Minimal inhibitory concentration of seven antimicrobials to Mycoplasma gallisepticum and Mycoplasma synoviae isolates from six European countries

ABSTRACT

Mycoplasma gallisepticum and Mycoplasma synoviae are bacterial pathogens that cause disease in poultry, adversely affecting their health and welfare, and are a financial burden on producers. This manuscript describes the results of the MycoPath project that is the first international antimicrobial susceptibility programme for mycoplasma pathogens isolated from poultry. Improved comparative analysis of minimal inhibitory concentration (MIC) results from participating countries was facilitated by using one laboratory determining all MICs. Chicken and turkey isolates were obtained from France, Germany, Great Britain, Hungary, Italy and Spain during 2014–2016. One isolate per farm was retained. The MIC of seven antimicrobial agents was determined using a broth microdilution method, with Friis Medium (M. gallisepticum) or Modified Chanock’s Medium (M. synoviae). Of the 222 isolates recovered, 82 were M. gallisepticum and 130 were M. synoviae. M. gallisepticum MIC_50/90 values were 0.12/0.5, 2/8, 0.5/4, 0.12/>64, 0.008/0.062, 0.008/32, 0.062/4 mg/l for doxycycline, enrofloxacin, oxytetracycline, spiramycin, tiamulin, tilmicosin and tylosin, respectively. For M. synoviae, the values were 0.5/1, 8/16, 0.5/1, 0.5/8, 0.25/0.5, 0.062/2 and 0.062/16 mg/l respectively. A bimodal MIC distribution for the fluoroquinolone (enrofloxacin) and the macrolides (spiramycin, tilmicosin and tylosin) indicate that both species have sub-populations that are less susceptible in vitro to those antimicrobials. Some differences in susceptibilities were observed according to host species, Mycoplasma species, and country of origin. This study provides a baseline of novel data for future monitoring of antimicrobial resistance in poultry Mycoplasma species. Additionally, this information will facilitate the selection of the antimicrobial agents most likely to be effective, thus ensuring their minimal use with targeted and correct therapeutic treatments.

Highlights

• First large-scale pan-European collection of representative Mg and Ms isolates.

• MIC values assessed in central laboratory for Mg and Ms from chickens and turkeys.

• Range of MIC values for 82 Mg and 130 Ms isolates to seven licenced antibiotics shown.

• Data can be used to help determine Mg and Ms veterinary-specific breakpoints.

KEYWORDS:

• Antimicrobial resistance
• MIC
• chicken
• turkey
• broth microdilution
• Mycoplasma gallisepticum
• Mycoplasma synoviae

Introduction

Mycoplasma gallisepticum and Mycoplasma synoviae are bacterial pathogens that cause disease in poultry (Ferguson-Noel, 2013). Their potential adverse health and economic impact on poultry production is so significant that these two Mycoplasma species are listed by the World Organization for Animal Health (OIE, 2018). M. gallisepticum is also included in the European Council Directive (European Council Directive 2009/158/EC) that facilitates trade between European countries. In 2018 the European poultry industry supplied 15,776,000 tons of poultry meat (AVEC, 2019) and 6,755,000 tons of eggs (European Commission, 2019).

M. gallisepticum causes chronic respiratory disease of domestic poultry, especially in the presence of management stresses and/or other respiratory pathogens. Disease is characterized by lachrymation, conjunctivitis, sneezing, cough, and sinusitis, particularly in turkeys and game birds. Airsacculitis and pneumonia are considered the gross lesions related to M. gallisepticum; M. gallisepticum infection can result in loss of production and increased carcass condemnation in meat poultry, and a loss of egg production in layers. Transmission of M. gallisepticum infection occurs either vertically (in ovo) from an infected breeder flock to the progeny, or horizontally by direct or indirect contact of susceptible birds with infected carriers or contaminated debris (Levisohn & Kleven, 2000). M. synoviae causes respiratory disease, synovitis, or may result in a silent infection. M. gallisepticum and M. synoviae strains vary in infectivity and virulence, and infections may sometimes be unapparent (Jordan, 1975; Kleven, 1998; OIE, 2018). In recent years, eggshell apex abnormalities have also been linked to M. synoviae infections (Catania et al., 2010; Feberwee et al., 2009). Catania et al. (2016) demonstrated a significant difference in daily egg mean weight and the number of eggs in M. synoviae experimentally infected birds.

The poultry industry has a number of approaches to maintaining healthy flocks (Mehdi et al., 2018). Disease freedom and biosecurity methods are considered as the major disease control approaches (Ferguson-Noel et al., 2020; Levisohn & Kleven, 2000). Maintaining flocks free of pathogenic mycoplasmas, means that replacement stocks need to be obtained from mycoplasma-free sources and those birds are then raised in a single-age all-in all-out farm management system. Good biosecurity and an effective monitoring system are necessary aspects of this approach (Kleven, 2008); however, disease control interventions are used as needed. Live vaccines are now commercially available for both M. gallisepticum and M. synoviae, but some studies report that they may not prevent infection (Feberwee et al., 2006). Even though the use of live vaccines is considered a good option in high prevalence areas, potential complications could arise in interpretation of seroconversion of the birds and ELISA positivity of the one-day-old pullets (Moronato et al., 2018). This increases the complexity in interpreting laboratory results of the breeder flocks, and there is also a potential risk of the vaccine reverting to a virulent form (Armour & Ferguson-Noel, 2015). Therefore, some producers prefer to rely on biosecurity and, as a consequence, antimicrobial agents are often needed for treatment and control of infections, in particular for M. synoviae. It is mostly smallholder poultry producers and “hobby/backyard” poultry keepers that rely on antimicrobial agents to treat mycoplasma infections, while the major producers will sacrifice/cull all flocks to maintain M. gallisepticum-free status so that the associated trades can be protected. Infections with M. synoviae have possibly been perceived as less important; however, some European countries have recognized the increased virulence of M. synoviae and are aiming to eliminate the infection (Landman, 2014; Michiels et al., 2016).

In 2010 the Centre Européen d`Etudes pour la Santé Animale (CEESA) introduced a MycoPath programme, which assesses the antimicrobial susceptibility of four different veterinary Mycoplasma species isolated from cattle, pigs and poultry (de Jong et al., 2013; Klein et al., 2017). The aim of the poultry programme was to create a pan-European collection of representative M. gallisepticum and M. synoviae isolates from clinical cases of diseased chickens and turkeys. The samples were only collected from poultry that had not recently been treated with antimicrobials, reducing the risk that any recent treatment would not have residual antimicrobials, or temporary genetic changes impacting on the minimal inhibitory concentration (MIC) levels obtained in this study. It is important to know if these pathogens are developing antimicrobial resistance, so that only effective antimicrobial agents are used for therapy, thus ensuring minimal use of antimicrobials by using targeted and correct treatments. The antimicrobials selected for this study are the most relevant antimicrobials which are licenced for commercial poultry use in Europe. In the AVEC annual report representing the European poultry meat sector, it is stated that they are “committed to minimising the use of antibiotics in poultry production” but “zero use is neither ethical nor sustainable and poultry farmers and veterinarians need to have antibiotics to maintain the health and welfare of birds” (AVEC, 2019). Therefore, it is essential that the effectiveness of antimicrobial agents is monitored and maintained. However, recent antimicrobial susceptibility data for M. gallisepticum and M. synoviae isolates is very limited and national resistance monitoring surveys, such as GERM-Vet (2018) or Resapath (Anses, 2017), do not include Mycoplasma species isolated from poultry.

This study provides new MIC data for M. gallisepticum and M. synoviae isolates recovered from six different European countries against seven licenced antimicrobial agents. The isolates were obtained from poultry that had not been treated with antimicrobials in the previous 15 days. This procedure ensures that the antimicrobial susceptibility data are representative for diseased birds without a history of antimicrobial therapy, and prevents the collection of isolates biased in favour of resistance. The testing was carried out at a central laboratory which is novel in providing standardized MIC tests which facilitated direct comparison of the isolates from different countries. The tests used culture media suitable for optimal growth of these two Mycoplasma species. The broth microdilution method used in this study essentially followed the guidelines of Hannan (2000) and CLSI (2011).

Materials and methods

Collection of Mycoplasma gallisepticum and Mycoplasma synoviae isolates

Mycoplasma isolates were collected from France, Germany, Great Britain, Hungary, Italy and Spain. Samples were collected based on specific criteria and were sent to national laboratories for culture, isolation and identification of M. gallisepticum and M. synoviae. The samples were collected during 2014–2016 and came from field cases reporting respiratory disease. Only one isolate per farm per clinical episode in a three-month period was allowed, which minimizes the risk of epidemiologically related strains. Isolates had to be from geographically spread areas within each country and from chickens or turkeys that had no antimicrobial treatment in the previous 15 days. In an attempt to achieve identical numbers of isolates from the participating countries, a fixed target number of 24 M. gallisepticum and 24 M. synoviae isolates was indicated for each country. The participating laboratories followed their standard Mycoplasma culture isolation, filter cloning and molecular identification procedures for obtaining pure cultures (Catania et al., 2014; Cisneros-Tamayo et al., 2020; Kreizinger et al., 2017; Mattison et al., 1995). The purity of all cultures was confirmed in the participating laboratories using their own methods, which would either have been by using fluorescent antibody tests on agar grown cultures (Bradbury, 1998), or by other sensitive methods that can detect other contaminating bacterial species, which included MALDI-TOF mass spectrometry (Spergser et al., 2019), or universal PCR and sequencing (Lauerman et al., 1995), or universal PCR followed by denaturing gradient gel electrophoresis fingerprinting (McAuliffe et al., 2003). Isolates were stored at temperatures below −50°C, before transfer to the central laboratory (Don Whitley Scientific, Bingley, UK) on dry ice, or at ambient temperature as lyophilized cultures, together with a case report form for each isolate. Culture of the isolates at the central laboratory was examined for typical M. gallisepticum and M. synoviae growth characteristics and additional identity checks were performed by the central laboratory on a random selection of 12 isolates (5.4%). The 12 isolates originated from Germany (1), Great Britain (3), Hungary (4), Italy (2), and Spain (2). These selected M. gallisepticum and M. synoviae isolates were re-identified using a duplex PCR method, with primer pairs for the detection of each of these species (Buim et al., 2009). The identity of all 12 isolates was confirmed.

Antimicrobial testing

Antimicrobial susceptibility testing for all M. gallisepticum and M. synoviae isolates was carried out at the central laboratory. The isolates were checked for viability using the culture medium to provide optimum growth for each species with M. gallisepticum in Friis Medium (Friis, 1975) and M. synoviae in Modified Chanock’s Medium (Bradbury, 1998), both with phenol red as an indicator and without the addition of antimicrobial agents. M. synoviae requires nicotinamide adenine dinucleotide provided in the Modified Chanock’s Medium which can be inhibitory to other Mycoplasma species, hence the different medium for each Mycoplasma species. Each isolate was incubated in broth medium until a distinctive colour change was produced, then divided into aliquots and frozen at −70°C ± 10°C. The viable count in one aliquot was determined by serial dilution and plating onto agar medium. During subsequent MIC tests, aliquots were thawed and diluted to a cell density of 10⁶ colony forming units (CFU) per ml, to produce a final inoculum density of nominally 5 × 10⁵CFU/ml in the MIC plates. M. gallisepticum NCTC 10115 (ATCC 19610) and M. synoviae NCTC 10124 (ATCC 25204) were used as quality control strains to monitor the performance of the MIC test.

MIC determinations were performed using a broth microdilution method. Seven antimicrobial agents, all EU-approved for medication in poultry, from four different antimicrobial classes (fluoroquinolones, macrolides, pleuromutilins, and tetracyclines) were tested. For each antimicrobial agent, a stock solution containing 1280 mg/l of the active ingredient was prepared using the appropriate solvents and diluents as specified in CLSI (2018) and dilutions were made in Friis Medium (M. gallisepticum) or Modified Chanock’s Medium (M. synoviae) to give a final test concentration range from 0.001-64 mg/l.

To determine the MIC for each isolate, 100μl of the appropriate antimicrobial solution was distributed into the conical wells of polystyrene microtitre plates, before 100μl of culture (thawed, pre-incubated for 1h and then diluted as described above) was added to each well to give a final cell concentration of approximately 5 × 10⁵CFU/ml. For each strain, a positive (growth) control well contained no antimicrobial with 100μl of sterile medium in its place and a single well with 200μl of sterile medium served as a negative uninoculated control. Immediately after inoculation, microtitre plates fitted with polystyrene lids were placed in a humidified atmosphere and incubated at 35°C ± 1°C. Plates were examined every 24 h. If no growth was evident in the positive control wells, plates were re-incubated for up to 5 days. For each isolate, MIC results were read as soon as adequate growth (unambiguous colour change) was visible in the positive control wells. All MIC plates were read against a white background to facilitate identification of colour changes in the medium from red (no growth) to orange/yellow (growth). The MIC of each antimicrobial was recorded as the lowest concentration that completely inhibited growth. For the test to be considered valid, it was necessary for a definite colour change to be visible in the positive control well and for the negative control well to remain unchanged. The reproducibility of the test was demonstrated by ensuring that the MIC results of the quality control strains of this study fell within ± one doubling dilution around a central value. In cases where the MIC results obtained for an antimicrobial agent against one or more strains deviated markedly from the MICs obtained against the majority of strains, the MIC test was repeated on two further occasions. In such cases, the reported MIC value was obtained on at least two separate occasions.

The MIC ranges, MIC distributions, MIC₅₀ and MIC₉₀ values were determined for each antimicrobial and Mycoplasma species, and for each country. The MIC₅₀ and MIC₉₀ values are percentiles calculated from the complete set of MIC results for a given substance against a specified group of mycoplasma isolates. MIC₅₀ is the lowest concentration of an antimicrobial agent at which growth is inhibited for 50% of tested strains. MIC₉₀ is the lowest concentration of an antimicrobial agent at which growth is inhibited for 90% of tested strains.

Results

Collection of isolates

The number of isolates collected, from either chickens or turkeys, varied between the participating countries for both M. gallisepticum and M. synoviae (Table 1). For M. gallisepticum 16 isolates (19.5%) were recovered from samples taken in 2014; 20 isolates (24.4%) from 2015, and 44 isolates (53.7%) from 2016. For M. synoviae, these figures were 5 (3.8%), 47 (36.2%) and 78 (60.0%), respectively. Demographic data were available for 90.1% of the isolates recovered. The vast majority of the samples in each country were from geographically spread farms from untreated flocks with mycoplasma-like clinical signs. The sample origin for chickens was 42.9% from layers, 15.0% from broilers and 42.1% from breeders; for turkeys these figures were 89.6% from fattening turkeys and 10.4% from breeders. The ages of the chicken layers and turkey fatteners ranged from 98 to 686 days and 77 to 140 days old, respectively. The size of the chicken flocks was 500–120,000 for layers, 2,500–40,000 for broilers, and 19–27,000 for breeders; for turkeys, the size of fattening turkey flocks varied from 500-42,000, and for breeder flocks from 18,000-36,000.

Table 1. Summary of Mycoplasma gallisepticum and Mycoplasma synoviae isolates from chickens and turkeys from each participating country during 2014–2016*.

Country	Mycoplasma gallisepticum			Mycoplasma synoviae
	Chicken	Turkey	Total	Chicken	Turkey	Total
France	3	0	3	13	0	13
Germany**	0	5	5	4	6	10
Great Britain	20	2	22	7	0	7
Hungary***	8	0	8	20	21	51
Italy	20	10	30	34	5	39
Spain	14	0	14	20	0	20
Total	65	17	82	96	34	130

*Two M. gallisepticum isolates were recovered from samples collected in 2013.

**Four M. synoviae isolates were recovered from joint samples collected from turkeys.

***Four isolates were recovered from samples collected in the Czech Republic, three isolates were from Austrian samples and three isolates were from Romanian samples.

Minimal inhibitory concentration results

The distribution of chicken and turkey MIC results is detailed separately for the M. gallisepticum and M. synoviae isolates in Table 2, providing an overall comparison of the MIC distribution for both of these avian Mycoplasma species against the seven antimicrobials. Table 3 summarizes the same results as MIC₅₀, MIC₉₀ and MIC range separately for chicken and turkey isolates of M. gallisepticum and M. synoviae as well as a combined MIC value for chicken and turkey isolates. The MIC ranges of the M. gallisepticum (n = 4–6) and M. synoviae (n = 5–10) control strains are also presented in Table 3. As different and low numbers of isolates were collected from each country, direct comparison between countries is difficult, but three countries (Great Britain, Italy and Spain) had more than 10 M. gallisepticum isolates from chickens, so their MIC values are detailed and compared in Table 4. Similar comparisons are made for M. synoviae from chickens for France, Hungary, Italy and Spain (Table 5).

Table 2. Minimal inhibitory concentrations (mg/l) of seven antimicrobials for 82 Mycoplasma gallisepticum (Mg) and 130 M. synoviae (Ms) isolates from chickens and turkeys obtained from European countries.

Antimicrobial	Species	<0.001	0.001	0.002	0.004	0.008	0.016	0.031	0.062	0.12	0.25	0.5	1	2	4	8	16	32	64	>64
Enrofloxacin	Mg							3	19	7	5		3	6	19	18	2
	Ms							1	2	2	9	18	20	5	4	26	41	2
Doxycycline	Mg						1	1	21	30	12	13	4
	Ms								11	17	34	40	21	7
Oxytetracycline	Mg								4	16	15	16	10	5	10	4	1	1
	Ms							3	6	9	29	42	33	5	2			1
Spiramycin	Mg						1	9	20	16	2	1	2		2	2	3		5	19
	Ms									3	28	42	27	13	2	3	6	3	1	2
Tylosin	Mg				2	2	15	19	9	3	2	2	1	8	11	8
	Ms					13	11	20	21	13	11	12		4	1	7	7	6	2	2
Tilmicosin	Mg	7		10	16	13	2	3	1			1				3	4	15	7
	Ms				4	6	10	25	33	13	8	6	9	9	2	5
Tiamulin	Mg			2	13	29	8	12	11	5	1	1
	Ms				1		1		5	41	43	31	7	1

Table 3. MIC₅₀ and MIC₉₀ (mg/l) for chicken and turkey M. gallisepticum and M. synoviae isolates obtained from European countries. The table includes the MIC range for the NCTC control strains.

Antimicrobial	Host species	Mycoplasma gallisepticum 65 chicken and 17 turkey isolates			Mycoplasma synoviae 96 chicken and 34 turkey isolates
		MIC50	MIC90	Range	MIC50	MIC90	Range
Enrofloxacin	Chickens Turkeys All Control	2 4 2 N/A	8 8 8 N/A	0.031–16 0.062–8 0.031–16 0.016–0.031 (0.01)	8 1 8 N/A	16 8 16 N/A	0.062–32 0.031–16 0.031–32 0.12–0.5 (0.5)
Doxycycline	Chickens Turkeys All Control	0.12 0.12 0.12 N/A	0.5 0.5 0.5 N/A	0.16–1 0.062–1 0.16–1 0.062–0.25	0.5 0.25 0.5 N/A	1 1 1 N/A	0.062–2 0.062–1 0.062–2 0.062–0.25
Oxytetracycline	Chickens Turkeys All Control	0.25 1 0.5 N/A	4 4 4 N/A	0.062–32 0.25–4 0.062–32 0.12–0.25 (0.1)	0.5 0.25 0.5 N/A	1 1 1 N/A	0.031–32 0.031–4 0.031–32 0.12–0.25 (0.1)
Spiramycin	Chickens Turkeys All Control	0.12 16 0.12 N/A	>64 >64 >64 N/A	0.031–>64 0.016–>64 0.016–>64 0.062–0.125	0.5 0.5 0.5 N/A	16 2 8 N/A	0.12–>64 0.25–32 0.12–>64 0.25–1
Tylosin	Chickens Turkeys All Control	0.031 2 0.062 N/A	4 8 4 N/A	0.004–8 0.008–8 0.004–8 0.031–0.031 (0.01)	0.25 0.062 0.062 N/A	32 0.12 16 N/A	0.008–>64 0.008–16 0.008–>64 0.031–0.12 (0.025)
Tilmicosin	Chickens Turkeys All Control	0.008 32 0.008 N/A	32 64 32 N/A	<0.001–64 0.002–64 <0.001–64 0.004–0.008	0.062 0.062 0.062 N/A	2 0.12 2 N/A	0.004–8 0.008–2 0.004–8 0.016–0.12
Tiamulin	Chickens Turkeys All Control	0.008 0.031 0.008 N/A	0.062 0.062 0.062 N/A	0.002–0.5 0.002–0.12 0.002–0.5 0.008–0.016 (0.0025)	0.25 0.12 0.25 N/A	0.5 0.5 0.5 N/A	0.004–2 0.016–1 0.004–2 0.062–0.5 (0.1)

Note: N/A = not applicable.

Data from Hannan (2000) for control strain in brackets.

Table 4. Comparison of MIC_50, MIC₉₀ and MIC range (mg/l) for Mycoplasma gallisepticum chicken isolates obtained from Great Britain (20 isolates), Italy (20 isolates) and Spain (14 isolates).

Antimicrobial	Country	MIC50	MIC90	Range
Enrofloxacin	Great Britain Italy Spain	0.062 4 4	0.12 8 16	0.031–0.12 0.062–8 0.062–16
Doxycycline	Great Britain Italy Spain	0.12 0.12 0.5	0.12 0.25 1	0.062–0.25 0.016–0.5 0.12–1
Oxytetracycline	Great Britain Italy Spain	0.25 0.5 4	0.5 4 8	0.062–0.5 0.062–8 0.12–8
Spiramycin	Great Britain Italy Spain	0.062 0.12 >64	0.12 64 >64	0.031–0.12 0.031–>64 0.062–>64
Tylosin	Great Britain Italy Spain	0.031 0.12 4	0.062 8 8	0.004–0.062 0.031–8 0.016–8
Tilmicosin	Great Britain Italy Spain	0.004 0.008 32	0.008 64 32	<0.001–0.008 <0.001–64 0.002–32
Tiamulin	Great Britain Italy Spain	0.008 0.008 0.031	0.016 0.031 0.12	0.004–0.031 0.002–0.062 0.004–0.12

Table 5. Comparison of MIC_50, MIC₉₀ and MIC range (mg/l) for Mycoplasma synoviae chicken isolates obtained from France (13 isolates), Hungary (20 isolates), Italy (34 isolates) and Spain (20 isolates).

Antimicrobial	Country	MIC50	MIC90	Range
Enrofloxacin	France Hungary Italy Spain	16 1 8 16	16 8 16 16	0.062–16 0.062–16 0.12–32 8–32
Doxycycline	France Hungary Italy Spain	0.5 0.12 0.5 1	1 0.5 1 2	0.12–2 0.062–1 0.12–1 0.25–2
Oxytetracycline	France Hungary Italy Spain	1 0.25 0.5 0.5	1 1 0.5 2	0.25–32 0.031–1 0.12–2 0.062–4
Spiramycin	France Hungary Italy Spain	0.5 0.5 0.5 1	8 2 16 16	0.12–64 0.25–4 0.12–>64 0.12–4
Tylosin	France Hungary Italy Spain	0.12 0.016 0.5 2	0.5 0.25 32 64	0.031–>64 0.008–2 0.008–64 0.062–>64
Tilmicosin	France Hungary Italy Spain	0.031 0.031 0.25 0.25	1 0.25 2 8	0.016–8 0.004–2 0.008–8 0.031–1
Tiamulin	France Hungary Italy Spain	0.5 0.12 0.25 0.25	1 0.5 0.5 1	0.12–2 0.062–0.5 0.12–1 0.12–1

Both M. gallisepticum and M. synoviae showed a bimodal distribution for the fluoroquinolone enrofloxacin, although the two main MIC peaks were slightly less for M. synoviae (Table 2). For the two tetracyclines, doxycyline and oxytetracycline, MIC values showed a monomodal distribution, although for oxytetracycline the M. gallisepticum isolates tended towards a bimodal distribution. In contrast, a bimodal distribution was observed for the three macrolides, spiramycin, tilmicosin and tylosin. For all macrolides the MIC values exhibited a broad range. The distribution of the MIC values for tiamulin showed two peaks for M. gallisepticum, while predominantly an even distribution was observed for M. synoviae.

With the exception of oxytetracycline at 0.5 mg/l and enrofloxacin at 2 mg/l, the MIC₅₀ values obtained for M. gallisepticum were low, at 0.12 mg/l or less (Table 3). However, the MIC₉₀ values were considerably higher, at 4, 4, 8, 32 and >64 mg/l for oxytetracycline, tylosin, enrofloxacin, tilmicosin and spiramycin, respectively. The bimodal populations and the high MIC values obtained with some isolates could suggest that M. gallisepticum has developed antimicrobial resistance against these antimicrobials. Doxycycline, enrofloxacin, oxytetracycline and tiamulin exhibited only slightly higher MIC₉₀ as compared to MIC_50.

Comparison of the MIC values of M. gallisepticum chicken isolates from Great Britain (20 isolates), Italy (20 isolates) and Spain (14 isolates), resulted in potential trends (Table 4). Spain had markedly higher MIC_50/90 values for all seven antimicrobial agents; all MIC₉₀ values for Italy were higher than those for Great Britain. Compared with Great Britain, Spain and Italy consistently had a bimodal population to all of the tested antimicrobial agents, with the second phase having higher MIC values. For M. synoviae, comparisons were made between France (13 isolates), Hungary (20 isolates), Italy (34 isolates) and Spain (20 isolates); differences among countries were less apparent (Table 5), but Hungary usually had lower MIC_50/90 values and lower maximum MIC values than the three other countries. M. synoviae isolates from Great Britain were not included for comparison because of the low number (n = 7). The reason for these between-country differences in MIC values is not known, but one could speculate about the different levels of antimicrobial agent used for the treatment of poultry, or it may just be differences in the management of the poultry farms from which the isolates originated.

The MIC comparison of turkey and chicken isolates is likely to be influenced by the low and different number of isolates from the participating countries (Tables 4 and 5). Using isolates from just one country, Italy could provide an acceptable comparison. Italian isolates included 20 M. gallisepticum from chickens and 10 from turkeys, with 34 M. synoviae isolates from chickens and five from turkeys (Table 6). From the Italian samples, some trends emerged. The M. synoviae isolates from both chickens and turkeys had higher MIC values for tiamulin than the M. gallisepticum isolates. For tilmicosin, although there was a bimodal population for all isolates and more isolates from chickens than turkeys, only the isolates from chickens had MIC values above 16 mg/l. A higher number of M. gallisepticum turkey isolates (six out of 10) had MIC values for spiramycin at 64 mg/l and above compared to four out of 20 from chickens. The M. gallisepticum turkey isolates had no oxytetracycline MICs below 0.5 mg/l, whereas values for nine of the 20 chicken isolates were below 0.5 mg/l. The M. synoviae isolates had slightly higher enrofloxacin MICs (maximum 32 mg/l) than the M. gallisepticum isolates (maximum 8 mg/l).

Table 6. Minimal inhibitory concentration distribution (mg/l) for Mycoplasma gallisepticum (Mg) and M. synoviae (Ms) isolates obtained from chickens and turkeys from Italy.

Antimicrobial	Species (n)	<0.001	0.001	0.002	0.004	0.008	0.016	0.031	0.062	0.12	0.25	0.5	1	2	4	8	16	32	64	>64
Enrofloxacin	Ms Turkey (5)										1			1	1	1	1
	Ms Chicken (34)									1		6		2	2	9	13	1
	Mg Turkey (10)										1				4	5
	Mg Chicken (20)								3	1	1			2	7	6
Doxycycline	Ms Turkey (5)										2	1	2
	Ms Chicken (34)									1	8	16	9
	Mg Turkey (10)								1	7	2
	Mg Chicken (20)						1		6	7	4	2
Oxytetracycline	Ms Turkey (5)											3	1		1
	Ms Chicken (34)									1	3	14	15	1
	Mg Turkey (10)											5	2	2	1
	Mg Chicken (20)								2	5	2	2	5	1	2	1
Spiramycin	Ms Turkey (5)										3						1	1
	Ms Chicken (34)									1	13	10	2	1	1		3	1		2
	Mg Turkey (10)									3	1									6
	Mg Chicken (20)							2	7	4					1	1	1		3	1
Tylosin	Ms Turkey (5)								1	1						1	2
	Ms Chicken (34)					1		3	2	3	3	9		1		4	3	4	1
	Mg Turkey (10)						3		1					2	2	2
	Mg Chicken (20)				1	1	3	6	1	1			1	1	2	3
Tilmicosin	Ms Turkey (5)						1			1		1		2
	Ms Chicken (34)				3	1												3	3
	Mg Turkey (10)						3		1					2	2	2
	Mg Chicken (20)	1		4	4	2	2									2	1	1	3
Tiamulin	Ms Turkey (5)									2	1	1	1
	Ms Chicken (34)									10	13	9	2
	Mg Turkey (10)				1	3	1	2	2	1
	Mg Chicken (20)			1	5	7	2	4	1

Discussion

Whilst it is useful to compare the data obtained from this study with those published elsewhere, the available literature is focused mainly on one Mycoplasma species (M. synoviae), and is limited to restricted geographical areas. In addition there are no recent European data for M. gallisepticum. This study is unique in many aspects as it is the first study that has the MIC values for the most important Mycoplasma species that affect poultry collected in major European poultry production countries. The isolates were collected by the European national laboratories during a set timeframe during 2014–2016; the isolates had not been exposed to antimicrobials within the previous 15 days; the MIC testing of the seven licenced antimicrobials was all carried out at a central independent laboratory. Therefore, this study permits some comparisons between the participating countries and these two Mycoplasma species. Other studies which are included in this discussion may have some limitations, such as just testing isolates from one country, testing a specific class of antimicrobials, or having long incubation times, potentially assessing mycoplasmacidal effect rather than the MIC. These organisms are not included in any of the European national, ongoing programmes, such as GERM-Vet (Germany), RESAPATH (France) or UK-VARSS (Great Britain). Nhung et al. (2017) stated it is necessary to increase efforts to harmonize testing practices, to promote free access to data on antimicrobial resistance, which will improve treatment guidelines, and monitor the evolution of antimicrobial resistance in poultry bacterial pathogens including the avian Mycoplasma species. However, all published information is useful in assessing any changes in the in vitro effectiveness of antimicrobials against these important avian disease-causing Mycoplasma species. It also strengthens the need for the introduction of standardized MIC testing and the development of breakpoints.

Although all participating countries were requested to collect equal numbers of M. gallisepticum and M. synoviae isolates, the number of isolates collected by each country varied. This is unlikely to be related to the size of the avian mycoplasma infections in the countries, but may be due to differences in practitioners requesting mycoplasma molecular tests in preference to culture and different time constraints of the practitioners. In addition, participating countries may use slightly different procedures for mycoplasma culture, isolation and identification which may have influenced the isolation rate.

At the outset it is important to emphasize that all of the data have great merit; however, care needs to be taken in the interpretation of the results. The numbers of isolates per Mycoplasma species per country were small and were distributed between chickens and turkeys. Similarly, numbers between chickens and turkeys were, in a few instances, too low to draw definitive conclusions. Note that several factors may affect the rate of resistance, including age of the birds, production system, different antimicrobial agent usage, source (diagnostic laboratory versus abattoir) or disease. Nevertheless, some differences were observed and are worthy of including in the discussion.

Although only 17 M. gallisepticum isolates were from turkeys, some differences between isolates from chickens and turkeys were observed; mainly the MIC₅₀ values of spiramycin and tilmicosin were higher in turkeys at 16 and 32 mg/l compared with 0.12 and 0.008 mg/l for the 65 chicken isolates (Table 3). From these results one may speculate that farmers could use more macrolides in turkeys probably because the life cycle is longer in this species; or that the hosts exert different antimicrobial resistance selection pressures on the Mycoplasma species. It may also be due to isolates being obtained from different countries. When similar comparisons are made for M. synoviae, with 34 of those isolates from turkeys and 96 from chickens, the enrofloxacin and tylosin MIC₅₀ values are higher for chickens at 8 and 0.25 mg/l, respectively, compared with 1 and 0.062 mg/l for turkeys. The MIC₉₀ values of macrolides are also higher for M. synoviae chicken isolates: spiramycin, tylosin and tilmicosin at 16, 32 and 2 mg/l, respectively, compared with 2, 0.12 and 0.12 mg/l for turkey isolates. This is in contrast to the data observed for M. gallisepticum. For M. gallisepticum it could be speculated that this is due to possible different treatment approaches. Usually, turkeys are kept for meat, so they are kept for longer than chickens and no vaccines are available; antimicrobial treatment can be repeated during the production cycle. The difference in MIC results obtained for M. synoviae may be a result of treatment usually being applied in the layer sector to avoid egg production losses, or to reduce the presence of eggshell apex abnormalities (Catania et al., 2010; Feberwee et al., 2009). In broiler breeder production this approach may be justified to help contain the infection and reduce the risk of spreading infection in the broilers by vertical transmission. Early detection of infection and adequate knowledge of antimicrobial effectiveness, such as MIC data, can reduce the amount of antimicrobial agents used and still improve and increase broiler production (Fincato et al., 2019).

Whilst the comparisons made between chicken and turkey isolates is informative in observing trends and potential risks of the avian mycoplasmas developing antimicrobial resistance, care is needed to not over-interpret these in vitro tests in relation to the in vivo situation. Although all the testing was carried out in a central laboratory, essentially following the same procedures, comparisons between MIC values for M. gallisepticum and M. synoviae isolates may be affected by the use of the different growth media, which are needed to provide the optimal growth conditions required for these two different Mycoplasma species. In the Hannan (2000) recommendations for MIC testing against veterinary Mycoplasma species, the same control strains were used for M. gallisepticum and M. synoviae against enrofloxacin, oxytetracycline, tiamulin and tylosin; a comparison with this study’s controls is included in Table 3.

During the 1980s and 1990s several European studies describing the in vitro susceptibility of avian mycoplasmas were published, but recent European reports on MICs are non-existent for M. gallisepticum, and scarce for M. synoviae (Catania et al., 2019; Dufour-Gesbert et al., 2006; Kreizinger et al., 2017; Landman et al., 2008). Dufour-Gesbert et al. (2006) carried out MIC determinations on 36 M. synoviae isolates originating from the respiratory tract of French layers without pathological lesions and obtained between 2002 and 2003. They tested six of the same antimicrobial agents used in this study, not tilmicosin, and all of the MIC values were ≤1 mg/l, which is lower than some of the MIC values obtained for French isolates in this study. Landman et al. (2008) tested 17 M. synoviae Dutch isolates obtained between 2000 and 2004; 14 were from the respiratory tract and three from joints, two of which were from turkeys. For enrofloxacin, difloxacin, doxycycline, tylosin and tilmicosin they recorded final MIC values at 7 and 14 days, whereas it is recommended to determine the real MIC value when growth in the control without antimicrobial is observed. With the exception of two isolates where MIC values were 2 or 4 mg/l for enrofloxacin, all other results were below 1 mg/l at 7 days. The 14 d results were higher, but that may be a minimal mycoplasmacidal concentration, rather than a MIC and could be due to antimicrobial activity decreasing during the longer incubation time. Both of these studies are now over 15 years old, although potentially valuable in observing any changes in antimicrobial susceptibility over time (Dufour-Gesbert et al., 2006; Landman et al., 2008); however, the isolates were from a small area, or single country, with a limited number of isolates tested (36 and 17, respectively).

In a more recent study, Kreizinger et al. (2017) tested 41 M. synoviae isolates from both chicken and turkey tracheas, mainly isolated between 2014 and 2016. Most isolates (26) were obtained from Hungary, the others came from Austria, Czech Republic, Slovenia, Ukraine, Russia and Serbia. These isolates were tested against a range of antimicrobial agents using a broth microdilution method. They reported some differences in MIC values between chicken and turkey isolates, with chickens having higher enrofloxacin MIC values. Overall, similar MIC₅₀ values for the same antimicrobial agents tested in this study were reported as well as elevated MIC₉₀ values for most of the antimicrobial agents tested here except for tylosin, which had a MIC₉₀ value of ≤0.25 mg/l compared with 16 mg/l in this study. A direct comparison between Kreizinger et al. (2017) with our study cannot be made as different growth media were used and their isolates were not collected exclusively from birds not recently treated with antimicrobials. They also used unofficial breakpoints to interpret their MIC data. In a recent study, Catania et al. (2019) investigated the antimicrobial susceptibility of 154 M. synoviae isolates from broiler chickens, layers and turkeys obtained between 2012 and 2017 in Italy. They tested them against seven of the same antimicrobial agents used in this study but with different mycoplasma culture media, and it is not stated if the isolates were from non-treated birds. The MIC ranges and MIC₅₀ results were similar to those obtained in this study, but this current study had higher MIC₉₀ values for spiramycin and tylosin both at 32 mg/l compared with 4 and 1 mg/l respectively, but lower for tilmicosin at 2 mg/l compared with >32 mg/l. Future MIC studies are needed to understand the observed differences.

Outside Europe, as early as 1994, Lin et al. (1994) reported high MIC₅₀ values of oxytetracycline and spiramycin at >32 mg/l from Taiwanese M. gallisepticum isolates. Gharaibeh & Al-Rashdan (2011) showed increased MIC₅₀ values for Jordanian M. gallisepticum isolates from 2007–2008 compared with isolates from 2004–2005. In comparison to 2004–2005 and 2007–2008, the values for tilmicosin, tylosin, enrofloxacin, doxycycline and oxytetracycline increased from ≤0.031 to 2, ≤0.031 to 0.125, ≤0.031 to 2, ≤0.031 to 0.062, and 0.062 to 2 mg/l, respectively. It should be noted that the antimicrobial usage patterns in Jordan might be much different from the European countries. Gautier-Bouchardon (2018) reviewed the MIC values obtained for M. gallisepticum but comparative data for “old strains” and “new strains” is limited: the main finding is an increase in M. gallisepticum maximum MIC value of enrofloxacin from 1 to 10 mg/l and oxytetracycline from 0.5 to 4 mg/l.

A major concern is that the development of antimicrobial resistance may lead to antimicrobial treatment against avian mycoplasma infections being ineffective. Several authors have reported that in vitro passaging of isolates in antimicrobial agents at sub-lethal levels rapidly induces resistance in as little as 10 passages. Zanella et al. (1998) demonstrated this with M. gallisepticum exposed to spiramycin, tylosin and enrofloxacin; Takahashi et al. (2002) to tylosin; and Wu et al. (2005) to tylosin and tilmicosin. Similarly, Gautier-Bouchardon et al. (2002) showed rapid development of antimicrobial resistance by both M. gallisepticum and M. synoviae to enrofloxacin and tylosin. Interpretation of these higher antimicrobial susceptibility levels as antimicrobial resistance raises some issues as no defined epidemiological cut-off values or clinical breakpoints are set for these avian Mycoplasma species. However, Gerchman et al. (2011) linked the decreased macrolide susceptibility to mutations in the 23S rRNA gene in their study where 50% of the 50 M. gallisepticum isolates from Israel had resistance to tylosin and tilmicosin. Lysnyansky et al. (2015) also described mutations in the 23S rRNA gene of M. synoviae associated with high MIC values. Le Carrou et al. (2006) and Lysnyansky et al. (2013) associated high enrofloxacin MIC values in M. synoviae with amino acid substitutions in the parC gene. Data from another Mycoplasma species, Mycoplasma bovis, has also demonstrated that high MIC values are linked with genetic mutations associated with antimicrobial resistance (reviewed in Lysnyansky & Ayling, 2016). Therefore it is likely that the high MIC values reported by previous workers are indicative of genuine antimicrobial resistance.

The in vivo effectiveness of antimicrobials has been the subject of many studies (Garmyn et al., 2019; Jordan & Horrocks 1996; Kempf et al., 1997; Riazuddin et al., 2017). However, Hinz & Rottmann (1990) investigated the re-isolation of M. gallisepticum following treatment with enrofloxacin, tylosin and tiamulin and they stated that tylosin proved to be inadequate, whereas enrofloxacin was highly effective. Despite this, Reinhardt et al. (2005) showed persistence of M. gallisepticum in chickens after treatment with enrofloxacin without development of resistance. Few cases document the use of antimicrobial agents to eradicate avian Mycoplasma infections from a flock; however, Hong et al. (2015) described the eradication of M. synoviae from a multi-age broiler-breeder farm using intensive antimicrobial treatment which consisted of the continuous administration of tilmicosin, after two rounds of treatment with chlortetracycline, doxycycline and enrofloxacin.

The focus of this study has been on M. gallisepticum and M. synoviae, arguably the most pathogenic avian Mycoplasma species, certainly those that potentially have the most economic impact on poultry production. However, many different Mycoplasma species have been isolated from poultry and most of those are thought to be commensal or possibly opportunist pathogens. Therefore, treatment of poultry with antimicrobial agents may inadvertently lead to antimicrobial resistance in these non-pathogenic (Beylefeld et al., 2018) or potentially opportunist Mycoplasma species, as well as in other commensal or zoonotic bacteria such as Salmonella and Campylobacter.

The criteria for relating in vitro MIC data to clinical breakpoints for Mycoplasma species from poultry, including M. gallisepticum and M. synoviae, have not been determined by CLSI or EUCAST, although guidelines have been published (Hannan, 2000). However, some standards have been established for three Mycoplasma species that cause significant clinical infections in humans (Mycoplasma pneumoniae, Mycoplasma hominis, and Ureaplasma urealyticum) (CLSI, 2011; Waites et al., 2012). The growth requirements of those Mycoplasma species differ significantly from M. gallisepticum and M. synoviae, thus the growth media and control strains that are suitable for testing these human-derived Mycoplasma species cannot be applied to this study. To try to address these shortcomings, a EUCAST Veterinary Subcommittee on Antimicrobial Susceptibility Testing (VetCAST) has recently been established. One of the remits of VetCAST is to initiate and coordinate EU research aimed at filling the current gaps on veterinary-specific breakpoints including epidemiological cut-off values (Toutain et al., 2017). The MIC distributions presented in this study can support the development of tentative epidemiological cut-off values. Until pharmacokinetic and clinical data become available, these epidemiological cut-off values can be used for monitoring the development of antimicrobial resistance and help determine clinical breakpoints.

In conclusion, the use of a central laboratory to determine MICs against M. gallisepticum and M. synoviae in chickens and turkeys from several European countries gave useful comparative MIC data. This has allowed the assessment of in vitro susceptibility of antimicrobial agents in treating the economically important mycoplasmas of poultry. It also provides a baseline for future comparisons to assess the development of antimicrobial resistance. An awareness of current MIC values facilitates the initial selection of the most optimal antimicrobial treatment of these infections. This study demonstrates that some isolates of both M. gallisepticum and M. synoviae have high MIC values indicating that antimicrobial resistance is a risk, and further studies are required to determine their efficacy in vivo. Selection and use of antimicrobial agents to effectively treat avian mycoplasmoses require knowledge of the organisms’ antimicrobial resistance status. Indeed, the lack of clinical breakpoints emphasizes the need for establishing avian mycoplasma-specific clinical breakpoints to ensure a correct interpretation of the susceptibility results. The availability of interpretive criteria will assist veterinarians in minimizing antimicrobial usage and to promote targeted treatment options that will avoid development of more resistant strains.

Acknowledgements

The MycoPath project is a pan-European programme dedicated to the collection and to the monitoring of antimicrobial susceptibility of veterinary mycoplasmas from diseased food-producing birds. MycoPath is an initiative of, and is coordinated by the Executive Animal Health Study Center (CEESA). CEESA’s membership is composed of international pharmaceutical companies researching and producing veterinary medicinal products. This study was funded by Bayer Animal Health GmbH (Germany), Boehringer Ingelheim (Germany), CEVA Santé Animale (France), Elanco Animal Health (UK), Merial (France), MSD Animal Health Innovation (Germany), Vetoquinol S.A. (France), Virbac (France) and Zoetis (Belgium). The authors thank the national co-ordinators and the national microbiological laboratories involved in the sampling and isolation procedures.

Disclosure statement

Some authors are connected with pharmaceutical companies; however, the testing, interpretation of results and preparation of the manuscript have been carried out independently.

Additional information

Funding

This study was funded by Bayer Animal Health GmbH (Germany), Boehringer Ingelheim (Germany), CEVA Santé Animale (France), Elanco Animal Health (UK), Merial (France), MSD Animal Health Innovation (Germany), Vetoquinol S.A. (France), Virbac (France) and Zoetis (Belgium) [CEESA].