Comparison of Broth Microdilution Method Using Haemophilus Test Medium and Agar Dilution Method for Susceptibility Testing of Eikenella corrodens

Abstract

Susceptibility testing of Eikenella corrodens is usually performed by a Mueller-Hinton sheep blood agar dilution (AD) method. However, this method is impractical for testing only a few strains. We compared AD with the broth microdilution method using Haemophilus test medium (HTM) in order to determine the susceptibility of 36 clinical isolates of E. corrodens to eight antimicrobial agents. MICs obtained by the HTM method yielded 95.5 and 84% agreement (within 2 and 1 log₂ dilutions, respectively) with those obtained by AD. The HTM method with incubation in CO₂ for 48 h was highly reproducible and constitutes an easy alternative for antimicrobial susceptibility testing of E. corrodens.

Eikenella corrodens is a fastidious, slow-growing, gram-negative rod that has been increasingly recognized as a pathogen in a wide variety of infections (2, 12). Increased isolation of β-lactamase-producing strains and frequent resistance to many antimicrobial agents make necessary the routine determination of the susceptibility of this pathogenic bacterium to different antimicrobial agents (2, 3, 5, 6, 9, 10, 13, 14). Susceptibility testing has usually been performed by a sheep-blood-based agar dilution (AD) method which is tedious, time-consuming, and impractical for laboratories testing only a few strains (3). In this study we compare the AD method with a broth microdilution method using Haemophilus test medium (HTM) to determine the susceptibility of E. corrodens to eight frequently used antimicrobial agents.

(This study was presented in part at the 37th Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, Ontario, Canada, 28 September to 1 October 1997.)

Thirty-six nonduplicate, recent clinical isolates of E. corrodens were tested. All strains were identified by standard microbiological methods (2, 7). The sources of the isolates were as follows: abscesses (n = 13); wounds (n = 9); blood (n = 4); peritoneal fluid (n = 4); tissue biopsy (n = 3); and pleural fluid, umbilical tissue, and whitlow (n = 1 each). β-Lactamase-producing strains were detected by the chromogenic cephalosporin test (nitrocefin) (Cefinase; BBL Microbiology Systems, Cockeysville, Md.). The isolates were maintained in 10% skim milk at −70°C until ready for use and then subcultured three times on 5% sheep blood Columbia agar plates for 48 h at 35°C in CO₂ before susceptibility studies were performed. These cultures were used for inoculum preparation. The following antimicrobial agents were used as powders of known potency: ampicillin, cephalothin, cefotaxime, and tetracycline, obtained from Sigma Chemical Co., St. Louis, Mo.; amoxicillin and clavulanic acid, obtained from SmithKline Beecham, Worthing, United Kingdom; cefoxitin and imipenem, obtained from Merck, Sharp and Dohme, Madrid, Spain; and ciprofloxacin, obtained from Bayer AG, Barcelona, Spain. All antimicrobial solutions were freshly prepared before use. AD MICs were determined by the standard procedure (3, 11) with Mueller-Hinton agar plates supplemented with 5% sheep blood incorporating twofold increments of the various antimicrobial agents, with concentrations varying from 128 to 0.001 μg/ml depending on the antimicrobial agent being tested. Inocula consisting of 10⁴ CFU/ml/spot were applied to the surfaces of the plates with a Steers replicator. Control plates without antimicrobial agents were also inoculated before and after each series of antibiotic-containing plates. All plates were incubated at 35°C in 5% CO₂ for 48 h and then examined. Staphylococcus aureus ATCC 29213 and Enterococcus faecalis 29212 were used as control strains.

MICs were also determined by the broth microdilution method according to National Committee for Clinical Laboratory Standards guidelines (11) with HTM broth made in-house. Plastic microdilution trays contained the antimicrobial agents in serial twofold dilutions from 128 to 0.001 μg/ml, depending on the antimicrobial agent being tested. Inocula were prepared in HTM broth from cultures grown on Columbia agar with 5% sheep blood. The final concentration was 5 × 10⁵ CFU/ml. All microdilution trays were prepared in duplicate and incubated at 35°C; one series was incubated in air, and the other was incubated in 5% CO₂. Readings for both series were performed at 24 and 48 h. In order to assess reproducibility, 16 strains were tested in triplicate. Reproducibility was calculated as the percentage of MICs within a range of 3 log₂ dilutions. Haemophilus influenzae ATCC 49247 was used as a control strain.

The antimicrobial susceptibilities of E. corrodens as determined by AD and by broth microdilution under two different incubation conditions (air and CO₂) and with two different reading times (24 and 48 h) are shown in Table 1. HTM broth provided good support for the growth of all E. corrodens isolates, especially when the microtrays were incubated in CO₂ for 48 h (C48-HTM). Only two strains failed to grow in air after 24 h. Results obtained by both methods show that all strains were uniformly susceptible to amoxicillin-clavulanate, cephalothin, cefoxitin, cefotaxime, imipenem, tetracycline, and ciprofloxacin according to National Committee for Clinical Laboratory Standards criteria (11); however, cephalothin was the least-active compound, with more than 50% of strains requiring 8 μg/ml for inhibition. There were two β-lactamase-producing strains (5%); the MICs of ampicillin for these isolates were 16 and 64 μg/ml. AD and HTM broth microdilution MICs were generally similar for all agents tested, especially with C48-HTM. Overall, MICs obtained by C48-HTM yielded 95.5% agreement (within 2 log₂ dilutions) with those obtained by AD. Agreement was 84% within 1 log₂ dilution (Table 2). Analysis of percentage of agreement for individual antimicrobial agents indicated no significant tendencies for amoxicillin-clavulanate and cephalothin; however, ampicillin, cefoxitin, and cefotaxime MICs tended to be 1 log₂ dilution lower with C48-HTM than with AD; conversely, tetracycline and ciprofloxacin MICs tended to be 1 log₂ dilution higher. Imipenem MICs varied widely, ranging from −4 to +3 dilutions, although 77.8% of the strains were within 2 log₂ dilutions. Reproducibility was 93%, regardless of incubation conditions and the antimicrobial agent used.

TABLE 1.

Comparative susceptibilities of 36 strains of E. corrodens to antimicrobial agents by the AD and HTM broth microdilution methods

Antimicrobial agent	AD		HTM broth microdilution
	Ranged	MIC90a	24 h, O2b		48 h, O2		24 h, CO2		C48-HTM
			MIC90	≤2 dilc	MIC90	≤2 dil	MIC90	≤2 dil	MIC90	≤2 dil
Ampicillin	0.25–64	2	1	53	1	75	1	94	1	97
Amoxicillin-clavulanate	0.12–1	0.5	0.25	92	0.25	97	0.5	100	0.5	100
Cephalothin	4–16	16	8	94	8	100	8	100	16	100
Cefoxitin	0.25–4	2	0.25	59	0.5	64	0.5	69	1	83
Cefotaxime	0.006–0.06	0.03	0.006	65	0.01	100	0.01	92	0.01	100
Imipenem	0.12–0.5	0.5	0.25	47	0.5	69	0.5	67	0.5	78
Tetracycline	0.25–2	1	1	91	2	86	0.5	92	2	97
Ciprofloxacin	0.006–0.01	0.01	0.006	94	0.01	97	0.01	97	0.01	97
Overall				74		86		89		94

^aMIC₉₀, the MIC at which 90% of the isolates are inhibited, expressed in micrograms per milliliter. 

^bTwo strains failed to grow in air after 24 h. 

^c≤2 dil, percentages of correlation of MICs obtained by the AD and HTM broth microdilution methods within 2 dilutions of each other. 

^dExpressed in micrograms per milliliter. 

TABLE 2.

Distribution in differences of MICs of antimicrobial agents determined by AD and HTM broth microdilution with C48-HTM

Antimicrobial agent	% of isolates for which MICs differ by indicated dilutions(s)							% within 1 log2 dilution
	≤−3	−2	−1	0	+1	+2	≥+3
Ampicillin	0	8.3	47.2	36.1	8.3	0	0	91.6
Amoxicillin-clavulanate	0	2.8	11.1	55.6	27.8	2.8	0	94.5
Cephalothin	0	2.8	25	52.8	19.4	0	0	97.2
Cefoxitin	13.9	8.3	30.6	36.1	8.3	2.8	0	75
Cefotaxime	0	13.9	47.2	36.1	2.8	0	0	86.1
Imipenem	19.4	11.1	25	11.1	8.3	19.4	2.8	44.4
Tetracycline	0	0	19.4	47.2	19.4	13.9	0	86.1
Ciprofloxacin	0	0	16.7	44.4	36.1	2.8	0	97.2

The AD technique is currently the method of choice for susceptibility testing of E. corrodens; however, it is tedious and time-consuming and impractical for laboratories testing only a few strains. The broth microdilution method has become increasingly popular, and many laboratories test susceptibility of most microorganisms by this technique (1). In 1983, Goldstein et al. (4) described a broth microdilution technique with a filtrate of laked sheep blood which compared favorably with the AD method. However, the procedure was tedious, since it required lysed sheep blood. HTM was developed in 1987 in order to satisfy the complex growth requirements of Haemophilus spp. (8). This medium also allows the growth of many fastidious microorganisms; however, it has never been used for antimicrobial susceptibility testing of E. corrodens. In our study, the HTM broth microdilution method provided good support for growth of E. corrodens strains and was easily interpretable and highly reproducible. The MICs showed a high correlation with those obtained by the AD technique, especially with C48-HTM. However, the HTM broth microdilution method has not proved to be reliable for testing susceptibility to imipenem and cefoxitin. Although the HTM broth used in the present study was prepared in-house, this medium can be purchased commercially. On the basis of our results and the commercial availability of the HTM broth, we recommend the HTM broth microdilution method for antimicrobial susceptibility testing of E. corrodens.