Frequency of Resistance in Obligate Anaerobic Bacteria Isolated from Dogs, Cats, and Horses to Antimicrobial Agents

Abstract

Clinical specimens from dogs, cats, and horses were examined for the presence of obligate anaerobic bacteria. Of 4,018 specimens cultured, 368 yielded 606 isolates of obligate anaerobic bacteria (248 from dogs, 50 from cats, and 308 from horses). There were 100 specimens from 94 animals from which only anaerobes were isolated (25 dogs, 8 cats, and 61 horses). The most common sites tested were abdominal fluid (dogs and cats) and intestinal contents (horses). The most common microorganism isolated from dogs, cats, and horses was Clostridium perfringens (75, 13, and101 isolates, respectively). The MICs of amoxicillin with clavulanate, ampicillin, chloramphenicol, metronidazole, and penicillin were determined using a gradient endpoint method for anaerobes. Isolates collected at necropsy were not tested for antimicrobial susceptibility unless so requested by the clinician. There were 1/145 isolates tested that were resistant to amoxicillin-clavulanate (resistance breakpoint ≥ 16/8 μg/ml), 7/77 isolates tested were resistant to ampicillin (resistance breakpoint ≥ 2 μg/ml), 4/242 isolates tested were resistant to chloramphenicol (resistance breakpoint ≥ 32 μg/ml), 12/158 isolates tested were resistant to clindamycin (resistance breakpoint ≥ 8 μg/ml), 10/247 isolates tested were resistant to metronidazole (resistance breakpoint ≥ 32 μg/ml), and 54/243 isolates tested were resistant to penicillin (resistance breakpoint ≥ 2 μg/ml). These data suggest that anaerobes are generally susceptible to antimicrobial drugs in vitro.

INTRODUCTION

Obligate anaerobic bacteria are part of the normal flora of the skin and gastrointestinal tract of animals and are frequently associated with opportunistic infection. They are isolated from various infection sites in domestic animal species, including the liver, gallbladder, bile, draining tracts, abscesses, abdominal and pleural fluid, the respiratory tract, the mouth, feces, and (rarely) the urinary tract (1). The most frequently isolated anaerobes in dogs and cats are of non-spore-forming, Gram-negative genera, including Fusobacterium, Peptostreptococcus, Porphyromonas, and Bacteroides, and spore-forming, Gram-positive genera, particularly Clostridium (1, 2). Anaerobic infections are most often treated empirically because performing culture and antimicrobial susceptibility testing on anaerobes is technically difficult and expensive and may require extended time before results are available (1). In many cases when a culture is performed, subsequent antimicrobial susceptibility testing is unsuccessful due to the poor growth of the isolated anaerobes. Even when antimicrobial susceptibility results are obtained, there can be questions as to the in vivo applicability of in vitro results, as clinical results are not always consistent with antimicrobial agent susceptibility testing (3, 4). There is very limited information available on the interpretive criteria of antimicrobial agent susceptibility to correlate in vitro MICs with clinical efficacy in veterinary anaerobes (5). Many of the antimicrobial drug choices that are made for veterinary patients are based on extrapolations from human medical literature (6).

Choosing an antibiotic empirically for anaerobic infections is reported to be increasingly difficult due to the increasing development of resistance of anaerobic microorganisms to antimicrobial drugs (7). Resistance to beta-lactam antibiotics is common in Bacteroides spp. due to beta-lactamase production (8). The proportion of anaerobic isolates from dogs and cats that are Bacteroides spp. has increased, as has the prevalence of resistant isolates among members of that genus (1). Resistance to clindamycin and tetracyclines in the dog has been reported (1, 5). Resistance to metronidazole in bacterial isolates from dogs or cats has not been commonly reported but has been shown in isolates from horses, particularly Clostridium difficile isolates from foals with diarrhea (9). Compounding the difficulty in antimicrobial drug selection, the infections where anaerobes are culpable are often polymicrobial, with other bacteria (both aerobic and facultatively anaerobic bacterial species) isolated from the same site (1, 2, 10).

Due to the limited data and lack of recent reports on antimicrobial agent susceptibility patterns of anaerobes isolated from dogs, cats, and horses, there is a need for further investigation. The objectives of this study were to quantify the frequency with which anaerobes were cultured in dogs, cats, and horses and to characterize their in vitro antimicrobial susceptibilities.

MATERIALS AND METHODS

A search of the electronic database of Texas A&M University Veterinary Medical Teaching Hospital (TAMU VMTH) records was performed to identify patients with positive anaerobic cultures collected between 1 January 2005 and 1 June 2009, including searching invoices for the terms “anaerobic,” “culture,” and “aerobic” and searching by microorganism name. Data were gathered from each patient record, including case number, species, admitting hospital service, date of admission, site cultured, whether a specimen was taken at necropsy, date of culture, number of anaerobic isolates cultured, number of aerobic isolates cultured, identity of anaerobic isolates, whether antimicrobial agent susceptibility testing was performed on isolates, results of antimicrobial agent susceptibility testing, and history of antimicrobial drug administration to the patient before admission (prior to culture and after culture results). Data were entered into web-based survey software (SelectSurvey.NET 1.5.1; ClassApps, Overland Park, KS) that allowed extraction of data in a spreadsheet (Microsoft Excel; Microsoft, Redmond, WA) for statistical analysis.

Microbiology methods and protocols.

Patient specimens submitted to the Clinical Microbiology Laboratory of the TAMU VMTH with a request for anaerobic culture were uniformly processed following the laboratory standard operating procedure. Specimens from normally sterile sites (e.g., internal organs such as the liver) were used to inoculate one brucella agar plate supplemented with 5% sheep blood, hemin, and vitamin K (brucella plate) (Remel, Lenexa, KS) and one brain heart infusion (BHI) broth (BD, Franklin Lakes, NJ) treated with a commercially available enzyme to remove oxygen (Oxyrase, Inc., Mansfield, OH). Fecal specimens and specimens from the intestinal tract were inoculated onto a brucella plate, a Clostridium difficile selective agar (CDSA) plate (Remel, Lenexa, KS), and a phenylethyl alcohol agar (PEA) plate supplemented with 5% sheep blood (BD, Franklin Lakes, NJ). No broth enrichment was performed for fecal or intestinal specimens. All plates were incubated anaerobically in an anaerobic chamber (Bactron II; Shel Lab, Cornelius, OR) at 37°C. Anaerobiosis was confirmed using an anaerobic indicator strip (BD GasPak; BD, Franklin Lakes, NJ). Plates were examined 48 h after inoculation, and microorganisms were isolated for identification and antimicrobial susceptibility testing by subculture to brucella plates and confirmed as anaerobes by simultaneous subculture to a Trypticase soy agar (TSA) plate supplemented with 5% sheep blood which was incubated aerobically. Microorganisms that failed to divide aerobically were presumptively considered anaerobes and were identified through the use of one of two commercial identification systems, a Vitek ANI card system (bioMérieux, Durham, NC) or a RapID ANA II system (Remel, Lenexa, KS)). Isolates for which there was insufficient growth to inoculate the commercial identification kits, according to the manufacturer's instructions, were identified only to the level of Gram stain reaction and morphology. Laboratory policy is to preferentially perform antimicrobial agent susceptibility testing rather than identification when bacterial growth is limited.

Antimicrobial susceptibility testing to determine the MIC was performed using a gradient endpoint method (the Etest method; AB Biodisk, Solna, Sweden) rather than the agar dilution method, as described in Clinical and Laboratory Standards Institute document M11-A7 (11). Briefly, isolated colonies of each tested microorganism were suspended in a 0.8% saline solution to achieve a density corresponding to a 1.0 McFarland turbidity standard or approximately 3 × 10⁸ CFU/ml. A cotton swab was used to inoculate the surface of a brucella plate, and a gradient diffusion test strip containing amoxicillin with clavulanate, ampicillin, benzylpenicillin, chloramphenicol, clindamycin, or metronidazole was placed on the surface according to the manufacturer's protocol. Plates were then incubated anaerobically at 37°C and examined at 24 h and 48 h to measure the zone of inhibition and determine the MIC for each antimicrobial drug. Results were reported with MIC and interpretation in the TAMU VMTH computerized record system. Isolates from specimens that were collected at necropsy were not tested for antimicrobial susceptibility unless so requested by the clinician.

Analysis.

Summary statistics were performed to characterize the number, type, and antimicrobial susceptibility of isolates, as well as to quantify isolates by species of animal and body site of isolation (IBM SPSS Statistics, version 21).

RESULTS

There were a total of 12,835 requests for culture from specimens collected from canine, feline, and equine patients admitted between 1 January 2005 and 1 June 2009. Among these, there were 4,018 anaerobic culture requests. These included 2,943 requests for anaerobic culture of specimens from live veterinary patients, 904 requests for anaerobic culture of tissues collected at necropsy from deceased patients, and 171 requests for anaerobic blood culture. Of these, 368 cultures (368/4,018 or 9.2%) were positive for the growth of anaerobes. A total of 606 anaerobic isolates were identified in the records search, 248 from dogs, 50 from cats, and 308 from horses (see Table 1). Microorganisms isolated at necropsy represented 23.8% of dog isolates, 32% of cat isolates, and 62.7% of horse isolates. Although identification to the species level was possible for members of several genera, they were grouped by genera for this report for the purpose of brevity, with the exception that members of the Bacteroides fragilis group, Clostridium difficile, and C. perfringens were reported as separate species. The most common species isolated from dogs was Clostridium perfringens (75 isolates, 41 from live animal specimens), followed closely by other Clostridium spp. (57 isolates, 47 from live animal specimens). The next most common isolates from dogs were Bacteroides spp. (14 isolates, 12 from live animal specimens) and Bacteroides fragilis (13 isolates, 11 from live animal specimens). The most common species isolated from cats was Clostridium perfringens (13 isolates, 8 from live animal specimens), and the next most common species isolated from cats was Bacteroides fragilis (nine isolates). The most common species isolated from horses was Clostridium perfringens (101 isolates, 30 from live animal specimens). The next most common species isolated from horse specimens were Clostridium spp. (22 isolates).

Table 1.

Species of anaerobic bacteria isolated from dogs, cats, and horses

Organism(s)	Animal species	No. of live animals	No. of necropsy samples	Total
Actinomyces odontolyticus	Dog	1	0	1
Bacteroides fragilis	Dog	11	2	13
	Cat	9	0	9
	Horse	8	5	13
Bacteroides spp.	Dog	12	2	14
	Cat	2	1	3
	Horse	7	5	12
Clostridium difficile	Dog	7	1	8
	Cat	1	1	2
	Horse	10	13	23
Clostridium perfringens	Dog	41	34	75
	Cat	8	5	13
	Horse	30	71	101
Clostridium spp.a	Dog	47	10	57
	Cat	5	2	7
	Horse	22	30	52
Eubacterium aerofaciens	Horse	1	0	1
Fusobacterium spp.	Dog	13	3	16
	Cat	0	3	3
	Horse	6	9	15
Peptostreptococcus spp.	Dog	10	0	10
	Cat	3	4	7
	Horse	8	11	19
Porphyromonas spp.	Horse	2	0	2
Gram-negative rods	Dog	12	3	15
	Cat	3	0	3
	Horse	6	20	26
Gram-positive cocci	Dog	4	1	5
	Cat	2	0	2
	Horse	3	3	6
Gram-positive rods	Dog	18	3	21
	Horse	8	10	18
Prevotella spp.	Dog	9	0	9
	Cat	1	0	1
	Horse	4	15	19
Propionibacterium spp.	Dog	4	0	4
	Horse	0	1	1
Subtotal	Dog	189	59	248
	Cat	34	16	50
	Horse	115	193	308
Total		338	268	606

^aClostridium species data include all clostridial organisms except Clostridium difficile, C. perfringens, C. tetani, and C. botulinum. There were no isolates of C. tetani or C. botulinum collected during the period of this study.

The body sites from which anaerobic bacteria were isolated, along with the species of the isolate, are shown in Table 2. The most common sites of isolation in dogs were abdominal fluid or swab (53 isolates) and intestinal contents, tissue, or swab (29 isolates). The most common site of isolation in cats was abdominal fluid or swab (11 isolates). The most common sites of isolation from horses were intestinal contents (72 isolates) and abscesses (47 isolates).

Table 2.

Body sites from which anaerobic bacteria were isolated from dogs, cats, and horses

Specimen site	Organism(s)	No. of animals			Total
		Dog	Cat	Horse
Abdominal fluid or swab	Bacteroides fragilis	6	4	0	10
	Bacteroides spp.	7	1	0	8
	Clostridium difficile	1	1	1	3
	Clostridium perfringens	9	2	0	11
	Clostridium spp.	16	3	3	22
	Gram-negative rods	4	0	2	6
	Gram-positive cocci	3	0	0	3
	Gram-positive rods	5	0	1	6
	Peptostreptococcus spp.	1	0	1	2
	Prevotella spp.	1	0	0	1
	Propionibacterium sp.	0	0	1	1
Skin, subcutaneous tissue, and muscle	Actinomyces odontolyticus	1	0	0	1
	Bacteroides fragilis	3	3	3	9
	Bacteroides spp.	0	1	4	5
	Clostridium difficile	2	1	0	3
	Clostridium perfringens	19	1	14	34
	Clostridium spp.	17	2	5	24
	Fusobacterium spp.	4	0	4	8
	Gram-negative rods	3	0	11	14
	Gram-positive cocci	0	2	4	6
	Gram-positive rods	6	0	9	15
	Peptostreptococcus spp.	7	2	3	12
	Porphyromonas spp.	0	0	1	1
	Prevotella spp.	4	1	7	12
Bone	Bacteroides fragilis	0	1	0	1
	Bacteroides spp.	1	0	1	2
	Clostridium difficile	2	0	0	2
	Clostridium perfringens	5	1	0	6
	Clostridium spp.	8	0	9	17
	Fusobacterium spp.	5	0	2	7
	Gram-negative rods	2	0	3	5
	Gram-positive cocci	1	0	0	1
	Gram-positive rods	2	0	2	4
	Peptostreptococcus spp.	1	1	6	8
	Prevotella spp.	1	0	3	4
	Propionibacterium spp.	3	0	0	3
Brain or cerebrospinal fluid	Bacteroides fragilis	0	0	1	1
	Bacteroides spp.	1	0	0	1
	Clostridium perfringens	1	0	2	3
	Clostridium spp.	0	0	1	1
	Fusobacterium spp.	0	0	1	1
	Gram-positive cocci	0	0	1	1
	Gram-positive rods	0	0	2	2
	Peptostreptococcus spp.	0	0	1	1
Circulatory system	Clostridium spp.	1	0	1	2
	Fusobacterium spp.	0	0	1	1
Gastrointestinal tract and feces	Bacteroides fragilis	2	0	4	6
	Bacteroides spp.	2	0	2	4
	Clostridium difficile	2	0	19	21
	Clostridium perfringens	20	2	57	79
	Clostridium spp.	3	1	13	17
	Eubacterium aerofaciens	0	0	1	1
	Fusobacterium spp.	1	0	2	3
	Gram-negative rods	2	3	5	10
	Gram-positive rods	2	0	1	3
	Peptostreptococcus spp.	1	1	1	3
	Prevotella spp.	1	0	3	4
	Propionibacterium spp.	1	0	0	1
Joint, tendon, or tendon sheath	Bacteroides fragilis	0	0	4	4
	Bacteroides spp.	0	0	1	1
	Clostridium perfringens	1	0	3	4
	Clostridium spp.	0	0	3	3
	Gram-negative rods	0	0	1	1
	Peptostreptococcus spp.	0	0	1	1
Liver, gallbladder, or bile	Bacteroides fragilis	1	1	0	2
	Bacteroides spp.	2	0	3	5
	Clostridium difficile	1	0	1	2
	Clostridium perfringens	10	5	12	27
	Clostridium spp.	6	0	5	11
	Fusobacterium spp.	3	1	4	8
	Gram-negative rods	1	0	1	2
	Gram-positive cocci	1	0	0	1
	Gram-positive rods	2	0	1	3
	Peptostreptococcus spp.	0	1	4	5
	Prevotella spp.	1	0	2	3
Pancreatic cyst	Gram-negative rods	1	0	0	1
	Gram-positive rods	1	0	0	1
Pleura, thorax, or mediastinum	Bacteroides spp.	1	0	1	2
	Clostridium difficile	0	0	1	1
	Clostridium perfringens	0	0	1	1
	Clostridium spp.	1	0	1	2
	Gram-positive rods	1	0	0	1
	Porphyromonas spp.	0	0	1	1
	Prevotella spp.	1	0	0	1
Reproductive tract—female	Bacteroides fragilis	1	0	1	2
	Clostridium perfringens	2	0	2	4
	Clostridium spp.	1	0	2	3
	Fusobacterium spp.	0	0	1	1
Reproductive tract—male	Clostridium perfringens	1	0	0	1
Respiratory tract	Clostridium perfringens	0	0	1	1
	Clostridium spp.	0	0	2	2
	Fusobacterium spp.	1	0	0	1
	Gram-negative rods	0	0	1	1
	Gram-positive cocci	0	0	1	1
	Gram-positive rods	0	0	2	2
	Peptostreptococcus spp.	0	0	1	1
Spleen and lymph node	Clostridium difficile	0	0	1	1
	Clostridium perfringens	1	1	7	9
	Clostridium spp.	1	0	6	7
	Fusobacterium spp.	2	1	0	3
	Gram-negative rods	0	0	2	2
	Gram-positive rods	2	0	0	2
	Peptostreptococcus spp.	0	2	1	3
	Prevotella spp.	0	0	4	4
Urinary tract (bladder, kidney, urine)	Bacteroides spp.	0	1	0	1
	Clostridium perfringens	6	1	2	9
	Clostridium spp.	3	1	1	5
	Fusobacterium spp.	0	1	0	1
	Gram-negative rods	2	0	0	2
Total		248	50	308	606

There were 100 specimens collected from 94 veterinary patients in which only anaerobes were isolated (25 dogs, 8 cats, and 61 horses), with a range of one to six anaerobes isolated from each of the specimens collected from various anatomic sites or tissues (Table 3). The majority of specimens that yielded only anaerobes were collected from the gastrointestinal tract followed by skin and abdominal fluid. In 268 cultures, one or more aerobes were isolated in addition to the anaerobes (138 specimens from dogs, 20 from cats, and 110 from horses).

Table 3.

Specimen collection sites for cultures that only yielded obligate anaerobes

Specimen site	No. of cultures			Total no. of cultures
	Dog	Cat	Horse
Gastrointestinal tract	9	0	42	51
Skin, subcutaneous tissue, and musclea	6	0	8	14
Abdominal fluid	2	5	2	9
Boneb	2	1	3	6
Liver or bile	4	0	0	4
Thoracic or pleural fluid	1	0	2	3
Uterus	1	0	1	2
Lung and spleen	0	1	1	2
Brain			1	1
Unknown	0	1	1	2
Total	25	8	61	94

^aIncludes wounds and abscesses.

^bIncludes tympanic bullae and maxillary sinus.

Table 4 and Table 5 show the MICs of all the isolates from live animal specimens. Using published interpretive criteria (12), 1/145 isolates tested were resistant to amoxicillin-clavulanate (resistance breakpoint ≥ 16/8 μg/ml), 7/77 isolates tested were resistant to ampicillin (resistance breakpoint ≥ 2 μg/ml), 4/242 isolates tested were resistant to chloramphenicol (resistance breakpoint ≥ 32 μg/ml), 12/158 isolates tested were resistant to clindamycin (resistance breakpoint ≥ 8 μg/ml), 10/247 isolates tested were resistant to metronidazole (resistance breakpoint ≥ 32 μg/ml), and 54/243 isolates tested were resistant to penicillin (resistance breakpoint ≥ 2 μg/ml). The MICs for Actinomyces odontolyticus were 0.5 μg/ml for amoxicillin-clavulanate, 0.25 μg/ml for metronidazole, and 0.125 μg/ml for penicillin. The MICs for the single Propionibacterium sp. tested were 0.5 μg/ml for amoxicillin-clavulanate, 2 μg/ml for chloramphenicol, 0.064 μg/ml for clindamycin, 0.064 μg/ml for metronidazole, and 0.125 μg/ml for penicillin.

Table 4.

MICs of Gram-negative anaerobic bacteria isolated from dogs, cats, and horses

Organism(s) and antibiotic(s)	No. of isolates tested	MIC (μg/ml)
		<0.016	0.016	0.023	0.032	0.047	0.064	0.094	0.125	0.19	0.25	0.38	0.5	0.75	1	1.5	2	3	4	6	8	12	16	24	32	48	>48
Bacteroides spp.
Amoxicillin-clavulanate	27					1		2	4	8	4	2	2		1		2	1
Ampicillin	12	1			2	1			1			1							1		2		1	1		1
Chloramphenicol	43												1	1	5	12	10	10	2	1	1
Clindamycin	30	2			1		2	1	2	1	2	2	1	2	2	6	2	1	2	1
Metronidazole	43		1				2	1	2	9	13	4		5	2	1			1				1				1
Penicillin	42		1	2	1	1		2						1		1	2		1	3	10	4	7	2	1	1	2
Fusobacterium spp.
Amoxicillin-clavulanate	9	3		1	2						2				1
Ampicillin	3			1	1	1
Chloramphenicol	13						1					1	4	3	1	1	1				1
Clindamycin	8	2	1	1		1		1						1				1
Metronidazole	14	3			1	2	4	1		1							1					1					1
Penicillin	14	3	1	3		1		1		1																	3
Porphyromonas spp.
Ampicillin	2		2
Chloramphenicol	2											1	1
Metronidazole	2						1			1
Penicillin	2		2
Prevotella spp.
Amoxicillin-clavulanate	6				1			2	2																		1
Ampicillin	1									1
Chloramphenicol	8											1	1	3	2	1
Clindamycin	6	3	1						1	1
Metronidazole	8			1	2		1		3				1
Penicillin	7	1		1			1						1						1								2
Gram-negative rods
Amoxicillin-clavulanate	2								1	1
Ampicillin	4					1				1				1	1
Chloramphenicol	10										3	2		1	1	1	1				1
Clindamycin	6			2		1					1					1			1
Metronidazole	10	1	1	2			1	1	1	1	1																1
Penicillin	8	1				1		1	3			1									1
Total
Amoxicillin-clavulanate	44	3		1	3	1		4	7	9	6	2	2		2		2	1									1
Ampicillin	22	1	2	1	3	3			1	2		1		1	1				1		2		1	1		1
Chloramphenicol	76						1				3	5	7	8	9	15	12	10	2	1	3
Clindamycin	50	7	2	3	1	2	2	2	3	2	3	2	1	3	2	7	2	2	3	1
Metronidazole	77	4	2	3	3	2	9	3	6	12	14	4	1	5	2	1	1		1			1	1				3
Penicillin	73	5	4	6	1	3	1	4	3	1		1	1	1		1	2		2	3	11	4	7	2	1	1	7

Table 5.

MICs of Gram-positive anaerobic bacteria isolated from dogs, cats, and horses

Organism(s) and antibiotic(s)	No. of isolates tested	MIC (μg/ml)
		<0.016	0.016	0.023	0.032	0.047	0.064	0.094	0.125	0.19	0.25	0.38	0.5	0.75	1	1.5	2	3	4	6	8	12	16	24	32	48	>48
Clostridium difficile
Amoxicillin-clavulanate	8								2	1	1	1	2	1
Ampicillin	5						1			1	1		1	1
Chloramphenicol	14									1	1	1	1	1	2	2		2	1				1		1
Clindamycin	7		1						1				1				2		1	1
Metronidazole	14		1	1	1	1	1		1	3		1	1	1		1	1
Penicillin	13					1	2		3		1		1	3		2
Clostridium perfringens
Amoxicillin-clavulanate	34		3	6	7	4	2	3	2	2	2		1		1			1
Ampicillin	23	2	7	4	1	1	1	2			3				2
Chloramphenicol	62					1									1	6	18	32	4
Clindamycin	39			1		3	4	3	2				2	1	1	5	6	2	4	2	1	1					1
Metronidazole	62					2		2		1	2	3	4	14	12	16	1		2	2	1
Penicillin	62				2	12	14	12	1	7	5		1	1	1		1		3		1			1
Clostridium spp.
Amoxicillin-clavulanate	35	1	1	3	4	6	2	1	4	4	3	2			1	2		1
Ampicillin	15	1			1	2	1	1		3	2	2		1					1
Chloramphenicol	55		1								1	2	4	4	8	14	8	6	2	2	1		1				1
Clindamycin	40			2	1	1	1	1		2	2	2	1			2	6	1	7	5	1	1	1				3
Metronidazole	57			2	2	5	6	3	8	4	3	4	5		2	1	2	1	1		1		1	1	1		4
Penicillin	56	3	4	4	3	3	2	5	7	4	4	5	4		2	3					1						2
Peptostreptococcus spp.
Amoxicillin-clavulanate	10		1					1	2	1	1		2	1		1
Ampicillin	7							2	3	1	1
Chloramphenicol	15											2	3	2	1	4	2		1
Clindamycin	8	1			1		3						1		1		1
Metronidazole	17		1		1	1	2	2		1	2	1	2			1	1										2
Penicillin	17		1		3	1	1	2	2	2	2	1	2
Gram-positive cocci
Amoxicillin-clavulanate	3						1			1				1
Ampicillin	2								1				1
Chloramphenicol	5												1			2	1										1
Clindamycin	3		1													1											1
Metronidazole	5						1			1	1	1					1
Penicillin	5								2	1			1													1
Gram-positive rods
Amoxicillin-clavulanate	9		1		2		2				3					1
Ampicillin	3				1			1				1
Chloramphenicol	14			1								1	1		2	1	5	1	1								1
Clindamycin	10	2	1	1	1						1					1				1							2
Metronidazole	13	1				2				1		3	3		1		1						1
Penicillin	14	1	1	2			3		1	1	1	1							1	1	1
Total
Amoxicillin-clavulanate	99	1	6	9	13	10	7	5	10	9	10	3	5	3	2	4		2
Ampicillin	55	3	7	4	3	3	3	6	4	5	7	3	2	2	2				1
Chloramphenicol	165		1	1		1				1	2	6	10	7	14	29	34	41	9	2	1		2		1		3
Clindamycin	107	3	3	4	3	4	8	4	3	2	3	2	5	1	2	9	15	3	12	9	2	2	1				7
Metronidazole	168	1	2	3	4	11	10	7	9	11	8	13	15	15	15	19	7	1	3	2	2		2	1	1		6
Penicillin	167	4	6	6	8	17	22	19	16	15	13	7	9	4	3	5	1		4	1	3			1		1	2

DISCUSSION

There were some important limitations to our study. First, the limited number of isolates in this study precludes the making of broad statements regarding differences between animal species and differences from human patients in the anaerobes cultured and their antimicrobial agent susceptibility patterns. Clindamycin was not included in the antimicrobial susceptibility testing for anaerobes from horses as the use of this drug in the equine species is lethal. Due to the qualitative nature of the history noted in the medical record, previous antibiotic use was not consistently recorded in all cases and therefore was not included in this study. Additionally, 96 of the 338 anaerobes cultured were not identified beyond the Gram stain reaction and morphology. This limitation was due in part to the laboratory policy of preferentially performing antimicrobial susceptibility testing rather than identification on anaerobes that divide poorly (i.e., in cases of adequate growth for antimicrobial agent susceptibility testing but insufficient amounts of microorganism for phenotypic identification tests). We anticipate that future adoption of systems that allow identification with less material such as DNA sequencing or that use the newer mass spectrometry (e.g., matrix-assisted laser desorption ionization–time of flight) instruments will allow identification of these microorganisms and improve identification results.

In our study, anaerobes were predominantly cultured from sites associated with the gastrointestinal tract or with translocation from the gastrointestinal tract (e.g., abdominal fluid) in 226 isolates (91 isolates from dogs, 18 isolates from cats, and 117 isolates from horses) followed by those from the skin (64 isolates from dogs, 14 isolates from cats, and 59 isolates from horses). This differs from other studies which more commonly found isolates from the respiratory tract (13). These differences may relate to differences in the types of specimens submitted as well as changes in clinical veterinary practice over the intervening 30 years. Similar to other studies, Clostridium perfringens and other Clostridium spp. were the predominant microorganisms isolated from dogs and cats followed by Bacteroides spp. (14).

In general, the MICs for the anaerobes tested were low and would have been considered to represent susceptibility whether evaluated using the CLSI criteria for aerobes or the criteria for anaerobes published in the 10th edition of the Manual of Clinical Microbiology (12) which is used by the laboratory. Whether these values reflect clinical efficacy of these drugs was not evaluated by this study.

There has been some debate over the use of the gradient diffusion test strip method in preference to the agar dilution method as described by CLSI for anaerobe antimicrobial agent susceptibility testing (15, 16). A recent study comparing two gradient diffusion methods with the agar dilution method found that the gradient diffusion method may not recognize (may underestimate) resistance to some antimicrobial agents in some anaerobes (16). In that study, penicillin and metronidazole had reduced MICs, with very major errors in agreement between the methods observed for the Bacteroides fragilis group and the Clostridium species isolates. As these microorganisms represent the majority of microorganisms isolated from dogs, cats, and horses, care should be taken when evaluating clinical response to these drugs when they are predicted to be susceptible using a gradient diffusion method. It remains possible that use of the gradient diffusion method in our study underestimated the MIC values for the isolates tested, as no comparison was made between this method and the CLSI agar dilution method and isolates are not retained by the laboratory, thus precluding retrospective analysis. The laboratory uses the gradient diffusion method rather than the agar dilution method because of the increased labor required in the agar dilution method.

Despite our original expectation that antimicrobial resistance might preclude effective empirical decisions, these data suggest that the empirical choice of antimicrobial drug based on animal species and location of infection remains a reasonable clinical decision for treatment of suspected anaerobic infections in veterinary medicine in cases where anaerobic culture is not an option due to cost or patient considerations. Additional studies are needed to correlate clinical outcome with in vitro MIC determinations for anaerobes and for comparison of gradient diffusion and agar dilution methods for testing veterinary anaerobic isolates.