TEXT

Top Abstract Text References

A variety of antibiotics are applied in animal husbandry at subtherapeutical levels as antimicrobial growth promoters (AGPs). Previous studies have shown that the use of AGPs may select resistant bacteria among the normal intestinal flora of animals (2, 3, 6). In some countries, therapeutically important antimicrobial agents or related substances are licensed for use as AGPs. For instance, avoparcin, a member of the glycopeptide family, and tylosin, a member of the macrolide family, have been widely used in Europe as AGPs. These two substances have consequently been associated with a high prevalence of vancomycin-resistant enterococci (VRE) and of macrolide-resistant enterococci in the intestinal flora of pigs and poultry (2, 3, 6, 8). The transmission of bacteria between animals and humans is not limited to agents of zoonotic diseases. Therefore, the selection of a reservoir of resistant opportunistic human pathogens and of possibly transmissible resistance determinants through the use of AGPs may have undesirable consequences for human health. For these reasons, and with the noticeable exception of tylosin, all the antimicrobial agents belonging to substance classes used for therapeutic treatments in humans (including tetracyclines) have been forbidden as AGPs since 1972 in Switzerland. For the same reasons, the use of avoparcin was also forbidden in Switzerland and in the European Union in February and April 1997, respectively. Finally, the use of any kind of AGPs was completely forbidden in Switzerland at the beginning of 1999, with a transitional period for completion of the ban ending on 30 June 1999.

In order to investigate the short-term effect of the AGP ban on antimicrobial resistance among fecal enterococci from pigs in Switzerland, a study was undertaken in farms using tylosin as an AGP before and after the complete cessation of AGP use. The evolution of resistance to macrolides, to vancomycin, and to other therapeutic antimicrobial agents not used as AGPs was examined. After the farmers were informed of the aim of the study and the procedures to be used, fecal samples from 16 farms that were still using tylosin as an AGP shortly before the AGP ban were investigated in a first step (farmers and feed producers were contacted to verify that tylosin was used regularly at this time). Samples from 13 of these farms were obtained and examined again 183 to 287 days later (median, 225 days). These samples were obtained 153 to 259 days (median, 161 days) after the final compulsory cessation of AGP use at the end of June 1999. For this purpose, two fresh fecal samples collected by the farmers on the same day in two different pens were sent immediately to the laboratory, mixed together in equal proportions, and inoculated on selective agar plates as follows. The fecal samples were homogeneously resuspended in a 1/5 proportion in brain heart infusion broth (Difco Laboratories, Detroit, Mich.), and approximately 200 µl of this suspension was used for isolation of enterococci on kanamycin-esculin azide agar (Merck, Darmstadt, Germany), on KF-streptococcus agar (Merck), and on Enterococcosel agar (Becton Dickinson Co., Cockeysville, Md.), following the instructions of the manufacturers. Four to six enterococcus-like colonies randomly chosen to represent all the morphologies observed on the plates were further examined for each sample. For samples with homogeneous colonies, a total of four to six colonies were randomly selected from the different selective media used. Esculin-positive, catalase-negative, and gram-positive cocci were identified at the species level using Rapid ID32 Strep galleries following instructions of the manufacturer (BioMerieux SA, Marcy-l'Etoile, France). A total of 96 and 59 Enterococcus sp. isolates were obtained before and after the AGP ban, respectively (Table 1). These 155 isolates consisted of 72 Enterococcus faecium isolates, 63 Enterococcus hirae isolates, 10 Enterococcus faecalis isolates, 7 Enterococcus gallinarum isolates, 2 Enterococcus durans isolates, and 1 Enterococcus casseliflavus isolate. Susceptibility testing was performed for all the isolates by disk diffusion using erythromycin, spiramycin, clindamycin, vancomycin, penicillin, tetracycline, chloramphenicol, high-level streptomycin, and high-level gentamicin disks (Sanofi Diagnostics Pasteur, Marnes-la-Coquette, France) following the NCCLS standard protocols for enterococci (9). Except for spiramycin and clindamycin, the interpretation criteria used were those of the NCCLS norm (10). For spiramycin (100-µg disks), strains with zone diameters smaller than 16 mm, between 16 and 22 mm, or larger than 22 mm were interpreted as resistant, intermediate, or susceptible, respectively. For clindamycin (2-µg disks), strains with a zone diameter smaller than 16 mm or larger than 18 mm were considered resistant or susceptible, respectively. The overall susceptibility to these agents before and after the ban as well as for the major species E. faecium and E. hirae are reported in Table 1. A clear decrease in resistance of enterococci to erythromycin, spiramycin, and clindamycin was visible at all levels in samples obtained after the ban. A strong association was observed between resistance to macrolides and resistance to tetracycline (P < 0.0001 for chi-square test and 95% confidence interval for odds ratios of 8.1 to 43.8 for spiramycin resistance, 3.4 to 16.1 for erythromycin resistance, and 2.1 to 16.5 for clindamycin resistance). A decreasing trend was consequently also visible for tetracycline resistance after the ban of AGPs (Table 1).

Four vancomycin-resistant E. faecium isolates (VRE) were recovered from two farms during the present study. Three of them were obtained from a single farm (two in the first fecal sample and one in the second sample, obtained 224 days later). All four VRE isolates were simultaneously resistant to penicillin, erythromycin, spiramycin, clindamycin, and tetracycline. The presence of the vanA gene in all four isolates was confirmed by PCR following the method of Dutka-Malen and collaborators (7). One VRE isolate from each of the two farms was examined for the presence of the ermB and tet(M) genes and transferability of vancomycin resistance as previously described (1). Both isolates contained the tet(M) and ermB genes, and transfer of vancomycin resistance was associated with transfer of erythromycin resistance but not tetracycline resistance. The four VRE and four other epidemiologically unrelated E. faecium isolates from four other pig farms were typed by pulsed-field gel electrophoresis (PFGE). DNA plugs were prepared by standard procedures, and the DNA was digested using the restriction enzyme SmaI. The electrophoresis was done in SeaKem gold agarose (FMC Bioproducts, Rockland, Maine) and 0.5× Tris-borate-EDTA at 12°C in a CHEF III electrophoresis unit (Bio-Rad, Hercules, Calif.). An electrical field of 6 V/cm with an angle of 120° and pulses ranging from 1 to 25 s was applied to the gel for a period of 16 h. The results presented in Fig. 1A show that all three VRE isolates from the same farm were identical (lanes 2 to 4) and clearly had more bands in common with the fourth VRE isolate of the other farm (lane 1) than with vancomycin-susceptible isolates originating from unrelated farms (lanes 6 to 9). The PFGE profiles of the VRE isolates found in the present study were compared to those of representative VRE from Danish pigs (1). Their profiles shared many bands, and the profiles of the two Swiss VRE isolates differed in only one to five bands from the profiles of the Danish VRE isolates (Fig. 1B), thus showing that VRE from pigs in Switzerland and in Denmark belong to the same E. faecium clonal lineage. A fourth E. faecium isolate was recovered after the ban of AGPs in the Swiss farm where the three VRE were found. Similarly to the three VRE, this isolate was resistant to penicillin, erythromycin, spiramycin, clindamycin, and tetracycline but not to vancomycin, and it lacked the vanA gene. The PFGE profile of this particular vancomycin-susceptible isolate (Fig. 1A, lane 5) presented clear similarities with the VRE isolates. Hence, E. faecium isolates belonging to the particular clonal lineage containing VRE isolates from Swiss and Danish pigs are not systematically resistant to vancomycin.

View larger version (73K): [in this window] [in a new window]   FIG. 1.   SmaI restriction profiles of vancomycin-resistant and -susceptible E. faecium isolates. (A) M, molecular weight marker. Lanes 1 to 4, vancomycin-resistant isolates from Switzerland; lanes 5 to 9, vancomycin-susceptible isolates. The isolates in lanes 2 to 5 came from the same farm, and those in lanes 6 to 9 came from four epidemiologically unrelated farms. (B) M, molecular weight marker. Lanes 1 and 2, same isolates as in panel A; lanes 3 to 6, representative vancomycin-resistant isolates from Danish pigs. Note that the Swiss isolate in lane 2 differs in only one to five bands from the Danish isolates.

Data on the evolution of antimicrobial resistance after discontinuation of AGP use are scarce (4). The present study suggests that the complete AGP ban applied in Switzerland had a relatively rapid effect on antimicrobial resistance in enterococci and support similar data obtained recently in Denmark (4). Resistance to macrolides and lincosamides drastically decreased shortly after the ban enforcement. In addition, a parallel decrease in tetracycline resistance was observed, which suggests that a ban on AGPs may also have positive effects on the susceptibility of enterococci to unrelated therapeutically important antimicrobial agents not used as AGPs (Table 1). These results strongly support the precautions taken by European countries in order to reduce the frequency of antimicrobial resistance in bacteria of general public health relevance from animals. Despite the avoparcin ban in 1997, VRE could still be isolated without the use of any vancomycin-containing selective culture media in Swiss pigs at the end of 1999. This finding suggests that VRE are still present at a significant level in Swiss pigs more than 2 years after the discontinuation of avoparcin use. A similar observation had been made with Danish pigs after the ban of avoparcin (4, 5). Because of the frequent presence of macrolide and glycopeptide resistance genes on the same genetic element in E. faecium isolates from pigs (1), this persistence may be due to a selection of VRE by other AGPs (tylosin was used as an AGP in Switzerland until June 1999). The recovery of the same VRE clone in the same farm at an interval of more than 7 months without any avoparcin selection but with initial tylosin use supports this hypothesis. One can therefore hope that the total ban of all AGPs enforced in Switzerland and in some other European countries may be a more effective measure in reducing the frequency of VRE and multiresistant enterococci in general in the animal reservoir than a ban of only selected AGPs.

Interestingly, the VRE isolates found in Swiss pigs in the course of the present study belonged to the same broad E. faecium clonal lineage as VRE from Danish pigs. A comparison with epidemiologically unrelated vancomycin-susceptible isolates shows that the observed similarities between Swiss and Danish VRE is not due to a lack of discrimination among E. faecium isolates from pigs but is indicative of a true genetic and probably epidemiological relatedness. Therefore, the present report suggests that, as in the case of the multiresistant clone of Salmonella enterica serovar Typhimurium phage type DT104, a broad VRE clone apparently associated with pigs (1) has spread among pigs in different countries and may be widely distributed in Europe. The epidemiology of antimicrobial resistance in commensal bacteria from the animal reservoir and the effects of the use of AGPs should therefore be examined not only at the local level but also at a more global and international level. Finally, the exact reasons for the apparent confinement of the vancomycin resistance determinants to this single E. faecium clonal lineage in pigs despite their documented transferability (1) remain to be elucidated.