INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Following the emergence of vancomycin (glycopeptide) resistance in enterococci, a discussion on the use of growth-promoting antibacterials has arisen. The fear of possible transfer of resistance genes from animal bacteria to human bacteria led to the ban on the use in animal feeds of most growth-promoting antibiotics in the European Community (6).

Reports of resistance in enterococci against growth-promoting agents are rare (1, 5, 15). Most data on resistance against growth promoters concern the prevalence of resistance against glycopeptides in enterococci from farm animals. Only one report has dealt with resistance rates in enterococci from pet animals against therapeutic antibiotics not used in animal feeds (21). However, only a limited number of strains and origins were investigated.

Since the prevalence of resistance against antibiotics in enterococci isolated from pet animals is largely unknown, we compared the resistance situation of enterococci from farm animals to that from pets. Most growth-promoting antibacterials are not used in the latter category of animals. The data obtained are discussed in respect to the major trends in antibiotic usage in the different animal groups. Resistance prevalences of enterococci from animals are compared to recent data on resistance prevalences among human enterococci.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains. A total of 146 strains of Enterococcus faecium and 166 strains of Enterococcus faecalis were isolated in Belgium from 1998 to 1999. They originated from the feces of different pet animals (divided into strains originating from mammalian pets and those from avian pets) and farm animals (broiler chickens, bovines, and fattening pigs), as listed in Table 1. Among mammalian pet animals were cats, dogs, hamsters, squirrels, monkeys, rabbits, rats, and horses. Among the pet birds were pigeons, geese, ducks, parrots, parakeets, swans, canaries, and some other passeriformes. There was no connection between the strains from pets and the strains from farm animals. Strains from pet animals were derived from pets attending the different clinics at the Faculty of Veterinary Medicine, University of Ghent, or were derived from animals brought in for necropsy at the Department of Pathology of the Faculty of Veterinary Medicine, University of Ghent. Each strain was representative of a single origin: a single farm or owner originating from the Flemish provinces of Belgium. No data on previous antibiotic usage (for therapy or for growth promotion) were collected. In general, Belgian farm animals are fed a diet containing a growth promoter. This study was performed after the banning of avoparcin (banned in 1997), and some strains were collected shortly after the banning of bacitracin, tylosin, spiramycin, and virginiamycin (banned in January 1999). Fecal samples were inoculated on Columbia CNA blood agar (Gibco, Paisley, United Kingdom), a selective medium containing colistin and nalidixic acid, and incubated overnight in air supplemented with 5% CO₂ or on Slanetz and Bartley agar (Oxoid, Basingstoke, United Kingdom). Enterococcus-like colonies were purified and identified as described earlier (12, 13, 14). Two control strains, E. faecalis ATCC 29212 and Staphylococcus aureus ATCC 29213, were included in the tests, as recommended in the National Committee for Clinical Laboratory Standards (NCCLS) standard procedures for testing therapeutic antibiotics (NCCLS M31-A) (18). The E. faecium type strain, LMG 11423^T, was included as an additional control.

Antibacterials. The growth-promoting antibacterials were obtained as laboratory standard powders. The following antibiotic preparations were tested: avoparcin (Roche, Basel, Switzerland), virginiamycin (Pfizer, Rixensart, Belgium), bacitracin (67,000 IU/g; Sigma, St. Louis, Mo.), tylosin (Sigma), avilamycin (Eli Lilly, Indianapolis, Ind.), and narasin (Eli Lilly). The antibiotics were dissolved in appropriate solvents to make stock solutions containing 1,000 µg/ml, or 1,000 IU (USP)/ml in the case of bacitracin, and then further diluted in sterile distilled water according to the methods recommended by the NCCLS (18). The antibiotics not listed in M31-A were avoparcin and bacitracin dissolved in distilled water, virginiamycin dissolved in ethanol, and avilamycin and narasin dissolved in methanol. Further dilutions were performed as prescribed by the NCCLS.

Susceptibility tests. MIC tests were carried out on Mueller-Hinton II agar (Becton Dickinson, Cockeysville, Md.) containing doubling dilutions of the antibiotics (7). Concentrations from 0.06 to 256 µg/ml were tested for all antibiotics except streptomycin and gentamicin, for which only the concentrations 500, 1,000, and 2,000 µg/ml were tested. Antibiotic-free agar plates were included as controls. Inocula were prepared by diluting overnight BHI (Oxoid) cultures in buffered saline to a density of 0.5 on the McFarland turbidity scale and were diluted 40-fold before inoculation. Plates were seeded with approximately 10⁵ CFU. The plates were incubated in air at 37°C for 24 h. The strains showing resistance to avoparcin were tested for the presence of van genes by using multiplex PCR for the detection of the vanA, vanB, vanC₁, and vanC₂ genes as described before (14). Strains for which the ampicillin MICs were 16 µg/ml were tested for the production of -lactamase by the nitrocefin test (Oxoid) (19). High-level resistance against streptomycin and gentamicin was defined as a MIC of >2,000 and 500 µg/ml, respectively, as indicated by the NCCLS guidelines (M31-A) (18). Breakpoints of other therapeutically used substances were derived from the same source.

Statistical analysis. Differences in prevalence of resistances among E. faecium and E. faecalis from different origins were analyzed by the Fischer exact test. A significance level of  0.05 was used.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

E. faecium and growth-promoting antibiotics. All strains were naturally resistant to flavomycin (MIC at which 90% of isolates are inhibited [MIC₉₀], >256 µg/ml), although strains from broiler chickens and swine had a wider MIC distribution range (Table 2) than strains from other origins. No resistance was found against monensin (MIC range, between 0.12 and 4 µg/ml). Three strains were resistant to the glycopeptide antibiotic avoparcin. One strain originated from a broiler chicken, one from a pet rabbit, and one from a bovine. Their resistance was mediated by the vanA gene, as demonstrated by PCR. Lowered susceptibility to narasin was mainly evident in strains from broiler chickens and swine. High avilamycin MICs were only found among the farm animal strains.

E. faecium and antibiotics used for both growth promotion and therapy. Tylosin-resistant strains were isolated from all origins. Significantly more resistant strains were found in farm animals than in pet animals (P < 0.001). Virginiamycin-resistant strains (MIC, 16 µg/ml) were isolated from broilers, swine, and a pet. Strains for which the bacitracin MICs were 16 IU/ml or higher were found mainly in the collections from broilers and swine.

E. faecium and therapeutic antibiotics. Ampicillin resistance was seen in strains from all origins except from pet birds. No -lactamase production was detected in these strains. Resistant strains were significantly more frequent in the collection from mammalian pets than in those from farm animals and from avian pets (P < 0.001). Resistance against streptomycin (MIC, >2,000 µg/ml) was not detected in strains isolated from ruminants. No statistically significant differences were seen among the different origins for this type of resistance. Gentamicin-resistant E. faecium strains (MIC, 500 µg/ml) were isolated from all origins but mainly from swine. Almost all strains from broilers, pigs, and ruminants were resistant to oxytetracycline, while strains from pet animals were significantly less often resistant to this antibiotic (P < 0.001). A monomodal distribution of the enrofloxacin MICs was obtained, ranging from 0.5 to 8 µg/ml. The enrofloxacin MIC for the E. faecium type strain (LMG11423^T [equivalent to ATCC 19434^T]) was consistently in the higher range (8 µg/ml).

E. faecalis and growth-promoting antibiotics. No resistance was found against the antibiotics flavomycin, avoparcin, and monensin. Resistance against narasin occurred in strains from all origins, but mainly in poultry strains. Avilamycin resistance was found in one strain each from a broiler chicken, a pig, and a duck.

E. faecalis and antibiotics used for both growth promotion and therapy. Resistance against tylosin in E. faecalis strains from all origins was high. No significant differences were noted among the different animal origins. High virginiamycin MICs were seen for strains from swine and broilers only. High bacitracin MICs occurred for strains from all origins, but strains from broilers were more often affected than strains from other origins.

E. faecalis and therapeutic antibiotics. No resistance was found against ampicillin. High-level streptomycin resistance (MIC, >2,000 µg/ml) occurred significantly more often among strains from swine than in strains from pets and ruminants (P < 0.01), and significantly more streptomycin-resistant strains were isolated from broiler chickens than from pet birds (P < 0.01). In contrast, high-level gentamicin resistance (MIC, 500 µg/ml) was not found among broiler strains. There were no significant differences in gentamicin resistance prevalence among strains from other origins. Tetracycline resistance was extremely frequent in strains from all origins. Not a single tetracycline-susceptible strain was present in the porcine collection. The enrofloxacin MIC₉₀ was 2 µg/ml, while a few strains were inhibited only by 8 to 64 µg/ml (Table 2), indicating the presence of resistance.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Although antibiotics used solely for growth promotion are not included in national and international standards for susceptibility testing, and universally recognized breakpoints are not available, certain indications are useful. This is notably the case with virginiamycin, bacitracin, and avilamycin. For virginiamycin, MICs of 16 µg/ml can be considered as indicating resistance, because resistance genes have been demonstrated only in E. faecium strains not inhibited by concentrations of 16 µg/ml and higher (16). The bacitracin MICs for E. faecium strains show a rather extended range. This indicates that different resistance mechanisms might be present. However, nothing is known about the mechanisms or about the genes encoding this type of resistance in enterococci. Avilamycin-resistant strains seemingly have two types of resistance: a low-level type and a high-level type. This was also evident in a Danish study (1). Resistance against everninomycin, a related antibiotic, has been shown to be mediated by a mutation in the ribosomal L16 protein in Streptococcus pneumoniae (2). If resistance in enterococci is also mediated by point mutations, a double mutation or a mutation combined with a specific resistance gene might explain the two levels of resistance observed in the present study.

No acquired resistance was found against flavomycin and monensin, two antibiotics used solely for growth promotion. Monensin results obtained in this study were similar to the results obtained in a Danish investigation (1). In the latter study, a few E. faecium strains for which the flavomycin MICs were low were reported. However, it must be noted that identification errors may occur when phenotypic identification is used (3), and it has been shown that several enterococcal species can be classified as naturally flavomycin resistant while others are naturally susceptible (4). The flavomycin MIC for the E. faecium type strain used as an additional control in the present investigation was consistently 256 µg/ml.

Only three E. faecium strains were resistant to avoparcin, a glycopeptide antibiotic with full cross-resistance with vancomycin (6). This relatively low resistance rate is due to the fact that no selective enrichment and isolation in media containing vancomycin was applied. This factor strongly influences the isolation frequencies of resistant strains (9).

Ionophores are included routinely in broiler feed for the prevention of coccidiosis. In swine, salinomycin is allowed for growth promotion, and in ruminants, monensin is used for the same purpose. These products are not used for pet animals. This might explain the differences seen between the enterococci from different animal origins. Narasin resistance, fully cross-resistant with salinomycin (8), was found only in E. faecium from farm animals, mainly broilers. Since the resistance mechanism and resistance genes are unknown, no specific additional test is available for confirming resistance. In contrast, narasin resistance in E. faecalis was present in strains from all origins, but resistant strains were isolated mainly from broilers.

The prevalences of resistances to antibiotics used both therapeutically and for growth promotion were generally much higher than those for resistances to the antibiotics used solely for growth enhancement. However, it should be noted that for the antibiotics virginiamycin and bacitracin, only topical preparations are available. These are only sporadically used in mammalian pet animals and not at all in pet birds. Bacitracin-resistant enterococci were found to occur in several animal species, including pet birds. The reason for this remains unclear. Tylosin and related macrolides are the only antibiotics in this group that are commonly applied in the therapy of pet and farm animals. This is reflected by the high resistance prevalence in pet animals as well as in farm animals. Nevertheless, significantly more tylosin-resistant E. faecium strains were isolated from farm animals than from pets, possibly reflecting the more extensive use of this antibiotic in farm animals.

For the therapeutically most important antibiotics in human enterococcal infections, certain differences were noted between the resistance situation in animal and in human clinical strains. The resistance rates among animal strains were lower than those found among human E. faecium strains isolated from hospital patients in a recent European study (20), while the rates among ruminant and mammalian pet strains were similar to those found in human strains in 1993 in Belgium (22). High-level streptomycin resistance, defined as a MIC of >2,000 µg/ml, has been reported to be generally more frequent than high-level gentamicin resistance, defined as a MIC of 500 µg/ml (17, 20). This was not evident in our study except for the poultry strains, in which gentamicin resistance was rare. The prevalence of gentamicin resistance in E. faecium strains from swine, ruminants, and avian pets was higher than that recently reported in human strains (20), while the opposite was true for E. faecalis.

Enterococcal infections are rare in animals and are not treated as such (10, 11). However, as intestinal inhabitants, enterocci are under selective pressure from every antibiotic administered that is active against them. Tetracyclines are still used frequently in animals, and resistance rates were very high in both the E. faecium and E. faecalis collections, especially in strains originating from swine. Only one porcine E. faecium strain was found to be susceptible, and all porcine E. faecalis strains were resistant. Enrofloxacin is only marginally active against E. faecium. Although all strains were in the intermediate or resistant category of the NCCLS susceptibility criteria, they cannot be considered as having acquired resistance, since not only the enrofloxacin MICs for the field strains were monomodally distributed. The breakpoint of this antibiotic is the median of the monomodal distribution of the MICs for the E. faecium strains tested here. Moreover, the MIC for the E. faecium type strain used was 8 µg/ml, which is above the breakpoint.

In conclusion, resistance against growth-promoting antibiotics and antibiotics used for both growth promotion and therapy, as well as resistance against therapeutic antibiotics, was present mainly in strains from food animals. -Lactam resistance, more frequently found in strains from mammalian pets, was the only exception in the last category of agents.