Nosocomial infections due to carbapenem-resistant Enterobacteriaceae (CRE) have been a worldwide problem in the last few decades. Falagas et al.1 have recently reported that the number of deaths was significantly higher in patients with CRE infections in comparison to those with carbapenem-susceptible Enterobacteriaceae infections.

Carbapenemase production has been considered as a major challenge for microbiological laboratories due to the vast number of genes that encodes for production of carbapenemase, high ability to disseminate, and difficulties in detecting.2–5

The objective of this study was to evaluate three different types of phenotypic tests for the detection of CRE from rectal swabs collected from hospitalized patients.

As part of the routine surveillance, 102 rectal swabs for CRE screening were consecutively obtained from patients admitted to intensive care units at Hospital Mãe de Deus, Porto Alegre, from January through February 2014.

Each swab was initially plated on MacConkey agar (bioMérieux, Brazil) for growth, and combined-disk test (CDT) was applied using 10μg meropenem (MER) disk (Oxoid, UK) alone, a MER disk plus 10μl of 40mg/ml phenylboronic acid (PBA) (Sigma–Aldrich, Germany) for KPC inhibition, and a MER disk plus 10μ of 0.1M EDTA (Sigma–Aldrich, Germany) for MBL inhibition. Bacterial growth was screened at 24h (CDT-24h with PBA or EDTA) and 48h (CDT-48h with PBA and EDTA) of incubation at 37°C in ambient air. The results of inhibition were interpreted according to a previous report.6 Original swabs were inoculated on selective/chromogenic ChromID agar (bioMérieux, Brazil) for reading after 24h (ChromID-24h) and 48h (ChromID-48h) of incubation. Suspect colonies were submitted to identification and susceptibility testing. Also, each swab was subsequently suspended in 5ml of tryptic soy broth to which a 10μg ertapenem (ERT) disk was added (selective enrichment broth – SEB test). The inoculated broth was incubated overnight period at 37°C for later plating on MacConkey agar containing 10μg ertapenem and 10μg MER disks.7 Enterobacterial colonies growing around MER and ERT disks were picked up, subcultured, identified to the species level, and subjected to susceptibility testing by the MicroScan automated system (Siemens, USA).

When a positive result was obtained from at least one of the tests, polymerase chain reaction (PCR) for the detection of blaNDM-1, blaOXA-48, blaKPC, blaSPM-1, blaIMP, and blaGES genes was applied.8

Statistical analyses were carried out using SPSS for Windows, version 13.0 (SPSS Inc., Chicago, IL). Kappa coefficient and 95% confidence intervals (CIs) were determined for each category (positive and negative results), in order to determine agreement among the distinct phenotypic tests.9

CRE was detected in 31 (30.4%) of 102 rectal swabs evaluated. For all, Klebsiella pneumoniae was the sole specie that presented a positive result in the phenotypic tests evaluated while blaKPC was the sole carbapenemase gene detected. Positive results were more often observed in CDT-48h with PBA (31 observations), followed by Chromid-48, CDT-24 with PBA, SEB and Chromid-24 (Table 1). After 48h of incubation, a false positive result (none gene detected by PCR) was noted for two samples in CDT with PBA. Also, one sample with no carbapenemase gene grew on ChromID-48h. For all methods, we verified the occurrence of false-negative results (negative result but presence of KPC gene detected), mainly in ChromID-24h (12 cases) and SEB test (6 cases).

Some particular characteristics of each method should be evaluated prior its application as a surveillance method. CDT with PBA shows to be an excellent and rapid method to predict the presence of KPC-producing CRE and 24h incubation was enough to produce a reliable result. In study by Pournaras et al.6 DCT with PBA was able to detect and to differentiate KPC and/or MBL production, with the advantage of obtaining results within one day. It is of note that this type of test does not favor the recovery of the isolate for later analyses, such as species identification, antimicrobial susceptibility testing, or molecular typing. Vrioni et al.3 found 95.1% accuracy with ChromID-24h. However, in our study, one ESBL-producing K. pneumoniae grew on ChromID-48h, resulting in a false-positive detection compared with ChromID-24h. Despite some observed discrepancies, all methods tested showed an almost perfect agreement, as assessed by the kappa coefficient (kappa=0.82; 95% CI 0.75–0.88; p<0.001).

It should be pointed out that the SEB test protocol is recommended by the CDC10 and with modification (by using ERT as inhibitor substrate) by the Brazilian Health Surveillance Agency – ANVISA.11 It has been adopted as standard protocol by Brazilian laboratories in an effort to prevent CRE dissemination.

Finally, our results demonstrate that different tests have similar performance to detect CRE obtained from surveillance rectal swabs. CDT with PBA proved to be a good test, with the limitation of not allowing for bacterial isolation. Use of a selective/chromogenic medium, such as ChromID, may represent a useful tool for microbiology labs, especially after 48h of incubation.