Establishment of Epidemiological Resistance Cut-Off Values of Aquatic Aeromonas to Eight Antimicrobial Agents

The abuse of antibiotics in aquaculture has led to the increasing rate of antibiotic resistance of aquatic bacteria including Aeromonas, which is an increasing threat to environmental and human health. To date, no epidemiological cut-off values (COWT) for Aeromonas spp. have been established by the Clinical and Laboratory Standards Institute nor the European Commission on Antimicrobial Susceptibility Testing. In this study, commercially prepared minimum inhibitory concentration (MIC) test 96-well plates (dry-form plates) were used to determine the MIC of eight antimicrobial agents against 556 Aeromonas strains. The obtained MIC distributions were simulated and analyzed by NRI and ECOFFinder to obtain tentative COWT values for Aeromonas spp. The COWT values of eight kinds of representative antimicrobial agents including trimethoprim–sulfamethoxazole, erythromycin, doxycycline, neomycin, colistin, florfenicol, enrofloxacin, and ceftazidime for Aeromonas spp. were established and were 0.25, 64/32, 4/2, 8, 4, 1, 0.062/0.125, and 0.5 μg/mL, respectively. Results showed that Aeromonas spp. had a very high proportion of non-wild-type strains to enrofloxacin, florfenicol, and doxycycline, which are the most widely used antimicrobials in aquaculture. The COWT values for Aeromonas spp. obtained in this study can contribute to the final establishment of COWT for Aeromonas spp. internationally.


Introduction
Antimicrobial resistance (AMR) has become a growing threat to global public health [1]. The continuous use of antibiotics has led to the emergence of drug resistance in some bacteria. Due to the difficulty of clinical treatment and high mortality, it poses a huge threat to human health [2]. According to the World Health Organization, antibiotic resistance is one of the top 10 global health threats and one of the major global health challenges in the 21st century [3]. Every year, 700,000 people die from drug-resistant bacterial infections around the world. It is estimated that by 2050, bacterial resistance will cause 10 million deaths worldwide each year [4].
The aquatic environment may provide an ideal environment for the acquisition and spread of antibiotic resistance because it is closely related to human activities [5], such as aquaculture, wastewater discharge, hydraulic engineering, and agricultural measures. Recent studies have also shown that aquatic ecosystems are important reservoirs of antibiotic China is the most important country in the world for aquaculture, with high aquaculture production and complex use of antimicrobial drugs [19]. Although many scholars have investigated the AMR status of fish pathogenic bacteria in China, especially the pathogenic Aeromonas, the CBPs used in these studies mainly referred to ones of CLSI, which are unsuitable for the clinical treatment of fish diseases nor suitable for the analysis of the prevalence of AMR in aquaculture. And whether the CO WT values established in the above literature [12] are equally applicable to water bodies and the aquaculture environment in China needs to be verified.
Therefore, this study aimed to establish CO WT criteria of Aeromonas spp. to investigate and analyze the current AMR status of Aeromonas spp. in typical aquaculture ponds, a reservoir, and a natural river in China.

Isolation and Identification of Bacteria
From November 2020 to October 2021, sick and healthy fish, water, or sediment samples from fish ponds in Jingzhou City, Hubei Province; sick and healthy fish samples from fish ponds in Wuhan City, Hubei Province, and Xiashan reservoir in Weifang City, Shandong Province; and healthy fish samples from Liuchong River in Bijie City, Guizhou Province, were collected and transported to the laboratory through ice boxes to perform bacterial isolation. The clinical symptoms of the sick fish were swirling on the surface of the water body, multiple congestions on the body surface, body ulceration; yellowing of the body, increased mucus, swelling of the abdomen with ascites; split head, khaki liver; ulcerative erythema on the body surface; anal swelling and gently press on the abdomen that had yellow mucus or blood coming out of the anus, with a shriveled gut and congested gut walls. And fish without any of the above symptoms was judged as healthy fish. The intestinal, kidney, liver, spleen, brain, or blood of the sampled fish were streaked onto sterile ampicillin MacConkey agar (i.e., Aeromonas selection medium). The water sample or sediment sample diluted with 0.86% sterile normal saline was pipetted onto the sterile ampicillin MacConkey agar and spread evenly with a triangular glass rod, and all plates were incubated at 28 • C for 12-24 h. The pink single colonies were picked and cultivated overnight in a 1 mL BHI broth medium for bacterial identification.
The above bacterial suspension was used as the DNA template. The bacterial universal primer pair 27F (5 -AGAGTTTGATCCTGGCTCAG-3 ) and 1492R (5 -GGTTACCTTGTTAC GACTT-3 ) was used for 16S rRNA gene PCR amplification. Amplification conditions were as follows: 94 • C for 5 min at pretreatment; 30 cycles of 94 • C for 30 s, 55 • C for 30 s, 72 • C for 90 s, and final extension of 72 • C for 7 min. The PCR products were sent to Wuhan Aikangjian Biotechnology Co., Ltd. (Wuhan China) for sequencing. Sequence homology analysis (bacterial identification) was performed using the Nucleotide BLAST in NCBI (https://www.ncbi.nlm.nih.gov/) (accessed on 8 February 2022). The obtained Aeromonas bacterial suspensions were then stored in a −80 • C refrigerator with 20% glycerol.
The MIC was determined according to the instructions of the MIC assay plate. The specific method is as follows: pick fresh single colonies suspended in cation-adjusted Mueller-Hinton broth, adjust the concentration of the bacterial solution to 0.5 McFarland concentration, and then dilute at a ratio of 1:200 so that the inoculum concentration of the bacterial solution is about 5 × 10 5 CFU/mL. About 100 µL of this bacterial solution was added to each well of the MIC assay plate, and positive and negative controls were set up simultaneously. The plates were incubated at 28 • C for 24-28 h, and then MIC values were recorded. The quality control strain used was ATCC 25922, and the results were considered valid only if the MIC of the quality control strain was within the specified quality control range.

Determination of Tentative Epidemiological Cut-Off Values (CO WT )
CO WT values were determined using the free and automated EXCEL spreadsheet programs ECOFFinder (available online: https://clsi.org/meetings/microbiology/ecoffinder/) and NRI (available online: http://www.bioscand.se/nri/) (both were accessed on 11 February 2022). These two methods were proposed by Turnidge et al. [28] and Kronvall [29], respectively, which will be referred to as Turnidge method and Kronvall method hereinafter. The principle was to use the method of nonlinear regression analysis to fit the cumulative distribution of MICs, inputting the distribution of MICs according to the operation instructions, and the program will automatically fit and calculate the CO WT values of bacteria. The NRI program directly provides CO WT values, whereas the ECOFFinder program selects the MICs with a confidence interval of 95.0% or 97.5% as the final CO WT values. ECV and EOCFF are used for epidemiological cut-off values set by two international institutions, CLSI and EUCAST, respectively. To avoid confusion, the epidemiological cut-off values are abbreviated as CO WT in this study.

Calculation of MIC 50 , MIC 90 , and Resistance Rate (NWT%)
MIC 50 is the MIC required to inhibit 50% bacterial growth, and MIC 90 is the MIC required to inhibit 90% bacterial growth. The strains with MIC values greater than the breakpoint value of the 95% or 97.5% confidence interval are all NWT strains. The strains with MIC values less than or equal to the breakpoint value of the 95% or 97.5% confidence interval are WT strains. The resistance rate is equal to the ratio of the number of NWT strains to the total number of strains tested in the present study.

Aeromonas Diversity
A total of 556 strains of Aeromonas were assayed for their MIC, and the proportion of strains in four sampling areas was as follows: 369 strains in Jingzhou City, Hubei Province; 30 strains in Wuhan City, Hubei Province; 16 strains in Weifang City, Shandong Province; and 141 strains in Bijie City, Guizhou Province. The 556 Aeromonas strains were derived from 183 fish, 14 water samples, and 5 sediment samples. The specific source and quantity distribution of the strains are shown in Table 1. Among them, 14 species were identified at the species level, and the proportions are shown in Figure 1. Among them, A. veronii (n = 328, 59.0%), A. sobria (n = 75, 13.5%), A. hydrophila (n = 52, 9.4%), and A. allosaccharophila (n = 36, 6.5%) were very abundant, accounting for 88.3% of the total number of Aeromonas.

MIC Measurement Results
A total of four batches of MIC assay plates were used in this study. Quality control tests were performed for each batch using the standard strain ATCC 25922, and the results were all within the quality control range except for neomycin and erythromycin, for which no quality control range was available. The MIC distributions of 556 Aeromonas strains to eight drugs are shown in Table 2, and the MICs of seven drugs exceeded the dilution range of the drugs. Drugs with MIC lower than or equal to the lowest concentration determined included trimethoprim-sulfamethoxazole (≤0.062, n = 212), florfenicol (≤0.125, n = 11), enrofloxacin (≤0.031, n = 99), and ceftazidime (≤0.062, n = 56). Drugs with MIC higher than the highest concentration measured were trimethoprim-sulfamethoxazole, erythromycin, neomycin, colistin, florfenicol, and ceftazidime. Only one drug, doxycycline, had MICs that were always within the drug dilution range.

MIC Measurement Results
A total of four batches of MIC assay plates were used in this study. Quality control tests were performed for each batch using the standard strain ATCC 25922, and the results were all within the quality control range except for neomycin and erythromycin, for which no quality control range was available. The MIC distributions of 556 Aeromonas strains to eight drugs are shown in Table 2, and the MICs of seven drugs exceeded the dilution range of the drugs. Drugs with MIC lower than or equal to the lowest concentration determined included trimethoprim-sulfamethoxazole (≤0.062, n = 212), florfenicol (≤0.125, n = 11), enrofloxacin (≤0.031, n = 99), and ceftazidime (≤0.062, n = 56). Drugs with MIC higher than the highest concentration measured were trimethoprim-sulfamethoxazole, erythromycin, neomycin, colistin, florfenicol, and ceftazidime. Only one drug, doxycycline, had MICs that were always within the drug dilution range. MIC values equal to or lower than the lowest concentration tested are presented as the lowest concentration. The dilution range of the drug and the number of strain distribution contained in the white background was used for the formulation of CO WT values, and HR (%) represents the number (percentage) of strains beyond the drug test range.

Establishment of Tentative Epidemiological Cut-Off Values (CO WT )
The Kronvall and Turnidge method nonlinear regression fitting distributions of eight drugs against Aeromonas spp. are shown in Figures 2 and 3, respectively. As can be seen, the nonlinear regression fitting distribution results of the eight drugs against Aeromonas spp. obtained by the Kronvall and Turnidge methods were almost identical. The MIC distribution of enrofloxacin against Aeromonas spp. was very wide and multimodal, with two peaks, one large and one small, on the right side. The MICs for doxycycline and florfenicol were widely distributed and showed no obvious bimodal distribution. The MIC distributions for the remaining five drugs were narrow and unimodal. The CO WT values can be preliminarily observed with the naked eye, which is the upper limit of the 95% or 97.5% confidence interval of the abscissa of the fitting curve. Erythromycin had the largest CO WT value, followed by neomycin and colistin, doxycycline, florfenicol, ceftazidime, trimethoprim-sulfamethoxazole, and enrofloxacin.  MIC values equal to or lower than the lowest concentration tested are presented as the lowest concentration. The dilution range of the drug and the number of strain distribution contained in the white background was used for the formulation of COWT values, and HR (%) represents the number (percentage) of strains beyond the drug test range.

Establishment of Tentative Epidemiological Cut-Off Values (COWT)
The Kronvall and Turnidge method nonlinear regression fitting distributions of eight drugs against Aeromonas spp. are shown in Figures 2 and 3, respectively. As can be seen, the nonlinear regression fitting distribution results of the eight drugs against Aeromonas spp. obtained by the Kronvall and Turnidge methods were almost identical. The MIC distribution of enrofloxacin against Aeromonas spp. was very wide and multimodal, with two peaks, one large and one small, on the right side. The MICs for doxycycline and florfenicol were widely distributed and showed no obvious bimodal distribution. The MIC distributions for the remaining five drugs were narrow and unimodal. The COWT values can be preliminarily observed with the naked eye, which is the upper limit of the 95% or 97.5% confidence interval of the abscissa of the fitting curve. Erythromycin had the largest COWT value, followed by neomycin and colistin, doxycycline, florfenicol, ceftazidime, trimethoprim-sulfamethoxazole, and enrofloxacin.  The CO WT , MIC 50 , and MIC 90 (µg/mL) and NWT (%) of 556 Aeromonas strains against eight drugs are shown in Table 3. The CO WT of Aeromonas spp. against each drug obtained by the NRI or ECOFFinder table processing program was consistent with that observed with the naked eye. The CO WT values calculated by the Kronvall and Turnidge methods were the same for the five drugs, including trimethoprim-sulfamethoxazole, neomycin, colistin, florfenicol, and ceftazidime. The CO WT values calculated by the two methods for the other three drugs had only a difference of one step of drug dilution concentration. The similarity rate of CO WT obtained by the two methods could reach 62.5%. Using the CO WT established in this study to calculate the number (percentage) of NWT strains among all strains collected in the present study, NWT (%), it could be concluded that enrofloxacin had the highest proportion of NWT strains of Aeromonas, exceeding 50%, followed by florfenicol. The two drugs with lower proportions were ceftazidime and neomycin. The difference between the MIC 50 and MIC 90 of trimethoprim-sulfamethoxazole was the largest, and its MIC 90 exceeds the dilution range of the drug. The difference between the MIC 50 and MIC 90 of neomycin and ceftazidime was the smallest, with only two drug dilution gradients. The drug with the largest MIC 50 and MIC 90 was erythromycin, followed by enrofloxacin and doxycycline, and ceftazidime. The COWT, MIC50, and MIC90 (μg/mL) and NWT (%) of 556 Aeromonas strains against eight drugs are shown in Table 3. The COWT of Aeromonas spp. against each drug obtained by the NRI or ECOFFinder table processing program was consistent with that observed with the naked eye. The COWT values calculated by the Kronvall and Turnidge methods were the same for the five drugs, including trimethoprim-sulfamethoxazole, neomycin, colistin, florfenicol, and ceftazidime. The COWT values calculated by the two methods for the other three drugs had only a difference of one step of drug dilution concentration. The similarity rate of COWT obtained by the two methods could reach 62.5%. Using the COWT established in this study to calculate the number (percentage) of NWT strains among all strains collected in the present study, NWT (%), it could be concluded that enrofloxacin had the highest proportion of NWT strains of Aeromonas, exceeding 50%, followed by florfenicol. The two drugs with lower proportions were ceftazidime and neomycin. The difference between the MIC50 and MIC90 of trimethoprim-sulfamethoxazole was the largest, and its MIC90 exceeds the dilution range of the drug. The difference between the MIC50 and MIC90 of neomycin and ceftazidime was the smallest, with only two drug dilution gradients. The drug with the largest MIC50 and MIC90 was erythromycin, followed by enrofloxacin and doxycycline, and ceftazidime. COWT K/T represents the epidemiological cut-off values calculated by the Kronvall method or Turnidge method, the numbers in square brackets represent the breakpoints of the same drug studied by Baron et al. [12], no square brackets represent nothing can refer to. NWT (%) indicates the   Table 4. The CO WT of A. veronii was the closest to that of Aeromonas spp., followed by A. sobria. The CO WT of the four Aeromonas spp. differed for the eight drugs. A. veronii obtained the highest CO WT similarity rate by Kronvall and Turnidge methods (up to 87.5%), and A. allosaccharophila obtained the lowest CO WT similarity rate by the two methods. The largest difference between MIC 50 and MIC 90 was trimethoprim-sulfamethoxazole of A. veronii, and the smallest difference was trimethoprim-sulfamethoxazole of A. sobria. The MIC 50 and MIC 90 of eight drugs against A. sobria were smaller. For these four Aeromonas spp., the MIC 50 of erythromycin was relatively high, whereas those of trimethoprim-sulfamethoxazole, ceftazidime, and enrofloxacin were relatively low. n is the number, CO WT K/T is the epidemiological cut-off values calculated by the Kronvall method or the Turnidge method, and only one value means the CO WT values obtained by the two methods are the same. The numbers in square brackets represent the breakpoints of the same drug studied by Baron et al. [12], the numbers in parentheses represent the CLSI breakpoints, and the absence of square brackets or parentheses means the reference breakpoints are not available. The asterisk indicates SD exceeds the limit of the NRI program.

Comparative Analysis of the Resistance Rate of Aeromonas spp. in Different Regions
Using the above obtained CO WT K and CO WT T of Aeromonas spp. against eight drugs, the resistance rate was calculated according to the MIC distributions of Aeromonas spp. against eight drugs in each region. The results are shown in Figures 4 and 5. Aeromonas spp. in culture ponds in Jingzhou City, Hubei Province, had high resistance rates to enrofloxacin, florfenicol, and trimethoprim-sulfamethoxazole. Aeromonas spp. in culture ponds in Wuhan City, Hubei Province, had high resistance rates to enrofloxacin, florfenicol, erythromycin, and doxycycline, but zero resistance to ceftazidime. Aeromonas spp. in the two ponds had the highest resistance rate to enrofloxacin (up to 82.7%), followed by florfenicol. The resistance rates of Aeromonas spp. in Xiashan Reservoir of Shandong Province to five drugs including trimethoprim-sulfamethoxazole, erythromycin, neomycin, florfenicol, and ceftazidime were 0, but the resistance rates of Aeromonas spp. to enrofloxacin and colistin were high. Aeromonas spp. in Liuchong River, Guizhou Province, had a high resistance rate to colistin, but their resistance rates to the other seven drugs were all low.

Discussion
The establishment of resistance breakpoints is influenced by different laboratories in terms of assay protocols, reagents, and human factors, and the results of the established breakpoints may vary greatly. Therefore, only the breakpoints established following the standard experimental protocols, such as that of CLSI, can be widely accepted and recognized. Many researchers have used Aeromonas as indicator bacteria for antimicrobial susceptibility in aquatic environments [8,30,31]. Aeromonas spp. are ubiquitous and pathogenic to many aquatic animals and can be isolated in freshwater systems at any time of the year [8]. A. hydrophila, A. veronii, A. sobria, and many other Aeromonas species are the most common pathogenic bacteria in aquaculture in China and have a negative impact on the Chinese aquaculture industry [32]. Therefore, we chose Aeromonas spp. as the first aquatic bacterial species to establish the resistance threshold. This is the first study of this type conducted using Chinese isolates.
In this study, Aeromonas were isolated from fish collected from a natural non-aquaculture river and aquaculture ponds in China. The bacterial strains from these two sources are defined as environmental or pathogen strains. The protocols reported in the literature [12] were used to establish the COWT values for Aeromonas. The results obtained in the present study can be used for comparison with those reported in the literature [12] to ver-

Discussion
The establishment of resistance breakpoints is influenced by different laboratories in terms of assay protocols, reagents, and human factors, and the results of the established breakpoints may vary greatly. Therefore, only the breakpoints established following the standard experimental protocols, such as that of CLSI, can be widely accepted and recognized. Many researchers have used Aeromonas as indicator bacteria for antimicrobial susceptibility in aquatic environments [8,30,31]. Aeromonas spp. are ubiquitous and pathogenic to many aquatic animals and can be isolated in freshwater systems at any time of the year [8]. A. hydrophila, A. veronii, A. sobria, and many other Aeromonas species are the most common pathogenic bacteria in aquaculture in China and have a negative impact on the Chinese aquaculture industry [32]. Therefore, we chose Aeromonas spp. as the first aquatic bacterial species to establish the resistance threshold. This is the first study of this type conducted using Chinese isolates.
In this study, Aeromonas were isolated from fish collected from a natural non-aquaculture river and aquaculture ponds in China. The bacterial strains from these two sources are defined as environmental or pathogen strains. The protocols reported in the literature [12] were used to establish the CO WT values for Aeromonas. The results obtained in the present study can be used for comparison with those reported in the literature [12] to verify the reproducibility of the method and to determine whether CO WT values are influenced by the geographical origin of the bacterial strains. Additionally, it can provide breakpoints for determining the epidemiological characteristics of Aeromonas AMR to antimicrobials of greatest concern to the Chinese aquaculture industry. Commercially prepared MIC test 96-well plates containing dehydrated antimicrobial agents (dry-form plates) were used in this study, this standardized method made the experimental results more accurate and reproducible. During the actual operation with NRI and ECOFFinder, the MICs of the 95% and 97.5% confidence intervals of ECOFFinder were the same, and most of the CO WT values obtained by selecting the 95% or 97.5% confidence intervals of ECOFFinder were the same or similar to the CO WT values obtained by NRI.
The CO WT obtained in this study was compared with the results reported in CLSI [17]. The results showed that the CO WT T values of A. hydrophila to erythromycin and florfenicol were consistent with the corresponding results of CLSI, and other CO WT T values or CO WT K values were similar to the corresponding result of CLSI. The maximum difference did not exceed two drug dilution gradients. The CO WT obtained in this study was compared with the results reported by Baron et al. [12]. The CO WT T values and CO WT K values of Aeromonas spp. to trimethoprim-sulfamethoxazole and ceftazidime were exactly the same with the corresponding results reported by Baron Baron et al., except for the CO WT T of A. sobria to enrofloxacin. The maximum difference did not exceed two drug dilution gradients. In conclusion, the CO WT T values obtained in this study were closer to the corresponding results reported by CLSI and Baron et al. In addition, the results obtained in this study are the same or similar to those reported by CLSI and Baron et al., which indicates that the experimental methods referenced in this study are reliable and credible. On the other hand, it also shows that if a standard set of experimental methods for establishing breakpoints is used to establish CO WT values, the CO WT values obtained are generally the same or similar and are not greatly affected by regions.
In general, the MIC distribution of the drug against the strain presents a unimodal or bimodal distribution. For a typical bimodal distribution diagram, the right peak represents the MIC distribution of NWT strains, and the left peak represents the MIC distribution of WT strains. If the MIC distribution of a certain drug to the strain has an obvious bimodal distribution, CO WT can be observed with the naked eye, that is, the upper limit of the MIC distribution of the WT strain. If the MIC distribution of the strain exhibits insignificant bimodal distribution or continuous multimodal characteristics, the fitting analysis of the MIC distributions can be performed with ECOFFinder or NRI to obtain more accurate CO WT values [28]. In this study, the results showed that the MIC of enrofloxacin on Aeromonas spp. had a wide range and a multimodal distribution, indicating the existence of a higher proportion of enrofloxacin-resistant strains. Aeromonas spp. indeed had the highest resistance rate to enrofloxacin, as high as 61.5%/53.8%. The MICs of doxycycline and florfenicol against Aeromonas spp. were widely distributed and showed an insignificant bimodal distribution, indicating that a certain proportion of strains were resistant to doxycycline and florfenicol. The resistance rate of Aeromonas spp. to doxycycline and florfenicol was higher (9.3%/20.5% and 22.7%, respectively). This study found that the resistance rates of Aeromonas spp. collected from Xiashan reservoir and Liuchong river were low for most drugs compared with those of Aeromonas strains collected from aquaculture ponds, which may be related to the fact that these two natural water bodies have almost no drugs present. However, the resistance rates of these two water bodies to colistin or enrofloxacin were high. This was probably due to the surrounding industrial and agricultural production wastewater and domestic pollution. Aeromonas spp. in culture ponds had higher resistance rates to enrofloxacin, florfenicol, doxycycline, and trimethoprim-sulfamethoxazole. More importantly, enrofloxacin, doxycycline, and florfenicol are commonly used antibacterial drugs in aquaculture [22,26,27]. Therefore, more attention should be paid to the resistance changes of Aeromonas to these three kinds of drugs, and these antibacterial drugs should be rationally used in aquaculture practice to avoid the outbreak of drug-resistant strains caused by the abuse of antibiotics.
Compared with previous studies [12,[33][34][35], the MIC 50 or MIC 90 of Aeromonas for the same drug in different studies was mostly different or quite different. This may be related to different human factors such as drug use, artificial breeding, and water pollution in different regions. However, the MIC 50 and MIC 90 of Aeromonas spp. and A. sobria to ceftazidime in this study was the same as the corresponding results of Baron et al. [12], and the resistance rates of this study and research of Baron et al. for Aeromonas spp. to ceftazidime were low, at 4.1% and 9.4%, respectively. This may be related to the fact that ceftazidime was less used in these two regions, and most of the existing Aeromonas were naturally sensitive to ceftazidime. The MIC 50 and MIC 90 of eight drugs against Aeromonas spp. obtained in this study can provide a theoretical reference for the usage and dosage of drugs in aquaculture.
The CLSI [36] document stipulates that more than 100 strains are required to establish the CO WT values of bacteria of a species, whereas at least 500 bacteria need to be collected to establish the CO WT values of a genus. However, the CLSI [37] document and Smith and Kronvall [38] suggested that only 30 strains may be required to statistically establish bacterial CO WT values. The number of strains collected in this study met the above requirements.
The CO WT values of Aeromonas for a variety of antimicrobial agents obtained in this study still need to be compared and integrated with data obtained by other laboratories to obtain internationally accepted CO WT values for Aeromonas [39], which will facilitate the surveillance and epidemiological study of AMR in aquaculture environments.

Conclusions
In this study, the optimized micro-broth dilution method was used for the first determination of the MIC of eight drugs including trimethoprim-sulfamethoxazole, erythromycin, doxycycline, neomycin, colistin, florfenicol, enrofloxacin, and ceftazidime against aquatic Aeromonas. The tentative epidemiological cut-off values (CO WT values) of these eight drugs against Aeromonas spp. were established by ECOFFinder and NRI methods, which were 0.25, 64/32, 4/2, 8, 4, 1, 0.062/0.125, and 0.5 µg/mL, respectively. Using these breakpoints as the standard, the proportion of NWT strains of Aeromonas spp. in different water bodies in China was analyzed. The results showed that Aeromonas spp. had a higher proportion of NWT strains against enrofloxacin, doxycycline, and florfenicol. Therefore, it is necessary to focus on the resistance changes of Aeromonas to these three drugs in the aquaculture environment and to conduct early monitoring of the resistance of Aeromonas to avoid the spread of drug-resistant strains.