Antimicrobial Susceptibility of Udder Pathogens Isolated from Dairy Herds in the West Littoral Region of Uruguay

A total of 522 strains belonging to streptococci, enterococci and staphylococci isolated from sub-clinical and clinical cases of bovine mastitis from the west littoral region of Uruguay were analysed for their susceptibility to several antimicrobial agents. The susceptibility patterns were studied by agar disk diffusion methods (ADDM) and broth micro-dilution to determine the minimum inhibitory concentration (MIC). The concentration that inhibits 90% (MIC90) of the analysed strains reported in micrograms per millilitre, for Staphylococcus aureus were > 8, 8, ≤ 0.5, ≤ 4, ≤ 1, ≤ 0.5, > 64, ≤ 0.25, 0.5, ≤ 1 and ≤ 1 to penicillin, ampicillin, oxacillin, cephalotin, gentamicin, erythromycin, oxitetracycline, enrofloxacin, trimethoprim/sulfamethoxazole, neomycin, and clindamycin, respectively. Coagulase-negative staphylococci (CNS) had different values for penicillin (4) and ampicillin (2), while the other antimicrobial agents had the same MIC90 values as reported for S. aureus. The MIC90 values for streptococci were 0.12, 0.25, ≤ 4, 16, ≤ 0.25, 0.5, 0.25 for penicillin, ampicillin, cephalotin, gentamicin, erythromycin, oxytetracycline and trimethoprim-sulfamethoxazole, whereas MIC90 for enterococci were 4, 4, 4, ≤ 0.5, 2, > 8 for penicillin, ampicillin, gentamicin, erythromycin, oxytetracycline and trimethoprim-sulfamethoxazole, respectively. Of 336 strains of S. aureus, 160 (47.6%) were resistant to penicillin. For 41 CNS strains, 10 (27%) presented penicillin-resistance. All the streptococcal strains were susceptible to penicillin, while 3 (7%) of the 43 enteroccocal strains were resistant. Non significant statistical differences were found between the results obtained by ADDM and broth micro-dilution for classifying bacterial isolates as susceptible or resistant according to the National Committee of Clinical Laboratory Standards.


Introduction
Bovine mastitis is the major problem for milk producers throughout the world and responsible for substantial losses of revenue annually. Antibiotic therapy is an important tool in the scheme of mastitis control. The treatments are more effective when directed by veterinarians; for example correct drug selection can be enhanced using an appropriate antimicrobial sus-ceptibility test. The misuse or intensive use of antibiotics can lead to the development of resistance among different bacterial strains and contamination of foodstuff, with animal and human health implications (Lingaas 1998). The antimicrobial resistance is the result of mutations or exchange of genetic material such as plasmids and transposons (Neu 1992). Such resis-tance determinants most probably are acquired by pathogenic bacteria from a pool of resistance genes in other microbial genera present in different environments (Davies 1994). Increased resistance of Staphylococcus aureus and coagulase-negative staphylococci (CNS) isolated from bovine mastitis cases to antimicrobial agents has been reported by Gentilini et al. (1995), Aarestrup & Jensen (1998) and Myllys et al. (1998). Milk production in Uruguay (South America) is important with a total of 410.000 dairy cows, yielding 1462 millions litres in 1999 (OPYPA 2000). In spite of the importance of this sector, only 3 surveys to evaluate the resistance of udder pathogens to antibiotics have been performed in Uruguay using agar disk diffusion (ADDM, Bauer & Kirby 1966): 1) Del Baglivi et al. (1976) testing S. aureus and Streptococcus agalactiae isolated from subclinical cases obtained from 43 dairy farms in the southern dairy region of Uruguay showed that 53% of S. aureus and 100% of Str. agalactiae were sensitive to penicillin. 2) Herrera et al. (1982) found 78% of S. aureus strains susceptible to penicillin in the dairy area around Tacuarembó city (north of Uruguay). 3) Bouman et al. (1999) studied the resistance patterns of S. aureus and CNS isolated in the laboratory routine during 4 years from milk samples collected in the southwestern region of Uruguay for penicillin, cloxacillin, nafcillin, rifampin and tetracycline obtaining: 58%, 16%, 5%, 6%, 29% of resistance for S.aureus and 75%, 42%, 17%, 12%, 26% for CNS, respectively. The sale of antibiotics is free in Uruguay, while the mastitis treatment is usually performed by the herd dairyman, and the antimicrobial agents most commonly used are tetracyclines, betalactams, macrolides, and aminoglycosides. The methods for susceptibility testing used to choose the appropriate drug are ADDM qualitative test and quantitative determinations by means of microdilution to determine the minimum inhibitory concentration (MIC) (Amsterdam 1996, Acar & Goldstain 1996. These methods can be interpreted following the National Committee for Clinical Laboratory Standards criteria (NCCLS 1999) or guidelines proposed by other national antibiogram committees (Acar & Goldstain 1996). The purposes of this work were: To determine the phenotypic expression of in vitro susceptibility of antimicrobials for pathogens (staphylococci, streptococci, and enterococci) isolated from dairy herds in Uruguay, and to compare the results obtained by the ADDM vs. broth micro-dilution method according to the NCCLS criteria.

Materials and methods Sample
A total of 522 strains including streptococci, enterococci and staphylococci were used in the study. The strains were isolated from sub-clinical and clinical cases of bovine mastitis from a survey carried out in the west littoral region of Uruguay (Gianneechini 2001), where quarter foremilk samples from 1077 milking cows and 40 milk samples from clinical cases detected in one month were collected in 29 randomly selected dairy farms. All strains were identified according to the procedures of the laboratory at the Department of Mastitis and Diagnostical Products, National Veterinary Institute (SVA), Uppsala, Sweden (National Veterinary Institute 1998). The isolates were maintained frozen at -20°C in Trypticase soy broth (Difco Laboratories, Michigan, USA) containing 10% glycerol until testing.

Susceptibility testing
Prior to the susceptibility testing all isolates were sub-cultured on Blood-esculin agar and incubated for 24 h at 37°C. Two different tests were carried out to determine the drug susceptibility for all strains: 1 -The ADDM was conducted and interpreted according to the recommendations and criteria of the NCCLS for bacteria isolated from animals (NCCLS 1999). The following disks (Becton Dickinson Microbiology System, Cockeysville, Maryland, USA) were used: penicillin, 10 µg; ampicillin, 10 µg; oxacillin, 1 µg; amoxicillin -clavulanic acid, 20 µg + 10 µg; cephalotin, 30 µg; gentamicin, 10 µg; erythromycin, 15 µg; enrofloxacin, 5 µg; tetracycline, 30 µg; neomycin, 30 µg; trimethoprimsulfamethoxazole, 1.25 µg + 23.75 µg. The staphylococci were tested against all the drugs above, while the streptococci against only 6 of these antimicrobial agents (penicillin, ampicillin, cephalotin, gentamicin, erythromycin and tetracycline), and enterococci against penicillin, ampicillin, gentamicin, erythromycin and tetracycline. The medium used was Mueller-Hinton Agar (Difco Laboratories, Detroit, USA) for sthaphylococci and Mueller-Hinton agar supplemented with 5% sheep blood for streptococci. S. aureus ATCC 25923, E. coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853 were included as quality control strains. The plates were read after 18 h incubation at 37°C under aerobic conditions. The isolates were categorised as susceptible, intermediate and resistant by measuring the inhibition zone. 2 -The MIC was determined using a commercially available microdilution system (VetMIC TM +/-panels, SVA, Uppsala, Sweden). The tests were performed by manufacturer's instruction and interpreted according to international standards (NCCLS 1999) using Mueller-Hinton broth (Oxoid Limited, Basingstoke Hants, England) and S. aureus ATCC 29213, Enterococcus faecalis ATCC 29212 and E. coli ATCC 25922, as quality control strains.
When the streptococcal strains were tested, 100 µl were inoculated in each well with erythromycin to obtain the following dilution: 0.25; 0.5; 1 and 2 µg/ml of the antimicrobial agent. These modifications were carried out to adapt to the breakpoints suggested by NCCLS (1999) for erythromycin. All panels were read on the same conditions as in the ADDM. The lowest dilution with no visible growth was considered as MIC for each strain. The concentration at which 50% and 90% of the isolates were inhibited, as well as the minimum and maximum range were determined. The breakpoints suggested by the NCCLS (1999) for kanamycin were used for neomycin in both tests.

Oxacillin resistance testing
In order to confirm the presence of oxacillin resistance among staphylococci, VetMIC TM GP_mo panels (SVA, Uppsala, Sweden) were used as recommended by NCCLS (1999). The procedures were conducted following the manufacturer's recommendations: the inoculum was prepared with colony material directly from the plate incubated 24 h before. A 1 µl loop with colony material was suspended in 4 ml of distilled water plus 0.02% Twin 80. From this suspension 100 µl were transferred to 10ml Mueller Hinton Broth + 2% NaCl (Baker et al, 1994), which achieved about 10 3 to 10 4 cfu/50 µl. Each oxacillin and control well of the panel was inoculated with 50 µl of this final bacterial suspension. The panel was incubated at 30°C during 24 h under aerobic conditions. The strains S. aureus ATCC 29886 and S. aureus ATCC 29887 were included as negative and positive control strains, respectively.

ß-Lactamase Testing (Cloverleaf Method)
The assay to determine the production of ß-lactamase by staphylococci was described previously by Franklin & Wierup (1982). Briefly, the non-ß-lactamase-producing S. aureus Oxford strain 209 is used as indicator. This strain is inoculated on PDM II agar plates (AB Biodisk, Solna, Sweden) to yield an almost confluent growth on the agar surface. In the centre of the agar plate a disk containing 10 µg of penicillin G (PDM Antibiotics Sensitive II, AB Biodisk) is placed in order to induce ß-lactamase production in the studied strain. The staphylococci to be tested were streaked in a line from the edge of the plate towards the centre of the penicillin disk. When the investigated strain was positive ß-lactamase producer, the indicator strain grew alongside this strain towards the penicillin disk, into the inhibited one. The S. aureus strains ATCC 29213 and ATCC 25923 were included as positive and negative control respectively.

Statistics analyses
The Z-test (Milton 1992) was performed to compare the proportions of resistant strains to each antimicrobial agents obtained by means of both test.

Results
All values obtained with control strains in both tests were within the expected ranges for all antimicrobial agents analysed. The ranges of MIC of each of the antimicrobial agents tested, MIC 50 and MIC 90 of the tested strains, and the percentage of resistance obtained by both micro-dilutions and ADDM are presented here for S. aureus (Table 1), CNS (Table 2), Str. agalactiae (Table 3), Streptococcus dysgalactiae (Table 4), Streptococcus uberis (Table 5) and Enterococcus sp ( Table 6). The differences found between both tests corresponding to each antimicrobial agent were not significant (p > 0.05). Of 336 strains of S. aureus, 215 (64%) were resistant to one or more antimicrobial agents in both tests. There was no resistance to oxacillin, cephalotin, gentamicin, enrofloxacin, clindamycin, and the combination of amoxicillin-clavulanic acid, whereas 160 (47.6%), 157 (46.7%), 45 (13.4%), 10 (3%), 2 (0.6%) and 1 strain (0.3%) were resistant to penicillin, ampicillin, tetracycline, erythromycin, neomycin and trimethoprim-sulphametoxazole, respectively. One hundred and fifty-six S. aureus isolates (46.4%) were ß-lactamase producers. While of 41 CNS strains, 10 (27%) presented resistance to penicillin and 9   and narrow spectrum. This is an important tool to limit the development of antibiotic resistance as much as possible (Aarestrup & Jensen 1998). In our study 47.6% of S. aureus were classified as penicillin resistant, MIC ≥0.25 µg/ml ( Table 2), 96% of which produced ß-lactamase. This was the same comparing the proportion of resistance (47%) as obtained by Del Baglivi et al. (1976) in the southern dairy area of Uruguay. The comparison between these results obtained in Uruguay over the years demonstrated that the situation in general has not changed during the last 25 years in relation to penicillin resistance. Whereas, the prevalence of resistance to penicillin was similar in Argentina (40%) (Gentilini et al. 2000) and Finland (50.7%) (Myllys et al. 1998). However, it was higher than in Norway, 4.2% from clinical cases and 18% from sub-clinical cases (Hofshager et al. 1999), and Sweden, 6% (Franklin 1998). In relation to CNS, 27% of 41 isolates were penicillin resistant (Table 1). Results from Finland 37% (Myllys et al. 1998) and Norway 26% (Hofshager et al. 1999) agree with our findings. The MIC 90 of penicillin was 4 µg/ml for our survey, and another study determined 0.5 µg/ml in New Zealand (Salmon et al. 1998).
The detection of ß-lactamase production in staphylococci is a useful and rapid method to detect penicillin resistance. At the National Veterinary Institute, Uppsala, ß-lactamase results are used as rapid screen to indicate penicillin resistance (National Veterinary Institute 1998). In this study 96% and 90% of penicillin resistant strains of S.aureus and CNS were positive as indicated by the cloverleaf method. Test for ß-lactamase producing should always be done to obtain the true picture of resistance to penicillin in staphylococci. The streptococci and enterococci showed high susceptibility (streptococci 100%) to penicillin in our study (Tables 3, 4, 5 and 6). This agree with the results from monitoring studies done in the Scandinavian countries, where the streptococci populations isolated from mastitis were highly susceptible to penicillin (Pyörälä & Myllys 1995). Only 7% of the Enterococcus sp. strains were classified as resistant against penicillin (>8 µg/ml). The MIC 90 of penicillin for Str. agalactiae, Str. dysgalactiae, and Str. uberis was 0.12 µg/ml in each case and for enterococci 4 µg/ml . Oxacillin was included here as recommended by the NCCLS (1999) to detect methicillin-resistant strains of S. aureus and CNS. In our study oxacillin resistance was not found in staphylococci. However, CNS strains with higher MIC than >0.5 *g/ml of oxacillin should be tested for possible carriage of the mecA gene, in order to verify the occurrence of this gene (Hussain et al. 2000).
Cephalotin was included to determine the resistance against the first-generation cephalosporin class for all bacterial species except Enterococcus sp. All tested microorganisms were 100% sensitive to cephalotin and MIC 90 was ≤4 µg/ml. Despite their structural differences macrolide and lincosamides antimicrobials have similar biological properties, including their mechanism of action against the 50S subunit of the bacterial ribosome. These common properties easily allow the development of cross-resistance (Prescott 2000). Erythromycin and clindamycin were included here to evaluate the resistance against these groups. Clindamycin was used in our survey to test resistance againts lincosamides in staphylococci and no strains were resistant. For our strains the MIC 90 value was <1 µg/ml, a result remarkably different as compared with 8 µg/ml obtained by De Oliveira et al. (2000) for S. aureus strains isolated in the United States. For erythromycin our findings (Tables 1 and 2) showed scarce resistance in S. aureus (3%) and in CNS (0%), similar to the result (2.4%) reported by Del Baglivi et al. (1976). The results were lower than reported by Gentilini et al. (2000) for S. aureus in Argentina (11.6%). In Finland, Myllys et al. (1998) found 2.6% and 11.5% resistance among S. aureus and CNS, respectively, while in Sweden, Franklin (1998) reported 1% resistance in S. aureus. The MIC 90 of erythromycin in our study was ≤0.5 µg/ml to staphylococci. Streptococci showed high erythromycin sus-ceptibility, only 3.4% of Str. agalactiae and 4.6% of Enterococcus sp. were resistant in our study. Substantial differences were found in relation to results obtained in Finland (17%) for enterococci (Myllys et al. 1998), but no differences with respect to the erythromycin susceptibility result in streptococci (2,8%) obtained by Del Baglivi et al. (1976) (Myllys et al. 1998), but higher than the results obtained in Norway for S. aureus (0.2 %) and CNS (3%) (Hofshager et al. 1999). The results were twofold higher than the 6% reported by Del Baglivi et al. (1976) in Uruguay. A possible explanation for this phenomenon could be that for many years tetracyclines have been the most widely antimicrobial class used by the farmers to treat any infection.
In general the streptococci and enterococci were susceptible to oxytetracycline, with the exception of Str. dysgalactiae (Table 4).
Pyörälä & Myllys (1995) stated that Str. dysgalactiae strains are less susceptible to tetracyclyne than Str. uberis strains, as also reported by Brown & Roberts (1991). Staphylococci and streptococci were susceptible to trimethoprim-sulfamethoxazole, whereas enterococci were resistant (Table 6). Enrofloxacin is approved for systemic administration to treat bovine mastitis in some Scandinavian countries. We found a high susceptibility in staphylococci (Tables 1 and 2) and a similar situation was found by Myllys et al. (1998). Both antimicrobial susceptibility tests, ADDM and broth microdilution, used in this survey were performed according to the approved standard for bacteria isolated from animals and the interpretative criteria for veterinary use according to NCCLS (1999). The ADDM is most commonly used in the veterinary laboratories in Uruguay and many other countries. There were no significant differences between the methods when classifying bacterial isolates as susceptible or resistant according to NLCCS (Tables 1,  2, 3, 4, 5 and 6). The results from ADDM could be influenced by several factors, such as: compositions of agar medium, pH, inoculum density, agar depth, timing of drug applications, incubation time, etc (Acar & Goldstain 1996). However, Myllys et al. (1992) have obtained high correlation coefficient (0.875 to 0.975) between both methods in agreement with our results. Kielbauch et al. (2000) considered ADDM as a useful tool when the level of compliance with NCCLS guidelines was evaluated periodically.

Conclusion
This study did not show changes with respect to the penicillin and erythromycin resistance level of udder pathogens (staphylococci and streptococci) during the last 25 years in Uruguay, while a clear increase in tetracycline resistance was found for S. aureus. The Agar Disk Diffusion Method was a good tool, inexpensive, and readily available for regional veterinary laboratories. However, considering the necessity to maintain the surveillance over antimicrobial resistance in a country, it is important to periodically evaluate the compliance with guidelines such as National Committee for Clinical Laboratory Standards guidelines. It is also important to monitor regularly the minimum inhibitory concentrations for the isolated strains from different regions of the country. A responsible antibiotic policy would be highly relevant in a future programme for mastitis control and udder health in Uruguay.