Antimicrobial susceptibility testing of Campylobacter jejuni: a comparison between Etest and agar dilution method

1. Introduction 

Campylobacter jejuni is one of the most frequent causes of bacterial diarrhea in developing countries (Nachamkin, 1995). Because of the self limiting nature of Campylobacter infections, antibiotic treatment is often not necessary and antibiotic susceptibility testing of C. jejuni is usually not performed in clinical laboratories (Blaser, 2000). Agar dilution is the recommended method for susceptibility testing (Tenover et al., 1992). Etest (AB BIODISK, Solna, Sweden) has been demonstrated to be an easy alternative method for Campylobacter spp, as it is for other bacteria Baker 1992, Tenover et al 1992.

In this study, we determined the susceptibility of C. jejuni strains against nine antimicrobial agents in a comparative study with Etest and standard agar dilution method to further investigate the correlation between these methods. The 50 strains used have been isolated and identified from stool samples of patients with diarrhea between April and October 1998 using standard microbiologic methods (Nachamkin, 1995). The strains were stored in tryptic soy broth with 20% glycerol at −70°C until analyzed. The control strains used were Escherichia coli ATCC 25922 and Staphylococcus aureus ATCC 29213. Ampicillin, ampicillin-sulbactam (2/1 ratio), clarithromycin, tetracycline, gentamicin, azithromycin, chloramphenicol, ciprofloxacin and erythromycin were the agents used for susceptibility testing. The antimicrobial powders were obtained from the manufacturers. Etest strips were purchased from AB BIODISK. The inoculum was prepared from 48 h growth of C. jejuni on 5–7% horse blood agar. The colonies were suspended in Mueller Hinton broth (Oxoid, Basingtoke, UK) to achieve a turbidity equivalent to a standard MacFarland of 0.5. The determination of MICs by agar dilution was performed according to NCCLS guidelines using Mueller Hinton agar (Oxoid, Basingtoke, UK) supplemented with 5–7% horse blood and the breakpoints given for organisms other than Hemophilus, Neisseria gonorrhoeae and Streptococcus pneumoniae were used (NCCLS, 2000 and 2002). Etest was set up on the same agar according to the manufacturer’s instructions Brown and Brown 1991, Huang et al 1992, Funke et al 1993. Plates were incubated at 37°C in candle jar for 24 h. The incubation time was extended to 48 h for agar dilution test if necessary.

C. jejuni strains were found to be highly susceptible to the antibiotics tested. The highest resistance rate was seen with tetracycline and ciprofloxacin; 12% by Etest and 16% by agar dilution for tetracycline and 14% for ciprofloxacin by both methods. For most of the strains (13/50), lower MIC values were obtained with Etest (Table 1). The ranges, MIC50, MIC90 and the resistance rates are outlined in Table 1. The agreement between the two methods analyzed as log2 dilution differences is presented in Table 2. The discrepancies between the methods evaluated as minor, major and very major errors. The highest discrepancy between Etest and agar dilution (>2 dilution) were seen with chloramphenicol (48%), ampicillin-sulbactam (44%), clarithromycin and tetracycline (40%). For erythromycin which is considered to be the drug of choice for Campylobacter infections, Etest results were outside the accepted ranges (±2 dilution) for 34% of the strains tested. The essential agreements between two methods were 66.6% (±1 dilution) and 85.5% (±2 dilution). The agreement of susceptibility categories were higher at 94.4%. There were no major and very major discrepancies. The minor discrepancy rates were 6% or less for all the agents tested except for erythromycin with 30% minor discrepancy. For the control strain, there was correlation between the two methods. The difference between the susceptibility percentages determined by each test were not statistically significant (P > 0.05) except for erythromycin which was 80% by Etest and 58% by agar dilution test: 11 strains susceptible by Etest were intermediate by agar dilution. The category interpretation of erythromycin results is considered as a problem when using current NCCLS (2002) breakpoints.

Table 1. The MIC ranges, MIC50, MIC90 values and resistance rates

	Antimicrobial agentsa	Methoda	MIC range (mg/L)	MIC50 (mg/L)	MIC90 (mg/L)	Resistance %
	AM	E	0.064–32	2	8	2
		AD	0.032–32	2	8	2
	AB	E	0.064–8	1	4	0
		AD	0.064–8	2	8	0
	EM	E	0.032–4	0.25	1	0
		AD	0.064–4	0.5	2	0
	CH	E	0.016–4	0.25	1	0
		AD	0.064–16	0.5	2	2
	AZ	E	0.016–16	0.064	0.25	4
		AD	0.032–32	0.125	0.25	4
	GM	E	0.016–16	0.5	2	2
		AD	0.064–4	0.5	2	2
	TC	E	0.064–32	0.25	8	12
		AD	0.032–256	0.25	16	16
	CL	E	0.25–8	1	4	0
		AD	1–8	2	8	0
	CI	E	0.016–32	0.064	8	14
		AD	0.016–32	0.125	8	14

a E, Etest; AD, Agar dilution; AB, Ampicillin-sulbactam; CH, Clarithromycin; AM, Ampicillin; EM, Erythromycin; GM, Gentamicin; AZ, Azithromycin; TC, Tetracycline; CL, Chloramphenicol; CI, Ciprofloxacin.

Table 2. The agreement between two methods analysed as log2 dilution differences

	Antimicrobial agentsa	<−2	−2	−1	0	+1	+2	>+2	Out of accepted range (%)
	AM	5	2	9	15	14	1	4	24
	AB	5	11	15	8	5	-	6	44
	EM	7	6	15	14	4	3	1	34
	CH	6	2	15	11	4	7	10	40
	AZ	-	3	26	13	6	1	1	10
	GM	4	5	17	11	8	3	2	28
	TC	-	9	10	11	9	4	7	40
	CL	4	19	15	7	4	1	-	48
	CI	4	5	16	15	3	2	10	32
	All Agents	35	63	136	105	57	22	41	33

a E, Etest; AD, Agar dilution; AB, Ampicillin-sulbactam; CH, Clarithromycin; AM, Ampicillin; EM, Erythromycin; GM, Gentamicin; AZ, Azithromycin; TC, Tetracycline; CL, Chloramphenicol; CI, Ciprofloxacin.

In this group of relatively small number of Campylobacter spp. strains collected from a single geographic area, there was acceptable agreement between Etest and the agar dilution for susceptibility categorization. Since Etest is neither dependent on special equipment nor time-consuming, it may be a useful alternative for susceptibility testing of Campylobacter spp. in clinical laboratories and this finding is in parallel with a study in Taiwan (Li et al., 1998).