Comparison of two commercial carbapenemase gene confirmatory assays in multiresistant Enterobacteriaceae and Acinetobacter baumannii-complex

Multidrug-resistant Gram-negative bacilli (MDR-GNB) producing carbapenemases are increasing at an alarming speed. Rapid confirmation of carbapenemase type will be an important diagnostic step in clinical microbiology laboratories not only to reduce the risk of transmissions but also for optimising antibiotic therapy in the future. We compared diagnostic reliability of two commercially available molecular assays (Check-Direct CPE vs. AID line probe assay) for detection and typing of carbapenemase genes in 80 well-characterized isolates of MDR-GNB. Respective strains were isolated in various clinical specimens at our clinical microbiology laboratory. The reference standard included confirmation of carbapenemase-production at the molecular level at the German National Reference Laboratory for Multidrug-resistant Gram-negative bacteria (Ruhr-University Bochum, Germany). 53 Enterobacteriaceae and 27 members of the A. baumannii-complex were used in this study. The tested assays appeared highly reliable to confirm carbapenemase-producing Enterobacteriaceae (CPE) with respective sensitivities of 97.7%, but are currently unsuitable for analysis of members of the A. baumannii-complex. Both assays are easy to perform and rapid tools for confirmation and typing of the most common carbapenemase genes in Enterobacteriaceae. Implementation should be possible for any clinical microbiology laboratory with Check-Direct CPE being easier to handle and having less technological requirements.

Treatment of patients suffering from infections with carbapenemase-producing MDR-GNB is difficult as only a few alternative treatment options remain [7]. Being one of the last-line options, the consumption of polymyxins, particularly colistin, almost doubled in Europe between 2009 and 2013 [2]. Carbapenemases are capable to hydrolyze most β-lactams, including carbapenems, and most enzymes are not inhibited by clinically available β-lactamase inhibitors [1,3]. Avibactam, a non-β-lactam β-lactamase inhibitor, recently became commercially available and inhibits class A and partially class D carbapenemases. It is not active against class B carbapenemases [8]. Infections due to carbapenemase-producers show high mortality [9]. Thus, rapid confirmation of carbapenemase type will be an important diagnostic step in clinical microbiology laboratories not only to reduce the risk of transmissions but also for optimising antibiotic therapy for these difficult-to-treat bacteria in the future. A wide array of newer techniques for detection of carbapenemases has become available recently [4]. The goal of this study was to compare diagnostic reliability of two commercially available molecular assays for detection and typing of carbapenemase genes in cultured MDR-GNB-isolates.

AID line probe assay
The AID carbapenemase line probe assay RDB2290 (GenID 1 , Straßberg, Germany) allows the genetic detection of various carbapenemases by reverse hybridization ( Table 2). The assay is able to discriminate between 13 carbapenemases (

Check-Direct CPE
Check-Direct CPE (Check-Points Health, Wageningen, The Netherlands) detects the presence of the carbapenemases KPC, NDM, VIM and OXA-48 by multiplex real-time PCR ( Table 2). The assay discriminates between KPC, OXA-48, and NDM/VIM, it does not report a specific variant of the carbapenemase gene families. As NDM and VIM are detected by the same fluorochrome, it is not possible to differentiate between these two types of carbapenemases. According to the user manual, crude DNA was extracted from bacterial cell suspension by heating at 98˚C for 10 minutes followed by centrifugation at 21255 g for 2 minutes. PCR reactions were performed using a Rotor-Gene Q/Corbett Rotor-Gene 6000 (Qiagen, Hilden, Germany) according to the proposed Rotor-Gene Q program. Results were interpreted per the Check-Points instructions with the Rotor-Gene Q Software (Qiagen, Hilden, Germany). The total time to perform Check-Direct CPE from cultural isolate to result is approximately 150 minutes.

Diagnostic reliability
Among Enterobacteriaceae OXA-48 was the most frequently detected carbapenemase. Among A. baumannii-complex OXA-23 was the predominant carbapenemase. Both assays missed IMP-14 with IMP-14 being in the proposed spectrum of AID line probe assay but not Check-Direct CPE (Table 3). ID and phenotype of the respective isolate (Enterobacter cloacae, MIC imipenem 1 μg/ml, MIC meropenem 4 μg/ml) were confirmed after analysing the organisms from fresh stock cultures. Both assays correctly confirmed 42/43 (97.7%) of the carbapenemase-producing Enterobacteriaceae-strains. One of twenty-five (4%) carbapenemase-positive members of A. baumanniicomplex were detected by AID line probe assay, while Check-Direct CPE detected none (Table 3). Both tests showed no false-positive reactions, neither in carbapenemase-negative isolates, nor in carbapenemase-producing MDR-GNB (e.g. different than those detected by the reference analysis).

Discussion
The main goal of the current study was to find a reliable method for confirmation and typing of carbapenemases in phenotypically suspicious cultured isolates in order to provide guidance for antibiotic therapy and to rapidly institute infection control measures for these difficult-totreat bacteria. Thus, we compared the commercially available AID line probe assay with the Check-Direct CPE multiplex-PCR in well-characterized cultured MDR-GNB-isolates.
Our results show, that the tested assays are highly reliable to confirm CPE as respective sensitivities were 97.7%. Both assays failed to detect IMP-14 with IMP-14 being in the proposed spectrum of AID line probe assay but not Check-Direct CPE.
In line with current epidemiology both tests appeared unsuitable for analysis of members of the A. baumannii-complex. It appears highly desirable to augment these assays with the capability of detecting OXA-23, as this enzyme may also emerge in Escherichia coli [10,11]. However, no false-positive reactions occurred, neither in carbapenemase-negative isolates, nor in carbapenemase-producing MDR-GNB. As the number of carbapenemase-negative A. baumannii-complex strains was very low, the A. baumannii strains with carbapenemases not targeted by the assays are negative specificity controls for this line of bacteria. Table 3. Tested carbapenemase genes and results for gold standard, AID line probe assay and Check-Direct CPE.

Enterobacteriaceae
OXA Assuming that both assays are able to detect their predicted carbapenemase targets, we calculated expected sensitivities using the nationwide German data of the NRL Bochum of the year 2015 [6]. Expected sensitivities were high for both assays (data not shown) being higher for Check-Direct CPE as it possesses a broader spectrum among the most prevalent carbapenemases OXA-48-like, KPC, NDM, and VIM, respectively (Table 2), but this awaits further study.
A recently published study showed also excellent diagnostic reliability using AID line probe assay for cultured isolates. In contrast, direct molecular testing of urine samples revealed problems with specificity/positive predictive values as positive AID line probe assay-results could not be confirmed by culture methods [12]. This fact is also true for using Check-Direct CPE in primary specimens (e.g. perirectal swabs). In a published clinical study, in only 16% of the Check-Direct CPE-positive perirectal swabs a corresponding carbapenemase-producing organisms could be identified by culture methods. Thus, the positive predictive value (PPV) was only 21% [13]. A significant false-positive rate and low PPV are also described in a similar study evaluating Check-Direct CPE for direct analysis of rectal swabs [14].
In general, false-positive results seem to be appearing in many studies performing molecular carbapenemase testing in stool samples or rectal swabs [13,15,16]. Interpretation of such molecular results can be difficult as the epidemiological and clinical relevance is not known [15].
The major strength of our study was the utilization of well-characterized MDR-GNB-isolates that were analysed by the German NRL for Multidrug-resistant Gram-negative bacteria as reference standard. There, confirmation of carbapenemase-production was performed on a molecular level. By intentionally using cultured isolates our aim was to focus not primarily on the hygiene and contact isolation issue using the assays for screening of primary specimens, but on confirmation of carbapenemase-production in phenotypically suspicious isolates of a relevant clinical sample (e.g. blood culture, urine culture), i.e. reduced carbapenem susceptibility and/or positive modified hodge test. So, we suggest that a combination of culture plus susceptibility testing and molecular methods is the ideal workflow.
A limitation of this study is, that the number of tested carbapenemase-negative isolates of the A. baumannii-complex was low as prevalence of carbapenemase-production (92.6%) was high in these MDR-GNB (Table 1).
Summarizing, both assays are easy to perform and rapid tools ideally allowing same day confirmation and typing of the most common carbapenemase genes in Enterobacteriaceae. Implementation should be possible for any clinical microbiology laboratory. Check-Direct CPE seems to be easier to handle in regard to extraction method, technician time, turnaround-time and technological needs.
Among CPE, Check-Direct CPE showed comparable sensitivity. Check-Direct CPE is not capable to distinguish between the metallo-β-lactamases (class B) NDM and VIM because both are detected by the same fluorochrome. In regard to the goal concerning the current needs, differentiation between class A, B and D will be sufficient, as the newly introduced βlactamase inhibitor avibactam shows no activity against class B and partial activity against class D [8]. Currently, we would prefer implementing Check-Direct CPE for confirmation of carbapenemase-producing Enterobacteriaceae in our laboratory. Continuous surveillance of carbapenemase epidemiology is required as the sensitivity of both assays may change unfavourably if some of the rarer carbapenemase-types become more prevalent.