Suitability of Carbapenem Inactivation Method (CIM) for Detection of IMP Metallo-β-Lactamase-Producing Enterobacteriaceae

LETTER

Carbapenem-resistant Enterobacteriaceae (CRE) strains have become globally distributed in the past decade, resulting in concern over the control of hospital infections and antimicrobial therapies (1, 2). The majority of CRE isolates are carbapenemase-producing Enterobacteriaceae (CPE) strains, so early detection of CPE strains is essential for providing optimal antimicrobial therapies and preventing horizontal transmission. In 2015, van der Zwaluw et al. reported the carbapenem inactivation method (CIM), a new method for detecting carbapenemase producers with high sensitivity and specificity (3). However, it was unclear whether CIM can identify IMP producers with a low carbapenem MIC or non-CPE strains that are highly carbapenem resistant. In this study, we evaluated whether CIM can identify CPE strains independently of their carbapenem MICs.

Were used 233 CPE and 51 non-CPE strains isolated in general hospitals across Japan from 2012 to 2016 and stocked in our laboratory. The strains were collected from blood, sputum, wounds, urine, and feces. Their antibiotic susceptibilities were determined by the agar dilution method in accordance with the recommendations of the CLSI (4). The CPE strains included 191 IMP, 20 KPC, 17 NDM, and 5 OXA-48-like (OXA-48 and OXA-244) producers. The non-CPE strains included 31 extended-spectrum beta-lactamase (CTX-M, SHV, and TEM) producers and 5 plasmid-mediated AmpC β-lactamase (DHA, CMY, and CFE) producers. They were identified by DNA sequence amplification as previously described (5–10). Of the non-CPE strains, 13 highly carbapenem-resistant strains were identified as producing cephalosporinases and lacking porin function by SDS-PAGE, DNA sequencing, and quantitative reverse transcription-PCR (11).

The CIM was conducted as previously described (3). The isolates were cultured on Mueller-Hinton agar (MHA) plates. A full 10-μl inoculation loop of each strain was suspended in 400 μl of sterile distilled water, and a 10-μg meropenem (MEM) susceptibility-testing disk (E-DF85; EIKEN) was immersed in the solution. After incubation at 35°C for 2 h, the disk was removed and placed on an MHA plate inoculated with a 0.5 McFarland standard of Escherichia coli strain ATCC 25922 with a sterile cotton swab. Finally, the plate was incubated overnight at 35°C and the inhibition zone around each disk was measured. Inhibition circles <10 mm in diameter were judged to indicate CIM positivity.

The CIM showed a sensitivity of 100% (233/233) for CPE strains and a specificity of 96.1% (49/51) for non-CPE strains (Table 1). The MICs of MEM for the 191 IMP producers ranged from 0.125 to 32 μg/ml (MIC₅₀ = 0.5 μg/ml, MIC₉₀ = 2 μg/ml), and those of imipenem (IPM) ranged from 0.06 to 8 μg/ml (MIC₅₀ = 0.125 μg/ml, MIC₉₀ = 0.5 μg/ml). All IMP producers were CIM positive, regardless of their low IPM and MEM MICs. However, 13 of the non-CPE strains with porin loss were CIM negative, although their MEM MIC was higher than that of the IMP producers (Table 1). The two non-carbapenemase-producing strains showed false-positive results (one formed no inhibition circle, and the other's circle of inhibition was only 8 mm in diameter); no carbapenemases were detected by carbapenemase-multiplex PCR or loop-mediated isothermal amplification (detecting IMP, VIM, NDM, SPM, AIM, DIM, GIM, SIM, KPC, BIC, OXA-48-like, SME, IMI, and GES) (10–12).

TABLE 1.

Results of CIM and antibiotic sensitivities of CPE and non-CPE strains

Species and β-lactamase	Total no. of isolates	No. of isolates CIM:		MIC50/MIC90
		Positive	Negative	MEM	IPM
CPE isolates
Escherichia coli
IMP	147	147	0	0.5/2	0.125/0.25
KPC	5	5	0	2/64	4/16
NDM	6	6	0	2/32	4/16
OXA-48-like	1	1	0	0.06/0.06	0.5/0.5
Klebsiella pneumoniae
IMP	42	42	0	2/4	0.25/1
KPC	10	10	0	4/32	4/32
NDM	4	4	0	4/8	4/4
OXA-48-like	3	3	0	0.25/0.5	0.5/1
Othersa
IMP	2	2	0	0.5/4	0.125/1
KPC	5	5	0	4/64	4/128
NDM	7	7	0	4/8	4/8
OXA-48-like	1	1	0	0.06/0.06	0.5/0.5
Total IMP producers	191	191	0	0.5/2	0.125/0.25
Total KPC producers	20	20	0	4/64	4/32
Total NDM producers	17	17	0	4/8	4/8
Total OXA-48-like producers	5	5	0	0.25/0.5	0.5/1
Total CPE isolates	233	233	0
Non-CPE isolates
Escherichia coli
ESBL	21	1	20	0.06/8	0.125/4
AmpC	2	0	2	0.06/0.06	0.5/0.5
None	4	0	4	0.06/0.06	0.125/0.125
Klebsiella pneumoniae
ESBL	10	1	9	2/16	4/16
AmpC	2	0	2	0.125/0.125	2/2
None	9	0	9	4/16	4/16
Othersb
ESBL	1	0	1	0.5/1	0.06/0.06
AmpC	1	0	1	0.06/0.06	1/1
None	1	0	1	0.06/0.06	0.25/0.25
Total ESBL producers	32	2	30	2/8	0.5/8
Total AmpC-β-lactamase producers	5	0	5	0.06/0.125	0.5/2
Total non-β-lactamase producers	14	0	14	1/4	0.5/4
Porin loss strains	13	0	13	4/16	8/16
Total non-CPE isolates	51	2	49

^aOthers (CPE isolates) contains Citrobacter freundii, Enterobacter cloacae, Enterobacter asburiae, Klebsiella oxytoca, Providencia rettgeri, and Serratia marcescens.

^bOthers (non-CPE isolates) contains Enterobacter aerogenes, Morganella morganii, and Serratia marcescens.

Several methods for detecting carbapenemase, including the double-disk synergy test, the modified Hodge test, and the Carba NP test, are used in clinical laboratories (13–15). In particular, the Carba NP test is known to detect carbapenemase producers rapidly, i.e., in <2 h. These methods, however, are reported to show false-negative and false-positive results for some CPE strains, such as IMP producers (16, 17). IMP metallo-β-lactamases are widespread in some regions of Asia, especially Japan. We previously reported that the frequency of IMP among clinical isolates of metallo-β-lactamase-producing E. coli was >90% in Japan, and they showed good susceptibility to both IPM and MEM. Thus, IMP producers are often overlooked in clinical laboratories and are also a major problem at medical institutions (18–20).

The CIM has been evaluated worldwide by methods similar to those used in our study (21–23). Some papers have pointed out that the CIM cannot detect some portion of carbapenemase producers or bacterial species, and several modified CIMs have been reported (24, 25). In our study, however, the CIM could effectively detect CPE strains with high specificity, even strains with low carbapenem MICs, such as IMP producers. Furthermore, CRE and non-CPE strains with porin loss, which showed slightly higher IPM and MEM MICs, were all CIM negative. CIM was useful for the identification of CPE strains independently of the carbapenem MIC, requiring only simple procedures and providing clear results, indicated by an inhibitory circle. In addition, the low price and ease of use of the CIM are noteworthy. Thus, the CIM can help to identify CRE infections and provide effective infection control surveillance.