Evaluation of the double-disk synergy test for New Delhi metallo-β-lactamase-1 and other metallo-β-lactamase producing gram-negative bacteria by using metal-ethylenediaminetetraacetic acid complexes



Momoko Fujisaki, Biochemical Research Laboratories, Research and Development Division, Eiken Chemical, 143 Nogi, Nogi-machi Shimotsuga-gun, Tochigi, 329-0114 Japan.

Tel: +81 280 57 0718; fax: +81 280 57 0712; email: momoko_fujisaki@eiken.co.jp


New Delhi metallo-β-lactamase-1 (NDM-1), one of the metallo-β-lactamases (MBLs), has been identified from clinical isolates worldwide. Rapid detection of NDM-1 producers is necessary to prevent their dissemination. Seven types of EDTA complexes were evaluated as MBL inhibitors in double-disk synergy tests (DDSTs), resulting in detection of the first isolate of NDM-1-producing Escherichia coli (NDM-1 Dok01) in Japan. NDM-1 Dok01 was detected when EDTA magnesium disodium salt tetrahydrate (Mg-EDTA), EDTA calcium disodium salt dihydrate, EDTA cobalt disodium salt tetrahydrate and EDTA copper disodium salt tetrahydrate were used as MBL inhibitors. The sensitivity and specificity of DDSTs using Mg-EDTA for 75 MBL producers and 25 non-MBL producers were 96.0% and 100%, respectively. These findings indicate that the DDST method using Mg-EDTA can detect MBL-producing strains, including NDM-1 producers.

A. baumannii

Acinetobacter baumannii


American Type Culture Collection




EDTA calcium disodium salt dihydrate

C. freundii

Citrobacter freundii


EDTA cobalt disodium salt tetrahydrate


EDTA copper disodium salt tetrahydrate


double-disk synergy test

E. cloacae

Enterobacter cloacae

E. coli

Escherichia coli


EDTA iron sodium salt trihydrate








EDTA magnesium disodium salt tetrahydrate


double-disk synergy tests using 10 mg EDTA magnesium disodium salt tetrahydrate with ceftazidime


double-disk synergy tests using 10 mg EDTA magnesium disodium salt tetrahydrate with imipenem


EDTA manganese disodium salt trihydrate


New Delhi metallo-β-lactamase-1




sodium mercaptoacetic acid


EDTA zinc disodium salt tetrahydrate

Metallo-β-lactamases are Ambler class B enzymes and hydrolyze broad-spectrum β-lactam agents, including third generation cephalosporins and carbapenems. Since the early 1990s, researchers all over the world have reported new MBL-encoding genes in gram-negative bacilli, most commonly Pseudomonas spp., Acinetobacter spp., and Enterobacteriaceae [1]. MBL antimicrobial resistance genes are carried on mobile genetic elements, allowing transfer of the resistance genes to various strains and species of bacteria. The MBL genes may spread rapidly to clinically important pathogens; nosocomial outbreaks caused by MBL-producing K. pneumoniae have been reported [2].

New Delhi metallo-β-lactamase-1 was first identified in 2008 in a single isolate of K. pneumoniae that had been recovered from a patient who was transferred to Sweden after treatment in a hospital in New Delhi [3]. Following the initial report from Sweden, NDM-1-producing organisms have been isolated from clinical specimens in many countries including the UK, India, Pakistan, USA, Canada, Belgium, France, Australia and others [4-9]. NDM-1 positive bacteria have been found not only in clinical specimens, but also in drinking water and seepage in New Delhi [10]. The first case of a NDM-1 producing E. coli (NDM-1 Dok01) infection in Japan was reported in 2010 [11]. This organism was isolated from the blood culture of a patient who had been hospitalized in India. The complete sequence of the NDM-1-bearing plasmid was also reported (GenBank accession number AP012208) [12].

Rapid detection of MBL-producing strains, including NDM-1 producers, is necessary to prevent their dissemination and associated nosocomial infections. Researchers have developed several phenotypic methods to detect MBL production. These tests include DDSTs using 2-mercaptoacetic acid or EDTA, combined disk tests with dipicolinic acid or EDTA, Etest MBL (BioMérieux, Durham, NC, USA) and the modified Hodge test [13-17]. The DDSTs using SMA with CAZ or IPM disks are simple methods and commonly used in clinical laboratories in Japan. However, the growth-inhibitory zone does not enhance sufficiently when the DDST using SMA with CAZ is performed for NDM-1 Dok01. In addition, with IPM disks, the results are equivocal because the enhancement of the zone of inhibition is only 4 mm, which researchers have interpreted as negative [11]. Aoki et al. reported that calcium disodium EDTA, a metal–EDTA complex that incorporates calcium ions into EDTA, is an effective inhibitor of MBL [18]. The purpose of this study was to evaluate the efficacy of detection of MBL, including NDM-1, of DDSTs using seven kinds of metal-EDTA complexes.


Bacterial strains

NDM-1 Dok01 was isolated at Dokkyo Medical University Hospital. K. pneumoniae ATCC BAA-2146 was used as a quality control strain that produces NDM-1. Strains evaluated were stock cultures of known MBL-producing strains of 46 P. aeruginosa, 7 A. baumannii, 5 P. putida, 3 E. coli, 2 Achromobacter xylosoxidans, 2 E. cloacae, 2 Serratia marcescens, 2 K. pneumoniae, 1 K. oxytoca, 1 Citrobacter freundii, 1 Pseudomonas spp., and 1 Acinetobacter spp. Non-MBL producing strains of 7 K. pneumoniae, 1 K. oxytoca, 6 E. coli, 3 C. freundii, 4 P. aeruginosa, and 4 A. baumannii were also evaluated.

Antimicrobial susceptibility testing

Minimum inhibitory concentrations were determined by the broth microdilution method, which was performed on Dry Plate Eiken DPD1 (Eiken Chemical, Tokyo, Japan) according to the manufacturer's instructions.

Comparison of metallo-β-lactamase inhibitors

Sodium mercaptoacetic acid and seven types of metal-EDTA complexes were used as MBL inhibitors. Metallo-β-lactamase SMA Eiken (SMA disk; Eiken Chemical) contains 3 mg of SMA. Ca-EDTA, Mg-EDTA, Co-EDTA, Cu-EDTA, Mn-EDTA, Fe-EDTA and Zn-EDTA were purchased from Dojindo Laboratories (Dojindo Laboratories, Kumamoto, Japan). These seven metal-EDTA complexes were dissolved in water at concentrations that provided maximum solubility. The concentrations of each metal-EDTA complex were as follows: 480 mg/mL Ca-EDTA, 480 mg/mL Mg-EDTA, 480 mg/mL Co-EDTA, 600 mg/mL Cu-EDTA, 400 mg/mL Mn-EDTA, 80 mg/mL Fe-EDTA and 480 mg/mL Zn-EDTA.

Phenotypic detection of metallo-β-lactamase-producing strain

The DDSTs were performed as described previously [13, 19]. A 0.5 McFarland bacterial suspension was inoculated on a Mueller Hinton agar plate (Eiken Chemical). Antimicrobial disks containing either 30 µg CAZ, 10 µg IPM, 10 µg panipenem, 10 µg meropenem, 10 µg biapenem, 10 µg doripenem or 10 µg tebipenem (Eiken Chemical) were used as substrates. Two disks of an antimicrobial agent were placed at least 30 mm apart on a Mueller Hinton agar plate and a blank or SMA disk placed either 7, 10, 15, or 20 mm from the antimicrobial disks (measured from center to center). Twenty-five microliters of each metal-EDTA solution was added to a blank disk. After incubation at 35°C for 16–18 hrs, the appearance of a ≥5 mm enhanced zone around the antimicrobial disk near the inhibitor disk was classified as positive (Fig. 1).

Figure 1.

DDST using inhibitor (metal-EDTA complexes or SMA) disks with antimicrobial disks 15 mm (center to center) apart. NDM-1 Dok01 was used as a test strain. When enhancement of growth inhibition (d = a − b) is more than 5 mm, the test is judged positive.


Using an SMA disk and seven types of metal-EDTA disks, DDSTs were performed for seven MBL producers carrying NDM-1, IMP-1, VIM-2 and IMP-11 and three non-MBL producers carrying KPC, CTX-M-2 and chromosomal AmpC (Table 1). CAZ or IPM disks were placed 15 mm from the metal-EDTA disks and the resultant enhancement of the zone of growth inhibition evaluated. Two NDM-1 producers showed negative results when SMA disks were used. However, DDSTs using Mg-EDTA, Ca-EDTA, Co-EDTA or Cu-EDTA were positive for NDM-1 producers when IPM disks were used. Regarding IMP-1, VIM-2 and IMP-11 producers, Mg-EDTA and Cu-EDTA inhibited all five MBLs in the DDSTs using CAZ. There were no false positive results for the three non-MBL producers. Because P. aeruginosa 7117 was positive only when Mg-EDTA and IPM were used, Mg-EDTA was selected for further studies.

Table 1. DDST results using SMA or seven metal EDTAs with CAZ or IPM for seven MBL producers and three non-MBL producers
 StrainType of β-lactamaseAntimicrobial diskDDST result
SMA (3 mg)Mg-EDTA (12 mg)Ca-EDTA (12 mg)Co-EDTA (12 mg)Mn-EDTA (10 mg)Cu-EDTA (15 mg)Fe-EDTA (2 mg)Zn-EDTA (12 mg)MIC (µg/mL)
  1. +, enhancement of growth inhibition ≥5 mm; −, enhancement of growth inhibition <5 mm.
E. coliDok01NDM-1IPM++++8
K. pneumoniaeATCC BAA-2146NDM-1IPM+++++32
K. pneumoniae4635IMP-1IPM+++2
P. aeruginosa7102VIM-2IPM++++++>32
P. aeruginosa7103VIM-2IPM++++++>32
P. aeruginosa7116IMP-11IPM>32
P. aeruginosa7117IMP-11IPM+>32
K. pneumoniaeATCC BAA-1705KPCIPM16
E. coli6663CTX-M-2IPM≤1
P. aeruginosa7149Chromosomal AmpCIPM2

First, the appropriate concentration of Mg-EDTA for detecting MBL when a Mg-EDTA disk was placed 15 mm from an IPM disk was evaluated. A. baumannii 7170 carrying blaIMP-1 was negative when 8 mg Mg-EDTA disks were used with IPM disks and positive when 10 mg Mg-EDTA disks were used with IPM disks. Therefore, a disk content of 10 mg Mg-EDTA was selected for the subsequent experiments.

Next, the optimal distance between antimicrobial and Mg-EDTA disks was evaluated. K pneumoniae ATCC BAA-2146 was used as a positive control strain for NDM-1 producers, and A. baumannii 7170 as a weak positive control strain for IMP-1 producers. Two strains producing either NDM-1 or IMP-1 were positive when 10 mg Mg-EDTA disks were placed 15 mm away from the IPM disks; however, they were negative when the Mg-EDTA disks were placed 20 mm away from the IPM disks. Therefore, it was decided that the Mg-EDTA and IPM disk would be placed 15 mm apart for the subsequent experiments.

To evaluate the efficiency of Mg-EDTA disks, 75 stock cultures carrying the various MBL genes and 25 stock cultures carrying other β-lactamase genes were tested by DDSTs using 10 mg MgEDTA–IPM or MgEDTA–CAZ. Positive results for MgEDTA–CAZ were obtained in 69 test strains (92.0%) and for MgEDTA–IPM in 58 test strains (77.3%) (Table 2). The results for MgEDTA–IPM and MgEDTA–CAZ were discordant for 16 MBL producers (Table 3). There were no false positive results for MgEDTA–IPM and MgEDTA–CAZ. Two P. aeruginosa carrying VIM-2 and one E. cloacae carrying IMP-1 had negative results with MgEDTA–IPM and MgEDTA–CAZ (Table 4); they were also negative by the SMA disk method. However, two false negative P. aeruginosa became positive when biapenem and doripenem were used with Mg-EDTA, and one false negative E. cloacae became positive when panipenem and meropenem were used as substrates.

Table 2. DDST results using 10 mg Mg-EDTA with CAZ or IPM on 75 MBL producers and 25 non-MBL producers
StrainType of β-lactamaseNo. testedNo. of isolates with positive result evaluated by DDST
10mg Mg-EDTA
MBL producer
P. aeruginosaIMP-127272427
 Chromosomal AmpC + IMP-15555
P. putidaIMP-15555
Pseudomonas spp.IMP-11101
A. baumanniiIMP-12222
 Chromosomal AmpC + IMP-12122
Acinetobacter spp.IMP-11111
A. xylosoxidansIMP-12202
E. coliIMP-13313
K. pneumoniaeIMP-12202
K. oxytocaIMP-11111
S. marcescensChromosomal AmpC + IMP-12222
E. cloacaeChromosomal AmpC + IMP-12101
C. freundiiChromosomal AmpC + IMP-11101
Total 75695872
Non-MBL producer     
P. aeruginosaChromosomal AmpC4000
A. baumanniiChromosomal AmpC1000
 Chromosomal AmpC + CTX-M3000
K. pneumoniaeSHV-261000
 DHA-1 + CTX-M-91000
E. coliCTX-M-22000
K. oxytocaDHA-1 + RbiA1000
C. freundiiChromosomal AmpC2000
 Chromosomal AmpC + CTX-M1000
Total 25000
Sensitivity  92.0%77.3%96.0%
Specificity  100%100%100%
Table 3. Strains with discordant results between MgEDTA-IPM and MgEDTA-CAZ by DDST
Strain Type of β-lactamaseAntimicrobial diskDDST resultb: zone diameter(mm)MIC (µg/mL)
  1. +, enhancement of growth inhibition ≥ 5 mm; −, enhancement of growth inhibition < 5 mm.
E. coli7142IMP-1IPM198
E. coli7143IMP-1IPM222
K. pneumoniae7144IMP-1IPM214
E. cloacae7172Chromosomal AmpC + IMP-1IPM222
C. freundii7174Chromosomal AmpC + IMP-1IPM174
A. xylosoxidans7112IMP-1IPM16>32
A. xylosoxidans7113IMP-1IPM1432
P. aeruginosa7116IMP-11IPM0>32
P. aeruginosa7118IMP-11IPM0>32
P. aeruginosa9156IMP-1IPM0>32
P. aeruginosa9184IMP-1IPM0>32
P. aeruginosa9989IMP-1IPM0>32
Pseudomonas spp.9995IMP-1IPM064
E. coliDok01NDM-1IPM+178
K. pneumoniaeATCC BAA-2146NDM-1IPM+1632
A. baumannii7170Chromosomal AmpC + IMP-1IPM+164
Table 4. DDST results using Mg-EDTA or SMA with ceftazidime or six carbapenems negative either by MgEDTA-IPM or MgEDTA-CAZ
Strain Antimicrobial diskb: zone diameter (mm)MgEDTA, 10 mgSMA, 3 mg
a: enhanced zone (mm)d: difference (mm)DDST resulta: enhanced zone (mm)d: difference (mm)DDST result
  1. +, enhancement of growth inhibition (d) ≥5 mm; −, enhancement of growth inhibition (d) <5 mm.
P. aeruginosa7123Ceftazidime12131131
P. aeruginosa7124Ceftazidime00000
E. cloacae7173Ceftazidime00000


After NDM-1 Dok01 was reported, two NDM-1-producing K. pneumoniae were identified by government-instigated surveillance in Japan. These isolates were collected from elderly people who had not recently traveled abroad and had had no contact with the Indian subcontinent. Although NDM-1 producers from clinical isolates are rare in Japan, accurate screening methods to detect them are needed to prevent their further transmission in both hospitals and communities.

Many clinical laboratories perform confirmatory tests for MBL production against carbapenem-resistant strains [20]. The DDST using SMA is the most convenient of the phenotypic MBL detection methods. However, the growth-inhibitory zone between IPM and the SMA disks is not large enough to be classified as positive with NDM-1 Dok01 [11]. In contrast to SMA disks, DDSTs using IPM disks and Mg-EDTA, Ca-EDTA, Co-EDTA or Cu-EDTA detected two NDM-1 producers. In addition, the DDSTs using Mg-EDTA had high sensitivity (96.0%) and specificity (100%) for 75 MBL producers and 25 non-MBL producers. Galani et al. reported that combined disk test with CAZ and EDTA (750 µg), and DDSTs with IPM disks 10 mm away from EDTA disks have high sensitivity (97.9–100%) and specificity (91.9–96%) in Enterobacteriaceae [14]. That we obtained similar sensitivity and specificity demonstrates that Mg-EDTA can be used as a MBL inhibitor.

Several reports have indicated that AmpC β-lactamase may cause false negative results in DDSTs using SMA [20, 21]. Arakawa et al. also reported that some MBL-producing gram-negative bacilli are difficult to detect. Because they have a low level of resistance to IPM, the expansion of the zone of inhibition is inconclusive [13]. In our study, only 3 of 75 strains were false negative by both MgEDTA–CAZ and MgEDTA–IPM; these three strains were also false negative in DDSTs using SMA. Two false negative P. aeruginosa strains were resistant to six carbapenems and one false negative E. cloacae was resistant to CAZ but susceptible to six carbapenems. Carbapenem resistance in P. aeruginosa is considered to be associated with loss of OprD outer membrane proteins and/or overexpression of active efflux systems in combination with strong expression of AmpC β-lactamase [22]. Furthermore, IPM induces expression of AmpC β-lactamase in P. aeruginosa more strongly than does doripenem [23]. Other resistant mechanisms, such as AmpC β-lactamase, may cause false negative results; such additional mechanisms could mask MBL inhibition by Mg-EDTA. Further studies are needed to investigate the reasons for false positives.

We found that the sensitivity of MgEDTA–CAZ is higher than that of MgEDTA–IPM. This makes CAZ preferable to IPM as a substrate in DDSTs. However, one IMP-1-producing A. baumannii and two NDM-1-producing Enterobacteriaceae were positive when IPM was used, but negative when CAZ was used. Kim et al. have reported that, because these organisms have other CAZ resistant mechanisms such as ESBL and AmpC β-lactamase production, DDSTs using CAZ have difficulty detecting MBL-producing Acinetobacter [21]. Therefore, DDSTs using Mg-EDTA should use both IPM and CAZ disks as substrates in order to further reduce false negative results.

False positive results reportedly also occur with MBL phenotypic methods using EDTA and IPM. It is believed that such false positive results are attributable to increasing membrane permeability caused by chelating agents [24, 25] and the anti-bacterial activity of EDTA [19, 24, 25]. DDSTs using Mg-EDTA yielded no false positive results among 25 non-MBL producers. The disk content of Mg-EDTA was 10 mg, this concentration being higher than that of the EDTA was used in previous reports. Because false positive results were confirmed for P. aeruginosa and Acinetobacter spp. by the Etest MBL and combined disk test, DDST using Mg-EDTA should be evaluated for specificity using non-MBL-producing P. aeruginosa or Acinetobacter spp.

In conclusion, this is the first report to evaluate several metal-EDTA complexes as inhibitors of MBL. Use of Mg-EDTA in DDSTs is the most useful phenotypic method for detecting MBL producers, including NDM-1 producing strains, in clinical laboratories. Because we tested only two NDM-1 producers by the Mg-EDTA DDST method, other NDM-1 producers should be confirmed by subsequent studies in actual clinical practice.


M. Fujisaki and S. Sadamoto are employees of Eiken Chemical.