Metallo-β-lactamases among Enterobacteriaceae from routine samples in an Italian tertiary-care hospital and long-term care facilities during 2008


Corresponding author and reprint requests: D. M. Livermore, Antibiotic Resistance Monitoring and Reference Laboratory, Health Protection Agency, Centre for Infections, 61, Colindale Avenue, London NW9 5EQ, UK


Clin Microbiol Infect 2011; 17: 181–189


The emergence of metallo-β-lactamase (MBL)-producing Enterobacteriaceae is a serious public health concern. Producers have been repeatedly isolated from patients and long-term care facility (LTCF) residents around Bolzano, and we sought to assess their prevalence and clinical impact. All routine Enterobacteriaceae isolates from a Bolzano tertiary-care hospital and associated long-term care facilities in 2008 (n = 5500) were screened for MBLs, with case details reviewed for the source patients. In total, 36 producers were obtained from 29 patients, comprising 14 Escherichia coli, six Klebsiella pneumoniae, four Klebsiella oxytoca, four Citrobacter freundii, two Enterobacter cloacae and two Morganella morganii, as well as single Citrobacter amalonaticus, Enterobacter aerogenes, Providencia stuartii and Proteus mirabilis isolates. All were PCR-positive for blaVIM and 25 were PCR-positive for qnrS; 19 non-K. pneumoniae had blaSHV and one had blaCTX-M-group1; 13 were from 12 LTCF residents and 23 were from 17 acute-care patients. All these patients had serious underlying diseases with prolonged hospitalization or LTCF stay; only seven had infections due to the MBL producers, comprising four urinary tract infections, two catheter-related bloodstream infections and one patient with both a surgical site infection and pneumonia. Five patients had more than one MBL-producing organism. Pulsed-field gel electrophoresis identified a cluster of six related E. coli, whereas pairs of K. pneumoniae and C. freundii isolates had >85% profile similarity. Transformants prepared from two isolates were shown to be PCR-positive for blaVIM, qnrS and blaSHV; their plasmids gave similar restriction fragment length polymorphism patterns, and blaVIM-1, qnrS1 and blaSHV-12 were detected by sequencing.


The repeated isolation of metallo-β-lactamase (MBL)-producing Enterobacteriaceae from sites of infection and the gut flora of long-term care facility (LTCF) residents in Bolzano is a major concern [1,2]. MBLs hydrolyse carbapenems and virtually all β-lactams except aztreonam, to which many producers are resistant for other reasons. IMP and VIM are the main transferable MBL types in Europe [3], with VIM MBLs frequent among Klebsiella pneumoniae in Greece [4]. The MBLs so far characterized from Enterobacteriaceae in Bolzano likewise have all proved to be VIM types, as have those from elsewhere in Italy [5–11]. This repeated detection led us to study their distribution in Bolzano more comprehensively by examining the microbiological and epidemiological features of all the MBL-producing isolates recovered from routine samples in 2008 in relation to the clinical details of the patients affected.

Materials and Methods

Screening for MBL-producing Enterobacteriaceae

All Enterobacteriaceae isolates from inpatient and outpatient specimens at the regional microbiology laboratory in Bolzano in 2008 were screened with the Vitek 2 system and the AST-GN015 card (bioMérieux, Marcy l’Etoile, France), which includes a specific ESBL-confirmation test. The laboratory serves an 850-bed hospital, associated LTCFs and a local population of 200 000. Modified Hodge tests with imipenem (10 μg) discs and Escherichia coli ATCC 25922 as the indicator strain [12] and MBL Etests (AB Biodisk, Solna, Sweden) were performed on: (i) all isolates (except Proteeae) with imipenem MICs ≥2 mg/L; (ii) on Proteeae with imipenem MICs ≥8 mg/L; and on (iii) all isolates, irrespective of species, with cefotaxime, ceftazidime or cefepime MICs ≥2 mg/L and a negative ESBL test.

Carbapenemase-producing isolates with aztreonam MICs ≥2 mg/L were subjected to ESBL detection tests with amoxicillin/clavulanic acid (20 μg/10 μg) and aztreonam (30 μg) discs, 15 and 20 mm apart. MICs for MBL producers were confirmed by British Society for Antimicrobial Chemotherapy agar dilution methodology [13].

Medical records for all patients with MBL-producing Enterobacteriaceae were reviewed retrospectively by an infectious diseases specialist. The data collected included age, sex, department and reason for hospitalization, hospitalizations and antibiotic treatments during the preceding 6 months, comorbidities, source and date of isolation of the carbapenemase producer, infected or colonized status [14], treatment outcome on day 7, and overall outcome at 6 months.

Molecular methods

Isolates were typed by pulsed-field gel electrophoresis (PFGE) of XbaI-digested genomic DNA [15], with banding patterns analyzed using BioNumerics software (Applied Maths, Sint-Martens-Latem, Belgium). They were considered to be related if there was ≥85% profile similarity [16]. Previous routine [1] and active-surveillance [2] isolates served as comparators.

Multiplex PCR for blaCTX-M was performed with published primers [17]. Isolates with MBL phenotypes were tested by PCR with consensus primers for blaVIM-1/2 [18] and blaVIM-1 [19]. Published primers were used to seek qnrS [20] and blaSHV [21].

Sequencing of blaVIM and qnrS PCR products was performed with the same primers as used for amplification. To sequence blaSHV, primers SHV-c, ATGCGTTATATTCGCCTGTG, and SHV-d, CTTAGCGTTGCCAGTGCTCG, were used together with those previously described [21]. In all cases, products were first purified with the Geneclean Turbo for PCR Kit (Q-BIOgene, Cambridge, UK) with subsequent sequencing by the GenomeLab Dye terminator Cycle Sequencing system, using the Quick Start Kit (Beckman Coulter, High Wycombe, UK) and a Beckman Coulter CEQ 8000 Genetic Analysis System.

Mega-X E. coli DH10B T1 electrocompetent cells (Invitrogen, Paisley, UK) were transformed by electroporation, using a Bio-Rad Gene-Pulser II (Bio-Rad, Hercules, CA, USA) at 2.0 kV, 200 Ω and 25 μF, with plasmids extracted by the method of Kado and Liu [22] and precipitated twice with ethanol. Transformants were selected on LB agar containing 2 mg/L cefotaxime. Plasmids for characterization were extracted from these transformants and digested with HpaI (Promega, Southampton, UK) and BamHI/SacI (Roche, Mannheim, Germany). The resulting fragments were separated by electrophoresis on 0.7% agarose, with pKPN25 [1] and pKOX105.1 [2] from E. coli DH10B transformants as controls.

Plasmid typing was by PCR (inc/rep PCR) for the major incompatibility groups [23]. Phylogenetic groups of E. coli were determined by PCR [24].

Statistical analysis

Imipenem MICs for isolates from colonized and infected patients were compared using Fisher’s exact test.


Epidemiology and clinical characterization

In total, 36 Enterobacteriaceae from a total of 5500 (0.65%) tested had phenotypes suggesting MBL production, although one E. coli was lost during subculture. These isolates were from 29 patients (Table 1) and belonged to ten species, comprising 14 E. coli, six K. pneumoniae, four Klebsiella oxytoca, four Citrobacter freundii, two Enterobacter cloacae, two Morganella morganii, one Citrobacter amalonaticus, one Enterobacter aerogenes and one Providencia stuartii. One MBL-positive Proteus mirabilis isolated in December 2007 was also included. The proportions of MBL producers among inpatient, LTCF and other outpatient isolates were 23/2300, 13/300 and 0/2400, respectively, with 0/500 among isolates from various small private clinics and other sources.

Table 1.   Epidemiological and clinical characteristics of patients infected or colonized with blaVIM-positive isolates
IsolateSex/age (years)ComorbitiesDepartmentHospitalization in last 6 monthsIn hospital at diagnosisMBL-positive SpecimenClinical relevance of MBL-producerPrior carbapenem therapy (within 6 months)Ongoing antimicrobial therapySpeciesTargeted antimicrobial therapyTreatment outcome (on day 7)Overall outcome (within 6 months)
  1. AMK, amikacin; CIP, ciprofloxacin; COL, colistin; LZD, linezolid; MEM, meropenem; TIG, tigecycline; TZP, piperacillin + tazobactam; VAN, vancomycin; TEC, teicoplanin; LVX, levofloxacin; SXT, trimethoprim + sulphamethoxazole; COL, colistin; NIT, nitrofurantoin; AMP, ampicillin; NET, netilmicin; CAZ, ceftazidime; MTZ, metronidazole; AF, atrial flutter; AMI, acute myocardial infarction; ABMT, allogeneic bone marrow transplantation; COPD, chronic obstructive pulmonary disease; CRBSI, catheter-related bloodstream infection; AML, acute myelogenous leukaemia; CMV, cytomegalovirus; CRF, chronic renal failure; CHF, chronic heart failure; DM, diabetes mellitus; NHL, non-Hodgkin’s lymphoma; HCV, hepatitis C virus infection; SSI, surgical site infection; UTI, urinary tract infection, ICU, intensive care unit; LTCF, long-term care facility, NICU, neonatal intensive care unit; NAV, not available; NAP, not applicable (colonization and not infection).

370M/81DM, senile dementiaLTCFYesNoUrineColonizationNAVNAVProteus mirabilisNAPNAPDead
402F/63Multiple sclerosis, UTILTCFNoNoUrineUTINoYesEscherichia coliNAVNAVAlive
406, 424M/50Demyelinizing encephalopathy, autoimmune thyroiditis, multifocal intestinal infarction, HCV+Intensive care unit, rehabilitation unitYesYesUmbilical drainage, bronchial aspirateSurgical site infection, pneumoniaYesAerosolized COL + TZPEnterobacter cloacae, Klebsiella oxytocaTGC + AMK + LVX + MEMSuccessDead
420F/81AF,CHF, Parkinson’s diseaseLTCFNoNoUrineColonizationNAVNAVEscherichia coliNAPNAPDead
425F/95Gastric cancer, hypertensionLTCFNoNoUrineColonizationNoNoneEscherichia coliNoneNAPDead
426M/61AMI, necrotizing enterocolitisIntensive care unitNoYesBronchial aspirateColonizationNoTZP + LVX + MTZKlebsiella pneumoniaeNoneNAPDead
428M/84StrokeLTCFYesNoUrineColonizationNAVNAVEscherichia coliNAPNAPDead
429F/49Perinatal hypoxia, congenital heart disease, hypothyroidism, Hepatitis CInternal medicineNoYesUrineColonizationNoTIG + LVXCitrobacter freundiiTGC + LVXNAPAlive
450, 451, 520M/5Prune-belly syndromePaediatricsYesNoGland swab, urineColonizationNoNoneEscherichia coli, Klebsiella pneumoniae, Morganella morganiiNoneNAPAlive
430F/69Kidney transplantation type 2 DM, HypertensionNephrologyYesNoUrineColonizationNoNoneCitrobacter freundiiNoneNAPAlive
445, 446M/94COPD, hemiplegiaLTCFYesNoUrineColonizationNAVNAVMorganella morganii, Providencia stuartiiNAPNAPDead
457, 462, 474F/63AML, ABMT, intracranial haemorrhagesHaematologyYesYesBlood, wound swab, vaginal swabCRBSI, wound infection, vaginitisYesTZP + MEM + LZDKlebsiella pneumoniae, Citrobacter amalonaticus, Citrobacter freundiiTGCSuccessAlive
458F/88Senile dementia, UTILTCFNoNoUrineColonizationNAVNAVEscherichia coliNAPNAPAlive
464F/78CHF, aortic bioprosthesis, hypertension, autoimmune thyroiditisInternal medicineYesNoUrineUTINoLVXEscherichia coliSXTSuccessAlive
472F/80AMI, CHF, hypertensionLTCFNoNoUrineColonizationNoNoneKlebsiella oxytocaNAPNAPDead
476F/78Obesity, hip prosthesis, colonic diverticolosisOrthopaedicsNoNoWoundColonizationNoLVX + TECEscherichia coliNoneNAPAlive
480M/76Biliary ducts cancer, hypertension, strokeLTCFYesNoUrineUTINoNoEscherichia coliNITNAVAlive
481M/65AML, AMIHaematologyNoNoBloodCRBSIYesVAN + MEMEnterobacter aerogenesTZP + LVXSuccessAlive
488F/5 monthsPremature infant, necrotizing enterocolitisNICUYesYesUrineColonizationYesCAZ + TECEnterobacter cloacaeNoneNAPAlive
490F/87Hypertension, COPD, senile dementiaLTCFYesNoUrineColonizationNAVNAVEscherichia coliNAPNAPAlive
491F/69CRF, Hypertension, UTI, CMV-infection, vasculitis, Clostridium difficile enterocolitisNephrologyYesYesUrineColonizationNoNoneKlebsiella pneumoniaeNoneNAPAlive
497F/87AMI, CHF, intestinal B-NHLOrthopaedicsNoYesUrineColonizationNoNoneKlebsiella pneumoniaeNoneNAPNAV
505F/79Stroke, type 2 DM, UTI, AMI, breast cancer, hypertensionInternal medicineYesYesUrineColonizationNoNoneEscherichia coliNoneNAPNAV
508F/76Alzheimer’s diseaseLTCFYesNoUrineColonizationNAVNAVKlebsiella pneumoniaeNAPNAPAlive
512, 513M/59HCV+, hypertension, intracranial haemorrhagesNeurosurgeryNoYesBronchial secretionColonizationNoNoneKlebsiella oxytoca, Citrobacter freundiiNoneNAPAlive
515F/84CRF, type 2 DM, anaemia, cerebral vasculopathyInternal medicineYesYesUrineColonizationNoNoneEscherichia coliNoneNAPAlive
518M/78AF, CHF, type 2 DM, brain infarction, urinary incontinenceInternal medicineYesYesUrineUTINoNoneEscherichia coliSXTSuccessAlive
519M/17 daysPremature infantNICUNoYesUrineColonizationNoAMP + NETKlebsiella oxytocaNoneNAPAlive
Not studiedM/89Senile dementia, recurrent UTILTCFNoNoUrineColonizationNAVNAVEscherichia coliNAPNAPDead

Demographic and clinical data for the 29 affected patients are shown in Table 1; 13 isolates were from 12 LTCF residents and 23 from 17 patients in acute-care hospital departments. All the patients had serious underlying diseases with prolonged hospitalization or LTCF stay; only seven had clinical symptoms signifying infection by the MBL-producing isolates: four had urinary tract infections (UTIs), two had catheter-related bloodstream infections and one had both a surgical-site infection and pneumonia; the remaining 22 were considered to be colonized. Multiple MBL-producing species were obtained from five patients (Table 1). One haemodialysis patient, with C. freundii 430 in 2008, had repeatedly yielded MBL-producing C. freundii from urine since 2006; similarly, the paediatric patient with urinary isolates E. coli 450, K. pneumoniae 451 and M. morganii 520 had previously yielded MBL-producing urinary E. coli, K. oxytoca and C. freundii in 2005 and 2006 [1]. Only four patients had been administered carbapenems in the 6 months before the isolation of MBL-producing Enterobacteriaceae.

Trimethoprim plus sulphamethoxazole treatment was successful in the patients with UTIs due to E. coli 464 and 518, both of which were susceptible in vitro. Success was achieved with various antibiotic combinations (Table 1) in the two patients with catheter-associated bloodstream infections: one had K. pneumoniae 457, C. amalonaticus 462 and C. freundii 474, whereas the other had E. aerogenes 481. The patient with surgical-site infection caused by E. cloacae 406 and pneumonia due to K. oxytoca 424 died of underlying disease, although combination therapy appeared initially to be successful. Outcome data were unavailable for the UTI patients with E. coli isolates 402 and 480.

PFGE of the MBL-producing isolates

PFGE (Fig. 1) identified a cluster of six MBL-producing E. coli isolates with ≥85% profile similarity. These were from three residents of a single LTCF (isolates 402, 428 and 490), one orthopaedic unit patient (isolate 476) and two internal medicine ward patients (isolates 515 and 518). Another E. coli isolate (no. 520), from a paediatrics patient, had >85% similarity to a staff-carriage isolate in a previous LTCF surveillance [2]. The cluster of six E. coli isolates all belonged to phylogenetic group B2 as did another three E. coli isolates, whereas the other five MBL-positive E. coli isolates belonged to other groups. Two K. pneumoniae isolates from nephrology and orthopaedic unit patients (isolates 491 and 497) had >90% similarity in PFGE patterns, whereas two of the five C. freundii isolates (nos. 429 and 430 from internal medicine and nephrology unit patients, respectively) were >85% similar. There was no clustering among the E. cloacae or the M. morganii isolates. Single E. coli (isolate 480) and K. oxytoca (isolate 512) were nontypeable by PFGE.

Figure 1.

 Dendrograms illustrating the relatedness of the blaVIM-positive Enterobacteriaceae; the phylogenetic types for the E. coli isolates are also shown. Long-term care facility surveillance culture isolates [2] or previously obtained routine isolates [1] run for comparison are underlined. The gradation on the scale represents the degree of similarity.

MICs for MBL-producing isolates

The MIC ranges of imipenem (2 to >128 mg/L), meropenem (≤0.06 to >32 mg/L) and ertapenem (≤0.12 to >16 mg/L) were very wide for the MBL producers, with most values below the current CLSI susceptibility breakpoint, or in the intermediate range (Table 2). EDTA decreased imipenem MICs by at least four-fold for all except the single P. stuartii isolate and, for 30/35 isolates, this reduction was eight-fold or greater. Imipenem MICs were >8 mg/L for five of ten isolates from patients with infection vs. 5/25 from colonized patients (0.11).

Table 2.   MICs (mg/L) and genes found for MBL-producing isolates
  1. AMC, amoxicillin/clavulanic acid 2:1; AZT, aztreonam; CTX, cefotaxime; CTXC, cefotaxime + 4 mg/L clavulanic acid; CAZC, ceftazidime + 4 mg/L clavulanic acid; CPR, cefpirome; CPRC, cefpirome + 4 mg/L clavulanic acid; FOX, cefoxitin; PIP, piperacillin; TZP, piperacillin + 4 mg/L tazobactam; IME, imipenem + 400 mg/L EDTA; IMI, imipenem; ERP, ertapenem; TOB, tobramycin; GEN, gentamicin; MIN, minocycline; AZT–DDS, double disc synergy test with aztreonam and amoxicillin/clavulanic acid; blaSHV, blaCTX-M, qnrS, PCR for the respective genes; ND, not determined. All other abbreviations are as in Table 1.

406Enterobacter cloacae>64>6464>256>32>256>32>64>64>64164>32>16>8.0>3288>3221
488Enterobacter cloacae>64>640.5256>32>256>32>64>64>6418880.2541140.51NDNDND
481Enterobacter aerogenes>64>64>64>256>32>256>32>64>64>6416>128>32>162324440.54+++
462Citrobacter amalonaticus>6464>64>256>32>256>3264>64>641821184240.51+++
429Citrobacter freundii>6464>6412832256>32>64>64641410.25>8.0162840.51+++
430Citrobacter freundii>6464>64128>32>256>3264>64>640.5820.5>8.082241≤0.5+++
474Citrobacter freundii>64648256>32>256>32>64>64>640.25844>8.081181≤0.5+++
513Citrobacter freundii>6464≤0.12128>32>256>32>64>64>6411641182240.51NDNDND+
402Escherichia coli>6432>64256>32>256>32>64>64>640.251684>8.016441≤0.251++
420Escherichia coli>64320.564>32256>32>64>64>640.1240.50.5>8.082216≤0.251NDNDND+
425Escherichia coli>643264328643232>64160.2520.25≤0.120.58241≤0.25≤0.5+++
428Escherichia coli>6432>64128>32>256>32>64>64>640.12842>8.016442≤0.251++
458Escherichia coli>6432>64128>32256>32>64>64>640.2584218222≤0.25≤0.5+++
464Escherichia coli>6464>6464>32>256>32>64>64>64140.25≤0.12182240.51+++
476Escherichia coli>646464128>32>256>32>64>64>640.251684>8.0>32882≤0.251+++
480Escherichia coli>6464>64>256>32256>3232>64>640.540.250.25>8.0164416≤0.251++
490Escherichia coli>6464646432128>3232>64640.2540.250.25>8.08411≤0.251++
505Escherichia coli>64646464161283232>64320.1240.25≤0.120.54422≤0.251+++
515Escherichia coli>6464>6464>32256>32>64>64320.25421>8.08421≤0.251+++
518Escherichia coli>6464>64128>32>256>32>64>64>640.251682>8.016422≤0.25≤0.5++
520Escherichia coli>64640.25128>32>256>32>64>64>640.25844>8.0>328>324≤0.251NDNDND+
424Klebsiella oxytoca>6432>64256>32>256>32>64>64>640.25161616>8.04≤0.50.51621+++
472Klebsiella oxytoca>6464>646432>256>3264>64>640.581121622411+++
512Klebsiella oxytoca>6464≤0.12128>32>256>32>64>64>640.251684132882≤0.252NDNDND+
519Klebsiella oxytoca>6464>64128>32>256>32>64>64>640.25882≤0.1216242≤0.251++
426Klebsiella pneumoniae>64324128>32>256>32>64>64>640.12844>8.032>6481611+NDND+
450Klebsiella pneumoniae>6432≤0.1264>32256>32>64>64>640.12821182220.5≤0.5NDNDND+
457Klebsiella pneumoniae>643213232>256>32>64>64>640.2540.50.5182120.5≤0.5NDNDND+
491Klebsiella pneumoniae>6464>64128>32>256>3232>64>640.25810.52822411+ND+
497Klebsiella pneumoniae>64>64>6432166432>64>646418180.5422881+ND
508Klebsiella pneumoniae>6464>646432>256>32>64>64640.58112841421+ND+
445Morganella morganii>64>64>64256326432>64>6444160.50.254164>32>322>32+++
451Morganella morganii>64>64≤0.12168483216843220.25≤0.1242142>32NDNDND
370Proteus mirabilis>64640.588828328280.25≤0.12>8.0162481>32NDNDND+
446Providencia stuartii>64>644841644>641622≤0.06≤0.12>8.0814322>32+NDND+

The MIC distribution of aztreonam was strongly bimodal, with nine values ≤1 mg/L and 23 at ≥64 mg/L: 22 of the latter 23 were ESBL producers with synergy between aztreonam and amoxicillin/clavulanic acid; the remaining one, E. cloacae 406, lacked blaSHV and blaCTX-M and had a phenotype implying derepressed AmpC.

Seventeen MBL-producing isolates were highly resistant to ciprofloxacin, with MICs ≥8 mg/L, whereas MICs for six were ≤0.5 mg/L. Aminoglycoside MICs were widely scattered, with more isolates resistant to tobramycin than to amikacin or gentamicin. Most isolates were resistant to minocycline but 28/35 were susceptible to tigecycline at the EUCAST breakpoint of 1 mg/L, with six of the remaining seven inhibited at 2 mg/L. With the exception of E. aerogenes 481 (MIC, 4 mg/L) and inherently-resistant Proteeae, all the isolates were susceptible to colistin, with MICs ≤2 mg/L (Table 2).

Construction of transformants

C. amalonaticus 462 and C. freundii 474 were selected as donors for transformation. Both were from the vaginal secretions of a haematology patient, were PCR-positive for blaVIM-1, qnrS and blaSHV and had ESBLs (Table 2). Approximately 10 000 colonies were obtained in each transformation and three representatives from each were analyzed. Plasmids from C. amalonaticus 462 and C. freundii 474 were designated pCAM462.1 and pCFR474.1 and were PCR-positive for blaVIM-1, qnrS and blaSHV. A second transformant of C. freundii 474 acquired a plasmid, designated pCFR474.3, PCR-positive for blaVIM-1 but negative for qnrS and blaSHV, as also seen for a plasmid from an active-surveillance LTCF K. oxytoca isolate [2].

Imipenem MICs for the transformants were in the range 2–16 mg/L compared to 0.5 mg/L for the E. coli DH10B recipient, and were reduced by at least eight-fold by EDTA; aztreonam MICs for the recipient and the blaSHV-negative transformant with pCFR474.3 were ≤0.125 mg/L, whereas those for the blaSHV-positive transformants with pCAM462.1 and pCFR474.1 were >64 and 0.5 mg/L, respectively. Sequencing showed that the former plasmid encoded SHV-12, whereas the latter encoded an SHV-12 variant with a single amino acid substitution (see below). Ciprofloxacin MICs for the qnrS-positive transformants were 0.25 vs. ≤0.125 mg/L for the recipient and the qnrS-negative transformant. Gentamicin and amikacin MICs were not significantly changed relative to that for E. coli DH10B, whereas the MIC of tobramycin increased from 0.5 to 2–8 mg/L.

Plasmid sizes were: pCAM462.1, 55 kb; pCFR474.1, 60 kb; and pCFR474.3, 30 kb. Restriction fragment length polymorphism after digestion with BamHI/SacI yielded two bands for pCFR474.3 and three bands for the other two plasmids, whereas digestion with HpaI yielded three bands for pCFR474.3, seven for pCAM462.1 and eight for pCFR474.1. Two HpaI digestion-generated bands were common to all of the plasmids and were present also in blaVIM-encoding plasmids of a transformant of an LTCF-active-surveillance K. oxytoca isolate [2] and in a transformed plasmid from a clinical K. pneumoniae strain isolated in 2005 [1]. All these present and previous blaVIM plasmids from Bolzano belonged to IncN.

Sequencing of blaSHV, blaVIM and qnrS in transformants

The blaSHV genes from pCAM462.1 and pCFR474.1 were partially sequenced. The sequence (772 out of 861 nucleotides) in pCAM462.1 corresponded to blaSHV-12, a known ESBL determinant [25], whereas that of pCFR474.1 (655 nucleotides sequenced) differed by an A→G transition, leading to a Ser238Gly substitution. The blaVIM and qnrS genes of pCFR474.1 were partially sequenced and were identical to blaVIM-1 and qnrS1.


Enterobacteriaceae with VIM MBLs are still rare in Italy, but producers have been found repeatedly since 2002 [5–7,9,10,26]. Clonal outbreaks involving K. pneumoniae and E. cloacae with VIM MBLs occurred in Genoa in 2004–5 [8] and in Rome in 2007–8 [11] whereas, in Bolzano, multiple species with IncN blaVIM plasmids have been recorded since 2005 [1,2].

The present study found a 0.65% prevalence of MBL producers among clinical Enterobacteriaceae in Bolzano. All the present source patients had spent >72 h in hospital or in an LTCF before isolation of the producer(s), and all had severe underlying diseases. Only seven had clinically-defined infection with their MBL producer(s), sometimes involving multiple species, whereas 22 were colonized.

Whether the small number of serious infections reflects low virulence or epidemiological reasons remains uncertain; only four source patients had received carbapenems in the 6 months before isolation of an MBL producer.

The treatment of infections due to MBL producers remains controversial. One Greek study found high 14-day mortality in bloodstream infections with VIM β-lactamase-positive K. pneumoniae, associated with ineffective empirical therapy, although only in cases where the imipenem MIC was >4 mg/L [27], whereas another found no mortality difference contingent on the imipenem MIC [28]. Imipenem MICs were >4 mg/L for isolates from seven of ten of the infected patients in the present study, without attributable mortality, although (and in contrast to the Greek series) few patients had life-threatening infections. Thus, both haematology patients with central-venous catheter-(CVC)-related bacteraemias survived despite (in accordance with internal protocols) being empirically administered meropenem plus linezolid (one patient) or vancomycin (the other). One of these two patients, an elderly woman with allogeneic bone marrow transplantation for acute myelogenous leukaemia (AML), was given tigecycline as targeted therapy, whereas piperacillin/tazobactam plus levofloxacin was administered to the other, a 65-year-old male with AML with E. aerogenes bacteraemia. Neither patient showed signs and symptoms of severe sepsis, and CVC removal probably contributed to successful outcomes. Co-trimoxazole therapy was successful in two patients with UTIs due to susceptible strains.

Although some clusters were identified by PFGE (notably six E. coli with >85% similarity), remarkable diversity was apparent among the MBL-producing strains, suggesting horizontal gene transfer as the major factor for dissemination of VIM MBLs. As also found previously in Bolzano [1] and for K. pneumoniae isolates from Greece [29,30], blaVIM-1 was located on various IncN plasmids. Coniugative transfer was not attempted in the present study but was demonstrated previously for several such plasmids [1].

Eight of the 14 VIM-positive E.  coli were from LTCF residents, as were five further isolates belonging to different species. This association led to an active surveillance in the largest LTCF in Bolzano in October 2008, revealing 6.3% of residents and one staff member as being colonized by MBL producers [2].

Twenty-five of the 36 MBL producers also had ESBLs: one E. coli isolate had blaCTXM-group-1, whereas the others mostly had blaSHV, identified in the transformant studied in detail as blaSHV-12. One isolate, C. freundii 474, with aztreonam-clavulanate synergy and a relatively low (8 mg/L) aztreonam MIC, yielded an SHV-12 variant with a Ser238Gly substitution. This may explain reduced aztreonam resistance, because Ser238 is on a key β-strand of the catalytic site, where glycine (as in SHV-1) is associated with much weaker hydrolysis of oxyimino-cephalosporins than serine, as in SHV-12 [31]. An association between blaVIM and blaSHV-12 [5,7,11] or blaSHV-5 [8] was previously reported for clinical Enterobacteriaceae from Italy, although the genes were not on the same plasmid, as in the present study. Of the 36 MBL-producing isolates, 71% had qnrS, one copy of which was partially sequenced and identified as qnrS1. In Europe, qnrS is mainly reported in Salmonella isolates from outside Italy but was recorded in an E. coli isolate from an Italian chicken in 2006 [32]. qnrS1 has been repeatedly found among Enterobacteriaceae in Bolzano, both in those with VIM MBLs and those lacking MBLs but producing ESBLs [1,2], whereas a study in Taiwan found a high prevalence of qnr genes (78.6%) co-existing on plasmids with blaIMP-8 [33]. Such linkages of qnrS1 with MBL and ESBL genes—on plasmids that have spread among species—present a major therapeutic and public health threat.


We wish to thank M. Warner for providing valuable advice and technical assistance, as well as P. Innocenti, L Moroder, R. Meyer, B. Ladinser and the laboratory technicians for their excellent assistance with the screening of MBL-producing enterobacteria. We express our gratitude to E. Pagani and V. Pasquetto for their help with the molecular confirmation of the isolates.

Transparency Declaration

No external funding was provided for this project. DML has shareholdings, or acts as enduring attorney for a shareholder, in AstraZeneca, Dechra, EcoAnimal Health, GlaxoSmithKline, Merck and Pfizer; he has had research contracts, or conference finance in the past 3 years from AstraZeneca, Calixa, Cerexa, Johnson & Johnson, Merck, Novartis, Novexel, Pfizer, Phico, Theravance and Wyeth. He is employed by the Health Protection Agency and is also influenced by their views on antibiotic use. The other authors have no conflicts of interest to declare.