($\ge$35 amino acids), sequence similarity ($\ge$90$\%$), and E-value ($\le$1 $\times$ 10⁻⁵), as described previously [19]. To address the variations in sequencing depth, we applied data normalization using Reads Per Kilobases per Million (RPKM).

2.4. Statistical Analysis

To analyze the diversity of ARGs within each sample, alpha diversity was employed using the Observed, Shannon, and Simpson indices using the microbiome (1.24.0) R package [20]. Moreover, Principal Coordinate Analysis (PCoA) was used to describe the beta diversity of ARGs between different samples by implementing the Bray–Curtis and Jaccard indices using the vegan (2.6.6.1) R package [21]. The permutational multivariate analysis of variance (PERMANOVA) was applied to determine the significance of PCoA distance separation between samples, with a value of R² $>$ 0 and a p-value $<$ 0.001 considered statistically significant. To identify indicator ARGs for countries and continents, we employed the multipatt function from the indicspecies (1.7.15) R package [22] and stats (4.3.2) [23] R package, utilizing 999 permutations. ARGs with a p-value $<$ 0.05 were considered significant indicators. The abundance of identified ARGs and indicator ARGs across different countries and continents was represented as PRKM using the ggplot2 (3.5.1) R package [24].

3. Results

3.1. Retrieved Metagenomic Data

A total of 217 metagenomic sequence datasets were initially identified through a comprehensive literature search. However, 4 samples were excluded due to errors with the associated FastQ files. This resulted in a final dataset of 213 metagenomic sequences retrieved from 10 studies [25,26,27,28,29,30,31,32,33,34]. The datasets represent wastewater samples collected from 5 countries across 3 continents: China (n $=$ 9) from Asia, the United Kingdom (n $=$ 8) and Russia (n $=$ 1) from Europe, and the United States (n $=$ 194) and Canada (n $=$ 1) from North America (Figure 1). Detailed information, including accession numbers, for included datasets is provided in Supplementary Table S1.

3.2. Diversity of ARGs in Global Wastewater

Alpha diversity analysis revealed that Russia and China had the highest observed ARGs (1118 and 523.6 $\pm$ 335.64, respectively) compared to other countries (UK: 395.63 $\pm$ 78.54, USA: 376.46 $\pm$ 74.65, and Canada: 195), although their sample sizes were smaller. At the Shannon and Simpson indices, Russia displayed a more diverse and even distribution of ARGs (4.95 and 64.07, respectively), followed by the USA (4.64 $\pm$ 0.27 and 49.96 $\pm$ 20.9), China (4.42 $\pm$ 0.6 and 44.06 $\pm$ 23.53), UK (4.22 $\pm$ 0.53 and 38.5 $\pm$ 17.9), and Canada (2.88 and 6.78). Continentally, Asia exhibited the highest observed ARGs (523.6 $\pm$ 335.64), followed by Europe (475.8 $\pm$ 251.75) and North America (375.53 $\pm$ 75.58). In contrast, samples from North America were more diverse and even distributed in ARGs at the Shannon and Simpson indices (4.63 $\pm$ 0.3 and 49.74 $\pm$ 21.1, respectively) compared to Asia (4.42 $\pm$ 0.6 and 44.1 $\pm$ 23.54) and Europe (4.3 $\pm$ 0.55 and 41.5 $\pm$ 17.89). Due to uneven sample sizes, statistical comparisons between countries were not feasible. The alpha diversity of ARGs is demonstrated in Figure 2A across countries and continents.

Beta-diversity analysis, using the Bray–Curtis index, revealed a clear separation between European and North American samples (R² $=$ 0.997). In contrast, samples from different countries exhibited less distinct clustering (R² $=$ 0.132). The Jaccard index, however, did not show a clear separation between samples at either the country or continent level (R² $=$ 0.06 and 0.043, respectively). Despite these findings, the observed dissimilarities between samples were statistically significant (p-value $<$ 0.001, PERMANOVA). Figure 2B demonstrates the beta diversity of ARGs across different countries and continents.

3.3. Abundance of ARGs in Global Wastewater

A total of 2483 ARGs were identified across 213 metagenomic samples, categorized into 23 ARG classes (Figure 3). Of these ARGs, the resistance genes of msrE, mphE, and qacL were the most abundant across all samples (6.2$\%$, 5.03$\%$, and 4.13$\%$, respectively), followed by APH(3″)-Ib (2.77$\%$), sul1 (2.4$\%$), APH(6)-Id (2.12$\%$), tet(Q) (2.08$\%$), mef(C) (2.08$\%$), and mphG (2$\%$). However, among the identified ARGs 19 were detected in all samples, including sul1, sul2, qacEdelta1, qacL, APH(6)-Id, Bado_rpoB_RIF, MexB, MexF, MexK, MexI, MuxB, MuxC, AAC(6′)-Ib7, rpoB2, Bbif_ileS_MUP, smeE, AxyY, oqxB, and ceoB. The abundance of ARGs identified across different countries is presented in Supplementary Table S2.

By grouping ARGs into antibiotic classes, multidrug-resistant (MDR) genes were the predominant (28.1$\%$), followed by the resistance genes related to macrolide–lincomycin–streptogramin (MLS) (16$\%$), aminoglycoside (10.64$\%$), tetracycline (8.14$\%$), carbapenem (7.24$\%$), disinfecting (6.41$\%$), sulfonamide (5.31$\%$), beta-lactam (2.86$\%$), and fluoroquinolone (2.5$\%$). Interestingly, MDR genes were the dominant resistance genes in all countries except Russia, where the resistance genes related to MLS were the most prevalent (21.85$\%$), followed by aminoglycoside (14.01$\%$) and MDR (12.11$\%$). Moreover, carbapenem-resistance genes were more prevalent in Russia (35.67$\%$) and the UK (23.1$\%$), while sulfonamide-resistance genes were abundant in China (42.57$\%$) and Russia (38.85$\%$). In contrast, MLS, aminoglycoside, and tetracycline-resistance genes were commonly observed in most samples.

At the continental level, MDR genes were the most abundant (27.6$\%$), followed by the resistance genes related to MLS (17.02$\%$), aminoglycoside (12$\%$), tetracycline (7.8$\%$), disinfecting (7.68$\%$), carbapenem (7.48$\%$), and sulfonamide (5.45$\%$). However, MDR genes accounted for 44.74$\%$ of all ARGs detected in Europe. Notably, the resistance genes related to MLS and aminoglycoside were more prevalent in North America (23.24$\%$ and 15.83$\%$, respectively) and Asia (16.24$\%$ and 11.7$\%$, respectively) than in Europe (10$\%$ and 6.7$\%$, respectively). The abundance of ARGs identified across different continents is presented in Supplementary Table S3.

3.4. Country-Specific Indicator ARGs in Global Wastewater

The indicator ARGs, which are unique ARGs to a particular country or continent based on their occurrence and abundance, were identified. A total of 113 ARGs were detected as country-specific indicators (p-value $<$ 0.05) (Figure 4A; Supplementary Table S4). Russia exhibited a higher number of unique ARGs (n $=$ 22) compared to Canada (n $=$ 3), the UK (n $=$ 3), and China (n $=$ 2). Additionally, Russia shared a higher number of ARGs with China (n $=$ 20) compared to other countries: Russia with Canada (n $=$ 11), Russia with the USA (n $=$ 8), and Russia with the UK (n $=$ 6), while there were no shared ARGs with other country pairs. However, these results should be validated with additional samples due to the limited sample size. The indicator ARG distribution across countries is presented in

Table A1. Unique indicator ARGs across countries.

Russia		Canada		UK		China
ARG	p-Value *	ARG	p-Value *	ARG	p-Value *	ARG	p-Value *
MCR-5.1	0.009	CrcB	0.011	Rm3	0.006	erm (46)	0.039
VEB-14	0.012	OprA	0.015	THIN-B	0.007	tet (42)	0.047
MCR-5.2	0.013	CMY-157	0.046	ParS	0.021
Ccol_ACT_CHL	0.014
VEB-5	0.017
aadA4	0.017
cmlA5	0.018
BES-1	0.019
AAC6_IB_HZ	0.02
OXA-836	0.02
dfrA13	0.027
linG	0.029
VEB-1	0.03
CMY-42	0.03
lnuF	0.031
dfrA7	0.031
OXA-928	0.037
OXA-921	0.04
PAC-1	0.043
VEB-9	0.044
ANT(2′′)-Ia	0.045
VEB-7	0.046

* p-value $<$ 0.05 considered statistically significant.

for unique indicators and

Table A2. Shared indicator ARGs across countries.

China and Russia		Russia and Canada		Russia and USA		Russia and UK
ARG	p-Value *	ARG	p-Value *	ARG	p-Value *	ARG	p-Value *
VEB-3	0.001	AAC(2’)-Ic	0.009	OXA-504	0.002	mexP	0.003
dfrA1	0.009	mef(B)	0.009	OXA-780	0.007	OXA-118	0.008
aadA16	0.011	Erm(38)	0.01	AxyX	0.01	ESP-1	0.012
arr-3	0.015	tap	0.01	MOX-2	0.024	TriB	0.018
pp-flo	0.015	efpA	0.01	CMY-48	0.026	OXA-198	0.033
AAC(6′)-Ib9	0.017	QnrVC4	0.014	CepS	0.03	OXA-20	0.044
dfrA27	0.019	OXA-56	0.015	MOX-13	0.033
tet(59)	0.02	Rv2856	0.036	OXA-724	0.035
EreB	0.024	OXA-912	0.036
AAC(6′)-31	0.026	OXA-7	0.04
dfrA17	0.026	FosA8	0.043
CARB-12	0.029
OXA-21	0.03
AAC(6′)-Ib	0.031
EreA	0.033
tet(33)	0.038
sul4	0.038
tet(L)	0.039
AAC(6′)-IIa	0.044
dfrA14	0.045

* p-value $<$ 0.05 considered statistically significant.

for shared indicators.

When indicator ARGs were grouped by antibiotic class, distinct patterns emerged across countries (Figure 4B). In Canada, the resistance genes related to rifamycin (54.03%) and MDR (33.56%) were dominant indicators. China exhibited a higher prevalence of aminoglycoside- (21.89%) and phenicol (17.94%)-resistance genes. Russia demonstrated a prevalence of resistance genes related to MLS (22.38%) and aminoglycoside (19.37%). Remarkably, the UK showed a higher prevalence in carbapenem-resistance genes (43.62%), followed by disinfecting (23.38%) and MDR (21.96%). In the USA, resistant genes related to MLS (51.72%) and carbapenem (17.34%) were predominant indicators. Notably, carbapenem-resistance genes were prevalent in all countries except Canada, where they accounted for only 0.68% of identified indicator ARGs. In contrast, carbapenem-resistance genes represented a substantial proportion of indicator ARGs in the UK (43.62%), USA (17.34%), Russia (14.1%), and China (7.86%). The country-specific indicator ARGs grouped by antibiotic class are presented in Supplementary Table S5.

3.5. Continent-Specific Indicator ARGs in Global Wastewater

A total of 293 continent-specific indicator ARGs were detected (Supplementary Table S6). Asia exhibited the highest number of indicator ARGs (n = 100), followed by Europe (n = 38) and North America (n = 18). Additionally, 113 ARGs were shared between Asia and Europe, 11 between Asia and North America, and 13 between Europe and North America. The indicator ARG distribution across continents is presented in

Table A3. Unique indicator ARGs across continents.

Asia		Europe		North America
ARG	p-Value *	ARG	p-Value *	ARG	p-Value *
AAC(6′)-Ib’	0.001	LAQ-1	0.001	AAC(6′)-Ic	0.019
AAC(6′)-Ii	0.001	LRA-1	0.001	CMY-79	0.001
AAC(6′)-31	0.001	FOX-9	0.003	CepS	0.001
AAC(6′)-Ib8	0.001	OXA-20	0.001	CMY-48	0.001
AAC6_30_AAC6_Ib	0.001	OXA-228	0.001	SRT-2	0.033
APH(2′′)-If	0.001	OXA-257	0.001	CMY-105	0.039
aphA15	0.001	CPS-1	0.001	CMY-41	0.042
AAC_6_IB_Su	0.001	OXA-669	0.001	OXA-724	0.001
APH(3′)-IIa	0.001	OXA-274	0.001	AMZ-1	0.017
AAC(6′)-IIa	0.001	OXA-37	0.001	cphA2	0.037
AAC_3Ib_AAC_6Ib	0.001	Rm3	0.001	QnrB47	0.005
AAC(6′)-Ib9	0.001	OXA-198	0.001	AxyX	0.001
AAC6_Ie_APH2_Ia	0.001	THIN-B	0.001	OprZ	0.001
aadA16	0.001	OXA-667	0.001	golS	0.033
AAC(6′)-IIc	0.002	OXA-355	0.001	MCR-3.4	0.034
ANT(3′′)-IIa	0.002	OXA-229	0.001	MCR-3.6	0.036
aad(6)	0.004	SGM-6	0.001	tet(X1)	0.001
ANT3II_ANT6II	0.006	OXA-668	0.001	tet(41)	0.014
CrcB	0.009	ESP-1	0.001
APH(6)-Ic	0.011	OXA-5	0.001
SCO-1	0.001	OXA-118	0.001
TEM-72	0.001	OXA-119	0.001
TEM-116	0.001	LEN-9	0.005
CTX-M-101	0.002	OXA-209	0.006
CTX-M-42	0.002	PDC-9	0.009
IreK	0.002	PDC-133	0.012
CTX-M-130	0.003	OXA-513	0.026
VEB-3	0.005	PDC-62	0.026
CTX-M-155	0.008	TriB	0.001
EC-14	0.019	TriA	0.001
TEM-183	0.033	QnrB8	0.012
EC-15	0.046	cfrC	0.001
CARB-12	0.001	ParS	0.001
GES-44	0.001	mexP	0.001
OXA-45	0.001	ParR	0.001
OXA-3	0.001	opmE	0.002
OXA-21	0.001	rosB	0.009
OXA-417	0.002	tet(30)	0.001
OXA-1	0.002
GES-12	0.002
PER-4	0.003
OXA-320	0.003
OXA-496	0.005
OXA-96	0.005
RAD-1	0.006
OXA-926	0.006
ACT-34	0.007
OXA-282	0.014
GES-3	0.014
ACT-25	0.022
TEM-102	0.022
TEM-198	0.027
LAP-2	0.031
dfrB4	0.001
dfrA16	0.001
dfrA27	0.001
dfrA17	0.001
dfrA1	0.001
QnrD1	0.001
QnrS8	0.001
QnrS1	0.003
TLA-2	0.001
lsaA	0.001
efrB	0.001
efrA	0.002
Abau_AmvA	0.006
AAC(6’)-Ib-cr1	0.011
Erm(51)	0.001
ErmC	0.001
erm(46)	0.001
lnuA	0.001
Erm(47)	0.001
ErmT	0.001
msrC	0.001
mef(F)	0.001
msr(G)	0.002
LnuP	0.005
EreA	0.008
ErmX	0.009
EreB	0.012
Erm(52)	0.013
msrF	0.018
ErmQ	0.022
SAT-4	0.02
catB2	0.001
pp-flo	0.001
cmlA4	0.005
catQ	0.006
Abau_AbaF	0.002
arr-3	0.001
sul3	0.001
sul4	0.001
tet(K)	0.001
tet(59)	0.001
tet(Z)	0.001
tet(42)	0.001
tet(33)	0.001
tet(43)	0.001
tet(L)	0.001
tet(36)	0.002

* p-value $<$ 0.05 considered statistically significant.

for unique indicators and

Table A4. Shared indicator ARGs across continents.

Asia and Europe		Europe and North America		Asia and North America
ARG	p-Value *	ARG	p-Value *	ARG	p-Value *
novA	0.001	aadA7	0.001	APH(9)-Ic	0.002
aadA4	0.001	OXA-780	0.001	CMY-114	0.001
AAC(3)-IIe	0.001	OXA-504	0.001	CfxA3	0.03
AAC(6’)-Ib	0.001	MOX-13	0.006	CMY-116	0.031
APH(3’)-Ib	0.003	OXA-726	0.01	MIR-2	0.043
AAC(3)-Ia	0.004	imiH	0.03	adeF	0.039
aadA27	0.004	OXA-34	0.034	mphF	0.001
aadA15	0.014	OXA-681	0.04	MCR-3.3	0.014
ANT(9)-Ia	0.016	QnrB19	0.026	Ecol_catII	0.033
RanA	0.016	MuxA	0.001	tet(D)	0.001
APH(3’)-VIa	0.018	MexV	0.011	tet(B)	0.001
aadA3	0.024	MCR-9.1	0.001
AAC(3)-IIb	0.028	MCR-3.17	0.041
ANT(6)-Ib	0.037
VEB-7	0.001
PJM-1	0.001
AIM-1	0.001
VEB-5	0.002
RAHN-1	0.011
CTX-M-88	0.013
VEB-9	0.017
BEL-1	0.017
FOX-3	0.019
VEB-1	0.024
VEB-14	0.025
LCR-1	0.031
SGM-1	0.001
OXA-296	0.001
JOHN-1	0.001
CARB-14	0.001
OXA-420	0.001
OXA-47	0.001
OXA-4	0.001
CARB-5	0.001
RCP-1	0.001
OXA-129	0.001
OXA-31	0.001
OXA-392	0.002
OXA-134	0.002
OXA-58	0.003
BKC-1	0.003
OXA-275	0.004
OXA-333	0.004
OXA-164	0.006
OXA-9	0.008
SHV-18	0.011
SHV-24	0.011
OXA-650	0.011
GES-14	0.012
blaF	0.015
CGA-1	0.016
AER-1	0.02
GES-17	0.029
ORN-1	0.041
OXA-651	0.044
OXA-727	0.046
dfrB10	0.001
dfrA14	0.003
dfrA7	0.005
QepA4	0.001
QepA1	0.001
QepA2	0.002
Abau_AbaQ	0.002
lfrA	0.002
qnrE1	0.019
adeN	0.001
aadT	0.002
Rv2856	0.004
abeM	0.004
EstT	0.027
AAC(6’)-Ib-cr3	0.033
Erm(42)	0.001
lnuF	0.001
lnuG	0.001
oleC	0.001
linG	0.001
msr(I)	0.001
mef(J)	0.001
lmrD	0.001
lnuB	0.004
lsaE	0.008
vatB	0.011
msrA	0.027
Erm(38)	0.03
lnuD	0.039
vanS_in_vanO_cl	0.001
vanR_in_vanO_cl	0.001
LpsB	0.001
ICR-Mo	0.007
vanW_in_vanG_cl	0.031
cmlA1	0.001
cmx	0.001
floR	0.001
cmlB1	0.004
Ccol_ACT_CHL	0.008
catP	0.028
catB11	0.047
cmlA5	0.048
FosXCC	0.003
Nfar_rox	0.001
Sven_rox	0.001
rphA	0.001
HelR	0.001
rphB	0.004
tet(Y)	0.001
otr(A)S.rim	0.001
tet(X6)	0.001
tet(H)	0.001
tet(S)	0.001
tet(X5)	0.001
tet(V)	0.001
tap	0.004
tetA(p)	0.015

* p-value $<$ 0.05 considered statistically significant.

for shared indicators.

When indicator ARGs were grouped by antibiotic class, carbapenem-resistance genes were more prevalent in Europe (56.8%) and North America (25.4%), while aminoglycoside and phenicol-resistance genes were more abundant in Asia (21.2% and 16.5%, respectively). Overall, Asia demonstrated a higher abundance of indicator ARGs compared to Europe and North America, as shown in Figure 4C. However, due to uneven sample sizes, further studies with uniform sampling are needed to strengthen these findings. The abundance of continent-specific indicator ARGs grouped by antibiotic class is presented in Supplementary Table S7.

4. Discussion

Wastewater-based surveillance (WBS) emerged in the 1990s as a valuable tool for tracking infection outbreaks and pathogens circulation within communities [35]. A few years later, WBS expanded its applications to monitor illicit drug use, pharmaceutical consumption, pesticide application, and even community dietary habits [1,5,35,36]. Given the potential of WBS for monitoring public health threats, we investigated the prevalence and diversity of ARGs in global wastewater and the potential impact of the COVID-19 pandemic on ARG trends.

Our diversity analysis of ARGs in 213 samples from five countries revealed notable variations in observed genes and their richness across countries and continents. Russia and China exhibited the highest observed ARGs, suggesting that excessive or improper antibiotic usage may contribute to increased diversity of resistance genes in these countries. China, in particular, is the country that showed a rapid increase in antibiotic resistance and was reported in 2010 as the second-largest consumer of antibiotics worldwide [37,38]. In addition to its high number of observed ARGs, Russia displayed a more diverse and even distribution of resistance genes than the other countries. A study by Zakharenkov et al. reported a relatively low consumption of antibiotics in Russia from 2017 to 2019, followed by a sharp increase in 2020, likely due to the heavy use of antibiotics during the COVID-19 pandemic [39].

Continently, Asia (represented by China) displayed the highest observed ARGs, followed by Europe and North America, reflecting the pattern observed at the country level, where China (Asia) and Russia (Europe) have the highest observed ARGs. These findings suggest a potential association between the fate of ARGs and population density, possibly attributed to anthropogenic activities in densely populated regions like Russia and China [40,41]. However, North American samples demonstrated a more diverse and even distribution of ARGs, potentially influenced by its larger sample size compared to Asia and Europe.

Dissimilarity analysis of ARG composition revealed distinct clustering between European and North American samples, indicating significant differences in ARG profiles across continents. In alignment with previous studies [34,42,43,44], these differences may be attributed to variations in antibiotic usage, healthcare practices, regulatory policies, environmental conditions, and socioeconomic factors. However, clustering between countries showed a weaker separation, suggesting more homogenous ARG profiles within countries, potentially influenced by factors like population movements and shared environmental conditions.

Our abundance analysis detected 2483 ARGs categorized under 23 antibiotic classes. A core set of 19 ARGs was consistently detected across all sampling sites, including sul1, sul2, qacEdelta1, qacL, APH(6)-Id, Bado_rpoB_RIF, MexB, MexF, MexK, MexI, MuxB, MuxC, AAC(6′)-Ib7, rpoB2, Bbif_ileS_MUP, smeE, AxyY, oqxB, and ceoB. The persistence of these genes across the sampled regions might reflect a shared pattern of antibiotic consumption. While these genes were universally found in our study, only three (sul1, sul2, and APH(6)-Id) were among the 30 most prevalent ARGs reported by Munk et al. in their global wastewater metagenomic study conducted between 2016 and 2019 [45]. The discrepancy between our findings and those of Munk et al. may be attributed to several factors, including differences in sampling regions, sampling timeframe, and, most importantly, the impact of the COVID-19 pandemic on antibiotic usage and healthcare practices.

The observed distribution of resistance genes to certain antibiotic classes across different countries was closely aligned with the continental pattern, with a predominance of MDR genes, highlighting the growing global threat of antibiotic resistance. This trend is particularly alarming in Europe, where nearly 45% of all detected ARGs were MDR. The highly probable spread of bacteria that have developed resistance to multiple antibiotics poses a significant threat to global public health due to limited therapeutic options, leading to increased risk of morbidity and mortality [46]. Infections caused by pathogens, such as MDR Klebsiella pneumoniae, Acinetobacter baumannii, and Pseudomonas aeruginosa, can be challenging to treat with available antibiotics and, in some cases, may be resistant to all known antibiotics [47]. Addressing the MDR threat requires intense efforts and comprehensive strategies that involve optimizing antibiotic use, strengthening infection prevention and control measures, and developing novel therapeutic options.

While MDR genes were prevalent in Europe, resistance genes associated with MLS and aminoglycosides were more abundant in North America and Asia. Despite the known adverse effects of aminoglycosides and their restricted usage in clinical practices [48], the pandemic may have contributed to increased consumption and self-medication of these antibiotics, leading to their misuse. However, our findings contrast with those reported by Deshpande et al., who observed a negative correlation between COVID-19 cases and aminoglycoside-/sulfonamide-resistance genes [49]. On the other hand, MLS antibiotics are widely used for treating bacterial infections, and their consumption was increased during the pandemic to treat secondary infections associated with COVID-19 [50,51]. A longitudinal study from 1997 to 2017 observed seasonal variations in MLS consumption in many countries, suggesting potential inappropriate prescribing [52].

The resistance genes to carbapenems, a last-resort antibiotic for treating severe bacterial infections, especially those caused by multidrug-resistant bacteria [53], were notably higher in Russia and the UK compared to other countries. The growing prevalence of carbapenem-resistance genes across European countries is alarming due to the potential for increased rate of treatment failures, limited therapeutic options, increased healthcare costs, and higher mortality rates [53,54,55]. A recent systematic review of literature from Eastern European countries revealed a significant prevalence of carbapenem-resistant Gram-negative bacteria [55], highlighting the urgent need for effective treatment strategies and infection-control measures to combat this growing public health threat.

In China and Russia, sulfonamide-resistance genes were more prevalent compared to other countries. According to available data, sulfonamides accounted for 12% of total antibiotic consumption in China, both for human and livestock use [56]. Despite the historical widespread of sulfonamides, their use in human medicine has drastically decreased since 1980 due to increasing bacterial resistance [46,57]. However, they remain widely utilized in veterinary medicine [57]. The higher prevalence of sulfonamide-resistance genes in these countries could be attributed to the indiscriminate use of these antibiotics to treat secondary bacterial infections associated with COVID-19 symptoms, potentially contributing to the increase in resistance genes.

Our analysis of indicator ARGs revealed a geographically diverse landscape of antibiotic resistance. We identified 113 ARGs with varying prevalence across different countries, and 293 ARGs were specific to particular continents. These regional disparities likely reflect differences in antibiotic usage patterns, healthcare practices, and regulatory frameworks. To effectively address this global challenge, we need targeted interventions that consider the specific factors driving resistance in each region. This may include implementing tailored antibiotic stewardship programs, improving infection-control practices, and promoting the responsible use of antibiotics in agriculture. Additionally, identifying specific ARGs in domestic wastewater can serve as a valuable indicator of antibiotic usage patterns and the frequency of antibiotic use within the community.

Public health surveillance based on wastewater offers a promising tool, enabling early detection of outbreaks and broader pathogen monitoring [7,8,36]. Unlike the current surveillance systems, WBS anonymously analyzes entire communities, providing real-time insights into emerging threats [7,8,35,36]. Furthermore, current antibiotic-resistance-surveillance systems typically rely on data generated through clinical microbiology laboratories, limiting most data to phenotypic results for specific pathogens, not including carriers or healthy individuals [5,35]. Although clinical surveillance will remain fundamental to infectious disease response, WBS can provide a complementary tool for antibiotic-resistance-monitoring systems [1,35]. Moreover, the expansion of metagenomic wastewater sequencing efforts enables broad pathogen detection and genomic characterization [5].

Our meta-analysis contributes to a growing body of evidence supporting the potential of WBS as a valuable tool for antibiotic-resistance monitoring. Despite limitations in sample size, particularly in Russia and China, our findings provide valuable insights into the global landscape of ARGs in wastewater. While these countries have shown unique ARG profiles, the smaller sample sizes and the heterogeneity among studies may affect the generalizability of our findings. Future research should address these limitations by expanding the timeframe of literature to include pre-pandemic data to gain a more comprehensive understanding of ARG trends before and after the COVID-19 pandemic and explore additional factors that may impact ARG dynamics.

5. Conclusions

Our metagenomic meta-analysis provides an overview of the global landscape of ARGs in wastewater during and after the COVID-19 pandemic. We identified regional variations in ARG profiles, with distinct patterns observed between Europe, North America, and Asia. These findings underscore the urgent need for region-specific strategies to combat the global antibiotic-resistance threat. By identifying region-specific indicator ARGs, we can help develop tailored strategies, such as targeted antibiotic stewardship programs, improved infection-control measures, and responsible antibiotic use in agriculture. While the long-term impact of COVID-19 on ARG profiles requires further investigation, our findings suggest potential consequences.

Our study highlights the value of WBS as a powerful tool for monitoring the emergence and dissemination of antibiotic resistance. By employing WBS, we can track emerging resistance trends, inform public health policies, and guide targeted interventions to mitigate the threat of antibiotic resistance.