Antifungal Drug Susceptibility Testing of Yeast: A Primer for Beginners
Expires: April 25, 2025
Nancy Wengenack, Ph.D.
Professor of Laboratory Medicine and Pathology
Professor of Microbiology
Division of Microbiology
Mayo Clinic, Rochester, Minnesota
Hello, I am Dr. Nancy Wengenack, and I am the Director of the Mycology and Mycobacteriology Laboratory in the Division of Clinical Microbiology at the Mayo Clinic in Rochester, Minnesota.
Today’s “Hot Topic” presentation is for those who are considering performing susceptibility testing of yeast in their laboratory, perhaps for the first time, or for those who choose to send their yeast isolates out to a reference laboratory for susceptibility testing but who want to understand how the reference laboratory generates the data that comes back to them on the susceptibility reports.
I have no disclosures relevant to this "Hot Topic."
In today’s "Hot Topic," I will review the most common methods used by clinical laboratories for antifungal drug susceptibility testing of yeast isolates grown from microbiologic culture. In addition, I will discuss how breakpoints are established and highlight the antifungal agents for which standardized interpretive criteria and breakpoints are currently available. Finally, we will look at a value known as the epidemiological cutoff value, or ECV, and I will discuss how ECV values are established, which yeast they are available for at this time, and I will discuss how ECVs should be used.
So, when should I order susceptibility testing for yeast recovered in a culture? Antifungal drug susceptibility testing of yeast isolates is important to help guide clinicians in the choice of antifungals that may be useful in treating clinically significant disease. However, that doesn’t mean that susceptibility testing should be performed on every yeast isolate recovered from a culture in the clinical microbiology laboratory. Susceptibility testing should be performed if the isolate is deemed clinically significant by the physicians who are caring for the patient. Yeast culture isolates from normally sterile sources, such as blood or CSF, are almost always considered significant and susceptibility testing should be performed. However, not every yeast recovered from a non-sterile source such as sputum or a mouth swab or from the skin is clinically significant, and performing susceptibility testing on these isolates could be a waste of limited resources.
In addition, one should think about whether the susceptibility testing result will be actionable. As we will discuss in this "Hot Topic," the reference method and interpretive criteria are intended for the most commonly encountered yeast, namely certain Candida species and Cryptococcus species. Susceptibility testing of other genera or other species will result in reporting of minimal inhibitory concentrations, or MICs, without accompanying interpretive criteria such as “susceptible” or “resistant.” Lack of clear interpretive criteria makes use of the antifungal susceptibility results challenging for physicians. Your Infectious Diseases team or your antimicrobial stewardship group can help to establish some guidance for your laboratory on when to perform susceptibility testing of yeast.
Once the decision is made to perform antifungal susceptibility testing, there are at least two major international organizations that provide guidance to clinical laboratories on how to perform standardized antifungal drug susceptibility testing of yeast. Those two organizations are the European Committee on Antimicrobial Susceptibility Testing (or EUCAST) and the Clinical and Laboratory Standards Institute (also known as CLSI). Many of the recommendations of the two groups are harmonized, but some differences do exist. In this "Hot Topic," I will focus on the CLSI method and interpretive criteria, but I would encourage you to also visit the EUCAST website at www.eucast.org for more information on their standards.
Okay. So, let’s talk a little bit about the various methods that can be used for antifungal drug susceptibility testing of yeast. Antifungal drug susceptibility testing can be performed using either phenotypic or genotypic methods. Genotypic methods such as targeted PCR or DNA sequencing are used for some combinations of fungi and antifungal drugs, such as examination of fks, the glucan synthase gene, for hotspot mutations that cause echinocandin resistance in Candida species. However, genotypic methods for yeast susceptibility prediction are not standardized and are still largely relegated to research or public health laboratories.
Most clinical laboratories currently utilize phenotypic methods for antifungal susceptibility testing, and there are a number of different phenotypic methods which can be used. The accepted reference method is broth dilution testing, and this method is recognized as the reference method by both EUCAST and CLSI. Instead of the reference method, many hospital laboratories may choose to use a microbroth dilution method with a colorimetric endpoint such as that provided by the YeastOne Sensititre plates from ThermoFisher, or they may choose to use an automated microbroth dilution instrument like the Vitek from bioMérieux. Often these alternative methods are selected because the susceptibility panels are commercially available, making them convenient for busy laboratories who would prefer to purchase rather than make their own broth dilution panels. Finally, the disk diffusion method or gradient MIC strips like the E-test strips or gradient strips from Lilofilchem are also commercially available and are utilized by some laboratories.
If you intend to perform yeast susceptibility testing in your laboratory using the broth dilution method, there are two CLSI documents that provide very useful information. CLSI document M27 provides the methodologic framework for performing the reference broth dilution method. M27 covers selection and preparation of antifungal agents, the required QC procedures, and it details how to perform the test procedure itself. A companion document, M60, provides tables with the standardized breakpoints and interpretive criteria that are available for yeast. I believe that the M60 document will be renumbered in the near future and published as a supplement to the M27 document, so please be on the lookout for that document numbering change when it occurs. The breakpoints and interpretive criteria document are updated annually by CLSI as new data becomes available. The reference method document needs less frequent revision and is therefore updated every three years or so, depending on developments in the field.
The M27 document is intended to provide methodology for susceptibility testing of Candida species and Cryptococcus species, but it has not yet been vetted for use with other genera of yeast or with the yeast forms of the endemic dimorphic pathogens such as Histoplasma capsulatum or Blastomyces dermatitidis. This method is also not intended for use with filamentous molds such as Aspergillus species and there is a separate CLSI standard for mold susceptibility testing, which is CLSI document M38.
As I have already mentioned, the recognized reference method for antifungal susceptibility testing of yeast is broth dilution. The method was developed and agreed upon through an international consensus process that tapped the knowledge and experience of mycology experts in clinical microbiology hospital laboratories and government entities such as the CDC, the FDA, and the Canadian Public Health Services, as well as experts from the pharmaceutical industry and the device manufacturers. The use of a reference method greatly facilitates interlaboratory agreement when testing yeast for resistance to antifungal agents and enhances patient care by providing robust susceptibility results regardless of the geographic location of the patient. As mentioned, the reference method is broth dilution, but either macrobroth or microbroth dilution can be performed and equivalent results can be achieved regardless of which format is selected.
Many laboratories choose to perform microbroth dilution testing of yeast rather than macrobroth dilution testing because the microbroth dilution method can be performed in microtiter plates, which saves on reagent costs and on incubator space. In addition, as already mentioned, microbroth dilution plates are commercially available. In order to perform the microbroth dilution method, an isolate of the yeast to be tested should be selected from an agar plate containing a pure culture of the organism. Care must be taken to ensure that mixtures of yeast or yeast mixed with bacteria are not tested, as the MIC result obtained from a mixture may be inaccurate. A standardized suspension of the yeast is prepared in sterile saline or sterile water to a concentration equivalent to a 0.5 McFarland standard. From this, a working inoculum is prepared by diluting the yeast suspension into RPMI broth to a final standardized concentration. The inoculum is then added to each well of the microtiter plate most often using a repeating or multichannel pipettor. The microtiter plate already contains the antifungals to be tested, often provided in a lyophilized or frozen format, and the antifungals are distributed on the plate in rows with the concentration of the antifungal increasing by doubling dilutions in a gradient across the plate or sometimes down the column depending on the layout of the plate. After inoculation with the yeast isolate to be tested, the plate is sealed and incubated at 35°C for 24 to 48 hours when testing Candida species. Cryptococcus neoformans and Cryptococcus gattii generally grow more slowly, so incubation for up to 72 hours may be needed for Cryptococcus species.
Following incubation for the required time period, the panel can be read either manually using a light box or using a semiautomated plate reader. The endpoint is determined by comparison of the growth in each well compared with the control well. For the reference broth dilution method, the minimal inhibitory concentration, or MIC, is the lowest concentration that prevents growth of the yeast for the antifungal amphotericin B or it is the lowest drug concentration that produces a 50% reduction in growth as visualized by a prominent reduction in turbidity for the azoles, the echinocandins, and flucytosine.
The commercial microbroth dilution panels, such as the one shown on this slide, have a colorimetric indicator that assists with endpoint determination, and some people feel that these endpoints are easier to read because the color change corresponds to the growth reduction that is required for each drug class using the reference method. This eliminates the need for laboratory staff to determine the 50% growth reduction level, which can be tricky for the inexperience reader. The color blue indicates a lack of growth, and pink or purple indicates growth of the yeast isolate in that microtiter well. The well in position A1 in the upper-left corner is the positive growth control well, and that well is pinkish-purple in color, indicating sufficient growth of the control organism so the plate is ready to be read. The echinocandin drugs (anidulafungin, micafungin, and caspofungin) are shown in the top three rows, with the endpoint for each drug circled. For anidulafungin, the endpoint is 0.03 mcg/mL; for micafungin the endpoint is 0.008 mcg/mL; and for caspofungin the endpoint is read as 0.12 mcg/mL. For the echinocandins, there has been some literature suggesting that in vitro susceptibility testing of caspofungin may be technically challenging and that results do not always correlate with in vivo activity of the drug. Therefore, some experts recommend using anidulafungin or micafungin susceptibility testing results as surrogates for the echinocandins. This is still under study and it does not mean that caspofungin should not be used clinically if indicated; it simply means that in vitro laboratory testing of caspofungin may not be representative of its utility in vivo.
Looking at the other drugs on the plate, the endpoint for 5-flucytosine in row D is 0.06 mcg/mL; posaconazole in row E has an endpoint of 0.5 mcg/mL; voriconazole in row F has a MIC of 0.12 mcg/mL; itraconazole in row G has an MIC of 0.25 mcg/mL; and fluconazole in row H has an MIC of 8 mcg/mL. Amphotericin B is oriented from top to bottom of the plate in column 12 and its endpoint is 0.5 mcg/mL. The accepted reproducibility of the broth dilution method is plus or minus one 2-fold or doubling dilution. So, what does this mean in practical terms? If we use voriconazole and itraconazole as an example, the voriconazole MIC is read as 0.12 mcg/mL. If we retested it a second time, the MIC result for the second test might be 0.06, 0.12, or 0.25 mcg/mL, and these would essentially be the same as we obtained for the first test since the inherent variability of the test is plus or minus 1 doubling dilution. In addition, if we compare the voriconazole MIC and the itraconazole MIC, they are within plus or minus 1 doubling dilution from each other at 0.12 mcg/mL and 0.25 mcg/mL, so both drugs would be considered as having essentially an equivalent endpoint. Having a 1 doubling dilution lower MIC does not mean that voriconazole would be a better choice than itraconazole in this example. Other factors that we will discuss in addition to the MIC play an important role in the selection of an antifungal for clinical use. Finally, an important point to remember is that yeast must be fully identified to the species level in order to select the correct breakpoints and interpretive criteria. We will also explore this in more depth in the upcoming slides.
Species-specific breakpoints and interpretive criteria are not available at this time for all antifungals or for all Candida species. At the time of this recording, breakpoints are available for the echinocandins (namely caspofungin, micafungin, and anidulafungin) and two azoles (fluconazole and voriconazole). For these five drugs, species-specific breakpoints are available for six different Candida species, which are Candida albicans, Candida glabrata (also known by its new name of Nakaseomyces glabrata), Candida tropicalis, Candida krusei (which also has a new name as a Pichia species), Candida parapsilosis, and Candida guilliermondii. Species-specific breakpoints are not available for other Candida species or for other yeast, including the Cryptococcus species. For Candida species without established breakpoints and for Cryptococcus species, laboratories should report the MIC value only without an interpretation and can consider adding a reporting comment that indicates something such as, “no standardized interpretive criteria are available for this drug.” Interpretive criteria for additional genera and species of yeast and for additional antifungals will be added to the M60 document as they become available from the CLSI Antifungal Susceptibility Testing Subcommittee. This group meets regularly to review new data as it becomes available and to establish additional species-specific breakpoints for the M60 document.
This slide presents a portion of the interpretive criteria that are available from the M60 document. In this example, we are looking at the MIC breakpoints and interpretive criteria for the echinocandins anidulafungin and caspofungin. As you can see from this table, the MIC breakpoints differ depending upon which Candida species is tested. For example, a Candida albicans with an MIC less than or equal to 0.25 mcg/mL would be considered susceptible to anidulafungin, while a Candida glabrata with an MIC of 0.25 mcg/mL would be considered as having intermediate susceptibility to anidulafungin. The MIC breakpoints for resistance also differ with Candida albicans resistant to anidulafungin at an MIC greater than or equal to 1mcg/mL, while Candida glabrata is considered resistant if the MIC is greater than or equal to 0.5 mcg/mL. I should also mention that the name of Candida glabrata has been recently changed to Nakaseomyces glabrata but the CLSI documents have not yet been updated with this name change. The key takeaway to remember from this slide is that each Candida species has its own MIC breakpoints and interpretive criteria for each drug that is tested.
One might ask, is there any other information available to help guide my choice of antifungal agents if my patient has an infection caused by a Candida species other than those listed in M60 or if the isolate is Cryptococcus neoformans or Cryptococcus gattii, because there are no established breakpoints or interpretive criteria for these yeast? To answer this question, it is helpful to think about how breakpoints are established. Standardized breakpoints are set using a collection of data including the MIC value distributions for the species of interest against the antifungal agent as determined by the reference method, through evaluation of pharmacokinetic (PK) and pharmacodynamic (PD) data, through the use of clinical trial outcome data, and by using post-marketing antifungal susceptibility surveillance. For many fungal species and antifungal drug combinations, obtaining each of these pieces of data is not practical. It could, for example, require years or even decades to complete clinical trials that enroll a sufficient number of patients with a particular fungal infection. So, clinical outcome data obtained through well-controlled clinical trials is rarely available for many types of fungal infections. Therefore, unfortunately, breakpoints may never be able to be established for many fungal species and antifungal agent combinations.
But there is another piece of data that is beginning to become available for certain bacteria and fungi that can help to provide insight into whether a particular isolate has an MIC that is typically associated with wild-type isolates of this species or whether the isolate may have either acquired or intrinsic resistance to a particular antifungal agent which may make it less likely to have a favorable response. That piece of data is known as the epidemiological cutoff value, or ECV. Sometimes it can also be referred to as the ECOFF value.
Epidemiological cutoff values represent the MIC value that separates microbial populations into those with and without acquired or mutational resistance based on their phenotypes. The ECV defines the upper limit of the MIC range for the wild-type population of the microbe. ECVs are reported in mcg/mL and their interpretation is reported as either “wild-type” or “non-wild-type.” A wild-type ECV defines isolates with no mechanisms of acquired resistance or mutational resistance for the antifungal agent being tested. A non-wild-type ECV defines isolates that have a presumed or known mechanism of resistance to the antifungal agent.
There are two additional CLSI documents which provide very useful information about epidemiological cutoff values for antifungal susceptibility testing. CLSI document M57 provides the rationale and method for establishing ECVs. A companion document, M59, provides tables with the EVCs that have been established for both yeasts and molds. The M59 document may be renumbered in the future and published as a supplement to the M57 document, so please be on the lookout for that document numbering change should it occur. The M59 document containing the ECVs can be updated annually by CLSI as new data becomes available. The M57 method document needs revision less frequently and is updated every three years or so, depending on developments in the field.
So how are epidemiological cutoff values established? ECVs are established by reference standard setting organizations like EUCAST and CLSI, and they are based solely on in vitro laboratory MIC data collected using the reference broth dilution method. ECVs are established by pooling MIC data from at least three independent laboratories and that MIC data is generated by testing at least 100 distinct isolates of the yeast species of interest. No single laboratory should contribute more than 50% of the MIC data to avoid introducing any lab-specific bias to the ECV. Having more than three laboratories contributing MIC data from larger numbers of isolates adds additional confidence and robustness to the ECV value determination. The isolates tested need to be unequivocally identified to the species level, and this is most often done using molecular identification methods such as DNA sequencing rather than using phenotypic identifications like microscopic morphology because morphology can be misleading. Obtaining the needed MIC data for at least 100 isolates of the more common yeast species is reasonably easy to do but it can be quite challenging to obtain 100 well-characterized isolates of some of the less common species, and that is one of the reasons why ECVs are only available for certain genera and species of yeast at this time.
So how are the epidemiological cutoff values established? The MIC data from the 100-plus well-characterized isolates for the species of interest are plotted on a log-scale histogram like the one shown on the right side of this slide. Typically, the MIC data will apportion into a normal distribution as shown in the light blue bars on the left side of the plot. An iterative statistical method is then used to fit a normal distribution line to the MIC data, as shown in the slide and a mode for the normal or wild-type distribution of isolates is determined. In the example on this slide, the normal distribution ranges from about 0.5 mcg/mL up to 8 mcg/mL, with a mode of 2 mcg/mL. Typically the ECV is set as the MIC, which encompasses 97.5% of the normal or wild-type distribution of isolates. In the example on this slide, the ECV is statistically determined to be 8 mcg/mL. Any isolates with an MIC below the established ECV of 8 mcg/mL would be considered wild-type isolates, and any isolates with an MIC above the ECV of 8 mcg/mL, such as those with an MIC of 32 mcg/mL in the brown bar on this example, would be considered non-wild-type isolates. Where available, molecular studies have demonstrated that isolates with an MIC above the ECV often harbor genetic mutations that could cause them to be resistant to the drug tested.
Epidemiological cutoff values are based only on in vitro laboratory data. By themselves, ECVs cannot be used to predict clinical outcome or suggest whether a particular antifungal agent should be used. An ECV should not be thought of as equivalent to a breakpoint. For example, a Candida krusei isolate with a fluconazole MIC equal to 16 mcg/mL is considered a “wild-type” isolate using the ECV, but we know that wild-type isolates of Candida krusei are resistant to fluconazole and so this agent should not be used to treat Candida krusei infections even though it has a wild-type ECV.
So, what are ECVs useful for if they cannot be used as breakpoints? ECVs can be useful to assist the clinician with knowledge of whether the patient’s isolate has presumed or acquired mutations that might make it less likely to respond to an antifungal agent. A wild-type ECV doesn’t mean that the drug will or won’t work against the yeast. It simply means that the isolate has no known presumed or acquired resistance mechanisms. Another way to think about the wild-type ECV is that is does not provide any information about additional factors that are needed to predict the utility of the antifungal agent for the yeast under study. To determine whether the antifungal will be useful, one also needs to know pharmacokinetic factors, such as is the drug absorbed and distributed to the body compartment where the infection is located, or how quickly is it metabolized or excreted by the body? In addition, knowledge of pharmacodynamic factors, such as the desired and undesirable effects of the drug on the body or the drug-drug interactions that should be considered for each individual patient, also need to be considered. A non-wild-type ECV, however, may be useful in that it should give a clinician pause and may suggest that caution should be used when considering selection of this particular antifungal agent because the isolate has already demonstrated the presence of a resistance mechanism through an elevated MIC that is above that for the normal wild-type population of this species.
This slide presents a portion of the ECV data that is available from the CLSI for the antifungal agent amphotericin B against a number of Candida species. As you can see from this table, the ECVs differ depending upon which Candida species is tested. For example, a Candida albicans with an MIC less than or equal to 2 mcg/mL would be considered wild-type, while a Candida dubliniensis with an MIC above 0.5 mcg/mL would be considered as non-wild type, suggesting the presence of a resistance mechanism. At the time of this recording, species-specific ECVs are currently available for the most common Candida species as well as Cryptococcus neoformans, Cryptococcus deuterogatti, and Cryptococcus gattii. Additional ECVs are continuously being considered by standards-setting organizations.
To summarize today’s presentation, yeast susceptibility testing against antifungal agents is something that many clinical laboratories are able to perform routinely using a standardized broth dilution method. There are international consensus standards from the CLSI and EUCAST, which provide the methodology and recognized interpretive criteria. Species-specific interpretive criteria are available for the most common Candida species against the echinocandins and two azoles. In instances where breakpoints and interpretive criteria are not available, the use of epidemiological cutoff values, or ECVs, may assist the clinician to know whether the isolate is defined as “wild-type” or whether it has intrinsic or acquired resistance mechanisms that make the antifungal less likely to be effective. It is anticipated that additional breakpoints and interpretive criteria along with additional ECVs will become available over time as sufficient data is collected for other yeast genera and species across multiple laboratories.
If you are interested in learning more about antifungal susceptibility testing or epidemiological cutoff values, I have included a couple of excellent papers at the end of this presentation for your reference.
Contact us: email@example.com
Image credit: Shutterstock