- Email: [email protected]
Antimicrobial resistance in Campylobacter from broilers: association with production type and antimicrobial use
Veterinary Microbiology 96 (2003) 267–276
Antimicrobial resistance in Campylobacter from broilers: association with production type and antimicrobial use Laetitia Avrain a , Florence Humbert b , Rolande L’Hospitalier c , Pascal Sanders d , Christine Vernozy-Rozand e , Isabelle Kempf a,∗ a b
Unité de Mycoplasmologie—Bactériologie, Agence française de Sécurité Sanitaire des Aliments, BP 53, F-22440 Ploufragan, France Unité d’Hygiène et Qualité des Produits Avicoles et Porcins, Agence française de Sécurité Sanitaire des Aliments, BP 53, F-22440 Ploufragan, France c Service des Systèmes d’Information, Agence française de Sécurité Sanitaire des Aliments, BP 53, F-22440 Ploufragan, France d Agence française de Sécurité Sanitaire des Aliments, F35302 Fougères, France e Unité de Microbiologie Alimentaire et Prévisionnelle, Ecole Nationale Vétérinaire de Lyon, BP 83, F-69280 Marcy-l’Etoile, France Received 27 February 2003; received in revised form 14 July 2003; accepted 25 July 2003
Abstract The isolation and antimicrobial resistance of Campylobacter jejuni and Campylobacter coli strains from broilers arriving in French slaughterhouses, were analysed according to production types (i.e. standard, export or free-range) and antimicrobial (i.e. coccidiostats, growth promoters or therapeutic agents) administration in flocks. Prevalence was 56.6% in standard, 51.3% in export and 80.0% in free-range broilers. Three hundred and ninety-three strains were identified. Two-thirds of the strains belonged to the species C. jejuni. The others were C. coli. Antimicrobial susceptibility testing was carried out for ampicillin, nalidixic acid, enrofloxacin, tetracycline, erythromycin and gentamicin according to a dilution method. The percentages of resistant strains were, 23, 25, 17, 57, 0.3 and 0% for C. jejuni and 29, 43, 40, 70, 31 and 0% for C. coli. Statistical analysis revealed significant difference in distribution of C. jejuni and C. coli and antimicrobial resistance according to production type or antimicrobial administration. © 2003 Elsevier B.V. All rights reserved. Keywords: Antimicrobial resistance; Campylobacter spp.; Chicken
∗
Corresponding author. Tel.: +33-2-96-01-62-81; fax: +33-2-96-01-62-73. E-mail address: [email protected] (I. Kempf). 0378-1135/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.vetmic.2003.07.001
268
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
1. Introduction Campylobacter is recognised as a commonly encountered microbe responsible of diarrhoeal diseases and foodborne gastro-enteritis (Solomon and Hoover, 1999; Uyttendaele et al., 1999). Campylobacter infections usually occur following ingestion of undercooked animal products, raw milk or contaminated water (Solomon and Hoover, 1999). One of the main reservoirs seems to be the gastro-intestinal flora of chickens (Ermel et al., 1997; Shane, 2000; Solomon and Hoover, 1999). Human Campylobacter jejuni infection may be quite asymptomatic or on the contrary may provoke inflammatory diarrhoeas and different serious sequels such as Guillain-Barré syndrome or Reiter syndrome (Shane, 2000). In patients in whom treatment is necessary, fluoroquinolones or macrolides are currently used (Le Noc et al., 1993; Ruiz et al., 1998; Binotto et al., 2000). Thus data concerning antimicrobial resistance in recent Campylobacter isolates are needed. This study was realised as part of a French surveillance programme of resistance in commensal bacteria (Escherichia coli and Enterococcus) and zoonotic bacteria (Salmonella and Campylobacter) of poultry. France is one of the main poultry producing countries in the European Union. The broiler production is divided into three main types of management. Standard and export chickens are bred inside farm buildings. They are, respectively, slaughtered at approximately 35 and 30 days of age. Free-range chickens have access to an outdoor hen run during at least half of the breeding period and are slaughtered at 80 days of age. In December 1998, the European Council voted in favour of the ban of four antimicrobial growth promoters, because of the fear that continued use of growth promoters could weaken the efficacy of certain antimicrobials in human medicine (European Economic Community, 1998). Flavophospholipol, avilamycin and coccidiostats ionophores such as monensin, salinomycin or synthesised products such as nicarbasin, and diclazuril remained authorised in poultry productions. In this context, the isolation frequency and the antimicrobial susceptibility of C. jejuni and Campylobacter coli isolated from French broilers were determined and compared according to the type of flock management and to consumption of antimicrobial agents used either for growth promotion, prophylactic or therapeutic purposes.
2. Materials and methods 2.1. Bacterial sampling During the 6-month period from January to July 1999, 600 chicken’s caecal samples were obtained during the evisceration process in 10 French slaughterhouses. The selection was representative of the three main types of French production: standard (65%), export (25%) and free-range (10%) broilers. The meat inspection staff who collected the samples also recorded data concerning antimicrobial consumption of the flock reported on the breeding document accompanying birds to slaughterhouse (coccidiostats, growth promoters or antimicrobial treatments). Only one chicken per flock was included.
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
269
2.2. Campylobacter isolation All Campylobacter cultures were incubated under microaerobic atmosphere (Campypak H2 CO2 , AES, Combourg, France). On arrival in the laboratory, 10 g of each cloacal content were weighted and diluted 1/10 in 25% glycerol peptone broth (GPB) (AES). A 10 l of this suspension was then streaked on Karmali (AES) and Butzler no. 2 (Merck, Fontenay-sous-Bois, France) selective media. After 48–72 h of incubation at 42 ◦ C, one characteristic colony was selected and streaked on each medium and incubated for 24–48 h. A total of 816 strains were thus obtained and stored in GPB at −70 ◦ C. Four hundred and eight randomly selected strains were thawed, streaked on Mueller-Hinton agar plate (AES) supplemented with 5% sheep blood (MHSB), and incubated for 48–72 h. For each MHSB plate, one characteristic colony was isolated on MHSB plate for purification, incubated for 24–48 h and this culture was used for identification and antimicrobial susceptibility testing. 2.3. Identification by m-PCR Identification of the genus Campylobacter and of C. jejuni and C. coli, the two main species isolated from poultry was performed using a recently developed multiplex PCR (m-PCR) with slight modifications (Denis et al., 1999; Thomas et al., 1999). About 8–10 colonies were suspended in 200 l of 10 mM Tris–HCl and 1 mM EDTA (TE 10.1 buffer), and then DNA preparation was performed by heating the cells at 95 ◦ C for 10 min. After centrifugation at 2000 × g for 2 min, 10 l of the supernatant were diluted in 990 l of TE 10.1 buffer and stored at −20 ◦ C. Briefly, the m-PCR consisted of a simultaneous amplification in the same tube, of several DNA fragments. Two sets of primers were used to identify the species C. jejuni and C. coli: MDmapA1 and MDmapA2, specific of the mapA gene to identify a 589 bp fragment for C. jejuni and COL3 and MDCOL2, specific of the ceuE gene to identify a 462 bp fragment for C. coli. Amplification reactions were performed in a 30 l mixture containing 3 M of dNTPs solution (Eurobio, Les Ulis, France), 0.11 M of MD16S1, MD16S2, 0.42 M of MDmapA1, MDmapA2, COL3 and MDCOL5 (Oligo express, Paris, France), 0.8 U of Taq polymerase 5 U/l (Roche Diagnostics, Meylan, France), 3 l of buffer 10× with MgCl2 (300 M Tris–HCl; 45 M MgCl2 ; 1500 M KCl) and 3 l of lysate. Amplification reactions were performed in a Perkin Elmer 9600 thermocycler: 10 min at 95 ◦ C, 35 cycles each consisting of 30 s at 95 ◦ C, 1 min 30 s at 59 ◦ C, 1 min at 72 ◦ C, one final cycle for 10 min at 72 ◦ C and a cooling down to 4 ◦ C. Ten microlitres of each amplified product were analysed on a 1.5% agarose gel (Eurogentec, Angers, France) containing 0.3% ethidium bromide and after 2 h at 100 V of migration, the fragments were revealed under UV light (302 nm). Only isolates identified as C. jejuni and C. coli were subjected to antimicrobial susceptibility testing. 2.4. Antimicrobial susceptibility testing Antimicrobial susceptibility testing was done according to the NCCLS document M7-A4. Six antimicrobials were tested: ampicillin (2–32 g/ml) (Sigma, Saint-Quentin-Fallavier,
270
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
France), nalidixic acid (2–128 g/ml) (Sigma), enrofloxacin (0.25–16 g/ml) (Bayer, Puteaux, France), tetracycline (0.25–16 g/ml) (Sigma), erythromycin (0.25–32 g/ml) (Sigma) and gentamicin (0.03–16 g/ml) (Sigma). Reference strains (E. coli (ATCC 25922), Staphylococcus aureus (ATCC 29213), Pseudomonas aeruginosa (ATCC 27853) and Enteroccocus faecalis (ATCC 29212)) and Campylobacter strains were suspended in 4 ml Mueller-Hinton broth (DIFCO, BD Biosciences, Le Pont de Claix, France), incubated at 37 ◦ C for 16–20 h, adjusted to 0.5 McFarland with a densitometer and diluted 1/10 in water. One microlitre of each suspension containing about 104 Colony Forming Units/spot was inoculated on each Mueller-Hinton with 5% sheep blood plate supplemented with antimicrobial agent with a multipoint inoculator and incubated at 37 ◦ C for 48 h (Aarestrup et al., 1997). The Minimum Inhibitory Concentration (MIC) was defined as the lowest concentration that achieved inhibition of visible bacterial growth. Results were recorded after checking that the MIC observed for the control strains corresponded to expected values, previously determined for each strain and each antimicrobial, on Mueller-Hinton agar and under microaerobic conditions. Susceptibility categorisation for fast-growing aerobic bacteria was carried out for Campylobacter according to the statement 1999 of the Antibiogram Committee of the French Society for Microbiology. 2.5. Statistical analysis Data concerning flocks (type of production, consumption of antimicrobial growth promoters or treatments) were recorded in an Open Access 1997 database. Distributions of species or of resistant strains were compared by means of the χ2 test and Fisher’s exact test. A significance level of 5% was used.
3. Results 3.1. Antimicrobial consumption Consumption of the different antimicrobial families in the three production types is summarised in Table 1. Samples and antimicrobial consumption data were obtained during the first semester of 1999, a transitional period for the use of antimicrobial growth promoters in France. According to the Council Regulation (EC) No. 2821/98 of 17 December 1998, bacitracin zinc, spiramycin, virginiamycin and tylosin phosphate remained authorised in France until 30 June 1999. Indeed the breeding documents reported that the use of these compounds as feed growth promoter was already very rare in export production but up to 19% of standard flocks had received bacitracin. In free-range production, according to French legislation, the use of antimicrobials as growth promoters was not reported (Approbation interministérielle par arrˆeté du 22 février, 1996). Avilamycin, one of the two remaining authorised compounds, was the most commonly administrated growth promoter in export production whereas bacitracin was predominant in standard production. Therapeutic administration of macrolides was never mentioned. According to data recorded by meat inspection staff, all the flocks had received one or more coccidiostat. Reported data concerning their use were consistent with
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
271
Table 1 Antimicrobial consumption according to production type Production type Standard, N = 253
Export, N = 83
Free-range, N = 57
Coccidiostats
Ionophore Synthesis
36.4a 18.4
79.5 78.3
52.6 54.4
Growth promoters
Bacitracin MLS as growth promotersb Avilamycin Flavophospholipol
19 4.7 39.1 4.7
0 1.2 79.5 0
0 0 0 0
Antimicrobials used as treatment
-Lactams Quinolones Tetracyclines Colistin Trimethoprime-sulphonamides
2.8 6.7 11.9 5.9 2.8
13.2 32.5 24.1 10.8 14.5
12.3 0 8.8 7 8.8
a b
Percentage of treated flocks. MLS: macrolides, lincosamides or streptogramins given as growth promoters.
previous observations (J.M. Repérant, AFSSA Ploufragan, personal communication). Therapeutic administration of macrolides was never mentioned. Tetracyclines (oxytetracycline and doxycycline) and quinolones (mainly oxolinic acid) were the two most frequently administrated antimicrobial families given for therapy. 3.2. Prevalence and identification of Campylobacter species During the sampling period, 816 bacterial strains were isolated on Karmali or Butzler no. 2 media from 600 caecal samples. One or two strains per sample were randomly selected for further analysis. Seven isolates did not grow after thawing. According to the m-PCR test, six strains belonged to other bacterial genera, 297 strains were identified as C. jejuni (75.6%), 96 as C. coli (24.4%) (Table 2), and 2 showed atypical profiles. The frequency of isolation of Campylobacter strains was 56.6% in standard broilers (221 out of 390 samples), 51.3% in export broilers (77 out of 150 samples) and 80.0% in free-range broilers (48 out of 60 samples). 3.3. Distribution of Campylobacter species according to type of production and antimicrobial administration C. jejuni was the predominant species in standard and export productions (81.4 and 88.0%, respectively), whereas the frequency of this species was significantly lower in free-range production (31.6%) (P ≤ 0.05). Within each production, links between use of antimicrobial agents and distribution of Campylobacter species were examined. In the standard production, the percentage of C. coli was significantly higher (28.6% (26/91)) when birds were
272
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
273
given ionophores compared to 13.0% (21/162) when ionophores were not administrated (P ≤ 0.05). In the same production, when birds were given colistin, the percentage of C. coli strains reached 46.6% (7/15) compared to 16.8% (40/238) in the absence of this compound. In the export production, the percentage of isolation of C. coli was significantly higher (4/11 vs. 6/72) when birds received -lactams (P ≤ 0.05). In the free-range production, no significant effect of administration of antimicrobials on the distribution of Campylobacter species could be observed. 3.4. Antimicrobial susceptibility The distributions of MICs of C. jejuni and C. coli strains are listed in Table 2. The percentages of strains resistant to nalidixic acid, enrofloxacin and erythromycin were significantly higher for C. coli than for C. jejuni (P ≤ 0.05). All strains were sensitive to gentamicin whereas more than 50% of the strains were resistant to tetracycline. The percentages of C. jejuni and C. coli resistant to -lactams or fluoroquinolones ranged from 17 to 40%. Most C. jejuni strains were sensitive or intermediately sensitive to erythromycin. 3.5. Distribution of resistant strains according to production types Statistical tests were performed to research significant differences of distribution of resistant strains of C. jejuni and C. coli according to production type. For one antimicrobial only, i.e. tetracycline, differences of distribution were observed. Eighty-one percent and 90% of C. coli strains isolated from standard and export productions, respectively, were tetracycline resistant compared to 51% in free-range production. 3.6. Distribution of resistant strains according to antimicrobial agents administrated in flocks In standard production, 55/85 (64%) flocks treated with avilamycin were contaminated by tetracycline resistant C. jejuni strains compared to 58/121 (48%) in non-treated flocks (P < 0.05). In the same production, when ionophores were used, the percentage of nalidixic acid resistant C. coli strains (16/26 (61%)) was significantly higher than in absence of ionophores (6/21, 29%) (P ≤ 0.05). In the free-range production, when synthetic anticoccidials were used in association with ionophores, the percentages of ampicillin, nalidixic acid and tetracycline resistant C. jejuni strains reached 7/10 (70%), 8/10 (80%) and 7/10 (70%), respectively, whereas the rates were 0% when flocks were not treated. In the same production the percentage of ampicillin and tetracycline resistant C. coli was also significantly higher when birds were given coccidiostats. In export production, the percentage of enrofloxacin resistant C. jejuni was 37.5% (6/16) when strains were isolated from flocks receiving tetracycline as compared to 12% (7/57) in non-medicated flocks (P ≤ 0.05). Considering the three productions together, the percentage of Campylobacter tetracycline resistant strains was significantly higher when broilers were treated with oxytetracycline or doxycycline (P ≤ 0.05).
274
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
4. Discussion This study is unusual in that it relates antimicrobial resistance of C. jejuni and C. coli in three different types of production and particularly to different patterns and usage of antimicrobial drugs. The frequency of isolation of Campylobacter was significantly higher in free-range than in the two other productions (P ≤ 0.05). Considering the total French production, C. jejuni was more frequent than C. coli in sampled broilers, as was previously observed by Aarestrup et al. (1998) and by Saenz et al. (2000) in other countries. However in free-range production, the ratio C. jejuni/C. coli was inverted. The high frequency of isolation and the predominance of C. coli in the free-range production had already been reported by Rivoal (2000) and could be attributed to very different breeding conditions such as possible outside runs, non-authorisation of antimicrobial growth promoters, limited use of antimicrobials for therapy and age at slaughter. The increase of the relative frequency of C. coli in the case of administration of -lactams could result from a lower susceptibility of this species to these compounds, as suggested by the MIC50 or MIC90 of ampicillin for the two Campylobacter species. Links between increased occurrence of C. coli and administration of ionophore coccidiostats or colistin in flocks could not be explained. The number of strains studied (393) was chosen to enable the detection of one resistant strain with a probability of 0.99 for a resistance level of 5%. This sampling level was too low to detect resistant strains only in the case of gentamicin. Our results were based on analysis of only one or two isolates per flock, whereas it has previously been demonstrated that flocks may harbour several strains of each species (Rivoal, 2000). In this study, when two C. coli strains were isolated from the same chicken, they shared a similar MIC pattern whereas two C. jejuni isolates obtained from the same sample could exhibit different MIC profiles (data not shown). The susceptibility of Campylobacter strains isolated from broilers may not reflect exactly the susceptibility of strains obtained from food or human pathology. Indeed Saenz et al. (2000), in Spain, showed that the percentages of C. coli strains resistant to clindamycin, ampicillin and fosfomycin were higher in human than in broiler isolates whereas for other antimicrobial agents, Campylobacter resistance was usually less frequent in human strains. Such different resistance profiles may result from a possible filter effect of food transformation processes, differences in pathogenicity or resistance selections in human flora. However, clones of mild pathogenicity may contain resistance genes (i.e. tetracycline, aminoglycosamides), transferable to other more virulent Campylobacter strains or to other pathogenic bacterial species (Taylor and Courvalin, 1988). Comparison with other published studies revealed that the percentage of quinolone resistant strains was higher in France than in Denmark but lower than in Spain (Aarestrup et al., 1998; Saenz et al., 2000). The frequency of tetracycline resistance was high in France whereas this compound seemed still rather active against Campylobacter in Spain or in Denmark. Spanish C. coli strains were frequently reported resistant to gentamicin. Differences in MIC determination methods or categorisation breakpoints may partly explain these differences. As drug use is usually regarded as a major source of drug resistance, the most obvious associations (resistance to ampicillin, erythromycin, enrofloxacin and tetracycline) in case of
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
275
administration of -lactams, macrolides, fluoroquinolones and tetracycline were examined but revealed only significant increase of tetracycline resistance when birds were treated with tetracycline (Aarestrup, 1999). The association between use of avilamycin in export production and resistance to tetracycline is noteworthy. Avilamycin is a compound of the everninomycin family, which binds to the ribosome (Bryskier, 1999). Until recently, an everninomycin product was under development for human therapeutic use, but this everninomycin product is no longer being developed. However, the study of the susceptibility of some of our strains to avilamycin was conducted in order to investigate the relations between resistance to these two ribosome-binding agents. No correlation between resistance to tetracycline and resistance to avilamycin could be demonstrated (data not shown). Results showed that the frequency of Campylobacter isolation, the C. jejuni/C. coli ratio or the antimicrobial resistance depended on the type of husbandry. Use of several antimicrobial agents, either as growth promoters or as treatment seemed to influence the resistance level of Campylobacter but confounding factors could not be disregarded. The influence of the European ban of the antimicrobial growth promoters might not be obvious immediately after their withdrawal and it will be important to monitor the trends in resistance over time to assess of antimicrobial resistance related to the use of antimicrobials in animal production.
Acknowledgements The authors are grateful to J.M. Repérant for information concerning coccidiostats and to M. Le Fellic, F. Eono and S. Queguiner for technical help. This work was supported by DGAl grant S98/29 from the French Ministry of Agriculture and Fishery.
References Aarestrup, F.M., Nielsen, E.M., Madsen, M., Enberg, J., 1997. Antimicrobial susceptibility patterns of thermophilic Campylobacter spp. from Humans, pigs, cattle, and broilers in Denmark. Antimicrob. Agents Chemother. 41, 2244–2250. Aarestrup, F.M., Bager, F., Jensen, N.E., Madsen, M., Meyling, A., Wegener, H.C., 1998. Surveillance of antimicrobial resistance in bacteria isolated from food animals to antimicrobial growth promoters and related therapeutic agents in Denmark. APMIS 106, 606–622. Aarestrup, F.M., 1999. Association between the consumption of antimicrobial agents in animal husbandry and the occurrence of resistance bacteria among food animals. Int. J. Antimicrob. Agents 12, 279–285. Antibiogram Committee of the French Society for Microbiology, 1999. Official statement 1999. Pathol. Biol. 47, 845–872. Approbation interministérielle par arrˆeté du 22 février, 1996. J.O. du 25 février. Binotto, E., McIver, C.J., Hawkins, G.S., 2000. Ciprofloxacin-resistant—Campylobacter jejuni infections. Med. J. Aust. 172, 244–245. Bryskier, A., 1999. Everninomycines. In: Bryskier, A. (Ed.), Antibiotiques, agents antibactériens et antifongiques. Ellipses ed., Paris, p. 1060. Denis, M., Soumet, C., Rivoal, K., Ermel, G., Blivet, D., Salvat, G., Colin, P., 1999. Development of a m-PCR assay for simultaneous identification of Campylobacter jejuni and C. coli. Lett. Appl. Microbiol. 29, 406–410. Ermel, G., Laisney, M.J., Rivoal, K., Ragimbeau, C., Gorin, S., Salvat, G., Colin, P., 1997. Les Campylobacter et la filière avicole. Sci. Techn. Avicoles. 20, 3–6.
276
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
European Economic Community, 1998. Council Regulation (EC) No. 2821/98 of 17 December 1998 on authorisation ban of certain antibiotics concerning the additives in animals food. Off. J. Eur. Communities L351/4. Le Noc, P., Croizé, J., Le Noc, D., 1993. Activité Antibactérienne Comparée d’une Nouvelle Fluoroquinolone, la Sparfloxacine (AT 4140, RP 64206) et de 4 autres Fluoroquinolones sur 332 Souches de Bactéries entéropathogènes. Pathol. Biol. 41, 294–301. Rivoal, K., 2000. Les Campylobacter dans la filière avicole. Caractérisation génomique et origine de la contamination dans les élevages. Ph.D. Thesis. Université de Bretagne Occidentale. Ruiz, J., Goni, P., Marco, F., Gallardo, F., Mirelis, B., Jimenez De Anta, T., Vila, J., 1998. Increased resistance to quinolones in Campylobacter jejuni: a genetic analysis of gyrA gene mutations in quinolone-resistant clinical isolates. Microbiol. Immunol. 42, 223–226. Saenz, Y., Zarazaga, M., Lantero, M., Gastanares, M.J., Baquero, F., Torres, C., 2000. Antibiotic resistance in Campylobacter strains isolated from animals, foods, and Humans in Spain in 1997–1998. Antimicrob. Agents Chemother. 44, 267–271. Shane, S.M., 2000. Campylobacter infection of commercial poultry. Rev. Sci. Tech. Off. Int. Epiz. 19, 376–395. Solomon, E.B., Hoover, D.G., 1999. Campylobacter jejuni: a bacterial paradox. J. Food Safety 19, 121–136. Taylor, D.E., Courvalin, P., 1988. Mechanisms of antibiotic resistance in Campylobacter species. Antimicrob. Agents Chemother. 32, 1107–1112. Thomas, C., Gibson, H., Hill, D.J., Mabey, M., 1999. Campylobacter epidemiology: an aquatic perspective. J. Appl. Microbiol. 85 (Suppl.), 168S–177S. Uyttendaele, M., Detroy, P., Debevere, J., 1999. Incidence of Salmonella, Campylobacter jejuni, Campylobacter coli, and Listeria monocytogenes in poultry carcasses and different types of poultry products for sale on the Belgian retail market. J. Food Prot. 62, 735–740.
Antimicrobial resistance in Campylobacter from broilers: association with production type and antimicrobial use Laetitia Avrain a , Florence Humbert b , Rolande L’Hospitalier c , Pascal Sanders d , Christine Vernozy-Rozand e , Isabelle Kempf a,∗ a b
Unité de Mycoplasmologie—Bactériologie, Agence française de Sécurité Sanitaire des Aliments, BP 53, F-22440 Ploufragan, France Unité d’Hygiène et Qualité des Produits Avicoles et Porcins, Agence française de Sécurité Sanitaire des Aliments, BP 53, F-22440 Ploufragan, France c Service des Systèmes d’Information, Agence française de Sécurité Sanitaire des Aliments, BP 53, F-22440 Ploufragan, France d Agence française de Sécurité Sanitaire des Aliments, F35302 Fougères, France e Unité de Microbiologie Alimentaire et Prévisionnelle, Ecole Nationale Vétérinaire de Lyon, BP 83, F-69280 Marcy-l’Etoile, France Received 27 February 2003; received in revised form 14 July 2003; accepted 25 July 2003
Abstract The isolation and antimicrobial resistance of Campylobacter jejuni and Campylobacter coli strains from broilers arriving in French slaughterhouses, were analysed according to production types (i.e. standard, export or free-range) and antimicrobial (i.e. coccidiostats, growth promoters or therapeutic agents) administration in flocks. Prevalence was 56.6% in standard, 51.3% in export and 80.0% in free-range broilers. Three hundred and ninety-three strains were identified. Two-thirds of the strains belonged to the species C. jejuni. The others were C. coli. Antimicrobial susceptibility testing was carried out for ampicillin, nalidixic acid, enrofloxacin, tetracycline, erythromycin and gentamicin according to a dilution method. The percentages of resistant strains were, 23, 25, 17, 57, 0.3 and 0% for C. jejuni and 29, 43, 40, 70, 31 and 0% for C. coli. Statistical analysis revealed significant difference in distribution of C. jejuni and C. coli and antimicrobial resistance according to production type or antimicrobial administration. © 2003 Elsevier B.V. All rights reserved. Keywords: Antimicrobial resistance; Campylobacter spp.; Chicken
∗
Corresponding author. Tel.: +33-2-96-01-62-81; fax: +33-2-96-01-62-73. E-mail address: [email protected] (I. Kempf). 0378-1135/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.vetmic.2003.07.001
268
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
1. Introduction Campylobacter is recognised as a commonly encountered microbe responsible of diarrhoeal diseases and foodborne gastro-enteritis (Solomon and Hoover, 1999; Uyttendaele et al., 1999). Campylobacter infections usually occur following ingestion of undercooked animal products, raw milk or contaminated water (Solomon and Hoover, 1999). One of the main reservoirs seems to be the gastro-intestinal flora of chickens (Ermel et al., 1997; Shane, 2000; Solomon and Hoover, 1999). Human Campylobacter jejuni infection may be quite asymptomatic or on the contrary may provoke inflammatory diarrhoeas and different serious sequels such as Guillain-Barré syndrome or Reiter syndrome (Shane, 2000). In patients in whom treatment is necessary, fluoroquinolones or macrolides are currently used (Le Noc et al., 1993; Ruiz et al., 1998; Binotto et al., 2000). Thus data concerning antimicrobial resistance in recent Campylobacter isolates are needed. This study was realised as part of a French surveillance programme of resistance in commensal bacteria (Escherichia coli and Enterococcus) and zoonotic bacteria (Salmonella and Campylobacter) of poultry. France is one of the main poultry producing countries in the European Union. The broiler production is divided into three main types of management. Standard and export chickens are bred inside farm buildings. They are, respectively, slaughtered at approximately 35 and 30 days of age. Free-range chickens have access to an outdoor hen run during at least half of the breeding period and are slaughtered at 80 days of age. In December 1998, the European Council voted in favour of the ban of four antimicrobial growth promoters, because of the fear that continued use of growth promoters could weaken the efficacy of certain antimicrobials in human medicine (European Economic Community, 1998). Flavophospholipol, avilamycin and coccidiostats ionophores such as monensin, salinomycin or synthesised products such as nicarbasin, and diclazuril remained authorised in poultry productions. In this context, the isolation frequency and the antimicrobial susceptibility of C. jejuni and Campylobacter coli isolated from French broilers were determined and compared according to the type of flock management and to consumption of antimicrobial agents used either for growth promotion, prophylactic or therapeutic purposes.
2. Materials and methods 2.1. Bacterial sampling During the 6-month period from January to July 1999, 600 chicken’s caecal samples were obtained during the evisceration process in 10 French slaughterhouses. The selection was representative of the three main types of French production: standard (65%), export (25%) and free-range (10%) broilers. The meat inspection staff who collected the samples also recorded data concerning antimicrobial consumption of the flock reported on the breeding document accompanying birds to slaughterhouse (coccidiostats, growth promoters or antimicrobial treatments). Only one chicken per flock was included.
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
269
2.2. Campylobacter isolation All Campylobacter cultures were incubated under microaerobic atmosphere (Campypak H2 CO2 , AES, Combourg, France). On arrival in the laboratory, 10 g of each cloacal content were weighted and diluted 1/10 in 25% glycerol peptone broth (GPB) (AES). A 10 l of this suspension was then streaked on Karmali (AES) and Butzler no. 2 (Merck, Fontenay-sous-Bois, France) selective media. After 48–72 h of incubation at 42 ◦ C, one characteristic colony was selected and streaked on each medium and incubated for 24–48 h. A total of 816 strains were thus obtained and stored in GPB at −70 ◦ C. Four hundred and eight randomly selected strains were thawed, streaked on Mueller-Hinton agar plate (AES) supplemented with 5% sheep blood (MHSB), and incubated for 48–72 h. For each MHSB plate, one characteristic colony was isolated on MHSB plate for purification, incubated for 24–48 h and this culture was used for identification and antimicrobial susceptibility testing. 2.3. Identification by m-PCR Identification of the genus Campylobacter and of C. jejuni and C. coli, the two main species isolated from poultry was performed using a recently developed multiplex PCR (m-PCR) with slight modifications (Denis et al., 1999; Thomas et al., 1999). About 8–10 colonies were suspended in 200 l of 10 mM Tris–HCl and 1 mM EDTA (TE 10.1 buffer), and then DNA preparation was performed by heating the cells at 95 ◦ C for 10 min. After centrifugation at 2000 × g for 2 min, 10 l of the supernatant were diluted in 990 l of TE 10.1 buffer and stored at −20 ◦ C. Briefly, the m-PCR consisted of a simultaneous amplification in the same tube, of several DNA fragments. Two sets of primers were used to identify the species C. jejuni and C. coli: MDmapA1 and MDmapA2, specific of the mapA gene to identify a 589 bp fragment for C. jejuni and COL3 and MDCOL2, specific of the ceuE gene to identify a 462 bp fragment for C. coli. Amplification reactions were performed in a 30 l mixture containing 3 M of dNTPs solution (Eurobio, Les Ulis, France), 0.11 M of MD16S1, MD16S2, 0.42 M of MDmapA1, MDmapA2, COL3 and MDCOL5 (Oligo express, Paris, France), 0.8 U of Taq polymerase 5 U/l (Roche Diagnostics, Meylan, France), 3 l of buffer 10× with MgCl2 (300 M Tris–HCl; 45 M MgCl2 ; 1500 M KCl) and 3 l of lysate. Amplification reactions were performed in a Perkin Elmer 9600 thermocycler: 10 min at 95 ◦ C, 35 cycles each consisting of 30 s at 95 ◦ C, 1 min 30 s at 59 ◦ C, 1 min at 72 ◦ C, one final cycle for 10 min at 72 ◦ C and a cooling down to 4 ◦ C. Ten microlitres of each amplified product were analysed on a 1.5% agarose gel (Eurogentec, Angers, France) containing 0.3% ethidium bromide and after 2 h at 100 V of migration, the fragments were revealed under UV light (302 nm). Only isolates identified as C. jejuni and C. coli were subjected to antimicrobial susceptibility testing. 2.4. Antimicrobial susceptibility testing Antimicrobial susceptibility testing was done according to the NCCLS document M7-A4. Six antimicrobials were tested: ampicillin (2–32 g/ml) (Sigma, Saint-Quentin-Fallavier,
270
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
France), nalidixic acid (2–128 g/ml) (Sigma), enrofloxacin (0.25–16 g/ml) (Bayer, Puteaux, France), tetracycline (0.25–16 g/ml) (Sigma), erythromycin (0.25–32 g/ml) (Sigma) and gentamicin (0.03–16 g/ml) (Sigma). Reference strains (E. coli (ATCC 25922), Staphylococcus aureus (ATCC 29213), Pseudomonas aeruginosa (ATCC 27853) and Enteroccocus faecalis (ATCC 29212)) and Campylobacter strains were suspended in 4 ml Mueller-Hinton broth (DIFCO, BD Biosciences, Le Pont de Claix, France), incubated at 37 ◦ C for 16–20 h, adjusted to 0.5 McFarland with a densitometer and diluted 1/10 in water. One microlitre of each suspension containing about 104 Colony Forming Units/spot was inoculated on each Mueller-Hinton with 5% sheep blood plate supplemented with antimicrobial agent with a multipoint inoculator and incubated at 37 ◦ C for 48 h (Aarestrup et al., 1997). The Minimum Inhibitory Concentration (MIC) was defined as the lowest concentration that achieved inhibition of visible bacterial growth. Results were recorded after checking that the MIC observed for the control strains corresponded to expected values, previously determined for each strain and each antimicrobial, on Mueller-Hinton agar and under microaerobic conditions. Susceptibility categorisation for fast-growing aerobic bacteria was carried out for Campylobacter according to the statement 1999 of the Antibiogram Committee of the French Society for Microbiology. 2.5. Statistical analysis Data concerning flocks (type of production, consumption of antimicrobial growth promoters or treatments) were recorded in an Open Access 1997 database. Distributions of species or of resistant strains were compared by means of the χ2 test and Fisher’s exact test. A significance level of 5% was used.
3. Results 3.1. Antimicrobial consumption Consumption of the different antimicrobial families in the three production types is summarised in Table 1. Samples and antimicrobial consumption data were obtained during the first semester of 1999, a transitional period for the use of antimicrobial growth promoters in France. According to the Council Regulation (EC) No. 2821/98 of 17 December 1998, bacitracin zinc, spiramycin, virginiamycin and tylosin phosphate remained authorised in France until 30 June 1999. Indeed the breeding documents reported that the use of these compounds as feed growth promoter was already very rare in export production but up to 19% of standard flocks had received bacitracin. In free-range production, according to French legislation, the use of antimicrobials as growth promoters was not reported (Approbation interministérielle par arrˆeté du 22 février, 1996). Avilamycin, one of the two remaining authorised compounds, was the most commonly administrated growth promoter in export production whereas bacitracin was predominant in standard production. Therapeutic administration of macrolides was never mentioned. According to data recorded by meat inspection staff, all the flocks had received one or more coccidiostat. Reported data concerning their use were consistent with
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
271
Table 1 Antimicrobial consumption according to production type Production type Standard, N = 253
Export, N = 83
Free-range, N = 57
Coccidiostats
Ionophore Synthesis
36.4a 18.4
79.5 78.3
52.6 54.4
Growth promoters
Bacitracin MLS as growth promotersb Avilamycin Flavophospholipol
19 4.7 39.1 4.7
0 1.2 79.5 0
0 0 0 0
Antimicrobials used as treatment
-Lactams Quinolones Tetracyclines Colistin Trimethoprime-sulphonamides
2.8 6.7 11.9 5.9 2.8
13.2 32.5 24.1 10.8 14.5
12.3 0 8.8 7 8.8
a b
Percentage of treated flocks. MLS: macrolides, lincosamides or streptogramins given as growth promoters.
previous observations (J.M. Repérant, AFSSA Ploufragan, personal communication). Therapeutic administration of macrolides was never mentioned. Tetracyclines (oxytetracycline and doxycycline) and quinolones (mainly oxolinic acid) were the two most frequently administrated antimicrobial families given for therapy. 3.2. Prevalence and identification of Campylobacter species During the sampling period, 816 bacterial strains were isolated on Karmali or Butzler no. 2 media from 600 caecal samples. One or two strains per sample were randomly selected for further analysis. Seven isolates did not grow after thawing. According to the m-PCR test, six strains belonged to other bacterial genera, 297 strains were identified as C. jejuni (75.6%), 96 as C. coli (24.4%) (Table 2), and 2 showed atypical profiles. The frequency of isolation of Campylobacter strains was 56.6% in standard broilers (221 out of 390 samples), 51.3% in export broilers (77 out of 150 samples) and 80.0% in free-range broilers (48 out of 60 samples). 3.3. Distribution of Campylobacter species according to type of production and antimicrobial administration C. jejuni was the predominant species in standard and export productions (81.4 and 88.0%, respectively), whereas the frequency of this species was significantly lower in free-range production (31.6%) (P ≤ 0.05). Within each production, links between use of antimicrobial agents and distribution of Campylobacter species were examined. In the standard production, the percentage of C. coli was significantly higher (28.6% (26/91)) when birds were
272
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
273
given ionophores compared to 13.0% (21/162) when ionophores were not administrated (P ≤ 0.05). In the same production, when birds were given colistin, the percentage of C. coli strains reached 46.6% (7/15) compared to 16.8% (40/238) in the absence of this compound. In the export production, the percentage of isolation of C. coli was significantly higher (4/11 vs. 6/72) when birds received -lactams (P ≤ 0.05). In the free-range production, no significant effect of administration of antimicrobials on the distribution of Campylobacter species could be observed. 3.4. Antimicrobial susceptibility The distributions of MICs of C. jejuni and C. coli strains are listed in Table 2. The percentages of strains resistant to nalidixic acid, enrofloxacin and erythromycin were significantly higher for C. coli than for C. jejuni (P ≤ 0.05). All strains were sensitive to gentamicin whereas more than 50% of the strains were resistant to tetracycline. The percentages of C. jejuni and C. coli resistant to -lactams or fluoroquinolones ranged from 17 to 40%. Most C. jejuni strains were sensitive or intermediately sensitive to erythromycin. 3.5. Distribution of resistant strains according to production types Statistical tests were performed to research significant differences of distribution of resistant strains of C. jejuni and C. coli according to production type. For one antimicrobial only, i.e. tetracycline, differences of distribution were observed. Eighty-one percent and 90% of C. coli strains isolated from standard and export productions, respectively, were tetracycline resistant compared to 51% in free-range production. 3.6. Distribution of resistant strains according to antimicrobial agents administrated in flocks In standard production, 55/85 (64%) flocks treated with avilamycin were contaminated by tetracycline resistant C. jejuni strains compared to 58/121 (48%) in non-treated flocks (P < 0.05). In the same production, when ionophores were used, the percentage of nalidixic acid resistant C. coli strains (16/26 (61%)) was significantly higher than in absence of ionophores (6/21, 29%) (P ≤ 0.05). In the free-range production, when synthetic anticoccidials were used in association with ionophores, the percentages of ampicillin, nalidixic acid and tetracycline resistant C. jejuni strains reached 7/10 (70%), 8/10 (80%) and 7/10 (70%), respectively, whereas the rates were 0% when flocks were not treated. In the same production the percentage of ampicillin and tetracycline resistant C. coli was also significantly higher when birds were given coccidiostats. In export production, the percentage of enrofloxacin resistant C. jejuni was 37.5% (6/16) when strains were isolated from flocks receiving tetracycline as compared to 12% (7/57) in non-medicated flocks (P ≤ 0.05). Considering the three productions together, the percentage of Campylobacter tetracycline resistant strains was significantly higher when broilers were treated with oxytetracycline or doxycycline (P ≤ 0.05).
274
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
4. Discussion This study is unusual in that it relates antimicrobial resistance of C. jejuni and C. coli in three different types of production and particularly to different patterns and usage of antimicrobial drugs. The frequency of isolation of Campylobacter was significantly higher in free-range than in the two other productions (P ≤ 0.05). Considering the total French production, C. jejuni was more frequent than C. coli in sampled broilers, as was previously observed by Aarestrup et al. (1998) and by Saenz et al. (2000) in other countries. However in free-range production, the ratio C. jejuni/C. coli was inverted. The high frequency of isolation and the predominance of C. coli in the free-range production had already been reported by Rivoal (2000) and could be attributed to very different breeding conditions such as possible outside runs, non-authorisation of antimicrobial growth promoters, limited use of antimicrobials for therapy and age at slaughter. The increase of the relative frequency of C. coli in the case of administration of -lactams could result from a lower susceptibility of this species to these compounds, as suggested by the MIC50 or MIC90 of ampicillin for the two Campylobacter species. Links between increased occurrence of C. coli and administration of ionophore coccidiostats or colistin in flocks could not be explained. The number of strains studied (393) was chosen to enable the detection of one resistant strain with a probability of 0.99 for a resistance level of 5%. This sampling level was too low to detect resistant strains only in the case of gentamicin. Our results were based on analysis of only one or two isolates per flock, whereas it has previously been demonstrated that flocks may harbour several strains of each species (Rivoal, 2000). In this study, when two C. coli strains were isolated from the same chicken, they shared a similar MIC pattern whereas two C. jejuni isolates obtained from the same sample could exhibit different MIC profiles (data not shown). The susceptibility of Campylobacter strains isolated from broilers may not reflect exactly the susceptibility of strains obtained from food or human pathology. Indeed Saenz et al. (2000), in Spain, showed that the percentages of C. coli strains resistant to clindamycin, ampicillin and fosfomycin were higher in human than in broiler isolates whereas for other antimicrobial agents, Campylobacter resistance was usually less frequent in human strains. Such different resistance profiles may result from a possible filter effect of food transformation processes, differences in pathogenicity or resistance selections in human flora. However, clones of mild pathogenicity may contain resistance genes (i.e. tetracycline, aminoglycosamides), transferable to other more virulent Campylobacter strains or to other pathogenic bacterial species (Taylor and Courvalin, 1988). Comparison with other published studies revealed that the percentage of quinolone resistant strains was higher in France than in Denmark but lower than in Spain (Aarestrup et al., 1998; Saenz et al., 2000). The frequency of tetracycline resistance was high in France whereas this compound seemed still rather active against Campylobacter in Spain or in Denmark. Spanish C. coli strains were frequently reported resistant to gentamicin. Differences in MIC determination methods or categorisation breakpoints may partly explain these differences. As drug use is usually regarded as a major source of drug resistance, the most obvious associations (resistance to ampicillin, erythromycin, enrofloxacin and tetracycline) in case of
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
275
administration of -lactams, macrolides, fluoroquinolones and tetracycline were examined but revealed only significant increase of tetracycline resistance when birds were treated with tetracycline (Aarestrup, 1999). The association between use of avilamycin in export production and resistance to tetracycline is noteworthy. Avilamycin is a compound of the everninomycin family, which binds to the ribosome (Bryskier, 1999). Until recently, an everninomycin product was under development for human therapeutic use, but this everninomycin product is no longer being developed. However, the study of the susceptibility of some of our strains to avilamycin was conducted in order to investigate the relations between resistance to these two ribosome-binding agents. No correlation between resistance to tetracycline and resistance to avilamycin could be demonstrated (data not shown). Results showed that the frequency of Campylobacter isolation, the C. jejuni/C. coli ratio or the antimicrobial resistance depended on the type of husbandry. Use of several antimicrobial agents, either as growth promoters or as treatment seemed to influence the resistance level of Campylobacter but confounding factors could not be disregarded. The influence of the European ban of the antimicrobial growth promoters might not be obvious immediately after their withdrawal and it will be important to monitor the trends in resistance over time to assess of antimicrobial resistance related to the use of antimicrobials in animal production.
Acknowledgements The authors are grateful to J.M. Repérant for information concerning coccidiostats and to M. Le Fellic, F. Eono and S. Queguiner for technical help. This work was supported by DGAl grant S98/29 from the French Ministry of Agriculture and Fishery.
References Aarestrup, F.M., Nielsen, E.M., Madsen, M., Enberg, J., 1997. Antimicrobial susceptibility patterns of thermophilic Campylobacter spp. from Humans, pigs, cattle, and broilers in Denmark. Antimicrob. Agents Chemother. 41, 2244–2250. Aarestrup, F.M., Bager, F., Jensen, N.E., Madsen, M., Meyling, A., Wegener, H.C., 1998. Surveillance of antimicrobial resistance in bacteria isolated from food animals to antimicrobial growth promoters and related therapeutic agents in Denmark. APMIS 106, 606–622. Aarestrup, F.M., 1999. Association between the consumption of antimicrobial agents in animal husbandry and the occurrence of resistance bacteria among food animals. Int. J. Antimicrob. Agents 12, 279–285. Antibiogram Committee of the French Society for Microbiology, 1999. Official statement 1999. Pathol. Biol. 47, 845–872. Approbation interministérielle par arrˆeté du 22 février, 1996. J.O. du 25 février. Binotto, E., McIver, C.J., Hawkins, G.S., 2000. Ciprofloxacin-resistant—Campylobacter jejuni infections. Med. J. Aust. 172, 244–245. Bryskier, A., 1999. Everninomycines. In: Bryskier, A. (Ed.), Antibiotiques, agents antibactériens et antifongiques. Ellipses ed., Paris, p. 1060. Denis, M., Soumet, C., Rivoal, K., Ermel, G., Blivet, D., Salvat, G., Colin, P., 1999. Development of a m-PCR assay for simultaneous identification of Campylobacter jejuni and C. coli. Lett. Appl. Microbiol. 29, 406–410. Ermel, G., Laisney, M.J., Rivoal, K., Ragimbeau, C., Gorin, S., Salvat, G., Colin, P., 1997. Les Campylobacter et la filière avicole. Sci. Techn. Avicoles. 20, 3–6.
276
L. Avrain et al. / Veterinary Microbiology 96 (2003) 267–276
European Economic Community, 1998. Council Regulation (EC) No. 2821/98 of 17 December 1998 on authorisation ban of certain antibiotics concerning the additives in animals food. Off. J. Eur. Communities L351/4. Le Noc, P., Croizé, J., Le Noc, D., 1993. Activité Antibactérienne Comparée d’une Nouvelle Fluoroquinolone, la Sparfloxacine (AT 4140, RP 64206) et de 4 autres Fluoroquinolones sur 332 Souches de Bactéries entéropathogènes. Pathol. Biol. 41, 294–301. Rivoal, K., 2000. Les Campylobacter dans la filière avicole. Caractérisation génomique et origine de la contamination dans les élevages. Ph.D. Thesis. Université de Bretagne Occidentale. Ruiz, J., Goni, P., Marco, F., Gallardo, F., Mirelis, B., Jimenez De Anta, T., Vila, J., 1998. Increased resistance to quinolones in Campylobacter jejuni: a genetic analysis of gyrA gene mutations in quinolone-resistant clinical isolates. Microbiol. Immunol. 42, 223–226. Saenz, Y., Zarazaga, M., Lantero, M., Gastanares, M.J., Baquero, F., Torres, C., 2000. Antibiotic resistance in Campylobacter strains isolated from animals, foods, and Humans in Spain in 1997–1998. Antimicrob. Agents Chemother. 44, 267–271. Shane, S.M., 2000. Campylobacter infection of commercial poultry. Rev. Sci. Tech. Off. Int. Epiz. 19, 376–395. Solomon, E.B., Hoover, D.G., 1999. Campylobacter jejuni: a bacterial paradox. J. Food Safety 19, 121–136. Taylor, D.E., Courvalin, P., 1988. Mechanisms of antibiotic resistance in Campylobacter species. Antimicrob. Agents Chemother. 32, 1107–1112. Thomas, C., Gibson, H., Hill, D.J., Mabey, M., 1999. Campylobacter epidemiology: an aquatic perspective. J. Appl. Microbiol. 85 (Suppl.), 168S–177S. Uyttendaele, M., Detroy, P., Debevere, J., 1999. Incidence of Salmonella, Campylobacter jejuni, Campylobacter coli, and Listeria monocytogenes in poultry carcasses and different types of poultry products for sale on the Belgian retail market. J. Food Prot. 62, 735–740.