Occurrence of Antimicrobial Resistance in Canine and Feline Bacterial Pathogens in Germany under the Impact of the TÄHAV Amendment in 2018

The occurrence of antimicrobial resistance due to the use of antimicrobials is considered to be a main cause for treatment failure of bacterial infections in humans and animals. The right of German veterinarians to use and prescribe medications such as antimicrobials is regulated by the Regulation of Veterinary Pharmacies (TÄHAV). The aim of this study was to investigate the impact of the second amendment to the TÄHAV in 2018 on the occurrence of antimicrobial resistance in selected bacterial pathogens isolated from dogs and cats in Germany. For this purpose, we analyzed antimicrobial susceptibility data from 38 German small animal practices gathered between 2015 and 2021 in cooperation with Laboklin (Labor für klinische Diagnostik GmbH & Co.KG, Bad Kissingen, Germany). Annual cumulative susceptibility data of eight bacterial species were analyzed and compared. The mean value of resistant isolates was determined for each year and supplemented by 95% confidence intervals. Encouraged by the amendment, an increase in sample submissions was observed in Germany. The highest resistance rates to the analyzed substances penicillin G, ampicillin, amoxicillin-clavulanic acid, cefovecin, and enrofloxacin were found for Staphylococcus pseudintermedius (S. pseudintermedius), S. aureus, and Escherichia coli (E. coli). In contrast, resistance rates were low for Pasteurella multocida (P. multocida) and β-hemolytic streptococci. Significant resistance trends (p < 0.05) assumed as influenced by the TÄHAV amendment could be the significant decreases in resistance rates of S. pseudintermedius against penicillin G to 67% (n = 322/479), and ampicillin to 63% (n = 286/453), as well as S. felis against amoxicillin-clavulanic acid and cefovecin to 2% (n = 2/109), furthermore, the reduction in the occurrence of resistance of S. aureus against enrofloxacin to 4% (n = 3/76) in 2021. Moreover, for all species, the efficacy against the analyzed substances was maintained over the study period.


Introduction
Antimicrobials are an essential part of the successful treatment of bacterial infections [1], and the loss of their efficacy due to antimicrobial resistance is currently one of the greatest global threats to public health [2,3]. Any use of antimicrobials creates a selective pressure resulting in the emergence of resistance [2,4]. The foundation for containing the spread of resistance is to optimize the use of antimicrobials [2,[4][5][6]. In line with the One Health approach, the use of antimicrobials in veterinary medicine is also a relevant factor in the global development of resistance. This is particularly relevant because antimicrobials used to treat companion animals are almost identical to those used in human medicine [7] and the intensive physical contact between pet and owner enhances the possibility of transmission of drug resistant microorganisms [8][9][10][11]. In this context, 3rd/4th generation cephalosporins and fluoroquinolones have been classified as highest priority critically important antimicrobials (HPCIA) to human health, as their efficiency is without alternative in the treatment of certain human diseases [4,12]. To reduce the use of HPCIA, it is particularly indicated to use the support of bacterial diagnostics and antimicrobial susceptibility testing (AST) to guide the selection of appropriate antimicrobials for therapy [2,4,6,13].
In Germany, the permission to dispense and prescribe medications is regulated in the Regulation on Veterinary Pharmacies (Verordnung über Tierärztliche Hausapotheken (TÄHAV)). This regulation was revised in 2018 with the aim to enhance the prudent use of antimicrobials in order to maintain the efficacy of antimicrobial agents. Therefore, AST, performed according to internationally recognized procedures, is mandatory for the use of 3rd/4th generation cephalosporines and fluoroquinolones ( §12c, d TÄHAV). While these innovations provide veterinarians with a basis for good scientific practice, they are often perceived as limiting therapeutic options [14].
Standardized resistance data could be used to develop further methods to reduce the risk of resistance selection and to evaluate the impact of these interventions on antimicrobial resistance as well as animal health and welfare [2,7,15,16]. However, the results of AST from commercial laboratories have currently not been used for a nationwide scientific evaluation in Germany.
The aim of this study was to analyze the occurrence of antimicrobial resistance in companion animal pathogens in Germany under the impact of the TÄHAV amendment in 2018.

Results
Data from 38 practices were included. Eighteen participants were classified as small practices with one to two veterinarians employed and a daily patient load of less than 30. Of these, seven practices were located in small towns, two were in cities with populations of 20,000-100,000 (medium-sized cities), and nine were located in large cities. Large veterinary practices employed more than 3 veterinarians and treated more than 30 patients daily. These practices were also located in small towns (8), in medium-sized towns (3), and large cities (5). The number of small and large practices was evenly distributed across 11 of 16 German states (Schleswig-Holstein, Mecklenburg-Western Pomerania, Brandenburg, Berlin, Saxony, Bavaria, Baden-Württemberg, Rhineland-Palatinate, Hesse, North Rhine-Westphalia, Lower Saxony). In addition, four clinics participated in the data analysis, two of which were in medium-sized cities and two in large cities. They were located in four different German states (Bavaria, Rhineland-Palatinate, Saxony, North Rhine-Westphalia).
During the previous survey [17], 79% (n = 30/38) of the current participants stated to use HPCIA less frequently due to the TÄHAV amendment. In addition, 66% (n = 25/38) reported using penicillins generally as initial therapy due to the revision. AST were requested more frequently by 76% (n = 29/38) of participants, but 41% (n = 11/27) reported testing mostly in parallel with antibiotic therapy and only 15% (n = 4/27) of participants actually waited for the test result to select an appropriate antimicrobial agent.

Antimicrobial Resistance of Staphylococcus felis
The highest resistance rates were observed for penicillin G and ampicillin. At the beginning of the study, 72% (n = 34/47) of isolates were evaluated as resistant to penicillin G and 59% (n = 27/47) as resistant to ampicillin. A significant (p < 0.05) decrease in resistance was observed for both agents until 2017 (39%, n = 25/65). A slight non-significant decrease in resistance was also observed in the following years (2020 30%, n = 38/127). Lower rates of resistance were identified to amoxicillin-clavulanic acid as well as cefovecin. While in 2015, 23% (n = 11/47) and 28% (n = 13/47) of isolates were resistant against amoxicillinclavulanic acid and cefovecin, a significant reduction (p < 0.05) of resistant isolates was found for both agents in 2016 (8%, n = 5/65 and 6%, n = 4/65, respectively). A further significant decrease (p < 0.05) in resistance for those two agents was found in 2021 (2%, n = 2/109). A similar finding was shown for resistance to enrofloxacin. A significant decrease

Antimicrobial Resistance of Canine E. coli Isolates from the Urinary Tract (UTI)
The highest resistance rates were found against ampicillin. The percentage of resistant bacteria varied between 39% (2017, n = 18/46) and 20% (2019, n = 12/59). No significant trend change in the occurance of resistance was observed during the study period. Likewise, no significant changes in resistance rates were identified to amoxicillin-clavulanic acid. Resistance rates ranged from 2% (2016, n = 1/44) to 11% (2017, n = 5/46). While

Antimicrobial Resistance of Escherichia coli Isolates from Skin and Soft Tissue (SST)
Only the occurrence of resistance against cefovecin and enrofloxacin was investigated, as, according to CLSI, isolates of Enterobacterales from body sites other than urine should not be reported as susceptible to aminopenicillins or amoxicillin-clavulanic acid due to low tissue concentrations achievable. The evaluation of cefovecin resistant isolates revealed significant (p < 0.05) fluctuations in direct consecutive years. However, since 2015 (28%, n = 50/179), a significantly decreasing trend (p < 0.05) in the occurrence of resistant E. coli was observed until 2021 (11%, n = 35/305). In 2015, 17% (n = 30/179) of isolates were resistant to enrofloxacin, and in 2018, a significant (p < 0.05) decrease in resistance was observed (8%, n = 27/349) ( Figure 6).

Antimicrobial Resistance of Escherichia coli Isolates from Skin and Soft Tissue (SST)
Only the occurrence of resistance against cefovecin and enrofloxacin was investigated, as, according to CLSI, isolates of Enterobacterales from body sites other than urine should not be reported as susceptible to aminopenicillins or amoxicillin-clavulanic acid due to low tissue concentrations achievable. The evaluation of cefovecin resistant isolates revealed significant (p < 0.05) fluctuations in direct consecutive years. However, since 2015 (28%, n = 50/179), a significantly decreasing trend (p < 0.05) in the occurrence of resistant E. coli was observed until 2021 (11%, n = 35/305). In 2015, 17% (n = 30/179) of isolates were resistant to enrofloxacin, and in 2018, a significant (p < 0.05) decrease in resistance was observed (8%, n = 27/349) ( Figure 6).

Antimicrobial Resistance of Pasteurella multocida
Less than 7% of the isolates were resistant to one or more of the investigated antimicrobial agents.

Antimicrobial Resistance of Pasteurella multocida
Less than 7% of the isolates were resistant to one or more of the investigated antimicrobial agents.

Antimicrobial Resistance of Pasteurella multocida
Less than 7% of the isolates were resistant to one or more of the investigated antimicrobial agents.

Antimicrobial Resistance of β-Hemolytic Streptococcus spp.
No resistance to β-lactam antimicrobials was detected in this study. Only in 2020, 2 of 202 isolates showed resistance to enrofloxacin. Thus, no significant change in the occurrence of resistance was detected for the tested agents.

Discussion
The World Health Organization compiled a list of critically important agents whose efficacy must be maintained for the treatment of human diseases [19]. The use of these agents is being restricted by an increasing number of guidelines, including the second amendment to the TÄHAV in Germany in 2018, in order to optimize the use of these antimicrobials to maintain their effectiveness [20]. From the HPCIA group, fluoroquinolones and 3rd generation cephalosporins were mainly used in small animal medicine due to their good efficacy and ease of use [14,21,22]. According to studies, their use was less than 10% of the applied antibiotics in Germany and the TÄHAV amendment was described as a reason for reduced use [17,23]. The most commonly applied agents in small animal medicine belong to the group of aminopenicillins [17,23,24], which, due to their broad spectrum of activity, are increasingly used as an alternative medication to HPCIA since the amendment of the TÄHAV, also to avoid AST [17]. Even if an influence of antibiotic application quantities on the development of resistance rates is assumed [25], this factor can only be reliably included into analyses from 2026 onwards with the legally required reporting of antibiotic consumption quantities.
As of 2018, a significant increase in sample submissions was observed in Germany. Therefore, it may be concluded that the amendment had an impact on the testing behavior of the participating veterinarians [17], assuming that the number of patients examined in the participating practices remained steady over the analyzed time period. According to previous surveys, the patient owner's willingness and ability to pay for AST influenced the treatment decision greatly [26,27]. By implementing the antibiogram requirement for the use of fluoroquinolones and 3rd/4th generation cephalosporines, veterinarians were provided with a rationale for good veterinary practice [14]. Especially among the most frequently sent materials, a large proportion of samples was found from which no bacteria could be isolated and thus an antibiotic therapy could be identified as avoidable.
The number of submitted samples for microbiological examination and susceptibility testing was significantly higher for dogs than for cats. However, the proportion of dogs and cats presented for examination in our participating practices is not known. Nevertheless, according to the literature, diseases of the skin and ears require antibiotic therapy particularly often which occur more frequently in dogs than cats [4,13,17,28] and are associated with increasingly resistant bacteria [29][30][31].

Staphylococcus spp.
Staphylococci are opportunistic pathogens bacteria that form part of the natural microbiome of skin and mucosa in animals [10,32,33]. The literature recommends aminopenicillins as first-line therapy for the treatment of infections caused by staphylococci [1,34]. However, resistance to penicillin G and ampicillin was found to be most common among Staphylococcus spp. as also described in the literature [31,32,35]. According to CLSI, an isolate resistant to penicillin G can be considered resistant to all penicillinase labile agents, such as aminopenicillins [36]. In contrast to M100, there are separate species-specific breakpoints for ampicillin used for evaluation in Vet01, which may have led to inconsistent interpretations in 133 Staphylococcus isolates during the analysis from 2015 to 2021.
Current resistance rates of Staphylococcus spp. determined during this study are comparable to values reported in the literature [31,35].
However, in contrast to the data of the GERM-Vet study, S. aureus and S. felis isolates were increasingly resistant against various antimicrobial agents in 2015 and 2016, respectively [37]. These deviations could be related to changes in internal laboratory processes that can no longer be tracked. However, an increase in the emergence of antimicrobial resistance could not be detected for the investigated active substances. Therefore, the efficacy of these antimicrobials has been maintained since the amendment of the TÄHAV; the GERM-Vet study also confirms these results and even supports the tendency of a decreasing occurrence of resistance [37]. Moreover, decreasing resistance to enrofloxacin was observed for all Staphylococcus spp. during this study. While the proportion of resistant S. aureus isolates only decreased significantly in 2021 compared to previous years, this trend can already be observed for S. felis as early as 2018 and even for S. pseudintermedius from 2017 onwards. As well, the Bundesamt für Verbraucherschutz und Lebensmittelsicherheit (BVL) found a significant reduction in resistance of S. aureus to enrofloxacin from 32% in 2012 to 11% in 2020. In contrast to our data, the GERM-Vet study described increasing levels of enrofloxacin resistance in S. pseudintermedius from 6% (2018) to 11% in 2020 in dogs pretreated with antimicrobials. Previous studies found tendencies that antibiotic pretreatment of animals could have an effect on the increased incidence of resistant isolates [31,37]. Information on the pretreatment of the animals was not available in the current study. These results could also be biased by different testing behavior of veterinarians as a result of the TÄHAV amendment as indicated by the rising number of samples submitted per year [17].

Enterobacterales
Enterobacteria are Gram-negative, facultative anaerobic rod-shaped bacteria. They are part of the natural microbiome of the intestinal tract of humans and animals [38]. However, in addition to physiological colonization, some strains cause severe intestinal and extraintestinal diseases [32,39].
Due to their natural occurrence in large numbers, members of this family with potentially zoonotic properties represent a major reservoir of resistance genes that may be transferred horizontally and are of public health concern [29,[40][41][42].
While aminopenicillins do not reach concentrations high enough for effective therapy in the tissue [18], these agents are excreted renally unchanged and are highly concentrated in urine, allowing its application for therapy of lower urinary tract infections caused by E. coli [43]. Therefore, a separate evaluation of isolates from infections of the urinary tract was performed.
The proportion of resistant isolates against HPCIA is equally low in all indications. This was supported by the results of the GERM-Vet study [37], while another study examining samples from antimicrobial pretreated animals described significantly higher resistance rates in E. coli [44]. The observed trend of declining resistance rates against enrofloxacin for both SST and UTI isolates was also confirmed for canine UTI E. coli in the GERM-Vet study data from 2012 to 2019 [37] and could reflect a more conscious use of antimicrobials in the treatment of cystitis [14].
Although the data from the current study found decreasing resistance rates against cefovecin for all species tested, it is important to be aware when evaluating these results that UTI breakpoints were also applied to SST isolates by the investigating laboratory, however, the accumulation in soft tissue is distinctly different from the concentration in urine. Furthermore, as of 2018, the CLSI provided an adjusted breakpoint for cefovecin UTI, but the time of implementation by the laboratory is unknown.
For the comparison of resistance in Klebsiella spp. and P. mirabilis in dogs and cats, the study of the Bundesverband für Tiergesundheit (BfT) from 2007 was used, as susceptibility testing was performed using the same method as in the current study. However, the BfT study described a much higher proportion of enrofloxacin resistant pathogens with 27% of P. mirabilis and 29% of Klebsiella spp. being resistant [45]. Only UTI isolates were examined for the BfT study, whereas in the current study isolates of various origins were analyzed. Additionally, Harada et al. found significant differences in resistance between a variety of indications for P. mirabilis [41].
Even though the generally observed decreasing incidence of resistance as of 2018 could be associated with the strongly increased number of Enterobacterales isolates tested per year since the TÄHAV amendment, due to changes in testing behavior of the participating veterinarians, the GERM-Vet study supports the finding of maintenance of antimicrobial efficiency [37].
3.3. Pasteurella multocida P. multocida are Gram-negative, rod-shaped bacteria that are part of the natural microbiome of the mucosa of the upper respiratory tract of cats but also dogs [46,47].
Infections caused by P. multocida are especially common in cats with bite wound infections or infections of the respiratory tract [46]. Literature recommends antibiotic treatment with aminopenicillins [47]. Data from this study as well as further studies showed good susceptibility of those agents used for an initial therapy [37,48]. However, the use of Convenia ® (Zoetis, Berlin, Germany), a 3rd generation cephalosporin with depot effect, was described in several studies due to the application advantages in cats [4,17,21,22]. The TÄHAV amendment may have led to more frequent bacteriological testing of bite wound infections and respiratory diseases, to be allowed to apply Convenia ® . Those cases would have been successfully treated without AST prior to the legal obligation. Therefore, an increase of susceptible isolates in the data could have occurred and led to the significant decrease in resistance observed during this study.

β-Hemolytic Streptococcus spp.
Streptococci are opportunistic pathogens of a wide variety of hosts, causing different infectious diseases, especially otitis and pyoderma [49][50][51]. While β-hemolytic streptococci are among the most commonly isolated pathogens in dogs, they are only exceptionally associated with resistance as found in the current data as well as further studies [30,52].

Limitations
It can be assumed that veterinarians who participated in this study and agreed to share the results of AST have a greater interest and awareness of the antimicrobial resistance problem, which may have resulted in data not reflecting the actual resistance situation in Germany. Inclusion of veterinary practices from all over Germany was done to avoid bias in resistance data due to geographic or practice-specific influences. In order to make the data more comparable and standardized, we decided to work with only one laboratory. Since retrospective data were evaluated, practice internal processes such as the decision to collect samples, the sample collection procedure, and sample processing, as well as information on pretreatment of patients are not available. It cannot be ruled out that these factors may have led to bias in the results. Especially since the number of submitted samples significantly increased as of the TÄHAV amendment, but the reasons for this are unknown. More frequent testing of uncomplicated infections since the TÄHAV amendment might have enhanced the number of susceptible isolates in the data pool and thus led to the significant decrease in resistance observed in this study. Although the results of our participants' survey suggest more prudent use of antibiotics, with HPCIAs used less frequently and penicillins deliberately selected as alternative antimicrobials, the first significant changes in the occurrence of resistance due to legal adjustments as the TÄHAV amendment are expected no earlier than three to five years after implementation [53]. Therefore, only significant trends in resistance as of 2021 could be assumed as influenced by the TÄHAV amendment such as the significant decreases in resistance rates of S. pseudintermedius against penicillin G and ampicillin, as well as the reduction in the occurrence of resistance of S. aureus against enrofloxacin.
Furthermore, in addition to AST results, the general health and condition of the immune system is also crucial when treating animals. AST should therefore only be considered as a supportive tool for treatment decisions.

Recruitment
To prevent bias in the data due to sample processing and evaluation, it was decided to work with only one laboratory, Laboklin (Labor für klinische Diagnostik GmbH & Co.KG, Bad Kissingen, Germany). During a previous survey, 378 German veterinarians were asked about their use of antimicrobials [17]. Forty-one of those veterinary practices collaborated with Laboklin for bacteriological examination and AST and provided their contact information after completing the questionnaire. All 41 practices gave their consent to provide susceptibility data for analysis. For three practices, no AST data were available from the period before the TÄHAV amendment in 2018, therefore, they were excluded from the analysis.

Bacterial Identification and Antimicrobial Susceptibility Testing (AST)
The bacteriological examination was performed by the commercial laboratory Laboklin. Bacterial identification was performed via MALDI TOF (Microflex LT and Biotyper sirius one, Bruker Daltonik GmbH, Bremen, Germany). For susceptibility testing, the minimum inhibitory concentration (MIC) was determined by broth microdilution using microtiter plates with individual customized layouts (MERLIN Micronaut System, Bornheim-Hersel, Germany). The microtiter plates were read photometrically. Testing and evaluation of the results was performed according to internationally recognized standardized procedures of the Clinical and Laboratory Standards Institute (CLSI) M100, Vet01, and Vet08 [18,36]. The results of bacterial identification and qualitative classification of the MIC values, together with information on the sample type and patient species were provided by the laboratory.

Coding
The qualitative classification of isolates was provided in Microsoft Excel ® 2018 (Microsoft Corporation, 2018, Microsoft Excel. Retrieved from https://office.microsoft.com/ excel, accessed on 30 June 2021) tables for 2015 to 2021. In Excel, recoding took place, summarizing the laboratory-provided names of the submitted samples according to synonyms and content-similar keywords.
Some bacterial species have been grouped together under their genus name; therefore Klebsiella spp. includes K. aerogenes, K. oxytoca, K. pneumoniae, and K. variicola.

Descriptive Analysis
Data obtained by the laboratory included the interpretation (sensitive (S)-intermediate (I)-resistant (R)) of the MIC. The analysis in this study was limited to relevant antimicrobials, particularly those whose efficiency may have been influenced by the TÄHAV amendment [17] and relevant pathogen-agent indication. The CLSI is the only organization that currently provides animal species-specific clinical breakpoints for some pathogen-drug combinations [18]. Nevertheless, for certain antimicrobial agents, no species-specific clinical breakpoints are available [54]. In the absence of veterinary-specific clinical breakpoints for isolates from dogs and cats, human-specific clinical breakpoints from CLSI document M100 were applied. For metric description, a sensitive result was assigned the numeral one, intermediate results the numeral two, and resistant results were represented by a three. Through these numerical assignments, an annual mean value representing the resistance rate was calculated. The development of the mean values was determined for each pathogen-agent combination in a line graph. Since the data are only a sample, the 95% confidence interval (CI) was calculated around each mean value. The mean value of 2015, the first year of this study, was set as the reference. If the mean values and their CIs of the following years were completely outside this reference line, a significant change in the resistance rate was assumed since p-values were less than 0.05 in these cases. The mean value of the year with the significant shift was set as reference for the assessment of changes in resistance of further years. All data were analyzed by descriptive analysis. The results are shown in the Supplementary Materials.

Conclusions
Veterinarians in Germany test pathogens more frequently for susceptibility as of 2018, which may lead to the assumption that the amendment of the TÄHAV has encouraged a more prudent use of antimicrobials. For all the bacterial species, the efficacy of analyzed substances was maintained over the study period. Nevertheless, the development and spread of antibiotic resistance still poses a public health risk and therefore requires further monitoring. Future research could focus on making commercially produced AST usable. In combination with the already legally required reporting of antibiotic consumption levels in dogs and cats from 2026, valuable information could be collected for the evaluation of legal measures.

Data Availability Statement:
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.