Antimicrobial Susceptibility Testing of Bilophila wadsworthia Isolates Submitted for Routine Laboratory Examination

Abstract

MICs of antibiotics against Bilophila wadsworthia isolates were measured by agar and broth microdilution with pyruvic acid and by Etest. The inoculum size influenced greatly agar dilution. Despite discrepancies in MICs depending on the measurement method used, clindamycin consistently showed potent activity. Broth microdilution and Etest appear to be candidates for laboratory susceptibility testing.

Bilophila wadsworthia was recently recognized as an anaerobic gram-negative rod in a study of abdominal infections in patients with gangrenous or perforated appendicitis (1). B. wadsworthia is the third most common anaerobe isolated from such patients (2, 5) and is involved in polymicrobial infections (1, 3, 9). B. wadsworthia requires strictly anaerobic conditions and supplementary factors for optimal growth in vitro and exhibits characteristic slow colony formation (4). For these reasons, results of conventional susceptibility tests, which rely on adequate growth in a standardized medium, are difficult to interpret when anaerobic species are assayed (11). Results obtained with standardized methods may also overestimate resistance; initial susceptibility studies of B. wadsworthia indicated significant resistance to several antibiotics, including imipenem, cefoxitin, penicillin G, and other β-lactam antibiotics (14). Recently, however, a new antibiogram for B. wadsworthia has been established by the use of aerobic triphenyltetrazolium chloride (TTC) (13) to facilitate endpoint determinations with the agar dilution method, suggesting lower MICs than those found with conventional methods (11). Although no β-lactamase production was demonstrated in the previous study (1), it was recently reported that 87% of clinical isolates of B. wadsworthia were β-lactamase positive (11). However, the methods of antimicrobial susceptibility testing of B. wadsworthia in routine clinical laboratory examinations have not been fully investigated.

Therefore, in the present study, we compared antimicrobial susceptibilities of clinical isolates of B. wadsworthia to several antibiotics by the agar dilution method with TTC, the broth microdilution method supplemented with pyruvic acid, and Etest (7).

Clinical isolates of B. wadsworthia used in this study are listed in Table 1 with clinical information. Strains NBW3 and NBW6 were isolated at Nagasaki University Hospital, Nagasaki, Japan, whereas the other strains were isolated at its affiliated medical facility, St. Francis Hospital, Nagasaki, Japan. All strains were isolated from clinical specimens by using Bacteroides bile esculin (BBE) agar (Becton Dickinson Microbiology Systems, Cockeysville, Md.) showing transparent colonies with a black center. The isolates were positive for catalase and were identified in accordance with the Wadsworth Anaerobic Bacteriology Manual (12). All isolates were also identified as B. wadsworthia by the RapID ANAII system (Innovation Diagnostic Systems, L.P., Norcross, Ga.). For quality controls, B. wadsworthia GAI95508 (= WAL7959) and Bacteroides fragilis GAI5562 (= GM7000), kindly provided by the Institute of Anaerobic Bacteriology, Gifu University School of Medicine, Gifu, Japan, were used.

TABLE 1.

B. wadsworthia isolates used in this study

Strain	Date of isolation (yr/mo/day)	Patient data				Other isolate(s) from patient:
		Age (yr)/sexa	Specimen	Type of infection	Underlying condition	Anaerobic	Aerobic
NBW1	1995/6/9	60/F	Pus	Abscess	Cholelithiasis	B. fragilis	None
NBW2	1995/8/11	69/F	Resected tissue	Peritonitis	Appendicitis	B. fragilis	Escherichia coli, Streptococcus spp.
NBW3	1995/10/31	53/F	Pus	Liver abscess	Cholangiocarcinoma	None	Klebsiella pneumoniae, Staphylococcus epidermidis
NBW4	1996/7/22	82/M	Pus	Abscess	Ileus following diverticulitis	None	E. coli, Pseudomonas aeruginosa, Enterococcus faecalis
NBW5	1996/7/31	79/F	Pus	Abscess	Colon cancer	B. fragilis	K. pneumoniae
NBW6	1996/11/18	4/M	Peritoneal fluid	Peritonitis	Appendicitis	B. fragilis, Peptostreptococcus prevotii	E. coli

^aF, female; M, male. 

Antimicrobial susceptibility tests for B. wadsworthia isolates were performed by the agar dilution method, the broth microdilution method, and Etest. Antibiotics used for both dilution methods were ampicillin (Pfizer Pharmaceuticals Inc., Tokyo, Japan), sulbactam-ampicillin (Pfizer), ceftizoxime (Fujisawa Pharmaceutical Co., Ltd., Osaka, Japan), imipenem (Banyu Pharmaceutical Co., Ltd., Tokyo, Japan), and clindamycin (Sigma Chemical Co., St. Louis, Mo.).

The agar dilution method was performed with brucella HK agar (Kyokuto Pharmaceutical Industrial Co., Ltd., Tokyo, Japan) supplemented with 5% sheep blood and 1% pyruvic acid (Wako Pure Chemical Industries Ltd., Osaka, Japan) to enhance the growth of B. wadsworthia organisms (11). Plates were incubated for 48 h with the Anaero Pack system (Mitsubishi Gas Chemical Co., Inc., Tokyo, Japan). For endpoint determination of MICs, 0.2% TTC (Sigma) in 0.1 M sodium phosphate buffer (pH 7.3) was mixed with an equal volume of molten 2% agar and applied over the growth on the plates within 5 min of exposure to air (11). The MICs were read at 5 min after TTC application to allow time for the reduction of TTC to red formazan by viable bacteria. To evaluate the influence of the inoculum size on MICs, bacteria were inoculated at 10⁴, 10⁵, and 10⁶ CFU/spot.

The broth microdilution method was performed with anaerobic bacterial culture medium (ABCM) broth (Eiken Chemical Co., Ltd., Tokyo, Japan) supplemented with 1% pyruvic acid and 3% lysed horse blood. To evaluate the influence of the inoculum size, bacteria were inoculated at 10⁵, 10⁶, and 10⁷ CFU/well. The MICs were determined after a 48-h incubation, as described above.

Susceptibility testing by Etest was performed by applying Etest strips (AB Biodisk, Solna, Sweden), corresponding to the antibiotics used for the other methods, to both BBE agar and supplemented brucella HK agar swabbed evenly with 10⁹ CFU of bacteria per ml. The Etest MICs were also determined after a 48-h incubation, as described above.

The ability of isolates to produce β-lactamase was examined in growth from 48-h cultures on supplemented brucella HK agar by the chromogenic cephalosporin method with a nitrocefin disk (Cefinase; Becton Dickinson).

Table 2 shows the MICs of several antibiotics tested for clinical isolates of B. wadsworthia by the agar dilution method with TTC, the broth microdilution method, and Etest. MICs determined by the agar dilution method were generally higher than those determined by other methods. Particularly, the MICs of all antibiotics other than imipenem and clindamycin were 128 μg/ml or more for the agar dilution method when 10⁵ or 10⁶ CFU of bacteria per spot were inoculated. The influence of the inoculum size was apparently less in the broth microdilution test than in the agar dilution test. The differences of MICs measured by the broth microdilution method between 10⁵ and 10⁷ CFU/ml of inoculation were within twofold in most assays, while those of ampicillin against NBW2 and NBW3 and of imipenem against NBW2, NBW3, and NBW5 differed by over threefold. The combination of sulbactam and ampicillin was more active against NBW3, NBW5, and NBW6 for the agar dilution method and against all isolates for the microdilution method than was ampicillin alone.

TABLE 2.

MICs of antibiotics for clinical isolates of B. wadsworthia, as determined by agar and broth microdilution methods and Etest

Antibiotic	Strain	MIC (μg/ml) by:
		Agar dilution (CFU/spot)			Broth microdilution (CFU/well)			Etest (agar)
		104	105	106	105	106	107	Brucella HK	BBE
Ampicillin	NBW1	16	>128	>128	8	8	8	1.5	3
	NBW2	16	>128	>128	16	16	>128	4	4
	NBW3	16	>128	>128	16	16	>128	16	12
	NBW4	16	>128	>128	8	16	16	4	3
	NBW5	16	>128	>128	16	16	16	12	8
	NBW6	16	>128	>128	8	8	8	4	4
Sulbactam-	NBW1	16	>128	>128	4	4	4	0.125	0.125
ampicillin	NBW2	16	>128	>128	8	8	8	1.5	0.75
	NBW3	2	>128	>128	8	8	8	0.125	0.125
	NBW4	16	>128	>128	4	4	16	0.75	0.25
	NBW5	2	>128	>128	4	8	8	0.25	0.125
	NBW6	2	128	>128	4	4	4	0.75	0.38
Ceftizoxime	NBW1	16	>128	>128	1	1	2	8	4
	NBW2	>128	>128	>128	0.5	0.25	2	48	12
	NBW3	16	>128	>128	1	0.5	0.5	3	1.5
	NBW4	>128	>128	>128	1	2	2	12	8
	NBW5	32	>128	>128	0.5	0.5	2	48	12
	NBW6	16	>128	>128	1	0.5	0.5	8	3
Imipenem	NBW1	0.25	0.25	>128	16	32	32	0.125	0.5
	NBW2	0.125	0.25	>128	8	16	64	0.125	0.25
	NBW3	0.25	0.25	>128	8	16	64	0.125	0.19
	NBW4	0.125	0.25	>128	32	32	64	0.125	0.19
	NBW5	0.25	0.25	>128	16	128	128	0.125	0.25
	NBW6	0.06	0.125	>128	32	32	32	0.094	0.38
Clindamycin	NBW1	0.125	0.25	>128	≦0.06	≦0.06	0.25	0.064	0.125
	NBW2	0.25	0.25	>128	≦0.06	≦0.06	0.25	0.094	0.064
	NBW3	0.25	0.25	>128	0.25	0.25	0.5	0.064	0.064
	NBW4	0.25	0.25	>128	≦0.06	0.125	0.125	0.125	0.064
	NBW5	0.125	0.25	>128	0.25	0.25	0.25	0.047	0.047
	NBW6	0.25	0.25	>128	≦0.06	0.25	0.25	0.125	0.19

Etest MICs between brucella HK agar and BBE agar were comparable, with a tendency to be slightly higher in brucella HK agar than in BBE agar. Generally, it was easier to determine MICs visually in BBE agar than in brucella HK agar (data not shown). Both MICs by Etest were lower than those determined by the agar and broth microdilution methods except in the cases of ceftizoxime and imipenem. MICs did not correspond well with those obtained by the agar and broth microdilution methods, especially for ceftizoxime. In all antimicrobial susceptibility testing, clindamycin was the most potent for all clinical isolates of B. wadsworthia.

Production of β-lactamase was confirmed within 5 min for NBW2, NBW3, NBW5, and NBW6, while the reaction was negative for NBW1 and NBW4 by 1 h.

In this study, B. wadsworthia strains were isolated from clinical specimens, as in previous studies (1–3, 5, 6, 9, 10), including pus of patients with abscesses and resected tissue and peritoneal fluid from those with appendicitis. Also, all isolates were involved in polymicrobial infections, as previously described (1, 3, 9).

Agar dilution with TTC is an excellent method for determining MICs against B. wadsworthia (11); however, a more simple and convenient technique may be required for clinical laboratory examinations. In this study, MICs measured by agar dilution with TTC were generally higher than those by other methods when the inoculum size was large. Although B. wadsworthia organisms did not grow visually in ABCM broth alone (data not shown), they did grow—and MICs against them could be easily judged after a 48-h incubation like other anaerobic bacteria, such as B. fragilis—if pyruvic acid was added to the ABCM broth. These findings suggest that microdilution in the presence of pyruvic acid can be used in clinical laboratory testing.

Etest also proved to be a simple and convenient quantitative method for susceptibility testing of B. wadsworthia in the present study. Since B. wadsworthia grew better on BBE agar than on supplemented brucella HK agar and showed black colonies on the former agar, judgement of MICs was easier with BBE agar than with supplemented brucella HK agar. However, the MICs between organisms grown on BBE agar and those grown on supplemented brucella HK agar were comparable. Our findings indicate that Etest is also a candidate for susceptibility testing of B. wadsworthia isolates in a clinical setting. In this study, MICs determined by Etest were generally lower than those determined by other methods, in spite of the fact that high concentrations of bacteria (10⁹ CFU/ml) were swabbed onto the surface of the agar because of their slow growth and small colony formation. Clindamycin showed antimicrobial activity by all methods, while sulbactam-ampicillin and imipenem did so by Etest only and by both the agar dilution method with TTC and Etest, respectively. At present it is unknown whether MICs determined by microdilution using broth supplemented with pyruvic acid or those determined by Etest are the more accurate.

Finegold et al. reported that 2% of B. wadsworthia strains are resistant to clindamycin (9). In our study, all isolates were susceptible to clindamycin, which was the most potent of the antibiotics studied regardless of the method. In contrast to the previous study (1), recent studies indicate that B. wadsworthia organisms usually produce β-lactamase (9, 11). Four of six isolates of B. wadsworthia produced β-lactamase in our study, supporting this finding.

There are significant differences in the MICs of antimicrobials tested against B. wadsworthia strains depending on the testing method used. The difference is so great that there is major discord. Although the size of the inoculum is a possible factor, the discrepancy cannot be explained solely by the inoculum size. Other factors that may contribute to discrepancies in MIC interpretations include production of a heavy haze (11) and tailing (7), in addition to the fact that the organism is fastidious and its growth is light even on the control plate. Since this study is a preliminary one with a small number of isolates, more strains need to be tested before definitive statements can be made. It is difficult to investigate the correlation of in vitro MICs and the clinical efficacy of antibiotics since B. wadsworthia is isolated in patients with polymicrobial infections; the accumulation and analysis of samples from many cases of B. wadsworthia infection are important and necessary.