The risk factors for infection with multidrug-resistant organisms include a previous invasive procedure, diabetes mellitus, solid tumors, tracheostomy, urinary catheter insertions, and receipt of antipseudomonal penicillin [9]. Because the above factors are common in intensive care unit (ICU) patients, the detection of CRO or CPO colonization, which easily leads to true infections or to the horizontal transfer of carbapenem resistance determinants to other species [10], is very important for infection prevention. According to the study reported in 2012, the prevalence of fecal carriage of carbapenem-resistant Enterobacteriaceae (CRE) was 0.3%; however, to date, there have been no carbapenemase-producing Enterobacteriaceae (CPE) reported in the ICUs of Korean tertiary care hospitals [11]. Furthermore, a Chinese study in 2012 reported the prevalences of CRE and CPE, including Klebsiella pneumonia carbapenemase (KPC)-2, imipenem (IMP)-4, and New Delhi metallo-β-lactamase (NDM)-1 producers, to be 6.6% and 2.6%, respectively [12]. However, these previous studies utilized conventional methods, which were time-consuming and laborious. Recently, the Xpert CARBA-R assay (Cepheid, Sunnyvale, CA, USA) has been introduced for the detection of CPO from clinical samples. This assay is based on a multiplex real-time PCR technique and can detect bla_VIM, bla_IMP, bla_NDM, bla_KPC, and bla_OXA-48-like alleles. However, reports on its clinical application have been scarce [13]. Hence, we applied the CARBA-R assay to determine the colonization rate of CPO in ICU patients in a tertiary care hospital in Korea and compared its results with those of conventional culture methods.

From July to August 2013, a total of 102 clinical samples were collected from ICU patients. Of these, the samples from patients in neonatal ICUs or patients with less than five-day stay in ICU were excluded. Additionally, duplicate samples from the same patients were excluded. In total, 40 clinical samples, including 23 stool samples and 17 rectal swabs, from 40 patients were evaluated. One of the rectal swabs was inoculated into the sample reagent and loaded into the cartridge of the CARBA-R assay according to the manufacturer's instructions, while another swab was inoculated in MacConkey broth containing 1 µg/mL meropenem (MEM) and incubated for 24 hr. The organisms from the broth containing MEM were then sub-cultured on MacConkey agar plates, each containing an MEM disk. Colonies growing near the MEM disks were sub-cultured on MacConkey agar containing 2 µg/mL MEM to ascertain resistance to carbapenem. Species were identified by using Matrix-Assisted Laser Desorption Ionization-Time-of-Flight Mass Spectrometry (Bruker Daltonics, Bremen, Germany). For carbapenemase screening, a modified Hodge test using an IMP disk and a double disk synergy (DDS) test using aminophenyl boronic acid (APBA) and dipicolinic acid (DPA) vs. MEM were performed [14].

Three out of 40 samples (7.5%) tested positive by the CARBA-R assay, i.e., they showed one bla_VIM-, one bla_IMP-, and one bla_KPC-positive signal, indicating that the prevalence of CPO was substantial in these ICU patients (Table 1). By conventional culture, 16 out of 40 samples showed bacterial growth close to the MEM disk on MacConkey agar. Of these, seven isolates (three Klebsiella pneumoniae, three Pseudomonas aeruginosa, and one Pseudomonas monteilii) were resistant to MEM with a minimum inhibitory concentration (MIC) over 2 µg/mL. Among the seven MEM-resistant strains, two K. pneumoniae and three P. aeruginosa were isolated from samples that tested negative by Xpert CARBA-R assay. For these five strains, the modified Hodge test was negative. In addition, enhanced inhibition zones were observed around APBA disks but not around DPA disks for four strains, and one strain showed no inhibition zone around either the APBA and DPA disks. This indicated that over-expression of Amp-C β-lactamase along with porin loss might be the causes of carbapenem resistance in these strains.

One P. monteilii isolate from a bla_VIM-positive sample was negative for the modified Hodge test using an IPM disk but positive for the DDS test, indicating the limitations of the modified Hodge test in carbapenemase screening owing to its relatively low sensitivity [151617]. Additionally, no bacterial growth was observed from the sample showing a positive signal for bla_IMP-, indicating that the CARBA-R assay is sensitive enough to detect CPO even in a rectal swab sample with a very low bacterial concentration. Alternatively, this might indicate a limitation of this study: we followed the MEM resistance cut-off value for Enterobacteriaceae in the CLSI guidelines [18]; however, this cut-off can cause low levels of CPO resistance to be missed. In addition, a recent report showed that the CARBA-R assay does not perform well in the detection of OXA-48-producing Escherichia coli [19]. Therefore, further studies are required to determine the prevalence of CPO in specific settings and to determine the accuracy of the CARBA-R assay with organisms that exhibit a low level of resistance to carbapenems.

Regarding the turn-around time, while over three days were required for the conventional culture, the CARBA-R assay required only about one hour, including 48 min running time.

The prevalences of intestinal colonization by CRO and CRE were 17.5% and 7.5%, respectively, in the ICUs of a tertiary care hospital in Korea, which is higher than those previously reported in similar settings [1112]. However, owing to the limited number of samples, further study will be needed to determine the true prevalence of CPO and CPE in the guts of ICU patients. The Xpert CARBA-R assay was found to be an easy-to-use assay, considering the labor and processing time required for conventional culture. The Xpert CARBA-R assay t should be adopted for surveillance and the determination of CPO colonization rates in clinical settings.