Evaluation of a novel procedure for rapid detection of carbapenemase-producing Enterobacteriaceae (CPE) using the LightMix^® modular carbapenemase kits

Objectives

Evaluation of the LightMix^® modular carbapenemase kits for the rapid detection of carbapenemase-producing Enterobacteriaceae (CPE) and the application of these kits to the direct detection of colonized patients and bacteraemias.

Methods

The modular multiplex PCR kits targeting bla_KPC, bla_NDM, bla_VIM, bla_IMP and bla_OXA-48-like carbapenem resistance genes were evaluated in terms of sensitivity and specificity for carbapenemase resistance in a set of 118 labelled clinical isolates. Among these, 96 were CPE genotypically characterized by PCR and sequencing. The limits of detection were calculated for the different carbapenem resistance genes in terms of cfu/mL. In addition, the kits were used to evaluate colonization of patients by CPE by comparing this assay with the Xpert^® Carba-R Kit on 127 rectal, perirectal and pharyngeal samples. Blood cultures from bacteraemias (4) and spiked blood cultures (23) with genotypically characterized isolates were also evaluated.

Results

The overall sensitivity and specificity of the multiplex PCR assay was 99% and 100%, respectively. The limit of detection for bla_KPC, bla_VIM, bla_IMP and bla_OXA-48-like is 60 cfu/mL and for bla_NDM 500 cfu/mL. The colonization and bacteraemia studies revealed a 100% agreement between the results obtained by this assay and the ones obtained by GeneXpert^®.

Conclusions

The LightMix^® modular carbapenemase kits are highly reliable and utilizable assays for both colonized and septic patients, and can help in the improvement of infection control. Their modular design facilitates cost-effective detection of CPE in hospital settings.

Introduction

Carbapenemase-producing Enterobacteriaceae (CPE) are emerging worldwide.^1,2 Accurate detection of carriers is required, allowing rapid implementation of infectious control measures. Detection of CPE in clinical samples is also important, which requires highly reliable methods compatible with use in a clinical setting, which must be easy to perform, fast and cost effective.³

Classical methods for antimicrobial resistance detection are based on measuring the phenotype and expression levels,^4–6 other tests are based on enzyme hydrolysis by the antibiotics (e.g. proteomic methods), or the Carba-NP test,^7–9 and others use immunological capture assays (e.g. the K-SeT assay).^10,11 However, molecular methods are able to detect specific carbapenemase genes directly from rectal swabs, stools or other colonization sources and have proven excellent for the surveillance of carriers.^12–15

The aim of this study was to evaluate whether the LightMix^® modular carbapenemase kits (TIB Molbiol, Berlin, Germany) can be used to identify CPE by using a modular multiplex PCR targeting bla_KPC, bla_NDM, bla_VIM, bla_IMP and bla_OXA-48-like carbapenem resistance genes, and to evaluate the possible application of these kits as a novel diagnostic method for CPE detection.

Materials and methods

Analytical evaluation

For initial studies of the evaluation of the kits, a total of 118 non-repeat Enterobacteriaceae (Tables 1 and 2), characterized by PCR and DNA sequencing,^16–18 were streaked on a Blood Agar Plate (Becton Dickinson, NY, USA). After an overnight incubation at 37°C, DNA was extracted, by boiling for 30 min at 96°C. The LightMix^® modular assays were run on a LightCycler^® 480 II instrument (Roche Diagnostics).

According to the manufacturer, the KPC assay will detect KPC-1 to KPC-11, the NDM assay detects NDM-1 to NDM-13, NDM-16 and NDM-17, the VIM assay detects variants 1–36, the IMP assay detects most described variants, but will miss IMP-9, -16, -18, -22 and -25-type, and the OXA-48 assay detects major OXA-48-type members, in particular OXA-162, -163, -244, -245, -247 and most likely OXA-181, -204 and -232.

The experimental procedure was always performed with positive and negative controls contained in the kit. Reactions were performed in a LightCycler^® Multiwell Plate 96 (Roche Diagnostics). For each reaction, 15 μL reaction mixture consisted of 1 μL of PCR-grade water (Roche Probes Master Kit), 1 μL of Uracil-DNA Glycosylase (Roche Diagnostics), 0.5 μL of each reagent containing primers and probes (VIM, NDM, OXA-48, KPC and IMP), 0.5 μL of control reaction and 10 μL of Master (LC480 Probes Master, Roche Diagnostics). To 15 μL of reaction mixture, pipetted into each well, 5 μL of control or sample was added, to give a final volume of 20 μL. The plate was sealed, centrifuged and the run started. The protocol consists of three steps: denaturation at 95°C for 5 min, then cycling for 45 cycles of 95°C for 5 s, 60°C for 15 s and 72°C for 15 s, and finally cooling at 40°C for 30 s. The multiplex PCR was completed in <90 min.

For reading results, the analysis was performed using the second derivative maximum method with colour compensation. The channels used for each resistance gene were: 488 for VIM, 510 for NDM, 580 for OXA-48, 610 for KPC and 640 for IMP. For interpretation of results, a run was considered valid when the internal control was detected at a quantification cycle (Cp) <37 cycles. All samples were considered as negative when Cp >37. The negative control must show no signal. For high positive samples (Cp <25) the internal control can be expected not to be visible.

The isolates were subjected to analysis to evaluate the sensitivity and specificity of the multiplex PCR assay. One isolate of each carbapenemase gene group (Klebsiella pneumoniae bla_VIM-1, K. pneumoniae bla_NDM-1, Escherichia coli bla_OXA-48, Enterobacter cloacae bla_KPC-2 and K. pneumoniae bla_IMP-22) was randomly selected to determine the limit of detection (LOD) in terms of cfu/mL. Linear regression analysis was performed in serial 10-fold dilutions in water, from 10⁷ to 10 cfu/mL, analysing each one five times. These dilutions were tested in the same way as the samples. LOD values were determined at a threshold of 40 Cp.

Clinical evaluation for CPE patient colonization screening

In total, 127 rectal, perirectal and pharyngeal samples were analysed during a period of 2 months from February to March 2016 in the Complejo Hospitalario Universitario A Coruña using the LightMix^® modular carbapenemase kits. The results were compared with those obtained using the Xpert^® Carba-R Kit (Cepheid, Sunnyvale, USA).^13,19 When two swabs were collected from the same location and patient, one was used for testing by GeneXpert^® following the manufacturer's instructions and the second was used for testing by the LightMix^® modular carbapenemase kits. The second swab was vortexed in 600 μL of PBS, from which 200 μL was used for extraction. If only one swab was available, 200 μL of sample was collected from the Xpert Carba-R Sample Reagent (5 mL, Cepheid).

Evaluation for CPE bacteraemia confirmation

Twenty-three blood cultures spiked with genotypically characterized isolates were tested, plus four blood cultures from bacteraemic ill patients. Spiked blood cultures were generated by inoculating blood culture flasks (Bactec TM, Aerobic/F Culture Vials, Becton Dickinson, Germany) containing 10 mL of human blood and 200 μL of a suspension of the respective isolate derived from a fresh overnight culture (with a turbidity equivalent to that of a 0.5 McFarland standard). The spiked isolates were four bla_KPC, three bla_VIM, four bla_NDM, four bla_IMP, five bla_OXA-48 and three negative controls: one bla_SHV-12, one bla_CMY-2 and one bla_CTX-M-15. Incubation was performed in an automated system (Bactec FX, Becton Dickinson, Germany) until the flask was flagged positive. In blood cultures obtained directly from the patients, the procedure was applied after the flask flagged was positive and once the GeneXpert procedure had been positive.

To 200 μL of sample (PBS from swab or blood culture) were added 180 μL of MagNa Pure Bacteria Lysis Buffer (Roche Diagnostics) and 20 μL of Proteinase K (20 μg/μL, Sigma–Aldrich, Germany). After incubation at 65°C for 10 min the solution was submitted to automatic DNA extraction, using the MagNa Pure Compact Nucleic Acid Isolation Kit (Roche Diagnostics). After 25 min, the extraction was complete and the multiplex PCR was performed as explained above.

Results and discussion

The analytical evaluation revealed complete agreement between the results obtained for isolates tested by PCR and sequencing and the LightMix^® modular kits, with an overall sensitivity of 99% and specificity of 100%. The assay was able to detect even bla_NDM-7, bla_IMP-22, bla_OXA-181 and bla_OXA-232, which the manufacturer did not claim the kit could detect, but was unable to detect bla_OXA-204. Enterobacteriaceae expressing non-carbapenemase resistance mechanisms, used as negative controls, remained negative with the LightMix^® modular assay.

According to the manufacturer, using a plasmid DNA dilution, the LightMix^® modular assay detects 5 genome equivalent copies per reaction or less, and about 10 copies for NDM genes. We were able to detect the resistance genes in all serial suspensions down to a dilution of 60 cfu/mL, except for bla_NDM-1 with a LOD of 500 cfu/mL. Although the sensitivity is lower than those provided by the manufacturer, all swab and blood culture samples carrying a carbapenemase enzyme, even an NDM-type, tested positive.

Of the 127 rectal, perirectal and pharyngeal samples analysed, 1 was positive for bla_VIM, 3 were positive for bla_NDM and 28 were positive for bla_OXA-48-like, and all the remaining samples were negative. The LightMix^® modular carbapenemase kits and the GeneXpert^® Carba-R Kit were in 100% agreement. Positive and negative predictive values of the multiplex PCR assay were 100%.

Spiked blood cultures were analysed with a 100% agreement between the LightMix^® modular assay and the PCR sequencing. The four blood cultures from bacteraemic patients were considered as bla_OXA-48-like by both the LightMix^® Modular assay and the Xpert^® Carba-R assay.

A limitation of the LightMix^® modular carbapenemase kits (as well for the Xpert^® Carba-R Kit) is that they give false positive results for bla_OXA-163, bla_OXA-405 and will certainly give these for bla_OXA-247 enzymes, which are OXA-48-like enzymes devoid of any carbapenemase activity.²⁰ Another limitation is the need for DNA extraction before running the PCR, extending the hands-on time of the procedure and the time taken for detection, and therefore increasing the secondary costs. However, the main advantage of the LightMix^® modular carbapenemase kits is the direct cost of the technique, which is 75% lower than for the Xpert^® Carba-R Kit (14€/determination compared with 55€ for the Xpert^® Carba-R Kit). In addition, their modular design allows fewer targets to be run, according to the user's choice, lowering the price per test. If only one target is detected, the price is 5€/determination, increasing by 2€/determination up to the five targets for the full panel of carbapenem gene detection.

In conclusion, the LightMix^® modular carbapenemase kits offer a reliable and novel pathway to detect CPE directly from rectal, perirectal and pharyngeal swab samples and from blood cultures, and should prove a valuable diagnostic tool in healthcare settings.

Funding

This work was supported by the Fondo de Investigación Sanitaria (grant number PI15/00860 to G. B.) integrated in the Plan Nacional de I+D and funded by the Instituto de Salud Carlos III (ISCIII). We also thank the Spanish Network for Research in Infectious Diseases (REIPI RD12/0015/0014 to G. B.) and Subdirección General de Redes y Centros de Investigación Cooperativa funded by the Ministerio de Economía y Competitividad, ISCIII, and co-financed by European Development Regional Fund “A Way to achieve Europe” ERDF. M. O. is financially supported by the Río Hortega Programme of the ISCIII (CM15/00155).

Transparency declarations

None to declare.

Acknowledgements

We thank Roche Diagnostics for proving us with the LightMix^® modular carbapenemase kits and for technical support.

References