Antimicrobial susceptibility of staphylococci from bovine milk samples in routine microbiological mastitis analysis in Finland

The most frequent reason for antimicrobial use in dairy herds is mastitis and knowledge about mastitis-causing pathogens and their antimicrobial susceptibility should guide treatment decisions. The overall objective of this study was to assess antimicrobial resistance (AMR) of staphylococci in mastitic milk samples in Finland. MALDI-ToF MS identified a total of 504 Staphylococcus isolates (260 S. aureus and 244 non-aureus staphylococci, NAS) originating from bovine mastitic milk samples. Phenotypic susceptibility against cefoxitin, ceftiofur, enrofloxacin, gentamycin, oxacillin, penicillin, and tetracycline was evaluated by disk diffusion method and the presence of blaZ, mecA, and mecC genes investigated by PCR. Nitrocefin test assessed these isolates’ beta-lactamase production. The most common NAS species were S. simulans, S. epidermidis, S. chromogenes, and S. haemolyticus. In total, 26.6% of the isolates (18.5% of S. aureus and 35.2% of all NAS) carried the blaZ gene. Penicillin resistance, based on disk diffusion, was lower: 18.8% of all the isolates (9.3% of S. aureus and 28.9% of all NAS) were resistant. Based on the nitrocefin test, 21.5% of the isolates produced beta-lactamase (11.6% of S. aureus and 32.0% of all NAS). Between the Staphylococcus species, the proportion of penicillin-resistant isolates varied, being lowest in S. simulans and highest in S. epidermidis. Resistance to antimicrobials other than penicillin was rare. Of the eight NAS isolates carrying the mecA gene, six were S. epidermidis. One S. aureus isolate carried the mecC gene. Agreement beyond chance, assessed by kappa coefficient, between phenotypic and genotypic resistance tests, was moderate to substantial. Some phenotypically penicillin-susceptible staphylococci carried the blaZ gene but isolates without blaZ or mec genes rarely exhibited resistance, suggesting that the more reliable treatment choice may depend upon genotypic AMR testing. Our results support earlier findings that penicillin resistance is the only significant form of antimicrobial resistance among mastitis-causing staphylococci in Finland.


Introduction
The most frequent reason for antimicrobial use in dairy herds is mastitis (1).Microbiological diagnosis and knowledge of the antimicrobial susceptibility of pathogens form the basis for effective treatment of intramammary infections.Antimicrobial resistance has been a growing global concern during recent decades (2,3), with solutions to combat the problem requiring joint efforts within both human and veterinary medicine, in a true One Health spirit (4,5).Antimicrobial use can lead to resistant strains, and resistance mechanisms may be transmitted from one bacterial species or strain to another through mobile genetic elements (6).
Monitoring the antimicrobial resistance of mastitis-causing pathogens is important for ensuring the continued availability of efficacious treatments.
In Finland, prudent antimicrobial use guidelines for mastitis therapy are followed: Bacteriologic analysis of milk samples before initiation of antimicrobial treatment is a common practice and penicillin is the drug of choice in mastitis caused by Gram-positive bacteria (7).Additionally, selective dry cow therapy (SDCT) has always been implemented, with only approximately one-fourth of cows receiving antibiotic dry cow treatment at the end of lactation (8).Since 2010, a mastitis diagnosis in Finland has relied predominantly on PCR methodology after the main dairy co-operative in the country switched to PCR in its mastitis laboratory.Later, also a private veterinary diagnostic laboratory and the clinical laboratory at the Department of Production Animal Medicine at the University of Helsinki adopted this technology.Veterinarians occasionally use bacterial culture and selective agar plates in their clinic laboratories to achieve a quick bacterial diagnosis during weekends and on-call work.The commercial PCR kit (PathoProof™ Complete-16 PCR Assay, Thermo Fisher Scientific, Waltham, MA, United States) used in these diagnostic laboratories targets the 15 main mastitis pathogens plus the staphylococcal blaZ gene which codes for beta-lactamase production.Russi et al. (9) reported a discrepancy between a finding of the blaZ gene and phenotypic penicillin resistance, a phenomenon also observable in the clinical laboratory of Production Animal Hospital, Faculty of Veterinary Medicine, University of Helsinki.After adopting the PCR methodology, the laboratory has used both PCR and phenotypic beta-lactamase production tests for determining the antimicrobial susceptibility of staphylococci.Discrepancies in results make treatment decisions challenging.
Staphylococcus spp.are the most prevalent causal agents of intramammary infections in Finland, constituting approximately 64% of all pathogens isolated from IMIs (10).Resistance of staphylococci against antimicrobials other than beta-lactams is infrequently monitored in Finland, since penicillin is the drug of choice for treatment of infections caused by Gram-positive cocci.In routine diagnostics, staphylococcal species, except for Staphylococcus aureus, are not identified beyond genus level in the PCR-based mastitis diagnostics, although some studies have shown differences between the species in antimicrobial susceptibility (11,12).
Methicillin resistance coded by the mecA and mecC genes is of great concern in both human and veterinary medicine, due to the zoonotic nature and role of S. aureus in both community-acquired and nosocomial infections (13,14).Although methicillin resistance is rare in staphylococci isolated from bovine mastitis (12,15,16), methicillinresistant S. aureus (MRSA) have appeared rather commonly in swine production (17,18), in horses, and in companion animals (19)(20)(21).
The main objective of this study was to evaluate the antimicrobial susceptibility of mastitis-causing staphylococci against those antimicrobial agents most frequently used in treatment of bovine mastitis and other infectious diseases of cattle.The other objectives were to estimate the prevalence of methicillin resistance, and to compare the genotypic and phenotypic beta-lactamase resistance in these staphylococci.

Milk samples
The study material comprised milk samples submitted for bacteriologic analysis to the mastitis laboratory of the main Finnish dairy company Valio Ltd.The company is a farmer-owned cooperative with members across the entire country; it processes approximately 80% of all the milk produced in Finland.Finnish dairy farmers and veterinarians frequently send milk samples from clinical mastitis cases to that laboratory for bacteriologic analysis.They typically receive results electronically the next day by text message, by email, or through a herd-health program of the dairy cooperative.Additionally, milk samples from subclinically infected, high-SCC quarters are often collected before treatment, e.g., at dry-off, to ensure selection of the most appropriate and effective treatment.For its mastitis diagnosis, the Valio laboratory uses PathoProof™ Complete-16 PCR Assay (Thermo Fisher Scientific), which targets 15 mastitis-causing microbial species and the staphylococcal blaZ gene.Each week between August and November 2017, the laboratory personnel saved and froze milk samples which according to the PCR assay contained substantial amounts of the DNA of S. aureus (cycle threshold, Ct-value ≤30.0) or non-aureus staphylococci, NAS (Ct-value ≤31.0) for later culturing and antimicrobial susceptibility testing.
Assuming 20% of S. aureus to be penicillin resistant (17), the Epitools calculator, using 80% power and 5% significance level, provided a sample size of 246 S. aureus isolates to estimate the proportion of penicillin-resistant S. aureus with 5% precision (22).Our goal was to collect the same number of NAS isolates.Only one isolate of any staphylococcal species detected in a herd was included in the data analysis, to ensure the observations' epidemiologic independence.If the staphylococcal species, however, differed from each other, multiple samples and isolates may have been included from the same cow or the same farm,

Isolate culture and speciation
In the laboratory of the Department of Production Animal Medicine, Faculty of Veterinary Medicine, University of Helsinki, the milk samples were thawed, and 10 μL of milk was streaked on sheep-blood agar plates (Oxoid, Thermo Fisher) and incubated for 24 h at +37°C.Potential staphylococcal colonies were identified based on colony morphology and hemolysis and then MALDI-ToF MS (Bruker Maldi Biotyper, Bruker Daltonics Scandinavia AB, Solna, Sweden) determined the species by the direct transfer method (23).The PCR assay identifies NAS only at group level.If, however, MALDI-ToF MS identified more than one NAS species in a sample, isolates of differing species were considered for the study.Similarly, if a sample contained both S. aureus and NAS, both were considered in the study.However, samples positive for differing species and the blaZ gene were excluded from the study, because it is not possible to know which species carried the blaZ gene.Samples with >2 different colony types were excluded from the study.

Susceptibility testing
The antimicrobial susceptibility to cefoxitin, ceftiofur, enrofloxacin, gentamycin, oxacillin, penicillin, and tetracycline was evaluated using the disk diffusion method.The disks (Oxoid, Thermo Fisher) contained cefoxitin 30 μg, ceftiofur 30 μg, enrofloxacin 5 μg, gentamycin 10 μg, oxacillin 1 μg, penicillin 10 μg, and tetracycline 30 μg.For the cut-off/breakpoint values used and their sources see Table 1.The epidemiologic cut-off (ECOFF) values for susceptibility of staphylococcal species were used when available (24).For ceftiofur, the values were CLSI clinical breakpoints for mastitis in cattle (25).For benzylpenicillin, epidemiologic cut-off values exist only for 1 μg disks, whereas we used 10 μg disks, and thus used a CLSI clinical breakpoint value for human staphylococci.Beta-lactamase production of the isolates was tested by use of liquid nitrocefin prepared from nitrocefin powder as recommended by the manufacturer (Nitrocefin Solution, Oxoid, Thermo Fisher).

Statistical analyses
Descriptive statistics (median, minimum, and maximum inhibitory zone diameters and proportions of isolates resistant for each antimicrobial drug) were calculated separately for S. aureus, for the four most prevalent NAS species, and for the other NAS species together as a group.Proportions of the bacteria carrying blaZ or mec genes were calculated.Agreement beyond chance between phenotypic (based on a disk diffusion method and nitrocefin test) and genotypic (carriage of blaZ gene) resistance to penicillin was assessed by kappa coefficient.Statistical analyses were conducted with the Statistical Analysis System, v. 9.4 (SAS Inst Inc., Cary, NC, United States) and Epitools Calculator (22).

Results
The antimicrobial susceptibility of 504 isolates was evaluated and included in the data analysis.These isolates originated from 497 quarter milk samples from 466 cows in 396 herds.Of these isolates, 260 were S. aureus, and 244 belonged to the NAS group.Altogether, 21 differing NAS species were detectable, including six isolates from three species previously included in the genus Staphylococcus: S. lentus (2), S. sciuri (3), and S. vitulinus (1), which recently have been reassigned to a novel genus, Mammaliicoccus (27).The most common NAS species were S. simulans, S. epidermidis, S. chromogenes, and S. haemolyticus, these accounting for approximately 74% (180/244) of all detectable NAS (Table 2).Isolates identified as S. haemolyticus may have included some isolates of a novel species, S. borealis, because MALDI-ToF MS does not differentiate between them (28).Of 28 samples with more than one staphylococcal species, six each contained S. aureus and a NAS species, and two had three different NAS species; the rest, 20 samples, each had two different NAS species.Of the 28 samples, in 21, the blaZ gene was also detectable, and these samples were excluded, because-based on the commercial PCR assay used-distinguishing which bacterial species carried the blaZ gene was impossible.Five blaZnegative samples contained two different NAS species, one contained S. aureus and a NAS, and one sample had three different NAS species.All other samples had only one species each.

Resistance to beta-lactams
Based on the detection of the blaZ gene in the PathoProof™ Complete-16 assay, of all 504 staphylococcal isolates, 134 (26.6%), this constituting 8.5% of S. aureus and 35.2% of all NAS, carried the blaZ gene.These could thus be considered resistant to penicillin.Penicillin resistance was lower based on disk diffusion: Of all the isolates, 18.8% (9.3% of S. aureus and 28.9% of all NAS) were penicillin resistant.
Inhibition-zone diameters (median, min, and max) of penicillin, cefoxitin, ceftiofur, and oxacillin for S. aureus, for the four most prevalent NAS species, and for the rest of the NAS group are in Tables 2, 3. Penicillin resistance based on disk diffusion testing was most common in S. epidermidis, of which 70.2% were resistant.The proportion of blaZ-positive S. epidermidis isolates was even higher, 79.6% (Table 2).Distribution of penicillin-inhibition zones, both for S. aureus and all NAS species, was bimodal, and the distribution of wild-type (isolates without acquired resistance) and resistant isolates was largely in agreement with the blaZ-gene occurrence (Figures 1, 2).In contrast to penicillin-inhibition zones, distribution of cefoxitinand oxacillin-inhibition zones for all staphylococcal species was unimodal, and the blaZ-gene carriers were largely susceptible, especially to cefoxitin (Figures 3, 4).Based on the nitrocefin test, 21.6% of the isolates produced beta-lactamase (11.6% of S. aureus and 32.0% of all NAS; Table 2).

mecA or mecC genes and agreement between different tests on beta-lactam resistance
In total, 158 staphylococcal isolates (84 S. aureus and 74 NAS) were tested for carriage of mecA and mecC.Eight of the NAS isolates carried mecA, six of these being S. epidermidis, one S. haemolyticus, and one  M. lentus (Table 2).No S. aureus isolates carried mecA, but one was mecC positive.The penicillin-inhibition zones of the mecA-positive NAS isolates ranged from 6 to 24 mm, indicating penicillin resistance (Figure 2), but much variability existed in the cefoxitin-and oxacillin-inhibition zones.
All mec-positive isolates were classified as resistant to penicillin in the disk diffusion test (Figures 1, 2).Of the eight mecA-positive isolates, four, and also the mecC-positive isolate, were resistant to cefoxitin (Figure 3).The mec carriage of the other four isolates would have remained undetected had the screening been based only on cefoxitin disk diffusion and the recommended cut-off values (24).
Of the nine mec-positive isolates, eight were resistant to oxacillin (Figure 4), and thus only one mecA-positive isolate remained undetected by oxacillin disk diffusion and the recommended cut-off value.On the other hand, of the 68 isolates classified as oxacillin resistant, only nine  were mec-gene positive.Thus, 59 isolates were classified as oxacillin resistant based on disk diffusion but were mec-negative.Of the eight mecA-positive NAS isolates, seven also carried the blaZ gene, detectable by the PathoProof™ Complete-16 assay.The mecC-positive S. aureus isolate carried no blaZ, but its inhibition zone for penicillin was 19 mm, for ceftiofur 16 mm, for cefoxitin 18 mm, and for oxacillin 14 mm, all indicating its being resistant to all beta-lactam antimicrobials.
Of all isolates, 12 were blaZ PCR-negative but phenotypically resistant to penicillin (inhibition zones ≤28 mm).Two of these carried either the mecA or the mecC gene, but 10 were mec-gene negative.Conversely, 50 blaZ-positive isolates had penicillin inhibition zones ≥29 mm, ranging from 29 to 51 mm, indicating penicillin susceptibility.Five of these isolates had oxacillin inhibition zone diameters ≤19 mm, ranging from 14 to 18 mm (Figure 4).Frequency distribution of cefoxitin-inhibition zones for all staphylococcal isolates, occurrence of blaZ gene (no = blue; yes = red) and carriage of mecA (indicated with *) or mecC gene (indicated with +).The cut-off for cefoxitin resistance in the disk diffusion test was 22 mm for S. aureus and 25 mm for non-aureus staphylococci.Frequency distribution of oxacillin inhibition zones for all staphylococcal isolates, occurrence of blaZ gene (no = blue; yes = red) and carriage of mecA (indicated with *) or mecC gene (indicated with +).The cut-off for oxacillin resistance in the disk diffusion test was 20 mm.Agreement beyond chance between phenotypic penicillin resistance based on the disk diffusion test and occurrence of the blaZ gene was moderate for S. aureus (kappa = 0.56) and was substantial for NAS species (kappa = 0.67), as well as for all staphylococcal species assessed together (kappa = 0.64; Table 4).Proportion of negative agreement (ranging from 0.89 to 0.94) was generally higher than proportion of positive agreement.Agreement beyond chance between blaZ occurrence and the nitrocefin test result was moderate (kappa = 0.62); 35.8% of the blaZ-positive isolates were nitrocefin-test negative, and 6.2% of the blaZ -negative isolates were nitrocefin-test positive.Overall, more commonly genotypically resistant isolates (those carrying blaZ or mec genes) were phenotypically susceptible than genotypically susceptible isolates (those without blaZ-or mec genes) were phenotypically resistant.Genotypic tests may be more reliable than phenotypic tests, especially when one is choosing treatment options.

Resistance to other antimicrobials
Resistance to other than beta-lactam antimicrobials was rare (Table 5).None of the isolates were resistant to gentamycin.Resistance to tetracycline was detectable in only six isolates (1.2%), of which three were mecA-positive S. epidermidis, one was S. kloosii, one S. simulans, and one S. xylosus.Resistance to enrofloxacin was detectable in three isolates (0.6%): S. simulans, a mecA-positive S. haemolyticus, and a mecA-positive M. lentus.Overall, the nine mecpositive isolates were the most resistant.All three ceftiofur-resistant isolates (S. aureus, S. haemolyticus, and M. lentus) were mec-gene carriers.In addition to beta-lactam resistance, three of the mecpositive isolates were resistant to tetracycline (all S. epidermidis) and two to enrofloxacin (S. haemolyticus, M. lentus).

Discussion
Our study of the antibiotic resistance of mastitis-causing staphylococci in dairy cows found that the most common form of resistance was resistance to penicillin.This is of practical importance because in Finland, penicillin is the drug of choice in mastitis treatments.When mastitis results from penicillin-resistant staphylococci, cloxacillin intramammary treatment is used (7), but antimicrobial treatment of penicillin-resistant staphylococcal mastitis, especially involving S. aureus, is generally not recommended because of the poor cure rate (30)(31)(32).Penicillin resistance of all isolates based on disk diffusion was 18.8%, if based on a nitrocefin test was 21.5%,   10.3389/fvets.2023.1235417 Frontiers in Veterinary Science 08 frontiersin.orgbut when based on blaZ gene carriage was higher, 26.6%.The Finnish Food Authority monitored antimicrobial resistance of staphylococci isolated from mastitic milk samples in 2005 and 2012 (15, 33).In 2005, 25% of S. aureus and in 2012, 23% of S. aureus, and 36% of NAS produced beta-lactamase, so it appears that the antimicrobial resistance of mastitis-causing staphylococci in Finland has been declining slightly.Lower or similar rates of penicillin resistance as in the current study have been detectable in three other Nordic countries (Norway, Sweden, and Denmark).In Sweden, based on beta-lactamase production, 3% of S. aureus isolates and 30% of NAS isolates from clinical mastitis cases were resistant to penicillin; figures based on MIC values were similar or higher: 3% for S. aureus and 38% for NAS (16).In another Swedish study, 34% of NAS isolates from subclinical mastitis were beta-lactamase positive (34).In Norway, 5% of S. aureus and 23% of NAS isolates were penicillin resistant based on disk diffusion (12).In Denmark, 18% of S. aureus and 22% of NAS isolates were penicillin resistant based on MIC values (35).Penicillin resistance levels for NAS, based on MIC values, were lower in Canada, 10% (36), and in Korea, 14% (37).In some other countries, higher proportions of penicillin-resistant staphylococci isolated from bovine milk have also been detectable.In France, of almost 7,000 coagulasepositive Staphylococcus isolates (mainly S. aureus) collected during 2006-2016, 40% were penicillin resistant based on disk diffusion (38).In Brazil, 36% of Staphylococcus isolates from bovine mastitis, mainly S. aureus, were blaZ positive (39).In South Africa, 63% of 142 NAS isolates from bovine subclinical mastitis were penicillin resistant based on disk diffusion (40).
It is worthwhile to note, however, that a direct comparison of the results between studies is difficult, as study populations, sample collection, and sources of the isolates and methods for antimicrobial susceptibility testing differ.Another explanatory factor for differences in penicillin resistance figures may be related to the distribution of NAS species.In a Swedish study by Nyman et al. (34), S. epidermidis was the most common species, involving 26% of the 783 NAS isolates; similarly, in our study, S. epidermidis comprised about 20% of all NAS species.In contrast, in a Canadian study, where the reported penicillin resistance was lower, only 4% of over 1700 NAS isolates were S. epidermidis, and the three most prevalent species: S. chromogenes, S. simulans, and S. xylosus covered 70% of all isolates (36).
Huge differences in penicillin resistance between the Staphylococcus species were detectable both in ours and other studies.In our study, the lowest proportion of penicillin-resistant isolates was in S. simulans, and the highest in S. epidermidis, based on disk diffusion, nitrocefin production, and blaZ carriage.Many other studies have also shown S. simulans to be mainly susceptible to penicillin and shown S. epidermidis to be the staphylococcal species most resistant to penicillin and to several other antimicrobials (12,36,37).Multidrug resistance has also been most common in S. epidermidis (11,40).In addition, S. epidermidis is the Staphylococcus species that most commonly carries the mecA gene coding for methicillin resistance (11,36,37), and consistently, of the eight mecA-positive isolates in the current study, six were S. epidermidis.
The prevalence of mec-positive isolates was low in S. aureus (0.4%), and slightly higher (3.3%) in NAS.In FINRES-Vet antimicrobial resistance monitoring in 2012, no S. aureus but five NAS (5.7%) carried the mecA gene (15).In the study by Gindonis et al. (41) which utilized three different Finnish samplings of staphylococci from bovine mastitis from previous studies, 1.5% of S. aureus and 1.8-5.2% of NAS isolates were mec-gene positive.Although in Finland methicillin resistance in mastitis-causing staphylococci is low, it may, in rare cases, cause problems in clinical work.The occurrence of the mec genes is not routinely tested for in mastitis cases.Some blaZnegative isolates may be mec positive and thus penicillin resistant.The only mecC-positive S. aureus isolate in our study was blaZ negative and was incorrectly classified as penicillin susceptible.Penicillin treatment of these cases would have led to treatment failure.
Besides resulting in ineffective treatments, methicillin-resistant staphylococci can transfer from cows to humans.In general, although staphylococcal lineages are host-specific, host shifts may occur (42,43).Currently, methicillin resistance of staphylococci is no problem in Finnish dairy production, but in swine production MRSA has emerged.In a year-long survey in 2016-2017, MRSA emerged in 77% of slaughter batches (17).In the same time period, MRSA infections caused by the livestock-associated CC398 have emerged in Finnish swine farmers (44).Some of the mecA-positive isolates in our study were also resistant to tetracycline or enrofloxacin.Although in Finland tetracycline is not the drug of choice for mastitis therapy, it is commonly used for treating other infectious diseases of cattle, especially respiratory diseases.Enrofloxacin is, among antimicrobials, the drug of choice in Finland for severe coliform mastitis.
The oxacillin disk-diffusion test better indicated mecA carriage than did the cefoxitin disk-diffusion test.However, many of the oxacillin disk-diffusion test results were false positive, i.e., mecnegative isolates showing resistance to oxacillin.Neither of the tests was, however, perfect for detecting mec-gene carriage.The European Committee on Antimicrobial Susceptibility Testing (EUCAST) and the Clinical and the Laboratory Standards Institute (CLSI) recommend cefoxitin disk for screening of mecA-mediated beta-lactam resistance in Staphylococcus aureus and NAS (25,45), but in our study it did not detect potential mecA-positive isolates very effectively.The cefoxitin disk is reported to perform better for mecC screening than does the oxacillin disk (46,47).The only mecC-positive S. aureus isolate in our study was detectable by both methods.We did not test for the recently detected mec-gene variant mecB (48), and other resistance mechanisms may also exist, such as mutations in the gdpP gene (49).
In Finland, veterinarians choose an antimicrobial drug for treatment of staphylococcal mastitis largely based on the blaZ result in the PCR test (PathoProof™ Complete-16 PCR Assay).After purchasing the equipment for analyzing milk samples with the q-rt PCR methodology, the clinical laboratory at the Department of Production Animal Medicine, University of Helsinki, analyzed milk samples both with PCR and with conventional culturing and detection of beta-lactamase production in staphylococci (Nitrocefin test, ThermoFisher Scientific).The discrepancy between these results has led to a challenge in decision-making regarding the management of an IMI case: what is the likelihood of cure and which antimicrobial to choose, if the case is treated with antimicrobials?Or is it preferable to dry off the infected quarter or even to cull the cow?The discrepancy between blaZ-gene carriage and the phenotypic penicillin resistance detected in our and other studies in veterinary (9, 50) and human (51,52) medicine is an interesting phenomenon.BlaZ-negative but phenotypically penicillin-resistant S. aureus isolates are scarce ( 9), but blaZ-positive isolates that are phenotypically penicillin sensitive commonly exist (9,50,51).In our study, the blaZ gene was identified with PathoProof™ Complete-16 PCR Assay (Thermo Fisher Scientific) directly from milk, not from cultivated isolates.Therefore, we excluded from the study the samples positive for more than one staphylococcal species and the blaZ gene, as in samples with more than one species it was not possible to distinguish which bacterial species carried the blaZ gene.Different tests typically show different sensitivities for detection of beta-lactamase production.Pitkälä et al. (53) compared six tests, using the blaZ PCR as the reference method.At least one method was always positive, supporting the potential for beta-lactamase production of the blaZ-positive isolates.Some authors report that tests based on detection of beta-lactamase production (e.g., nitrocefin test, clover leaf test) correlate better with the occurrence of the blaZ gene than does the agar dilution method (54).Others have reported lower sensitivities for these tests (51,55).The results of the agar dilution and MIC methods depend on the set cut-off values.In our study, the agreement beyond chance between phenotypic penicillin resistance based on the disk diffusion test and occurrence of the blaZ was moderate to substantial, depending on the Staphylococcus species.One possible reason for phenotypically penicillin susceptible but blaZ-positive isolates is impaired function of the blaZ or its regulators, the blaI and blaR genes, because of sequence mutations (56).Whether all blaZ-positive isolates produce beta-lactamase in vivo remains to be solved.Meanwhile, blaZ-positive results should be interpreted as originating from potentially penicillin-resistant isolates.In practice, the possible discrepancy between occurrence of the blaZ gene and penicillin susceptibility may cause problems in antimicrobial treatment of mastitis.
Resistance to enrofloxacin, gentamycin, and tetracycline was rare for us, similarly to findings in most other studies.None of our isolates was resistant to gentamycin, and most of the few isolates resistant to enrofloxacin or tetracycline carried the mecA gene.Fergestad et al. (12) found, among 100 Staphylococcus isolates from clinical mastitis in Norway, only three isolates resistant to gentamycin, and three isolates resistant to tetracycline.Duse et al. (16) found no gentamycin resistance in Swedish bovine mastitis isolates.All of their NAS isolates were susceptible to enrofloxacin and tetracycline, and tetracycline resistance was also rare for S. aureus (16).Chehabi et al. (35) found no gentamycin resistance among 63 Danish S. aureus isolates, with only one isolate being tetracycline resistant.Of 49 Danish NAS isolates, one was gentamycin resistant, and 5 isolates (10%) were tetracycline resistant.The level of resistance among NAS species varies considerably.Nobrega et al. (36) reported 10% tetracycline resistance for their entire NAS group in Canada, but the percentage of resistant isolates of their most common NAS species ranged from 2% in S. chromogenes to 31% in S. xylosus and 32% in S. epidermidis.Among similar findings in South Korea by Kim et al. (37) were tetracycline resistance in 4% of S. chromogenes and in 26% of S. epidermidis.

Conclusion
Our results support earlier findings that penicillin resistance is the only significant form of antimicrobial resistance among mastitiscausing staphylococci in Finland, and the proportion of resistant isolates has not increased.Based on blaZ-gene carriage, less than one-third of all Staphylococcus isolates were resistant to penicillin, and in phenotypic testing, penicillin resistance was even lower.The difference in proportions of penicillin resistance between Staphylococcus species, however, was considerable, lowest in S. simulans and highest in S. epidermidis.Methicillin resistance was rare, except in S. epidermidis, of which 12% were mecA positive.Some phenotypically penicillin-susceptible staphylococci carried the blaZ gene but isolates without blaZ or mec genes rarely exhibited resistance, suggesting that the more reliable treatment choice may depend upon genotypic AMR testing.The drug of choice for mastitis treatments in Finland is penicillin and resistance to antimicrobials other than penicillin was rare.With targeted treatment decisions, based on the knowledge on infection-causing pathogens, it is possible to keep levels of resistance to antimicrobials low.

FIGURE 1 Frequency
FIGURE 1Frequency distribution of penicillin-inhibition zones for Staphylococcus aureus, occurrence of blaZ gene (no = blue; yes = red) and carriage of mecC gene (indicated with +).The cut-off for penicillin resistance in the disk diffusion test was 29 mm.No S. aureus carried mecA gene.

FIGURE 2 Frequency
FIGURE 2Frequency distribution of penicillin-inhibition zones for non-aureus staphylococci (NAS), occurrence of blaZ gene (no = blue; yes = red) and carriage of mecA gene (indicated with *).The cut-off for penicillin resistance in the disk diffusion test was 29 mm.No NAS isolates carried mecC gene.

TABLE 1
Epidemiologic cut-off and clinical breakpoint values (mm) for categorizing bacterial isolates as susceptible or resistant in disk diffusion test.

TABLE 2
Percentages of isolates carrying the blaZ gene encoding beta-lactamase production, and zone-inhibition diameters for penicillin in the most common staphylococcal species causing intramammary infections in dairy cows.Only those isolates with oxacillin-inhibition zone ≤20 mm or cefoxitin-inhibition zone ≤28 analyzed for mecA or mecC genes.Numbers are percentages carrying mec of the total number of isolates.

TABLE 3
Zone-inhibition diameters for cefoxitin, ceftiofur, and oxacillin in the most common staphylococcal species 1 in dairy-cow intramammary infections.

TABLE 5
Zone inhibition diameters for enrofloxacin, gentamycin, and tetracycline among the most common staphylococcal species 1 isolated from bovine milk.
1 Species identification by MALDI-TOF MS. 2 Percentage of isolates classified as non-susceptible (including resistant and intermediate), based on disk diffusion test using EUCAST epidemiologic cut-off values when available or CLSI clinical breakpoints (see text for details).