|Home | About | Journals | Submit | Contact Us | Français|
An international multilaboratory collaborative study was conducted to develop standard media and consensus methods for the performance and quality control of antimicrobial susceptibility testing of Mycoplasma pneumoniae, Mycoplasma hominis, and Ureaplasma urealyticum using broth microdilution and agar dilution techniques. A reference strain from the American Type Culture Collection was designated for each species, which was to be used for quality control purposes. Repeat testing of replicate samples of each reference strain by participating laboratories utilizing both methods and different lots of media enabled a 3- to 4-dilution MIC range to be established for drugs in several different classes, including tetracyclines, macrolides, ketolides, lincosamides, and fluoroquinolones. This represents the first multilaboratory collaboration to standardize susceptibility testing methods and to designate quality control parameters to ensure accurate and reliable assay results for mycoplasmas and ureaplasmas that infect humans.
Methods for in vitro antimicrobial susceptibility testing of mycoplasmas were first described in the 1960s (6). Despite numerous publications during the ensuing years that have reported the activities of antimicrobial agents against these organisms, there have been no universally accepted or standardized broth dilution- or agar-based methods designating the optimum testing conditions, pH, media, length of incubation, quality control (QC) MIC reference ranges, or reference strains. The lack of a consensus method for MIC determination, coupled with complex cultivation requirements, has resulted in considerable confusion regarding the antimicrobial activities of various drugs against these fastidious organisms.
To address the need for a standard method for performing and validating in vitro susceptibility tests for human mycoplasmas and ureaplasmas, the Clinical and Laboratory Standards (CLSI) Subcommittee on Antimicrobial Susceptibility Testing of Human Mycoplasmas devised a series of studies involving a total of 10 laboratories from 3 different countries representing academia, industry, and government. Sequential evaluations of both broth- and agar-based methods were done with Mycoplasma hominis, Mycoplasma pneumoniae, and Ureaplasma species. The methods included commercial and individual laboratory-produced media and the testing of multiple reference strains for evaluation as QC strains. The QC strains and their respective MIC reference ranges were designated for several drug classes, including macrolides, ketolides, lincosamides, tetracyclines, and fluoroquinolones for each organism.
The CLSI mandates specific protocols and numbers of participating laboratories for determining MIC reference ranges for QC purposes and for the actual measurement of MICs. These requirements have evolved over the years and have now become quite stringent (2). However, the fastidious nature, complex media and incubation requirements, and relatively slow growth for some mycoplasmal species necessitated some modifications in the protocols that were used in the present studies. These modifications were performed with CLSI knowledge and approval. Even though a total of 10 laboratories contributed data to these investigations, some individual laboratories were unable to participate in every one of the agar and broth dilution studies for all 3 species.
The CLSI Human Mycoplasma Susceptibility Testing Subcommittee reviewed various protocols for broth microdilution and agar dilution and developed a consensus method for each technique, including rigorous QC parameters. The actual testing protocols that were used are described in the supplemental material. Once the actual methodology was approved, it was possible to plan studies to investigate the optimal media for use in the respective assays, to determine assay reproducibility within and among participating laboratories, and to designate reference QC strains with the tightest MIC ranges for the greatest number of drugs.
A preliminary study performed at the University of Alabama at Birmingham (UAB) tested 3 M. hominis reference strains against 8 drugs using a single lot of laboratory-prepared SP4 broth supplemented with arginine and a single batch of laboratory-prepared modified Hayflick's mycoplasma broth (MHMB) with arginine (see the supplemental material for medium formulations). No differences in the numbers of CFU recovered or the MICs were detected between these two media, so they were deemed equivalent. A second study performed in 6 laboratories tested the same 3 reference strains against 8 drugs using a single lot of commercially prepared SP4 glucose broth plus arginine (Remel, Lenexa, KS, available by special order) and 6 batches of MHMB with arginine prepared individually in each laboratory using the assay conditions described in the supplemental material. Due to problems in the detection of a sharp color change endpoint with commercial SP4 broth plus arginine in some participating laboratories, MHMB, with its sharper MIC endpoints, simpler composition, and lower cost than SP4, was chosen as the primary medium for performing broth microdilution assays for M. hominis. For M. pneumoniae, 5 laboratories tested 3 reference strains against 7 drugs using a single lot of commercial SP4 glucose broth (Remel) and 5 batches of MHMB with glucose prepared individually in each laboratory. No differences in the MIC color change endpoints were detected between these media for any of the antimicrobial agents. Testing to determine the MIC QC ranges in the present investigation was performed using Remel SP4 broth with glucose due to its commercial availability. For the Ureaplasma spp., 5 laboratories tested 3 reference strains against 7 drugs using a single lot of commercially prepared 10B broth (Remel) and 6 batches of 10B broth prepared individually in each laboratory. Due to problems with the detection of sharp MIC color change endpoints in the laboratory-prepared 10B broth, commercial 10B broth was chosen for Ureaplasma spp. However, since the formulation of the commercial 10B broth is essentially the same as the laboratory-prepared 10B broth, either may be used if definitive color change endpoints can be detected.
Agar dilution MIC assays must use noncommercially prepared agars to which the antimicrobials are added, because it is not possible to purchase such materials commercially. Furthermore, the generally accepted 72-h shelf life for media containing antibiotics requires that such media be made in an individual testing laboratory. Modified Hayflick's mycoplasma agar (MHMA) with arginine was chosen for testing M. hominis since the broth equivalent gave satisfactory performance in the preliminary evaluation for broth microdilution. For M. pneumoniae, MHMA with glucose was used for QC reproducibility testing, but SP4 glucose agar is also acceptable for MIC determination, since levels of growth of the organisms and the MICs are equivalent based on data described previously using the corresponding broths and direct comparisons performed at the University of Alabama at Birmingham (data not shown). A8 agar is the primary agar medium used for the cultivation of ureaplasmas, so it was used for the agar dilution MIC assays. The individual formulations for noncommercial media used in the broth and agar dilution assays are provided in the supplemental material.
Since the inoculum can influence MIC values, it is important to quantify accurately the numbers of organisms used for a broth- or agar-based MIC system. Due to their small cellular dimensions, mycoplasmas and ureaplasmas do not produce turbidity in liquid media, so use of the turbidity method for the determination of organism concentrations is not possible. In order to quantify these organisms for use in MIC determinations, it is first necessary to grow them in the appropriate broth medium until color change occurs due to pH changes brought about by the hydrolysis of urea (Ureaplasma spp.), arginine (M. hominis), or glucose (M. pneumoniae) using phenol red as a pH indicator. The culture is then frozen in multiple aliquots at −80°C overnight. Afterward, an aliquot is thawed and the CFU per ml determined by performing serial 10-fold dilutions in broth, plating each dilution on agar, incubating, and then counting colonies as described in the supplemental material. Those results can be used to determine the proper dilution that must be made in the appropriate broth to yield 104 to 105 CFU/ml for the MIC assay. The volume of inoculum required for MIC assays will depend on the number of drugs being tested in duplicate and the range of dilutions of each drug to be tested. The MIC inoculum is incubated at 37°C for 2 h to allow the organisms to become metabolically active. Ureaplasma spp. should be incubated for only 1 h prior to setting up the MIC assay due to their higher growth rate.
The reference standard powders of each drug were obtained from the manufacturers or from Sigma Chemical (St. Louis, MO) for nonproprietary compounds. The same lot number of each drug was used by all of the participating laboratories. The drugs tested included clindamycin, erythromycin, tetracycline, azithromycin (Pfizer), telithromycin (Sanofi-Aventis), levofloxacin (Ortho-McNeil), and moxifloxacin (Bayer). The powdered drugs were weighed and dissolved according to the manufacturer's instructions as described in the supplemental material, taking into account the purity of the drugs. The stock solutions for each drug were prepared on the days the MIC assays were performed in each participating laboratory in accordance with CLSI procedures (3, 4). The dilutions of the stock solutions for use in the individual MIC assays were also prepared in accordance with published CLSI procedures (5).
The organisms considered for designation as reference strains for QC purposes included type strains obtained from the American Type Culture Collection (ATCC) and clinical isolates derived from patient cultures in Birmingham, Alabama. Three different strains of each of the organisms underwent reproducibility testing for inclusion as the type strain for MIC determination for each respective species. For M. hominis we tested strains ATCC 43521 (M132), ATCC 23114 (PG21), and the clinical isolate MH 5155, for M. pneumoniae we tested strains ATCC 29342 (M129), ATCC 15531 (FH), and the clinical isolate MPN 834, and for Ureaplasma spp. we tested strains ATCC 27815 (U. parvum serovar 3), ATCC 27618 (U. urealyticum serovar 8), and ATCC 33175 (U. urealyticum serovar 9).
The broth microdilution MIC assay employed a 96-well microdilution plate into which a defined inoculum of the organism to be tested was added to wells containing doubling dilutions of the antimicrobial agents using a multichannel pipette. The microdilution plates were incubated in ambient air at 37°C until the positive growth control well changed color due to the phenol red pH indicator. Growth of the Ureaplasma spp. in 10B broth will cause a color change from yellow to pink and growth of M. hominis in MHMB or SP4 broth will cause a color change from pink to deep red, while the growth of M. pneumoniae will cause MHMB or SP4 broth to change from pink to yellow. For the Ureaplasma spp., this color change occurs in approximately 16 to 18 h of incubation, whereas M. hominis typically requires 48 to 72 h, and M. pneumoniae may require 4 to 6 days to show a color change in broth. The MIC endpoint was then determined as the lowest concentration of an antimicrobial that did not show any color change at the time the growth control showed a color change. It is particularly important to read the MIC endpoint at the first appearance of a color change in the growth control well due to the tendency of the MIC to shift after a longer incubation time, thus giving a falsely elevated MIC. Details of the broth microdilution procedure and its interpretation along with the following agar dilution procedure are described in the supplemental material.
The agar dilution method for the determination of MICs was based on the incorporation of doubling dilutions of the antimicrobial agents into molten agar plates, with each plate containing a different concentration. The antimicrobial dilutions were prepared for the entire concentration range desired for each drug using sterile ultrapure clinical laboratory reagent water (CLRW) as the diluent. There were typically 8 dilutions of each drug tested. The instructions for weighing out drug powders, accounting for purity, and preparing stock solutions are the same as for the broth microdilution. The appropriate dilutions of the drugs were dispensed in 2-ml volumes in 50-ml sterile polystyrene culture screw-cap tubes to facilitate mixing of the antibiotics with the molten agar. After the agar plates solidified, 10 μl of a defined organism inoculum of 104 to 105 CFU/ml prepared in the appropriate broth was added to the agar plates using a Steers replicator. The plates were incubated in ambient air plus 5% CO2 at 37°C. The MIC was read as the lowest concentration of the antimicrobial agent that prevented colony formation when examined under a stereomicroscope at the same time that the antimicrobial-free control plate demonstrated growth of approximately 30 to 300 CFU per spot of inoculum. The length of time until MICs could be read is similar to what is typically observed with the broth microdilution assay. Due to difficulties sometimes encountered in obtaining the correct inoculum, it may be useful to perform the MIC assay using an undilute, a 1:10 dilution, and a 1:100 dilution of the stock inoculum and use data from the dilution that provides the desired number of CFU on the antimicrobial-free control plate.
Six laboratories tested 10 replicates of the 3 reference strains of M. hominis against 8 drugs prepared from separate inocula using 3 separate batches of MHMB with arginine prepared in each laboratory. The tests were performed over a minimum of 3 days with a maximum of 4 replicates of each strain tested per day. For M. pneumoniae, the testing was performed similarly to the testing with M. hominis, except that the MICs were determined for 7 drugs and SP4 glucose broth was used in 7 participating laboratories. For the Ureaplasma spp., 5 laboratories tested 7 drugs using 10B broth.
Agar dilution testing was performed with M. hominis being tested against 8 drugs by 6 laboratories using MHMA with arginine. M. pneumoniae was tested by 8 laboratories using MHMA with glucose, and testing of the Ureaplasma spp. was performed by 6 laboratories using A8 agar.
All of the MIC data points obtained for the 10 replicates of each reference strain from each laboratory were reviewed in aggregate, so a total of 60 to 80 replicate MICs were reviewed for each organism/drug combination. The reference strain of each species that had the most MIC data points that occurred within a 3 to 4 2-fold-dilution range for the greatest number of drugs was selected as the designated QC strain. The QC ranges for individual drugs for the designated reference strains were calculated using the method established by the CLSI (2). This method includes tabulating the MIC mode for each antimicrobial/strain combination and expanding the range to include 1 log2 dilution on each side of the mode, giving a 3-dilution QC range. If there was a bimodal distribution of the MICs, then a 4-dilution range was selected. In both cases, the QC range must include ≥95% of the MIC replicate data points determined in aggregate for all the participating laboratories, excluding outliers as determined by the method of Turnidge and Bordash (7). No QC range was proposed for the drugs for which the MIC ranges exceeded 4 dilutions.
The M. hominis strain ATCC 23114 (PG21), M. pneumoniae strain ATCC 29342 (M129), and U. urealyticum strain ATCC 29342 (serovar 9) demonstrated the most reproducible MIC data points and the tightest ranges for the most drugs (data not shown) and were designated as the reference strains to be used for QC purposes for both agar and broth microdilution assays. Therefore, reproducibility data and MIC ranges derived from replicate testing are shown for these strains only.
Tables 1, ,2,2, and and33 show the combined results for the MIC data points obtained by replicate testing for the designated reference strains of M. hominis, M. pneumoniae, and U. urealyticum, respectively. Table 4 shows the MIC QC ranges for several drugs that were derived from these data. These QC ranges and reference strains were approved by the CLSI Subcommittee on Antimicrobial Susceptibility Testing and were recently published in a CLSI document (3). Overall, the reproducibility of the MIC determinations within individual laboratories and among the laboratories was best for the mycoplasmacidal fluoroquinolones and worst for the macrolides, telithromycin, and clindamycin. Despite the efforts of the participating laboratories to determine MIC ranges for the respective reference strains for all currently available antimicrobial agents relevant for testing against these organisms, there was lack of consensus for the assignment of ranges for some drugs by agar and/or broth methods due to excessive variation in the MICs obtained among the participating laboratories (Tables 1 to to3).3). It was not possible to designate a 3- to 4-dilution range for azithromycin for any organism by either MIC method. However, given the importance of azithromycin for the treatment of M. pneumoniae infections and the increasing prevalence of macrolide resistance in this organism, a single cutoff value of ≤0.06 μg/ml was chosen for the broth microdilution tested against M. pneumoniae strain ATCC 29342. This cutoff readily distinguishes macrolide-susceptible from macrolide-resistant strains for which azithromycin MICs may exceed 32 μg/ml (1). A similar decision was made for the tetracycline tested by agar dilution against U. urealyticum strain ATCC 33175. Even though no 4-dilution range could be established, a single cutoff MIC of ≥8 μg/ml was selected for this tetracycline-resistant strain, which is known to contain the tet(M) transposon that mediates tetracycline resistance in this organism (1). Tetracycline-susceptible strains usually have MICs of ≤2 μg/ml (1, 9).
Standardized antimicrobial susceptibility testing methods and designated QC parameters for human mycoplasmas and ureaplasmas are needed, because a culture is seldom performed for diagnostic purposes, and in vitro testing of individual isolates is even more rarely obtained. Antimicrobial susceptibilities can vary geographically and in response to selective antimicrobial pressure. Moreover, clinically significant acquired drug resistance, potentially affecting multiple antimicrobial classes, can occur in all of the mycoplasmal and ureaplasmal human pathogens (1). Since most mycoplasmal and ureaplasmal infections are treated empirically, obtaining accurate and reproducible antimicrobial resistance surveillance data for currently available drugs and new investigational agents is important. In the event of a systemic infection, particularly in an immunosuppressed host, an individual who has failed previous treatment, or someone who is known to harbor a resistant organism, susceptibility testing can be valuable for patient management (8, 9). As new antimicrobials are developed, it is also important to have knowledge of their antimycoplasmal activities relative to those of existing drugs.
The data described in this publication represent the first systematic multilaboratory project to standardize test methodology and demonstrate interlaboratory reproducibility for MIC determinations for human mycoplasmas and ureaplasmas, thus allowing the designation of specific type strains for M. pneumoniae, M. hominis, and U. urealyticum with defined QC reference MIC ranges for several antimicrobial agents by using a study design and methods approved by the CLSI. The fastidious nature of these organisms and the relative difficulty in obtaining an accurate and reproducible inoculum were probably at least partially responsible for our inability to get the sufficient consensus of results necessary to generate 3- to 4-dilution QC ranges for some drugs. Fluoroquinolones gave tighter QC ranges overall than did the other drug classes, perhaps because they are bactericidal against mycoplasmas, while the other drug classes are bacteriostatic and are known to have shifting MICs, especially for broth microdilution (8).
Standard methods, media formulations, and QC parameters have been described for both broth microdilution and agar dilution methods. The relative advantages and disadvantages of each technique for testing human mycoplasmas and ureaplasmas have been discussed elsewhere (8), and either method can be performed depending on the preferences and needs of individual laboratories.
Due to the inherent differences in their cultivation requirements and growth rates, no single procedure, pH, or medium was considered sufficient for testing all of the clinically important species, and the differences in the proposed methods and test conditions reflect this observation. For example, the slow growth of M. pneumoniae mandates that several days of incubation must take place before the MICs can be determined. Additionally, the low medium pH of 6.0 required to detect the growth of ureaplasmas can affect MIC results since the activity of some drugs, such as macrolides, are highly pH dependent (8). It is also important to note that there has been no attempt to generalize these methods for application to other mycoplasmal species of human or animal origin, which may have very different growth and testing requirements. Therefore, these procedures and reference ranges should be limited to testing only the species for which they are described.
Individual laboratories can prepare their own SP4 glucose and 10B media for broth microdilution MIC determinations for M. pneumoniae and Ureaplasma species, respectively, using the formulations provided in the supplemental material as long as the QC strains yield MICs with a sufficiently sharp endpoint to allow visual interpretation, and the values fall within the designated reference ranges. Alternatively, laboratories can purchase these media from commercial sources, although it may require a special order. Reproducibility testing and the determination of QC ranges for reference strains by agar dilution for M. pneumoniae were performed using MHMA with glucose. However, it is also acceptable to perform these assays using SP4 glucose agar prepared as described in the supplemental material, as long as the QC reference strain MICs fall within the designated ranges.
The performance of the studies described here took place in accordance with procedures approved by the CLSI, so it is anticipated that these procedures and MIC ranges for the reference strains will be widely adopted for future work. The use of these CLSI-approved methods for broth microdilution MIC determination has recently been proven accurate for the detection of resistance to macrolides and fluoroquinolones in organisms in which resistance markers have been genotypically characterized (10, 11). Detailed protocols for the performance of antimicrobial susceptibilities, QC, and interpretive MIC breakpoints for several drugs are included in Methods for Antimicrobial Susceptibility Testing for Human Mycoplasmas. Approved Guideline M43-A, published by the CLSI (3).
This work was performed by the Clinical and Laboratory Standards (CLSI) Subcommittee on Antimicrobial Susceptibility Testing of Human Mycoplasmas. Financial support was provided by Abbott Laboratories, AstraZeneca Pharmaceuticals, Bayer Pharmaceuticals, Bristol-Myers-Squibb, Ortho-McNeil Pharmaceuticals, Pharmacia & Upjohn, Pfizer Animal Health, and Sanofi-Aventis Pharmaceuticals. Remel Laboratories provided commercially prepared SP4 and 10B broth media.
Technical and administrative support from Tracy Dooley and the CLSI, Donna Crabb, Danuta Kovach, Dena Hensey-Rudloff, Hélène Renaudin, Jeanette Jones, W. Lanier Thacker, Bill Kabat, and Laura Houston is gratefully acknowledged.
Published ahead of print 22 August 2012
Supplemental material for this article may be found at http://jcm.asm.org/.