Breakpoints are a critical component of modern microbiological testing, used to define the susceptibility and resistance of bacterial strains to antimicrobial agents. Depending on the assay method, breakpoints may be expressed as concentrations (mg/L or µg/mL) or as inhibition zone diameters (mm). In general, all antimicrobial susceptibility tests require breakpoints—also referred to as interpretive criteria—to classify results as susceptible, intermediate, or resistant and communicate these findings to clinicians. While some experienced clinicians may prefer MIC values and pharmacokinetic–pharmacodynamic parameters over breakpoints to optimize antimicrobial selection and dosing, given the large volume of specimens processed by clinical microbiology laboratories and the varying levels of expertise among the clinicians they serve, categorical interpretation of susceptibility results remains practical and more readily accepted by the majority of clinicians.

The term “breakpoint” has multiple interpretations in the literature. The first interpretation refers to the MIC value used to distinguish between wild-type bacterial populations and those exhibiting acquired or selective resistance (“wild-type breakpoint,” sometimes also called the microbiological breakpoint). Data for such breakpoints are derived from medium- to large-scale in vitro MIC studies that adequately characterize the wild-type population. Wild-type strains are defined as those that do not harbor any acquired or selectively induced resistance to the test agent or to agents with the same mechanism of action.

The second type of breakpoint is referred to as the clinical breakpoint; it is used to distinguish between pathogenic microorganisms with a favorable prognosis and those associated with treatment failure. This breakpoint is derived from prospective clinical studies in infected patients, established by comparing the clinical outcomes of pathogens with different MICs.

The third type of breakpoint is the drug concentration calculated using pharmacodynamic theory and pharmacodynamic parameters that predict in vivo drug activity; this is referred to as the pharmacokinetic/pharmacodynamic (PK/PD) breakpoint. Such data are derived from animal models and extrapolated to humans through mathematical or statistical methods. To avoid confusion arising from the multiple meanings of “breakpoint,” the European Committee on Antimicrobial Susceptibility Testing (EUCAST) has replaced the term “microbiological breakpoint” with a new term, “epidemiological (or wild-type) breakpoint.” Today, the term “breakpoint” is more broadly used to describe three distinct types of breakpoints—such as the wild-type breakpoint, the PK/PD breakpoint, and the clinical breakpoint—while the word “breakpoint” itself is reserved for the final report issued by clinical laboratories.

Mechanism for setting the inflection point

The process of establishing breakpoints varies considerably across different antimicrobial susceptibility testing methods, and this process is often not explicitly described in the corresponding method protocols. For methods that provide breakpoints but do not explain how they were determined, we may assume that the classification of susceptible, intermediate, and resistant is based on wild-type cutoff values.

Only two international standard-setting organizations—the Clinical and Laboratory Standards Institute (CLSI, formerly known as NCCLS) and the European Committee on Antimicrobial Susceptibility Testing (EUCAST)—have published guidelines on data sources and data‑use criteria for breakpoint determination. Table 1 compares the susceptibility testing methods and the parameters required for breakpoint setting employed by CLSI and EUCAST. The U.S. Food and Drug Administration also establishes breakpoints; harmonizing these breakpoints will be a goal for the future.

Data required to set the inflection point

The breakpoint‑setting process requires a comprehensive set of data to define breakpoints, including in vitro microbiological data, pharmacokinetic/pharmacodynamic data from animal and human studies, and clinical/bacteriological outcome data. No single data source can provide all the information needed for decision‑making.

The following four types of data are essential for setting the correct breakpoints:

① MIC distribution and the cutoff values for wild-type strains;

② In vitro resistance markers, including phenotypic and genotypic ones;

③ PK/PD data derived from animal models and human studies;

④ Derived from appropriate clinical studies and clinical and bacteriological outcome data based on pathogen MICs. To establish disk diffusion breakpoints, we need to establish a linear relationship between inhibition zone diameters and MICs.

The most critical—and also the most challenging—aspect of using these data is determining how to strike the optimal balance among diverse datasets, tailored to specific pathogens and the corresponding types and severities of infections. At present, no universal formula exists to quantify which data are most relevant in any given clinical scenario. In practice, final decisions are reached through consensus within the standard‑setting committee. This approach is inherently difficult to reconcile, as scholars differ in their breadth and depth of knowledge, leading to varying assessments of the relative importance of different data types. Consequently, it is essential that all members of the breakpoint‑establishment committee possess a comprehensive understanding of all four categories of data mentioned above. Since infections can occur in various anatomical sites, in theory one would assign distinct breakpoints for each infection type—such as bloodstream infections, cellulitis, meningitis, lower urinary tract infections, osteomyelitis, pneumonia, and so on. However, this would substantially increase the complexity of breakpoint determination; therefore, nearly all methods rely on a single set of breakpoints, with occasional adjustments only when drug concentrations at the site of infection vary markedly—for example, in urinary tract infections or meningitis. Generally, bloodstream infections are regarded as the most common form of complicated infections; accordingly, pharmacokinetic parameters of the drug in blood are used to establish the breakpoints.

MIC distribution and wild-type breakpoints

Establishing the MIC distribution for each strain–antimicrobial agent combination is the first step in setting breakpoints. The MIC of the drug can be determined using broth or agar dilution methods, and a MIC distribution histogram can be generated, as shown in Figure 1. The histogram provides a visual indication of whether wild-type strains or strains with abnormally elevated MICs are present.

Figure 1. MIC distribution of ampicillin against Escherichia coli. EUCAST has published wild-type MIC distribution histograms for numerous bacterial–drug combinations, which are available on its website. www.eucast.org These data were obtained from large‑scale domestic and international studies. The breakpoint is generally not set within the wild‑type MIC range to prevent strains lacking resistance mechanisms from being classified as resistant. COWT: the critical value separating wild‑type and resistant isolates based on MICs.

Phenotypic and Genotypic Resistance Markers

In addition to MIC determination, strains are often assessed for resistance mechanisms using other approaches, including phenotypic and genotypic assays. If simple, specific phenotypic assays (beyond MIC determination) or genotypic assays can detect acquired resistance, they may be employed to validate wild-type breakpoints. Strains with MICs near the wild-type breakpoints should be subjected to these specialized phenotypic or genotypic tests to evaluate their relevance. When necessary, wild-type breakpoints should be adjusted in light of the underlying resistance mechanisms.

PK/PD parameters

Pharmacokinetics (PK) and pharmacodynamics (PD) are two essential components of pharmacology. PK describes the absorption, distribution, metabolism, and excretion of a drug in the body, while PD examines how drug dosage influences therapeutic efficacy and the clinical outcomes of treatment. Key PK/PD parameters include T > MIC, Cmax/MIC, and AUC/MIC (AUIC), as shown in Figure 2.
Ⅰ. T > MIC: Refers to the duration during which the plasma drug concentration remains above the MIC after administration. It is typically expressed as a percentage of the dosing interval.
II. Cmax/MIC: The ratio of the peak plasma concentration of an antimicrobial agent to its minimum inhibitory concentration (MIC).
III. AUIC: Refers to the ratio of the 24-hour AUC to the MIC in the plasma concentration–time curve.

Antimicrobial PK/PD studies integrate pharmacokinetic and in vitro pharmacodynamic parameters. Based on the relationship between antimicrobial activity and plasma drug concentrations or exposure duration, these agents can generally be classified into three categories: concentration‑dependent, time‑dependent (with a short post‑antibiotic effect, PAE), and time‑dependent (with a long PAE). This classification provides important theoretical guidance for optimizing dosing regimens for different antimicrobial agents, as shown in the table.

Clinical research data

Some organizations devote substantial effort to prospective clinical studies aimed at establishing breakpoints. The advantage of such studies is that they minimize the use of multiple antibiotic classes, allowing for the independent evaluation of each individual antibiotic’s activity. However, their drawback is:

① Most studies pre-defined “drug-resistant strains” and excluded patients with infections caused by such strains.

② For certain bacterial species, infections caused by strains with higher MICs are underrepresented in the case count, thereby increasing the bias in clinical breakpoints.

③ Currently, there is no consensus on the cure rate or bacterial clearance rate, which has resulted in the inability to standardize the endpoints used to validate clinical trial outcomes.

Revision of the breakpoint and its implications for clinical practice

In 2010, CLSI revised the breakpoints for cefazolin, ceftazidime, cefotaxime, ceftizoxime, ceftriaxone, and aztreonam against Enterobacteriaceae, and added corresponding dosing regimens following these breakpoints; see the table.

The revised breakpoints are based on microbiological, pharmacological, and clinical data. Many of the breakpoints currently in routine clinical use date back 25 years and would be unacceptable by today’s regulatory and quality‑assurance standards. The updated breakpoints in M100‑S20 more accurately reflect the true efficacy of antimicrobial agents when treating infections caused by specific bacterial strains according to currently recommended regimens. Findings from studies on ESBL‑producing strains played a pivotal role in this revision of the breakpoints. Initially, CLSI recommended performing screening and confirmatory testing for ESBLs and stipulated that, for ESBL‑producing isolates, susceptibility results for penicillins, cephalosporins, and aztreonam should be reclassified from “susceptible” to “resistant.” This guideline was based on the following considerations:

① The study observed that, for certain ESBL-producing strains, the MICs of the aforementioned agents were elevated but remained within the susceptible range (using the old breakpoints).

② Limited clinical observations have indicated that patients with infections caused by ESBL-producing strains tend to have a poorer prognosis.  

The ESBL assay is recommended as a short-term solution for addressing a single emerging resistance mechanism. Subsequently, additional resistance mechanisms—such as novel ESBLs and AmpC enzymes—have been identified, and an increasing number of strains produce multiple enzymes, further complicating ESBL detection.

These facts, together with the growing understanding that PK–PD parameters of cephalosporins and monobactams play a decisive role in treatment outcomes, have led to revisions of the breakpoints. Following these updates, ESBL screening and confirmatory testing will no longer be required to guide therapeutic decisions. The MIC values of bacterial isolates correlate more strongly with clinical prognosis than do the resistance mechanisms they harbor. CLSI believes that the new breakpoints will provide clinicians with more reliable guidance for patient management while reducing uncertainty and workload in clinical laboratories.

In addition, the breakpoints for intravenous cefuroxime, cefepime, cefotetan, and cefoxitin against Enterobacteriaceae have also been reassessed but remain unchanged. Cefoxitin and cefotetan belong to the cephamycin class; they are not hydrolyzed by ESBLs, and a positive confirmatory test for ESBLs does not necessitate reclassifying these agents as resistant. The breakpoint for cefepime has not been revised, based on clinical trial data and PK–PD assessments.

Currently, several clinical trials have demonstrated that cefepime is effective in treating patients infected with ESBL-producing strains that remain susceptible to the drug (MIC ≤ 8 µg/mL). Pharmacokinetic–pharmacodynamic analyses indicate that a daily cefepime dose exceeding 3 g—specifically, 1 g every 8 hours or 2 g every 12 hours—can achieve target exposure levels corresponding to the established breakpoints. Furthermore, retrospective data suggest that no revision of the current breakpoints for cefuroxime (for injection) is necessary; however, it should be noted that these breakpoints apply only to doses of 1.5 g every 8 hours or higher.

The CLSI’s revision of the susceptibility breakpoints for Enterobacteriaceae to certain cephalosporins represents a significant milestone in the ongoing updating of antimicrobial susceptibility standards. Consequently, the question of how to define breakpoints for other antimicrobial agents—such as those for carbapenems, β‑lactam/β‑lactamase inhibitor combinations in Enterobacteriaceae, and the susceptibility breakpoints for Pseudomonas aeruginosa—remains an area requiring continuous refinement and revision.

As evolving resistance mechanisms, bacterial species distribution, and our understanding of pharmacokinetic–pharmacodynamic parameters underlying clinical response continue to advance, breakpoints will become increasingly rationalized across different types of infections, patient populations, and dosing regimens.