Susceptibility Test Methods

The main objective of susceptibility testing is to predict the outcome of treatment with the

antimicrobial agents tested. Results are generally reported as categories of susceptibility.

The implication of the result category “susceptible” is that there is a high probability that the

patient will respond to treatment with the appropriate dosage regimen for that antimicrobial

agent. The result “resistant” implies that treatment with the antimicrobial agent is likely to

fail. One group has coined the term “90–60 rule”; that is, for many infections we can expect

treatment success about 90% of the time when the organism tests as “susceptible” to that

treatment, and success may still occur in around 60% of cases when the organism tests as

“resistant” to the agent used (41). The apparent 60% response rate to ineffective

antimicrobials is said to reflect the natural response to many bacterial infections in an

immunologically normal host (36).

Most test methods also include an “intermediate” category of susceptibility, which can have

several meanings. With agents that can be safely administered at higher doses (e.g.,

penicillins and cephalosporins), this category may imply that higher doses may be required

to ensure efficacy or that the agent may prove efficacious if it is normally concentrated in an

infected body fluid, e.g., urine. Conversely, for body compartments where drug penetration

is restricted even in the presence of inflammation (e.g., cerebrospinal fluid), it suggests that

extreme caution should be taken in the use of the agent. It may also represent a buffer zone

that prevents truly resistant strains from being incorrectly categorized as susceptible, and

vice versa.

A further aim of susceptibility testing is to guide the clinician in the selection of the most

appropriate agent for a particular clinical problem. In most clinical settings, susceptibility test

results are usually obtained 24 to 48 h or more after the patient has been given empirical

treatment. The test results may confirm the susceptibility of the organism to the drug

initially prescribed or may indicate resistance, in which case alternative therapy will likely be

required. The report describing the susceptibility testing results should provide the clinician

with alternative agents to which the organism is susceptible. These alternatives also may be

useful if the patient subsequently develops an adverse reaction to the initial antimicrobial

agent. There is a growing emphasis from professional societies and managed care

organizations to use susceptibility test results to direct therapy toward the most narrow

spectrum, least expensive agent to which the pathogen should respond. This is particularly

true for hospitalized patients, in whom the rate of antimicrobial resistance tends to be

higher, and it is easier to make therapeutic changes for inpatients than for out patients. This

makes the accuracy of susceptibility testing even more critical for effective patient care.

The clinical microbiology laboratory should perform susceptibility testing only for pathogens

for which well standardized methods are available and for pathogens whose resistance is

known or suspected to be a clinical problem; susceptibility testing should not be performed

on normal biota or colonizing organisms. Currently, routine susceptibility testing methods are

best standardized for the common aerobic and facultative bacteria and systemic antibacterial

agents. For some uncommon or highly fastidious bacteria and for most topical antibacterial

agents, simple routine test methods have not been standardized. Taking into account this

limitation, the Clinical and Laboratory Standards Institute (CLSI; formerly the NCCLS) has

released recommendations on how some of these bacteria may be tested and the results

interpreted (11). With some pathogens (e.g., Mycobacterium tuberculosis and invasive

fungi), routine testing is important for patient management, but testing is best performed by

specialized laboratories in which test volumes are sufficient to maintain technical proficiency

and where unusual or inaccurate results are likely to be recognized. Susceptibility testing

methods for certain other pathogens (e.g., mycoplasmas, chlamydiae, legionellae,

spirochetes, viruses, protozoa, and helminths) may not be well established at present and/or

are limited to a few specialty laboratories. A number of choices exist in antibacterial

susceptibility testing with respect to methodology and selection of agents for routine testing.

SELECTING AN ANTIMICROBIAL SUSCEPTIBILITY

TESTING METHOD Back to top

Clinical microbiology laboratories can choose from among several conventional or novel

methods of routine antibacterial susceptibility testing. These include the broth microdilution,

disk diffusion, antimicrobial gradient, and automated-instrument methods. In recent years,

there has been a trend toward the use of commercial broth microdilution and automatedinstrument

methods instead of the disk diffusion procedure. However, there remains ongoing

interest in the disk diffusion test because of its inherent flexibility in drug selection, its ability

to respond quickly to changes in interpretive breakpoints or when new agents are

introduced, and its low cost. The availability of numerous antibacterial agents and the

diversity of antimicrobial agent formularies of different institutions have made it difficult for

manufacturers of commercial test systems to provide standard test panels that fit everyone’s

needs. Thus, the inherent flexibility in drug selection that is provided by the disk diffusion

test is an undeniable asset of the method. The test is also one of the most established and

best proven of all susceptibility tests and continues to be updated and refined through

frequent (usually annual) CLSI publications (13, 14). Instrumentation is now available for

reading and interpreting zone diameters as well as for storing this information and may

reduce interobserver reading errors (31–33).

Advantages of the microdilution and agar gradient diffusion methods include the generation

of a quantitative result (i.e., an MIC) rather than a category result and the ability to test

accurately some anaerobic or fastidious species that may not be tested by the disk diffusion

method (7, 9, 11, 24, 27); ancillary benefits are computer systems that accompany many of

the microdilution and automated systems (24). Indeed, computerized data management

systems are very important in laboratories that may have limited or inflexible laboratory

information systems. However, an MIC method should not be chosen on the basis that MICs

are routinely more useful to physicians. There is no clear evidence that MICs are more

relevant than susceptibility category results to the selection of appropriate antibacterial

therapy for most infections (17).

A laboratory may choose to perform rapid, automated antibacterial susceptibility testing in

order to generate results faster than manual methods can generate them. The provision of

susceptibility results 1 day sooner than that provided by conventional methods seems a

logical advance in patient care. Three studies have demonstrated both the clinical and

economic benefits derived from the use of rapid susceptibility testing and reporting

(2, 21, 30), while a further study has not shown such a benefit (4). However, rapid

susceptibility testing results may not have substantial impact unless the laboratory uses

more aggressive means of communication to make physicians aware of the results (51). A

previously cited shortcoming of rapid susceptibility testing methods was the failure to detect

some inducible or subtle resistance mechanisms (23,29, 49, 50). However, the instruments

most notorious for such problems are no longer marketed, and the manufacturers of the

remaining instruments have made substantial efforts to correct earlier problems

(34, 40,44, 54) or to extend testing to include fastidious organisms (26). It is important to

emphasize that accuracy should not be sacrificed in an effort to generate a rapid

susceptibility testing result.

SELECTING ANTIBACTERIAL AGENTS FOR ROUTINE

TESTING Back to top

The laboratory has the responsibility to test antimicrobial agents and report on those that are

most appropriate for the organism isolated, the site of infection, and the clinical practice

setting in which the laboratory functions. The battery of antimicrobial agents routinely tested

and reported on by the laboratory will depend on the characteristics of the patients under

care in the institution and the likelihood of encountering highly resistant organisms (25). A

laboratory serving a tertiary-care medical center, which specializes in the care of

immunosuppressed patients, may need to test routinely agents that are broader in spectrum

than those tested by a laboratory that supports a primary-care outpatient practice in which

antibiotic-resistant organisms are less commonly encountered (46).

When a laboratory’s routine susceptibility testing batteries are determined, several principles

should be followed. First, the antimicrobial agents that are included in the institution’s

formulary and that physicians prescribe on a daily basis should be tested. Second, the

species tested strongly influences the choice of antimicrobial agents for testing. The CLSI

publishes tables that list the antimicrobial agents appropriate for testing against various

groups of aerobic and fastidious bacteria (14). The guidelines indicate the drugs that are

most appropriate for testing against each organism group and for treatment based upon the

specimen source (e.g., cerebrospinal fluid, blood, urine, or feces). The lists also include a few

agents that may be tested as surrogates for other agents because of the greater ability of a

particular agent to detect resistance to closely related drugs (e.g., the use of the cefoxitin

disk test to predict overall β-lactam resistance in staphylococci) (14). This initial list of

agents must be tailored to an individual institution’s specific needs through discussions with

infectious disease physicians, pharmacists, and committees concerned with infection control

and the institutional formulary (25).

A third important step in defining routine testing batteries is ascertaining the availability of

specific antimicrobial agents for testing by the laboratory’s routine testing methodology.

Certain methods (e.g., the disk diffusion, gradient diffusion, and in-house-prepared broth

and agar dilution methods) allow the greatest flexibility in the selection of test batteries. In

contrast, some commercial systems may have less flexibility or may experience delays in

adding the latest antimicrobial agents approved for clinical use. However, practicality limits

the maximum number of drugs that can be tested simultaneously with an isolate by any

susceptibility testing method. For example, a maximum of 12 disks can be placed on a 150-

mm-diameter Mueller-Hinton agar plate, and a similar number can ordinarily be

accommodated in a microdilution panel if full concentration ranges of each agent are to be

included for routine determination of MICs. Some commercial test panels attempt to resolve

this problem by testing a larger array of antimicrobial agents, although in a very limited

concentration range (perhaps 2 to 4 dilutions for each agent).

ESTABLISHING SUSCEPTIBILITY BREAKPOINTS Back to top

There is general agreement that the MIC is the most basic laboratory measurement of the

activity of an antimicrobial agent against an organism. It is defined as the lowest

concentration that will inhibit the growth of a test organism over a defined interval related to

the organism’s growth rate, most commonly 18 to 24 h. The MIC is the fundamental

measurement that forms the basis of most susceptibility testing methods and against which

the levels of drug achieved in human body fluids may be compared to determine breakpoints

for defining susceptibility.

The conventional technique for measuring the MIC involves exposing the test organism to a

series of twofold dilutions of the antimicrobial agent in a suitable culture system, e.g., broth

or agar for bacteria. The twofold-dilution scheme was originally used because of the

convenience of preparing dilutions from a single starting concentration in broth and agar

dilution methods. Subsequently, this system proved to be meaningful because an antibiotic’s

MICs for a single bacterial species in the absence of resistance mechanisms have a

statistically normal distribution when plotted on a logarithmic scale. This provides

investigators with the opportunity to examine the distributions of MICs for bacterial

populations and distinguish strains for which the MICS are abnormally high (potentially

resistant strains) from those for which MICs are normal (susceptible strains)

(28;http://www.srga.org/eucastwt/WT_EUCAST.htm).

MIC measurements are influenced in vitro by a number of factors, including the composition

of the medium, the size of the inoculum, the duration of incubation, and the presence of

resistant subpopulations of the organism. The in vitro test conditions also do not encompass

other factors that can have an influence on in vivo antimicrobial activity. These include sub-

MIC effects, postantibiotic effects, protein binding, effects on organism virulence or toxin

production, variations in redox potential at sites of infection, and the pharmacokinetic

changes resulting from different drug levels in blood and at the site of infection over time.

Nevertheless, if determined under standardized conditions, MIC measurements provide a

fixed reference point for the setting of pharmacodynamic breakpoints with the power to

predict efficacy in vivo. Pharmacological breakpoints can be applied directly to routine

dilution testing methods that generate MICs, such as broth microdilution, agar dilution,

gradient methods, and some automated instruments. They also provide reference values for

deriving breakpoints for disk diffusion methods.

Breakpoints (or interpretative criteria) are the values that determine the categories

susceptible, intermediate, and resistant. The approach to setting breakpoints varies by

organization or regulatory body. Depending on the approach taken, up to four sources of

data can be examined in establishing breakpoints (52).

MIC Distributions

Examination of MIC distributions can indicate the range of MICs of a population of strains

that lack any known mechanisms of resistance to a particular drug (wild-type population).

These distributions may aid in the recognition of new resistance mechanisms by highlighting

strains for which the MICs fall outside the normal distribution. However, distributions of MICs

have limited direct application since they vary between species, and for some strains for

which the MICs are outside the normal range, the MICs may be below clinically derived

breakpoints. Such strains may or may not respond to treatment. Knowledge of the presence

of specific resistance mechanisms that inactivate compounds of a particular drug class

assists in deriving microbiological breakpoints often called epidemiological cutoff values.

Pharmacokinetics and Pharmacodynamics

Pharmacokinetics examines the absorption, distribution, accumulation, and elimination

(metabolism and excretion) of a drug in the body over time. These parameters are usually

determined using healthy volunteers. A drug’s MICs can be compared with the concentration

of the drug achievable in blood or other body fluids (e.g., cerebrospinal fluid). In the past,

breakpoints were chosen generally such that the MICs for susceptible pathogens would be

exceeded by the drug level for most or all of the dosing interval. Newer data that are now

considered when establishing breakpoints include pharmacodynamic calculations.

Pharmacodynamics is the study of the time course of drug action against the microorganism.

For antimicrobial agents, the desired action is pathogen eradication. In vitro

pharmacodynamic studies have revealed that agents fall into three classes: those with

principally time-dependent antimicrobial action and no or short postantibiotic effects, those

with time-dependent action and long postantibiotic effects, and those with prominent

concentration-dependent action (17). For drugs with time-dependent action and no or short

postantibiotic effects, the critical determinant of bacterial killing in vivo is the percentage of

time in a dosing interval that the drug concentration is above the MIC. For the other two

classes, the important determinant is the ratio of the area under the concentration-time

curve to the MIC and/or the ratio of the peak concentration to the MIC. for β-lactams, shortacting

macrolides, and clindamycin, the relevant measure is the percentage of time that the

drug concentration is above the MIC, and the ratio of the area under the concentration-time

curve or of the peak concentration to the MIC is the relevant parameter for aminoglycosides,

long-acting macrolides, tetracyclines, glycopeptides, and fluoroquinolones (18). These values

can be used to calculate the maximum MICs or breakpoints that allow the achievement of

optimum efficacy with standard drug dosing schedules.

Clinical and Bacteriological Response Rates

During clinical trials, the clinical and/or bacteriologic eradication response rates of organisms

for which the MICs of new antimicrobial agents have been determined give an indication of

the relevance of breakpoints selected by using the MIC distributions and the

pharmacokinetic/pharmacodynamic properties of the drug. Response rates of at least 80%

may be expected for organisms classified as susceptible, although the rates can be lower

depending on the site and type of infection. While in some countries breakpoints are

determined primarily from clinical and bacteriological response rates, the CLSI evaluates

clinical and bacteriological response rates in conjunction with population distributions,

pharmacokinetics, and pharmacodynamics in establishing the breakpoints in an attempt to

provide the best correlation between in vitro test results and clinical outcome (8).

Inhibition Zone Diameter Distributions for Disk Diffusion

Methods

Once the MIC breakpoints are selected, disk diffusion breakpoints can be chosen by plotting

the inhibition zone diameters against the MICs derived from the testing of a large number of

strains of various species. A statistical approach that uses the linear regression formula may

be used to calculate the appropriate zone diameter intercepts for the predetermined MIC

breakpoints. An alternative, pragmatic approach to deriving disk diffusion breakpoints is the

use of the error rate-bounded method, in which the zone diameter criteria are selected on

the basis of the minimization of the disk interpretive errors, especially the very major errors

(5,38) (Fig. 1). Newer statistical techniques are being studied to improve the correlation with

MICs (16). The newest CLSI approach focuses on the rate of interpretive errors near the

proposed breakpoint versus rates of errors with MICs more than a single log2 dilution from

the MIC breakpoints (8). The concept is that errors that occur with isolates for which MICs

are very close to the MIC breakpoints are less of a concern than errors with more highly

resistant or susceptible strains.

Breakpoints derived by professional groups or regulatory bodies in various countries are

often quite similar. For instance, there is a small number of breakpoint discrepancies

between the CLSI and the U.S. Food and Drug Administration, which are under review by

both groups. However, there can be notable differences in the breakpoints used in different

countries or regions for the same agents. The reasons for the differences may be that certain

countries use different dosages or administration intervals for some drugs. In addition, some

countries are more conservative in assessing the susceptibility to antimicrobial agents and

place greater emphasis on the detection of emerging resistance, noted primarily by

examination of microorganism population distributions. Technical factors such as the

inoculum density, atmosphere of incubation, and test medium can also affect MICs and zone

diameters, thereby justifying different interpretive criteria in some countries. These technical

differences are summarized in chapter 68 of this Manual. Two non-U.S. methods minimize or

avoid the use of an intermediate category of susceptibility, based on the rationale that such

results are of little value to clinicians (3, 36). The lack of a buffer between susceptible and

resistant categories can result in higher rates of incorrect categorization. It may be safer for

a laboratory to employ a method that uses an intermediate category or, if not, to report

intermediate results as resistant.

Information on a range of international susceptibility testing methods and/or breakpoints can

be downloaded or purchased from the following websites:

· the CLSI website at http://www.clsi.org;

· the European Committee on Antimicrobial Susceptibility Testing website

at http://www.eucast.org (EUCAST now sets or harmonizes breakpoints for all its

member countries);

· the British Society for Antimicrobial Chemotherapy website at http://www.bsac.org.uk;

· the Deutsches Institut fur Normung website at http://www.beuth.de/;

· the Swedish Reference Group for Antibiotics website at http://www.srga.org;

· a website featuring the Danish commercial disk diffusion method at http://www.rosco.dk;

· a website featuring the Australian Calibrated Dichotomous Sensitivity disk diffusion

method at http://www. med.unsw.edu.au/pathology-cds/.

MOLECULAR DETECTION OF RESISTANCE Back to top

As highlighted in chapter 74, there is now a range of molecular techniques for the detection

of many resistance genes. While none are currently recommended for routine testing, some

have found a place in larger laboratories, where detection of certain important resistance

genes can be implemented in cost-effective manner. Examples include the detection

of mecA in Staphylococcus species, especially S. aureus, and detection of

the vanA and vanB genes in Enterococcus species. In addition, molecular techniques are

becoming available for the rapid detection of methicillin-resistant S. aureus from positive

blood culture bottles. These molecular tests have been increasingly valuable for infection

control purposes (48). Molecular methods are also valuable for confirming unusual

resistances and for determining which mechanism of resistance is present when this has

epidemiological significance. For example, there is currently great interest in the emergence

and spread of plasmid-mediated varieties of carbapenem resistance (53), quinolone

resistance (45), and resistance to aminoglycosides attributable to ribosomal methylases (6)

in enteric gram-negative bacteria.

SELECTED USE OF CONFIRMATORY AND SUPPLEMENTARY

TESTS Back to top

Besides performing routine susceptibility testing, laboratories will encounter isolates or tests

results that are unexpected, i.e., for which there is no testing guidance available or which

are considered to be of major epidemiological importance. The CLSI provides some guidance

on what resistances might be considered unexpected (either uncommon or never reported

[14]). Recommendations of how to proceed vary, but if results are not attributable to simple

laboratory errors and are reproducible, then testing by an alternative method and, if

necessary, referral to a reference laboratory is warranted. For some key resistances, e.g.,

carbapenem resistance in Enterobacteriaceae, the most effective confirmatory method is

resistance gene detection, as described in chapter 74. In other circumstances, one of the

special phenotypic tests described inchapter 70 will suffice.

Inducible resistance is considered to be clinically important for a small number of

antimicrobial-bacterial combinations. At present, only methods for detecting inducible

resistance to clindamycin in Staphylococcus andStreptococcus species have been sufficiently

evaluated and standardized to be recommended for routine laboratory use (12, 13).

For some clinical conditions, e.g., bacterial meningitis caused by Streptococcus

pneumoniae, susceptibility test interpretation and clinician guidance can be enhanced by

performing MIC measurements if these data were not generated by the laboratory’s routine

method. Customized or locally prepared microtiter trays or, alternatively, a commercial

gradient diffusion method may be used (39).

REPORTING OF RESULTS Back to top

The reporting of results is the crucial final step in susceptibility testing. There are no

universally agreed upon practices for generating reports, but the following elements should

be considered. Categorical reporting, in which the tests results are susceptible, intermediate,

or resistant, is standard practice and widely understood by clinicians. When available, MIC

data may be reported but should appear only along with a categorical interpretation.

Susceptibility test reports should be formatted in such a way that the results are

unambiguous in either printed or electronic form, especially if more than one organism is

being reported. Most importantly, so-called cascade reporting is recommended to reduce the

chance of the clinician choosing a broader-spectrum antimicrobial agent inappropriately (19).

Cascade reporting involves the withholding of results for broader-spectrum antimicrobials

from the report when the isolate tested is susceptible to narrower spectrum agents, e.g.,

withholding a vancomycin result when an isolate of Staphylococcus aureus tests as

susceptible to oxacillin or cefoxitin. Such reporting is considered to be an essential part of

hospital antimicrobial stewardship programs (20), as is the production of annual reports that

summarize overall susceptibility and resistance patterns (antibiograms) (10).

FUTURE DIRECTIONS AND NEEDS IN ANTIMICROBIAL

SUSCEPTIBILITY TESTING Back to top

Antimicrobial resistance is becoming widespread among a variety of clinically significant

bacterial species (47,55). Therefore, microbiology laboratories play a key role in the patient

management process by providing accurate data on which physicians can base therapy

decisions. Susceptibility testing results, however, are also used by investigators in

surveillance studies and by infection control practitioners to detect and control the spread of

antibiotic-resistant organisms (10, 43). Surveillance can be performed at laboratories at the

local, regional, national, and international level through direct interchange of data from

laboratory information systems to centralized databases (42). Thus, the accuracy of stored

results becomes almost as important as the accuracy of test performance and interpretation.

To meet these challenges and responsibilities, clinical microbiologists must continuously

assess and update their susceptibility testing strategies. The first priority is to use accurate

and reliable methods, whether they are conventional or perhaps newer molecular methods.

Then, careful monitoring of test performance with well-characterized control strains that

challenge the capabilities of the testing methods becomes essential. Today, laboratories

must use a variety of testing methods, each tailored specifically to a particular species or

group of organisms. It is not likely that a single method, whether conventional or

commercial, will be optimal for all antimicrobial agents, organisms, and resistance

mechanisms. This will require increased education and training for clinical microbiologists in

the future. Some assistance may be sought from the computer-based “expert” systems that

allow a rapid and accurate view of antimicrobial susceptibility profiles and recognition of

potentially aberrant results or novel resistance mechanisms (15, 44). Rapid progress is also

being made on molecular methods that are starting to have practical application in routine

clinical laboratories (1, 35, 37; see also chapter 74).

More-effective means of conveying critical antimicrobial susceptibility testing information to

clinicians in a time frame that allows efficient and effective management of patients and in a

format that is unambiguous to clinicians in various practice specialties are still needed.

Clinical microbiologists should become more proactive in the reporting of antimicrobial

susceptibility results and in cross-linking that information to other databases (e.g., those for

pharmacy prescriptions) to ensure that patients receive the most efficacious cost-effective

therapy.

Susceptibility Test Methods: Dilution and Disk Diffusion Methods

JEAN B. PATEL, FRED C. TENOVER, JOHN D. TURNIDGE, AND JAMES H. JORGENSEN

There are a number of methods for antimicrobial susceptibility testing of bacteria, and they

are categorized into dilution methods that generate MIC results and disk diffusion methods

that generate zone diameter results. Susceptibility testing methods can also be categorized

as generic reference methods, which are described by standards-setting organizations (e.g.,

the Clinical and Laboratory Standards Institute [CLSI], European Union Committee on

Antimicrobial Susceptibility Testing [EUCAST], and British Society for Antimicrobial

Chemotherapy [BSAC]), and commercial methods, which are mostly automated systems

(e.g., MicroScan [Siemens Healthcare Diagnostics, Deerfield, IL], Vitek [bioMerieux, Durham,

NC], Phoenix [BD, Sparks, MD], or Sensititre [Trek Diagnostics Systems, Cleveland, OH]) or

gradient diffusion methods (e.g., Etest [bioMerieux] or M.I.C.Evaluators [Oxoid, Cambridge,

United Kingdom]). Generic reference methods are those in which the reagents for testing can

be obtained from multiple sources and prepared in a laboratory without the need for

sophisticated manufacturing processes. The CLSI reference methods are broth macrodilution,

broth microdilution, agar dilution, and disk diffusion (20, 21).

The choice of methods to be used in individual laboratories is based on factors such as

relative ease of performance, cost, flexibility in selection of drugs for testing, availability of

automated or semiautomated devices to facilitate testing, and perceived accuracy of the

methodology (50). Reference dilution methods are typically used by pharmaceutical

companies to establish MIC data for new antimicrobial agents, by research laboratories and

device manufacturers as a standard to which new susceptibility testing methods are

evaluated, and by reference laboratories for confirming unusual susceptibility test results.

Although it has become increasingly uncommon, some clinical microbiology laboratories use

reference dilution methods for routine diagnostic testing. More frequently, clinical

microbiology laboratories use automated systems or a combination of MIC and disk diffusion

methods for routine susceptibility testing.

Interpretive categories for antimicrobial agent test results (i.e., susceptible, intermediate,

and resistant) are established for disk diffusion methods based on MIC data so that

interpretive errors between methods are minimized (18). Briefly, interpretive categories, or

breakpoints, are first established for MIC results generated by either the broth or agar

reference method. These breakpoints are based upon the normal MIC distributions for a

bacterial species, pharmacokinetic/pharmacodynamic modeling data, and data from clinical

outcome studies. Subsequently, disk diffusion breakpoints are set by comparing disk

diffusion data to MIC data and choosing breakpoints so that interpretive errors between

methods are within acceptable limits (a more detailed description is provided later in this

chapter). Interpretive categories for MIC methods and disk diffusion are established after

both intralaboratory and interlaboratory reproducibilities are verified for these methods.

Since interpretive criteria for disk diffusion data are set so that there is optimal correlation

with MIC results, in most cases, one method of susceptibility testing is not superior to the

other. However, there are specific examples of when one method may be preferred.

Daptomycin, polymyxin B, and colistin are antimicrobial agents that do not diffuse well in

agar, so in most cases disk diffusion is not an accurate method for these agents and MIC

methods are recommended (38, 47). MIC tests are also recommended for testing of

susceptibility of staphylococci to vancomycin because vancomycin susceptibility testing by

disk diffusion is not an accurate method for distinguishing vancomycin-intermediate from

vancomycin-susceptible staphylococci (70). Also, MIC testing is necessary for accurate

category assignment of Streptococcus pneumoniae isolates that produce zones of inhibition

of ≤19 mm around a 1-μg oxacillin disk (51). For some bacteria, there are limited or no disk

diffusion data available from well-controlled studies; thus, establishing interpretive criteria is

not feasible (e.g., Pseudomonas spp. other than Pseudomonas aeruginosa, Bacillus spp.,

and Corynebacteriumspp.). An example of a disk diffusion test being preferred to MIC testing

is the cefoxitin disk test that predictsmecA carriage by coagulase-negative staphylococci. The

disk test is more accurate than cefoxitin MIC testing for detecting mecA-mediated resistance

in these species. The MIC testing produced inaccurate results in part because some lots of

Mueller-Hinton broth did not adequately support growth of all coagulase-negative

staphylococcus isolates tested (71).

For MIC testing, the CLSI recommends reporting the MIC along with the interpretive

category, but for disk diffusion testing, only the interpretive category should be reported. In

most instances, the actual MIC does not affect patient management, except for cases of

endocarditis, osteomyelitis, and meningitis, when MICs close to the nonsusceptible

breakpoint can significantly influence the choice of antimicrobial agents. However, most

clinicians typically use only the interpretive category to make their treatment decisions. MIC

results also can be informative for epidemiological purposes for multiply resistant isolates.

For epidemiological purposes, emerging resistance mechanisms may be identified among

those isolates for which the antimicrobial agent MIC is above the normal MIC distribution for

isolates of the same species. A valuable source of normal MIC distributions for various

bacterium-antimicrobial agent combinations is the EUCAST website

(http://www.escmid.org/research_projects/eu_cast/). Isolates for which MICs are greater

than the normal distribution that still test as susceptible to the antimicrobial agent may

possess a resistance mechanism, such that a successful therapeutic outcome with the agent

cannot be predicted. There are several instances where information about elevated but

susceptible MICs is useful. For example, isolates of Enterobacteriaceae that show elevated

fluoroquinolone MICs but that are still in the susceptible category may possess a first-step

fluoroquinolone mutation or possibly a plasmid-mediated fluoroquinolone resistance

mechanism that may either reduce the effectiveness of the drug, as with Salmonella

enterica serovar Typhi infections, or allow the organism to survive long enough to develop

high-level resistance (56). This information would be particularly useful in an outbreak

setting in which fluoroquinolones were being considered for treatment or prophylaxis.

Similarly, an elevated but susceptible-range cephalosporin MIC in an isolate

of Enterobacteriaceae may indicate production of an extended-spectrum β-lactamase (ESBL).

When using the revised CLSI cephalosporin breakpoints for Enterobacteriaceae, tests for

ESBL detection are no longer required, but ESBL production is useful information for

epidemiological and infection control purposes. When treating an infection caused by a

multiresistant isolate, which may be susceptible only to a single indicated agent, clinicians

may consider using agents that give intermediate or even resistant results at an alternative

dose, a different route of administration to optimize drug concentrations at the site of

infection, or combinations of agents to try to effect a cure. An MIC result combined with

agent-specific pharmacokinetic/pharmacodynamic data may be used to guide this decision.

The selection of antibacterial agents for testing is complicated by the large number of agents

available today and the diversity of institutional formularies. Some of these compounds,

however, exhibit similar, if not identical, activities in vitro, so in some cases, one compound

can be tested as a surrogate to represent one or more closely related compounds. Such

extrapolations are listed in Table 1. Use of drug surrogates (also referred to as class

representatives) can substantially reduce the number of agents required for testing and in

some cases provide necessary flexibility in adapting commercial test systems for routine use

in a variety of institutions. For instance, the susceptibility of Staphylococcus spp. to oxacillin

(or cefoxitin) can be extrapolated to apply to all currently available penicillinase-stable

penicillins, most cephalosporins (with the exception of the cephalosporins against methicillinresistant

Staphylococcus aureus [MRSA], e.g., ceftaroline and ceftobiprole), and essentially

all carbapenems. Thus, it is unnecessary to test any of the agents in these chemical classes

(20). Other extrapolations are possible, especially if there is demonstrated susceptibility to a

less potent member of the chemical class of antimicrobial agent.

DILUTION TESTING: AGAR METHOD Back to top

Dilution of Antimicrobial Agents

The solvents and diluents needed to prepare stock solutions of most commonly used

antimicrobial agents are listed in the CLSI document on dilution testing (20).

Preparation, Supplementation, and Storage of Media

Mueller-Hinton agar is the recommended medium for testing most commonly encountered

aerobic and facultatively anaerobic bacteria (20). The dehydrated agar base is commercially

available and should be prepared as described by the manufacturer. Before sterilization, the

molten agar is usually distributed into screw-cap tubes in exact aliquots sufficient to dilute

the desired antimicrobial concentrations 10-fold. Tubes of agar, one for each drug

concentration to be tested, are sterilized by autoclaving at 121°C for 15 min, and the agar is

allowed to equilibrate to 48 to 50°C in a preheated water bath. Once the agar has

equilibrated, the appropriate volume of antimicrobial agent is added, the tube contents are

mixed by gentle inversion and poured into 100-mm-diameter round or square sterile plastic

petri plates set on a level surface, and the agar is allowed to solidify. For growth controls,

agar plates without antimicrobial agents are also prepared. All plates should be filled to a

depth of 3 to 4 mm (20 to 25 ml of agar per round plate and 30 ml per square plate), and

the pH of each batch should be checked to confirm the acceptable pH range of 7.2 to 7.4

(20).

After sterilization and temperature equilibration of the molten agar, any necessary

supplements are aseptically added to the Mueller-Hinton agar at the time of addition of the

drug solutions. For testing of streptococci, supplementation with 5% defibrinated sheep or

horse blood is recommended (20). However, sheep blood supplementation may antagonize

the activities of sulfonamide and trimethoprim with some organisms (8). The presence of

blood also affects results with novobiocin and nafcillin as well as the in vitro activities of

cephalosporins against enterococci (12, 67); therefore, blood supplementation should not be

used unless necessary for bacterial growth (see chapter 75 for acceptable methods for

testing of fastidious bacterial species). Performance standards for Mueller-Hinton agar have

been defined sufficiently such that calcium and magnesium supplementation is unnecessary

(23). The agar should be supplemented with 2% NaCl for testing of oxacillin against

staphylococci (42).

Once prepared, plates should be sealed in plastic bags and stored at 4 to 8°C. In general,

they should be used within 5 days of preparation or as long as the MICs for control strains

that are tested routinely are within the acceptable ranges. However, certain agents are

sufficiently labile that plates may not be stored prior to use, e.g., carbapenems, cefaclor, and

clavulanic acid. Before inoculation, plates that have been stored under refrigeration should

be allowed to equilibrate to room temperature and the agar surface should be dry.

Inoculation Procedures

Variations in inoculum size may substantially affect MICs; therefore, careful inoculum

standardization is required to obtain accurate results. The recommended final inoculum for

agar dilution is 104 CFU per spot (20). This may be achieved in either of two ways. Four or

five colonies are picked from overnight growth cultures on agar-based medium and

inoculated into 4 to 5 ml of suitable broth that will support good growth (usually tryptic soy

broth). Broths are incubated at 35°C until visibly turbid, and then the suspension is diluted

until it matches the turbidity of a 0.5 McFarland, barium sulfate (BaSO4), or latex particle

turbidity standard (ca. 108CFU/ml). The 0.5 McFarland standard may be purchased or the

barium sulfate standard may be prepared as described in the CLSI document (20). The

accuracy of the standard should be verified by using a spectrophotometer with a 1-cm light

path; for the 0.5 McFarland standard, the absorbance at 625 nm should be 0.08 to 0.13

(20). An alternative inoculum standardization method, one that is preferred by many

microbiologists, utilizes direct suspension of colonies from overnight growth cultures on a

nonselective agar medium in saline or broth to a turbidity that matches the 0.5 McFarland

standard. This approach eliminates the time needed for growing the inoculum in broth (20).

In either case, normal saline or sterile broth is used to make a 1:10 dilution of the

suspension to obtain an adjusted concentration of 107 CFU/ml (20).

Once the adjusted bacterial inoculum suspension is prepared, inoculation of the antimicrobial

agent plates should be accomplished within 15 min, since longer delays may lead to changes

in inoculum size. By using a pipette, a calibrated loop, or, more commonly, an inoculumreplicating

device, 0.001 to 0.002 ml (1 to 2 μl) of the 107-CFU/ml suspension is delivered to

the agar surface, resulting in the final desired inoculum of approximately 104 CFU per spot.

For convenience, use of a replicator is preferred because consistent inoculum volumes for up

to 36 different isolates are delivered simultaneously (20, 68). To use this device, an aliquot

of the adjusted inoculum for each isolate is pipetted into the appropriate well of an inoculum

seed plate and a multiprong inoculator is used to pick up and gently transfer 1 to 2 μl from

the wells to the agar surfaces. Replicators are also available that deliver only 0.1 to 0.2 μl

per spot and that do not require the 0.5 McFarland standard suspension to be diluted prior to

delivery to the agar surface (20). The surfaces of the agar plates must be dry before

inoculation, which should begin with a growth control plate that does not contain drug. Then

inoculation continues from plates with the lowest drug concentration to plates with the

highest drug concentration. Finally, a second growth control plate that does not contain drug

is inoculated to check for contamination or significant carryover of the antimicrobial agent.

All plates should be clearly marked so that the locations of the different isolates being tested

on each plate are known.

Incubation

Inoculated plates are allowed to stand for several minutes until the inoculum drops have

been completely absorbed by the medium; then they are inverted and incubated in air at

35°C for 16 to 20 h before results are read. To facilitate detection of vancomycin-resistant

enterococci and methicillin-resistant or vancomycin-resistant or -intermediate staphylococci,

plates containing vancomycin or oxacillin should be incubated for a full 24 h before results

are read (20). Incubation should not be carried out in the presence of an increased

CO2concentration unless a fastidious organism is being tested (see chapter 75).

Interpretation and Reporting of Results

Before reading and recording the results obtained with clinical isolates, those obtained with

applicable quality control strains tested at the same time should be checked to ensure that

their values are within the acceptable ranges (see “Quality Control” below), and the drugfree

control plates should be examined for isolate viability and purity. Endpoints for each

antimicrobial agent are best determined by placing plates on a dark background and

examining them for the lowest concentration that inhibits visible growth, which is recorded

as the MIC. A single colony or a faint haze left by the initial inoculum should not be regarded

as growth. If two or more colonies persist at antimicrobial concentrations beyond an

otherwise obvious endpoint or if there is no growth at lower concentrations but there is

growth at higher concentrations, the isolate should be subcultured to confirm purity and the

test should be repeated. Substances that may antagonize the antibacterial activities of

sulfonamides and trimethoprim may be carried over with the inoculum and cause “trailing,”

or less definite endpoints (8, 12). Therefore, the MICs of these antimicrobial agents should

be interpreted as the endpoints at which 80% or more diminution of growth occurs. Although

much less pronounced, trailing endpoints may also occur for some organisms with

bacteriostatic agents such as chloramphenicol, the tetracyclines, linezolid, and quinupristindalfopristin

(20).

The MIC of each antimicrobial agent is usually recorded in micrograms per milliliter, although

in Europe and in the international standard reference method, the values are expressed as

milligrams per liter (20, 43). These quantitative results should be reported with the

appropriate corresponding interpretive category (susceptible, intermediate, or resistant), or

the interpretive category may be reported alone. The MIC interpretive standards for these

susceptibility categories, as currently recommended by the CLSI (20), are provided inTable

3. For detailed instructions concerning the use of these criteria and categories, the latest

CLSI standards for dilution testing methods should be consulted (22).

DILUTION TESTING: BROTH METHODS Back to top

The general approaches for broth methods include broth macrodilution, in which the broth

volume for each antimicrobial concentration is ≥1.0 ml (usually 2 ml) contained in test

tubes, and broth microdilution, in which antimicrobial dilutions are most often in 0.1-ml

volumes in wells of 96-well microdilution trays.

Broth Macrodilution Methods

Dilution of Antimicrobial Agents

Stock solutions are prepared as discussed in the CLSI document on dilution testing (20) and

are the same as those used for agar dilution tests. As in the agar method, the actual

volumes used for the dilutions would be proportionally increased according to the number of

tests being prepared, with a minimum of 1.0 ml needed for each drug concentration.

Because addition of the inoculum results in a 1:2 dilution of each concentration, all final drug

concentrations must be prepared at twice the actual desired testing concentration (see

“Inoculation Procedures” below).

Preparation, Supplementation, and Storage of Media

Cation-adjusted Mueller-Hinton broth (CAMHB) is recommended for routine testing of

commonly encountered nonfastidious organisms (20). Adjustment of the cations Ca2+ (20 to

25 mg/liter) and Mg2+ (10 to 12.5 mg/liter) is required to ensure acceptable results when P.

aeruginosa isolates are tested with aminoglycosides and when tetracycline is tested with

other bacteria (6). However, for convenience and consistency, cation adjustment of Mueller-

Hinton broth is now recommended for testing of all species and antimicrobial agents (20).

Some manufacturers provide Mueller-Hinton broth that already has appropriate

concentrations of divalent cations, so the cation content of commercial dehydrated media

must be ascertained and care must be taken to supplement only those commercial broths

that have not already been adjusted. If adjustment is necessary, it can be accomplished by

the addition of suitable volumes of filter-sterilized, chilled CaCl2 stock (3.68 g of CaCl2 ・ 2H2O

dissolved in 100 ml of deionized water for a concentration of 10 mg of Ca2+ per ml) and

MgCl2 stock (8.36 g of MgCl2 ・ 6H2O in 100 ml of deionized water for a concentration of 10

mg of Mg2+ per ml) to the cooled broth (20). Insufficient cation concentrations result in

increased aminoglycoside activity (27), and excess cation content results in decreased

aminoglycoside activity against P. aeruginosa (27, 79). While the effects of inappropriate

calcium and magnesium ion contents are well recognized, other ions, including zinc and

manganese, may adversely affect the activities of some drugs, e.g., carbapenems (26). The

CLSI has initiated a consensus standard for manufacturers of Mueller-Hinton broth that

attempts to specify all known factors that determine performance of the medium (24). Other

supplements to CAMHB may be required for accurate susceptibility results with specific

agents. For example, accurate daptomycin testing requires a calcium supplement so that the

final concentration is 50 mg/liter (20, 37), and detection of staphylococcal resistance to

oxacillin requires that the CAMHB be supplemented with 2% NaCl (20, 76). Accurate

susceptibility testing of tigecycline requires that the CAMHB be prepared fresh on the day of

testing or frozen within 12 h of preparation (22).

To minimize evaporation and deterioration of antimicrobial agents, tubes should be tightly

capped and stored at 4 to 8°C until needed. With most agents, the dilutions should be used

within 5 days of preparation or as long as quality control ranges are maintained (see “Quality

Control” below). As in agar dilution testing, certain β-lactam agents are too labile for

prolonged storage at final test concentrations.

Inoculation Procedures

The recommended final inoculum for broth dilution testing is 5 × 105 CFU/ml. Isolates are

inoculated into a broth that will support good growth (such as tryptic soy broth) and

incubated until turbid. The turbidity is adjusted to match that of a 0.5 McFarland standard

(approximately 108 CFU/ml). Alternatively, four or five colonies from overnight growth

cultures on a nonselective agar plate may be directly suspended in broth to match the

turbidity of the 0.5 McFarland standard (20). This alternative approach is preferred for

testing of oxacillin against staphylococci (20). A portion of the standardized suspension is

diluted approximately 1:100 (to 106 CFU/ml) with broth or saline. When 1 ml of this dilution

is added to each tube containing 1 ml of the drug diluted in CAMHB, a final inoculum of 5 ×

105 CFU/ml is achieved. Broth not containing an antimicrobial agent is inoculated as a control

for organism viability (growth control). All tubes should be inoculated within 30 min of

inoculum preparation, and an aliquot of the inoculum should be plated to check for purity.

Incubation

Tubes are incubated in ambient air at 35°C for 16 to 20 h before MICs are determined.

Incubation should be extended to a full 24 h for the detection of vancomycin-resistant

enterococci and oxacillin-resistant or vancomycin-resistant or -intermediate staphylococci

(20). An atmosphere with increased CO2 should not be used.

Interpretation and Reporting of Results

Before MICs for the test strains are read and recorded, the growth controls should be

examined for viability, inoculum subcultures should be checked for contamination, and

appropriate MICs for the quality control strains should be confirmed (see “Quality Control”

below). Growth or lack thereof in the antimicrobial-agent-containing tubes is best

determined by comparison with the growth control. Generally, growth is indicated by

turbidity, a single sedimented button >2 mm in diameter, or several buttons with smaller

diameters. As with the agar method, trailing endpoints may be seen when trimethoprim or

sulfonamides are tested, and the concentration at which 80% or greater diminution of

growth, compared with that of the growth control, occurs should be recorded as the MIC

(20). Other interpretation problems include the “skipped tube” phenomenon, in which growth

is not observed at one concentration but is observed at lower and higher drug

concentrations. Most authorities suggest that when this occurs, the skipped tube should be

ignored and the concentration that finally inhibits growth at serially higher concentrations

should be recorded as the MIC. If more than one skipped tube occurs or if there is growth at

higher antimicrobial concentrations but not at lower ones, the results should not be reported

and the test for that drug should be repeated.

The lowest concentration that completely inhibits visible growth of the organism as detected

by the unaided eye is recorded as the MIC. The CLSI MIC interpretive standards in effect as

of the date of this writing (22) for the susceptibility categories are provided in Table 3. The

definitions of and comments concerning these categories that were given for the agar

method also pertain to the broth macrodilution method.

Advantages and Disadvantages

The broth macrodilution method is a well-standardized and reliable method that may be

useful for research purposes or for testing of one drug with a bacterial isolate. However,

because of the laborious nature of the procedure and the availability of more convenient

dilution systems (e.g., microdilution), this procedure is generally not practical for routine

susceptibility testing in most clinical microbiology laboratories.

Broth Microdilution Method

The convenience afforded by the availability of dilution susceptibility testing in microdilution

trays has led to the widespread use of broth microdilution methods. In fact, the broth

microdilution method is now considered the international reference susceptibility testing

method (43). The disposable plastic trays, containing a panel of several antimicrobial agents

to be tested simultaneously, may be prepared in-house or obtained commercially either

frozen or freeze-dried. When commercial systems are used, the manufacturer’s

recommendations concerning storage, inoculation, incubation, and interpretation should be

followed. The primary focus of this section is the in-house preparation and use of broth

microdilution panels. However, most of the principles and practices discussed here are

pertinent to the broth microdilution method regardless of the source of the antibiotic panels.

Dilution of Antimicrobial Agents

Antimicrobial stock solutions are prepared as outlined in the CLSI document on dilution

testing (20). The dilution scheme for preparing broth microdilution panels is the same as that

described for agar and broth macrodilution methods. Automated dispensing systems for

preparation of microdilution panels use tubes that contain from 10 to 200 ml or more of

broth containing each antimicrobial concentration. From the master tube dilutions, aliquots of

0.05 or 0.1 ml are simultaneously dispensed into the corresponding wells of each broth

microdilution tray by using a mechanized dispenser. If 0.05-ml volumes are dispensed,

allowances must be made for the 1:2 dilution of the final drug concentration that will occur

when the 0.05 ml of inoculum is added (see “Inoculation Procedures” below). When 0.1-ml

aliquots are dispensed, the volume of inoculum normally used is sufficiently small (≤0.005

ml) that adjustments in the antimicrobial dilution scheme are not needed. As a general rule,

when the inoculum volume is less than 10% of the broth volume in the well, dilution of the

antimicrobial concentration by the inoculum does not have to be taken into account (20).

Preparation, Supplementation, and Storage of Media

CAMHB is the recommended medium for broth microdilution testing of nonfastidious

organisms and should be prepared as discussed above for the broth macrodilution method.

Also, supplementation of the broth with 2% NaCl is required for detection of oxacillinresistant

staphylococci, and daptomycin testing requires that the broth be adjusted to 50 mM

Ca2+ (20). After the antimicrobial dilutions have been dispensed into the plastic trays, the

panels are stacked in groups of 5 to 10, with a tray lid or an empty tray placed on top to

minimize contamination and evaporation. Each stack is sealed in a plastic bag and frozen

immediately, preferably at −60°C or colder, or at −20°C if a −60°C freezer is not available.

At −20°C, preservation is ensured for at least 6 weeks with most drugs, but the shelf life

may be extended to months if the trays are stored at −60 to −70°C. Care must be taken in

storing highly labile compounds such as cefaclor, clavulanic acid, and carbapenems, which

may lose potency during storage. Trays with these agents should not be stored at

temperatures above −60°C. If thawed, panels must be used or discarded but not refrozen,

since freeze-thaw cycles cause substantial deterioration of β-lactam antibiotics. For this

reason, −20°C household-type freezers with self-defrosting units must not be used.

Inoculation Procedures

As with the macrodilution procedure, the final desired inoculum concentration is 5 ×

105 CFU/ml. The isolates may be grown in broth to match the turbidity of a 0.5 McFarland

standard (ca. 1 × 108 to 2 × 108 CFU/ml), or a suspension of that density can be made from

colonies grown overnight on a nonselective agar medium (5), which is the method preferred

for detecting oxacillin-resistant staphylococci (20). For broth microdilution procedures that

require 0.001- to 0.005-ml volumes to inoculate wells containing 0.1 ml of broth, a portion of

the 0.5 McFarland standard suspension is diluted 1:10 (107 CFU/ml) in sterile saline or broth.

Multipoint metal or disposable plastic inoculum replicators designed to collect and deliver

appropriate volumes are used to transfer the inoculum from the diluted suspension to the

wells of the broth microdilution tray, resulting in further dilutions ranging from 1:20 to 1:50.

Final inoculum concentrations should be 4 × 105 to 6 × 105 CFU/ml (4 × 104 to 6 × 104 CFU

per well). For protocols that use an inoculum volume of 0.05 ml to inoculate 0.05 ml of

broth, a 1:100 dilution of a 0.5 McFarland standard suspension (ca. 106 CFU/ml) is used.

When the inoculum is added to the wells, the 1:2 dilution of the 106-CFU/ml inoculum results

in a final inoculum concentration of 5 × 105 CFU/ml (5 × 104 CFU per well) and also halves

the antimicrobial concentration in each well. Special care should be taken to confirm the

inoculum density on a periodic basis to ensure that the appropriate density of inoculum is

achieved. Moreover, slight deviations from the initial 1:10 dilution described above may be

necessary to provide the target inoculum density with some species or organism groups.

Insufficient inoculum can be a significant problem with the inducible resistance mechanisms

of some organisms (such as β-lactamases), which may not be recognized as a problem

based on the MICs obtained for the very susceptible quality control strains.

Broth microdilution trays should be inoculated within 30 min of inoculum preparation, and an

aliquot should be subcultured to check the purity of the isolates. Finally, one well of each

panel not containing an antimicrobial agent should be inoculated and used as a growth

control, and a second uninoculated well serves as a sterility control.

Incubation

After inoculation, each tray should be covered with plastic tape or sealing film, sealed in a

plastic bag, or tightly fitted with a lid or an empty tray to prevent evaporation during

incubation. Trays are incubated in ambient air at 35°C for 16 to 20 h before results are read

and should not be incubated in stacks of more than four trays for uniform temperature

distribution. The incubator should be kept sufficiently humid to avoid evaporation but not so

humid that condensation results in contamination problems. A full 24 h of incubation is

recommended for the detection of vancomycin-resistant enterococci and oxacillin-resistant or

vancomycin-resistant or -intermediate staphylococci (20). Incubation in an atmosphere with

increased CO2 should not be used with nonfastidious organisms.

Interpretation and Reporting of Results

Before MICs for the clinical isolates are read and recorded, the growth control wells should

be examined for organism viability and the inoculum purity should be checked. The

appropriateness of the MICs obtained for the quality control strains should be confirmed if

tests of these strains were set up simultaneously with those of clinical isolates (see “Quality

Control” below). Various viewing devices are available and should be used to facilitate

examination of the broth microdilution wells for growth. The simplest and most reliable

method is the use of a parabolic magnifying mirror and tray stand that allow clear visual

inspection of the undersides of the microdilution trays. Growth is best determined by

comparison with that in the growth control well and generally is indicated by turbidity

throughout the well or by buttons, single or multiple, in the well bottom. The occurrence of

trailing endpoints with trimethoprim or sulfonamides should be ignored, and the MIC

endpoint should be based on ≥80% growth inhibition. Results for drugs with more than one

skipped well should not be reported, as with the broth macrodilution test.

The CLSI MIC interpretive criteria (22) for susceptibility categories are given in Table 3. It

should be noted that these values are published each year, and only the most recent tables

should be used for interpretation of results. The definitions of the interpretive categories and

the comments concerning the use of these standards for agar and broth macrodilution

methods are also applicable to broth microdilution methods.

Advantages and Disadvantages

Broth microdilution is a well-standardized reference method for antimicrobial susceptibility

testing. Inoculation and reading procedures allow convenient simultaneous testing of

multiple antimicrobial agents with individual isolates. Because few laboratories have the

facilities required for preparation of broth microdilution trays, several sources of

commercially prepared antimicrobial agent panels are available. Such products provide either

frozen or freeze-dried trays with wells containing prepared antimicrobial dilutions. Frozen

trays must be stored at least at −20°C in the laboratory, whereas dried panels can be stored

at room temperature. Most of these products are accompanied by multipoint inoculating

devices; however, the trays may be inoculated with multichannel pipettors. Results of testing

may be determined by visual examination or by use of semiautomated or automated

instrumentation.

Breakpoint Susceptibility Tests

“Breakpoint susceptibility testing” refers to methods by which antimicrobial agents are tested

only at the specific concentrations necessary for differentiating among the interpretive

categories of susceptible, intermediate, and resistant rather than in a range of five or more

doubling-dilution concentrations used to determine MICs. When two drug concentrations are

selected adjacent to the breakpoints defining the intermediate and resistant categories, any

one of the interpretive categories may be determined. Growth at both concentrations

indicates resistance, growth at only the lower concentration signifies an intermediate result,

and no growth at either concentration indicates susceptibility.

Like full-range dilution testing, breakpoint methods require the use of appropriately adjusted

and supplemented Mueller-Hinton broth or agar. In addition, the standard inoculation,

incubation, and interpretation procedures recommended for the full-range dilution methods

should be followed.

Considering the limited range of drug concentrations tested, a greater number and variety of

antimicrobial agents can be incorporated into a broth microdilution panel for breakpoint

testing than into panels designed for full-range dilution testing (32). However, convenient

quality control procedures to ensure that appropriate concentrations of each antimicrobial

agent are present are lacking for breakpoint panels. One possible approach is to use one

organism for which the modal MIC is equal to or no less than one doubling dilution less than

the lower or lowest concentration tested and a second organism for which the modal MIC is

equal to or no more than one doubling dilution greater than the higher or highest

concentration tested (20). One of these two quality control organisms should provide onscale

results (20). Despite the theoretical soundness of this approach, routine quality control

of breakpoint panels is difficult and not readily accomplished in the clinical laboratory.

Resistance Screens

In some circumstances, testing a single drug concentration may be a reliable and convenient

method for detecting antimicrobial resistance. The most clinically useful resistance screens

are those for resistance to oxacillin in S. aureus, resistance to vancomycin

in Enterococcus spp. and Staphylococcus spp., and high-level resistance to gentamicin and

streptomycin in enterococci (20). These practical and reliable methods are described

in chapter 74 of this Manual.

Gradient Diffusion Method

The Etest (AB Biodisk, bioMerieux) and the M.I.C.Evaluator (Oxoid) are commercial methods

for quantitative antimicrobial susceptibility testing that incorporate a preformed antimicrobial

gradient applied to one side of a plastic strip to provide drug diffusion into an agar medium.

Each test is performed in a manner similar to that for disk diffusion testing, in that a 0.5

McFarland standard suspension of a test isolate is generally swabbed onto the agar surface

for inoculation. Following incubation, the MIC is read directly from a scale on the top of the

strip at the point where the ellipse of organism growth inhibition intercepts the strip. Several

strips, each containing a different antimicrobial agent, can be placed radially on the surface

of a large round Mueller-Hinton agar plate, or they can be placed in opposing directions on

large rectangular plates. MICs determined by this method generally agree well with MICs

generated by standard broth or agar dilution methods (4, 41, 63). The Etest and

M.I.C.Evaluator combine the simplicity and flexibility of the disk diffusion test with the ability

to determine MICs of up to five antimicrobial agents on a single 150-mm agar plate.

However, both agar gradient diffusion strips are much more expensive than the paper disks

used for diffusion testing. Strengths of these methods include the simplicity of the procedure

itself and the ability to determine the MIC of an infrequently tested drug and to test

fastidious or anaerobic bacteria by applying the strips onto specialized enriched media

(see chapters 71 and 72 of this Manual).

QUALITY CONTROL Back to top

Quality control recommendations are designed for evaluation of the precision and accuracy of

test procedures, monitoring of reagent performance, and evaluation of the performance of

the individuals who are conducting the tests.

Reference Strains

A critical element of quality control is the selection and use of reference bacterial strains that

are genetically stable and for which MICs are in the mid-range of MICs of each antimicrobial

agent tested (20). That is, the dilutions in a series should ideally encompass at least two

concentration increments above and below the previously established MIC for the reference

strain. If there are four or fewer dilutions in a series or if nonconsecutive dilutions are tested

(e.g., in breakpoint susceptibility testing), quality control for the correct interpretive category

only rather than actual MIC ranges may be accomplished. Escherichia coli ATCC 25922, P.

aeruginosa ATCC 27853, Enterococcus faecalis ATCC 29212, and S. aureus ATCC 29213 are

the recommended reference strains for both agar and broth dilution methods (20). The β-

lactamase-producing strain E. coli ATCC 35218 is recommended only for penicillin–β-

lactamase inhibitor combinations (20). These organisms may be obtained from the American

Type Culture Collection or other reliable commercial sources. For proper storage and

subculture procedures, the recommendations of either the CLSI (20) or the commercial

provider should be followed.

MIC Ranges

The acceptable quality control MIC ranges for the various reference strains are given in the

CLSI document on dilution testing (20). Updates of these MIC ranges are published annually

(22) and should be readily available in each clinical laboratory. An out-of-control result is

defined as an MIC not within the acceptable range of values. Certain out-of-control results

can be directly related to the medium used for testing. High MICs of gentamicin for P.

aeruginosa ATCC 27853 suggest an inappropriately high divalent cation content or

excessively low pH of the Mueller-Hinton medium, and low MICs indicate an insufficient

divalent cation concentration or elevated pH. Although trimethoprim-sulfamethoxazole is not

recommended for therapy ofEnterococcus faecalis infections, results obtained with the ATCC

29212 strain can be useful for detecting excessive amounts of substances such as thymidine

in the testing medium that interfere with the in vitro activity of antifolate drugs.

Trimethoprim-sulfamethoxazole MICs of >0.5 to 9.5 μg/ml indicate the presence of such

interfering substances (20).

Batch and Lot Quality Control

Representative plates, panels, or trays from each new reagent batch, if prepared in-house,

or from each new lot, if obtained from a commercial source, should be subjected to quality

control and sterility testing. Antimicrobial agent MICs obtained by testing reference quality

control strains should be within acceptable CLSI ranges (22). If such accuracy is not

achieved, the batch or lot should be rejected or patient results obtained with the

antimicrobial agent(s) in question should not be reported (see below). Similarly, if selected

uninoculated plates or trays fail the sterility check after incubation, the batch or lot should be

rejected. In addition to these formal quality control procedures that use reference strains,

careful review of susceptibility results obtained during daily testing of clinical isolates is

important to identify unusual susceptibility patterns possibly indicative of reagent or

technical problems.

Quality Control Frequency

In addition to batch and lot testing, quality control tests should be performed daily, or at

least every day that the plates or trays are being used to test clinical isolates. When quality

control is performed on each day of testing, performance is considered satisfactory if no

more than 3 of 30 consecutive results for each drug-reference strain combination are outside

the acceptable limits. If this frequency is exceeded, the laboratory must perform corrective

action to determine the source of the error and to correct it as described below. However, if

daily quality control testing does not reveal an excessive rate of errors, daily testing may be

replaced by weekly testing as outlined below (20, 22).

To convert to a weekly quality control testing interval, each drug-reference strain

combination is tested for 20 or 30 consecutive testing days to obtain a total of 20 to 30 MICs

for each combination. If no more than 1 of 20 or 3 of 30 MICs per combination are outside

the accuracy range, weekly testing may replace daily testing. During weekly testing, a single

MIC outside the acceptable range requires that daily testing be performed for five

consecutive days unless there is an obvious source of error (e.g., contamination, use of an

incorrect reference strain or an incorrect inoculum density, testing of an incorrect

antimicrobial agent, or use of an incorrect atmosphere for incubation). In such a

circumstance, the quality control test need be repeated only once. If no obvious source of

error is noted, but all five MICs for a problem drug-organism combination are within the

accuracy range, weekly testing may be resumed. If one or more of the five MICs for the

problem drug-organism combination are outside the accuracy range, daily testing must be

initiated and further means to resolve the problem must be pursued. Returning to weekly

testing requires again documenting 20 or 30 consecutive days with no more than one to

three MICs outside the accuracy range. If more than the acceptable number of MICs for

organism-drug combinations are outside the accuracy range, daily quality control testing

must be continued while the problem is being resolved (20, 22).

DISK DIFFUSION TESTING Back to top

The disk diffusion method of susceptibility testing allows categorization of most bacterial

isolates as susceptible, intermediate, or resistant to a variety of antimicrobial agents. To

perform the test, commercially prepared filter paper disks impregnated with a specified

single concentration of an antimicrobial agent are applied to the surface of an agar medium

that has been inoculated with the test organism. The drug in the disk diffuses through the

agar. As the distance from the disk increases, the concentration of the antimicrobial agent

decreases logarithmically, creating a gradient of drug concentrations in the agar medium

surrounding each disk. Concomitant with the diffusion of the drug, the bacteria that were

inoculated onto the surface and were not inhibited by the concentration of the antimicrobial

agent in the agar continue to multiply until a lawn of growth is visible. In areas where the

concentration of the drug is inhibitory, no growth occurs, forming a zone of inhibition around

each disk.

The disk diffusion procedure has been standardized primarily for testing common, rapidly

growing bacteria (7,21). This method should not be used to evaluate antimicrobial

susceptibilities of bacteria that show marked strain-to-strain variability in growth rates, e.g.,

some fastidious or anaerobic bacteria. The test, however, has been modified to allow reliable

testing of certain fastidious bacterial species (discussed in chapter 75 of thisManual).

The diameter of the zone of inhibition is influenced by the rate of diffusion of the

antimicrobial agent through the agar, which may vary among different drugs depending upon

the size of the drug molecule and its hydrophilicity. The zone size, however, is inversely

proportional to the logarithm of the MIC, measured as discussed earlier in this chapter.

Criteria currently recommended for interpreting zone diameters and MIC results for

commonly used antimicrobial agents are listed in Table 3 and published annually by the CLSI

(22).

Establishing Zone-of-Inhibition Diameter Interpretive Criteria

The first step in determining interpretive criteria for the disk diffusion test is selection of MIC

breakpoints that define susceptibility and resistance categories for each antimicrobial agent.

Zone-of-inhibition diameters that correspond to these breakpoints are initially established by

testing 300 or more bacterial isolates by both dilution and disk diffusion methods and

correlating the diameters of the zones of inhibition with the MICs determined for each drug

tested (18). Isolates tested should include not only those commonly encountered in clinical

laboratories but also those with resistance mechanisms pertinent to the class of antimicrobial

agent being tested (18). Organisms evaluated should be those most likely to be tested with

the antimicrobial agent in question. The data from these studies are analyzed by preparing a

scattergram of values (see the example in chapter 67). By convention, each MIC (log2 scale)

is plotted on the y axis, and the corresponding zone-of-inhibition diameter (arithmetic scale)

is plotted on the x axis. Regression analysis can be performed, and a straight regression line

showing the best fit is drawn. From this line, an approximation of the MIC can be inferred

from any zone diameter. For antimicrobial agents to which isolates are either susceptible or

resistant and only infrequently intermediate, regression analysis is not valid. In such cases,

the data are plotted as a scattergram and the interpretive standards are selected so as to

allow optimal separation of resistant and susceptible populations (16, 18, 61). This approach,

often called the error rate-bounded method, may also be employed to minimize interpretive

errors that can ensue from strictly applying the linear regression formula to a data set (18).

Antimicrobial Agent Disks

The amounts of antimicrobial agents in the disks used for the disk diffusion method are

standardized, and in the United States, only a single concentration for each drug is

recommended (22). The optimal amount of an antimicrobial agent per disk is determined

early in the development of a new drug by testing disks with several different drug contents

that can be evaluated by using scattergrams and regression lines (18). The most desirable

concentration of a drug per disk is that which produces a zone-of-inhibition diameter of at

least 10 mm with resistant isolates and a zone diameter no larger than 30 mm with

susceptible isolates.

Commercially prepared antimicrobial disks usually are supplied in separate containers, each

with a desiccant. They must not be used beyond the specified expiration date and should be

stored under refrigeration (2 to 8°C) or frozen in a non-frost-free freezer at −20°C or colder

until needed. Disks containing a β-lactam agent should always be stored frozen to ensure

that they retain their potency, although a small supply may be stored in the refrigerator for

up to 1 week. Unopened disk containers should be removed from the refrigerator or freezer 1

to 2 h before use. This allows the disks to equilibrate to room temperature before the

container is opened, thus minimizing the amount of condensation that will occur when warm

air contacts the cold disks. A commercially available, mechanical disk-dispensing apparatus

can be used and should be fitted with a tight cover, supplied with an adequate desiccant,

stored in the refrigerator when not in use, and warmed to room temperature before being

opened.

Agar Medium for Disk Diffusion

The recommended medium for disk diffusion testing in the United States is Mueller-Hinton

agar (21). This unsupplemented medium has been selected by the CLSI for several reasons:

(i) it demonstrates good batch-to-batch reproducibility for susceptibility testing; (ii) it is low

in sulfonamide, trimethoprim, and tetracycline inhibitors; (iii) it supports the growth of most

nonfastidious bacterial pathogens; and (iv) years of data and clinical experience regarding its

performance have been accrued. Fastidious bacteria, such as Haemophilusspecies, Neisseria

gonorrhoeae, Neisseria meningitidis, and streptococci, do not grow satisfactorily on

unsupplemented Mueller-Hinton agar but can be tested by the disk method by using

supplemented or modified test media as discussed in chapter 71 of this Manual.

Plates of Mueller-Hinton agar may be purchased, or the agar may be prepared from a

commercially available dehydrated base according to the manufacturer’s directions. If the

agar is prepared, only formulations that have been tested according to, and have met

acceptance limits recommended by, the CLSI should be used (23). The prepared medium is

autoclaved and immediately placed in a 45 to 50°C water bath. When cool, it is poured into

round plastic flat-bottomed petri dishes on a level surface to give a uniform depth of about 4

mm (60 to 70 ml of medium for 150-mm-diameter plates and 25 to 30 ml for 100-mmdiameter

plates) and allowed to cool to room temperature. Agar deeper than 4 mm may

cause false resistance results (excessively small zones), whereas agar less than 4 mm deep

may be associated with excessively large zones and false susceptibility.

Each batch of Mueller-Hinton agar should be checked when the medium is prepared to

ensure that the pH is between 7.2 and 7.4 at room temperature, which means that the pH

must be measured after the medium has solidified. This can be done by allowing a small

amount of agar to solidify around the tip of a pH electrode in a beaker or a cup, by

macerating a sufficient amount of agar in neutral distilled water, or by using a properly

calibrated surface electrode. A pH outside the range of 7.2 to 7.4 may adversely affect

susceptibility test results. If the pH is too low, drugs such as the aminoglycosides,

macrolides, and fluoroquinolones will appear to lose potency, whereas others (for example,

the penicillins and tetracyclines) may appear to have excessive activity. The opposite effects

are possible if the pH is too high.

Freshly prepared plates may be used the same day or stored in a refrigerator (2 to 8°C); if

refrigerated, they should be wrapped in plastic to minimize evaporation. Just before use, if

excess moisture is visible on the surface, plates should be placed in an incubator (35°C) or,

with lids ajar, in a laminar-flow hood at room temperature until the moisture evaporates

(usually 10 to 30 min). At the time that the medium is to be inoculated, no droplets of

moisture should be visible on its surface or on the petri dish cover.

Various components of or supplements to Mueller- Hinton medium may affect susceptibility

test results; therefore, appropriate quality control procedures (see “Quality Control” below)

must be performed and zone diameters must be within acceptable limits. For example, media

containing excessive amounts of thymidine or thymine can reverse the inhibitory effects of

sulfonamides and trimethoprim, causing zones of growth inhibition to be smaller or less

distinct. Organisms may therefore appear to be resistant to these drugs when in fact they

are not. Variation in the concentrations of divalent cations, primarily calcium and

magnesium, affects results of aminoglycoside and tetracycline tests with P.

aeruginosa isolates (21). A cation content that is too high reduces zone sizes, whereas a

cation content that is too low has the opposite effect. Sheep blood should not be added to

Mueller-Hinton medium for testing of nonfastidious organisms, because the blood can

significantly alter the zone diameters with several agents and bacterial species (12).

Inoculation Procedure

To ensure reproducibility of disk diffusion susceptibility test results, the inoculum must be

standardized (7,21). The inoculum may be prepared by the growth method or by direct

suspension from colonies on the agar plate, as described above for dilution testing.

When trimethoprim-sulfamethoxazole is tested by the direct inoculum suspension method,

colonies from blood agar medium may carry over enough trimethoprim or sulfonamide

antagonists to produce a haze of growth inside the zones of inhibition with susceptible

isolates.

The Mueller-Hinton agar plate should be inoculated within 15 min after the inoculum

suspension has been adjusted. A sterile cotton swab is dipped into the suspension, rotated

several times, and gently pressed onto the inside wall of the tube above the fluid level to

remove excess inoculum from the swab. The swab is then streaked over the entire surface of

the agar plate three times, with the plate rotated approximately 60° each time to ensure

even distribution of the inoculum. A final sweep of the swab is made around the agar rim.

The lid may be left ajar for 3 to 5 min but no longer than 15 min to allow any excess surface

moisture to be absorbed before the drug-impregnated disks are applied.

Antimicrobial Disks

Within 15 min after the plates are inoculated, selected antimicrobial agent disks are

distributed evenly onto the surface, with at least 24 mm (center to center) between them.

Disks are placed individually with sterile forceps or, more commonly, with a mechanical

dispensing apparatus and then gently pressed down onto the agar surface to provide uniform

contact. No more than 12 disks should be placed onto one 150-mm-diameter plate and no

more than 5 disks should be placed onto a 100-mm-diameter plate to avoid overlapping

zones. Some of the antimicrobial agent in the disk diffuses almost immediately; therefore,

once a disk contacts the agar surface, the disk should not be moved.

Incubation

No longer than 15 min after disks are applied, the plates are inverted and incubated at 35°C

in ambient air. A delay of more than 15 min before incubation permits excess prediffusion of

the antimicrobial agents. The interpretive standards for nonfastidious bacteria are based on

results of test samples incubated in ambient air, and the zone-of-inhibition diameters for

some drugs, such as the aminoglycosides, macrolides, and tetracyclines, are significantly

altered by CO2; therefore, plates should not be incubated in atmospheres with increased

CO2. Testing isolates of some fastidious bacteria, however, requires incubation in 5% CO2,

and zone diameter criteria for those species have been established on that basis (see chapter

71 of this Manual).

Interpretation and Reporting of Results

Each plate is examined after incubation for 16 to 18 h for all nonfastidious bacterial isolates

except staphylococci and enterococci, which must be incubated for a full 24 h to allow

detection of resistance to oxacillin and vancomycin, respectively (21). If plates are inoculated

correctly, the diameters of the zones of inhibition are uniformly circular and the lawns of

growth are confluent. Growth that consists of individual isolated colonies indicates that the

inoculum was too light, and the test must be repeated. The diameters of the zones of

complete inhibition, including the diameter of the disk, are measured to the nearest whole

millimeter with calipers or a ruler. With unsupplemented Mueller-Hinton agar, the measuring

device is held on the back of the inverted petri dish, which is illuminated with reflected light

located a few inches above a black, nonreflecting background.

The zone margin is the area where no obvious growth is visible with the naked eye. When

isolates of staphylococci or enterococci are tested, any discernible growth (especially a haze

of pinpoint colonies) within the zone of inhibition around the oxacillin disk (for staphylococci)

or vancomycin disk (for enterococci) is indicative of resistance. For other bacteria, discrete

colonies growing within a clear zone of inhibition may indicate testing of a mixed culture that

should be subcultured, reidentified, and retested. However, the presence of colonies within a

zone of inhibition may also indicate selection of high-frequency mutants indicative of

eventual resistance to that agent, e.g., Enterobacter spp. with penicillins and cephalosporins.

WithProteus species, if a thin film of swarming growth is visible in an otherwise obvious zone

of inhibition, the margin of heavy growth is measured and the film is disregarded. With

trimethoprim, the sulfonamides, and combinations of the two agents, antagonists in the

medium may allow some minimal growth; therefore, the zone diameter is measured at the

obvious margin, and slight growth (20% or less of the lawn of growth) is disregarded.

The zone diameters measured around each disk are interpreted on the basis of guidelines

published by the CLSI, and the organisms are reported as susceptible, intermediate, or

resistant (or in some cases nonsusceptible when no resistance breakpoint has been defined)

to the antimicrobial agents tested (Table 3) (22). The clinical interpretation of the categories

of susceptible, intermediate, and resistant has already been provided above under “Dilution

Testing.” Computer programs are available that accompany some automated zone size

reading devices to allow MICs to be derived from the linear regression equation with selected

antimicrobial agents and bacterial isolates (11, 59) (see also chapter 69).

Advantages and Disadvantages

The disk diffusion test has several advantages: (i) it is technically simple to perform and very

reproducible, (ii) the reagents are relatively inexpensive, (iii) it does not require any special

equipment, (iv) it provides susceptibility category results that are easily understood by

clinicians, and (v) it is flexible regarding selection of antimicrobial agents for testing. The

primary limitation of the disk diffusion test is the spectrum of organisms for which it has

been standardized. There have not been adequate studies to develop reliable interpretive

standards for disk testing of bacteria not listed in the CLSI disk diffusion document (21) or

the 2006 CLSI guideline for infrequently isolated or fastidious bacteria (19). It is also

important to note that only certain drugs have been validated for disk diffusion testing

of Stenotrophomonas maltophilia and Burkholderia cepacia(22). The disk test is inadequate

for detection of vancomycin-intermediate S. aureus (21, 73, 74), does not detect daptomycin

resistance in staphylococci and enterococci (39, 55) or colistin resistance in gram-negative

bacilli (38), and in the past was reported to have difficulties in the detection of oxacillinheteroresistant

staphylococci (28) and enterococci with low-level (VanB-type) vancomycin

resistance (66, 74). A potential disadvantage of disk diffusion susceptibility testing is that it

provides only a qualitative result and a quantitative result indicating that the degree of

susceptibility (MIC) may be needed in some cases, e.g., those involving penicillin and

cephalosporin susceptibilities of S. pneumoniae and certain viridans group streptococci

(see chapter 71 of this Manual).

Quality Control

The goals of a quality control program for disk diffusion are to monitor the precision and

accuracy of the procedure, the performance of the reagents (medium and disks), and the

performance of persons who do the testing and read, interpret, and report results. To best

achieve these goals, reference strains are selected for their genetic stability and their

usefulness in the disk diffusion test.

Reference Strains

Reference strains recommended by the CLSI for quality control of the disk diffusion

procedure when nonfastidious bacteria are tested are E. coli ATCC 25922, P.

aeruginosa ATCC 27853, S. aureus ATCC 25923 (not the same strain used for quality control

of MIC tests), Enterococcus faecalis ATCC 29212, and E. coli ATCC 35218 (22, 25). E.

coli ATCC 35218 is recommended as a control only for β-lactamase inhibitor combinations

containing clavulanic acid, sulbactam, or tazobactam. Enterococcus faecalis ATCC 29212 can

be used to ensure that the levels of inhibitors of trimethoprim or sulfonamides in Mueller-

Hinton agar do not exceed acceptable limits and can also be used to control disks containing

a high concentration of gentamicin or streptomycin (seechapter 70 of this Manual).

The reference strains listed above should be obtained from a reliable source, and stock

cultures should be maintained in such a way that viability is ensured and the opportunity for

selection of resistant variants is minimal (25). The procedures for maintaining and storing

working stock cultures are described in the CLSI standard (21). If an unexplained result

indicates that the inherent susceptibility of the strain has been altered, a fresh subculture of

that organism should be obtained.

Zone-of-Inhibition Diameter Ranges

The ranges of zone diameters for reference strains used to monitor performance of the disk

diffusion test are updated frequently and published annually; therefore, readers should refer

to the most recent CLSI document for this information (22). Generally, results of 1 in every

20 tests in a series of tests might be out of the accepted limits. If a second result falls

outside the stated limits, corrective action must be taken. The action taken and the results of

that action must be documented.

Frequency of Testing

Each new batch or lot of Mueller-Hinton agar must be tested with the reference strains listed

above before the medium is released for use with clinical specimens, and quality control

must be done before a new lot of antimicrobial disks is introduced. Appropriate reference

strains also should be tested each day that the disk diffusion test is performed. The

frequency of testing, however, may be reduced if satisfactory performance is documented for

20 or 30 consecutive days of testing: for each combination of drug and reference strain, no

more than 1 of 20 or 3 of 30 zone-of-inhibition diameters may be outside the accepted limits

published by the CLSI (22). When this criterion is fulfilled, each reference strain need be

tested only once per week and any time a reagent component of the test is changed.

However, if the diameter of a zone of inhibition falls outside the acceptable control limits,

corrective action must be taken. If the problem appears to be caused by an obvious error

such as use of the wrong disk or the wrong reference strain, contamination of the reference

strain, or incubation in an incorrect atmosphere, repeating the test with the appropriate

parameter is acceptable. However, if a cause of the error is not obvious, quality control must

be performed daily for a period that will allow discovery of the source of the aberrant result

and documentation of how the problem was resolved. This may be accomplished by the

approach described above under “Quality Control” after the section on dilution testing.

Special Disk Tests

Two specialized applications of the disk test are described in detail in chapter 70. In brief,

disk testing with cefoxitin is now the preferred method for detection of mecA-mediated

oxacillin resistance in both S. aureus and coagulase- negative staphylococcal species

(21, 22, 72). Cefoxitin serves as a surrogate marker for the principal mechanism of oxacillin

resistance in staphylococci and provides more reliable results than oxacillin itself. Secondly,

inducible clindamycin resistance is not reliably detected by standard dilution or disk diffusion

susceptibility testing without induction of the expression of erm-mediated macrolidelincosamide-

streptogramin B resistance in staphylococci and hemolytic streptococci (54).

Such strains can be accurately detected only by induction of resistance expression by

exposure to a macrolide. A disk approximation test in which erythromycin and clindamycin

disks are placed in close proximity allows recognition of inducible resistance by truncating

the clindamycin zone and giving rise to a positive “D-zone test” (22, 36). When recognized

through disk approximation testing, such strains should be reported as resistant to

clindamycin (22, 54). A similar approach may be taken by incorporating a subinhibitory

concentration of erythromycin into broth or agar dilution tests with clindamycin. Broth-based

tests are available on several of the automated susceptibility testing systems.

ANTIBACTERIAL SUSCEPTIBILITY TESTING AND

INTERPRETIVE METHODS USED OUTSIDE THE UNITED

STATES Back to top

The CLSI is best known for developing laboratory testing standards for use in the United

States, including those for antimicrobial susceptibility testing (20–22). The CLSI standards

are recognized as U.S. national standards by the American National Standards Institute and

by federal regulations, including the Clinical Laboratory Improvement Amendments (40), and

as standard reference procedures by the Food and Drug Administration. However, the CLSI

procedures are also used by an increasing number of laboratories outside the United States,

including countries in North and South America and in several areas of Europe, Asia, and

Australia. Some countries have national standards or professional committees comprising

their own expert microbiologists that establish methods of susceptibility testing for their

countries and interpretive criteria for those tests that may or may not be the same as those

of the CLSI (15, 35), as described further in chapter 72.

Several variations on dilution and diffusion methods are used for routine susceptibility testing

outside the United States (Table 4). Most non-U.S. methods are specific to individual

countries, having developed and evolved locally over many years. Many of the non-U.S.

methods differ from the CLSI procedures in the choice of media, inoculum preparation

procedures, and, for diffusion methods, disk contents. There are also some variations

between these methods in breakpoints and the approaches to establishing the breakpoints

(78). Variation in test methods can cause considerable confusion in laboratories, especially if

both CLSI and non-CLSI methods are used for different organisms. Thus, it is important that

any method be followed in all its detail for valid application of specific breakpoints.