Verification Is an Integral Part of Antimicrobial Susceptibility Test Quality Assurance

LETTER

With interest we read the commentary of Kirby and colleagues, which advocated implementing antimicrobial susceptibility testing (AST) for new drugs without verification (1). We agree with the authors that new drugs should not be used clinically without an accompanying AST. Implementing AST for new drugs as soon as possible is a critical function of the clinical laboratory. However, we disagree that verification studies beyond quality control (QC) testing are superfluous. We posit that some form of verification is warranted when these tests are implemented to ensure that the laboratory is in compliance with national regulations, i.e., the Clinical and Laboratory Improvement Amendments (CLIA), which require verification of new test systems prior to their use for patients; far more importantly, the laboratory must have confidence that accurate results are obtained with the device. The CLIA view the addition of additional analytes to a test system (e.g., for AST or new drugs) as a unique test system that requires verification (2).

Many laboratories are paralyzed by the prospect of performing a verification study for AST, for the reasons outlined by Kirby: time and financial resources are restricted, organisms can be difficult to obtain, and guidance is lacking. Nonetheless, the answer to these hurdles is not to bypass the verification process as a whole. Rather, a risk-based approach to inform the extent of verification activities should be done, taking the organisms to be tested, the laboratory’s experience in performing the test for other drugs, and the availability of published information on the new drug AST into consideration. While the CLSI suggests the use of 30 isolates, this is a guideline, and the extent of testing is at the director’s discretion. Testing isolates that harbor resistance profiles targeted by the new drug and MICs near the clinical breakpoint may be helpful. An ongoing quality assurance program implemented alongside AST adds additional confidence, ensuring routine investigation of unexpected resistance or susceptibility.

Several factors other than those identified by Kirby et al. might impact an AST’s performance, new disks, and gradient strip performance: (i) the use of Mueller-Hinton agar of a brand different from that used in the studies conducted by the manufacturer for U.S. Food and Drug Administration (FDA) clearance, (ii) nuanced alteration of the AST method defined in the package insert (e.g., an extension of the incubation period for isolates with poor growth), (iii) testing organisms in a manner other than within the drug’s indications for use and for which the AST was not evaluated by the manufacturer (e.g., testing a member of the Enterobacterales that is not claimed in the drug label), (iv) application of clinical breakpoints to interpret the results that are different from those applied at the time of FDA clearance for the device (e.g., the use of CLSI breakpoints), (v) the prevalence of an antimicrobial resistance mechanism not well represented in the AST clearance studies, and (vi) differences in the quality of disk or gradient strip materials and lot-to-lot variability (3).

We caution that quality control (QC) testing does not sufficiently evaluate these variables. QC ranges are typically much lower than the clinical breakpoint. For example, the meropenem-vaborbactam QC ranges span from 0.008 to 0.5 μg/ml for Enterobacterales QC organisms. The Enterobacterales clinical breakpoints are ≤4 μg/ml (susceptible), 8 μg/ml (intermediate), and ≥16 μg/ml (resistant), several dilutions away from these QC ranges (4).

QC testing failed to identify issues with the piperacillin-tazobactam Etest used in the MERINO trial. Subsequent studies demonstrated that very major errors (VMEs, false susceptibility) were associated with patient mortality. The OXA-1 resistance mechanism was common in this study, associated with VMEs by the test, and not evaluated in the clinical trials for Etest (5, 6). Another example is a retracted paper on the performance of the ceftolozane-tazobactam Etest. In this case, the authors compared the Etest to reference broth microdilution, which was performed using drug obtained from the pharmacy and serially diluted, leading to an inappropriate concentration of tazobactam in the test and false resistance by the reference method (7, 8). Again, QC testing was within acceptable limits. Furthermore, QC does not evaluate all aspects of testing (i.e., preanalytical, analytical, and postanalytical components), whereas a verification allows the opportunity to confirm all components, from testing to interpretation and reporting.

AST is complex, with annual updates to standards and the emergence of new resistance mechanisms and new drugs, all of which much be addressed by the laboratory. Unfortunately, this critical function of the clinical microbiology laboratory is undervalued, from a reimbursement perspective (a typical Center for Medicare Services reimbursement is ∼$15 per test), an urgency perspective (as evidenced by the continued lack of FDA-cleared ASTs for current breakpoints and slow adoption of ASTs for new drugs by laboratories), and a technical competency perspective. Continuing to lower the bar on this critical function in the laboratory by eliminating verification studies will not serve to address the antimicrobial resistance crisis. Implementation of FDA-cleared devices for AST of new drugs is paramount to ensure that these drugs are used for the patients that so desperately need them. Laboratories should not allow the technical challenges of implementing these tests to prevent their adoption. We therefore call on leaders in our field to work with standards development organizations and regulatory and accreditation bodies to provide accessible, simplified, and risk-based tools for laboratories to implement AST devices. Many laboratories do not benefit from a full-time clinical microbiology medical director who can advocate for the resources needed to correctly implement these tests. We urge action toward a national plan that includes improved reimbursement for new drug ASTs and incentives for adoption. In addition, such action would help incentivize national reference laboratories to adopt ASTs for new drugs in a timely manner, with turnaround times that are in line with patient care needs. It is certain that antimicrobial resistance is portable, and coordinated efforts are needed to ensure that antimicrobials continue to be a national resource.