INTRODUCTION
Timely intervention in the treatment of bloodstream infection is of high significance to provide precise antimicrobial [1,2]. With prompt determination of antimicrobial susceptibility of causative agent, rapid AST can help selecting appropriate antimicrobial therapy. Furthermore, it is expected that rapid antimicrobial susceptibility test (AST) can contribute to the decrease of redundant laboratory tests, healthcare-associated expenses and prevention of empirical therapy [3-5]. In most of clinical settings, clinicians utilize phenotypic approaches for the antibiotic profiling of patients. Currently, most of clinical settings utilize phenotypic AST methods including manual methods such as broth microdilution method, disk diffusion method,E-test and commercially available automated systems based on turbidity measurement such as MicroScan WalkAway (Beckman Coulter, Brea, CA, USA), VITEK 2 (bioMerieux, Hazelwood, MO, USA) and PHOENIX (BD Bioscience, San Jose, CA, USA).
These conventional AST methods require relatively longer turnaround time (TAT) due to preceding preparation procedures including blood culture, Gram stain and overnight subculture onto solid medium of flagged positive culture samples [6]. For reducing TAT, various researchers have tried to directly process the positive blood culture (PBC) samples without subculture involving overnight culture [7]. AST methods utilizing direct inoculation from PBC had previously been reported with significant agreement rates compared to reference methods [8,9]. However, these previous studies reported that direct AST methods were insufficient to handle Gram-positive bacteria, mainly applicable for Gram-negative organisms [10-12]. The QMAC-dRAST (Quantamatrix Inc., Seoul, Republic of Korea) system has been introduced in the clinical microbiology laboratory for expedite treatment of bloodstream infection and antibiotic- resistant strain infections via a direct inoculation of PBC [13]. This system determines the antimicrobial susceptibility of bacteria within six hours without performing any additional separation processes and inoculum size measurement [13-15].
The objective of this study was to evaluate the accuracy of QMAC-dRAST AST system with faster protocol for staphylococci and enterococci cocci isolates originated from blood culture bottles. The accuracy was compared to that from MicroScan WalkAway plus system, commonly used as an automated AST system.
MATERIALS AND METHODS
1. Samples
The clinical evaluation of this study was conducted from June 2015 to June 2016 at Seoul National University Hospital (SNUH). For blood culture, BACTEC Plus Aerobic/F and Anaerobic/F culture bottles were used in BACTEC FX automated incubation system (Becton Dickinson Company, NJ, USA) and BacT/Alert FA Plus and SN bottles in a BacT/Alert 3D system (bioMerieux Inc., Marcy l'Étoile, France). PBC samples containing mono-bacterial infection with staphylococci and enterococci were included in the study. The PBC samples identified as Gram-negative rod, yeast and Streptococcus spp. were excluded from this study.
2. Quality control strains
Quality control testing of QMAC-dRAST panels was conducted when they were manufactured at the factory and delivered to the laboratory. For the quality control organisms, E. faecalis ATCC 29212 and S. aureus ATCC 29213 strains were used.
3. QMAC-dRAST system as the direct and rapid AST
A PBC sample was collected from the positive-flagged blood bottle with a 1 mL syringe and transferred to a sterile tube without any additional process. To achieve the optimum inoculation concentration for the QMAC-dRAST system, a 10 μL of the PBC sample was mixed with 4 mL of liquid-state 0.5% agarose at 37-40°C. Next, 10 μL of the mixture was inoculated in the 96-well format QMAC-dRAST Gram-positive (GP) panel containing 13 antimicrobials at various concentrations. Due to the capillary effect, the micro-patterned radial shape of the well helped the agarose mixture spread and form a disk shaped matrix in the entire well. After solidification of the agarose matrix at room temperature, 100 μL of culture medium was loaded into well of QMAC-dRAST GP panel to rehydrate the freeze-dried antimicrobial antibiotics. The culture medium with antibiotics was diffused into the agarose matrix. The QMAC-dRAST GP panel was incubated for 6 hours at a 35°C and images of each well were taken both at the time of inoculation (0 hour) and the experimental termination (after 6 hours). After acquisition and analysis of images from each well, the minimum inhibitory concentration (MIC) values and the interpretative determination were reported based on the interpretive criteria of the the Clinical & Laboratory Standards Institute (CLSI) [16].
4. MicroScan WalkAway plus system
For comparative data, MicroScan WalkAway plus system was processed in parallel with QMAC-dRAST system. A drop of PBC sample was inoculated onto blood agar plate. The inoculated blood agar plates were incubated at 35°C in 5% CO2 to enable bacterial growth. After 20 hours incubation, the bacterial colonies were formed on the agar plate of a pure culture, as clinical isolates.
5. Broth microdilution (BMD) method as the confirmatory AST
The BMD method, recommended by CLSI, was used as a reference method to establish consensus categorical and MIC discrepancies occurred between the QMAC-dRAST system and MicroScan WalkAway plus system for each organism-antimicrobial combination [16]. For the BMD method, the antibiotic solutions were prepared from the stock solution. The antimicrobial agents at the appropriate concentration, which was determined by CLSI recommendation, were loaded on the bottom of 96 microwell plates (SPL Life Sciences, Pocheon, Korea). After preparation of the BMD panel, 10 μL of clinical isolates at a final concentration of 5×105 CFU/mL was inoculated into well. The BMD tests were performed in triplicate. After 16-20 hours of incubation at 35°C, the MIC value of the BMD tests was determined as the concentration of the complete inhibition in growth as compared to the control. In case of results from erythromycin, linezolid, and tetracycline, MIC was determined in the lowest concentration where the trailing growth began [16]. For trimethoprim/sulfamethoxazole, MIC was the concentration in which there is ≥80% reduction in growth as compared to the control [16].
6. Identification of bacteria from PBC bottles
For the analysis of bacterial identification, a PBC sample was processed using the Sepsityper kit (Bruker Daltonics Inc., Billerica, MA, USA) prior to analysis using the Bruker MALDI-TOF Biotyper system (Bruker Daltonics Inc., Billerica, MA, USA) by trained staff at SNUH. In detail, 1.0 mL of sample from a PBC sample was transferred to a 1.5 mL centrifuge tube. A 200 μL aliquot of lysis buffer (provided) was added to the PBC sample, and the mixture was vortexed prior to centrifugation. Following centrifugation, the supernatant was removed, and the bacterial pellet was resuspended in 1.0 mL of wash buffer (provided), vortexed, and centrifuged. The supernatant was discarded, and the pellet was resuspended in 300 μL of deionized water. Then, 900 μL of 100% ethanol was added, and the mixture was vortexed and centrifuged at 13,000 rpm for 1 min. The ethanol was discarded, and the sample was again centrifuged. The pellet was allowed to dry completely. After drying, 70% formic acid (Sigma-Aldrich, St. Louis, MO, USA) at the same volume as the pellet (∼10 μL) was added and mixed. The same volume (∼10 μL) of acetonitrile (Sigmaldrich, St. Louis, MO, USA) was added, and the pellet was resuspended. One microliter of the final supernatant was dropped onto the MALDI plate and dried. It was then overlaid with 1 μL of the α-Cyano-4-hydroxycinnamic acid (HCCA) matrix solution. After drying the sample, MALDI-TOF MS analysis was performed on a Bruker Microflex MALDI-TOF mass spectrometer and interpreted by MALDI Biotyper software (version 3.1).
7. Data analysis
The MIC values analyzed from the QMAC-dRAST system and BMD method were translated into clinical categories (susceptible, intermediate or resistant) according to the interpretative criteria of the CLSI based on bacterial identification results. The MIC values from MicroScan WalkAway plus were determined according to the manufacturer’s instructions including the interpretation of results. The concordance in results from tests was recorded as agreement (Categorical agreement (CA), i.e., agreement of interpretive results between test method and the reference. Essential agreement (EA), i.e., agreement within plus or minus, one two-fold dilution of the new device under evaluation with the reference method) while susceptibility discrepancies were classified as very major errors (VME, i.e., sensitive with the test method but resistant with the reference method, false susceptible), major errors (ME, i.e., resistant with the automation method but susceptible with the reference method, false resistant), or minor errors (mE, i.e., susceptible or resistant versus intermediate).
RESULTS
1. Direct identification of PBCs with Gram-positive cocci for interpreting AST results
Bacterial identification of PBC samples of Gram-positive cocci was simultaneously performed by Sepsityper kit and MALDI Biotyper. Since it is possible to obtain identification results of PBC samples from Sepsityper kit and MALDI Biotyper within 2 hours, it was possible to interpret the MIC values at the termination of AST. Clinical samples identified with species level were only included for this study. The clinical samples identified with genus level or were not identical with identification of MicroScan WalkAway plus system were also excluded from in this evaluation. Total 115 positive blood culture samples of Gram-positive cocci were investigated and a complete list of tested organisms was presented in Table 1. There were 76 Staphylococcus spp. including 42 S. aureus, 23 S. epidermidis, 8 S. haemolyticus, and 3 S. hominis. There were 39 Enteoroccocus spp. including 24 E. faecium,, 14 E. faecalis and one E. gallinarum. These GPC isolates were used for interpreting MIC and interpretative AST results.
Table 1
Distribution of Gram-positive cocci from PBCs used in this study
2. The performance of the QMAC-dRAST system
The AST results of the QMAC-dRAST system for Gram-positive cocci from 115 (1,187 antimicrobial agent-microorganism combinations) PBC samples are shown in the Table 2. For 115 Gram-positive samples, through the QMAC-dRAST system, we observed 94.9% of CA and 98.3% of EA between QMAC-RAST system and MicroScan WalkAway plus system, respectively. The discrepancy rates of QMAC-dRAST were VME of 1.0% (5/485), ME of 1.2% (9/693) and mE of 4.0% (47/1,187), respectively. The highest CA was observed from ampicillin (100%), followed by linezolid (99.1%). Lowest CA was shown from erythromycin (84.3%) and ciprofloxacin (89.6%). The QMAC-dRAST system yielded over 90% of CA for each antimicrobial except two antimicrobial agents; erythromycin and ciprofloxacin. The majority of discrepancies was classified as mEs.
The discrepancy rates for to QMAC-dRAST results compared to those from MicroScan WalkAway plus system in species level are shown in Table 3. The QMAC-dRAST system yielded CA 96.8%, EA 97.8% in S. aureus, CA 90.9%, EA 98.7% in CNS and CA 96.5, EA 98.7% in Enterococcus spp., respectively (Table 3).
Specifically, four VMEs and two MEs were observed from S. aureus in the QMAC-dRAST. Regarding the lowest CA from erythromycin-S. aureus combination, majority of discrepancies were determined as mEs. One of 18 oxacillin-resistant S. aureus and one of 15 erythromycin-resistant S. aureus were not determined as resistant by the QMAC-dRAST system. From majority of antimicrobial agents except erythromycin, CA exceeded 95% in S. aureus. For CNS, only one VME and seven MEs were observed in the QMAC-dRAST results. Among the seven MEs, three MEs were detected from trimethoprim/sulfamethoxazole and mEs occurred mostly from ciprofloxacin, gentamicin, erythromycin, and vancomycin. For Enterococcus spp., no VME and ME were found, however, mE were mostly observed from erythromycin and vancomycin. All vancomycin-resistant Enterococcus were detected as resistant by the QMAC-dRAST system. In overall, mE accounted for the most of errors from the tests with all antimicrobial agents (Table 3).
The QMAC-dRAST results were available within 6 hours after inoculation into the QMAC-dRAST GP panel, which were concordant with previous research [13].
Table 2
Agreement rates in AST results between the QMAC-dRAST and MicroScan WalkAway plus system for Gram-positive cocci
Abbreviations: CA, category agreement; EA, essential agreement; VME, very major error; ME, major error; mE, minor error.
With the results obtained with the standard method as comparator. Errors between the results compared to the standard methods with MicroScan
WalkAway plus system were resolved using BMD method (see text for details).
Table 3
Discrepancy rates and agreement rates of two AST systems in this clinical evaluation classified into combination of antimicrobial agents and bacterial species
Abbreviations: S, susceptible; R, resistant; CA, category agreement; EA, essential agreement; VME, very major error; ME, major error; mE, minor error.
With the results obtained with the standard method as comparator. Discrepancies among results obtained from MicroScan WalkAway plus system were identified by the BMD method (see text for details).
The number of results within the intermediate category can be calculated by subtracting the resistant and susceptible results from the number of isolate-antibiotic combinations tested.
DISCUSSION
Rapid and accurate antimicrobial prescription is essential for successful management of BSI patients. In general, antimicrobial regimens are empirically chosen until AST results are available. Since empirical broad-spectrum antimicrobial treatment could induce the emergence of antimicrobial resistant microorganisms, it is important for clinicians to prescribe requisite antibiotics for reduction of mortality and morbidity. Faster identification of antimicrobial susceptibility can reduce TAT via expediting appropriate antimicrobial choices. Through expedite antibiotic profiling, patient management can be further improved by reducing numbers of redundant laboratory tests and procedures [5]. As the current AST systems necessitate longer TAT and manual procedures, there needs automated form of rapid AST system for both reducing TAT and redundant procedures [17].
However, the antimicrobial prescription process using current automated AST system takes more than two days from existence of causative agents in blood culture bottle, involving sequential experimental step; sub-culture to form the colony from the PBC samples and to perform conventional AST. To accelerate this process, many research groups have focused on the development of methodology by directly applying the PBCs into the current AST system.
One experimental approach has utilized inoculum directly from PBCs for AST, rather than cultured microbial colonies. Several studies have already compared this direct AST methods to the current methods used in the laboratory. These direct AST methods using inoculation from PBCs had quite acceptable results for Gram-negative rods, but not for Gram-positive cocci [8,9,11,12,18]. For higher accuracy of direct AST result for Gram-positive bacteria, many studies have attempted and chosen methods requiring multiple procedures, such as high speed centrifugation [5], reagent treatments [19], short incubation time on solid medium [20]. As a result, accuracy of direct AST results for Gram-positive cocci were quite acceptable with comparable agreement. However, these additional processes are not suitable for clinical application due to laborious works.
The other strategy is the application of novel techniques such as MALDI-TOF mass spectrometry, PCR-based techniques, microarrays, and microfluidics [7]. The QMAC-dRAST system detects the response of individual bacteria to antimicrobials by time-lapse microscopic observation and image analysis of micro- colony formation at the various concentrations of antimicrobial agents [13].
This study focused on whether the QMAC-dRAST system could be utilized routinely to reduce time for AST results of staphylococci and enterococci by comparing results between QMAC-dRAST system from PBCs and MicroScan WalkAway plus system.
As a result, the QMAC-dRAST system evaluated 1,187 antimicrobial Gram-positive cocci combinations, showing 94.9% overall CA and 98.3% overall EA, with only 1.0% VME and 1.3% ME. These results indicate that the antimicrobial agreement rate of QMAC-dRAST from PBC with Gram-positive cocci was comparable or superior to that of other studies [9,12,21]. It is noteworthy that the detection of oxacillin-resistant staphylococci with the QMAC-dRAST system was accurate even if one VME was observed in S. aureus, this result agreed with other standard AST reporting for oxacillin-resistance detection, for the agreement rates ranged from 95 to 100% [10,22]. Also, the QMAC-dRAST yielded high accuracy for the detection of vancomycin-resistant enterococci. Interestingly, no errors were observed for vancomycin and linezolid, the most frequently used antibiotics to treat systemic infections caused by staphylococci and enterococci. These results showed that QMAC-dRAST system can precisely detect the antimicrobial resistance of medically significant GPC.
However, there lie some limitations to this study. The QMAC-dRAST panel did not cover some concentrations of oxacillin necessary for interpreting CNS. The up-to-date version of panel fully cover concentrations for CNS. In addition, further studies are necessary to confirm the reliability of this method with various kinds of pathogens with different resistance phenotypes.
We hereby report the favorable performance of direct AST via the QMAC-dRAST system with PBC sample. It is expected that the QMAC-dRAST system may provide the essential information to accelerate therapeutic decisions in a timely manner by decreasing the time from flagged positive signal to obtain AST results for Gram-positive cocci. This could offer the possibility of having susceptibility test results available on the same day that a positive blood culture was detected.
