Evaluation study of using ampicillin susceptibility to predict imipenem susceptibility of E. faecalis and E. faecium

Objective: To evaluate ampicillin to predict activity of Enterococcus faecalis and Enterococcus faecium to imipenem. Methods: A total of 127 non-duplicated strains of Enterococcus faecalis and 124 strains of Enterococcus faecium were collected from 23 hospitals in China. The antimicrobial susceptibility testing was determinated using broth microdilution and disk diffusion. Results: For all E. faecalis , when using penicillin/ampicillin results to predict susceptibility to imipenem (called as penicillin-imipenem prediction mode and amipicillin-imipenem mode), the categorical agreement (CA) and major error (ME) rate was 88.9%/95.3% and 6.3%/0%, whereas it was 89.7%/96.8% and 8.7%/1.6%, when using that results of disk diffusion, respectively. No very major error (VME) rate was founded for both prediction modes. For penicillin susceptible, ampicillin susceptible E. faecalis , the CA rate of ampicillin-imipenem prediction mode based on results from broth microdilution and disk diffusion to was both 100%, and neither was founded with VME or ME rate. For penicillin resistant, ampicillin susceptible E. faecalis , the CA rate of ampicillin-imipenem prediction mode based on results from broth microdilution and disk diffusion was 57.1% and 81.8%, respectively. And neither was founded with VME or ME rate. For penicillin resistant, ampicillin resistant E. faecalis , the CA/ME/VME rate of ampicillin-imipenem prediction mode based on results from broth microdilution and disk diffusion was 100%/0%/0% and 77.8%/22.2%/0%, respectively. For all E. faecium , the CA rate of penicillin-imipenem and ampicillin-imipenem prediction mode based on results from broth microdilution was 100% and 99.2%, and it was both 99.2% based on results from disk diffusion. ME

faecalis related infections, because of its low adverse effects and high susceptibility.
However, with the emergence of major drug resistance mechanisms such as betalactamase production and penicillin binding protein (PBPs) mutation, the resistance of Enterococcus spp. to penicillin and ampicillin increased (3,4). Additionally, imipenem has been approved for the treatment of E. faecalis related infections by the FDA. For Enterococcus spp., the clinical breakpoint of imipenem is absence in the "Performance Standards for Antimicrobial Susceptibility Testing" of CLSI, but it is clearly mentioned that faecalis, which indicated that strains form different countries or regions may be different in antimicrobial susceptibility phenotype (8,9). In order to study the reliability of using ampicillin susceptibility result to predict imipenem susceptibility of E. faecalis and E. faecium isolated in China, 127 strains of E. faecalis and 124 strains of E. faecium were collected from 23 hospitals across the countries to performed with antimicrobial susceptibility tests of penicillin, ampicillin and imipenem, respectively.

Isolate collection.
A total of 127 nonduplicate isolates of E. faecalis and 124 nonduplicate isolates of E.

Antimicrobial susceptibility testing.
Penicillin, ampicillin and imipenem susceptibility testing and results interpretation for all Enterococcus isolates were performed using broth micro dilution (BMD) and disc diffusion according to the CLSI guidelines. In absence of imipenem breakpoint for Enterococcus spp.

Evaluation index.
Evaluation index include category agreement (CA), which is the percentage of strains whose susceptibility results (suceptible, inter-mediate, resistant) of the evaluated method are consistent with the reference method. Very major error (VME) means that the evaluated method misjudges resistant strains as susceptible strains. Major error (ME) means that the evaluated method misjudges susceptible strains as resistant strains.

Assessment of using penicillin and ampicillin susceptibility to predict imipenem susceptibility of E. faecalis and E. faecium.
For all E. faecalis, when using penicillin susceptibility result of broth microdiluton to predict susceptibility to imipenem (penicillin-imipenem prediction mode), the categorical agreement (CA) and major error (ME) rate was 88.9% and 6.3%, and it was 89.7% and 8.7%, when using that result of disk diffusion, respectively. The CA and ME rate was 95.3%, 96.8% and 0%, 1.6% for ampicillin-imipenem prediction when using results from broth microdiluton and disk diffusion, respectively. No very major error (VME) rate was found for either penicillin-imipenem or ampicillin-imipenem prediction. For penicillin susceptible-and ampicillin susceptible-E. faecalis (107/127), the CA rate of ampicillinimipenem prediction mode based on results from broth microdilution and disk diffusion was both 100%, and neither was founded with VME or ME rate. For penicillin resistant-and ampicillin susceptible-E. faecalis, the CA rate of ampicillin-imipenem prediction mode based on results from broth microdilution (14/127) and disk diffusion (11/127) was 57.1% and 81.8%, respectively. Neither was founded with VME or ME rate. For penicillin resistant-and ampicillin resistant-E. faecalis, the CA/ME/VME rate of ampicillin-imipenem prediction mode based on results from broth microdilution (6/127) and disk diffusion (9/127) was 100%/0%/0% and 77.8%/22.2%/0%, respectively.
For all E. faecium, the CA rate of penicillin-imipenem and ampicillin-imipenem prediction mode based on results from broth microdilution was 100% and 99.2%, and it was both 99.2% based on results from disk diffusion. ME and VME rate for all four prediction modes was 0%. For penicillin susceptible, ampicillin susceptible E. faecium, the CA rate of ampicillin-imipenem prediction mode based on results from broth microdilution (15/124) and disk diffusion (14/124) was both 100%. For penicillin resistant, ampicillin resistant E. faecium, the CA rate was 100% based on results from broth microdilution (108/124) and disk diffusion (109/124). Only one E. faecium isolate was resistant to penicillin, imipenem and susceptible to ampicillin. None of prediction mode was found with ME or VME rate.

Discussion
Enterococcus spp. is one of the most common opportunistic pathogens caused nosocomial infection after S. aureus in all gram-positive bacteria (13). It accounted for 8.42%  (14). In addition, studies also shown that imipenem is certain clinically effective against Enterococcus infection (15,16). The determination standard of imipenem susceptibility testing against the Enterococcus is absence at present, and CLSI only mentioned that the susceptibility result of ampicillin can be used to predict the susceptibility to imipenem of E. faecalis. According to study, whether ampicillin susceptibility can accurately predict imipenem susceptibility of E. faecalis is associated with the susceptibility result of penicillin.
Grouped based on penicillin susceptibility results, for E. faecalis with opposite susceptibility results of penicillin and ampicillin, the CA, VME and ME of using ampicillin susceptibility to predict imipenem susceptibility was different. For example, for penicillinresistant, ampicillin-susceptible E. faecalis, the CA of using ampicillin susceptibility to predict imipenem susceptibility tested by broth microdilution and disk dilution was 57.1% and 81.8%, respectively, and imipenem inter-mediate isolates accounted for 42.9% and 18.2%. For penicillin-resistant, ampicillin-resistant or penicillin-susceptible, ampicillinsusceptible E. faecalis, ampicillin susceptibility of broth microdilution could both accurately predict imipenem susceptibility (CA/VME/ME=100%/0%/0%). Conceicao N et al.
also reported that E. faecalis isolates susceptible to both penicillin and ampicillin were also susceptible to imipemem tested by broth microdilution, and it was consistent with our study (8). It should be noted that for E. faecalis resistant to both penicillin and ampicillin tested by disc diffusion, ampicillin susceptibility could not be used to predict imipenem susceptibility (the CA and ME of ampicillin-impenem prediction was 77.8% and 22.2%). In recent years, penicillin-resistant, ampicillin-susceptible E. faecalis has been reported frequently (17,18). As Metzidie et al. reported, the detection rate was 31.4%, higher than that in our study (11% tested by broth microdilution and 8.7% tested by disc diffusion), and most strains was resistant to imipenem (18). Weinstein MP et al. found that the CA of using penicillin susceptibility to predict imipenem susceptibility tested by broth microdilution was 95.2%; and the CA of using ampicillin susceptibility to predict imipenem susceptibility tested by disk diffusion was 99.8% (ME=0.2%) (7). The result ignores penicillin-resistant, ampicillin-susceptibile E. faecalis, leading us to mistakenly believe that ampicillin susceptibility can predict imipenem susceptibility of all E. faecalis with different susceptibility phenotypes.
In our study, the penicillin-resistant, ampicillin-susceptible E. faecium strains are rare (1/124), and for penicillin-susceptible, ampicillin-susceptible or penicillin-resistant, ampicillin-resistant E. faecium, the CA of using ampicillin susceptibility to predict imipenem susceptibility of was both 100%. And for all E. faecium, it was 99.2% tested by either broth microdilution or disk diffusion, which was both 98% in the study of Weinstein

Not Applicable
Authors' contributionsLG and DY contributed to study conception, data acquisition, analysis and interpretation, manuscript drafting and revising; YG, YZ, QS, YY contributed to data analysis; DZ and FH contributed to study conception and design, data acquisition, analysis and interpretation, manuscript revising. All authors read and approved the final manuscript.

Funding
This work was supported by the National Natural Science Foundation of China (grant no. 81871690).

Availability of data and materials
Except for participant identifying information, all data generated or analysed during this study are included in this published article.

Ethics approval and consent to participate
The study was approved by ethics review boards at HuashanHospital, Fudan University (KY2018-218,01, on Sep 1 2017).
All participants provided written informed consent to participate in the study.

Consent for publication
All participants provided written informed consent for their data to be disseminated, including publication in peer-reviewed journals.