Antimicrobial Susceptibility of Campylobacter Strains Isolated from Chicken Carcasses in Senegal

Thermophilic Campylobacter species, especially C. jejuni and C. coli, are among the main agents of gastrointestinal infections in developed countries (2, 12, 29). Campylobacter also poses an emerging public health problem in developing countries (13, 19, 20, 32). Most cases of Campylobacter gastrointestinal infections do not require antimicrobial treatment, being brief, clinically mild, and self-limiting. However, complications such as severe cases of enteritis or septicemia may occur and require treatment. Macrolides and fluoroquinolones (FQs) are the most useful antimicrobial drugs for Campylobacter infections, but resistance has been reported to be increasing (12, 19, 26, 29).


Summary
Campylobacter resistance to antimicrobial agents appears as an emerging public health problem in industrialized countries, but, on the other hand, only few data on the subject are available in developing countries. To assess antibiotic susceptibility of Campylobacter strains in Senegal, skin samples were collected from 250 chicken carcasses from January 2001 to October 2002. Among 204 Campylobacter strains isolated, two species were identified: C. jejuni (59%) and C. coli (41%). In vitro susceptibility to five antimicrobial drugs (amoxicillin, amoxicillin-clavulanic acid, erythromycin, nalidixic acid and ciprofloxacin) was determined by the E-test method. Minimum inhibitory concentrations (MICs) showed 34% of Campylobacter isolates were ciprofloxacin resistant with a high level of resistance (MIC ≥ 32 mg/l) in 25% of both species. Cross-resistance between nalidixic acid and ciprofloxacin was found in 96% of quinolone-resistant strains. The level of amoxicillin resistance was statistically higher for C. coli than for C. jejuni (20.2 versus 10.8%), but all the strains were susceptible to amoxicillin combined with clavulanic acid. Both species showed low resistance to erythromycin. Multiresistant phenotype to three of the drugs tested was found in 9.8% of the strains: 15.5% of C. coli strains and 5.8% of C. jejuni strains. No strain was resistant to four or more of the drugs studied. Further studies appear necessary to evaluate antibiotic resistance of Campylobacter isolated in human and animal samples in order to control the emergence of new multidrug-resistant strains in Senegal.
human population with animal protein. Poultry farmers use antibiotics to fight diseases. Because growing scientific evidence shows that the use of antimicrobial agents in veterinary medicine results in development of resistant pathogenic bacteria (1), the authors' objective was to assess the frequency of antimicrobial resistance of Campylobacter strains isolated from "modern" poultry meat.

Specimen collection and isolation
From January 2001 to October 2002, 250 chicken carcasses were collected from slaughterhouses and retail shops in Dakar. The carcasses were immediately brought back to the laboratory in cool boxes (+4°C). Samples of neck skin (10 g) were taken from each carcass and added to 90 ml of Preston broth with Preston antibiotic supplement 1

Identification of isolates
Isolates were identified using a commercial identification method (API Campy ® , bioMérieux, France). Identification of every isolate was confirmed by a multiplex PCR, using specific primers of Campylobacter genus (MD16S1, MD16S2), C. jejuni species (MDMapA1, MDMapA2) and C. coli species (COL3, MDCOL2) (10). Briefly, Campylobacter colonies from blood agar plate were suspended in 0.2 ml TE buffer. Cells were lysed by heating at 95°C for 10 min and cellular debris were removed by centrifugation at 5000 g for 10 min. The supernatant was used as source of template for DNA amplification. Each multiplex PCR tube contained 200 µM deoxynucleoside triphosphate, 2.5 µl of 10X reaction buffer, 20 mM MgCl2, 0.11 µM Campylobacter genus primers, 0.42 µM C. jejuni primers and 0.42 µM C. coli primers. Template DNA (3 µl) was added and the volume adjusted with sterile water to obtain 30 µl. DNA amplification was carried out in a Perkin-Elmer 9600 ® thermocycler using an initial denaturation step at 95°C for 10 min, followed by 35 cycles. Cycling conditions were as follows: denaturation, 95°C for 30 s; annealing, 59°C for 90 s; extension, 72°C for 1 min. After the last cycle, a final extension step at 72°C for 10 min was added. Ten microliters of PCR product were analyzed by gel electrophoresis (1.5% gel agarose). Gels were stained with ethidium bromide at 0.5 µl/ml and visualized by UV transillumination. A 100-bp DNA ladder (Amersham Biosciences, France) was used as size marker. Negative controls were added in each run. Positive PCR controls consisted of C. jejuni subsp. jejuni ATCC 49943 and C. coli ATCC 49941.

Susceptibility testing
Several colonies of each strain were suspended in 5 ml of Mueller-Hinton broth to achieve turbidity equal to 0.5 MacFarland standard. The suspensions were inoculated with sterile swabs onto Mueller-Hinton agar with 5% sheep blood. After application of E-test ® strips (AB Biodisk, Sweden) plates were incubated at 37°C for 48 h under microaerophilic atmosphere. Minimum inhibitory concentrations (MICs) were read according to the recommendations of the manufacturer by two different readers. Breakpoints were those recommended by the Antimicrobial Committee of the French Society for Microbiology (33). Breakpoints for resistance susceptibility were higher than 16 mg/l for amoxicillin, than 16/2 mg/l for amoxicillinclavulanic acid, than 4 mg/l for erythromycin, than 16 mg/l for nalidixic acid, and than 2 mg/l for ciprofloxacin.

Statistical analysis
Data were entered and analyzed with SPSS, version 10 (SPSS, Chicago, USA). The χ 2 test and Fisher's exact two-tailed test were used for statistical analysis of the significant difference of resistance rates. An α of 0.05 was used for statistical significance.

■ RESULTS
Campylobacter was isolated from 81.6% of the samples. C. jejuni was the most prevalent species (59% C. jejuni versus 41% C. coli).

Susceptibility testing
The results of antimicrobial susceptibility testing for C. coli and C. jejuni are shown in Table I

Table I
Antimicrobial susceptibility of Campylobacter coli and C. jejuni strains isolated from chicken carcasses in Senegal C. coli, MICs of all resistant strains were higher than 256 mg/l. Among C. jejuni, MICs of 46% of amoxicillin-resistant strains were higher than 256 mg/l. In addition, amoxicillin MIC 90 was statistically higher for C. coli than for C. jejuni. No strains resistant to amoxicillin-clavulanic acid were observed. For amoxicillin high level resistant strains (MIC > 256 mg/l), amoxicillin-clavulanic acid MIC was between 2/1 and 8/4 mg/l in both species. The overall resistance rate to erythromycin was 4.7% for C. coli and 3.3% for C. jejuni without any significant difference between the two species.
Resistance rates to nalidixic acid in C. coli and C. jejuni were 34.5 and 31.6%, respectively. Among these resistant strains, 55.2% of C. coli and 68.5% of C. jejuni exhibited MICs higher than 256 mg/l. For both species, MIC 90 was higher than 256 mg/l. Compared to nalidixic acid, quite similar resistance rates to ciprofloxacin were observed with 34.5 and 30.8% for C. coli and C. jejuni, respectively. Ciprofloxacin MIC was higher than 32 mg/l for 25% isolates in both species. Among Campylobacter-resistant strains, 75.8% of C. coli and 78.3% of C. jejuni had ciprofloxacin MICs higher than 32 mg/l. Furthermore, 90% of the isolates were inhibited with MICs higher than 32 mg/l for C. coli and C. jejuni. For both species, MIC 90 was higher than 32 mg/l.

Coresistant and multiresistant isolates
Drug resistance to one or more drugs was detected in over 38% of the strains (Table II). Among them, C. coli and C. jejuni, 41.7 and 36.7% of the strains, respectively, exhibited resistance to the tested antibiotics. Among C. coli and C. jejuni, 33.4 and 30%, respectively, were resistant to both quinolones (nalidixic acid and ciprofloxacin). A cross-resistance between nalidixic acid and ciprofloxacin was found for all isolates (96%) except for one strain of C. jejuni and one strain of C. coli, which showed a nalidixic acid-resistant ciprofloxacin-susceptible phenotype. Table III shows a comparison between nalidixic acid MICs and ciprofloxacin MICs for cross-resistant isolates: 46.5% of C. coli strains and 63.8% of C. jejuni strains exhibited a high level of resistance to both quinolones with MICs higher than 256 mg/l for nalidixic acid and 32 mg/l for ciprofloxacin. Multiresistance, defined as resistance to three or more of the drugs tested, was found in 9.8% of Campylobacter strains: 15.5% were C. coli strains and 5.8% were C. jejuni strains. No strain was resistant to four or five drugs. The only multidrug-resistant phenotype was amoxicillin, nalidixic acid and ciprofloxacin in both species.

■ DISCUSSION
In Senegal, beside traditional poultry production, modern poultry production is expanding around Dakar. Senegalese poultry farmers try to minimize the impact of diseases on their production with huge amounts of antibiotics, like in Europe or the United the States. Considering the growing prevalence of resistant strains in industrialized countries, antimicrobial drugs must be used very carefully in Senegal.
In the present survey, 81.6% of chicken carcasses were contaminated with Campylobacter. In developed countries, several studies have also reported high levels of Campylobacter isolation from chicken carcasses and retailed chickens: 46% in Germany (4), 46% in Japan (23) and from 73% to 100% in the USA (40). Although little information is available from developing countries, the present results are consistent with those from Kenya and China where thermophilic Campylobacter has been isolated from 77 and 76% of chicken samples, respectively (24,31). As reported by several studies, Campylobacter isolated from animals may contaminate humans either by direct ingestion of undercooked food or by cross-contamination of raw poultry to other foods caused by non hygienic handling such as unwashed hands or dirty utensils (18,22). As already reported (28), C. jejuni was more frequently isolated than C. coli (59 versus 41%). This phenomenon is easily explained by the fact that C. coli is mainly associated with pigs (3).
Because no great discrepancies have been observed between double dilution agar and E test according to Baker et al. (5) and Funke et al. (14), the authors used the E test for susceptibility testing. This method is reliable, technically simple, and needs no special equipment (13). In some countries FQ resistance rates were similar between strains isolated from poultry meat and from man (17,19,29), highlighting the need to carefully monitor resistance in human and animal samples.
In the present survey, high resistance rates to quinolones were observed in both species isolated from chicken samples. This high resistance rate was similar to those observed in several European countries (26) or in Japan (9). As reported by Saenz et al. in Spain (29), a cross-resistance was observed between nalidixic acid and ciprofloxacin. But unlike Thwaites and Frost in the United Kingdom (37), Campylobacter strains with a high level of resistance to both quinolones were predominant. Nevertheless, rates of resistance to ciprofloxacin observed in the present study appeared lower than those described in Belgium (39)  or Lebanon (35), where they ranged from 61 to 100%. By contrast, no Campylobacter strains were found ciprofloxacin resistant in Chile (13).
Since 1991, when Endzt et al. (11) identified the first quinoloneresistant Campylobacter strains in C. jejuni and C. coli in the Netherlands, Campylobacter resistance to FQ has been on the increase throughout the world (12,29). The percentage of ciprofloxacin-resistant strains in man increased in Spain from 0-3% in 1989 to 30-50% in 1991 after licensing the use of enrofloxacin in 1990 in animal husbandry (30). During this period, the same evolution was observed in The Netherlands (11), Finland (27) and Canada (15). Likewise in England, before the introduction of FQ in veterinary medicine, a study showed that domestically-bred poultry were less contaminated by ciprofloxacin-resistant strains than those imported from countries where the use of FQ was legal (16). Thus, this major development of FQ-resistant strains in humans and animals seems related to the introduction of FQ in veterinary medicine (11,29).
The speed at which the level of resistance to nonfluorinated (nalidixic acid) and fluorinated (ciprofloxacin) quinolones (34 versus 31%) has been reached may be related to the mechanisms by which quinolone resistance involves a single chromosomal mutation on the targets of quinolone action (DNA gyrase, DNA topoisomerase IV) (12,25). In Senegal, FQs (norfloxacin, enrofloxacin) were introduced in veterinary medicine in 1996 in poultry production to treat respiratory and intestinal diseases as salmonellosis or colibacillosis. Since 2000, they became first line molecules because of treatment failures with other antibiotic drugs. According to observations from other reports and despite the absence of previous studies on antibiotic resistance in Senegal, the high prevalence of FQ resistance could thus be related to the introduction of FQs in the country especially for poultry production (21). As in developed countries, FQ resistance may soon become a public health problem.
High levels of amoxicillin resistance were found, particularly in C. coli strains. Campylobacter resistance to penicillin A has been reported in previous studies with resistance rates never exceeding 39% (3,9,13,29). However, in Senegal, amoxicillin was rarely used for treatment of pasteurellosis or salmonellosis in poultry production. Resistance is usually associated with beta-lactamase production but other mechanisms of resistance could be involved, such as modified penicillin-binding proteins or impermeability (29, 34). As reported in other studies, most of the isolates of the present survey showed high susceptibility to amoxicillinclavulanic acid (35). Clavulanic acid inhibits some beta-lactamases and has been claimed to have intrinsic antibacterial activity against Campylobacter (36).
Among the stains isolated, 4.7% of C. coli and 3.3% of C. jejuni strains were resistant to erythromycin. This appeared much lower than results from Nigeria, where resistance rate reached 40.4% among animal strains (32). In the present study, erythromycinresistance rates were similar in both species, contrary to results from surveys conducted in Spain (17 and 83% of resistance in C. jejuni and C. coli, respectively; 29) or in Vietnam (17 and 50% of resistance in C. jejuni and C. coli, respectively; 19). This difference may be explained by the introduction in these countries of tylosin, especially used as growth promoter in the pig industry (39). In Senegal, macrolides were not used as growth promoters in poultry production; they were only used for treatment of some respiratory diseases such as chronic respiratory diseases. Resistance to erythromycin is chromosomally mediated and is due to the alteration of the ribosome (12), but natural resistance as efflux pump has been observed too (8).
To the five most relevant antimicrobial drugs selected for campylobacteriosis treatment, seven resistance patterns were observed, but only 9.8% of total strains were resistant to three drugs. None were resistant to four or more drugs. The present results were much lower than those reported from Belgium (39) or Spain (29). However, consistently with these studies, C. coli exhibited more multidrug-resistant strains than C. jejuni in the present survey.
This study is the first to highlight the decrease of antibiotic susceptibility of Campylobacter isolated from chicken in Senegal. Many reasons explain this situation; among others, there is the lack of information among poultry farmers and veterinary surgeons on the use of antibiotics and on drug resistance, the inordinate use of antibiotics due to the high prevalence of infectious diseases among flocks, and the difficulty to perform large antimicrobiological studies.
However, as everywhere else in the world, the increase of Campylobacter-resistant strains seems to be related to the amounts of antibiotics used in animals (38). Thus, to prevent transfer of resistant bacteria or resistance genes from animals to humans via the food chain (6) in Senegal, measures to be implemented are the same as the ones in developed countries: reduction of the use of antibiotics (antibiotics should not be available over the counter in order to avoid self medication), restricting their use to encourage narrow-spectrum specific antibiotic therapy instead of broad spectrum antimicrobials (7), and replacement of antibiotics with improvements in hygiene and flock management.