ABSTRACT
Rapid and accurate detection of carbapenemase-producing Enterobacteriaceae (CPE) is critical for appropriate treatment and infection control. We compared a rapid fluorogenic assay using a carbapenem-based fluorogenic probe with other phenotypic assays: modified carbapenem inactivation method (mCIM), Carba NP test (CNP), and carbapenemase inhibition test (CIT). A total of 217 characterized isolates of Enterobacteriaceae were included as follows: 63 CPE; 48 non-carbapenemase-producing carbapenem-resistant Enterobacteriaceae (non-CP-CRE); 53 extended-spectrum β-lactamase producers; and 53 third-generation-cephalosporin-susceptible isolates. The fluorogenic assay using bacterial colonies (Fluore-C) was conducted by lysing the isolates followed by centrifugation and mixing the supernatant with fluorogenic probe. In addition, for the fluorogenic assay using spiked blood culture bottles (Fluore-Direct), pellets were obtained via the saponin preparation method, which can directly identify the pathogens using matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS). The fluorescence signal was measured over 50 min using a fluorometer. The fluorescent signal of CPE was significantly higher than that of non-CPE in both Fluore-C (median relative fluorescence units [RFU] [range], 5,814 [240 to 32,009] versus 804 [36 to 2,480], respectively; P < 0.0001) and Fluore-Direct (median RFU [range], 10,355 [1,689 to 31,463] versus 1,068 [428 to 2,155], respectively; P < 0.0001) tests. Overall, positive and negative percent agreements of Fluore-C, mCIM, CNP, CIT, and Fluore-Direct were 100% and 98.7%, 98.3% and 97.5%, 88.1% and 100%, 96.4% and 98.7%, and 98.3% and 98.1%, respectively. The relatively lower positive percent agreement (PPA) of CNP was mainly observed in OXA-type CPE. The fluorogenic assay showed excellent performance with bacterial colonies and also directly from positive blood cultures. We included many non-CP-CRE isolates for strict evaluation. The fluorogenic assay will be a useful tool for clinical microbiology laboratories.
INTRODUCTION
Carbapenems are the last resort to control infections caused by Gram-negative bacteria. Currently, however, the emergence and spread of carbapenem-resistant Enterobacteriaceae (CRE) are a threat to global public health (1–3). Both carbapenemase-producing Enterobacteriaceae (CPE) and non-carbapenemase-producing CRE (non-CP-CRE) are associated with significant mortality, and CPE bacteremia is particularly associated with unfavorable patient outcomes compared with non-CP-CRE bacteremia (4). Moreover, the carbapenemase genes located in CPE can be easily spread by mobile elements (5). Therefore, it is important to distinguish CPE and non-CP-CRE as early as possible.
Laboratories use molecular or phenotypic techniques to detect carbapenemases. A molecular assay takes only a few hours to differentiate the genotypes but is expensive and requires specialized personnel. The phenotypic methods, such as the carbapenemase inhibition test (CIT) (6), carbapenem inactivation method (CIM) (7), and modified carbapenem inactivation method (mCIM) (8), take at least 18 to 24 h, which delays detection. The Carba NP test (CNP) requires only 2 h to obtain results (9), but its sensitivity for OXA-48-like carbapenemases is significantly low (10). In addition, the subjective interpretation of the color changes can be problematic.
Here, we introduce a fast and accurate method using a fluorogenic probe. Fluorogenic β-lactamase probes have attracted considerable attention because of their high sensitivity, operational simplicity, and relatively low costs (11–16). Moreover, several studies have utilized fluorescent probes that specifically detect carbapenemase-producing bacteria (17–20). They used the carbapenem moiety as a substrate for carbapenemases and utilized boron-dipyrromethene (BODIPY) or umbelliferone as a fluorophore. Recently, we developed a novel carbapenem-based fluorogenic probe consisting of a carbapenem moiety with umbelliferone connected by an active linker (21). We used benzyl ether to not only induce a cascade reaction with our probe upon enzymatic hydrolysis but also facilitate its binding with active sites of different types of carbapenemases for the detection of a wide range of CPE.
In this study, we evaluated the performance of the rapid fluorogenic assay using a carbapenem-based fluorogenic probe by comparing it with other phenotypic assays (mCIM, CNP, and CIT). In addition, as bacteremia is one of the most important causes of mortality and morbidity worldwide (22), we also investigated the performance of the fluorogenic assay with the pellet derived from positive blood culture, which can be used for direct identification with matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) and antimicrobial susceptibility testing with Vitek 2 (bioMérieux, Marcy l’Etoile, France) (23). To the best of our knowledge, this is the first demonstration of a fluorescence assay for the detection of carbapenemases not only from bacterial isolates but also directly from positive blood cultures.
MATERIALS AND METHODS
Bacterial isolates.A total of 217 previously characterized isolates of Enterobacteriaceae were included in this study, including 63 CPE isolates with the following carbapenemase genotypes: KPC (n = 36), GES (n = 3), NDM (n = 9), VIM (n = 5), IMP (n = 1), OXA (n = 6), KPC plus OXA (n = 2), and NDM plus OXA (n = 1). The other isolates included 154 non-CPE isolates including 48 non-CP-CRE, 53 extended-spectrum-β-lactamase (ESBL) producers, and 53 third-generation-cephalosporin-susceptible isolates (nonproducers). The isolates were collected from various clinical samples obtained from multiple Korean hospitals, and three CPE coproducing multiple carbapenemases were generous gifts from Patrice Nordmann (University of Fribourg, Switzerland). All isolates were stored at −70°C and were thawed and subcultured on a blood agar plate before testing. The MICs of imipenem, meropenem, and ertapenem were obtained using the broth microdilution method according to the CLSI guideline (24). Isolates were previously characterized by PCR amplification and sequencing of carbapenemase genes, including KPC, NDM, VIM, IMP, OXA, and GES, as previously described (25, 26). The presence of the CTX-M gene was investigated by multiplex PCR (27). Detailed characteristics of the isolates are listed in Table S1 in the supplemental material.
Fluorogenic probe.The novel fluorogenic probe 1 ((4S,5R,6S)-6-((R)-1-hydroxyethyl)-4-methyl-7-oxo-3-((4-(((2-oxo-2H-chromen-7-yl)oxy)methyl)phenoxy)methyl)-1-azabicyclo[3.2.0]hept-2-ene-2-carboxylic acid; C27H25NO8; molecular weight 491.50), which is composed of a carbapenem moiety, hydroxybenzyl ether linker, and umbelliferone fluorophore, was synthesized in several steps starting from the key intermediate 2 (21) (Fig. 1). When the β-lactam ring of probe 1 was selectively hydrolyzed by carbapenemase, the cascade reaction occurred rapidly to release hydroxybenzyl alcohol 3 and an anionic form of umbelliferone 4, which is responsible for high fluorescence. The 1 mM stock solution of synthetic probe 1 was prepared using 5% (vol/vol) methanol–phosphate-buffered saline (PBS) as a solvent.
Carbapenemase-specific fluorogenic probe 1 and its hydrolysis by carbapenemase (21).
PBS (pH 7.4) buffer plays a crucial role in the fluorescence assay. First, this buffer minimized the hydrolysis of probe 1, which was rapidly decomposed under basic conditions to generate umbelliferone and affect background fluorescence signal. Second, the buffer maintained a consistent pH (around 7.4) during the enzymatic reaction. Third, the buffer ensured that umbelliferone (generated by the enzymatic reaction) existed in an anionic form (deprotonated form), which induced fluorescence emission at 465 nm upon excitation at 360 nm. Our probe showed a strong absorbance at 325 nm in the beginning (red), but a high absorbance signal at 360 nm after treatment with NDM-1 (black) (Fig. S1). Based on this result, our fluorescence assay was performed at 360 nm.
Fluorogenic assay using bacterial colonies (Fluore-C).The fluorogenic assay was initiated by lysing a 2-μl loopful of bacterial isolates with 150 μl of B-PER II (bacterial protein extraction reagent; Thermo Scientific Pierce, Rockford, IL, USA). After vortexing for 1 min, the mixture was left at room temperature for 30 min and centrifuged at 10,000 × g for 5 min. The supernatant (30 μl) was mixed with 100 μl PBS and 13 μl of fluorogenic probe in a flat-bottom 96-well microplate (Greiner Bio-One GmbH, Germany). Subsequently, the fluorescent signal was measured (λex = 360 nm/λem = 465 nm) over 50 min using a fluorometer (Infinite F200pro; Tecan Group Ltd.). The fluorescent signal was calculated by subtracting the value at 0 min from the value at 50 min because our analysis was based on the fluorescence generated by the probe degraded by carbapenemase. For quality control, in every run, we included Escherichia coli ATCC 25922 as a negative control and umbelliferone (Sigma-Aldrich, St. Louis, MO, USA) as a fluorescence control that continuously emitted fluorescence regardless of the presence of carbapenemase.
Fluorogenic assay directly from positive blood culture (Fluore-Direct).The mixture of 1 ml (1.5 × 106 CFU/ml) of each strain and 4 ml of sheep blood was inoculated into Bactec aerobic/F blood culture bottles, and incubated via the Bactec FX (Becton, Dickinson, Franklin Lakes, NJ, USA) blood culture system. After the culture bottle was flagged positive, pellets were obtained using the saponin method with some modifications (23). In detail, 1 ml of culture-positive blood was added to 200 μl of 2% saponin (Sigma-Aldrich) solution and mixed well. After centrifugation at 13,000 × g for 1 min, the supernatant was removed, and the pellet was suspended in 1 ml of sterilized distilled water. The washing steps (from centrifugation to suspension) were repeated twice. Finally, the pellet was suspended in 100 μl of B-PER II. After vortexing for 1 min, it was left at room temperature for 30 min and centrifuged at 10,000 × g for 5 min. A 30-μl amount of supernatant was mixed with 100 μl PBS and 13 μl of fluorogenic probe. The fluorescence was measured in the same way as the Fluore-C assay.
Other phenotypic methods.The mCIM test was performed according to the CLSI guidelines (28) using a 10-μg meropenem disk (Becton, Dickinson).
The CNP was conducted as described in the CLSI guidelines (28) except for one modification as reported previously (29), in which we used another tube, c, which did not contain zinc sulfate. Solutions A and B were prepared as described in the CLSI guidelines. Solution C was prepared similarly to solution B except that no zinc sulfate was added to solution A. Therefore, the following solutions were used: solution A, solution B (solution A plus 6 mg/ml imipenem), and solution C (solution A without zinc plus 6 mg/ml imipenem). A 1-μl loopful of the test isolates was resuspended in 100 μl B-PER II in each of the three 1.5-ml microtubes, a, b, and c. After vortexing the tubes for 5 s, we added 100 μl of solution A to tube a, solution B to tube b, and solution C to tube c. The tubes were vortexed well, incubated at 35°C, and read every 30 min for 2 h. The test result was interpreted as follows: a carbapenemase producer was detected if the color of tubes b and c changed from red to yellow while the color of tube a remained red, an Ambler class B carbapenemase producer was detected if the color of tube b changed from red to yellow while the color of tubes a and c remained red, no carbapenemase was detected if the color of all tubes remained red, and the test was not interpretable if the color of tube a changed to yellow or the color of any tube changed to orange.
The CIT was performed as described previously (6, 30) using a 10-μg meropenem disk (Becton, Dickinson) and 10 μl of three different β-lactamase inhibitors: 60 mg/ml phenylboronic acid (PBA) (Sigma), 0.2 M EDTA (Sigma), and 75 mg/ml cloxacillin (CLX) (Sigma). An 0.5 McFarland suspension of the test isolate was prepared and inoculated into a Mueller-Hinton agar (MHA) plate. Four meropenem disks were placed on the plate at least 24 mm apart, and subsequently 10 μl of each inhibitor solution was added onto the meropenem disk. After incubation for 18 h at 35°C, the diameter of the growth-inhibitory zone around each meropenem disk containing inhibitor was compared with that around the meropenem disk alone. An increase of ≥5 mm in zone diameter was considered positive for each inhibitor. The test result was interpreted as follows: an Ambler class A carbapenemase was detected if positive only for PBA synergy, an Ambler class B carbapenemase was detected if positive only for EDTA synergy, and no carbapenemase was detected if the test was negative for all inhibitors or positive for both PBA and CLX synergy. However, the OXA-type CPE were not included in calculations of CIT performance, as there was no specific inhibitor.
Comparative evaluation.A consensus positive was defined as a positive result from at least three assays among five phenotypic assays: mCIM, CNP, CIT, Fluore-C, and Fluore-Direct. A consensus negative was defined as a negative result from at least three of the five assays. As we defined consensus results as described above, the estimates were called positive percent agreement (PPA) and negative percent agreement (NPA), rather than sensitivity and specificity.
Statistical analysis.The MedCalc statistical software version 18.9 (MedCalc Software bvba, Ostend, Belgium) was used to calculate the Cohen kappa coefficient (κ) and the Spearman correlation coefficient (ρ) and to determine the cutoff value of the fluorogenic assay using receiver operating characteristic (ROC) curve analysis. The Mann-Whitney U test and F-test were used for statistical analysis of fluorescent signals. A P value of <0.05 was considered statistically significant.
Ethics statement.This study was approved by the Institutional Review Board of Seoul St. Mary’s Hospital, Seoul, South Korea (KC17SCSI0569). No informed consent was needed as no personal information was used in this study.
RESULTS
Overall performance.Phenotypic assays (Fluore-C, mCIM, CNP, CIT, and Fluore-Direct) were conducted with 63 CPE and 154 non-CPE. The overall PPAs and NPAs of Fluore-C, mCIM, CNP, CIT, and Fluore-Direct were 100% and 98.7%, 98.3% and 97.5%, 88.1% and 100%, 96.4% and 98.7%, and 98.3% and 98.1%, respectively (Table 1). The results of phenotypic assays for the detection of CPE according to the genotypes and carbapenem MICs of the isolates are summarized in Table 2. The characteristics of 18 isolates which showed discrepant results between phenotypic assays are presented in Table 3. Three isolates (one OXA producer and two GES producers) were determined as consensus negative, even though they were CPE. Another OXA producer was not subject to any consensus result because it showed neither three positive nor three negative results.
Performance of fluorogenic assays, mCIM, CIT, and CNPc
Results of MICs, fluorogenic assays, mCIM, CNP, and CIT in 63 CPE and 154 non-CPE isolatesb
Characteristics of isolates showing discrepant results between phenotypic assaysa
Fluorogenic assay using bacterial colonies (Fluore-C).The PPA, NPA, and κ of Fluore-C were 100%, 98.7%, and 0.977, respectively (Table 1). The Fluore-C detected all isolates determined consensus positive. One isolate of the non-CPE and one of the GES producers showed false-positive results (Table 3).
The areas under the curve (AUCs) were the largest at 50-min fluorescent signal difference, and the value was 0.958 with Fluore-C. The cutoff value for positivity was determined as 1,971 relative fluorescence units (RFU) (see Fig. S2 in the supplemental material) according to the literature for optimal point (31). The distribution of fluorescence was plotted for each carbapenemase or group in Fig. 2A. The fluorescent signal of CPE was significantly higher than that of non-CPE (median RFU [range], 5,814 [240 to 32,009] versus 804 [36 to 2,480], respectively; P < 0.0001, Mann-Whitney U test). The differences in fluorescent signal between subgroups were as follows: KPC versus non-CPE, median RFU (range), 5,709 (3,024 to 8,677) versus 804 (36 to 2,480), respectively (P < 0.0001, Mann-Whitney U test), and OXA versus non-CPE, median RFU (range), 3,223 (240 to 7,282) versus 804 (36 to 2,480), respectively (P = 0.0966, Mann-Whitney U test).
Distribution of fluorescent signals in fluorogenic assays: Fluore-C (A) and Fluore-Direct (B). Abbreviations: RFU, relative fluorescence units; Fluore-C, fluorogenic assay using bacterial colonies; Fluore-Direct, fluorogenic assay directly from positive blood culture; nonproducer, third-generation-cephalosporin-susceptible isolates.
Correlation between fluorescent signal and imipenem, meropenem, and ertapenem MICs.The correlation between fluorescent signal and exponents of MICs was investigated in 63 CPE, which showed a weak correlation (Spearman’s correlation coefficient [ρ], 0.261, 0.232, and 0.123 for imipenem, meropenem, and ertapenem, respectively; P values, 0.0385, 0.0677, and 0.3349, respectively). In addition, we categorized CPE into three groups based on the results of Fluore-C and CNP and distributed them according to MICs (Fig. 3). Although the carbapenem MICs were broadly distributed in the three groups, they were higher in the group testing positive for both Fluore-C and CNP than the group positive for only Fluore-C.
Frequencies of CPE categorized with Fluore-C and CNP according to MICs of imipenem (A), meropenem (B), and ertapenem (C). Abbreviations: CPE, carbapenemase-producing Enterobacteriaceae; Fluore-C, fluorogenic assay using bacterial colonies; CNP, Carba NP test.
Modified carbapenem inactivation method (mCIM).The PPA, NPA, and κ of mCIM were 98.3%, 97.5%, and 0.943, respectively (Table 1). All the 36 KPC-producing, 15 MBL-producing, and 3 coproducing isolates yielded positive results. However, a false-negative result was observed in a GES producer and there were four indeterminate results among ESBL producers (Table 3) with an inhibitory zone diameter of 18 mm, which were similar after the retest.
Carba NP (CNP) test.The PPA, NPA, and κ of CNP were 88.1%, 100%, and 0.915, respectively (Table 1). All the 36 KPC-producing, 9 NDM-producing, and 3 coproducing isolates showed positive results. The CNP did not give false-positive results at all, but a total of seven false-negative results were found in 1 GES, 2 VIM, 1 IMP, and 3 OXA producers (Table 3).
Carbapenemase inhibition test (CIT).The PPA, NPA, and κ of CNP were 96.4%, 98.7%, and 0.951, respectively (Table 1). By using carbapenemase inhibitors, we detected the carbapenemase types of the isolates. The CIT showed two false-negative results among isolates harboring VIM and IMP. Class A results were observed in 100% (36/36) of KPC producers, 67% (2/3) of GES producers, and 100% (2/2) of KPC-OXA coproducers. However, the result of one GES producer was classified as false positive because the consensus result was negative. Class B results were observed in 100% (9/9) of NDM producers, 80% (4/5) of VIM producers, none (0/1) of the IMP producers, and 100% (1/1) of NDM-OXA coproducers. A false-positive class A result was found in non-CP-CRE (Table 3). The six OXA producers were not included in the evaluation of this assay.
Fluorogenic assay directly from positive blood culture (Fluore-Direct).The PPA, NPA, and κ of Fluore-Direct were 98.3%, 98.1%, and 0.954, respectively (Table 1). A single false-negative result was observed with a VIM-producing Citrobacter freundii isolate. Three false positives were found in one OXA producer and two GES producers. As the consensus results were negative, their results were classified as false positives (Table 3).
For Fluore-Direct, the AUC at 50-min fluorescent signal difference was 0.999 with a positive cutoff value of 2,155 RFU (Fig. S2). The distribution of fluorescence values was plotted in Fig. 2B. The Fluore-Direct test also showed a significant difference in fluorescent signals between CPE and non-CPE (median RFU [range], 10,355 [1,689 to 31,463] versus 1,068 [428 to 2,155], respectively; P < 0.0001, Mann-Whitney U test). The differences in fluorescent signal between subgroups were as follows: KPC versus non-CPE, median RFU (range), 11,532 (4,597 to 25,968) versus 1,068 (428 to 2,155), respectively (P < 0.0001, Mann-Whitney U test), and OXA versus non-CPE, median RFU (range), 5,669 (3,335 to 8,242) versus 1,068 (428 to 2,155), respectively (P < 0.0001, Mann-Whitney U test).
DISCUSSION
In this study, we evaluated a novel fluorogenic probe 1 comprising a partial structure of carbapenem conjugated with hydroxybenzyloxy umbelliferone. We compared the assay with other phenotypic assays such as mCIM, CNP, and CIT. The Fluore-C fluorogenic assay using bacterial colonies showed 100% PPA by detecting 59 CPE within 90 min. During our study, a fluorescent probe with a similar carbapenem/umbelliferone structure was reported by another research group (20). It showed good selectivity for MBL but was a poor substrate for KPC and OXA carbapenemases. We presumed that the presence of umbelliferone as a sole leaving group might diminish the sensitivity of their probe to KPC- and OXA-type carbapenemases.
In our study, however, our probe 1, containing hydroxybenzyl ether as an active linker between the carbapenem moiety and fluorophore umbelliferone, showed excellent performance in detecting KPC as well as MBL. The cutoff (1,971 RFU) of Fluore-C was determined by the ROC curve, and the fluorescent signal of CPE was significantly higher than that of non-CPE. As shown in Fig. 2A, the fluorescent signals of MBL were remarkably high, and those of KPC and OXA were also high enough to distinguish from the non-CPE group.
In contrast to a previous fluorogenic probe for third-generation cephalosporins, which showed a positive correlation with cefotaxime and ceftazidime MICs (32), the fluorescent signal of CPE showed weak correlations (ρ = 0.261, 0.232, and 0.123 for imipenem, meropenem, and ertapenem, respectively) with the exponents of MICs of carbapenems. In addition, the distribution of carbapenem MICs was broad among the three categories (Fig. 3) because the carbapenem resistance of CPE was attributed to other resistance mechanisms such as AmpC β-lactamase overproduction, efflux, and porin loss in addition to carbapenemase production contributing to increasing carbapenem MICs (33).
All discrepant results (despite retest) between phenotypic assays using characterized isolates are listed in Table 3. In class A carbapenemases, all KPC producers were detected in all phenotypic tests. For the three GES producers, although two of them were defined as negative, while CNP and mCIM did not detect any of them, Fluore-C, Fluore-Direct, and CIT detected two, three, and two isolates, respectively. This finding was partly in line with a previous study (34) where the sensitivity for GES-4 producers was 0% in CNP and 17% in CIM but 100% in mCIM. In another study involving 22 GES-6 producers, the CIM showed a sensitivity of 50% (35). This phenomenon could be related to the low hydrolytic properties or enzyme types of GES producers. Although GES-type carbapenemase producers are still rare, further studies with a larger number of strains including various GES enzyme types are necessary to elucidate the performance of phenotypic assays for GES producers.
The CLSI guideline stated that the CNP showed >90% sensitivity and mCIM showed >99% sensitivity when detecting class B carbapenemases, including NDM, VIM, and IMP, among Enterobacteriaceae isolates (28). Also, recent studies showed that the sensitivity for VIM producers was 85.7 to 100% with CNP, 71.4 to 100% with CIM, and 100% with mCIM (10, 36, 37) and the sensitivity for IMP producers was 93.8 to 100% with CNP, 93.8 to 100% with CIM, and 100% with mCIM (34, 38–40). Our results also coincided with these results. All nine NDM producers were detected in all phenotypic tests. The fluorogenic assay and mCIM detected all five VIM producers, but CNP and CIT missed two and one, respectively. In addition, although only a single IMP producer was included in our study, it was detected by fluorogenic assay and mCIM but not by CNP and CIT.
For OXA producers, the fluorogenic assay and mCIM showed higher PPA (100%) than CNP (25%). The low ability of the CNP to detect OXA producers is now well known (10, 36, 37, 41). In contrast, the mCIM detects OXA producers well as demonstrated by the CLSI and previous studies (8, 28, 42). Likewise, in our study, among the six OXA producers, five (83%) were detected by mCIM, four (67%) were detected by fluorogenic assay, and one (17%) was detected by CNP.
In addition, we conducted a fluorogenic assay using bacterial pellets obtained directly from positive blood cultures. Using the positive blood culture directly saves time by avoiding subcultures on solid medium. The Fluore-Direct showed high PPA and NPA (98.3% and 98.1%, respectively), suggesting that the fluorogenic assay can be used to detect CPE not only among bacterial isolates but also directly from positive blood cultures. Interestingly, the distribution of fluorescence signals among 36 KPC producers was broad in Fluore-Direct but narrow in Fluore-C (standard deviation, 5,491 RFU versus 1,485 RFU, respectively; P < 0.0001, F-test) for unexplained reasons.
Using the mCIM method, we obtained four indeterminate results among 53 ESBL producers. Their inhibitory zone diameters were 18 mm even after retest. With Fluore-C, we had one positive result of a non-CP-CRE isolate. The above five isolates were subjected to additional PCR to confirm rare carbapenemases such as AIM, GIM, SPM, SIM, DIM, and BIC (25), which tested negative (data not shown). In addition, with the CIT method, a single non-CP-CRE, Enterobacter cloacae, showed a positive result indicating class A carbapenemase. We performed another PCR to identify possible NmcA-type carbapenemase in E. cloacae (43). The PCR result was negative (data not shown), indicating a false synergy with PBA despite the absence of carbapenemase. We considered these false positives lacking in carbapenemase. Further investigation using whole-genome sequencing may be helpful in the future.
At the very beginning of this study, we performed mCIM using bacterial isolates in excess of a 1-μl loopful, and nearly half of non-CP-CRE showed indeterminate results (data not shown). They tested negative when reevaluated with an adequate 1-μl loopful of bacterial isolates, suggesting that adequate amounts of bacterial colonies are needed for the mCIM test as indicated in the CLSI guidelines.
Recently, new or modified methods such as CIM-plus, SMA-mCIM, and meropenem hydrolysis assays using MALDI-TOF MS have been continuously introduced for fast and accurate CPE detection (39–41, 44). However, in many laboratories, CNP is still the most commonly used method for rapid detection of CPE. We here demonstrated that the fluorogenic assay has significant advantages over the CNP. First, the fluorogenic assay showed higher PPA than CNP, especially for OXA producers. Second, it can be interpreted objectively according to the cutoff based on fluorescent signals. Last, the measurement of fluorescence takes only 90 min, which is shorter than CNP, and positive blood culture samples can also be tested without subculture. Although the performance of the CNP test was good for the detection of CPE in positive blood cultures, it required 3 h of incubation after the blood culture tested positive in addition to mechanical lysis via vigorous agitation using microbead tubes and vortexing for 30 min (45). Therefore, the fluorogenic assay with higher PPA than CNP especially for the OXA type can be used to detect CPE within 90 min not only from bacterial colonies but also from positive blood cultures.
In conclusion, we evaluated a novel fluorogenic assay using carbapenem-specific fluorogenic probes with 217 characterized Enterobacteriaceae isolates. We included many non-CP-CRE to evaluate the performance with high stringency. Nonetheless, the novel fluorogenic assay showed good agreement compared with mCIM, CNP, and CIT, and the results can be obtained within 90 min. Moreover, this novel fluorogenic assay can be used directly to test positive blood cultures with the pellet which can also be used for identification with MALDI-TOF MS and the antimicrobial susceptibility test and is, therefore, very useful for the clinical microbiology laboratory.
ACKNOWLEDGMENTS
This work was supported by grants of the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant numbers HI17C1037 and HI16C0443).
No potential conflicts of interest relevant to this article were reported.
FOOTNOTES
- Received 25 June 2019.
- Returned for modification 18 July 2019.
- Accepted 23 October 2019.
- Accepted manuscript posted online 30 October 2019.
Supplemental material is available online only.
- Copyright © 2019 American Society for Microbiology.
