Performance of QMAC-dRAST (Direct Rapid Antimicrobial Susceptibility Testing)-a Newcomer in Phenotypic Automatic AST

QMAC-dRAST is a phenotypic automatized Antibiotic Susceptibility Testing (AST) system based on microfluidic chip technology enabling observation of changes in a single bacterial cell under antibiotic treatment conditions. The 96 wells plate with dried antibiotics comprises 19 and 17 antibiotics for the Gram-Negatives (GNs) and Gram-Positives (GPs), respectively. Categorical (Sensitive, Intermediate or Resistant) results were compared to results obtained by our laboratory standard susceptibility testing procedure and given as Categorical Agreement (CA).


INTRODUCTION
In a recent review, estimates of the total burden of Blood Stream Infection (BSI) from population-based studies from North America and Europe were summarized [1]. The BSI incidence ranged between 113 and 220 per 100,000 populationbased on reports from eight countries [2]. The serious prognosis for many of these infections and the increasing (MALDI-ToF MS) made a revolution with respect to establishing accurate, rapid and cheap pathogen identification [5]. Rapidly growing micro organisms often may be convincing identified after 3-6 hour growth on agar plates. A great focus is on fast Antimicrobial Susceptibility Testing (AST) in bloodstream infections in order to help clinicians being able to give the most optimized antimicrobial treatment. QMAC-dRAST TM (Quanta Matrix Inc., Seoul, Republic of Korea) is a phenotypic AST system based on microfluidic chip technology that enables the observation of changes in a single bacterial cell under antibiotic treatment conditions, with a total turn around time of only 6 h from Gram stain reporting of recognized positive blood cultures [6]. The positive blood culture sample is Gram stained prior to loading the instrument for choosing the right antibiotic panel. On the instrument blood, culture material is automatically mixed with agarose and inoculated into a 96 wells plate with dried antibiotic agents. Each well is followed over time by microscopic detection of changes in bacterial colony formation in the agarose. The performance of this technique has been evaluated in recent publications showing promising categorical (Sensitive, Intermediate or Resistant) agreements (CAs) with standard methods [6 -10]. The patented 96 wells plate with dried antibiotics comprises, 19 and 17 antibiotics for the Gram Negatives (GNs) and Gram Positives (GPs), respectively. The size of an instrument is a minor refrigerator (width 59 cm, depth 76.5 cm and 109 cm high, weighting 180 kg), fully automated with random access of up to 12 samples (Quantamatrix.com).
In this study, AST results for 255 recognized positive blood culture bottles from as many patients obtained by QMAC-dRAST TM were compared to results obtained by the standard procedures performed in the laboratory in order to be able to comment on CAs, error rates, time to result and commenting on laboratory flow.

Study Population
This prospective study was conducted in a Danish Regional Department of Clinical Microbiology performing microbiological service to a region (Region Zealand) with 800,000 inhabitants and six hospitals with a 2353 bed capacity. On a yearly basis, 140,000 blood cultures are examined. In a 3month period (2019/2020), blood cultures detected positive in the morning were included. Only the first positive blood culture was included for each patient. Cases, where the pathogen was not identifiable via MALDI-ToF MS analysis, were excluded as was known off-panel strains of QMAC-dRAST TM , such as Gram-Positive bacilli, Streptococcus and Candida species. No polymicrobial cases were included.

Laboratory Procedures
Besides being routinely processed, the blood cultures detected positive were examined in a separate set up in order to compare CA between QMAC-dRAST TM and routine testing methods used in the laboratory (EUCAST recommendations, see below). Evaluation of categorical disagreements included Minimal Inhibitory Concentration (MIC) determinations (etest and Broth Micro Dilution (BMD) ). 5% horse-bloodagar plates were streaked with material from blood cultures detected positive and after approximately 3-6 hours incubation in 5% CO 2 -enriched atmosphere growth sufficient for MALDI-ToF MS examination was present. MALDI-ToF MS examination was performed as recommended by the manufacturer for direct examination of bacterial growth. The standard criteria for taxon confirmation used in the laboratory were applied. Briefly, identification results were considered reliable at the genus level when the score value was ≥ 1.7 and at the species level when the score value was ≥ 2.0 or ≥ 1.7 and the score value difference to the next best taxon match was ≥ 0.3.
QMAC-dRAST TM testing: A GP or GN panel was chosen for further AST according to the Gram stain result of positive blood culture. Briefly, after the identification of Gram negative and positive bacteria by direct smear examination about 300 µl of the culture was taken from the blood culture bottle, using a syringe, and added to a test tube. The test tubes and other kit components were placed in the QMAC-dRAST TM instrument and the following testing was done fully automated. The test was set up and performed according to the instructions given by the instrument. GN AST was performed using the panel card for the GN bacteria, including examining for susceptibility to amikacin, gentamicin, amoxicillin/clavulanic acid, ampicillin, ampicillin/ sulbactam, piperacillin/tazobactam, aztreonam, cefotaxime, ceftazidime, cefepime, ertapenem, imipenem, meropenem, colistin, trimethoprim/ sulfamethoxazole, and an ESBL test. For GP bacteria, the GP card included testing for susceptibility to penicillin, ampicillin, oxacillin, ciprofloxacin, levofloxacin, erythromycin, clindamycin, inducible clinda-mycin resistance, gentamicin, streptomycin, rifampicin, trimethoprim/sulfamethoxazole, vancomycin and linezolid. Cefoxitin screening was also performed in order to detect MRSA and methicillin-resistant Coagulase Negative Staphylococci (CoNS). MIC results were interpreted in accordance with EUCAST recommendations. A well testing for inducible clindamycin resistance was included in the GP panel.
In addition to QMAC-dRAST TM testing, positive blood cultures were processed using standard setups in the laboratory following EUCAST disc diffusion recommendations and for vancomycin MIC determinations by etest. Thereby the following result comparisons could be made:

Performance Evaluation of Tests
The terms for the AST accuracy evaluation were defined as follows: Categorical Agreement (CA), comparison of categorical result (Sensitive, Intermediate or Resistant) obtained with QMAC-dRAST TM and with EUCAST disc diffusion and/or etest/BMD MIC. Very major error (VME), false susceptibility of QMAC-dRAST TM compared to EUCAST disc diffusion and/or etest/BMD MIC. Major Error (ME), false resistance of QMAC-dRAST TM compared to EUCAST disc diffusion and/or etest/BMD MIC. Minor error (mE), intermediate susceptibility to QMAC-dRAST TM and susceptible or resistant according to EUCAST disc diffusion and/or etest/BMD MIC, or vice versa.
For all blood culture specimen runs, data measuring time from loading the instrument to available susceptibility report were extracted from the instrument.

Blood Culture Specimens Included
During the study period, 255 positive blood cultures from as many patients were examined. All were monobacterial infections. All strains were reliably identified by MALDI-ToF MS to the species level except two strains (strains belonging to the genera Proteus and Acinetobacter). GNs comprised 144 of the positive blood culture strains; 111 were GPs.

Taxons and Numbers of Strains Included No. of Comparable Tests for Each Taxon and Susceptibility Patterns Obtained on GNs and GPs by QMAC-dRAST Examination
Data on GNs and GPs included and no. of comparable tests are presented in Table 1. E. coli and Klebsiella species dominating among the GNs and S. aureus, CNS and enterococci among the GPs. Respectively, 1530 and 972 comparable results were obtained. In Tables 2 and 3 susceptibility patterns obtained on GNs and GPs by QMAC-dRAST examinations are given, except for enterococci, being comparable with the standard setup in our laboratory. In general, strains showed relatively high susceptibility to many antibiotics. Exceptions were ampicillin and amoxicillin/ clavulanic acid for the Gram-negatives, penicillin for S. aureus and for CoNS, more resistant phenotypes were typically seen. ESBL production was detected in six strains, methicillin resistance in one S. aureus strain. Among the 14 enterococcal strains, VanA, VanB and high-level gentamicin resistance in each one E. faecium strain were detected as well as resistance to ampicillin and linezolid in respectively eight and two strains.

Discrepancy and Categorical Agreement Data
Data are given in Table 4. All in all, discrepancies were found in approximately 4% of comparable data when looking on as well GNs as GPs. The percentage of VMEs were 0.7% for GNs and 2.2% for GPs. Among the GNs most frequent discrepancies were detected when comparing results for piperacillin/tazobactam, ceftazidime, cefotaxime, imipenem and ciprofloxacin. Of the 12 discrepancies noticed when comparing data for piperacillin/tazobactam susceptibility four were VMEs of which one was one dilution step from the breakpoint. Additional two VME`s were noticed, one with imipenem on a Ps. aeruginosa strain (one dilution step from the breakpoint) and one with cefotaxime on a E. coli strain (QMAC-dRAST TM MIC ≤1 µg/ml and etest MIC of 4 µg/ml). Among the Gram positives most frequent discrepancies were noticed when comparing results for penicillin, erythromycin, clindamycin and fusidic acid. Two VME discrepancies were noticed when testing S. aureus for penicillin susceptibility; in both cases, the QMAC-dRAST TM MIC was one dilution step below their breakpoint MIC. In 11 tests, QMAC-dRAST TM MIC`s were in the range 0.25-0.5 µg/ml and interpreted as resistant in contrast to results obtained with clover leaf testing [11]. For two VME`s seen when examining for erythromycin susceptibility, MIC differences were ≥ two dilution steps from the EUCAST breakpoint recommendation. Eleven of 13 testing results with discrepancy when examining susceptibility to clindamycin were VME`s; two when testing S. aureus strains and nine when examining CoNS strains. In nine testings these discrepancies were based on not detecting inducible clindamicin resistance by the QMAC-dRAST TM setup (QMAC-dRAST TM MIC values were one dilution step from the recommended breakpoint). In one testing, QMAC-dRAST TM detected in accordance with the routine method inducible clindamycin resistance. The four remaining VMEs were among E. coli strains and all related to susceptibility testing of ampicillin and amoxicillin/clavulanic acid; for all, MICs were within one MIC dilution step from the breakpoints given by EUCAST.

Data on Time to Instrument Reporting on Susceptibility Pattern
For 87% of blood culture specimens examined, susceptibility reports were available within 6-7 hours and all reports available within 8 hours. Results for strains of E. coli, Klebsiella species and S. aureus, being the most often detected taxons, were ready in 5-6 hours after blood culture material being loaded into the instrument.

DISCUSSION
The QMAC-dRAST TM set up tested in this study is a newcomer in AST testing of blood cultures detected positive. It is based on single-cell morphological analysis that can determine antimicrobial susceptibility by automatically analyzing and categorizing morphological changes in single bacterial cells under various antimicrobial conditions [12]. Determining factors for optimal antimicrobial treatment are correct bacterial identification, reliable performance and interpretation of AST results and the speed with which these informations can be obtained [13]. Combining the accuracy and speed of MALDI-ToF MS identifications with the accuracy and speed of QMAC-dRAST TM has proven to holdthe potential of improving the existing examination flow [7 -10].
Reports from high-income countries have documented key pathogens to have been and remain being S. aureus, E. coli, Klebsiella species, Ps. aeruginosa, enterococci, streptococci and CoNS [14]. The spectrum of organisms differs for community-acquired and healthcare associated (community and hospital onset) infections. Ps. aeruginosa and staphylococci are clearly associated with healthcare institutions, where as among community-acquired cases there are more typically infections due to Streptococcus pneumoniae and other streptococci, and due to E. coli [1].
The robustness and accuracy of MALDI-ToF MS identifications have been documented for the majority of BSI pathogens irrespective of being the setup from Bruker or from Bio Merieux, though especially non-hemolytic streptococci, including pneumococci, still are challenging with respect to species identification. Directly from recognized positive blood cultures the Sepsityper kit (Bruker Daltonics) has been used also in connection with testing the QMAC-dRAST TM strategy [7,8]. In only 18 of 346 (5.2%) incidences of monobacterial infections and in 7 of 38 (18.4%) incidences of polymicrobial infection no identifications were obtained [8], being comparable to results from previous studies [5]. Identifications have to be added to the instrument for finalizing the AST report in order to adjust for expert rules. In our laboratory, we streak positive blood culture material on agar-plates and incubates 3-6 hours making it possible, in by far the most cases, to have identifications and these being ready before AST testing is ready for reporting, which is a prerequisite for interpretation of obtained data by the included expert system based on EUCAST breakpoint recommendations. This is in agreement with the study by Sekercioglu et al. [15], on 1351 positive blood cultures,where a Columbia agar plate with 5% sheep blood was inoculated with one drop from the blood culture broth. After a 5-hour incubation period, a colony from the culture plate was submitted to MALDI-ToF MS. When manufacturer-recommended score values were taken into account, MALDI-ToF MS correctly identified 98.4% of the isolates to the species level with a score of > 2.0, 89.1% with a score between 1.7 and 2.0, and 75.4% with a score of < 1.7.
In the first report from 2014 by Choi et al [12] they compared 189 clinical hospital samples, including extendedspectrum β-lactamase-positive E. coli and K. pneumoniae, imipenem-resistant Ps. aeruginosa, methicillin-resistant S. aureus, and vancomycin-resistant enterococci with BMD testing. A CA of 91.5% with 6.51% mEs 2.56% MEs, and 1.49% VMEs were obtained. Since more studies have been added [7 -10]. In Table 5 published studies comparing data obtained by QMAC-dRAST TM and standard AST methods are given. Four additional studies, including our study have included as well GNs as GPs [6 -8], one study focused on GNs [10] and one on staphylococci and enterococci [9]. In the study by Grohs et al. and in our study QMAC-dRAST TM data  were compared mainly to disc diffusion AST; strains with disagreeing results were additionally examined with etest or BMD testing. In the other studies given in (Table 4) BMD ASTs were used for comparison. All in all CAs of 91.1-96.3% have been found, with our study being the one with highest CAs,%. Likewise, differences in VMEs, MEs and mEs were registered. Thus, comparisons have shown comparable results though with some differences in obtained VMEs, MEs and mEs. Especially the VMEs have to be careful evaluated with respect to possible improvements. In our study MIC determinations for most of discrepancies were close to the breaking points, thereby finding that the QMAC-dRAST TM setup seems very reliable for detecting susceptibilities for by far the most of the compared antimicrobials.
The genera and species QMAC-dRAST TM has been validated for includes the most frequent isolated GNs and GPs. Genera/species dominating in our study are in agreement with those dominating in the other studies, though strains in our study tends to have a more susceptible pattern. This exemplified in the study by Kim et al. [7] on 119 patients where more GNs were ESBL producing (E. coli and Klebsiella species) and carbapenem resistant (Ps. aeruginosa and Acinetobacter baumannii) as well as the GPs including more vancomycin and ampicillin resistant enterococci and methicillin resistant S. aureus. However, single strains were recognized with these susceptibility traits in our study except that no carbapenemase producing strains were recognized. Most often registered discrepancies, though relatively few, when testing GNs were for susceptibility to piperacillin/tazobactam, ceftazidime, cefotaxime, imipenem and ciprofloxacin. By far the most were within +/-one dilution step and the number of VMEs relatively small. This is close to be in accordance with the study by Kim et al. [7] where however no VMEs were observed for GNs with commonly used antibiotics such as beta-lactam/beta-lactamase inhibitors, broad spectrum cephalosporins and carbapenem. Huh et al. [9] studied 110 detected positive blood cultures with staphylococci and enterococci. AST was performed directly using the QMAC-dRAST TM and thereafter, colony isolates derived from subculture were used for the QMAC-dRAST TM , the VITEK-2 system and BMD. The QMAC-dRAST TM with colony isolates produced more reliable results for staphylococci and enterococci than the direct QMAC-dRAST TM blood cultures detected positive. On colony isolates, the QMAC-dRAST TM system performed comparably to BMD and the VITEK-2 system. However, the QMAC-dRAST TM seems relative robust taking the presumed differences in amount of bacterial cells present into consideration; this in agreement with the setup based on single-cell morphological analysis. When examining GPs, especially discrepancies when examining S. aureus strains for penicillin susceptibility and CoNS (and two S. aureus strains) for inducible clindamycin resistance were noticed. In both, different methods are applied by QMAC-dRAST TM and our routine for their determination. Regarding penicillin susceptibility, we use a combination of growth zone appearance and clover leaf testing [11], whereas QMAC-dRAST TM interpretation is based on MICs. Inducible clindamycin resistance we detect as recommended by TM The Open Microbiology Journal, 2021, Volume 15 49 EUCAST by demonstrating the D phenomenon, which does not seem as applicable when examining in a well as in QMAC-dRAST TM . Whether small colony variants of S. aureus, potentially responsible for chronicization of an infection [16], will challenge a method using microscopy as the detection method remains to be settled. When examining positive blood cultures, there is an exigent desire/need for the method also being validated to be used for susceptibility testing of other relevant organisms such as Gram positive rods, streptococci (as well pneumococci as hemolytic and non-hemolytic streptococci), Gram negative cocci (especially Neisseria meningitidis), the HACEK group of fastidious GNs, anaerobic bacteria and Candida species. Patient outcome is critically influenced by delayed effective therapy, wherefore fast and accurate pathogen diagnostics, including AST, decisively improves the care of patients [17]. Progress in BSI pathogen diagnostics is based on a bundle approach that includes optimization of the pre-analytical parameters, rapid start of incubation, the use of rapid methods, re-organization (e.g. 24/7, transportation service) and a close involvement of antimicrobial stewardship teams [17]. Time to available antibiotic susceptibility report for blood culture specimens detected positive plays a major role in diminishing the time to being able to institute the most optimal antibiotic treatment. The presented time to report data obtained by QMAC-dRAST TM in this study, and those mentioned in (Table 4), being of 3-8 hours for blood culture specimens examined, strongly support a further possible improvement in the workflow for handling blood stream infections.
Studies have reported on the importance of initial antibiotic coverage, coverage after guidance by Gram stain of recognized positive blood culture smears and improvement of coverage when guided by MALDI-ToF MS identifications [18]. Vlek et al. [18] reported on 253 episodes of blood stream infections that MALDI-ToF MS directly performed on positive blood culture broths reduced the time until definitive identification of bacterial species by 28.8 hours and increased the proportion of patients on appropriate antimicrobial therapy within 24 hours by 11.3% to 82%. In the study by Kim et al. [8] MALDI-ToF MS identification and QMAC-dRAST TM testing were applied on 359 BSI episodes, where involved pathogens were divided into pathogens having a susceptible or resistant pattern. When reporting based on MALDI-TOF MS results on susceptible/resistant pathogens, optimal targeted treatment was found respectively in 79%/63%, unnecessary broad-spectrum treatment in 16%/5%, suboptimal treatment in 4%/1%, and appropriate antibiotic treatment 100%/68%. Adding QMAC-dRAST TM results to the decision making raised the percentage of optimal antibiotic treatments to 98.2%. The global emergence of antibiotic resistant microorganisms makes it desirable being guided as quickly as possible in relation to antibiotic susceptibility [3]. In that respect phenotypic methods as the QMAC-dRAST TM seems desirable as they illustrate all resistance mechanisms expressed in a micro organism.
From a laboratory flow and technician aspect, the procedure about loading the instrument was very easy and intuitive. The QMAC-dRAST TM is a newcomer on the market and when the data transport from the instrument to a LIMS system will be done automatically the combination of easy loading, the automatized running procedure, easy testing termination, automatic data transport (in addition to automatic alarming when ready for reporting) seems to add important benefits to the current testing availabilities. The saved time from blood culture taken to availably of testing results being present will be able to improve antibiotic treatment of patients.

CONCLUSION
The genera and species QMAC-dRAST has been validated for includes the most relevant GNs and GPs in relation to positive blood culture specimens, though the desire for also being able to test other blood stream infection relevant pathogens, including streptococci, are exigent. The high CA for as well GNs as GPs are promising. The QMAC-dRAST is a newcomer on the market and when the data transport from the instrument to a LIMS system will be done automatically the combination of easy loading, the automatized running procedure, easy testing termination, automatic data transport (in addition to automatic alarming when ready for reporting) seems to add important benefits to the current testing availabilities. The presented time to report data obtained by QMAC-dRAST in this study being of 3-8 hours for blood culture specimens examined strongly support a further possible improvement in the workflow for handling blood stream infections.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE
Due to nature of the study ethics approval is not required.

HUMAN AND ANIMAL RIGHTS
No animals/humans were used for studies that are the basis of this research.

CONSENT FOR PUBLICATION
Not applicable.

AVAILABILITY OF DATA AND MATERIALS
The data supporting the findings of the article is available from the corresponding author [J. J] on reasonable request.

FUNDING
None.

CONFLICTS OF INTEREST
The authors declare no conflict of interest, financial or otherwise.