β-Hemolytic streptococcal isolates obtained from humans can be subdivided into large-colony and small-colony (<0.5 mm in diameter) formers. Large colony formers include Streptococcus pyogenes (Lancefield group A antigen), Streptococcus agalactiae (Lancefield group B antigen), and Streptococcus dysgalactiae subsp. equisimilis (Lancefield group C and G antigens) [1]. The small-colony-forming β-hemolytic strains with Lancefield group A, C, F, or G antigens are considered part of the viridans group streptococci (VGS). VGS also include Streptococcus mitis, Streptococcus oralis, Streptococcus mutans, Streptococcus salivarius, Streptococcus sanguinis, and Streptococcus bovis [1]. Although penicillin remains the drug of choice in the treatment of infections caused by large-colony-forming β-hemolytic streptococci (BHS), drug tolerance and clinical therapeutic failures have been reported [2]. Macrolides and lincosamides have been frequently used to prevent β-lactam allergies in patients. These agents are also used in empiric and preventive therapies for the treatment of BHS infections [3, 4]. However, recent studies have shown considerable changes in the susceptibility of BHS to erythromycin and clindamycin, although different resistance rates to these agents owing to geographical variation and investigators have been reported [5-7]. β-Lactam agents have been the treatment of choice for VGS infections; however, increase in the incidence of VGS with multidrug-resistance to penicillin and other agents, such as cephalosporins, macrolides, lincosamides, tetracycline, quinupristin-dalfopristin, and quinolones, has been reported [7, 8]. Moreover, CLSI has recommended that VGS isolated from normally sterile body sites should be tested for penicillin susceptibility by using a minimum inhibitory concentration (MIC) method and interpretive criteria [9]. Accurate susceptibility testing for BHS and VGS is required in order to guide appropriate antimicrobial therapy and to monitor further spread of resistant pathogens. Rising drug resistance of BHS and VGS has increased the need for accurate determination of antimicrobial susceptibility in a timely manner in clinical microbiology laboratories. Rapid reporting of the results of an antimicrobial susceptibility test (AST) has been shown to improve patient outcomes and reduce hospital costs [10, 11]. Because there are significant differences in the susceptibility of BHS and VGS to β-lactam agents, there are separate interpretive criteria for the susceptibility of the 2 groups of organisms to ampicillin, penicillin, cefotaxime, ceftriaxone, and cefepime [9].

Automated commercial susceptibility test systems for streptococci offer reliable AST results for MIC measurement and help accurately determine the antimicrobial susceptibility profile according to the Streptococcus group. However, most studies are focused on evaluating the AST performance of Streptococcus pneumoniae. To the best of our knowledge, no study has been conducted to evaluate the accuracy of the automated MicroScan (Siemens Healthcare Diagnostics, Sacramento, CA, USA) AST system for susceptibility testing of BHS and VGS. Therefore, this study was designed to evaluate the clinical usefulness of the MicroScan MICroSTREP plus antimicrobial panel (MICroSTREP) as a susceptibility testing system for BHS and VGS.

Using the agar dilution method, the VGS non-susceptibility rates were 57.6% for ampicillin (MIC ≥0.5 mg/L), 59.3% for penicillin (MIC ≥0.25 mg/L), 22.0% for cefotaxime (MIC ≥2 mg/L), 28.8% for meropenem (MIC >0.5 mg/L), 39.0% for erythromycin (MIC ≥0.5 mg/L), 22.0% for clindamycin (MIC ≥0.5 mg/L), and 13.6% for levofloxacin (MIC ≥4 mg/L). All BHS isolates were susceptible to ampicillin, penicillin, cefotaxime, and meropenem. Among the BHS isolates, the non-susceptibility rates for erythromycin, clindamycin, and levofloxacin were 28.0%, 37.3%, and 6.7%, respectively (Table 1).

The overall EAs and CAs determined for MICroSTREP and the CLSI reference method were 98.2% and 96.9%, respectively (Tables 2 and 3). For the BHS isolates, the EAs for individual antimicrobial agents ranged from 94.7% (erythromycin) to 100% (penicillin, cefotaxime, meropenem, and levofloxacin), while the EA for ampicillin, clindamycin, and vancomycin was 98.7%. The CA for erythromycin was 96.0%; for cefotaxime, meropenem, levofloxacin, and vancomycin, 100%; and for ampicillin, penicillin, and clindamycin, 98.7%. For the VGS isolates, the EAs ranged from 91.5% (penicillin) to 100% (cefotaxime, erythromycin, clindamycin, and levofloxacin), and the EA for ampicillin and vancomycin was 96.6%. The CAs ranged from 84.7% (penicillin) to 100% (erythromycin, clindamycin, and vancomycin). The CAs for ampicillin and meropenem were 88.1% and 86.5%, respectively, and the CA for cefotaxime and levofloxacin was 96.6%.

Of the total 134 isolates, all categorical errors were MIEs (Table 3). In the cases of 14 isolates for which MIEs were obtained for penicillin and/or ampicillin, MIEs for penicillin and ampicillin were obtained with 4 isolates (3 S. mitis isolates, 1 S. anginosus isolate); MIE only for penicillin was obtained with 6 isolates (4 S. mitis isolates, 1 S. anginosus isolate, and 1 S. agalactiae isolate); and MIE only for ampicillin was obtained with 4 isolates (2 S. mitis isolates, 1 S. anginosus isolate, 1 S. agalactiae isolate). The frequency of total MIEs obtained with BHS isolates was lower than that obtained with VGS isolates (Table 4).

Healthcare professionals are faced with a variety of significant, fastidious organisms, including S. pneumoniae, BHS, and VGS, that are increasingly showing resistance to commonly used antimicrobial agents. Although, in the past, many laboratories may have chosen to screen VGS isolated from normally sterile body sites, e.g. cerebrospinal fluid, blood, and bone, to determine penicillin resistance, the rapid spread of multidrug-resistant strains requires a more aggressive approach. In addition, certain antimicrobials (penicillin, ampicillin, ertapenem, meropenem, and daptomycin) that may be used to treat VGS infections cannot be reliably tested using the disk diffusion method [9]. The CLSI also recommends the inclusion of penicillin (or ampicillin), cefepime (cefotaxime or ceftriaxone), erythromycin, clindamycin, and vancomycin in a routine, primary testing panel [9].

The accuracy and efficiency of AST dictates timely and appropriate decisions in choosing the antibiotic therapy. Automated systems with built-in expert systems can potentially increase the reproducibility and reliability of test results, and thus can be expected to improve the quality of patient care. In this study, penicillin-intermediate, and/or ampicillin-intermediate, and VGS isolates that were not susceptible to meropenem accounted for the bulk of the total MIEs. Despite the elevated frequency of MIEs obtained with VGS, the high EA values for penicillin, ampicillin, and meropenem suggest that these errors can be largely attributed to the MICs being close to the interpretive breakpoints (Table 4). In particular, all categorical errors for meropenem occurred when the MICroSTREP MIC was 2-fold dilution lower than the reference. The overall frequency of MIEs obtained with BHS was lower than that obtained with VGS, but MIEs for erythromycin and clindamycin were only detected with BHS. The highest MIE rates obtained for penicillin were similar to those reported by Guthrie et al. [12], who reported that penicillin was responsible for the highest MIE rate obtained with S. pneumoniae. When the reference MIC was close to a breakpoint value, 2 MIEs were observed for several isolates [13]. For most antibiotics, the MICroSTREP MICs tended to be lower than the reference MICs, which greatly contributed to the number of MIEs. The exceptions were the MIC results for ampicillin, levofloxacin, and vancomycin obtained with BHS: MICroSTREP reported higher MICs in these cases (Table 2). No VMEs or MEs were detected, and relatively few MIEs were observed. The only observed exceptions to the performance standards were categorical disagreement in cases of ampicillin, penicillin, and meropenem for VGS. The more the isolate population was concentrated near the MIC breakpoint level, the greater the possibility of categorical discrepancies between the results obtained using the testing instrument and the reference method. In contrast, the more the isolate population was distributed far away from MIC breakpoint level, the lesser the possibility of categorical discrepancies. Although diagnostic performance of the MIC testing system in BHS and VGS has not been reported, Jorgensen et al. [14] reported that MICroSTREP did not result in any VMEs or MEs and that high EA and CA values were achieved in testing with S. pneumoniae. The high EA and CA values obtained by using MICroSTREP satisfied the minimal performance criteria (CA, ≥90%; EA, ≥90%; VMEs, ≤1.5%; and MEs, ≤3%) of the Food and Drug Administration (FDA) [15]. The FDA suggested that 50% susceptible and 50% resistant distribution is desired to generate meaningful statistics regarding the performance characteristics of a method or system [15]. However, it was unlikely that sequentially collected isolates from clinical samples would be anything similar to the 50% susceptible and 50% resistant distribution suggested by the FDA.

The performance of MICroSTREP affirms the capability of the instrument as a reliable and efficient diagnostic tool for determining the appropriate antimicrobial agents for use against VGS and BHS infections. In conclusion, this instrument was able to decrease the turnaround time to results because of the reduced hands-on time required in comparison to that required in conventional laboratory methods.