Carbapenem-Resistant Klebsiella pneumoniae Producing New Delhi Metallo-β-Lactamase at an

Acute Care Hospital, Colorado, 2012

Author(s): Erin E. Epson, MD; Larissa M. Pisney, MD; Joyanna M. Wendt, MD; Duncan R.
MacCannell, PhD; Sarah J. Janelle, MPH; Brandon Kitchel, MS; J. Kamile Rasheed, PhD; Brandi
M. Limbago, PhD; Carolyn V. Gould, MD; Alexander J. Kallen, MD; Michelle A. Barron, MD;
Wendy M. Bamberg, MD
Source: Infection Control and Hospital Epidemiology, Vol. 35, No. 4, Special Topic Issue:
Carbapenem-Resistant Enterobacteriaceae and Multidrug-Resistant Organisms (April 2014), pp.
390-397
Published by: The University of Chicago Press on behalf of The Society for Healthcare Epidemiology
of America
Stable URL: http://www.jstor.org/stable/10.1086/675607 .
Accessed: 28/05/2014 10:01

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

The University of Chicago Press and The Society for Healthcare Epidemiology of America are collaborating
with JSTOR to digitize, preserve and extend access to Infection Control and Hospital Epidemiology.

http://www.jstor.org

This content downloaded from 193.48.123.144 on Wed, 28 May 2014 10:01:12 AM

All use subject to JSTOR Terms and Conditions
 infection control and hospital epidemiology april 2014, vol. 35, no. 4

original article

Carbapenem-Resistant Klebsiella pneumoniae Producing New Delhi

Metallo-b-Lactamase at an Acute Care Hospital, Colorado, 2012

Erin E. Epson, MD;1,2 Larissa M. Pisney, MD;3 Joyanna M. Wendt, MD;1,4 Duncan R. MacCannell, PhD;5
Sarah J. Janelle, MPH;2 Brandon Kitchel, MS;4 J. Kamile Rasheed, PhD;4 Brandi M. Limbago, PhD;4
Carolyn V. Gould, MD;4 Alexander J. Kallen, MD;4 Michelle A. Barron, MD;3 Wendy M. Bamberg, MD2

objective. To investigate an outbreak of New Delhi metallo-b-lactamase (NDM)–producing carbapenem-resistant Enterobacteriaceae

(CRE) and determine interventions to interrupt transmission.
design, setting, and patients. Epidemiologic investigation of an outbreak of NDM-producing CRE among patients at a Colorado
acute care hospital.
methods. Case patients had NDM-producing CRE isolated from clinical or rectal surveillance cultures (SCs) collected during the period
January 1, 2012, through October 20, 2012. Case patients were identified through microbiology records and 6 rounds of SCs in hospital
units where they had resided. CRE isolates were tested by real-time polymerase chain reaction for blaNDM. Medical records were reviewed
for epidemiologic links; relatedness of isolates was evaluated by pulsed-field gel electrophoresis (PFGE) and whole genome sequencing
(WGS). Infection control (IC) was assessed through staff interviews and direct observations.
results. Two patients were initially identified with NDM-producing CRE during July–August 2012. A third case patient, admitted in
May, was identified through microbiology records review. SC identified 5 additional case patients. Patients had resided in 11 different units
before identification. All isolates were highly related by PFGE. WGS suggested 3 clusters of CRE. Combining WGS with epidemiology
identified 4 units as likely transmission sites. NDM-producing CRE positivity in certain patients was not explained by direct epidemiologic
overlap, which suggests that undetected colonized patients were involved in transmission.
conclusions. A 4-month outbreak of NDM-producing CRE occurred at a single hospital, highlighting the risk for spread of these
organisms. Combined WGS and epidemiologic data suggested transmission primarily occurred on 4 units. Timely SC, combined with
targeted IC measures, were likely responsible for controlling transmission.
Infect Control Hosp Epidemiol 2014;35(4):390-397

Mechanisms of decreased susceptibility to carbapenems
among Enterobacteriaceae include acquisition of genes en-
coding carbapenemase enzymes. These carbapenemases can
be transmitted between bacteria on mobile genetic elements,
leading to the potential for widespread transmission of de-
creased carbapenem susceptibility among multiple genera of
gram-negative bacteria.1 The New Delhi metallo-b-lactamase
(NDM) is a carbapenemase first described in 2009 in an
isolate from a Swedish patient who had received medical care
in India.2 NDM-producing carbapenem-resistant Enterobac-
teriaceae (CRE) have been reported primarily from South
Asia but have also been isolated from patients worldwide.3,4
NDM-producing CRE have been described rarely in the
United States, where NDM was first detected in an isolate
collected in 2009.5 Before August 2012, only 15 NDM-
producing isolates were identified from 8 US states (Centers
for Disease Control and Prevention [CDC], unpublished
data); all but 3 were from patients who had received inpatient
medical care in regions where these organisms are believed
to be more common.6 We describe a US outbreak of NDM-
producing CRE at a Colorado hospital. A collaborative in-
vestigation conducted by the hospital, Colorado Department
of Public Health and Environment (CDPHE), and the CDC
characterized multiple routes of healthcare-associated trans-
mission during 4 months and led to targeted infection control
measures to halt the outbreak.
methods
Epidemiologic Investigation
This investigation underwent review by the CDC Scientific
Education and Professional Development Program Office hu-

Affiliations: 1. Epidemic Intelligence Service, Centers for Disease Control and Prevention, Atlanta, Georgia; 2. Colorado Department of Public Health
and Environment, Denver, Colorado; 3. Division of Infectious Diseases, University of Colorado, Aurora, Colorado; 4. Division of Healthcare Quality
Promotion, Centers for Disease Control and Prevention, Atlanta, Georgia; 5. Antimicrobial Resistance and Characterization Laboratory, Centers for Disease
Control and Prevention, Atlanta, Georgia.
Received August 16, 2013; accepted November 8, 2013; electronically published March 6, 2014.
䉷 2014 by The Society for Healthcare Epidemiology of America. All rights reserved. 0899-823X/2014/3504-0010$15.00. DOI: 10.1086/675607

This content downloaded from 193.48.123.144 on Wed, 28 May 2014 10:01:12 AM

All use subject to JSTOR Terms and Conditions
 ndm-producing cre at a colorado hospital 391

man subjects protection coordinator and was determined to
be nonresearch, because the primary intent was disease con-
trol. The outbreak occurred at a tertiary-care academic hos-
pital in the Denver metropolitan area. A case patient had a
clinical or surveillance culture (SC) positive for an Entero-
bacteriaceae isolate that was polymerase chain reaction
(PCR)–positive for NDM while hospitalized during the pe-
riod January 1, 2012, to October 20, 2012. Hospital micro-
biology records were queried for Enterobacteriaceae isolates
that were nonsusceptible to any carbapenem cultured from
any patient specimen collected during the period January 1,
2012, to October 20, 2012. Available isolates that were non-
susceptible to all carbapenems and extended-spectrum ceph-
alosporins tested (ie, cefotaxime, ceftriaxone, ceftazidime, and
cefepime) were assessed for Klebsiella pneumoniae carbape-
nemase (KPC) and NDM by PCR.7 To identify asymptomatic
colonization with NDM-producing CRE among patients with
epidemiologic links to case patients, serial rounds of rectal
SCs were conducted among patients on hospital units where
patients with NDM-producing CRE had resided. SC rounds
were performed weekly until no additional case patients were
identified in 4 successive rounds.
Medical records from all confirmed case patients were re-
viewed for clinical and epidemiologic characteristics, includ-
ing dates of admissions, transfers, and hospital discharges;
locations within the hospital; radiographic studies; proce-
dures; use of invasive devices; microbiology results; and an-
timicrobials received. Interviews of case patients or their sur-
rogate were conducted to obtain travel history and exposures
to other healthcare settings during the 6 months before ad-
mission to the hospital; reviews of medical records of patients
who resided in the same hospital unit as the earliest identified
case patients were also conducted to identify NDM risk
factors.
Infection Prevention Assessment
Infection control (IC) protocols were reviewed, including
procedures for tracking patients with cultures positive for
multidrug-resistant organisms (MDROs) and notifying ac-
cepting facilities or units of the patients’ MDRO status upon
transfer. Interfacility transfers of case patients were investi-
gated, and implementation of IC measures at the receiving
facilities was assessed. Unit managers and nursing staff in
affected units; staff in perioperative, procedure, and radiology
areas; and environmental cleaning managers were inter-
viewed. Direct observations were conducted of hand hygiene
and contact precaution (CP) adherence, cleaning and dis-
infection of environmental surfaces in patient rooms and
other patient care areas, and use of shared medical equipment.
Laboratory Analysis
Antimicrobial susceptibility testing (AST) of Enterobacteri-
aceae isolates was performed by the hospital clinical labo-
ratory using Kirby-Bauer disc diffusion (KB testing) for iso-
lates cultured from nonurine sites and MicroScan
(MicroScan) for isolates cultured from urine. Clinical and
Laboratory Standards Institute (CLSI) interpretive criteria for
AST (CLSI document M100-S22)8 were used for isolates eval-
uated by KB testing, and M100-S19 CLSI criteria9 were ap-
plied to urine isolates evaluated by MicroScan.
AST was repeated at the CDC using broth microdilution

table 1. Characteristics of Patients with New Delhi Metallo-b-Lactamase (NDM)–

Producing Carbapenem-Resistant Enterobacteriaceae, Colorado, 2012
Case patients
Patient-specific risk factors (n p 8)
Male sex 4 (50)
Age, median (range), years 58 (23–75)
Charlson index,30 median (range) 2 (0–9)
Hospital days before incident NDM culture or screen, median (range) 18 (12–83)
Any device exposure 8 (100)
Endotracheal intubation 4 (50)
Tracheostomy 1 (13)
Central venous catheter 5 (63)
Urinary catheter 6 (75)
Any surgical drain 6 (75)
Any procedure or radiology exposure 8 (100)
Any operating room exposure 7 (88)
Any computed tomography 7 (88)
Any magnetic resonance imaging 6 (75)
Any echocardiography 6 (75)
Any antimicrobial exposure 8 (100)
Any carbapenem 3 (38)
Any third or fourth generation cephalosporin 5 (63)
Any fluoroquinolone 4 (50)
note. Data are no. (%) of patients, unless otherwise indicated.

This content downloaded from 193.48.123.144 on Wed, 28 May 2014 10:01:12 AM

All use subject to JSTOR Terms and Conditions
392 infection control and hospital epidemiology april 2014, vol. 35, no. 4

figure 1. Epidemiologic tracing and overlap of hospitalized patients with New Delhi metallo-b-lactamase (NDM)–producing carbapenem-
resistant Enterobacteriaceae (CRE), Colorado, May 27–September 26, 2012 (n p 8). The 3 patients identified from clinical cultures are
depicted as circles lettered A, B, and C; the 5 patients detected through surveillance cultures are depicted as squares numbered 1–5. Each
patient’s hospitalization is graphed as a horizontal bar, with different colors representing units where the patients resided during their
hospitalizations. The yellow bars indicate the dates of specimen collection for each patient’s index NDM-producing CRE culture. Overlap
in procedure and radiology rooms is denoted by the arrows between patient tracings. “Other units” included medical intensive care unit,
transplant unit, medicine unit, and cardiac intensive care unit. BICU, burn and trauma intensive care unit; CT, computed tomography;
MRI, magnetic resonance imaging; NSICU, neurosurgical intensive care unit; OR, operating room; SICU, surgical intensive care unit.

(CLSI document M07-A9),10 and isolates were tested at the
CDC for the presence of blaKPC and blaNDM with multiplex
real-time PCR.7
SCs were performed at the hospital clinical laboratory ac-
cording to a CDC-developed protocol.11 M100-S22 CLSI in-
terpretive criteria were applied to AST of these isolates.8
Case isolates and a convenience sample of epidemiologi-
cally linked NDM-negative extended-spectrum b-lactamase
(ESBL)–producing K. pneumoniae from the hospital were
characterized at the CDC by pulsed-field gel electrophoresis
(PFGE) and whole genome sequence (WGS) analysis. PFGE
was performed as described previously.12 Genome libraries
were constructed, and these were sequenced using Illumina
MiSeq (Illumina). Initial clustering of strains was performed
using kSNP 2,13,14 followed by manual reference mapping and
single-nucleotide polymorphism (SNP)–calling using open-
source software.15-18 MEGA 5 was used to review the final set
of parsimoniously informative SNPs and to perform boot-
strapped phylogenetic analysis and visualization.19 A putative
transmission map was constructed by overlaying routes of
transmission suggested by epidemiologic information and
WGS analysis.
results
Epidemiologic Investigation
During July–August 2012, carbapenem-resistant K. pneumo-
niae (CRKP) were isolated from respiratory specimens from
2 patients located in different hospital units. The isolates were
submitted to the CDC for carbapenemase testing; both pos-
sessed blaNDM. Retrospective microbiology records review sub-
sequently identified a third CRKP isolate, cultured from a
respiratory specimen collected from a patient in June 2012;
this isolate also possessed blaNDM. Two additional CRE isolates,
a Citrobacter species and an Enterobacter species, were iden-
tified during the microbiology review. Only the Citrobacter
isolate was available for testing; it was NDM negative, but
KPC positive.
The first round of rectal SCs was a limited survey per-
formed by hospital IC staff during the period August 27 to
August 29 among 14 (42%) of 33 patients on the burn and
trauma intensive care unit (BICU) and surgical intensive care
unit (SICU) who had resided in adjacent rooms or shared
nursing staff with the first 2 identified case patients during
August; all SCs were negative for CRE. On September 26,
SCs were performed for 37 (66%) of all 56 patients residing
on 4 units (BICU, SICU, and cardiology and rehabilitation
units), where the first 2 identified case patients had resided
during their hospitalizations. A total of 5 CRKP isolates were
detected, all possessing blaNDM; these isolates were from pa-
tients who resided in the SICU (1 isolate) and cardiology and
rehabilitation units (2 isolates each). During the period Oc-
tober 3–26, 4 rounds of SCs were performed for patients in
5 units where the initial 3 identified case patients had resided,
including the neurosurgical intensive care unit (NSICU),
BICU, SICU, and cardiology and rehabilitation units, as fol-
lows: October 3, 55 (87%) of 63 patients; October 9, 48 (76%)
of 63 patients; October 15–18, 54 (76%) of 71 patients; and
October 22–26, 49 (69%) of 71 patients. No additional NDM-
positive CRE isolates were identified.
Basic demographic and clinical characteristics of the 8 con-
firmed case patients are summarized in Table 1. The 3 initial
case patients had NDM-producing CRE isolated from clinical
respiratory cultures and were treated for NDM-producing
CRE infection; all survived their hospitalizations. The 5 col-
onized case patients were not treated. No deaths were attrib-
utable to NDM-producing CRE; 1 colonized patient died due
to unrelated causes.
Case patients had resided on 11 different units for a median
of 18 days (range, 12–83 days) before cultures positive for
NDM-producing CRE were obtained. Extensive overlap of
patient locations existed in hospital units as well as operating
rooms (ORs) and computed tomography and magnetic res-
onance imaging suites, which suggested multiple potential
routes of transmission (Figure 1; online only).
Six of the 8 case patients were interviewed, including the
earliest patient with NDM-producing CRE. One colonized
case patient reported hospitalization in the Philippines during
May 2012, before admission to the affected hospital in July.
Because he had no exposures to the affected hospital before
NDM-producing CRE was isolated from the earliest case pa-

This content downloaded from 193.48.123.144 on Wed, 28 May 2014 10:01:12 AM

All use subject to JSTOR Terms and Conditions

figure 2. Pulsed-field gel electrophoresis (PFGE) patterns of 8 New Delhi metallo-b-lactamase (NDM)–producing carbapenem-resistant
Enterobacteriaceae isolates (A–C, 1–5) and 4 epidemiologically linked extended-spectrum b-lactamase (ESBL)–producing Klebsiella pneumo-
niae isolates, Colorado, May 27–September 26, 2012.
tient in June, this patient is unlikely to be the index patient.
None of the case patients and none of the 197 patients ep-
idemiologically linked to the earliest case patients had a re-
ported history of travel or receipt of healthcare outside of
the United States.
Infection Prevention Assessment
Hospital IC staff received notification from microbiology lab-
oratory personnel of all positive cultures, flagged charts of
patients with cultures positive for MDROs, and reviewed hos-
pital census lists daily to ensure that flagged patients were
maintained on CP, including upon intrafacility transfers; pa-
tient case managers were responsible for communicating pa-
tient MDRO status upon interfacility transfers. Upon rec-
ognition of the outbreak in August, patients with
NDM-producing CRE were maintained on CP with a 1 : 1
nursing ratio. When additional case patients were identified
through SCs in September, all case patients who remained in
the hospital were cohorted on the same hospital unit and
assigned dedicated nursing staff and medical equipment, and
a visitor policy restricting young children and limiting visitors
to 2 at a time was implemented. Additionally, hospital IC
staff performed targeted education about CRE and the im-
portance of hand hygiene and CP adherence among staff on
units where case patients were residing.
Observations identified lapses in hand hygiene and ad-
herence to CP among physicians, nurses, and visitors in per-
ioperative areas. Of 102 observed opportunities for hand hy-
giene recorded in the BICU and NSICU, 28 (27%) were
missed by physicians, nurses, nursing assistants, and visitors;
routine hospital IC surveillance indicated an average 94%
and 90% compliance with hand hygiene in those units, re-
spectively, and similar compliance in hospital units involved
This content downloaded from 193.48.123.144 on Wed, 28 May 2014 10:01:12 AM
All use subject to JSTOR Terms and Conditions
ndm-producing cre at a colorado hospital 393
and uninvolved in the outbreak. Lack of documentation of
cleaning and disinfection of portable medical equipment be-
tween patients and infrequent cleaning of certain patient care
areas (eg, radiology rooms and a rehabilitation gym) were
also noted.
In 2 instances, patients with NDM-producing CRE were
transferred to other facilities during the outbreak period. Al-
though communication between hospital case managers and
the 2 receiving facilities reportedly occurred, the receiving
facilities did not maintain CP for portions of the patients’
stay. Two rounds of SCs were therefore recommended for
patients at each of these facilities, which did not identify any
CRE.
Laboratory Analysis
PFGE results are illustrated in Figure 2. All case patient iso-
lates were highly related, as were the convenience sample of
4 ESBL-producing, NDM-negative K. pneumoniae isolates
from the hospital. WGS analysis identified 67 parsimoniously
informative SNP differences among these 12 isolates. A den-
drogram constructed from these SNPs (Figure 3) indicated
3 clusters among these isolates that were indistinguishable by
PFGE. Applying epidemiologic information, these clusters
were associated with particular hospital units. ESBL-produc-
ing K. pneumoniae isolates were interspersed among clusters
2 and 3.
Putative Transmission Map
Transmission maps are included in Figure 4. In the map in
Figure 4A, direct epidemiologic overlap (ie, in which both
patients occupied the same unit or patient care area during
the same period before positive cultures were obtained) is
394 infection control and hospital epidemiology april 2014, vol. 35, no. 4
figure 3. Whole genome sequencing (WGS) dendrogram analysis
of 8 New Delhi metallo-b-lactamase (NDM)–producing Klebsiella
pneumoniae isolates (A–C, 1–5) and 4 epidemiologically linked
extended-spectrum b-lactamase (ESBL)–producing K. pneumoniae
isolates (ESBL1–ESBL4), Colorado, May 27–September 26, 2012.
Preliminary WGS analysis demonstrates 3 clusters of isolates that
correspond to units in the hospital. The patients in cluster 1 (patients
1, 2, 4, and 5) were associated with either the cardiology or reha-
bilitation units. The patients in cluster 2 (B and 3) had both resided
in the surgical intensive care unit (SICU). Patients in cluster 3 (A
and C) had both resided in the burn and trauma intensive care unit
(BICU).
indicated by solid gray arrows, and indirect overlap (ie, in
which patients were located on the same unit or patient care
area but separated by time) is indicated by the dashed gray
arrows. The directionality of the arrows indicates whether a
clear sequence of one patient’s NDM-producing CRE posi-
tivity versus another existed, as follows: unidirectional arrows
indicate when one patient resided on a unit and had an NDM
isolate before overlapping with the other patient; a bidirec-
tional arrow indicates when we cannot infer who transmitted
to whom, such as when the patients had overlapped before
both having cultures positive for NDM collected on the same
day. Patients with isolates in the same WGS cluster are linked
by bidirectional black arrows, because the directionality of
transmission cannot be inferred on the basis of genome se-
quence clustering. The map in Figure 4B includes only links
that were substantiated by both epidemiologic and WGS in-
formation, and priority was given to direct over indirect ep-
idemiologic links. This map illustrates strong evidence for
transmission occurring in 4 hospital units: BICU, SICU, car-
diology unit, and rehabilitation unit. Certain transmission
This content downloaded from 193.48.123.144 on Wed, 28 May 2014 10:01:12 AM
All use subject to JSTOR Terms and Conditions
routes suggested by epidemiologic information were not sup-
ported by WGS (eg, between patients A and B, who had
epidemiologic overlap in an OR). All transmission of NDM-
producing CRE was not explained by direct epidemiologic
overlap, suggesting that undetected colonized patients were
involved in some transmission routes.
discussion
We describe what was the largest outbreak of NDM-produc-
ing CRE in the United States to date, in a state where NDM
had not been reported previously. No travel or healthcare
exposures outside of the United States were identified among
case patients or among 197 patients epidemiologically linked
to the earliest case patients that explained how NDM was
introduced into the hospital. The earliest identified case pa-
tients thus likely acquired NDM-producing CRE through
healthcare-associated transmission arising from an unde-
tected index patient, and our epidemiologic and laboratory
analysis was consistent with subsequent transmission within
the hospital.
Although an index patient was not identified, it is likely
that, before this outbreak, a person colonized with an NDM-
producing organism through travel or healthcare exposure
outside of the United States entered the hospital. Before this
outbreak, 12 of 15 known patients with NDM-producing
organisms had a recent overnight stay in a healthcare facility
outside the United States (CDC, unpublished data). In re-
sponse to increasing reports of NDM-producing isolates, the
CDC recently disseminated a health advisory recommending
that CRE isolated from patients with such a history within
the previous 6 months be tested for carbapenemases (at a
minimum, KPC and NDM). In addition, screening these pa-
tients for CRE at admission should be considered.20
Our investigation revealed likely healthcare transmission
of NDM-producing CRE at a US hospital and therefore high-
lights the risk for transmission of NDM-producing organisms
among persons receiving medical care inside the United
States. Similar healthcare-associated NDM-producing CRE
outbreaks have been reported during recent years in other
countries.21-23 Preventing the spread of carbapenemases is a
priority and requires an aggressive approach when these or-
ganisms are encountered. Recommended interventions for
controlling CRE can be found in the CDC CRE toolkit.24
No single source of transmission was identified in this
outbreak. Affected patients had prolonged, complex hospi-
talizations with extensive exposures to medical devices and
procedures. Epidemiologic tracing identified extensive over-
lap among patients in multiple areas of the hospital through-
out a 4-month period. A combination of epidemiologic and
laboratory analyses, including WGS, suggested multiple trans-
mission events occurring in at least 4 hospital units during
this time. WGS analysis is an emerging technology in inves-
tigation of healthcare-associated outbreaks.25-27 and offers the
possibility of more finely discriminating among outbreak iso-
 ndm-producing cre at a colorado hospital 395

lates. Although laboratory costs might vary, currently WGS

costs $125–$200 per isolate26 compared with approximately
$20 for PFGE; however, the cost of WGS is decreasing. WGS
can be used to identify relationships and determine additional
information that PFGE cannot, such as high-resolution strain
type and virulence and resistance markers, which might oth-
erwise require additional costly testing. In this outbreak, com-
bining WGS with epidemiologic data enabled us to narrow
the number of areas where transmission was likely to have
occurred and identify areas where transmission was less likely
(eg, ORs). If applied in real time, WGS might have enabled
earlier clarification of likely transmission routes and allowed
for more timely and targeted control measures during the
initial phase of this outbreak.
Some NDM-negative K. pneumoniae isolates from the hos-
pital closely matched NDM-positive isolates by both PFGE
and WGS. We did not systematically collect and analyze
NDM-negative isolates from the hospital and therefore can-
not definitively describe the extent or reasons for this finding.
However, we believe there are at least 2 possible explanations:
(1) the NDM-negative isolates might have produced NDM
and represented additional outbreak cases, but the NDM ge-
netic element was lost either in the patient or in the laboratory
after collection, or (2) the NDM-negative isolates are a suc-
cessful clone representing the background K. pneumoniae in
the hospital, into which NDM was introduced and subse-
quently spread.
Certain transmission events that were unexplained by di-
rect epidemiologic overlap between patients suggested more
complex routes of transmission involving undetected colo-
nized intermediary patients. For instance, we identified only
2 patients who might have served as intermediaries in a trans-
mission event between the 2 earliest case patients in the BICU
(patients A and C). If SCs had been performed when the
earliest patient with CRE was identified, colonized interme-
diary patients might have been identified and isolated sooner, figure 4. Putative transmission maps based on epidemiologic and
possibly limiting the early propagation of the outbreak. SCs whole genome sequencing (WGS) analysis among patients with New
ultimately identified the majority of case patients (5 of 8) in Delhi metallo-b-lactamase (NDM)–producing Klebsiella pneumoniae
this outbreak and allowed targeted IC measures to be insti- isolates, Colorado, May 27–September 26, 2012 (n p 8). A, all
tuted. SCs were also essential for evaluating for transmission epidemiologic and WGS links. The 3 patients identified from clinical
at other healthcare facilities that received case patients. How- cultures are depicted as circles A, B, and C; the 5 patients detected
ever, in this outbreak, none of the 3 laboratories serving the through surveillance cultures (SCs) are depicted as squares 1–5. Gray
involved facilities was prepared to do SCs for CRE when the arrows are epidemiologic links between patients. Direct epidemio-
outbreak began, delaying the response. SCs have a crucial role logic overlap is indicated by solid arrows, and indirect overlap is
for rapidly identifying CRE transmission in regions where indicated by dashed arrows. Unidirectional arrows indicate one pa-
tient resided on a unit or patient care area and had an NDM isolate
CRE are uncommon. As more regions identify CRE, clinical
before overlap with second patient; bidirectional arrows indicate that
laboratories should validate and be ready to perform cultures direction of transmission cannot be inferred. Black arrows are WGS
of rectal swab samples to screen for CRE. In addition, simpler links. Patients with isolates closest to each other in the same WGS
and more efficient SC methods need to be developed and cluster are linked by black arrows between isolates; because direc-
validated. tionality of transmission cannot be inferred on the basis of the
No NDM-producing CRE were identified at other health- genome sequence clustering, all genome sequence links are bidirec-
care facilities that received case patients; however, failure to tional. B, transmission links supported by both epidemiologic and
continue appropriate IC measures after transfer put addi- WGS information. Routes not substantiated by both methods were
tional patients at risk and might have expanded the outbreak. removed. Remaining routes indicate units with strong evidence for
Although reasons that CP were not continued in this instance transmission, labeled by boxes overlying arrows.

This content downloaded from 193.48.123.144 on Wed, 28 May 2014 10:01:12 AM

All use subject to JSTOR Terms and Conditions
396 infection control and hospital epidemiology april 2014, vol. 35, no. 4
are unclear, these might include failure to communicate the
MDRO status to the accepting facility at transfer or failure
of the accepting facility to understand the necessary measures.
As in this outbreak, public health can have an important role
in facilitating communication among involved facilities and
serve as a resource to assist facilities in understanding the
appropriate IC measures. In addition, public health can pro-
vide regional situational awareness about outbreaks of
MDROs; in this instance, CDPHE disseminated a health alert
to healthcare facilities in the Denver metropolitan area with
recommendations for CRE screening of patients with expo-
sures to the affected facility.
A plan to address IC breaches at the hospital was jointly
developed among investigation collaborators and focused on
strengthening communication for interfacility transfers of pa-
tients with MDROs, monitoring and adherence to hand hy-
giene and CP, and cleaning and disinfection of portable med-
ical equipment and patient care areas. In the rehabilitation
gym, for example, therapy for NDM-positive patients was
rescheduled separate from non-NDM-positive patients at the
end of the day, and only equipment constructed of a material
that could be cleaned and disinfected was used. The gym is
now cleaned and disinfected on a nightly basis.
Limitations of this investigation included the delay between
identifying the initial patients and performing SCs, during
which patients that overlapped with case patients might have
been discharged from the hospital and were no longer avail-
able for SCs; in addition, the imperfect sensitivity of SCs in
detecting CRE colonization28 and the lack of feasibility for
performing SCs on all units in the hospital all limited our
ability to detect additional case patients and determine trans-
mission routes. Overlap of hospital staff caring for case pa-
tients was not assessed, and cultures of hospital staff were
not performed; however, to date, chronically colonized
healthcare providers have not been implicated as a source of
CRE transmission. Generalizability of our findings might be
limited, because detecting this outbreak relied on CRE mech-
anism testing that is not routinely performed in the United
States.
In this large outbreak of NDM-producing CRE, basic but
crucial measures—in particular, the use of SCs for identifying
colonized patients—were responsible for interrupting trans-
mission. A recent review similarly cites SCs and cohorting of
patients and staff on the basis of these results as the key
interventions.29 To prevent future outbreaks, particularly in
areas where CRE remain uncommon, clinicians, infection
preventionists, and public health practitioners should be
aware of the prevalence of CRE in their region, have systems
in place for identifying CRE, and be prepared to implement
appropriate control measures promptly. Although all carba-
penemase-producing CRE are epidemiologically important,
preventing the emergence of novel carbapenemases (eg,
NDM-producing organisms) will require recognition and
screening of patients at risk for these organisms and aggressive
action when those organisms are identified.
This content downloaded from 193.48.123.144 on Wed, 28 May 2014 10:01:12 AM
All use subject to JSTOR Terms and Conditions
acknowledgments
Financial support. W.M.B. and S.J.J. report grant support through the Cen-
ters for Disease Control and Prevention Emerging Infections Program grant.
Potential conflicts of interest. All authors report no conflicts of interest
relevant to this article. All authors submitted the ICMJE Form for Disclosure
of Potential Conflicts of Interest, and the conflicts that the editors consider
relevant to this article are disclosed here.
Address correspondence to Erin Epson, MD, Epidemic Intelligence Service
Officer, Colorado Department of Public Health and Environment, 4300
Cherry Creek Drive South, Denver CO 80246 (erin.epson@state.co.us).
references
1. Johnson AP, Woodford N. Global spread of antibiotic resistance:
the example of New Delhi metallo-beta-lactamase (NDM)-
mediated carbapenem resistance. J Med Microbiol 2013;62:499–
513.
2. Yong D, Toleman MA, Giske CG, et al. 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. Anti-
microb Agents Chemother 2009;53:5046–5054.
3. Nordmann P, Naas T, Poirel L. Global spread of carbapenemase-
producing Enterobacteriaceae. Emerg Infect Dis 2011;17:1791–
1798.
4. Wilson ME, Chen LH. NDM-1 and the role of travel in its
dissemination. Curr Infect Dis Rep 2012;14:213–226.
5. Centers for Disease Control and Prevention. Detection of En-
terobacteriaceae isolates carrying metallo-beta-lactamase:
United States, 2010. MMWR Morb Mortal Wkly Rep 2010;59:
750.
6. Nordmann P, Poirel L, Walsh TR, Livermore DM. The emerging
NDM carbapenemases. Trends Microbiol 2011;19:588–595.
7. Centers for Disease Control and Prevention (CDC). Multiplex
Real-Time PCR Detection of Klebsiella pneumoniae Carbape-
nemase (KPC) and New Delhi metallo-beta-lactamase (NDM-1).
Atlanta, Georgia: CDC, 2011. http://www.cdc.gov/HAI/pdfs
/labSettings/KPC-NDM-protocol-2011.pdf. Accessed August 6,
2013.
8. Clinical and Laboratory Standards Institute (CLSI). Performance
Standards for Antimicrobial Susceptibility Testing; Twenty-Second
Informational Supplement. Wayne, PA: CLSI, 2012. CLSI doc-
ument M100-S22.
9. Clinical and Laboratory Standards Institute (CLSI). Performance
Standards for Antimicrobial Susceptibility Testing; Nineteenth In-
formational Supplement. Wayne, PA: CLSI, 2009. CLSI document
M100-S19.
10. Clinical and Laboratory Standards Institute (CLSI). Methods for
Dilution Antimicrobial Susceptibility Tests for Bacteria That Grow
Aerobically: Approved Standard. 9th ed. Wayne, PA: CLSI, 2012.
CLSI document M07–M09.
11. Centers for Disease Control and Prevention (CDC). Laboratory
Protocol for Detection of Carbapenem-Resistant or Carbapene-
mase-Producing Klebsiella spp. and E. coli from Rectal Swabs.
Atlanta, Georgia: CDC, 2011. http://www.cdc.gov/hai/pdfs
/labsettings/klebsiella_or_ecoli.pdf. Accessed August 6, 2013.
12. Kitchel B, Rasheed JK, Patel JB, et al. Molecular epidemiology
of KPC-producing Klebsiella pneumoniae isolates in the United

States: clonal expansion of multilocus sequence type 258. An-
timicrob Agents Chemother 2009;53:3365–3370.
13. kSNP 2.0 ed. Livermore, California: Lawrence Livermore Na-
tional Lab, 2013.
14. Gardner SN, Hall BG. When whole-genome alignments just
won’t work: kSNP v2 software for alignment-free SNP discovery
and phylogenetics of hundreds of microbial genomes. PLOS
ONE 2013;8(12):e81760.
15. Li H, Durbin R. Fast and accurate short read alignment with
Burrows-Wheeler transform. Bioinformatics 2009;25:1754–1760.
16. Li H, Handsaker B, Wysoker A, et al. The sequence alignment/
map (SAM) format and SAMtools. Bioinformatics 2009;25:2078–
2079.
17. Koboldt D, Zhang Q, Larson D, et al. VarScan 2: somatic mu-
tation and copy number alteration discovery in cancer by exome
sequencing. Genome Res 2012;22(3):568–576.
18. Garrison E, Marth G. Haplotype-based variant detection from
short-read sequencing. eprint arXiv:12073907. 2012.
19. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar
S. MEGA5: Molecular evolutionary genetics analysis using max-
imum likelihood, evolutionary distance, and maximum parsi-
mony methods. Mol Biol Evol 2011;28:2731–2739.
20. Centers for Disease Control and Prevention (CDC). New Car-
bapenem-Resistant Enterobacteriaceae Warrant Additional Action
by Healthcare Providers. Atlanta, Georgia: CDC, 2013. http://
www.bt.cdc.gov/HAN/han00341.asp. Accessed August 6, 2013.
21. Borgia S, Lastovetska O, Richardson D, et al. Outbreak of car-
bapenem-resistant Enterobacteriaceae containing blaNDM-1,
Ontario, Canada. Clin Infect Dis 2012;55:e109–e117.
22. Lowe CF, Kus JV, Salt N, et al. Nosocomial transmission of New
Delhi metallo-beta-lactamase-1-producing Klebsiella pneumo-
This content downloaded from 193.48.123.144 on Wed, 28 May 2014 10:01:12 AM
All use subject to JSTOR Terms and Conditions
ndm-producing cre at a colorado hospital 397
niae in Toronto, Canada. Infect Control Hosp Epidemiol 2013;
34:49–55.
23. Escobar Perez JA, Olarte Escobar NM, Castro-Cardozo B, et al.
Outbreak of NDM-1-producing Klebsiella pneumoniae in a neo-
natal unit in Colombia. Antimicrob Agents Chemother 2013;57:
1957–1960.
24. Centers for Disease Control and Prevention (CDC). 2012 CRE
Toolkit: Guidance for Control of Carbapenem-resistant Entero-
bacteriaceae (CRE). Atlanta, Georgia: CDC, 2012. http://www
.cdc.gov/hai/organisms/cre/cre-toolkit/f-level-prevention.html.
Accessed August 6, 2013.
25. Snitkin ES, Zelazny AM, Thomas PJ, et al. Tracking a hospital
outbreak of carbapenem-resistant Klebsiella pneumoniae with
whole-genome sequencing. Sci Transl Med 2012;4:148ra16.
26. Koser CU, Holden MT, Ellington MJ, et al. Rapid whole-genome
sequencing for investigation of a neonatal MRSA outbreak. New
Engl J Med 2012;366:2267–2275.
27. Lewis T, Loman NJ, Bingle L, et al. High-throughput whole-
genome sequencing to dissect the epidemiology of Acinetobacter
baumannii isolates from a hospital outbreak. J Hosp Infect 2010;
75:37–41.
28. Thurlow CJ, Prabaker K, Lin MY, et al. Anatomic sites of patient
colonization and environmental contamination with Klebsiella
pneumoniae carbapenemase-producing Enterobacteriaceae at
long-term acute care hospitals. Infect Control Hosp Epidemiol
2013;34:56–61.
29. Munoz-Price LS, Quinn JP. Deconstructing the infection control
bundles for the containment of carbapenem-resistant Entero-
bacteriaceae. Curr Opin Infect Dis 2013;26:378–387.
30. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method
of classifying prognostic comorbidity in longitudinal studies:
development and validation. J Chronic Dis 1987;40:373–383.