Introduction

The dynamic spread of carbapenemase-producing Enterobacteriaceae, Pseudomonas spp. and Acinetobacter spp. has been observed all over the world in recent years. The predominant carbapenemases include β-lactamases of the Ambler class A (KPCs), metallo-β-lactamases (MBLs) of the class B (VIMs, IMPs, NDMs) and carbapenem-hydrolysing oxacillinases of the class D (OXA-23, OXA-24/40, OXA-48 or OXA-58 types) [1,2,3,4,5,6,7,8,9]. Carbapenemase producers notoriously exhibit multidrug resistance phenotypes, making treatment of infections extremely difficult [10,11,12,13]. Therefore, quick, unambiguous, easy and cheap identification of carbapenemase-producing bacteria has been one of the priorities of modern clinical microbiology. The biochemical Carba NP test for Enterobacteriaceae and Pseudomonas spp., and its variant CarbAcineto for Acinetobacter spp., were proposed to address this challenge in 2012 and 2014, respectively [1, 14, 15]. Both tests were reported to have high specificity and sensitivity for all of the major carbapenemase types [1, 14, 15]; however, it was shown that, for some enzymes, the sensitivity relies strongly on the type of microbial medium used for bacteria cultivation [16,17,18]. The original Carba NP test was soon followed by modified variants, and has been commercialised [19,20,21,22,23,24,25,26].

The aim of this study was to evaluate the Carba NP and CarbAcineto tests for the detection of carbapenemase-producing Enterobacteriaceae, Pseudomonas spp. and Acinetobacter spp. cultivated on different media. Moreover, the Carba NP/CarbAcineto tests were assessed in the routine diagnostics of carbapenemase producers conducted by the National Reference Centre for Susceptibility Testing (NRCST) in Poland in 2014.

Materials and methods

Assessment of the Carba NP and CarbAcineto tests: strain collection and cultivation media

For the initial evaluation of the Carba NP and CarbAcineto tests, a total of 107 strains were used: 81 carbapenemase producers that were well characterised with phenotypic methods, polymerase chain reaction (PCR) and sequencing, and 26 carbapenemase non-producers, including 21 meropenem-non-susceptible strains (Table 1). Both tests were performed by the in-house protocol as previously described [14, 18]. The results presented are those that were recorded after 2 h of incubation time. For cultivation of bacteria, a number of non-selective or screening media and the Mueller–Hinton ready-to-use agar plates, products of six manufacturers, were used (Tables 2 and 3).

Table 1 Isolates used in the Carba NP/CarbAcineto tests evaluation
Full size table
Table 2 The results of the Carba NP test for Enterobacteriaceae isolates (34 carbapenemase producers and 10 carbapenemase non-producers) cultivated on a set of microbial media
Full size table
Table 3 The Carba NP results for Pseudomonas spp. (15 carbapenemase producers and 6 carbapenemase non-producers) and the CarbAcineto results for Acinetobacter spp. (32 carbapenemase producers and 10 carbapenemase non-producers) grown on a set of microbiological media
Full size table

The Carba NP and CarbAcineto tests in routine reference diagnostics

In the second part of this work, the Carba NP and CarbAcineto tests were implemented into the routine activity of the NRCST, which is responsible for the reference diagnostics and surveillance of carbapenemase-producing Gram-negative pathogens in Poland. Both tests were performed along with the phenylboronic acid disc test [27], double-disc synergy test with EDTA (DDST-EDTA) [28], temocillin disc test [29] for the phenotypic detection of KPCs, MBLs and OXA-48 types, respectively [30], spectrophotometric assay for imipenem hydrolysis [31] and PCR for bla KPC-, bla VIM-, bla IMP-, bla NDM-, bla OXA-48-, bla OXA-23-, bla OXA-24/40- and bla OXA-58-like genes [32]. This analysis was done on 1282 isolates in total (Table 4), including 915 Enterobacteriaceae, 309 Pseudomonas spp. and 58 Acinetobacter spp. collected from April to the end of December 2014 in hospitals all over Poland. All these isolates were cultivated only on Columbia blood agar plates.

Table 4 Performance of the Carba NP/CarbAcineto tests in the analysis of 1282 surveillance isolates submitted to the National Reference Centre for Susceptibility Testing (NRCST) from April to December 2014
Full size table

Results

Assessment of the Carba NP and CarbAcineto tests and the cultivation media

For Enterobacteriaceae, the Carba NP test performed well, with the carbapenemase-producing strains cultured on most of the non-selective or screening media (Table 2). False-negative results were obtained only for single isolates with VIMs or OXA-48s grown on three types of chromogenic media. When bacteria were cultivated on Mueller–Hinton (MH) agars, the Carba NP test was positive for all KPC, IMP and all but one OXA-48 producers, while false-negative or questionable results were obtained for NDM- and VIM-producing strains. For a number of carbapenemase-negative strains picked up from some media, the results were questionable and for single strains were false-positive.

The results of the Carba NP test for Pseudomonas spp. strains are presented in Table 3. In the case of carbapenemase (VIM or IMP) producers, the best results were observed for the strains cultivated on the Columbia blood agar, chromID CPS, Brilliance UTI agar (Oxoid) and Brilliance CRE agar (Oxoid). The false-negative results were observed for some isolates cultivated on five other chromogenic media, and on MH (Oxoid) and MHF (bioMérieux) agars. For single carbapenemase-negative strains cultivated on five media, false-positive results were observed.

The results of the CarbAcineto test for Acinetobacter spp. strains are shown in Table 3. As in the case of Enterobacteriaceae and Pseudomonas spp., the impact of the cultivation medium on CarbAcineto was revealed; moreover, for isolates with OXA-like carbapenemases, the correct results were observed more often than for VIM-positive strains (Table 3).

The Carba NP and CarbAcineto tests in routine reference diagnostics

The results of the comparison of the Carba NP/CarbAcineto tests with the standard methodology used by the NRCST, obtained with 1282 bacterial isolates suspected of carbapenemase production, are presented in Table 4. The phenotypic tests and PCRs detected various carbapenemases in 451/915 Enterobacteriaceae, 118/309 Pseudomonas spp. and 54/58 Acinetobacter spp. isolates. For Enterobacteriaceae, the overall sensitivity and positive predictive value (PPV) of Carba NP were 95.8% [95% confidence interval (CI), 93.9–97.6%], and the specificity and negative predictive value (NPV) were 93.3% (95% CI, 90.9–95.6%). For Pseudomonas spp. the sensitivity and PPV were 97.5% (95% CI, 94.6–100%), and the specificity and NPV were 99.0% (95% CI, 97.6–100%). For Acinetobacter spp., the sensitivity and PPV of CarbAcineto were 88.9% (95% CI, 80.4–97.4%).

Discussion

The Carba NP test was described for the first time in 2012 by Nordmann et al. as a rapid, easy and cheap method for carbapenemase detection in Enterobacteriaceae and Pseudomonas spp., with high sensitivity and specificity [1, 15]. In previous studies, the test was evaluated by several groups, often confirming the original observations [2, 19, 33,34,35]. However, one study undermined the sensitivity in the case of MBL-producing strains [2]. The problem turned out to be associated with the cultivation medium used [16,17,18], and the follow-up study by Dortet et al. documented well the impact of microbiological media on the performance of Carba NP [18].

In the first part of this work, well-characterised Enterobacteriaceae, Pseudomonas spp. and Acinetobacter spp. strains with various carbapenemases were used, along with a broad panel of commercially available media. The results clearly illustrated the influence of the media on the Carba NP/CarbAcineto results. According to our experience, the Columbia blood agar is the best choice for the cultivation of bacteria for the Carba NP/CarbAcineto tests, providing the highest sensitivity and clarity of the results. The VIM or NDM producers harvested from most of the MH agars yielded a remarkable number of false-negative results, which were most probably related to lower zinc ions’ concentrations that is crucial for the MBL activity [18]. Owing to the low incidence of OXA-48 producers in Poland so far, only six such isolates were used in this work, and all of these were clearly positive in Carba NP, in contrast to some other reports [17, 25, 34,35,36,37]. This study has definitely confirmed the usefulness of the CarbAcineto test for the detection of the acquired carbapenemases in Acinetobacter spp. [8, 38, 39] and for distinguishing such isolates from hyperproducers of the natural OXA-51 types and from the carbapenemase-negative carbapenem-non-susceptible isolates, which is important for epidemiological reasons.

The Carba NP and CarbAcineto tests have been used in the NRCST from the end of 2013 and March 2014, respectively, and now are included in its routine algorithm for carbapenemase detection as reliable, inexpensive and easy screening tests, reducing remarkably the time for the first feedback information for clinical laboratories. Based on own experience, and in the context of the alarming spread of carbapenemase-producing Gram-negative pathogens in Poland, the NRCST has been recommending Carba NP/CarbAcineto tests for use in diagnostic microbiology laboratories all over the country.