Bathing in resistance: evidence of multidrug-resistant bacteria in recreational coastal waters in Niterói City, Rio de Janeiro, Brazil

Antibiotic resistance is a critical One Health issue, necessitating continuous investigation of bacterial diversity and resistance, particularly in non-clinical environments. This study investigated the presence of antimicrobial-resistant bacteria in waters collected from four sampling points of a recreational beach in the city of Niterói, Rio de Janeiro state. Water samples were collected from four points along the beach, and after serial dilutions, the samples were plated onto selective media for bacterial isolation. Bacterial identification was performed using matrix-assisted laser desorption ionization–time-of-flight (MALDI–TOF) mass spectrometry. Antimicrobial susceptibility was assessed using the disk diffusion method and interpreted according to Clinical & Laboratory Standards Institute (CLSI) standards. DNA from Gram-negative bacteria with detected phenotypic resistance was extracted, and PCR was used to investigate the presence of β-lactamase resistance genes. The bla GES gene was sequenced and deposited in GenBank. The multiple antibiotic resistance phenotype (MARP) and multiple antibiotic resistance index (MARI) were calculated based on the number of antibiotics each isolate was resistant to. The presence of the genes bla GES, bla SHV, bla TEM, and bla CTX-M, which confer resistance to clinically relevant β-lactams, was detected in four members of the order Enterobacterales. A bla GESS determinant was detected in an isolate of Citrobacter freundii in combination with phenotypic resistance to ciprofloxacin, erythromycin, sulfamethoxazole/trimethoprim, and tetracycline, characterizing a multidrug-resistant phenotype. This observation, when combined with findings from previous surveys of Guanabara Bay, indicated that the bla GES determinants are either established or continually introduced into this important ecosystem. The multiple antibiotic resistance indices ranged from 0.19 to 0.62, with a strain of Klebsiella variicola showing the highest value, based on resistance to 11 of the 16 antimicrobials examined. The data presented herein demonstrated the presence of an abundant diversity of antimicrobial-resistant bacteria in waters used for bathing, sailing, and fishing by a large population of humans and emphasizes the requirement for enhanced monitoring of recreational waters along the Guanabara Bay coastline.

Anthropogenic activities have contributed to the establishment of selective pressures and stimulated the dissemination of antimicrobial resistance in diverse environmental compartments, including communities and healthcare settings, making the treatment of infectious diseases increasingly difficult (Centers for Disease Control and Prevention (CDC), 2023). Environments such as hospitals, where antimicrobials and disinfectants are continuously used and which exert a continuous selective pressure for the emergence and dissemination of antimicrobial resistance, have contributed to an increase in the frequency of antimicrobial-resistant bacteria. These bacteria are released and subsequently concentrated within hospital sewage, where the opportunities for genetic interactions among them and environmental bacteria are manifold and, from there, enter the municipal sewage system, which in turn, may be released into water bodies, including rivers, lakes, and oceans, often without prior treatment (Davies and Davies, 2010). As such, monitoring antimicrobial resistance in aquatic environments is crucial in order to understand the epidemiology of specific clones, how resistance spreads and/or is maintained, and how animal and human exposure to resistant bacteria affects their health (Abdelrazik and El-Hadidi, 2023). Consequently, results from surveillance may contribute to future studies concerning the ecological and evolutionary processes of resistant microorganisms, including transformations caused by climate change. Evidently, any future assessment can only be made based on the data we present today.

Globally, the contamination of coastal and estuarine waters by domestic sewage, urban runoff, and hospital effluents has created aquatic environments heavily burdened with fecal indicator bacteria, pathogenic microorganisms, and antimicrobial-resistant (AMR) strains. Recent reviews have highlighted that ambient recreational waters act as both recipients and reservoirs for AMR bacteria and resistance genes, representing an important route of exposure for bathers and water-sport users (Nappier et al., 2020). In temperate coastal settings, ingestion of small volumes of seawater during recreation has been shown to result in measurable exposure to antibiotic-resistant Escherichia coli, including third-generation cephalosporin-resistant strains (Leonard et al., 2015). In Brazil, Guanabara Bay and adjacent coastal systems illustrate the combined impacts of chronic organic pollution and AMR: decades of untreated sewage discharge have led to eutrophication, high levels of fecal contamination, and the emergence of clinically relevant pathogens, including carbapenemase-producing Enterobacterales and other opportunistic bacteria (Fistarol et al., 2015). This scenario reinforces that recreational beaches located in highly urbanized coastal zones, such as Icaraí, are embedded in a broader One Health context in which contaminated surface waters may act as reservoirs and dissemination routes for AMR bacteria between human populations, aquatic ecosystems, and the built environment (Fistarol et al., 2015; Nappier et al., 2020; Martins et al., 2025).

In this context, a recreational beach located in Niterói city (Rio de Janeiro, Brazil) was used as a model in the present study. Icaraí Beach is a popular location for the practice of sports and recreation, but it is surrounded by wastewater treatment plants (Costa et al., 2023) and is part of Guanabara Bay, a very impacted and heterogeneous environment (Fistarol et al., 2015), and as such, those dissimilarities should be taken into consideration when planning and executing assessments and monitoring of water quality (Fistarol et al., 2015). Although previous publications have reported Enterobacterales, extended-spectrum β-lactamases (ESBLs), and even Guiana extended-spectrum (GES)-type carbapenemases in parts of Guanabara Bay, no study to date has specifically characterized the presence of multidrug-resistant (MDR) bacteria and resistance determinants at Icaraí Beach, one of the most densely frequented recreational beaches in Niterói and a critical interface for human–environment exposure. Given the environmental heterogeneity of Guanabara Bay, the diversity of pollution gradients, and the proximity of wastewater treatment plants to this beach, site-specific assessments are essential and cannot be inferred from data generated at other locations. The aim of this work was to determine the presence of bacteria carrying genetic determinants of antibiotic resistance and establish phenotypic antibiotic signatures among isolates collected from surface waters from four different points at the Icaraí Beach. By generating point-specific and gene-level data for this site, our study expands the current understanding of AMR distribution within Guanabara Bay and provides an initial reference for future temporal or longitudinal monitoring programs.

Sample collection and bacterial isolation from recreational surface waters

A single water sample was collected from different points on Icaraí Beach (P1: −22,9044936, −43,1180354; P2: −22,9042907, −43,1179717; P3: −22,9051329; P4: −22,9127509, −43,1100193) located in the city of Niterói, metropolitan region of Rio de Janeiro, on September 1, 2021. The four sampling points were selected to represent different types of environmental influence along Icaraí Beach. Point 1 (P1) is located near the eastern edge of the beach, close to discharge zones; Point 2 is located very close to Point 1, and, at this site, the presence of bathers is already noticeable; Point 3 (P3) lies in a more open and hydrodynamically exposed area with recreational and swimmer activities; and Point 4 (P4) is positioned near Ary Parreiras Canal, which is part of the drainage system of the Icaraí basin and conveys runoff from higher-elevation areas toward the coastline. Volumes of 500 mL −1 of water from each selected point were collected in sterile vials at a depth of approximately 20 to 100 cm. Water samples were collected and treated as described by Rice and Bridgewater (2012); APHA (2012), and ISO (2006) maintained at 4°C, immediately transported to the laboratory; and analyzed within 1 hour. Tenfold serial dilution (10 −1, 10 −2, and 10 −3) was carried out, then 200 μL (Montezzi et al., 2015) of the appropriately diluted sample was placed onto Petri dishes containing culture media: CHROMagar ESBL (Plastlabor, Brazil), CHROMagar mSuperCARBA (Plastlabor, Brazil), Superpolymyxin Agar (Plastlabor, Brazil), CHROMagar™ MRSA(Plastlabor, Brazil), and CHROMagar VRE (Plastlabor, Brazil). The plates were incubated at 37°C for 24 hours for bacterial growth and subsequent counting of colonies (CFU/mL). Distinct colony types were selected and subcultured on plates of Nutrient Agar (NA) in order to stock cultures in brain heart infusion (BHI) supplemented with 10% glycerol at −20°C.

Bacterial identification and antimicrobial susceptibility test

Bacterial identification was performed using the matrix-assisted laser desorption ionization–time-of-flight mass spectrometry (MALDI–TOF MS) technique using a MALDI–TOF apparatus MicroflexLT (Bruker Daltonics, Bremen, Germany). The obtained spectra were processed using the Biotyper software version 3.1 (Bruker Daltonics). In the current work, two tests were carried out for each sample, and E. coli DH5 alpha was used for calibration. The results were interpreted based on scores, with values between 2.00 and 3.00 indicating identification with high confidence and values between 1.70 and 1.99 indicating identification with low confidence. The level of consistency was considered (category A indicates high consistency and high-confidence identification; category B, low consistency; and category C, absence of consistency).

The phenotypic detection of resistance was performed using the disk diffusion method in agar (CLSI, 2022). The diameters of inhibition zones were compared with CLSI-established standard values. The following antibiotics were used, according to CLSI-recommended panels for each bacterial group: amikacin (AMK; 30 µg), gentamicin (GEN; 10 µg), ciprofloxacin (CIP; 5 µg), levofloxacin (LEV; 5 µg), cefotaxime (CTX; 30 µg), cefepime (CPM; 30 µg), ceftazidime (CAZ; 30 µg), ceftriaxone (CRO; 30 µg), meropenem (MPM; 10 µg), ertapenem (ERT; 10 µg), imipenem (IMI; 10 µg), tetracycline (TET; 30 µg), piperacillin–tazobactam (PPT; 100:10 µg), aztreonam (AZT; 30 µg), ampicillin–sulbactam (ASB; 10:10 µg), and sulfamethoxazole–trimethoprim (SXT; 23.75:1.25 µg).

DNA extraction and detection of antimicrobial-resistant genes

DNA from Gram-negative bacteria with phenotypic resistance detected was extracted from individual colonies using the Wizard ® Genomic DNA Purification kit (Promega - Madison, Wisconsin) according to the manufacturer’s instructions. The presence of genes encoding ESBLs (bla GIM, bla SHV, bla TEM, and bla CTX-M) and (bla NDM, bla KPC, bla IMP, bla VIM, bla GES, and bla OXA48) was investigated according to Poirel et al. (2001); Pagani et al. (2003); Dallenne et al. (2010), and Poirel et al. (2011). All amplification reactions were realized on a GeneAmp™ PCR System 9700 thermal cycler (Applied Biosystems - Waltham, Massachusetts).

Analysis of multiple antibiotic resistance phenotype and multiple antibiotic resistance index of bacterial isolates

The analysis of multiple antibiotic resistance phenotype (MARP) and multiple antibiotic resistance index (MARI) of bacterial isolates was performed according to Fadare et al. (2020), based on a mathematical equation reported by Krumperman (1983). The mathematical expression was MAR index = x/y (“x” is the number of antibiotics against which a bacterial species exhibited resistance, and “y’’ is the total number of antibiotics that bacterial species were exposed to). The MARP patterns for isolates that displayed resistance against more than three antibiotics were assessed (Osundiya et al., 2013). Strains showing phenotypic resistance to at least three classes of antimicrobials were classified as MDR based on the recommendations of Magiorakos et al. (2012).

Purification, sequencing, and analysis of PCR product

The bla GES amplicon from the PCR was purified using the PCR Product Purification Kit (Illustra GFX PCR DNA and Gel Band Purification Kit, GE Healthcare - Chicago, Illinois) according to the manufacturer’s instructions. The reaction was sent to the DNA-PDTIS Sequencing Platform of the Sequencing and Bioinformatics Laboratory of IOC-FIOCRUZ, which uses the 3730 DNA Analyzer sequencer (Applied Biosystems - Waltham, Massachusetts). Variance in the nucleotide sequence of the bla GES amplicon was performed on the Sequencher software version 5.4.6 (Gene Codes Corporation ©). The amplicon sequence has been deposited in GenBank under the accession number PP871353.

Statistical analysis

To evaluate significant differences concerning antibiotic resistance patterns among coliforms, Pseudomonas sp., Aeromonas sp., and Acinetobacter sp. isolated in this survey, we conducted the Kruskal–Wallis test by including only the resistant strains. For that, we considered the following variables: ratio between the number of antibiotics to which the strains exhibited resistance and the number of resistant isolates in the analyzed group, that is, in other words, in average, to how many antibiotics were the resistant strains resistant to (ARA); the ratio of the number of resistant strains to the total number of isolates tested (R/tot.); and the MAR index calculated.

We also aimed to compare the differences in antibiotic resistance patterns among the four sampling points. We performed the Kruskal–Wallis test computing the following variables: ratio between the number of antibiotics to which the strains exhibited resistance and the number of resistant isolates in the analyzed point, that is, in other words, in average, to how many antibiotics were the resistant strains resistant to (ARA); the ratio of the number of resistant strains to the total number of isolates tested (R/tot.); and the CFU log/mL on mSuperCARBA™ medium.

Identification of bacterial isolates and phenotypic assays for detection of antimicrobial resistance

A total of 145 pure bacterial cultures were obtained, and MALDI–TOF MS analyses provided the identification of 137 strains (n = 94.48%). The majority were representatives of the Enterobacterales (n = 45/32.85%), Acinetobacter spp. (n = 37/27.01%), and Pseudomonas spp. (n = 17/12.41%). The other bacteria identified were classified as Aeromonas spp. (n = 11/8.03%), Bacillus spp. (n = 4/2.92%), Comamonas spp. (n = 4/2.92%), Staphylococcus spp. (n = 3/2.19%), Chryseobacterium spp. (n = 2/1.46%), Photobacterium spp. (n = 2/1.46%), Sphingobacterium spp. (n = 2/1.46%), Vibrio spp. (n = 2/1.46%), Wautersiella spp. (n = 2/1.46%), Enterococcus spp. (n = 2/1.46%), Delftia spp. (n = 1/0.73%), Empedobacter spp. (n = 1/0.73%), Micrococcus spp. (n = 1/0.73%), and Stenotrophomonas spp. (n = 1/0.73%). The inability to identify approximately 5% of the isolates reflects the limited number of environmental bacterial species included in the Bruker database.

Figures 1–3 provide phenotypic resistance data to 16 antimicrobials obtained for the bacteria identified as Enterobacterales (n = 45), Acinetobacter spp. (n = 35), and Pseudomonas spp. (n = 17), respectively. Two Acinetobacter spp. isolated failed to regrow, preventing the performance of antimicrobial susceptibility testing. Overall, isolates exhibited diverse phenotypes, with frequent resistance across multiple antibiotic classes. Across sampling points, P1–P4 contained Enterobacterales isolates with resistant and intermediate phenotypes, with P1 and P2 showing the broadest distribution of resistance across drug classes. Notably, P1 displayed a greater number of Acinetobacter isolates with resistance to multiple classes, including carbapenems, while P2 and P4 showed more restricted resistance patterns. Resistance was primarily concentrated in ceftazidime, cefepime, aztreonam, and imipenem, especially in Pseudomonas isolates from P1 and P2. Some isolates also showed resistance or intermediate phenotypes to fluoroquinolones (ciprofloxacin and levofloxacin) and aminoglycosides, although less frequently.

Determination of the multiple antibiotic resistance index and the phenotypic pattern of multiple antibiotic resistance phenotype and statistical analysis

Table 1 and Figures 1–3 demonstrate the MARP patterns and provide details of the MARI indices exhibited by the bacterial isolates, which showed resistance to three or more of the 16 antibiotics evaluated. A total of 12 isolates displayed resistance against at least three of the tested antimicrobials and were classified as MDR. An additional 16 isolates displayed resistance to one or two. No predominant MARP was identified, and the MARI ranged from 0.19 to 0.62.

The statistical evaluation of differences, determined using the Kruskal–Wallis test, concerning antibiotic resistance patterns present among coliforms, Pseudomonas sp., Aeromonas sp., and Acinetobacter sp., considering the variables ARA, R/tot., and the MAR index, indicated no significant differences, with a p-value of 0.92. The Kruskal–Wallis results showed no significant differences, with p-values of 0.28 in antibiotic resistance patterns among the four sampling points, considering the variables ARA, R/tot., and CFU Log mSuperCARBA. Furthermore, we excluded sampling point 3, where no resistant isolates were found, and the Kruskal–Wallis test showed no significant difference with a p-value of 0.58. The overall average values observed for the variables characterizing antimicrobial resistance profiles—ARA, R/tot., MAR index, and Log CFU mSuperCARBA—were 2.36, 0.44, 0.26, and 2.53, respectively.

Assays for the detection of antibiotic resistance determinants

The presence of ESBL genes was detected in Enterobacter cloacae, Klebsiella variicola, and Kluyvera ascorbata, and a carbapenemase gene (bla GES) was identified in Citrobacter freundii (Table 2). In the present study, a bla GES resistance determinant was amplified from a single strain of C. freundii. The amplicon was sequenced and showed 99.7% sequence similarity (783 bases/786 bases) (Table 3) with the bla GES-16 gene.

In summary, Figures 4 and 5 summarizes the data obtained by the present study in the collection points in Icaraí Beach. At P1, located near an effluent discharge site, 82 bacteria were identified, with Enterobacterales representing the largest group (n = 32). Three isolates were classified as MDR, and two carried bla CTX-M genes. At P2, also influenced by effluent discharge, 33 bacteria were recovered, predominantly Acinetobacter spp. (n = 12). At P3, an area without apparent effluent influence, only three bacteria were isolated, mainly Bacillus spp. (n = 2), with no MDR isolates and no beta-lactamase genes identified. At P4, a third effluent-impacted site, 19 bacteria were identified, including Acinetobacter spp. (n = 6). Two isolates were MDR, and one carried bla TEM.

The State Institute for the Environment (INEA, Instituto Estadual do Ambiente) is responsible for verifying the suitability for bathing of the beaches and the water quality of Guanabara Bay using the parameters determined in the resolution (274/2000) of the National Council for the Environment (CONAMA-Conselho Nacional de Meio Ambiente). In order to verify the suitability for bathing on the beaches, the variables measured are the presence of thermotolerant coliforms (MPN/100 mL −1) and enterococci (MPN/100 mL −1). In addition, physico-chemical criteria, presence of thermotolerant coliforms (MPN/100 mL −1), and phytoplankton are used in the determination of the water quality in Guanabara Bay (INEA, 2023). According to their analysis on a collection point in Icaraí Beach during the same month that we collected the water samples (results are available online, from the 2020/2021 biennium http://www.inea.rj.gov.br/publicacoes/publicacoes-inea/boletins-de-qualidade-das-aguas/), the Water Quality Index indicated that the water was of a very poor category, with high coliform counts. Icaraí Beach is known for its generally poor water quality and receiving domestic waste effluents (Barreto et al., 2015; Fistarol et al., 2015; Costa et al., 2023), as well as being part of Guanabara Bay, which is recognized as being contaminated by untreated domestic effluents, industrial waste, and pollutants such as hydrocarbons, heavy metals, and suspended solids (Fistarol et al., 2015); despite this, it remains a popular recreational site for the local population. This persistence is likely due to a lack of alternatives and the necessity for accessible leisure spaces. Interestingly, despite the beach’s widespread use and known contamination, the present study represents the first to investigate in detail the presence of MDR bacteria in its surface waters.

Among 145 pure bacterial cultures, the majority of isolates were Enterobacterales (32.85%), Acinetobacter spp. (27.01%), and Pseudomonas spp. (12.41%). According to the WHO, this group of bacteria has acquired resistance to a large number of antibiotics, including carbapenems and also to third-generation cephalosporins (World Health Organization (WHO), 2017), and as such, these bacteria are included in a Global Priority Pathogens List (World Health Organization (WHO), 2017), which was updated by the WHO in 2024 (World Health Organization (WHO), 2024).

The presence of Enterobacter spp. and Klebsiella spp. resistant to cephalosporins, aztreonam, gentamicin, and trimethoprim–sulfamethoxazole, as previously detailed in river water flowing into Guanabara Bay (de Araujo et al., 2016), is consistent with the findings of the current work. The data in Figure 1 demonstrated that tetracycline resistance was the most prevalent (8/45) phenotype among the Enterobacterales collected. This observation was not considered surprising given that tetracyclines remain the most widely employed class of antimicrobial globally, principally in the animal health sector (Wang et al., 2024). Moreover, their persistence as bioactive compounds in both aquatic and soil environments, resulting in prolonged and repeated exposure of microbial communities to these agents, is well documented (Roberts, 2018).

Table 1 and Figures 1–3 show the MARP patterns and MARI indices of bacterial isolates resistant to three or more of the 16 tested antibiotics. A total of 12 isolates were classified as MDR, while 16 isolates exhibited resistance to one or two antibiotics. Importantly, a MAR index value of more than 0.2 is considered indicative of intensive usage of antibiotics in the environment under study and may be employed to predict an elevated risk for the promotion of antibiotic resistance (Osundiya et al., 2013). In spite of the inherent limitations of a relatively small sample size, the association of the novel MARI data with the microbial water quality data recorded by Inea and CONAMA for Icaraí Beach during the collection period indicated that the collection points examined in the current investigation represent high-risk environments wherein there was either selective pressure for the emergence of antibiotic resistance or persistent contamination.

ESBL genes were found in E. cloacae, K. variicola, and K. ascorbata, while a carbapenemase gene (bla GES) was identified in C. freundii (Table 2). The detection of ESBL-encoding genes in environmental isolates has been previously documented (Picão et al., 2013; Stasiak et al., 2021; Canellas et al., 2021; Costa et al., 2023), often attributed to the widespread use of β-lactams in both human and veterinary medicine. The bacterium K. variicola has been described as an eco-friendly organism (Duran-Bedolla et al., 2021), but importantly, it has also been reported as an emerging human pathogen (Rodríguez-Medina et al., 2019). Despite carbapenemase resistance not being detected in the K. variicola strain reported herein, the combined genotypic and phenotypic resistance data (Table 1, Figure 1) ranked it as the MDR isolate with the highest resistance index (0.62), based upon resistance to 11 of the 16 antimicrobials examined, including aminoglycosides, fluoroquinolones, β-lactams, and tetracyclines. When considered together with the observed co-occurrence of three ESBL genes in this isolate, the potential of this species to act as a reservoir for resistance, contributing to the dissemination of multiple resistant genes in both clinical and environmental settings, is indisputable.

The detection of MDR bacteria in aquatic environments is closely linked to the discharge of inadequately treated effluents from sewage treatment plants (STPs). In the city of Niterói, four STPs release their effluents into Guanabara Bay (Fistarol et al., 2015). Observations of collection sites revealed that collection points P1 and P4 received effluents that, in theory, should have undergone prior treatment before being released into the water bodies. However, due to the city’s rapid population growth and the corresponding increase in sewage volume, it is possible that the treatment processes were insufficient to eliminate a significant portion of microorganisms. It is important to note that these water treatment systems are designed to remove and/or neutralize most antimicrobial compounds present in the water, further ensuring its safety. Nevertheless, continuous monitoring and improvements in treatment processes are essential to guarantee the effectiveness of these systems in removing harmful substances. A study on hospital effluents found that the most frequently isolated bacterial groups were from the order Enterobacterales, followed by members of the genera Aeromonas and Acinetobacter (Picão et al., 2013). In the current study, many of the bacteria isolated were from the same groups, along with species pertaining to the genus Pseudomonas, despite differences in water salinity. Except for Aeromonas, these bacterial groups are strongly associated with both hospital- and community-acquired infections, which can be challenging to treat due to the continued emergence of antimicrobial resistance (Horcajada et al., 2019; Nasr, 2020; Tompkins and van Duin, 2021).

The number of studies reporting the presence of carbapenemase resistance in environmental sources, especially aquatic reservoirs (Montezzi et al., 2015; de Araujo et al., 2016; Picão et al., 2013; Chagas et al., 2011; Oliveira et al., 2014; Paschoal et al., 2017), has grown almost exponentially. A highly focused study (Paschoal et al., 2017), which sampled water from Botafogo and Flamengo beaches to detect carbapenemase resistance, was conducted in the same bay (Guanabara Bay) as the Icaraí Beach examined herein. Representatives of the bla GES class of resistance determinant identified in the MDR isolate of C. freundii (Figure 1) were also recorded in the study of Paschoal et al. (2017), albeit in Enterobacter spp. and Klebsiella spp. suggesting either their persistence or the continued input of bacteria carrying this group of resistance determinants into the Guanabara Bay ecosystem. The GES enzyme (GES enzymes) is classified as class A β-lactamases and functional group 2f, having a broad spectrum of hydrolysis of β-lactams (Queenan and Bush, 2007). The bla GES, class of determinants, was considered rare but has now been described globally among clinical and environmental strains, commonly present within integrons, being frequently associated with mobile genetic elements, which serve to explain its widespread distribution (Streling et al., 2018). According to the Beta-Lactamase DataBase: Structure and Function (http://bldb.eu/), until now, 49 GES variants have been described, and at least 14 are carbapenemases. In the present study, bla GES resistance showed higher sequence similarity with the sequence NG_049118 deposited in the GenBank as bla GES gene for carbapenem-hydrolyzing class A beta-lactamase GES-16, present in the plasmid pSM190 carried by a strain of Serratia marcescens, isolated from a human patient hospitalized in Rio de Janeiro, Brazil (Streling et al., 2018).

This bla GES allele was initially described in S. marcescens in 2005 from bronchial secretion and catheter blood from two patients admitted to a hospital located in Rio de Janeiro (Streling et al., 2018), and it is also disseminated widely in Guanabara Bay and in river water entering the Guanabara Bay (Montezzi et al., 2015; de Araujo et al., 2016). Moreover, C. freundii carbapenem-resistant isolates have been reported in hospital environments (Dziri et al., 2022). Evidently, our findings were not conclusive in linking ertapenem-resistant C. freundii to wastewater treatment (WWT)-associated pollution, but this possibility cannot be discounted. In this context, Guanabara Bay receives water from municipal and hospital wastewater treatment plants, which are considered a source of antibiotic-resistant bacteria (Fistarol et al., 2015; Roberts, 2018).

Guanabara Bay has been the subject of numerous studies due to its high degree of deterioration as a consequence of human activities (Silva et al., 2019). As mentioned previously, several studies conducting surveys to detect the presence of AMR bacteria demonstrated the presence of diverse bacterial species and resistance genes in this water body (Montezzi et al., 2015; de Araujo et al., 2016). Moreover, Costa et al. (2023), in a survey of E. coli producing extended-spectrum β-lactamases, highlighted the presence of this bacterium in waters with different levels of pollution in Guanabara Bay. Although limited to a survey conducted at a single time point, the current investigation provides a snapshot that reinforces the problem of antibiotic resistance in Guanabara Bay. Phenotypic resistance to all classes of antibiotics examined was detected throughout three of the four collection points, with statistical analyses demonstrating no significant differences in the degree of contamination among those locations. When combined with the findings of previous investigations of AMR in Guanabara Bay, the data presented herein reinforce the hypothesis that AMR has become established as a permanent characteristic of this ecosystem.

Taken as a whole, the data from studies performed in water collected from Guanabara Bay suggest that ESBL genes are well established in this environment, persisting through horizontal or vertical transmission among resident bacteria, or the bay water is being constantly contaminated by effluents from hospitals or even residences where this resistance mechanism is already established. The presence of bacteria containing antimicrobial resistance genes in the aquatic environment is not exclusive to the current scenario; resistance has been found naturally in these places at controlled levels. However, anthropogenic activities in the waters of seas, rivers, lakes, and others have served to stimulate gene exchange between circulating microorganisms, enhancing the spread of antimicrobial resistance (Larsson and Flash, 2022; Jurelevicius et al., 2021; Zhuang et al., 2021). In the absence of surveillance activities that monitor the quantitative and qualitative aspects of AMR, it will be challenging to determine what are the impacts of resistant bacteria and their associated resistance determinants upon environmental, animal, and human health. Accurate measurement of resistant bacteria profiles may provide a basis for the development of more effective prevention and control strategies in both healthcare and community settings. Examples of such strategies include implementing stringent hygiene measures, education on the use of antimicrobials, and continued and expanded research to develop new drugs and alternative therapies.

This study has several limitations that must be explicitly acknowledged. First, only a single water sample was collected at each of the four points on a single date. As a consequence, the results do not allow inference of permanence, establishment, seasonal variation, or temporal stability of resistant bacteria in this ecosystem. Second, the use of highly selective media was intentional to detect clinically relevant MDR organisms, but this approach inherently introduces bias and does not permit estimation of true prevalence or abundance in the total bacterial community. Therefore, the findings should be interpreted as evidence of presence. Furthermore, although recreational exposure pathways exist for swimmers, sailors, and fishers, the present study did not measure colonization, infection, or viable pathogen concentrations under realistic exposure scenarios. Thus, the results support the potential risk of human exposure rather than confirming direct health outcomes. Future studies incorporating quantitative microbial risk assessment, metagenomics, and longitudinal sampling would be required to address these aspects.

The current work highlights the presence of bacteria resistant to multiple classes of antimicrobial compounds in recreational waters observed at high tide, theoretically with a greater dilution of the components present than would occur during low tide. Comparative analysis of antimicrobial resistance profiles, examining both taxonomic groups and sampling locations (specifically sites 1, 2, and 4) using ARA, R/tot., MAR, and Log/mL mSuperCARBA, showed no statistically significant differences. Therefore, the elevated overall values observed for these metrics suggest the worrisome presence and potential dissemination of antimicrobial resistance within the studied environment. The direct impact on animals and humans in that location was not demonstrated. However, we suggest that there is a risk to the health of individuals who frequent that location, and as demonstrated previously, many of the resistant bacteria found show potential to cause infections in humans and animals and may negatively impact antimicrobial therapy. The data presented herein corroborate previous studies that have demonstrated the need for more surveillance research in water bodies, monitoring the presence of resistant bacteria, antimicrobial resistance genes, and antimicrobial compounds in that location, in addition to monitoring the efficiency of wastewater treatment plants.