If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Clinical Microbiology, University Hospital Limerick, Limerick, IrelandCentre for Interventions in Infection, Inflammation & Immunity (4i) and Graduate Entry Medical School, University of Limerick, Limerick, Ireland
School of Medicine National University of Ireland Galway, Galway, IrelandCarbapenemase-Producing Enterobacteriaceae (CPE) Reference Laboratory, Department of Medical Microbiology, University Hospital Galway, Galway, Ireland
School of Medicine National University of Ireland Galway, Galway, IrelandCarbapenemase-Producing Enterobacteriaceae (CPE) Reference Laboratory, Department of Medical Microbiology, University Hospital Galway, Galway, Ireland
Department of Clinical Microbiology, University Hospital Limerick, Limerick, IrelandCentre for Interventions in Infection, Inflammation & Immunity (4i) and Graduate Entry Medical School, University of Limerick, Limerick, Ireland
Carbapenemase-producing Enterobacteriaceae (CPE) may cause healthcare-associated infections with high mortality rates. New Delhi metallo-β-lactamase-1 (NDM-1) is among the most recently discovered carbapenemases.
Aim
To report the first outbreak of NDM-1 CPE in Ireland, including microbiological and epidemiological characteristics, and assessing the impact of infection prevention and control measures.
Methods
This was a retrospective microbiological and epidemiological review. Cases were defined as patients with a CPE-positive culture. Contacts were designated as roommates or ward mates.
Findings
This outbreak involved 10 patients with a median age of 71 years (range: 45–90), located in three separate but affiliated healthcare facilities. One patient was infected (the index case); the nine others were colonized. Nine NDM-1-producing Klebsiella pneumoniae, an NDM-1-producing Escherichia coli and a K. pneumoniae carbapenemase (KPC)-producing Enterobacter cloacae were detected between week 24, 2014 and week 37, 2014. Pulsed-field gel electrophoresis demonstrated similarity. NDM-1-positive isolates were meropenem resistant with minimum inhibitory concentrations (MICs) ranging from 12 to 32 μg/mL. All were tigecycline susceptible (MICs ≤1 μg/mL). One isolate was colistin resistant (MIC 4.0 μg/mL; mcr-1 gene not detected). In 2015, four further NDM-1 isolates were detected.
Conclusion
The successful management of this outbreak was achieved via the prompt implementation of enhanced infection prevention and control practices to prevent transmission. These patients did not have a history of travel outside of Ireland, but several had frequent hospitalizations in Ireland, raising concerns regarding the possibility of increasing but unrecognized prevalence of NDM-1 and potential decline in value of travel history as a marker of colonization risk.
Enterobacteriaceae are Gram-negative colonizers of the human gut. Carbapenemase-producing Enterobacteriaceae (CPE) are resistant to most classes of antimicrobials.
New Delhi metallo-β-lactamase-1 (NDM-1) is among the most recently discovered carbapenemase enzymes. The responsible blaNDM-1 gene is thought to have originated in the environment from plant pathogens and is plasmid-borne.
NDM-1 confers broad-spectrum β-lactam resistance mediated by hydrolysis of all β-lactam antimicrobials, with the exception of monobactams, such as aztreonam.
Since first reported as implicated in human disease, NDM-1-producing bacteria have been recovered from numerous infection sites including device-associated infections, intra-abdominal, urinary tract, bloodstream, and surgical wounds.
Characterization of a new metallo-beta-lactamase gene, bla (NDM-1), and a novel erythromycin esterase gene carried on a unique genetic structure in Klebsiella pneumoniae sequence type 14 from India.
Acquisition of NDM-1-producers has been reported as associated with travel to known reservoir areas, notably the Indian subcontinent (Pakistan, India, Sri Lanka) and the Balkan countries, where prevalence of community carriage is estimated to be 5–15%.
Prevalence of faecal carriage of Enterobacteriaceae with NDM-1 carbapenemase at military hospitals in Pakistan, and evaluation of two chromogenic media.
Does broad-spectrum beta-lactam resistance due to NDM-1 herald the end of the antibiotic era for treatment of infections caused by Gram-negative bacteria?.
The NDM isolates identified in Ireland prior to this outbreak were isolated or paired cases from several hospitals countrywide and generally with an identifiable link with travel. Dissemination of the blaNDM-1 gene, like other similar resistance mediators, is facilitated by inadequate infection prevention and control practice in healthcare settings, uncontrolled or poorly controlled antimicrobial use, inadequate practices related to food preparation and water treatment, and poor general sanitation.
The largest reported NDM outbreak to date in a non-endemic country was reported from Poland in 2015, where 374 cases of infection or colonization, with a variety of NDM-producing Enterobacteriaceae, were identified from 40 hospitals over a two-year period.
In this report, we describe what we believe to be the first outbreak of NDM-1-producing Enterobacteriaceae in Ireland, which occurred in 2014.
Methods
Setting
The Department of Clinical Microbiology at University Hospital Limerick (UHL) provides a centralized microbiology service for six acute hospital sites (800 beds; population circa 380,000 people). As an aid to contextualizing this outbreak, it is notable that 48 K. pneumoniae carbapenemase (KPC) and one imipenem-hydrolysing β-lactamase (IMI)-producing isolates were detected at UHL between February 2009 and May 2015, as previously published.
A CPE screening policy was in place before the outbreak began. Patients in our hospital group were screened on admission for CPE if: admitted to the intensive care unit (ICU) or high dependency unit (HDU) at UHL; transferred from another hospital in Ireland; have had an acute admission in the past 12 months to any hospital within our hospital group (except for paediatric, maternity, or orthopaedic); or hospitalized abroad. Haemodialysis patients are screened every three months. Patients in ICU and HDU are screened weekly until discharge.
Study definitions
Cases were defined as patients with a NDM-1 positive culture from any site during their hospitalization. Contacts were designated as room or ward mates.
Microbiological and molecular detection of NDM-1
Since 2011, CPE surveillance at UHL had been performed on stool samples or rectal swabs using KPC-producer selective chromogenic agar (CHROMagar™ KPC, Paris, France). Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (Bruker Diagnostics) identification was performed on all colonies, as previously described.
Antimicrobial susceptibility testing was performed using broth microdilution (ARIS Sensititre® system, Thermo Fisher Scientific, Inc., MA, USA). Elevated carbapenem minimum inhibitory concentrations (MICs) for meropenem and ertapenem were confirmed by E-test (AB Biodisk, Solna, Sweden) following the European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines; ertapenem resistance MIC >1 g/L, meropenem resistance MIC >8 g/L. Isolates with elevated carbapenem MICs were further evaluated using the modified Hodge test (MHT). Commercially available diagnostic kits (Rosco Diagnostica A/S, Taastrip, Denmark) consisting of meropenem discs supplemented with β-lactamase inhibitors (meropenem + dipicolinic acid; meropenem + boronic acid; meropenem + cloxacillin) were used to phenotypically distinguish CPE isolates. Isolates were referred to the National Carbapenemase-Producing Enterobacteriaceae (CPE) Reference Laboratory Service (CPERLS) at University Hospital Galway, Ireland for CPE confirmation by molecular methods. Genetic relationship of NDM-1 isolates was determined by pulsed-field gel electrophoresis (PFGE).
Details of NDM-1-positive patients
A retrospective chart review assessing clinical and epidemiologic characteristics was completed for all patients involved, including: dates of admission, transfers, and hospital discharges; locations within the hospital; procedures and operative notes; use of invasive devices; biochemical and haematological blood test results; antimicrobials received and documentation of a travel history.
Infection control interventions
The isolation of NDM-1-producing K. pneumoniae triggered initiation of the hospital's outbreak management protocol. Rectal swabs or stool samples were obtained from all contacts of the index case. Information leaflets were distributed to all patients and family as appropriate. The Public Health England CPE toolkit was implemented during the outbreak.
All infected or colonized patients were barrier-nursed using long-sleeved disposable gowns and gloves, and single rooms were used when available. Chlorhexidine gluconate wash-cloths were employed for bathing of patients. Dedicated equipment was prioritized for NDM-positive inpatients, both infected and colonized, but was not available for all NDM contacts. A semi-automated electronic data surveillance system, ICNet™ (Baxter, Gloucester, UK), was used to collate the records of the outbreak meetings and patients involved.
All patients identified as CPE positive or as CPE contact were flagged on the ICNet™ system and their medical charts were assigned a CPE alert sticker, placed on the front cover.
During the 13 weeks of this outbreak, an additional 13 KPCs and one oxacillin-hydrolysing carbapenamase (OXA-48) were identified. This was the first OXA isolated at UHL. At a practical level, staff were familiar with the term ‘KPC’, but the introduction of the terms ‘NDM’ and ‘OXA’ created confusion, and the concept of three different types of CPE circulating simultaneously generated alarm among clinical staff. Members of the infection prevention and control team provided additional education sessions at ward level to nursing staff and healthcare assistants. New CPE posters were designed explaining the different CPE types in simple and clear language, and these were placed in the doctors' residence (communal living space) and on all wards. Anecdotal feedback received regarding the posters was positive. An electronic link to the location of the CPE guideline on the hospital intranet was disseminated on a memo to all staff.
Hand hygiene audits were performed with greater frequency in affected areas, which involved twice weekly observational audits at ward level. Enhanced cleaning, twice daily, of all implicated clinical areas and patient equipment was instigated in parallel with increased auditing of cleaning practice. The index cases' room in the ICU and all ward areas, where positive NDM-1 patients had been admitted, underwent routine cleaning followed by hydrogen peroxide vapour decontamination post discharge. A deep clean of the emergency department (ED) including the waiting room and resuscitation areas was performed as seven patients involved in the outbreak had been admitted via the ED. High-touch surfaces such as door handles, bedside lockers, and chairs and bed rails were emphasized by the hospital hygiene nurse manager for cleaning to reduce cross-transmission. An ultraviolet torch was used to assess the quality of cleaning performed and face-to-face feedback regarding cleaning deficits was supplied to cleaning operatives. Environmental sampling was not performed during this outbreak.
A further initiative was introduced involving, on a daily basis, a joint pharmacist/clinical microbiologist handover of all in-house carbapenem prescriptions, and subsequent discussion by the either the microbiology consultant or registrar with clinical teams regarding alternative agents where appropriate.
Results
During the outbreak, between June and September 2014, nine patients with NDM-1-producing K. pneumoniae and one patient with both NDM-1-producing Escherichia coli and KPC-producing Enterobacter cloacae were detected. The isolates were detected in samples from UHL and from two affiliated regional hospitals, located 10 and 40 km away. Prior to this outbreak, no cases of NDM-producing Enterobacteriaceae had been identified in our laboratory. None of the patients in this outbreak were known to have previous colonization with extended-spectrum β-lactamase (ESBL) producers based on previous screening. Three of the patients had been screened for CPE before this outbreak, and had been found to be negative on those occasions.
Clinical specimens that were positive for NDM-1-producing Enterobacteriaceae included mid-stream urine samples (N = 2), rectal swabs (N = 8), and skin biopsy samples (N = 3). NDM-1-producing isolates were meropenem resistant with MICs ranging from 12 to 32 mg/L. All isolates were tigecycline susceptible (MIC ≤1 mg/L). One isolate was colistin-resistant (MIC 4.0 mg/L; mcr-1 gene negative). PFGE demonstrated that the K. pneumoniae isolates were closely related (Figure 1). Multi-locus sequence typing (MLST) was not performed.
Figure 1Pulsed-field gel electrophoresis of 12 Klebsiella pneumoniae isolates. ME 140282 is the index case.
The index case was a community-dwelling Irish female. In the summer of 2014, she was admitted with sepsis. Blood and peritoneal fluid cultures confirmed E. coli. Initial admission was to a six-bed bay in a general medical ward preceding transfer the following day to a single room in the intensive care unit (ICU). Admission screens confirmed prior colonization with meticillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant Enterococcus faecium (VRE). A rectal screen at that time did not detect CPE. Empiric therapy used intravenous (IV) ceftazidime 2 g every 8 h and gentamicin. As she did not respond to initial therapy, treatment was changed to meropenem 1 g IV every 8 h, metronidazole 500 mg IV three times per day and vancomycin 1.5 g IV every 12 h but was de-escalated to meropenem monotherapy.
Seven days post admission, a rectal CPE screen (routine for ICU patients) detected CTX-M extended spectrum β-lactamase (ESBL)-producing K. pneumoniae and NDM-1-producing K. pneumoniae. The patient subsequently developed a skin rash, and diagnostic biopsies were performed on the left forearm. An abscess developed at the skin biopsy site; CTX-M ESBL-producing K. pneumoniae and NDM-1-producing K. pneumoniae were isolated from the abscess discharge. IV tigecycline 50 mg every 12 h was initiated and the site was debrided surgically. The patient died two months following admission secondary to refractory soft tissue sepsis.
Evolution of the outbreak
The patient demographics are summarized in Table I; all were permanent Irish residents. During the outbreak, our routine screening programme was extended to include weekly testing of rectal swabs or stool specimens from patients with epidemiological and environmental links to confirmed CPE-positive patients. CPE screening was performed for a period of four weeks after no new cases of CPE colonization or infection had been detected.
Table IClinical characteristics of patients involved in an outbreak of New Delhi metallo-β-lactamase-1 (NDM-1) carbapenemase-producing Enterobacteriaceae
Patient
Age (years)
Sex
Place of residence
Date admitted
Admitting diagnosis
Treated with meropenem during admission
Specimen
Organism(s) isolated
Carbapenemase enzyme(s) detected
Date of culture
Infected/colonized
Outcome
Previous admission to hospital in previous 12 months
A
44
F
Community
Week 24 (2014)
Peritonitis
Yes
Rectal swab
Klebsiella pneumoniae
NDM-1
Week 25 (2014)
Infected
Died
Yes
B
78
F
Community
Week 24 (2014)
Streptococcus milleri bacteraemia
No
Mid-stream urine
K. pneumoniae
NDM-1
Week 25 (2014)
Colonized
Discharged to LTCF
Yes
C
61
M
Community
Week 25 (2014)
Urinary retention
No
Rectal swab
K. pneumoniae
NDM-1
Week 25 (2014)
Colonized
Discharged to the community
Yes
D
65
F
Residential care facility
Week 30 (2014)
Urinary infection
No
Rectal swab
K. pneumoniae
NDM-1
Week 32 (2014)
Colonized
Discharged to residential care
Yes
E
81
F
Public LTCF
Week 30 (2014)
Respiratory infection
No
Rectal swab
K. pneumoniae
NDM-1
Week 32 (2014)
Colonized
Discharged to LTCF
Yes
F
71
F
Community
Week 30 (2014)
Collapse
No
Rectal swab
K. pneumoniae
NDM-1
Week 32 (2014)
Colonized
Discharged to LTCF
Yes
G
89
F
Private LTCF
Week 32 (2014)
Congestive cardiac failure
No
Rectal swab
Escherichia coli, Enterobacter cloacae
NDM-1, KPC
Week 33 (2014)
Colonized
Discharged to LTCF
Yes
H
90
F
Community
Week 30 (2014)
Respiratory infection
No
Rectal swab
K. pneumoniae
NDM-1
Week 33 (2014)
Colonized
Died
No
I
75
M
Public LTCF
Week 35 (2014
Respiratory infection
No
Rectal swab
K. pneumoniae
NDM-1
Week 35 (2014)
Colonized
Discharged to LTCF
No
J
53
F
Community
Week 37 (2014)
Skin and soft tissue infection
No
Mid-stream urine
K. pneumoniae
NDM-1
Week 37 (2014)
Colonized
Discharged to the community
No
LTCF, long-term care facility; KPC, Klebsiella pneumoniae carbapenemase.
In June 2014, two contacts of the index case prior to her ICU admission were identified (patients B, D). Both had been on the medical ward with the index case during her 24 h admission before transfer to ICU. Patient B was screened, identified as NDM-1-producer positive and was isolated. Patient D was discharged prior to CPE screening. She re-presented for admission in July 2014 and was re-admitted to a six-bedded area on the same medical ward to which she had been admitted in June 2014. CPE screening confirmed that she was NDM-1-producer positive and she was isolated immediately. Patients E, F, and H were ward contacts of the index case, and patients G and I were identified from routine admission rectal CPE screens performed during the outbreak period. Both had been admitted to UHL in the previous 12 months; neither had had contact with the index case.
Two additional NDM-1-producer positive patients were identified during the outbreak period, but that had no apparent epidemiologic link with the outbreak cases. Patient C was identified as CPE positive from a screening rectal swab that was performed at a regional hospital 10 km from UHL. This CPE screen was performed because the patient had been admitted to UHL in the previous 12 months but his last admission to UHL had been almost five months before the outbreak was declared. Patient J was identified as CPE positive from a urine sample collected at a regional hospital 40 km away from UHL; this isolate was determined by PFGE (Figure 1) to be the isolate most distantly related to the other outbreak isolates.
Given the identification of NDM-1 at three different associated hospitals, a decision was made to perform contact tracing and screening of all contacts at each site. As a result of that exercise, during the outbreak, 2204 CPE screens, including contact tracing and routine CPE screening, were processed in our laboratory, which in addition to detecting the NDM-1 isolates, also identified 13 new KPCs and one OXA-48.
Carbapenem consumption
Only the index case had been prescribed meropenem during the current admission; the patient had received five days of meropenem before the isolation of an NDM-positive culture.
NDM-1 isolates identified post outbreak
In week 31, 2015 (i.e. 10 months after the 2014 outbreak ended), NDM-1-producing K. pneumoniae was identified in an mid-stream urine sample from an 81-year-old female residing in a private long-term care facility (LTCF). This patient was a contact of the index case during the outbreak but CPE was not detected at the time. PFGE demonstrated similarity to the 2014 isolates. In week 32, 2015, NDM-1-producing K. pneumoniae was identified from a rectal swab of a 71-year-old public LTCF patient who was known to have been colonized previously with KPC-producing Citrobacter freundii. This isolate did not demonstrate similarity to previous isolates. In week 42, 2015, NDM-1-producing K. pneumoniae was isolated by rectal swab from an admission CPE screen of a patient repatriated from Bosnia. She had not been admitted to UHL previously and had never had any specimens processed in the UHL microbiology laboratory. In week 48, 2015, NDM-1-producing E. coli was detected in a 60-year-old patient who had been recently hospitalized in India (Figure 2). Again, these isolates did not demonstrate similarity to the outbreak strain.
Figure 2Pulsed-field gel electrophoresis of the two NDM-1 Escherichia coli isolates.
The source of the index patient's NDM-1-producer acquisition remains uncertain although acquisition during the hospital admission of June 2014 is considered likely. The index case had been on haemodialysis for 16 years and switched to peritoneal dialysis in 2013 (five months prior to NDM-1 detection). In our care, all haemodialysis patients undergo surveillance CPE screening every three months, but the same screening is not conducted for peritoneal dialysis patients. The index case patient was screened for rectal CPE in December 2013, at which time CPE was not detected. Following her transition to peritoneal dialysis, no further CPE screening was performed prior to her transfer to ICU on this final admission to UHL, at which time CPE was likewise undetected. She had never worked in a healthcare setting nor lived with any healthcare workers. She had no known travel to NDM-1 endemic areas.
In this outbreak, international travel was not a recognized factor, implying that there may be a hospital and/or community burden of blaNDM-1 than had not been previously appreciated. Struelens et al. reviewed 77 NDM-1 producing Enterobacteriaceae reported from 13 European countries from 2008 to 2010.
Among 55 of the cases with recorded travel history, 31 had involved travel to, or admission to a hospital in, India or Pakistan, and five patients had been hospitalized in the Balkan region. Possible nosocomial acquisition accounted for 13 of 77 cases (17%). In contrast, our outbreak more closely resembled the outbreak reported by Borgia et al. that occurred in Brampton, Ontario, Canada where five patients were identified as carrying NDM-1–producing K. pneumoniae; all of them epidemiologically linked with each other, but none with a relevant travel history.
The dominant species in our NDM-1 outbreak was K. pneumoniae (Figure 1). A successfully controlled outbreak in Mexico City reported the isolation of NDM-1-producing E. coli and NDM-1-producing E. cloacae from the same patient in addition to three NDM-1-producing K. pneumoniae isolates derived from three other epidemiologically-related patients. In that outbreak, one plasmid (IncFII) was borne by all of the isolates.
In a large outbreak reported from South Africa (as in our case, also from three acute hospitals), which persisted for 16 weeks in 2012, K. pneumoniae was also the dominant species and accounted for 28/38 isolates (74%) with E. cloacae accounting for the 5/38 (13%).
Currently, there are no data available in Ireland regarding national prevalence of CPE in long-term care facilities (LTCFs) as a national point prevalence study of LTCFs relating to multidrug-resistant organisms has never been performed. However, three of the ten patients in this outbreak were permanent residents of three separate LTCFs (two public, one private) and one other patient was a permanent resident in a residential care facility for adults with learning disabilities (Table I). Such a study is needed, justified by our data and the fact that, between 2009 and 2015, 140 CPE isolates were identified from clinical specimens of which 12 CPE isolates originated in local public (N = 10 isolates) and private (N = 2 isolates) LTCFs.
Influenced by this outbreak, our antimicrobial stewardship has been modified. Overall hospital antibiotic consumption rate in defined daily doses (DDD) per 100 bed-days used (BDU) demonstrates a reduction in carbapenem consumption. Between 2014 and the end of 2015, carbapenem consumption decreased by 21% (2014: 4.43 DDD/100 BDU, 2015: 3.49 DDD/100 BDU). This compares very favourably with a 4% increase from 2013 (4.24 DDD/100 BDU) to 2014, 25% increase from 2012 (3.39 DDD/100 BDU) to 2013, a 36% increase from 2011 (2.50 DDD/100 BDU) to 2012.
In conclusion, this outbreak and our other sporadic isolates indicate the changing epidemiology of NDM-1 CPE. In Ireland, as elsewhere (e.g. Canada), a history of travel to a known endemic area is of decreasing value in identifying persons at risk of colonization or infection with NDM-1 producers. As the successful management of this outbreak demonstrates, prompt infection prevention and control practices are essential to prevent transmission. No staff or environmental screening was performed but extensive resources directed towards education, hand hygiene compliance, environmental disinfection, cleaning standards and reducing carbapenem consumption were successful in controlling rapid in-hospital transmission of NDM-1 producers. The subsequent detection of additional cases, in particular the related isolate from a nursing home resident in week 31, 2015, demonstrates both the difficulty of definitely eradicating these organisms once established in the ‘revolving door’ systems of nursing homes and hospitals, and the fact that these bacteria are becoming more prevalent.
Acknowledgements
We thank the staff of the Microbiology Laboratory at UHL for their expertise in successfully dealing with this outbreak, E. McGrath for performing pulsed-field gel electrophoresis at University Hospital Galway, and S. Guilfoyle, medical secretary at the UHL Department of Clinical Microbiology, for assisting with review of medical charts.
Conflict of interest statement
None declared.
Funding sources
This study was funded in part by an Irish Society of Clinical Microbiologists research bursary, which is supported by Pfizer Ireland.
References
Gupta N.
Limbago B.M.
Patel J.B.
Kallen A.J.
Carbapenem-resistant Enterobacteriaceae: epidemiology and prevention.
Characterization of a new metallo-beta-lactamase gene, bla (NDM-1), and a novel erythromycin esterase gene carried on a unique genetic structure in Klebsiella pneumoniae sequence type 14 from India.
Prevalence of faecal carriage of Enterobacteriaceae with NDM-1 carbapenemase at military hospitals in Pakistan, and evaluation of two chromogenic media.
Does broad-spectrum beta-lactam resistance due to NDM-1 herald the end of the antibiotic era for treatment of infections caused by Gram-negative bacteria?.
30319-X&id=gr1.jpg){kind=link}
30319-X&id=gr2.jpg){kind=link}