Professor of Pathology, Microbiology, and Immunology
Medical Director of the Microbiology Laboratory
Vanderbilt University Medical Center
Every year, the Clinical and Laboratory Standards Institute (CLSI) publishes an update to the standards, “Performance Standards for Antimicrobial Susceptibility Testing”, otherwise known as the M100.1 Globally, laboratories refer to this document as the reference for how to perform and interpret antimicrobial susceptibility tests (AST). My own laboratory diligently prints the document out upon publication, lovingly inserting each page in a page protector and displaying the document in handy proximity to culture benches, for frequent reference. It would be a lie to say that this ritual is not performed with some trepidation: what new protocol or breakpoint will we have to implement or adapt to, this year? I have learnt that not all technologists in the laboratory get quite the same satisfaction as I, pondering the nuances of antimicrobial susceptibility testing nor admiring the wily complexity of the bacteria we deal with day in and day out in the laboratory!
With this in mind, let’s briefly review the major changes to the M100, 32nd edition. I promise, these are not as intimidating as they may appear upon first review of the index of changes: updates include 3 major breakpoint updates and some expansion of the direct-from-blood culture disk diffusion (DD) method.
Arguably, the most impactful breakpoint change in M100 S32nd edition is an update to the piperacillin-tazobactam (TZP) MIC and DD breakpoint for the Enterobacterales. TZP is a frequently prescribed drug for Gram-negative infections and is tested and reported on virtually every isolate of Enterobacterales encountered by the clinical laboratory. TZP breakpoints were first published by CLSI in 1993 and since then, the number and prevalence of extended-spectrum beta-lactamases (ESBLs) produced by Enterobacterales has increased exponentially. Mounting evidence suggested that the TZP breakpoints were in need of a significant revision. Specifically, clinical data indicated TZP MICs >16 µg/mL were associated with increased risk of 30-day mortality for serious infections, and extensive pharmacokinetic-pharmacodynamic (PK/PD) data suggested low probability of attaining target exposures at MICs >8 µg/mL.
One study in particular, the MERINO trial, provided a clear signal of patient safety concerns with the M100 S31st Ed breakpoint. In this study, patients with bacteremia caused by ceftriaxone-intermediate or -resistant Escherichia coli or Klebsiella pneumoniae were randomized to treatment with meropenem or TZP. The study was halted early, as a clear safety signal demonstrated that TZP failed to demonstrate non-inferiority to meropenem, with a risk difference of 8.6% for 30-day mortality in the TZP arm in primary analysis and a risk difference of 6.8% in the per-protocol analysis2. Interestingly, post hoc evaluation of the study isolates showed that many were TZP-resistant when evaluated by reference broth microdilution (BMD) but had tested susceptible by the local laboratory’s routine AST methods. The authors demonstrated that this effect was far more common when the isolate co-expressed an ESBL and OXA-1 beta-lactamase than if the isolate harbored an ESBL alone.3 OXA-1 is an enzyme present in 30-60% of global ceftriaxone-resistant E. coli and K. pneumoniae,5 but has only weak affinity for tazobactam, meaning isolates with OXA-1 have elevated TZP MICs of 8-32 µg/mL, right at the historical susceptible/intermediate breakpoint – making assessment of susceptibility challenging, even by reference methods.4 When the authors evaluated clinical outcome by reference BMD MIC, a clear cut-off at an MIC of ≤16 µg/mL (91% survival, n=147 patients) versus ≥32 µg/mL (50% survival, n=10 patients) was established.3 However, as noted, isolates with OXA-1 are challenging to test and may yield variable MICs of 8 – 32 µg/mL, depending on the test method. As such, a result of 16 µg/mL generated by a typical clinical laboratory could be 32 µg/mL by reference testing, putting the patient at increased risk of treatment failure, and highlighting the value of a “buffer zone” category (e.g., intermediate) to account for this inherent testing variability.1
Extensive studies have been conducted to evaluate the PK/PD of TZP.5 These demonstrate that exposures, using standard FDA-approved label doses of TZP, are best for isolates with MICs ≤8-16 µg/mL. Exposure is improved if an extended infusion of the antimicrobial is given, as this keeps the serum concentration higher for a longer duration, resulting in more time at or above a concentration of 16 µg/mL than is possible with standard, 30-minute infusion times. Many hospitals have moved towards extended infusions of TZP for this reason, and led the CLSI AST Subcommittee to introduce a susceptible, dose-dependent category for TZP, based on a dose of 4.5 g q 6h as a 3-h infusion or 4.5 g q8 h as a 4-h infusion.1 This susceptible dose, dependent category also serves as the buffer zone for testing inaccuracies described above.
Laboratories should discuss immediate implementation of the CLSI Enterobacterales TZP breakpoint with their antibiotic stewardship program (ASP), as the update provides a significant improvement to predicting treatment outcome for TZP. However, by U.S. regulations, commercial manufacturers cannot implement the change to their test systems until such a time as the FDA recognizes the breakpoint. CLSI has submitted a rationale document to the FDA for consideration5, but evaluation of the breakpoint typically takes >6 months by the FDA. Laboratories may implement the breakpoint by performing a validation study for the breakpoint on their commercial AST systems and implementing this as a modification to their tests. Isolates available through the Center for Disease Control AR Bank6 may be a source of isolates for this update, as TZP reference BMD is available for those isolates. Further guidance on validation studies for this change can be found in the CLSI M52 guideline.7
A second notable breakpoint change includes updated cefiderocol MIC and DD breakpoints for Stenotrophomonas maltophilia. As a reminder, cefiderocol is a novel cephalosporin that chelates iron, entering the bacterial cell via active iron transport systems. Like other beta-lactams, cefiderocol binds penicillin-binding proteins (PBP3 in this case) and disrupts cell wall synthesis. Uniquely, cefiderocol is active against a variety of multi-drug resistant bacteria, including S. maltophilia. Prior to 2022, CLSI published investigational breakpoints for S. maltophilia based on in vitro cefiderocol MIC data and animal model data provided by the sponsor, Shionogi. In 2021, results of clinical studies, including APEKs-cUTI, APEKS-NP, CREDIBLE-CR and compassionate use cases, along with clinical PK/PD analyses were presented to the CLSI subcommittee.8 Very little clinical data are available for S. maltophilia due to the relative infrequency of infections caused by this organism. Furthermore, interpretation of clinical data is complex as patients who acquire S. maltophilia are normally very compromised (e.g., patients with hematological malignancies), making assessment of outcome due to infection versus underlying disease challenging, if not impossible. S. maltophilia is often isolated as part of a polymicrobial infection in respiratory specimens and it is often unclear in this context is the organism represents colonization or true infection.9
Nonetheless, in cases of true infection, S. maltophilia is difficult to treat due to its intrinsic multi-drug resistance, including against other beta-lactams due to the co-expression of two beta-lactamases: L1 (a metallo-beta-lactamase) and L2 (a serine beta-lactamase). Cefiderocol MICs are normally < 1 µg/mL, well below the 4 µg/mL that is predicted to be associated with good PK/PD target attainment. It should be noted that PK/PD assessments, while a valuable tool, do not replace clinical data and after much deliberation, a conservative breakpoint for S. maltophilia, at the epidemiological cut-off (i.e., wild type) MIC of ≤1 µg/mL was established. No intermediate or resistant categories were set, due to the rare occurrence of isolates with MICs above 1 µg/mL.1
The final major breakpoint adjustment to the M100, 32nd edition is revision of the amoxicillin-clavulanate breakpoint for Haemophilus spp. As background to this change, an ad hoc working group for the CLSI AST Subcommittee has undergone a systematic review of the aminopenicillin breakpoints over the course of the last several years. These include, ampicillin, ampicillin-sulbactam, amoxicillin and amoxicillin-clavulanate. Complicating this review is the myriad of clinical indications and doses for these antimicrobials and the near absence of any good clinical data. Ultimately, the working group was able to re-affirm nearly all the breakpoints and added the doses against which the breakpoints were assessed to the current version of M100. Review of the amoxicillin-clavulanate breakpoint for H. influenzae and H. parainfluenzae determined that an MIC of 4 µg/mL (susceptible by the M100, 31st edition breakpoint) was not achievable with oral dosing of amoxicillin-clavulanate (the only formulation available in the U.S.), whereas at an MIC of 2 µg/mL target attainment was improved, resulting in an adjustment of the susceptible breakpoint to ≤2 µg/mL. Of note, no data were available by which to set a DD breakpoint resulting in deletion of the DD breakpoint for these organisms in the M100, 32nd edition.
Laboratories that perform testing of Haemophilus by DD may consider dual use of an ampicillin disk and beta-lactamase test such as the nitrocefin disk method. Isolates that are susceptible by DD to ampicillin are predicted to be susceptible to amoxicillin and amoxicillin-clavulanate (i.e., not resistance mechanism). Isolates that are resistant to ampicillin by DD and negative for beta-lactamase are predicted to be resistant to amoxicillin-clavulanate, as the resistance mechanism in this case is mutation to the PBP, not a beta-lactamase.1 Isolates that are resistant to ampicillin and positive for beta-lactamase are most likely susceptible to amoxicillin-clavulanate, as the most common resistance mechanism associated with this phenotype is a beta-lactamase. However, rare isolates may harbor both beta-lactamase and PBP mutation, which would result in resistance to amoxicillin-clavulanate. This possibility should be discussed with the ASP and clinicians, to either inform laboratory policy, or on a case-by-case basis if amoxicillin-clavulanate results are needed for a given Haemophilus influenzae.
Additional changes to breakpoints include adjustments to several DD breakpoints: Enterobacterales cefiderocol, Acinetobacter spp. cefiderocol, Enterobacterales ceftolozane-tazobactam, H. influenzae lefamulin and S. pneumoniae lefamulin. These changes occurred as new data demonstrate the previous DD breakpoints yielded errors when compared to MIC results, and should be relatively easy to implement for those laboratories routinely using these DD tests.
Speaking of DD, the M100 S32nd edition includes an expansion to the breakpoints available for the direct-DD method for positive blood cultures. Major updates to this method include addition of breakpoints for Pseudomonas aeruginosa, and inclusion of breakpoints for some antimicrobials to be read at 8-10 hours, as opposed to the standard incubation time to 16-18 h.
As a refresher, the direct DD method is a method optimized by CLSI to enable use of positive blood culture broth as the inoculum for DD testing, as opposed to the standard method of preparing an inoculum from colony growth from subculture. This method is possible as most blood cultures flag positive at a bacterial concentration similar to that of a 0.5 McFarland standard, which is used for routine DD, and are often mono-microbial. The method is straightforward, by design, as the intent is to provide a more rapid time-to-results for AST accessible for all patients with bacteremia, a condition where each hour on ineffective therapy is associated with increased mortality.10 By the CLSI method described in M100, a positive blood culture displaying Gram-negative bacilli is tested within 8 hours of flagging positive. The culture is inverted to mix, and a 20-guage venting needle is inserted into the bottle’s septum. Four drops of blood culture broth are dispensed to a Mueller Hinton agar plate, and spread across the surface, as one would do for a standard DD test. Disks are applied, and the plate is incubated at 35 +/- 2 C for 8-10 hours, at which point a first read may be done for some drugs, and a final 16-18 hours, at which the remaining DD results may be evaluated. Further details on the procedure are available in the M100, S32nd edition, Table 3E.1
Importantly, the results can only be interpreted once organism identification is available: either by standard methods, or rapid identification methods like molecular tests or “scum” growth MALDI-TOF mass spectrometry. Additionally, if the culture is found to be mixed, the results cannot be interpreted. CLSI is actively reviewing additional antimicrobial agents for the direct blood culture DD method. Use of rapid methods for AST from blood cultures is becoming standard of care, not only to identify patients for whom antimicrobials must be escalated due to resistance, but also for the majority of patients, where antimicrobials can be narrowed and/or optimized based on AST results. The CLSI method provides an inexpensive and practical method for all laboratories to improve care of patients with sepsis and should be considered by those laboratories with no rapid AST methods for blood cultures.
In summary, a practical approach to implementing updates presented in the M100 32nd edition may be to start with implementation of the revised TZP breakpoints – this change is likely to be the most impactful to patient care at your institution. Laboratories should have cefiderocol testing available, either in-house, or as a pre-planned approach at a reference laboratory, as this antimicrobial is a drug of last resort for many multi-drug resistant organisms. If susceptibility testing for Haemophilus is performed in your laboratory, evaluate how to implement the amoxicillin-clavulanate breakpoint, either through an MIC method, or through use of an ampicillin-DD/nitrocefin approach, as described above. Minor changes to cefiderocol, ceftolozane-tazobactam and lefamulin DD breakpoints can be implemented with minimal verification/validation, if this test method is used by your laboratory.7 Finally, if not already established, laboratories should discuss rapid AST needs for patients with bacteremia, in conjunction with ASP programs, infectious diseases colleagues, pharmacists and other vested clinicians at your institution, as rapid detection of antimicrobial resistance (and susceptibility!) has the potential to significantly impact patient management and clinical outcomes.