Antimicrobial susceptibility testing of Brachyspira intermedia and Brachyspira pilosicoli isolates from Australian chickens

Abstract

Susceptibilities of predominantly Australian isolates of the pathogenic intestinal spirochaetes Brachyspira intermedia (n=25) and Brachyspira pilosicoli (n=17) from chickens were tested in agar dilution against four concentrations each of the antimicrobials tiamulin, lincomycin, tylosin, metronidazole, tetracycline and ampicillin. Based on available minimum inhibitory concentration (MIC) breakpoint values for Brachyspira hyodysenteriae or other Gram-negative enteric veterinary pathogens, isolates of both species generally were susceptible to tiamulin, lincomycin, metronidazole and tetracycline. Although not classed as resistant, four isolates of B. intermedia had an elevated MIC range for tiamulin (1 to 4 mg/l), 11 isolates of B. intermedia and five of B. pilosicoli had an elevated MIC range for lincomycin (10 to 50 mg/l), one isolate of B. pilosicoli had an elevated MIC range for tetracycline (10 to 20 mg/l), and one isolate of B. intermedia and five of B. pilosicoli had an elevated MIC range for ampicillin (10 to 50 mg/l). A clear lack of susceptibility to tylosin (MIC > 4 mg/l) was seen in 11 isolates each of B. intermedia and B. pilosicoli, and to ampicillin (MIC > 32 mg/l) in two isolates of B. pilosicoli. These data suggest that some resistance to common antimicrobials exists among intestinal spirochetes obtained from laying hens and supports the need of MIC data for clinical isolates before any treatment is considered.

Détermination de la sensibilité aux antibiotiques des souches de Brachyspira intermedia et Brachyspira pilosicoli isolées chez des poulets australiens

Les sensibilités des souches principalement australiennes de spirochètes pathogènes intestinaux Brachyspira intermedia (n = 25) et Brachyspira pilosicoli (n = 17) isolés chez des poulets ont été testées par la méthode de dilution en milieu gélosé vis-à-vis de quatre concentrations de chacun des antibiotiques suivants : tiamuline, lincomycine, tylosine, métronidazole, tetracycline et ampicilline. En se basant sur les valeurs seuil de concentrations minimales inhibitrices (CMI) disponibles pour Brachyspira hyodysenteriae ou d'autres agents pathogènes entériques Gram négatif d'origine vétérinaire, les souches des deux espèces se sont révélées généralement sensibles à la tiamuline, la lincomycine, le métronidazole et la tétracycline. Bien que non classées comme résistantes, quatre souches de B. intermedia avaient des valeurs élevées de CMI pour la tiamuline (1–4 mg/l), 11 souches de B. intermedia et cinq de B. pilosicoli avaient des valeurs élevées de CMI pour la lincomycine (10–50 mg/l), une souche de B. pilosicoli avait des valeurs élevées de CMI pour la tétracycline (10–20 mg/l), et une souche de B. intermedia ainsi que cinq de B. pilosicoli avaient des valeurs élevées de CMI pour l'ampicilline (10–50 mg/l). Un manque évident de sensibilité à la tylosine (CMI > 4 mg/l) a été observé pour 11 souches de B. intermedia et 11 de B. pilosicoli, et à l'ampicilline (CMI > 32 mg/l) pour deux souches de B. pilosicoli. Ces données suggèrent que des résistances à des antimicrobiens classiques existent parmi les spirochètes intestinaux isolés chez des poules pondeuses et sont en faveur de la nécessité de la détermination des CMIs pour les souches isolées de cas cliniques avant d'entreprendre un traitement.

Antimikrobielle Empfindlichkeitstestung von Brachyspira intermedia- und Brachyspira pilosicoli-Isolaten aus australischen Hühnern

Die Empfindlichkeit von überwiegend australischen Isolaten der pathogenen Darmspirochäten Brachyspira intermedia (n = 25) und Brachyspira pilosicoli (n = 17) aus Hühnern wurde in der Agarverdünnung gegen vier Konzentrationen jedes der Antiinfectiva Tiamulin, Lincomycin, Tylosin, Metronidazol, Tetrazyklin und Ampicillin getestet. Basierend auf bekannten minimalen Hemmkonzentrations (MHK)-Grenzwerten für Brachyspira hyodysenteriae oder andere gramnegative, veterinärmedizinisch bedeutsame Darmpathogene wurden beide Spezies als generell empfindlich gegenüber Tiamulin, Lincomycin, Metronidazol und Tetrazyklin eingestuft. Obwohl nicht als resistent klassifiziert, hatten vier B. intermedia-Isolate einen erhöhten MHK-Bereich für Tiamulin (1–4 mg/l), 11 B. intermedia-Isolate und fünf B. pilosicoli-Isolate für Lincomycin (10.50 mg/l), ein B. pilosicoli-Isolat für Tetrazyklin (10–20 mg/l) sowie ein B. intermedia-Isolat und fünf B. pilosicoli-Isolate für Ampicillin (10–50 mg/l). Eine eindeutige Unempfindlichkeit gegenüber Tylosin (MHK > 4 mg/l) wurde bei 11 Isolaten beider Spezies und gegenüber Ampicillin (MHK > 32 mg/l) bei zwei B. pilosicoli-Isolaten festgestellt. Diese Ergebnisse lassen vermuten, dass bei den aus Legehennen stammenden Darmspirochäten einige Resistenzen gegenüber den üblichen Antiinfektiva bestehen, und unterstützen die Notwendigkeit der Ermittlung von MHC-Werten für Isolate aus klinischen Fällen vor der Festlegung einer Behandlung.

Susceptibilidad antimicrobiana de aislados de Brachyspira intermedia y Brachyspira pilosicoli de pollos australianos

Se evaluó la susceptibilidad de los aislados australianos más importantes de las espiroquetas intestinales patógenas Brachyspira intermedia (n = 25) y Brachyspira pilosicoli (n = 17) de pollos en diluciones de ágar frente a cuatro concentraciones de cada uno de los antimicrobianos: tiamulina, lincomicina, tilosina, metronidazol, tetraciclina y ampicilina. En base a los valores de corte de las concentraciones mínimas inhibitorias (MIC) disponibles para Brachyspira hyodysenteriae y otro patógenos gram-negativos entéricos de origen animal, los aislados de ambas especies fueron en general susceptibles a la tiamulina, lincomicina, metronidazol y tetraciclina. Aunque no se consideraron como resistentes, cuatro aislados de B. intermedia presentaron un elevado valor de MIC para la tiamulina (1–4 mg/l), 11 aislados de B. intermedia y cinco de B. pilosicoli presentaron un elevado valor de MIC para la lincomicina (10–50 mg/l), un aislado de B. pilosicoli presentó un elevado valor de MIC para la tetraciclina (10–20 mg/l), y un aislado de B. intermedia y cinco de B. pilosicoli presentaron un elevado valor de MIC para la ampicilina (10–50 mg/l). No se observó susceptibilidad a la tilosina (MIC > 4 mg/l) en 11 aislados de B. intermedia and B. pilosicoli, y a la ampicilina (MIC > 32 mg/l) en dos aislados de B. pilosicoli. Estos datos sugieren que existe una cierta resistencia a los antimicrobianos más comunes entre las espiroquetas intestinales obtenidas a partir de gallinas de puesta y remarca la necesidad de valorar los datos de MIC de los aislados clínicos antes de considerar un tratamiento.

Introduction

Anaerobic intestinal spirochaetes of the genus Brachyspira commonly colonize the caeca and colon/rectum of various species of birds (Stephens & Hampson, 2001). Brachyspira species identified in birds include the potentially pathogenic Brachyspira intermedia, Brachyspira pilosicoli, Brachyspira alvinipulli, and Brachyspira hyodysenteriae, as well as the presumed commensal “non-pathogenic” species Brachyspira innocens, Brachyspira murdochii, and “Brachyspira pulli” (Jensen et al., 1996; McLaren et al., 1997; Stanton et al., 1998; Stephens & Hampson, 2001; Jansson et al., 2004). Intestinal spirochaete infections have been reported to cause problems in flocks of laying hens and broiler breeder hens, where the associated clinical condition has been called “avian intestinal spirochaetosis” (AIS) (Swayne & McLaren, 1997). AIS has been linked to reduced egg production, delayed onset of laying, reduced growth rates, increased feed consumption, and increased faecal moisture content that leads to wet litter (Davelaar et al., 1986; Griffiths et al., 1987; Dwars et al., 1989 1992; Swayne et al., 1992; Trampel et al., 1994; Smit et al., 1998). The two pathogenic species most commonly involved in AIS are B. intermedia and B. pilosicoli (Stephens & Hampson, 2001; Stephens et al., 2005). To date, B. alvinipulli has only been reported from diseased chickens in one flock in the USA (Swayne et al., 1992; McLaren et al., 1997), and infection with B. hyodysenteriae (the agent of swine dysentery) has not been reported as a natural cause of disease in commercial poultry.

In comparison with the situation with B. hyodysenteriae and B. pilosicoli infections in pigs, where a large body of literature is available, relatively little has been published on means to treat AIS (Swayne, 2003). Treatment of infected laying hens with antimicrobials is complicated by a lack of registered products, and potential problems of antimicrobial residues in eggs for human consumption. Antimicrobial agents that have been reported to temporarily reduce clinical signs of AIS in chickens include 5-nitroimidazole in the drinking water of breeder birds (Smit et al., 1998), and tiamulin, lincospectin or oxytetracycline in the drinking water of laying hens (Stephens & Hampson, 1999).

It is known that isolates of B. hyodysenteriae and B. pilosicoli that infect pigs may develop resistance if regularly exposed to antimicrobials (Karlsson et al., 2002 2004), and the same is likely to be true for Brachyspira spp. isolates from chickens. Consequently, prior to medicating for pathogenic Brachyspira spp. in chickens it is important to determine the probable effectiveness of available antimicrobial agents. To date there has been only one published report on the in-vitro antimicrobial susceptibilities of intestinal spirochaetes from chickens, and this involved small numbers of US isolates (Trampel et al., 1999). The two isolates of B. pilosicoli and the two of B. alvinipulli tested were susceptible to tiamulin, lincomycin and carbadox, resistant to streptomycin, and gave strain-dependent results for chlortetracycline, oxytetracycline, tylosin, bacitracin, erythromycin, neomycin, and penicillin.

The aim of the current study was to gain more information about the probable antimicrobial susceptibilities of intestinal spirochaetes from chickens. The antimicrobials tested were mainly those used to treat B. hyodysenteriae and B. pilosicoli infections in pigs, because it was presumed that these also would be effective for treating Brachyspira spp. in chickens. The type strain of B. hyodysenteriae (B78^T) was used as a positive control strain throughout the study, as MICs for this strain have been established in our laboratory (Karlsson et al., 2002), and no chicken reference strains of B. intermedia and B. pilosicoli of known antimicrobial susceptibility profiles are available. Since there have been no widely accepted MIC breakpoints for Brachyspira spp., other than for B. hyodysenteriae, breakpoints for B. hyodysenteriae were used where available.

Materials and Methods

Spirochaete isolates

Isolates of B. intermedia (n=25) and B. pilosicoli (n=17) from chickens were obtained as frozen stock from the culture collection held at the Reference Centre for Intestinal Spirochaetes at Murdoch University. Before being stored the isolates had been subcultured at least three times to ensure their purity. All isolates had been identified and typed using species-specific polymerase chain reactions and multilocus enzyme electrophoresis (MLEE) (Stephens & Hampson, 1999; Stephens et al., 2005). The electrophoretic types (ETs) identified in the MLEE study of Stephens et al. (2005) were used to help assist with interpretation of the results of the current study. Twenty-seven of the isolates of both species originated from five laying hen flocks and three broiler breeder flocks in Queensland (Stephens & Hampson, 1999), 10 isolates of B. intermedia were from nine assorted laying hen and broiler breeder flocks in Western Australia (WA) (McLaren et al., 1996 1997), three isolates of B. pilosicoli and one of B. intermedia were from chickens in The Netherlands, and one B. pilosicoli isolate was from a chicken in the USA (McLaren et al., 1997). The type strain of B. hyodysenteriae (B78^T) was included as a control strain for standardization of antimicrobial susceptibility testing.

Antimicrobial susceptibility testing

Stock solutions were made by mixing the following antimicrobial powders with sterile distilled water: tiamulin hydrogen fumerate (99.5% activity), lincomycin hydrochloride (808 u/mg), tylosin tartrate (906 mg/g), metronidazole (100%), tetracycline hydrochloride (950 mg/g), and ampicillin (100%). Tiamulin hydrogen fumarate was obtained from Novartis Animal Health (Sydney, Australia), and the other antimicrobials were from the Sigma Chemical Company (St Louis, Missori, USA). Stock solutions were stored at 4°C for less than 24 h before use.

Antimicrobial susceptibility plates were made using Trypticase Soy Agar (TSA), supplemented with 5% defibrinated ovine blood. Four concentrations of each antimicrobial were tested and the appropriate amount of each stock solution was added to the agar immediately before pouring the plates (Table 1). Antimicrobial sensitivity plates and control plates containing TSA with 5% ovine blood but no antimicrobials were dried for 15 min at 37°C before inoculation with the measured amount of test organisms. The inoculum was allowed to dry before incubating in an atmosphere of 94% H₂ and 6% CO₂ at 37°C.

Table 1.  Antimicrobial concentrations used in the agar dilution antimicrobial sensitivity test

Antimicrobial agent	Concentrations tested (mg/l)
Tiamulin	0.1, 1, 4^a, 10
Lincomycin	1, 10, 50, 100
Tylosin	4^a, 20, 50, 100
Metronidazole	0.1, 1, 10, 20
Tetracycline	1, 5, 10, 20
Ampicillin	1, 10, 50, 100
^aA concentration of 5 mg/l was tested initially.

For preparation of the inoculum, isolates were plated onto TSA with 5% defibrinated ovine blood and incubated in an atmosphere of 94% H₂, 6% CO₂ at 37°C for 5 days. The cells were gently resuspended in 1 ml sterile phosphate-buffered saline and then counted using a haemocytometer and phase-contrast microscope. A total of 10⁵ cells were drop-inoculated onto the control and sensitivity plates. Each isolate was tested at least in duplicate, and B. hyodysenteriae control strain B78^T was included in each batch of tests.

Growth of the strains on the control and sensitivity plates was checked visually after 5 days of incubation. Zones of weak haemolysis were present around growth on the control plates, and isolates were recorded as being susceptible to the antimicrobial concentration in the test plates if no such zones were observed. Any surface growth was scraped off the plate and examined under a phase contrast microscope to confirm purity and the endpoint. As only four dilutions of each antimicrobial were used, and in some cases these were quite widely separated, the minimum inhibitory concentration (MIC) values were recorded as being in the range between the highest sensitive concentration and the lowest resistant concentration. The MIC ranges of the antimicrobials at which 50 and 90% of the isolates were inhibited (MIC₅₀ and MIC₉₀, respectively) also were calculated for each species. MIC breakpoints used to assist interpretation of the results are presented in Table 2. For tiamulin, lincomycin, tylosin and metronidazole, values identified for B. hyodysenteriae by Rønne & Szancer (1990) were applied, while for tetracycline and ampicillin MIC breakpoints for Gram-negative enteric veterinary pathogens were used (National Committee for Clinical Laboratory Standards, 1999).

Table 2.  MIC breakpoints (mg/l) for in-vitro antimicrobial susceptibility tests as used to assist in interpretation in the study

Antimicrobial	Sensitive^a	Intermediate^a	Resistant^a
Tiamulin	≤1	>1 ≤ 4	>4
Lincomycin	≤4	>4 ≤ 36	>36
Tylosin	≤1	>1 ≤ 4	>4
Metronidazole	≤4	>4 ≤ 16	>16
Tetracycline	≤4	8	≥16
Ampicillin	≤8	16	≥32
^aBased on MICs determined for B. hyodysenteriae against tiamulin, lincomycin, tylosin and metronidazole, and for Gram-negative enteric veterinary pathogens against tetracycline and ampicillin.

Results

MIC ranges for the antimicrobials tested against B. hyodysenteriae strain B78^T were consistent between batches, and the values agreed with previous results obtained in our laboratory for this strain (Karlsson et al., 2002). The MIC ranges for the chicken B. intermedia and B. pilosicoli isolates are presented in Table 3.

Table 3.  Susceptibilities to six antimicrobials of 25 isolates of B. intermedia and 17 isolates of B. pilosicoli from hens in layer (L) and broiler breeder (BB) flocks in Queensland (Q), Western Australia (WA), The Netherlands and the USA

			MIC range (mg/l)
ET^a	Isolate name	Flock	Tiamulin	Lincomycin	Tylosin	Metronidazole	Tetracycline	Ampicillin
B. intermedia
42	QAW40	Q-L 2	0.1–1	10–50	4–20	0.1–1	<1	<1
43	QAW32	Q-L 3	0.1–1	10–50	4–20	0.1–1	<1	<1
44	histo 6	WA 1	0.1–1	<1	<4	<0.1	<1	1–10
45	ABB60-9	WA 2	1–4	<1	<4	0.1–1	<1	<1
45	B52iii	WA 3	0.1–1	<1	<4	<0.1	<1	<1
46	E2	WA 4	0.1–1	<1	<4	<0.1	<1	<1
46	A7	WA 5	0.1–1	<1	<4	<0.1	<1	<1
47	1380	The Netherlands	<0.1	<1	<4	0.1–1	<1	<1
48	HB60-1	WA 6	<0.1	1–10	<4	<0.1	1–5	<1
51	B37 iii	WA 3	0.1–1	<1	<4	<0.1	1–5	<1
52	QAW20	Q-L 7	0.1–1	10–50	4–20	0.1–1	<1	<1
53	22-6	WA 7	<0.1	<1	<4	<0.1	<1	<1
53	3B-1	WA 8	0.1–1	<1	<4	<0.1	<1	<1
53	P3	WA 9	0.1–1	<1	<4	<0.1	<1	<1
54	QAW4	Q-L 5	0.1–1	<1	<4	0.1–1	<1	<1
57	QAW15	Q-L 6	1–4	10–50	>100	0.1–1	1–5	<1
58	QAW17	Q-L 6	0.1–1	10–50	4–20	<0.1	<1	<1
59	QAW12	Q-L 6	0.1–1	10–50	50–100	0.1–1	1–5	<1
60	QAW1	Q-L 2	1–4	10–50	>100	0.1–1	1–5	<1
60	QAW2	Q-L 2	1–4	10–50	>100	0.1–1	1–5	<1
60	QAW3	Q-L 2	0.1–1	10–50	50–100	0.1–1	<1	<1
61	QAW18	Q-BB 4	0.1–1	10–50	20–50	0.1–1	1–5	<1
62	QAW21	Q-BB 4	0.1–1	10–50	20–50	<0.1	<1	<1
63	QAW7	Q-BB 1	<0.1	<1	<4	0.1–1	<1	<1
64	QAW36	Q-BB 2	<0.1	<1	<4	0.1–1	<1	10–50
B. pilosicoli
71	QAP3	Q-BB 4	<0.1	10–50	>100	0.1–1	<1	<1
73	QAP1	Q-BB 4	<0.1	1–10	50–100	0.1–1	1–5	>100
73	QAP5	Q-BB 4	<0.1	<1	<4	<0.1	<1	10–50
73	QAP6	Q-BB 4	<0.1	1–10	4–20	0.1–1	1–5	>100
73	QAP7	Q-BB 4	<0.1	1–10	<4	<0.1	<1	10–50
73	QAP10	Q-BB 4	<0.1	<1	<4	0.1–1	<1	10–50
74	QAP4	Q-L 2	<0.1	1–10	>100	0.1–1	<1	10–50
76	QAP8	Q-BB 1	<0.1	<1	4–20	0.1–1	<1	<1
76	QAP12	Q-BB 1	<0.1	<1	<4	0.1–1	<1	<1
76	QAP14	Q-BB 1	<0.1	<1	<4	0.1–1	<1	<1
76	QAP15	Q-BB 1	<0.1	<1	<4	0.1–1	<1	<1
76	QAP16	Q-BB 1	<0.1	10–50	>100	0.1–1	<1	10–50
78	QU-1	Q-BB 1	0.1–1	1–10	4–20	0.1–1	1–5	<1
79	4742	The Netherlands	0.1–1	10–50	>100	0.1–1	<1	<1
81	1772	The Netherlands	<0.1	<1	50–100	0.1–1	1–5	1–10
81	42167	USA	0.1–1	10–50	>100	<0.1	10–20	1–10
82	13316	The Netherlands	<0.1	10–50	>100	0.1–1	<1	<1
^aElectrophoretic type in multilocus enzyme electrophoresis: from Stephens et al. (2005).

Generally, isolates of both species were susceptible to most of the antimicrobials tested, although sometimes the MIC breakpoints fell within the MIC range, making interpretation difficult. For example, 11 isolates of B. intermedia and five of B. pilosicoli had an elevated MIC range for lincomycin (10 to 50 mg/l), spanning the value of 36 mg/l considered to represent resistance in B. hyodysenteriae (Rønne & Szancer, 1990). Four isolates of B. intermedia from three farms had an elevated MIC range for tiamulin (1 to 4 mg/l), one isolate of B. pilosicoli had an elevated MIC range for tetracycline (10 to 20 mg/l), and one isolate of B. intermedia and five of B. pilosicoli from three farms had an elevated MIC range for ampicillin (10 to 50 mg/l). In comparison, a clear lack of susceptibility to tylosin (MIC >4 mg/l) was seen for 11 isolates of B. intermedia (44%) and 11 isolates of B. pilosicoli (65%). Similarly, two isolates of B. pilosicoli (12%), both from farm Q-BB 4, were clearly not susceptible to ampicillin in vitro (MIC > 32 mg/l). The MIC₅₀ and MIC₉₀ values for the two species are presented in Table 4. These values reflect a tendency for the collection of B. intermedia isolates to be less susceptible to tiamulin than the B. pilosicoli collection, with the latter tending to be less susceptible to lincomycin and tylosin, and clearly less susceptible to ampicillin than the isolates in the B. intermedia collection.

Table 4.  MIC₅₀ and MIC₉₀ ranges (mg/l) for the six antimicrobials tested with the B. intermedia and B. pilosicoli isolates

	B. intermedia		B. pilosicoli
Antimicrobial	MIC50	MIC90	MIC50	MIC90
Tiamulin	0.1–1	1–4	<0.1	0.1–1
Lincomycin	<1	10–50	1–10	10–50
Tylosin	<4	>100	4–20	>100
Metronidazole	0.1–1	0.1–1	0.1–1	0.1–1
Tetracycline	<1	1–5	<1	1–5
Ampicillin	<1	<1	1–10	>100

Three farms had multiple isolates that belonged to the same Brachyspira species and had the same ET on MLEE. Farm Q-L 2 had three isolates of B. intermedia in ET 60, of which isolate QAW3 had lower MIC ranges for tiamulin, tylosin, and tetracycline than isolates QAW1 and QAW2. Farm Q-BB 4 had five isolates of B. pilosicoli in ET 73, and again there were differences between some isolates in their MIC ranges for the antimicrobials tested, except for tiamulin. Farm Q-BB 1 had five isolates of B. pilosicoli in ET 76, and differences were observed between some of them in MIC ranges for lincomycin, tylosin, and ampicillin. Other farms had multiple isolates belonging to closely related ETs, and again there was a tendency for some of these isolates to vary in their MIC ranges for some of the antimicrobials (e.g. isolates QAW15, QAW17 and QAW12 in related ETs 57, 58 and 59).

There were no obvious or consistent differences in the antimicrobial susceptibilities of isolates from laying hen flocks compared with those from broiler breeder flocks. The 10 WA isolates of B. intermedia were susceptible to tylosin (MICs < 4 mg/l), while 11 of 14 (79%) Queensland isolates of B. intermedia were not susceptible. There were no other consistent differences in susceptibility between the isolates from the two Australian States.

Discussion

This study was the largest to date examining antimicrobial susceptibilities in Brachyspira isolates from chickens. Isolates tested were mainly from Australia, as currently there are few other isolates available worldwide. The agar dilution technique has been widely used for antimicrobial susceptibility testing of porcine Brachyspira isolates (Hommez et al., 1998), usually using doubling dilutions of antimicrobials throughout the MIC range to obtain accurate MIC values. Due to constraints on resources, in the current study only four dilutions of each antimicrobial were tested for each isolate—results were therefore expressed as MIC ranges. Despite these limitations of geographic origin of isolates and the use of broad MIC ranges, the current data set increase our understanding of what antimicrobial agents might prove useful to treat AIS caused by one or other of the main two pathogenic Brachyspira species infecting chickens.

MIC ranges obtained for most isolates of both Brachyspira species for tiamulin, lincomycin, metronidazole, and tetracycline suggested that these antimicrobials generally are likely to be effective for treatment. This conclusion is supported from the results of other studies where these drugs have been used to control AIS in naturally infected flocks (Smit et al., 1998; Stephens & Hampson, 1999), and for tiamulin and lincomycin in hens experimentally infected with B. intermedia and B. pilosicoli (Hampson et al., 2002; Stephens & Hampson, 2002). Whether or not these antimicrobials can be used to treat AIS in infected flocks, particularly in hens laying eggs intended for human consumption, will also depend on adherence to national veterinary medical regulations.

The two Brachyspira species had identical MIC₅₀ and MIC₉₀ ranges for metronidazole and tetracycline, but there were trends for some species-related differences in susceptibility to the other four antimicrobials. Reduced susceptibility to tiamulin occurred more commonly among the B. intermedia isolates than among the B. pilosicoli isolates, while the reverse was true for ampicillin. The MIC₅₀ figures also suggested a trend for reduced susceptibility to lincomycin and tylosin among B. pilosicoli isolates compared with B. intermedia isolates. These trends emphasize the need to identify the spirochaete species involved in individual cases of AIS, and wherever possible to conduct antimicrobial susceptibility testing before selecting an antimicrobial for treatment.

In some cases, one-dilution differences in susceptibility ranges were detected among different isolates from the same farm that grouped into a single ET, and hence were presumed to represent a single strain or a closely related clonal group of strains. These differences occurred for both B. pilosicoli and B. intermedia isolates, and may represent the beginning of a loss of susceptibility to particular agents by members of a given clonal group on a farm. In future work it would be useful to confirm these trends by monitoring susceptibility of strains over a longer period and testing with more antimicrobial dilutions across the appropriate concentration ranges.

Resistance to tylosin was relatively common, being recorded in seven of the 13 (54%) B. pilosicoli isolates from Queensland, all four non-Australian B. pilosicoli isolates, and in 11 of the 14 (79%) isolates of B. intermedia from Queensland flocks. In comparison, all 10 isolates of B. intermedia from WA flocks were susceptible to tylosin (MIC < 4 mg/l). Again, these differences reinforce the need to undertake MIC testing before selecting an antimicrobial to treat AIS. Tylosin is sometimes used to treat mycoplasma infections in poultry, and it is possible that such exposure in the Queensland and non-Australian flocks had encouraged the development of resistance. The generally low levels of susceptibility to tylosin among the chicken isolates, and an associated tendency for these also to have elevated MIC ranges to lincomycin, mirrors the situation seen among porcine isolates of B. hyodysenteriae and B. pilosicoli (Karlsson et al., 2002 2004). Currently it is not known whether this resistance in the chicken isolates was caused by the common point mutations in the 23S rRNA described as being responsible for macrolide and lincosamide resistance among porcine isolates of B. hyodysenteriae and B. pilosicoli (Karlsson et al., 1999 2004).

Two isolates of B. pilosicoli clearly showed resistance to ampicillin, while five other B. pilosicoli and one B. intermedia isolate had elevated MIC ranges. Previously, certain human and porcine isolates of B. pilosicoli have been shown to be resistant to amoxicillin (Brooke et al., 2003). In that study resistance was associated with beta-lactamase enzyme activity, and it was suggested that this could have been acquired by horizontal transfer of beta-lactamase genes from other Brachyspira strains or bacterial species. The genetic basis of the ampicillin resistance observed in the chicken isolates, particularly in B. pilosicoli, deserves further investigation.

Acknowledgments

This study was supported in part by a grant from the Australian Chicken Meat Council and the former Australian Egg Industry Council, administered through the Rural Industries Research and Development Corporation. The authors thank the Queensland Department of Primary Industries and Fisheries and Murdoch University for provision of facilities.