The potential associations of antimicrobial resistance in clinical Escherichia coli and Staphylococcus aureus across the human-animal interfaces in Chongming Island, Shanghai

doi:10.21203/rs.3.rs-4281909/v1

Background： AMR is a significant concern within the One Health framework due to its ongoing spread across humans, animals, and the environment. Phenotypic data remains the primary form of surveillance data for AMR. Yet, there is a noticeable absence of thorough data analysis across various interfaces.

Methods: The AMR phenotypic data of clinical and food animal E. coli and S. aureusfrom Chongming Island over the past five years were analyzed to determine the key characteristics of AMR and its association at the human-animal interface.

Results: The clinical isolates of E. coli and S. aureus demonstrated a significant level of AMR rates, including penicillins (83.92%), cephems (63.05%), fluoroquinolones (62.21%), tetracyclines (57.77%) for E. coli and penicillinase-labile penicillins (90.89%), macrolides (51.51%), penicillinase-stable penicillins (43.96%), lincosamides (43.55%) for S. aureus.. ESBL-positive E. coli isolates account for 53.26% (1398/2526), with MRSA prevalence at 43.81% (435/993). Notably, there has been an increase in the number of E. coli isolates resistant to 8 to 12 antimicrobial classes, and that of S. aureus isolates resistant to 5 to 9 classes. Furthermore, certain MDR phenotypes were first identified in food animal isolates. No significant association was found in the identical AMR phenotype between food animals and clinical isolates. However, several MDR phenotypes shared by the two interfaces were identified, with 44 in E. coli and 12 in S. aureus. Through further co-occurrence analysis, 16 pairs of phenotypes were found in both clinical and food animal E. coli isolates. Moreover, two unique co-occurrence clusters of MDR were identified: Pen-βLac-Cep in clinical interface and Ami-Tet-Fpa in food animal interface. Meanwhile, three clusters were identified in clinical S. aureus, namely Ami-Mac-Lin, Ami-Mac-Flu-Lin, and PSPen-Ami-Flu-Lin and two clusters were identified in food animal isolates, namely PSPen-Mac and PSPen-Flu-Lin.

Conclusion: A heightened MDR prevalence in clinical E. coli and S. aureus have been discovered. Some MDR profiles manifest in food animal interface before clinical settings, highlighting the potential for AMR association and its transmission between humans and food animals. Within the One Health framework, integrating genomic data into AMR monitoring is an essential next step.

Multidrug resistance

Escherichia coli

Staphylococcus aureus

Clinical isolates

Food animals

Antimicrobial resistance (AMR) is a pressing concern in public health on a global scale, as it has resulted in increased rates of illness, mortality, and economic burdens worldwide [1–3]. Meanwhile, AMR has been recognized by the World Health Organization (WHO) as one of the ten significant global health threats to humanity [4]. According to the Global Research on Antimicrobial Resistance (GRAM) Project report, there were around 4.95 million bacterial drug-resistant infections in 2019 worldwide, resulting in 1.27 million deaths that were directly attributed to AMR [5]. The global AMR crisis has impacted nearly all countries, including China, the leading producer and consumer of antimicrobials, with an annual usage of approximately 210,000 tons [6, 7]. Two recent studies have shown an increasing pattern of antimicrobial usage over the next decade, predominantly in food animals and Asia [8, 9]. This suggests that addressing AMR will primarily center on the continent of Asia.

Escherichia coli (E. coli) and Staphylococcus aureus (S. aureus) are commonly regarded as the primary Gram-negative (G^-) and Gram-positive (G⁺) bacterial pathogens, respectively. E. coli plays a dual role in the gastrointestinal tract of humans and animals; it acts as a commensal, coexisting with both humans and animals, yet it can also cause severe infections that pose a life-threatening [10]. While S. aureus is a significant pathogen that can cause infections in hospital and community settings [11]. Moreover, it is the main factor behind bovine mastitis, leading to substantial financial repercussions in the agricultural sector [12]. It is reported that E. coli and S. aureus were identified as the leading pathogens causing over 250,000 deaths related to AMR in 2019 [5]. Meanwhile, E. coli and S. aureus are the most frequently isolated G^- and G⁺ bacteria in China's national AMR surveillance program [13, 14]. In recent years, there has been a growing global concern over the spread of extended-spectrum β-lactamase (ESBL)-producing E. coli and MRSA (methicillin-resistant S. aureus) [15–18]. It is crucial to note that ESBL-producing E. coli and MRSA frequently demonstrate resistance to numerous other classes of antimicrobial drugs [19, 20]. Significantly, specific resistance phenotypes or resistance genes, such as streptothricin-resistant phenotype [21], blaVIM-1 gene (in porcine E. coli) [22], plasmid-mediated mcr-1 gene [23], and methicillin-resistant phenotype [17], were initially discovered in the two pathogens. Therefore, these two bacterial species are frequently used as AMR surveillance bioindicators.

AMR is considered a significant concern within the One Health framework, as bacteria play a vital role in connecting the various interfaces [24, 25]. Antimicrobial-resistant bacteria (ARB) can spread to humans either by the food supply, direct contact with food or companion animals, or, more indirectly, through environmental pathways, including waterways, soils, and vegetables contaminated with human or animal waste, and vectors such as rodents, insects, and birds [26]. Human wastes from homes, hospitals, and offices also contaminate rivers and waterways with ARBs. Therefore, the environmental carriers, including soil and water bodies, represent a “hot spot” of ARBs [27]. Furthermore, the spread of ARBs can occur between farms through infected carrier animals, companion animals, or wildlife vectors. Additionally, patients and bacteria flow between community and hospital environments [26, 28]. Notably, plasmids serve as carriers that facilitate the transfer of ARGs (antimicrobial resistance genes) among bacteria in intra- and inter-microbiome contexts, even across genetically diverse species or genus [29].

Chongming Island stands out among the districts of Shanghai due to its distinct separation, thus rendering it an ideal spot for conducting One Health research and activities. Furthermore, Chongming Island remains to be characterized by an ongoing process of urbanization, preserving a significant portion of its original ecological system. [30]. The availability of medical facilities on Chongming Island, especially larger hospitals, is relatively limited compared to the Shanghai metropolitan area. Meanwhile, Chongming Island has a crucial role in the breeding and agricultural industry, supplying substantial food ingredients for Shanghai [31]. Thus, Chongming Island is highly suitable for investigating the potential relationship between the human and animal interfaces based on AMR phenotypic data. Here, we explored the characteristics of clinical interfaces by utilizing phenotypic data from two bacteria species, E. coli, and S. aureus, obtained from one prominent hospital in Chongming. Afterward, we investigated the association of AMR at the human-animal interface by incorporating the phenotypic data of isolates from food animals. We intend to provide additional advice for AMR control and aid in establishing a One Health surveillance network for AMR in Chongming Island.

Study design

A dataset comprising phenotypic data, obtained from nosodochium and food animals, was utilized to explore the AMR characteristics of clinical E. coli and S. aureus isolates over the period 2018 to 2022. The study further sought to identify potential associations between AMR in clinical and food animal isolates, aiming to establish a One Health surveillance framework for AMR and offer more insights for AMR control. The sampled farm is situated close to the medical institution.

Sample collection and identification

The clinical samples were collected from one prominent nosodochium located on Chongming Island, and these samples subsequently underwent bacterial culture isolation. Only the data of bacteria isolates with unique codes were recorded in the dataset, without any additional information about the host patients. The sample collection and culture were conducted following established routine protocols in the hospital setting. Briefly, the swabs taken from the patients were placed in Luria Bertani (LB) broth (ThermoFisher, Waltham, USA) and incubated overnight at 37℃ and 200 rpm. After that, the swab cultures were streaked on E. coli coliform chromogenic media or Staphylococcus chromogenic prepared plate medium (Hopebio biotech company, Qingdao, China) and incubated at 37 °C for 16 h. Colonies that exhibited similar morphological characteristics were selected and verified using the fully automatic microbial mass spectrometry detection system Auto ms1000 (Autobio company, Zhenzhou, China) utilizing matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS). The confirmed isolates were preserved at -80 °C using commercial magnetic bead strain preservation tubes (Pro-Lab Microbank, Canada).

Collecting and identifying food animal-derived isolates followed a similar procedure as above described with a few notable distinctions: 1) individual samples for E. coli were collected from rectum or cloaca of asymptomatic animals, and samples for S. aureus were isolated from raw milk; 2) A second culture was done using the nutrient agar medium after culturing in colorimetric medium; 3) the pure colonies were identified as E. coli or S. aureus using a microbial biochemical identification system (VITEK 2 Compact, Biomerieux, France).

Testing for antimicrobial susceptibility

The antimicrobial susceptibility test (AST) of clinical isolates was evaluated by determining the minimum inhibitory concentration (MIC) values of different kinds of antimicrobials following the recommended microbroth dilution protocol (CLSI M100, 32th Edition) of the Clinical & Laboratory Standards Institute (CLSI, United States). The AST of clinical E. coli included 29 antimicrobial agents from 13 different antimicrobial classes, and the AST of clinical S. aureus included 16 agents from 12 different classes (Table 1). In 2020, cefoxitin and aztreonam were incorporated into the clinical E. coli AST schedule, while in 2021, several agents, such as amoxicillin/clavulanic acid, ceftriaxone, ertapenem, tobramycin, tetracycline, moxifloxacin, chloramphenicol, nitrofurantoin, and colistin were added. In 2021, the clinical S. aureus AST schedule was also updated by introducing four agents: tigecycline, ciprofloxacin, moxifloxacin, and chloramphenicol. Quality control was performed using E. coli ATCC25922 and S. aureus ATCC25904. In addition, the AST of food animal E. coli isolates included 14 agents belonging to 10 antimicrobial classes, and that of food animal S. aureus isolates included 18 agents belonging to 14 antimicrobial classes (supplement Table 1).

The results were interpreted using the recommended breakpoints by CLSI for antimicrobial agents. When the CLSI breakpoint was unavailable, alternative breakpoints from the European Committee on Antimicrobial Susceptibility Testing (EUCAST) and the Shanghai Local Standard of Technical Specification for Antimicrobial Resistance Surveillance in Breeding of Livestock and Poultry (DB31/T 1160-2019) were utilized for interpretation. The breakpoints classified the isolates as susceptible or non-susceptible (intermediate and resistant).

Statistical analysis

All analyses in this study were conducted by R 4.3.2 (Lucent Technologies, Jasmine Mountain, USA) with given packages. The resistance rates of all antimicrobial classes, resistance prevalence of all agents, and rates of multidrug resistance (MDR) were calculated using the "dplyr" and "comparegroups" packages, and the bar charts, pie charts, and heatmaps were generated by package of “ggplot2” and “ggprism.” Meanwhile, scatter plot of annual dynamic changes of all class resistant rates, agent prevalence and MDR rates were displayed by “ggplot2”. The Mann-Kendall trend tests were performed using the "trend" package. The AMR carriage was defined as the total number of antimicrobial classes to which an isolate was phenotypically resistant. This measurement evaluated MDR dynamics changes over five years utilizing generalized linear mixed models (GLMMs). The GLMMs were implemented using the "lme4" package. Heatmaps were used to display all AMR carriage profiles (resistance profiles). The correction of AMR phenotypes and MDR rates between clinical and food animal isolates was calculated using the "psych" package and the results were visualized by heatmap using "pheatmap" package. Subsequently, the shared phenotypes between the clinical and food animal isolates were displayed by the Venn Diagram “VennDiagram” package. To further investigate the potential association, the pairwise co-occurrence matrix that identified the absence and presence of AMR phenotypes in both clinical and food animals were constructed aiming to uncover any identical co-occurrence pairwise at the human-animal interface. The co-occurrence analysis was performed using "polycor" package and visualized with "corrplot" package. The categorical data analysis was conducted using the chi-square test or Fisher exact test using the built-in function of R. The P-value of 0.05 was chosen as statistically significant.

The status of AMR in clinic E. coli and S. aureus isolates.

During the sampling period from 2018 to 2022, 2625 clinical isolates of E. coli and 994 clinical isolates of S. aureus were collected. The annual sampling quantity of the two bacteria species was shown in supplement Table 2.

The AST results revealed that a significant proportion of clinical E. coli isolates demonstrated resistance to penicillins (83.92%), cephems (63.05%), fluoroquinolones (62.21%), and tetracyclines (57.77%, Fig. 1A). A low proportion of clinical E. coli isolates were resistant to carbapenems (3.73%), fosfomycin (6.40%), nitrofuran (6.73%), polymyxins (0.74%, Fig. 1A). Out of the six antimicrobial classes that contain more than one agent, the resistance prevalence of distinct agents within the same class was evident discrepant except for classes of carbapenems and fluoroquinolones (Fig. 1C). Among the tetracyclines class, tigecycline exhibited a resistance prevalence nearly zero (0.34%), whereas tetracycline displayed a resistance prevalence exceeding fifty percent (57.77%).Moreover, of the 7 agents classified as cephems, the isolates exhibited the most significant resistance prevalence to first-generation agent (cefazolin [61.95%]), followed by second-generation agents (excluding cefoxitin, cefuroxime [55.09%]) and third-generation agents (cefotaxime [53.26%], ceftazidime [34.90%], and ceftriaxone [52.07%]). The isolates exhibited the lowest resistance prevalence to fourth-generation agents (cefepime [25.71%]).

The clinical S. aureus isolates showed high resistance to penicillinase-labile penicillins (90.89%), macrolides (51.51%), penicillinase-stable penicillins (43.96%), lincosamides (43.55%). Notably, all isolates of S. aureus were susceptible to glycopeptides and oxazolidinines (Fig. 1B). In addition, the clinical S. aureus exhibited low resistance to the folate pathway antagonists (5.84%) and ansamycins (2.72%). All classes of tetracyclines and fluoroquinolones contain three agents in the clinical S. aureus AST, and the resistance prevalence of distinct agents was also discrepant except ciprofloxacin and moxifloxacin (Fig. 1D). Furthermore, tigecycline (3.57%) and minocycline (7.35%), both from the tetracycline class, have resistance prevalence below 10%.

Subsequently, the AMR traits within the hospital were further elaborated. In short, the differences in resistance rates to most antimicrobial classes were not related to gender and inpatient or outpatient regardless of whether they were infected with E. coli or S. aureus (supplement Fig. 1 and 2, supplement Table 3 and 4). Furthermore, the isolates from neonate and neonatology demonstrated the lowest resistance rates to almost all antimicrobial classes (supplement Fig 3 and 4). Additionally, the resistance rates of isolates obtained from sputum, throat swabs, and urine are higher than those of other sampling positions (supplement Fig 5).

Trends in variation of resistance rates and prevalence

The clinical E. coli AMR dataset comprised 8 classes of continuous data spanning from 2018 to 2022. The linear trend test (2018-2022) revealed that resistance rates for six classes, namely β-Lactam combination agents, cephems, carbapenems, fluoroquinolones, folate pathway antagonists, and fosfomycin, exhibited linear trends. Of these, five classes showed an upward trend (R>0, P<0.05), while the class of folate pathway antagonist exhibited a downward trend (R=-0.072, P< 0.001, Fig. 2A). Notably, the change in resistance rate was relatively minor in magnitude (R<0.2). Furthermore, no discernible patterns in resistance prevalence were observed towards the 15 agents with continuous data (supplement Fig. 6).

Linear trend tests of resistance rate to 11 classes with 5 years of continuous data, were conducted in clinical S. aureus isolates. There was a slight five-year downward trend observed in resistance rates to penicillinase-labile penicillins, penicillinase-stable penicillins, fluoroquinolones, lincosamides, and aminoglycosides, with the most significant decrease observed in resistance rates to fluoroquinolones (R=-0.177, P<0.001, Fig. 2B). In contrast, the resistance rates of S. aureus to folate pathway antagonists and tetracyclines showed a slight upward trend (R=0.215 and 0.281, P<0.001). No isolate resistant to glycopeptides and oxazolidinones was detected over the five years, and resistance rates to macrolides and ansamycins had no linear change trend (P=0.213 and 0.606). All classes of tetracyclines and fluoroquinolones included three agents in the clinical S. aureus AST schedule, and three agents of tetracycline, minocycline, and levofloxacin has continuous data frame, also showed no significant change (supplement Fig. 7).

The prevalence of extended-spectrum beta-lactamases (ESBL) positive E. coli and methicillin-resistant S. aureus (MRSA)

The ESBL-positive isolates account for 53.26% (1398/2526) of all tested clinical E. coli isolates (Fig. 3A). Meanwhile, the linear trend test showed no significant change in prevalence over the five years (P=0.136, Fig. 3B). The prevalence of clinical MRSA was 43.81% (435/993, Fig. 3C), and the prevalence showed a slight downward trend based on the linear trend test (R=-0.106, P< 0.001, Fig. 3D).

The characteristics of ESBL-positive E. coli and MRSA were further analyzed regarding gender, age, hospitalization status, and visiting department of hospital patients. There was a significant difference in the prevalence of ESBL-positive E. coli between male and female patients (61.02%/48.13%, χ²=41.91, P< 0.001, supplement Fig. 8), while there was no difference in that of MRSA (44.35%/42.97%, χ²=0.185, P=0.667). The prevalence of ESBL-positive E. coli (53.77%/48.39%, χ²=2.609, P=0.106) and MRSA (44.02%/28.57%, χ²=1.339, P=0.247, supplement Fig. 8) in inpatient was higher compared to outpatients, but the difference was not statistically significant. Out of the four age groups, the geriatric had the highest prevalence of ESBL-positive E. coli (56.14%) and MRSA (51.17%), whereas the neonate group had the lowest prevalence (21.62%/13.64%, supplement Fig. 8). In addition, the Department of Neonatology had the lowest prevalence of ESBL-positive E. coli (20.59%) and MRSA (14.29%).

Status and dynamics of MDR in clinical E. coli and S. aureus isolates

Out of the 2625 clinical E. coli isolates tested, 1877 (71.50%) were MDR isolates, which were resistant to three or more antimicrobial classes (Fig. 4A). None of the E. coli isolates were resistant to all 13 antimicrobial classes, and only one isolate was resistant to 12 classes, with no colistin resistance. There was an upward trend in the proportion of MDR isolates from 2018 to 2022, mainly those that are resistant to 9 to 12 classes (Fig. 4B and 4C). Notably, a considerable number of isolates have been detected that are resistant to 8 and 9 classes since 2021, and the emergence of isolates resistant to 10 and 11 classes started in the same year. Furthermore, the isolates resistant to 12 classes were detected in 2022.

The proportion of MDR S. aureus isolates was 50.80% (505/994), with none of the isolates exhibiting resistance to all 12 classes (Fig. 4D). Despite the lack of a discernible trend in the proportion of MDR S. aureus over the past five years (P=0.885), the highest proportion of MDR isolates was observed in 2022 (61.69%, Fig. 4E and 4F). Moreover, the isolates resistant to 8 and 9 antimicrobial classes were all identified in 2021 and 2022. The more details of MDR proportion of S. aureus in gender, age, hospitalization status, and visiting department of hospital patients was shown in supplement Fig. 9A and 9D.

The Poisson GLMMs were used to investigate the variation of AMR carriage (the total number of antimicrobial classes to which an isolate was phenotypically resistant) over the five years. The AMR carriages of E. coli in 2020, 2021, and 2022 were found to be higher compared to 2018 (odds ratio [OR]= 1.09, 1.31, and 1.59, P< 0.05), suggesting a deteriorating trend of E. coli MDR (Table 2). In contrast to E. coli, the AMR carriages of S. aureus in 2019, 2020, and 2021 were lower than that of 2018 (OR=0.83/0.91/0.78, P=0.003/0.147/0.001, Table 2), suggesting a decline in the MDR situation of S. aureus. Nevertheless, the slightly elevated AMR carriage of 2022 (OR=1.06, P=0.354) suggested that the issue of MDR should not be disregarded.

The AMR association between clinical and food animal isolates based on sole phenotype data

The AMR characteristics of E. coli and S. aureus isolates from food animals (or raw milk) are depicted in supplement Fig. 10A-C. The correlation coefficient of resistance rates was calculated to assess the AMR association between clinical and food-animal interfaces. The correlation coefficient heatmap for E. coli showed a strong negative correlation in the resistance rate of cephems between the clinical and food animal isolates (r=-1.000, P<0.001). No one-to-one correlation has been established between other phenotypic resistance rates, including the MDR rate. The analysis of the phenotype correlation between clinical resistance rates of S. aureus and isolates from food animals has also revealed the same phenomenon (Fig. 5). These results suggest that establishing connections at the human-animal interface through a single AMR phenotype is exceedingly challenging.

The shared AMR profiles between clinical and food animal interfaces

323 unique AMR profiles were identified in the clinical and food animal E. coli isolates including 250 profiles in clinical E. coli isolates, 29 profiles in food animal isolates and 44 shared by both interfaces (Fig. 6A).

Out of the 294 profiles in clinical E. coli isolates, the profiles contained six antimicrobial classes (antibiogram lengths 6) had the highest number of unique profiles (51 kinds), followed by antibiogram lengths 5 (47 kinds) and antibiogram lengths 4 (43 kinds, supplement Table 5, Fig. 6C). Notably, there was a noticeable rise in the diversity of AMR profiles, and the profiles with antibiogram lengths ranging from 7 to 12 were almost observed in 2020, 2021, and 2022 (Fig. 6D). Interestingly, each antibiogram length had at least one dominant AMR profile (supplement Table 5), including Pen-Fpa (78/293) and Pen-Flu (63/293) with antibiogram length 2; Pen-Cep-Flu (70/361) with antibiogram length 3; Pen-βLac-Cep-Flu (81/446) and Pen-Cep-Flu-Fpa (75/446) with antibiogram length 4; Pen-Cep-Ami-Flu-Fpa (63/371) and Pen-βLac-Cep-Flu-Fpa (56/371) with antibiogram length 5; Pen-βLac-Cep-Ami-Flu-Fpa (105/333) with antibiogram length 6; Pen-βLac-Cep-Ami-Flu-Fpa-Mon (40/170) with antibiogram length 7; Pen-βLac-Cep-Ami-Tet-Flu-Fpa-Mon (29/103) and Pen-βLac-Cep-Tet-Flu-Fpa-Mon-Phe (21/103) with antibiogram length 8; Pen-βLac-Cep-Ami-Tet-Flu-Fpa-Mon-Phe (28/64) with antibiogram length 9; Pen-βLac-Cep-Ami-Tet-Flu-Fpa-Mon-Fos-Phe (10/20) with antibiogram length 10; Pen-βLac-Cep-Ami-Tet-Flu-Fpa-Mon-Fos-Phe-Nit with antibiogram length 11 (4/8).

Out of the shared AMR profiles in E. coli, there were 7 AMR profiles in clinical scenarios in 2018, 2019, and 2020, consisting of four profiles with antibiogram length 1, two with antibiogram length 2, and one with an antibiogram length 3. In 2021, 24 new AMR profiles emerged, and an additional 13 profiles were observed in 2022. (supplement Table 6). In contrast, the AMR profiles of E. coli isolates from food animals were frequently observed in the initial three years (30/44), with 17 profiles in 2018, 6 in 2019, and 7 in 2020 (supplement Table 6). The results suggest that specific AMR profiles, particularly those with antibiogram length of three or more (MDR isolates), may exist earlier in the E. coli from food animals.

The S. aureus isolates from both clinics and food animals exhibited 141 AMR profiles, including 107 in clinical isolates, 22 in food animal isolates, and 12 shared by both interfaces (Fig. 6B). Among the 119 profiles of clinical isolates, both antibiogram length 3 and 4 exhibited 22 different AMR profiles, followed by antibiogram length 6 (19 kinds) and antibiogram length 5 (18 kinds, Fig. 6C). The AMR profiles in 2022 (68) and 2021 (54) were significantly higher compared to those in 2018 (33), 2019 (27), and 2020 (29). Additionally, there was an increase in the proportion of antibiogram length exceeding 3 in 2021 (68.52%, 37/54) and 2022 (83.83%, 57/68, supplement Table 6). Similarly, each antibiogram length of clinical S. aureus had one dominant AMR profile, namely PLPen-PSPen with antibiogram length 2 (99/204); PLPen-Mac-Lin with antibiogram length 3 (72/129); PLPen-PSPen-Mac-Lin with antibiogram length 4 (87/146); PLPen-PSPen-Mac-Flu-Lin with antibiogram length 5 (14/58); PLPen-PSPen-Ami-Mac-Flu-Lin with antibiogram length 7 (62/99); PLPen-PSPen-Ami-Mac-Tet-Flu-Lin with antibiogram length 7 (51/69, supplement Table 7).

The shared AMR profiles in S. aureus consist of 3 profiles with antibiogram length 3 and 1 with antibiogram length 4 (supplement Table 8). Interestingly, the four profiles also emerged in clinical isolates in 2021 or 2022 and were observed during the initial three years in food-animal isolates. This observation aligned with that of E. coli, which could potentially strengthen the validity of our conclusion.

Co-occurrence analysis discovered two apparent clusters in the resistant phenotype dataset of E. coli. One cluster was penicillins-β-Lactam_combination_agents-cephems- monobactams (Pen-βLac-Cep-Mon) in clinical isolates, and the other was aminoglycosides-tetracyclines-folate_pathway_antagomists (Ami-Tet-Fpa) in food animal isolates (P< 0.01, correlation coefficient > 0.5, Fig. 7A and B) was found in food animal isolates. The cluster of Ami-Tet-Fpa was detected in 10 AMR profiles of food animal isolates, 11 profiles of clinical isolates, and four profiles shared by both interfaces. As the monobactams class was not adopted in food animal E. coli AST, an alternative cluster of Pen-βLac-Cep was scanned at the human-animal interface and was detected in 125 AMR profiles of clinical isolates, eight profiles of food animal isolates, and six profiles shared by both interfaces (Table 3). Furthermore, 13 pairs of phenotypes were found in clinical E. coli isolates. Six co-occurrence pairs (3 pairs were same with clinic isolates) in food animal isolates were identified. Most of them (except Cep-Fos and Fos-Phe) were detected at human and food animal interfaces (Table 3). The results provide evidence of the potential association between human and food animal interfaces based on MDR phenotypes.

Co-occurrence analysis in the clinical S. aureus phenotypic dataset identified three distinct clusters of Ami-Mac-Lin (correlation coefficient > 0.75), Ami-Mac-Flu-Lin, and PSPen-Ami-Flu-Lin (correlation coefficient > 0.5, Fig. 7C). The three clusters were all presented in clinical and food animal AMR profiles (24.37%/0.29%, 9.24%/0.29%, 5.04%/0.57%, Table 3). In addition, three pairs of co-occurrence AMR phenotypes were identified in clinical S. aureus isolates, with only PLPen-PSPen being found in both clinical (34.35%) and food animal (0.86%) interfaces (Table 3). The AST for S. aureus from food animals incorporated three more classes of β-Lactam combination agents, cephems, and diterpenes. Thus, although three clusters of PSPen-βLac-Cep, PSPen-βLac-Cep-Mac and PSPen-βLac-Cep-Flu-Lin-Dit were identified in phenotypic dataset of food animal S. aureus isolates, only PSPen-Mac (adjusted from PSPen-βLac-Cep-Mac, 31.09%/0.29%) and PSPen-Flu-Lin (adjusted from PSPen-βLac-Cep-Flu-Lin-Dit, 10.08%/0.57%) can be used to determine their presence or absence in both interfaces (Fig. 7D, Table 3). Furthermore, eight pairs of co-occurrence AMR phenotypes were identified in the food-animal phenotypic dataset, and 6 of these pairs were observed in both clinical and food-animal interfaces. The results obtained from S. aureus also provided insights into the potential use of MDR as a cross-interface association assessment tool for G+ bacteria.

Nowadays, AMR poses a significant challenge to global public health and clinical treatment, impacting nations of various economic statuses and geographical locations [1–3]. Meanwhile, AMR is a multifaceted issue encompassing interactions between humans, animals, and the environment [24, 25, 32]. Investigating the relationship between different interfaces, particularly utilizing common bacterial resistance phenotypes, is imperative and necessitates immediate attention. Here, we analyzed surveillance data on resistance phenotypes of E. coli and S. aureus from a primary hospital and surrounding farms of food animals on Chongming Island. We sought to grasp the AMR status and characteristics of the clinical interface and explore the potential association between AMR phenotypes of food animals and human populations on the same island. This study is the prevenient attempt to explore associations across different interfaces using AMR phenotype data within the framework of One Health.

The study revealed that the E. coli isolates exhibited a considerable resistance rate (>50%) to penicillins, cephems, fluoroquinolones, and tetracyclines, while the S. aureus isolates demonstrated a notable resistance rate (>40%) to penicillins, macrolides, and lincosamides. The results align with the resistance patterns observed in Shanghai and various other provinces throughout the country [13]. The high and stable resistance rates are primarily attributed to the widespread use of specific antimicrobial agents, including penicillin and cephalosporin, which are consistently preferred by China's public healthcare institutions, as detailed in Yang's research [33]. In recent years, there have been reports of G- bacteria exhibiting resistance to carbapenems, colistin, and tigecycline [23, 34, 35], and clinical S. aureus isolates have emerged with varying levels of resistance to vancomycin and linezolid [36, 37]. Our study reveals that clinical E. coli and S. aureus isolates exhibited a low resistance rate or were susceptible to these drugs, indicating their continued high effectiveness in Chongming Island. Furthermore, the zero-resistance rate to linezolid of S. aureus in our results was consistent with the previous research result of almost zero-resistance to this antibiotic worldwide [38]. However, the surveillance to those drugs should be strengthened as constantly occurring mutation, the spreading of isolate clones and intraspecific or interspecific horizontal transfer of resistance genes.

ESBL-producing bacteria emerged in the 1980s, and the high prevalence of ESBL-producing E. coli in the community contributed to their nosocomial emergence [39]. A recent global surveillance database showed the prevalence of ESBL-producing E. coli in China has been remarkably high at 60-70%, being significantly higher than that of Europe (15%), North America (10%), Southeast Asia, and East Asia (20-40%) [40]. Our results showed the prevalence of 53.26% for ESBL-producing isolates in Chongming Island, also demonstrating a high prevalence rate. Moreover, our analysis and previous research have consistently demonstrated that the global spread of ESBL-producing E. coli remains ongoing.

MRSA has emerged as a prevalent global resistant phenotype since 1961 [17]. According to China's national AMR report, the MRSA prevalence of Shanghai in 2022 was 43.8%, higher than the national average of 28.9%. The result of Chongming was consistent with the prevalence in Shanghai [13], being significantly higher than some developed countries with published data, such as Germany (7.6%) [41], Australia (17.65%) and Japan (35.85%), but were lower than that of African countries like Egypt (63%) [42] and Nigeria (46%) [43]. Encouragingly, the surveillance data revealed that the prevalence of MRSA have steadily decreased in Chongming and Shanghai over the past 5 years.

Our research has revealed a significant proportion of MDR E. coli and S. aureus, as indicated by several studies [44, 45]. During the 5 years, the MDR rate of E. coli showed an increasing pattern, whereas the MDR rate of S. aureus was no obvious trend of change. The generalized linear mixed models were used to evaluate the MDR changes of the two bacterial species, and the findings provided additional support for our above conclusion. Afterwards, a thorough analysis was performed on the 294 AMR profiles of E. coli and the 119 AMR profiles of S. aureus. The results showed that AMR profiles with antibiogram length exceeding 3 became more prevalent after 2020, especially those profiles with antibiogram length exceeding 6. The gradual buildup of AMR and higher diversity of AMR profiles in these bacterial species leads to a more serious situation of MDR and increases the likelihood of transitioning to extensive-drug resistance and pan-drug resistance. Nevertheless, the exact cause of this accumulation, whether it results from the coexistence of multiple plasmids or other genetic elements, or it results from multiple resistance genes on a single plasmid, remains uncertain. A co-occurrence analysis previously reported was conducted to infer if the MDR phenotype occurs on a single plasmid [46, 47]. The phenotypic dataset from E. coli showed the presence of 13 pairs of co-occurring phenotypes and a cluster of three phenotypes. Similarly, the dataset of S. aureus displayed 3 pairs of co-occurring phenotypes, 1 cluster of three phenotypes, and 2 clusters of four phenotypes. Notably, it is essential to employ genome sequence analysis to unambiguously validate these findings, thereby highlighting the importance of incorporating whole genome sequencing into AMR surveillance.

It is imperative to tackle the pressing challenge of assessing the association between various interfaces within the One Health surveillance framework of AMR without high-throughput sequencing data [48]. Our results indicated that it is difficult to discover the association between AMR in food animals and medical institutions using resistance rate of each specific antimicrobial class or agent. Subsequently, it was discovered that some of identical AMR profiles existed on both interfaces of human and food animals. A more significant point is that some AMR profiles manifest in food animal interface before clinical settings, potentially confirming once more that animals serve as a source of certain AMRs and highlighting the potential for AMR transmission between humans and animals; however, determining the association between the two interfaces and the AMR transmission remains challenging based on the current dataset. This is another important reason why we call for the prompt integration of whole genome sequencing into AMR surveillance, an initiative also actively explored and advocated by WHO [49, 50]. Currently, only a few countries have incorporated bacterial genome data into routine AMR surveillance. The United States has been a pioneer in incorporating genomic data into AMR monitoring since 2014, and the National Antimicrobial Resistance Monitoring System (NARMS) serves as a multi-sectoral, multidisciplinary, and multi-scenario monitoring framework of One Health [51–53]. As sequencing technology improves and becomes more affordable, an increasing number of countries will embrace this trend [54]. Our study provides a methodological framework for investigating the associations between various interfaces in One Health network using phenotypic data. Meanwhile, a series of MDR phenotypes that co-existing at both interfaces of human and food animals were identified, including some with a high probability of co-occurrence. However, genomic data must be rapidly integrated into the system to fulfill the monitoring requirements of the One Health framework.

AMR phenotypic data continues to be the primary form of surveillance data. Nevertheless, there is an urgent requirement for a more comprehensive and integrated surveillance system within the One Health framework. In our study, we first utilized phenotypic data of clinical E. coli and S. aureus to analyze the characteristics and trends of AMR on Chongming Island. We subsequently expanded our investigation to examine the potential association of AMR at the human-animal interface by incorporating phenotype data from food animals into the pooled dataset. The findings indicated that the MDR situation among G+ and G- bacteria in Chongming Island's clinical scenario was worsening, characterized by an increase in the diversity of AMR profiles accompanied by longer antibiogram lengths. Some MDR profiles manifest in food animal interface before clinical settings, highlighting the potential for AMR association and AMR transmission between humans and food animals. Meanwhile, several co-occurrence phenotypes that could serve as indicator tools for AMR association were identified in both human and food animal interfaces. Our findings and subsequent work indicate the pressing need for and importance of incorporating genomic data into AMR monitoring system.

Ethics approval and consent to participate

The experimental protocol was established, according to the ethical guidelines of the Helsinki Declaration and was approved by the Human Ethics Committee of Chongming Hospital Affiliated to Shanghai University of Medicine and Health Sciences (CMEC-2021-KT-15). Written informed consent was obtained from individual or guardian participants.

Consent for publication

Not applicable.

Availability of data and material

Data available within the article or its supplementary materials.

Competing interests

The authors declare that they have no competing interests.

Funding

This research was funded by a grant from the National Natural Science Foundation of China (32170141)

Authors' contributions

Sample collection and processing: JL, MJQ, HPY, BQS, JS, LZ. Laboratory work: JL, MJQ, HPY, BQS, YZ. Data analysis and figure drawing: CL, JWH, ML, NZ, ZLC, YWC, WYC. Manuscript writing and editing: CL, JL, XKG, JS, LZ, YZZ. Conceptualization and design: JS, LZ, YZZ. Final approved manuscript: All

Acknowledgments

The authors acknowledge Prof. Xiao-Nong Zhou and Prof. Jing Xu for their guidance on the data analysis.

Rehman S. A parallel and silent emerging pandemic: Antimicrobial resistance (AMR) amid COVID-19 pandemic. J Infect Public Health. 2023;16 (4):611–7.
Theuretzbacher U. Accelerating resistance, inadequate antibacterial drug pipelines, and international responses. Int J Antimicrob Agents. 2012;39:295–9.
WHO. Antimicrobial resistance. https://www.who.int/news-room/fact-sheets/detail/antimicrobial-resistance. [Cited 28 Jan 2024]
WHO. Ten health issues WHO will tackle this year. https://www.who.int/news-room/spotlight/ten-threats-to-global-health-in-2019. [Cited 28 Jan 2024].
Antimicrobial Resistance Collaborators. Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. Lancet. 2022;399 (10325):629–55.
Hao R, Zhao R, Qiu S, Wang L, Song H. Antibiotics crisis in China. Science. 2015;348 (6239):1100–1.
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 U S A. 2015;112 (18):5649–54.
Tiseo K, Huber L, Gilbert M, Robinson TP, Van Boeckel TP. Global Trends in Antimicrobial Use in Food Animals from 2017 to 2030. Antibiotics (Basel). 2020;9 (12):918.
Mulchandani R, Wang Y, Gilbert M, Van Boeckel TP. Global trends in antimicrobial use in food-producing animals: 2020 to 2030. PLOS Glob Public Health. 2023;3 (2):e0001305.
Kaper JB, Nataro JP, Mobley HL. Pathogenic Escherichia coli. Nat Rev Microbiol. 2004;2 (2):123–40.
Conlon BP. Staphylococcus aureus chronic and relapsing infections: Evidence of a role for persister cells: An investigation of persister cells, their formation and their role in S. aureus disease. Bioessays. 2014;36 (10):991–6.
Molineri AI, Camussone C, Zbrun MV, Suárez Archilla G, Cristiani M, Neder V, et al. Antimicrobial resistance of Staphylococcus aureus isolated from bovine mastitis: Systematic review and meta-analysis. Prev Vet Med. 2021;188:105261.
CARSS. 2022 National Bacterial Antimicrobial Resistance Surveillance Report (Brief Edition). https://www.carss.cn/Report/Details?aId=917. [Cited 29 Jan 2024].
Guo Y, Ding L, Yang Y, Han R, Yin D, Wu S, et al. Multicenter Antimicrobial Resistance Surveillance of Clinical Isolates from Major Hospitals - China, 2022. China CDC weekly. 2023; (52):1155-60.
Peirano G, Pitout JDD. Extended-Spectrum β-Lactamase-Producing Enterobacteriaceae: Update on Molecular Epidemiology and Treatment Options. Drugs. 2019;79 (14):1529–41.
Ramatla T, Mafokwane T, Lekota K, Monyama M, Khasapane G, Serage N, et al. “One Health” perspective on prevalence of co-existing extended-spectrum β-lactamase (ESBL)-producing Escherichia coli and Klebsiella pneumoniae: a comprehensive systematic review and meta-analysis. Ann Clin Microbiol Antimicrob. 2023;22(1):88.
Lee AS, de Lencastre H, Garau J, Kluytmans J, Malhotra-Kumar S, Peschel A, et al. Methicillin-resistant Staphylococcus aureus. Nat Rev Dis Primers. 2018;4:18033.
Ciapponi A, Bardach A, Sandoval MM, Palermo MC, Navarro E, Espinal C, et al. Systematic Review and Meta-analysis of Deaths Attributable to Antimicrobial Resistance, Latin America. Emerg Infect Dis. 2023;29 (11):2335–44.
Huang H, Wan P, Luo X, Lu Y, Li X, Xiong W, et al. Tigecycline Resistance-Associated Mutations in the MepA Efflux Pump in Staphylococcus aureus. Microbiol Spectr. 2023;11 (4):e0063423.
Dunn SJ, Connor C, McNally A. The evolution and transmission of multi-drug resistant Escherichia coli and Klebsiella pneumoniae: the complexity of clones and plasmids. Curr Opin Microbiol. 2019;51:51-6.
Hummel R, Tschäpe H, Witte W. Spread of plasmid-mediated nourseothricin resistance due to antibiotic use in animal husbandry. Journal of basic microbiology. 1986;26(8):461-6.
Fischer J, Rodríguez I, Schmoger S, Friese A, Roesler U, Helmuth R, et al. Escherichia coli producing VIM-1 carbapenemase isolated on a pig farm. J Antimicrob Chemother. 2012;67 (7):1793–5.
Liu YY, Wang Y, Walsh TR, Yi LX, Zhang R, Spencer J, et al. The 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):161–8.
Robinson TP, Bu DP, Carrique-Mas J, Fèvre EM, Gilbert M, Grace D, et al. Antibiotic resistance is the quintessential One Health issue. Trans R Soc Trop Med Hyg. 2016;110 (7):377–80.
Berg G, Rybakova D, Fischer D, Cernava T, Vergès M-CC, Charles T, et al. Microbiome definition re-visited: old concepts and new challenges. Microbiome. 2020;8 (1):103.
da Costa PM, Loureiro L, Matos AJF. Transfer of multidrug-resistant bacteria between intermingled ecological niches: the interface between humans, animals and the environment. Int J Environ Res Public Health. 2013;10 (1):278–94.
Banerjee S, van der Heijden MGA. Soil microbiomes and one health. Nat Rev Microbiol. 2023;21 (1):6–20.
Argudín MA, Deplano A, Meghraoui A, Dodémont M, Heinrichs A, Denis O, et al. Bacteria from Animals as a Pool of Antimicrobial Resistance Genes. Antibiotics (Basel). 2017;6 (2):12.
Castañeda-Barba S, Top EM, Stalder T. Plasmids, a molecular cornerstone of antimicrobial resistance in the One Health era. Nat Rev Microbiol. 2024;22 (1):18–32.
Liu SL, Wang XJ, Su XJ. Study on the construction of Chongming Ecological Island in Shanghai. Chin Nation Conditions Strength. 2020;09:56–9.[in Chinese]
Luo F, Lu GX, Shen XA, Lin YX. Countermeasures and measures for developing ecological animal husbandry in Chongming Island. Shanghai J Anim Husbandry Vet Med. 2013;06:52–3. [in Chinese]
Kim D-W, Cha C-J. Antibiotic resistome from the One-Health perspective: understanding and controlling antimicrobial resistance transmission. Exp Mol Med. 2021;53 (3):301–9.
Yang Y, Geng X, Liu X, Wen X, Wu R, Cui D, et al. Antibiotic Use in China’s Public Healthcare Institutions During the COVID-19 Pandemic: An Analysis of Nationwide Procurement Data, 2018-2020. Front Pharmacol. 2022;13:813213.
Potter RF, D’Souza AW, Dantas G. The rapid spread of carbapenem-resistant Enterobacteriaceae. Drug Resist Updat. 2016;29:30–46.
He T, Wang R, Liu D, Walsh TR, Zhang R, Lv Y, et al. Emergence of plasmid-mediated high-level tigecycline resistance genes in animals and humans. Nat Microbiol. 2019;4 (9):1450–6.
Cong Y, Yang S, Rao X. Vancomycin resistant Staphylococcus aureus infections: A review of case updating and clinical features. J Adv Res. 2020;21:169–76.
Crowe-McAuliffe C, Murina V, Turnbull KJ, Huch S, Kasari M, Takada H, et al. Structural basis for PoxtA-mediated resistance to phenicol and oxazolidinone antibiotics. Nat Commun. 2022;13 (1):1860.
Shariati A, Dadashi M, Chegini Z, van Belkum A, Mirzaii M, Khoramrooz SS, et al. The global prevalence of Daptomycin, Tigecycline, Quinupristin/Dalfopristin, and Linezolid-resistant Staphylococcus aureus and coagulase-negative staphylococci strains: a systematic review and meta-analysis. Antimicrob Resist Infect Control. 2020;9 (1):56.
Lob SH, Nicolle LE, Hoban DJ, Kazmierczak KM, Badal RE, Sahm DF. Susceptibility patterns and ESBL rates of Escherichia coli from urinary tract infections in Canada and the United States, SMART 2010-2014. Diagn Microbiol Infect Dis. 2016;85(4):459–65.
Chong Y, Shimoda S, Shimono N. Current epidemiology, genetic evolution and clinical impact of extended-spectrum β-lactamase-producing Escherichia coli and Klebsiella pneumoniae. Infect Genet Evol. 2018;61:185–8.
Rödenbeck M, Ayobami O, Eckmanns T, Pletz MW, Bleidorn J, Markwart R. Clinical epidemiology and case fatality due to antimicrobial resistance in Germany: a systematic review and meta-analysis, 1 January 2010 to 31 December 2021. Euro Surveill. 2023;28 (20):2200672.
Azzam A, Khaled H, Mosa M, Refaey N, AlSaifi M, Elsisi S, et al. Epidemiology of clinically isolated methicillin-resistant Staphylococcus aureus (MRSA) and its susceptibility to linezolid and vancomycin in Egypt: a systematic review with meta-analysis. BMC Infect Dis. 2023;23(1):263.
Ezeh CK, Eze CN, Dibua MEU, Emencheta SC. A meta-analysis on the prevalence of resistance of Staphylococcus aureus to different antibiotics in Nigeria. Antimicrob Resist Infect Control. 2023;12 (1):40.
Dunn SJ, Connor C, McNally A. The evolution and transmission of multi-drug resistant Escherichia coli and Klebsiella pneumoniae: the complexity of clones and plasmids. Curr Opin Microbiol. 2019;51:51–6.
Assis LM, Nedeljković M, Dessen A. New strategies for targeting and treatment of multi-drug resistant Staphylococcus aureus. Drug Resist Updat. 2017;31:1–14.
Muloi D, Kiiru J, Ward MJ, Hassell JM, Bettridge JM, Robinson TP, et al. Epidemiology of antimicrobial-resistant Escherichia coli carriage in sympatric humans and livestock in a rapidly urbanizing city. Int J Antimicrob Agents. 2019;54 (5):531–7.
Hassell JM, Ward MJ, Muloi D, Bettridge JM, Robinson TP, Kariuki S, et al. Clinically relevant antimicrobial resistance at the wildlife-livestock-human interface in Nairobi: an epidemiological study. Lancet Planet Health. 2019;3(6):e259-e269.
Muloi DM, Jauneikaite E, Anjum MF, Essack SY, Singleton DA, Kasudi MR, et al. Exploiting genomics for antimicrobial resistance surveillance at One Health interfaces. Lancet Microbe. 2023;4 (12):e1056–62.
Jauneikaite E, Baker KS, Nunn JG, Midega JT, Hsu LY, Singh SR, et al. Genomics for antimicrobial resistance surveillance to support infection prevention and control in health-care facilities. Lancet Microbe. 2023;4(12):e1040–6.
Baker KS, Jauneikaite E, Nunn JG, Midega JT, Atun R, Holt KE, et al. Evidence review and recommendations for implementing genomics for antimicrobial resistance surveillance: reports from an international expert group. Lancet Microbe. 2023;4(12):e1035–9.
USDA. FSIS NARMS Multi-Year Report – 2014-2019 | Food Safety and Inspection Service. http://www.fsis.usda.gov/news-events/publications/fsis-narms-multi-year-report-2014-2019. [Cited 29 Jan 2024].
Wu Bo, Yan Fei, Fu Chaowei. Comparison of bacterial resistance surveillance and governance between China and the United States. China Health Resources, 2021,24(05):615-618. [in Chinese]
Pan H, Li XL, Fang W, Yue Min. Analysis of significant human and foodborne pathogens and their resistance to antimicrobials in the USA in the past two decades: Implications for surveillance and control of antimicrobial resistance in China. J Zhejiang University (Agriculture & Life Sciences), 2018, 44(2): 237-246. [in Chinese]
Argimón S, Masim MAL, Gayeta JM, Lagrada ML, Macaranas PKV, Cohen V, et al. Integrating whole-genome sequencing within the National Antimicrobial Resistance Surveillance Program in the Philippines. Nat Commun. 2020;11 (1):2719.

Table 1 The antimicrobial classes and antimicrobial agents involved in the clinical E. coli and S. aureus antimicrobial susceptibility test.

E. coli		S. aureus
Antimicrobial classes	Antimicrobial agents	Antimicrobial classes	Antimicrobial agents
Penicillins (Pen)	Ampicillin	Penicillinase-labile Penicillins (PLPen)	Penicillin
β-Lactam combination agents (βLac)	Ampicillin/Dsulbactam	Penicillinase-stable Penicillins (PSPen)	Oxacillin
	Cefoperazone/Sulbactam	Glycopeptides (Gly)	Vancomycin
	Piperacillin/Tazobactam	Aminoglycosides (Ami)	Gentamicin
	Amoxicillin/Clavulanic acid**	Macrolides (Mac)	Erythromycin
Cephems (Cep)	Cefazolin	Tetracyclines (Tet)	Tetracycline
	Cefepime		Minocycline
	Cefotaxime		Tigecycline**
	Ceftazidime	Fluoroquinolones (Flu)	Levofloxacin
	Cefuroxime		Ciprofloxacin**
	Cefoxitin*		Moxifloxacin**
	Ceftriaxone**	Lincosamides (Lin)	Clindamycin
Carbapenems (Car)	Meropenem	Folate pathway antagonists (Fpa)	Trimethoprim/Sulfamethoxazole
	Imipenem	Ansamycins (Ans)	Rifampin
	Ertapenem**	Oxazolidinines (Oxa)	Linezolid
Aminoglycosides (Ami)	Amikacin	Phenicols (Phe)	Chloramphenicol**
	Gentamicin
	Tobramycin**
Tetracyclines (Tet)	Tetracycline**
	Tigecycline
Fluoroquinolones (Flu)	Levofloxacin
	Ciprofloxacin
	Moxifloxacin**
Folate pathway antagonists (Fpa)	Trimethoprim/Sulfamethoxazole
Fosfomycins (Fos)	Fosfomycin
Monobactams (Mon)	Aztreonam*
Phenicols (Phe)	Chloramphenicol**
Nitrofuran (Nit)	Nitrofurantoin**
Polymyxins (Pol)	Colistin**

* The antimicrobial agents were added to the AST schedule in 2020; ** The antimicrobial agents were added to the AST schedule in 2021. The characters in parentheses are our self-defined abbreviations.

Table 2 Results of a Poisson generalized linear mixed model examining the likelihood of antibiogram length within different years

Year	No. Isolates	Estimate	OR	SE	95% CI	Z score	P-value
E. coli
2018	590	Reference
2019	578	-0.08987	0.91	0.03596	0.85, 0.98	-2.499	0.0124
2020	516	0.08735	1.09	0.0359	1.02, 1.17	2.433	0.015
2021	555	0.27393	1.32	0.034	1.23, 1.41	8.058	<0.001
2022	386	0.46518	1.59	0.0359	1.48, 1.70	12.958	<0.001
S. aureus
2018	251	Reference
2019	198	-0.18199	0.83	0.06277	0.74, 0.94	-2.899	0.003739
2020	174	-0.09362	0.91	0.06452	0.80, 1.03	-1.451	0.146799
2021	217	-0.24664	0.78	0.06183	0.69, 0.88	-3.989	<0.001
2022	154	0.05942	1.06	0.0641	0.94, 1.20	0.927	0.353978

Table 3 The proportion of the co-occurrence cluster and pairwise phenotypes at the human-animal interface

Co-occurrence AMR phenotypes	E. coli			Co-occurrence AMR phenotypes	S. aureus
Co-occurrence AMR phenotypes	Proportion in clinic profiles (%)	Proportion in food animal profiles (%)	Proportion in shared profiles (%)	Co-occurrence AMR phenotypes	Proportion in clinic profiles (%)	Proportion in food animal profiles (%)	Proportion in shared profiles (%)
Clinic isolates				Clinic isolates
Pen-Flu	52.72 (155/294)	19.33 (23/119)	45.45 (20/44)	Ami-Tet	17.65 (21/119)	0.00 (0/34)	0.00 (0/12)
Pen-βLac^a	52.38 (154/294)	18.49 (22/119)	34.09 (15/44)	PLPen-PSPen	34.45 (41/119)	0.86 (3/34)	0.00 (0/12)
Pen-Phe	32.31 (95/294)	29.41 (35/119)	50.00 (22/44)	PLPen-Ans	17.65 (21/119)	-	-
Pen-Fpa^a	44.90 (132/294)	25.21 (30/119)	47.73 (21/44)	Ami-Mac-Lin	24.37 (29/119)	0.29 (1/34)	0.00 (0/12)
βLac-Flu	34.69 (102/294)	7.56 (9/119)	20.45 (9/44)	Ami-Mac-Flu-Lin	9.24 (11/119)	0.29 (1/34)	0.00 (0/12)
βLac-Fpa^a	29.25 (86/294)	10.92 (13/119)	25.00 (11/44)	PSPen-Ami-Flu-Lin	5.04 (6/119)	0.57 (2/34)	0.00 (0/12)
Ami-Flu	30.27 (89/294)	13.45 (16/119)	25.00 (11/44)	Food animal isolates		0.00 (/34)
Flu-Col	2.38 (7/294)	0.84 (1/119)	0.00 (0/44)	PSPen-Ami	15.13 (18/119)	0.57 (2/34)	0.00 (0/12)
Cep-Fos	23.47 (69/294)	-	-	PLPen-Mac	55.46 (66/119)	3.74 (13/34)	25.00 (3/12)
Fos-Phe	9.52 (28/294)	-	-	Ami-Lin	24.37 (29/119)	0.57 (2/34)	0.00 (0/12)
Tet-Phe	24.15 (71/294)	27.73 (33/119)	45.45 (20/44)	Flu-Lin	18.49 (22/119)	0.86 (3/34)	0.00 (0/12)
Ami-Fpa	26.19 (77/294)	21.01 (25/119)	27.27 (12/44)	Ami-Dit	-	0.57 (2/34)	-
Tet-Fpa	26.19 (77/294)	25.21 (30/119)	43.18 (19/44)	Lin-Dit	-	0.86 (3/34)	-
Pen-βLac-Cep-Mon^b	42.52 (125/294)	6.72 (8/119)	13.64 (6/44)	Mac-Flu	28.57 (34/119)	1.44 (5/34)	0.00 (0/12)
Food animal isolates				Mac-Lin	51.26 (61/119)	0.57 (2/34)	0.00 (0/12)
Flu-Phe	24.49 (72/294)	18.49 (22/119)	38.64 (17/44)	PSPen-βLac-Cep	-	0.86 (3/34)	-
Pen-Cep	65.65 (193/294)	14.29 (17/119)	27.27 (12/44)	PSPen-βLac-Cep-Mac^c	31.09 (37/119)	0.29 (1/34)	-
βLac-Ami	27.21 (80/294)	12.61 (15/119)	20.45 (9/44)	PSPen-βLac-Cep-Flu-Lin-Dit^d	10.08 (12/119)	0.57 (2/34)	-

The profiles shared by human and food animal isolates; b the proportion of the profile was obtained by the adjusted profile of Pen-βLac-Cep; c the adjusted profile of PSPen-Mac obtained the proportion of the profile; d the proportion of the profile was obtained by the adjusted profile of adjusted PSPen-Flu-Lin. Pen Penicillins, βLac β-Lactam combination agents, Cep Cephems, Ami Aminoglycosides, Tet Tetracyclines, Flu Fluoroquinolones, Fpa Folate pathway antagonists, Fos Fosfomycins, Mon Monobactams, Phe PheniPols, Nit Nitrofuran, PLPen Penicillinase-labile Penicillins， PSPen Penicillinase-stable Penicillins， Gly Glycopeptides， Mac Macrolides， Lin Lincosamides， Ans Ansamycins， Oxa Oxazolidinines.

No competing interests reported.

The potential associations of antimicrobial resistance in clinical Escherichia coli and Staphylococcus aureus across the human-animal interfaces in Chongming Island, Shanghai

Status:

Version 1

Abstract

Figures

Introduction

Materials and methods

Results

Discussion

Conclusion

Declarations

Reference

Tables

Additional Declarations

Supplementary Files

Status:

Version 1