Enterococcus faecalis and Penicillin Susceptibility Testing: The Challenge of Multiple Methods and Agent-to-Agent Predictions
Year: 2018
Type: article
Abstract: To the Editor: Enterococci are gram-positive aerobic bacteria that elicit various opportunistic infections. The most commonly encountered species Enterococcus faecalis is typically susceptible to cell wall active agents (including noncephalosporin β-lactams and vancomycin), many of which represent therapeutic mainstays. Nevertheless, empiric testing is needed to confirm the susceptibility profile of any given isolate. The CLSI has established guidelines for enterococci, to determine the minimum inhibitory concentrations (MICs) of antibiotics by using reference methods (broth or agar dilution) and their interpretation as susceptible or resistant (1). While most ampicillin-susceptible E. faecalis isolates are susceptible to penicillin, this association cannot be assumed, and explicit penicillin testing is required. Guidelines do stipulate that ampicillin results for non-β-lactamase–producing E. faecalis (the vast majority) predict the activity of amoxicillin, imipenem, and piperacillin-tazobactam (1), the latter of which is commonly used as a broad-spectrum intravenous agent. Nevertheless, these guidelines' robustness has been challenged by observations that ampicillin results do not predict piperacillin for the subset of ampicillin-susceptible/penicillin-resistant strains, which further emphasizes the value of penicillin testing (2). In practice, the antibiotics that a laboratory analyzes and reports for any given organism are a function of clinical needs and the antimicrobial susceptibility testing (AST)1 methodology in local use. Often, this technique is not a CLSI reference method, but instead it is a commercial platform with regulatory approval as an in vitro diagnostic, whose performance emulates reference methods. Our laboratory currently performs susceptibility testing for gram-positive aerobes by using the BD Phoenix-PMIC 106 cartridge, whose concentration range for penicillin does not allow interpretations for Enterococcus; explicit results are likewise not provided for piperacillin-tazobactam and must be inferred from ampicillin. Therefore, in clinical practice, our laboratory had conducted supplemental penicillin testing on ampicillin-susceptible E. faecalis isolates by using Etest gradient-diffusion strips (Biomerieux). Etest is a longstanding agar-based technique that is commonly used to measure MICs for ad hoc antimicrobials. The benzylpenicillin Etest carries in vitro diagnostic designation in the US; its indicated use broadly includes gram-positive aerobes, and its product insert cites CLSI breakpoints for Enterococcus (3). Over 12 months, 182 ampicillin-susceptible E. faecalis isolates from invasive sites of unique patients (MIC ≤8 μg/mL by Phoenix) were subjected to penicillin Etest, performed and read per explicit manufacture instructions (0.5 McFarland swab inoculum on Mueller–Hinton agar; read at complete inhibition after 16–20 h, 35 °C). In total, 26 isolates demonstrated penicillin MICs by Etest that correspond to “resistant” by CLSI criteria (≥16 μg/mL) (Table 1); concurrent quality controls demonstrated expected results. This number of ostensibly ampicillin-susceptible/penicillin-resistant strains was unanticipated, drawing our attention to a single previous editorial communication that stated categorical penicillin resistance can be overestimated for E. faecalis by Etest (4). Of note, the latter work did not comment on concomitant piperacillin-tazobactam susceptibility results. Artifactual penicillin resistance in E. faecalis was likewise recently reported for a different commercial testing platform (5). In light of these studies, we sought to determine whether a similar phenomenon might be at play with our isolates and, going further, to more completely probe the relationship between penicillin Etest results and piperacillin-tazobactam susceptibility. Antimicrobial susceptibility results for penicillin (Etest and BMD) and piperacillin-tazobactam (Etest only) for 24 ampicillin-susceptible/Etest-penicillin resistance isolates of E. faecalis, including MICs and susceptibility interpretations as described in the text. R, resistant; S, susceptible; WT, wild type. For 8 of these 24 strains, the penicillin Etest result was documented by the performing technologist as “resistant, MIC ≥ 16,” although the specific observed MIC (16, 32, or >32) was not indicated. Antimicrobial susceptibility results for penicillin (Etest and BMD) and piperacillin-tazobactam (Etest only) for 24 ampicillin-susceptible/Etest-penicillin resistance isolates of E. faecalis, including MICs and susceptibility interpretations as described in the text. R, resistant; S, susceptible; WT, wild type. For 8 of these 24 strains, the penicillin Etest result was documented by the performing technologist as “resistant, MIC ≥ 16,” although the specific observed MIC (16, 32, or >32) was not indicated. Following internal review board approval, broth microdilution (BMD) susceptibility testing was performed for penicillin and piperacillin-tazobactam at ARUP Laboratories per CLSI reference methodology (1). Along with 24 Etest-penicillin-resistant strains (2 were unavailable), 17 Etest-susceptible strains were included as comparators. Overall, all 24 Etest-penicillin-resistant strains demonstrated a susceptible penicillin MIC by BMD testing (categorizing the Etest results as “major errors” per standard CLSI definitions) (Table 1), as did 17/17 comparators. By considering all strains analyzed by either method, we can compare our actual reported results (for Etest) to hypothetical results (had BMD instead been used as our method-of-choice for the selected strains): 26/182 (14.3%; 95% CI, 9.8%–19.9%) and 0/41 (0%; 95% CI, 0%–5.9%) strains would be considered penicillin-resistant, respectively. Furthermore, 24/24 Etest-penicillin-resistant strains and all comparators demonstrated a piperacillin-tazobactam MIC ≤16/4 μg/mL by BMD. While an Enterococcus breakpoint for piperacillin-tazobactam has not been established, this MIC-range is considered wild type by European Committee on Antimicrobial Susceptibility Testing epidemiologic cutoff values (16/4 μg/mL; https://mic.eucast.org). Our observations therefore reinforce the inability of the penicillin Etest to categorically match BMD results for E. faecalis in clinical practice, at least in conjunction with CLSI breakpoints. And we caution against using these results to preclude therapy with penicillin or piperacillin-tazobactam (as occurred here), at least without replicate testing by preferably reference methods. Beyond this species, our experience highlights how surrogate AST methods—even techniques with formal regulatory approval—do not emulate reference results for all organism-agent combinations. We urge particular caution when a product's stated indication includes massive categories of organisms (e.g., “gram-positive aerobes”) whose member species could demonstrate disparate performance. Clinical consequences can be magnified when interpretations for one antibiotic are then used to predict others. antimicrobial susceptibility testing minimum inhibitory concentration broth microdilution.
Institutions: Vanderbilt University, Vanderbilt University Medical Center
Language: English
Subfield: Clinical Biochemistry
Domain: Life Sciences
Open Access status: bronze
OpenAlex ID: W2902610524