Maternal-Neonatal Reports
Overestimation of an Outbreak of Enterobacter cloacae in a
Neonatal Intensive Care Unit in Germany, 2015
Gyde Steffen, MPH,*†‡§ Michael Pietsch, MSc,¶ Martin Kaase, MD,║** Sören Gatermann, MD,║
Guido Werner, PhD,¶ Stephan Fuchs, PhD,¶ Yvonne Pfeifer, PhD,¶ Wolfgang Schmitt, MD,††
Henning Adam,†† Tim Eckmanns, MD, MSc,‡ and Sebastian Haller, MD, MPH, MSc‡§
Background: In August 2015, 17 neonates with Enterobacter cloacae
(E. cloacae) colonization were identified in a neonatal intensive care unit
(NICU) in Germany. Two developed severe brain abscesses. Despite tempo-
rary NICU closure in September, another infant with E. cloacae coloniza-
tion was detected in October 2015.
Methods: We defined potential cases as inpatients treated in the NICU or
any pediatric/maternity ward in 2015 with E. cloacae in any specimen before
molecular typing. Cases were at first confirmed by arbitrarily-primed-pol-
Downloaded from http://journals.lww.com/pidj by BhDMf5ePHKbH4TTImqenVCSTviUIIukZ1ngjgjM0KMJcOcdJP7RLpKylKlZDnCOH on 12/16/2019
ymerase-chain-reaction and later by XbaI-macrorestriction/pulsed-field gel
electrophoresis and next-generation-sequencing. Enhanced barrier precau-
tions and cohorting were implemented for all potential cases and microbio-
logic screening was extended from NICU to all pediatric/maternity wards.
Results: Of 41 potential cases (occurring between 08/04/2015 and
15/11/2015 in 4 wards), the isolates of 23 shared identical arbitrarily-
primed-polymerase-chain-reaction patterns; 3 without plausible epidemio-
logic link. Pulsed-field gel electrophoresis analyses verified only 10 cases
(all in the NICU); next-generation-sequencing analysis confirmed these
results. In addition 6 cases without isolates available for genotyping were
closely linked in place and time.
Conclusions: Forty-one suspected patients were cohorted and the NICU
was temporarily closed. Further analyses revealed that only 16 cases
belonged to the outbreak. Only close interdisciplinary collaboration and
highly discriminatory genotyping methods allowed to clearly differentiate
between cases and noncases in this E. cloacae outbreak.
Key Words: neonates, infection prevention, hygiene, antibiotic resistance, NGS
(Pediatr Infect Dis J 2019;38:631–637)
E
nterobacter cloacae (E. cloacae) is a commensal of the human
gastrointestinal tract and an important facultative pathogen
responsible for healthcare-associated infections.1,2 Early gut and
Accepted for publication November 2, 2018.
From the *Postgraduate Training for Applied Epidemiology (PAE), Robert Koch
Institute, Berlin, Germany; †European Programme for Intervention Epide-
miology Training (EPIET), ECDC, Sweden; ‡Department for Infectious
Disease Epidemiology, Robert Koch Institute, Berlin, Germany; §Institute
of Public Health, Charité-Universitätsmedizin Berlin, Berlin, Germany;
¶Department of Infectious Diseases, Robert Koch-Institute, Wernigerode
Branch, Germany; ║Department of Medical Microbiology, Ruhr University
Bochum, Germany; **Central Department of Hospital Hygiene and Infecti-
ology, University Medicine, Göttingen, Germany; and ††Local Public Health
Authority of Saarlouis District, Saarlouis, Germany.
This work was partly funded by a grant of the Federal Ministry of Education and
Health, Germany: No. 01KI1313F to M.P., Y.P. and G.W. and the Federal
Ministry of Health for the Reference Laboratory Network Project MRSAar-
PLUS to G.W., T.E. and S. H.
The authors have no conflicts of interest to disclose.
Address for correspondence: Gyde Steffen, MPH, Department for Infectious Disease
Epidemiology, Robert Koch-Institute, Berlin, Germany. E-mail: steffeng@rki.de.
Supplemental digital content is available for this article. Direct URL citations
appear in the printed text and are provided in the HTML and PDF versions of
this article on the journal’s website (www.pidj.com).
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
ISSN: 0891-3668/19/3806-0631
DOI: 10.1097/INF.0000000000002264
The Pediatric Infectious Disease Journal  •  Volume 38, Number 6, June 2019 www.pidj.com | 631
Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
nasopharyngeal colonization of neonates with E. cloacae is com-
mon in neonatal intensive care units (NICUs)3,4 and infants may
develop severe infections.5,6 Healthcare-associated outbreaks with
person-to-person transmission of E. cloacae have been described.7,8
In the majority of these outbreaks, patients colonized with
the outbreak strain are the source of the outbreak and transiently
contaminated hands of care takers act as vectors.9,10 Sensitivity
of microbiologic screening must be high to assure the identifica-
tion of all colonized individuals to allow isolation or cohorting of
cases. Low specificity may result in isolation of too many patients.
Unnecessary separation and enhanced barrier precautions may
lead to disturbance of important physical closeness, and has been
associated with adverse outcomes.11 Moreover misclassification of
cases in both directions may delay transmission interruption and
therefore influence clinical treatment and outcome of the affected
individuals.
In August 2015, a cluster of neonates with E. cloacae colo-
nization was identified in a NICU of a tertiary care hospital in Ger-
many. The outbreak led to media attention and to debates in the
regional parliament.12 Between September and October 2015, the
hospital temporarily closed the NICU, but after reopening another
infant was tested positive for E. cloacae. The primary objective of
our investigation was to identify all outbreak-related cases to under-
stand the mode of transmission and to implement targeted infection
control measures. Here we present the outbreak investigation to high-
light the challenges in outbreaks of commensals in hospitals.
METHODS
The outbreak investigation consisted of retrospective and
prospective elements, and a descriptive analysis of the collected
data was conducted.
Setting
The outbreak took place in the perinatal center of a 400-
bed tertiary care hospital in Germany. The perinatal center contains
parts of the pediatric department including a 13-cot NICU and a
maternity ward. Nine hundred eighty-two babies were born in the
perinatal center in 2014 and 380 received treatment in the NICU.
Case Definition
Inpatients treated in the NICU, maternity ward or any
other ward of the pediatric department between 01/01/2015 and
31/12/2015 were defined as the following:
•• confirmed cases, if E. cloacae were detected and molecular
typing revealed relatedness to other isolates from this outbreak
•• probable cases, if they had any record of E. cloacae in any
specimen and an epidemiologic link to a confirmed case in
admission ward and time of treatment, but no isolate was
available for molecular typing
•• potential cases, if they had any record of E. cloacae in any
specimen (independent from epidemiological link) before
molecular typing of the isolates could be performed
Steffen et al The Pediatric Infectious Disease Journal  •  Volume 38, Number 6, June 2019
Case Search
For prospective case search, we introduced a systematic
microbiological screening (rectal and nasal swabs) twice a week
including admission screening for all infants of the NICU from
18/10/2015 until the end of the outbreak was declared. Screen-
ing of mothers was implemented to rule out an outbreak within
the maternity ward. After detection of potential cases in pediatric
wards besides the NICU, we extended screening to all pediatric
wards (from 05/11/2015).
For retrospective case search, we collected information on
all E. cloacae positive patients in the pediatric department and in
the maternity ward in 2015. Additionally, the database “National
Surveillance of Healthcare-Associated Outbreaks” was screened13
to exclude supraregional E. cloacae outbreaks.
Environmental Investigations
More than 200 environmental samples from the NICU and
maternity ward were taken. Screening of stool samples, perianal
and/or nasal swabs from 121 healthcare workers (HCWs) was con-
ducted.
Outbreak Control Measures
Basic hygiene measures in the affected wards were
reviewed by the hospital hygienist and local public health services.
Hygiene education for HCWs and patients’ relatives was intensi-
fied. Enhanced barrier precautions were put in place and all cases
(potential, probable and confirmed) were cohorted in designated
areas of the affected wards.
Microbiological Investigations
Species Identification and Antimicrobial Susceptibility
Testing
VITEK 2 GN card (bioMérieux, Marcy-l’Étoile, France)
was used for species identification. Antimicrobial susceptibilities
were determined by microbroth dilution according to EUCAST
v6.0 (http://www.eucast.org/clinical_breakpoints). Susceptibilities
to tigecycline, colistin and imipenem and ertapenem were deter-
mined by VITEK2 (card AST-N248, bioMérieux, Nürtingen, Ger-
many) or Etest.
Molecular Typing
First, the hospital laboratories performed XbaI-restriction
of total DNA and subsequent pulsed-field gel electrophoresis
(PFGE). When no banding patterns could be found, the method
on site was switched to Arbitrary-Primed-PCR (AP-PCR)14 with
DNA isolated using MAXWELL16Dx (Promega). Amplification
products were compared by gel electrophoresis. In parallel, all
isolates were sent to the Robert Koch Institute (RKI) for next
generation sequencing (NGS)-based analyses. PFGE analysis was
repeated at RKI, with a modified method including thiourea (39
µM) added to the running buffer to avoid DNA degradation. Mac-
rorestriction patterns were interpreted according to the criteria of
Tenover et al.15
NGS was performed on a MiSeq instrument (Illumina,
USA) using the Nextera XT library (Illumina, USA) and the
MiSeq v3 reagent kit generating 2 × 300 base pair (bp) paired-
end reads. Sequence data were submitted to the European Nucle-
otide Archive (http://www.ebi.ac.uk/ena; PRJEB22570). Raw
read data were trimmed using Trimmomatic (v. 0.0.9; default
parameters except maxinfo 15:0.5) and assembled by A5-miseq
(v. 0.0.9 beta; default parameters).16,17 Contigs were submitted
to the Centre for Genomic Epidemiology and known antibiotic
resistance markers were extracted using ResFinder 2.1 and Plas-
midFinder 1.3, respectively (https://cge.cbs.dtu.dk/services/).18
632 | www.pidj.com
Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Multilocus sequence types (MLST) of the E. cloacae isolates
were identified using MLST 1.8.19,20 Three new sequence types
(STs) were submitted to https://pubmlst.org/ecloacae/. For phy-
logenetic analysis, paired-end reads were mapped to Enterobac-
ter cloacae complex ECNIH2 (NZ_CP008823.1) using BWA-
SW (v. 0.7.13-r1126; default parameters), achieving a mean
coverage of 61.8.21,22
Single nucleotide polymorphisms (SNPs) were identified
from the mapped reference genome using SNPfilter23 (https://git-
lab.com/s.fuchs/snpfilter). A maximum likelihood tree was calcu-
lated based on 3161 SNP positions (RAxML 7.2.8, GTR GAMMA
model, Rapid hill-climbing, 100 starting trees).24
To investigate the genetic basis of varying susceptibilities to
third-generation cephalosporins in isolates of the outbreak group,
de novo assembled contigs were analyzed for known ampC phe-
notype inducing sites by aligning to known wild-type sequences
of ampC (X07274.1), ampD (Z14003) and ampR (X04730 and
M27222.1), as well as ompF (KT780421.1), ompC (KT780422.1)
and corresponding promotor regions using Geneious (v. 10.0.5,
Biomatters Ltd, New Zealand).
Epidemiologic Analysis
For epidemiological analysis, numerous variables were col-
lected for each case from their clinical records and analyzed in
Stata 14.1 (College Station, TX).
Ethics
A formal ethical review was not required for the outbreak
investigation in accordance with Article 25, Section 1 of the Ger-
man Infection Protection Act of 2001.
RESULTS
Course of the Outbreak
We detected 41 potential cases that occurred between
08/04/2015 and 15/11/2015 including 21 infants in the NICU, 16
children of all ages on 2 other pediatric wards and 4 women on
the maternity ward (Fig. 1). All potential cases with antimicrobial
resistant E. cloacae were isolated immediately after detection.
From 08/2015 onwards, enhanced barrier precautions (see above)
were put in place.
When PFGE was not functioning at the hospital’s labora-
tories, AP-PCR was performed. Isolates from 23 of the potential
cases shared identical AP-PCR banding patterns. For 3 of these
cases, no epidemiological link could be identified. Thus, PFGE was
repeated at RKI with addition of thiourea and only isolates of 10
potential cases showed identical macrorestriction patterns. NGS-
analysis confirmed a close relationship between isolates of these
10 cases. Furthermore, 6 probable cases were attributed to the out-
break. Finally, the outbreak comprised 16 cases (occurring between
24/06/2015 and 22/10/2015) on 2 different wards (15 in the NICU
and 1 on a pediatric ward).
Descriptive Epidemiology of Confirmed Cases
Three infants (19%) developed severe infection with the out-
break strain [2 (12%) developed brain abscesses after known colo-
nization, 1 (0.1%) developed sepsis without previously detected E.
cloacae colonization].
Sixty-two percent (10/16 cases) were female (Table 1). The
median age at the time of E. cloacae detection was 10 days [range
1–86]. Preterm neonates (gestational age <37 weeks) represented
87% of the cases (14/16). The median birth weight was 1805 g
[range 1210–4040 g], 19% (3/16) were infants with very low birth
weight (<1500 g).
© 2019 Wolters Kluwer Health, Inc. All rights reserved.
The Pediatric Infectious Disease Journal  •  Volume 38, Number 6, June 2019 NICU Outbreak E. cloacae, Germany

FIGURE 1.  Epicurve: cases per

calendar week in dependence
of diagnostic method used,
01/01/2015–31/12/2015: (1) all 41
cases with E. cloacae ssp. dissolvens
(potential cases); (2) 23 E. cloacae
cases with AP-PCR confirmed
outbreak strain, 5 probable cases;
(3) 10 E. cloacae cases with PFGE/
NGS confirmed outbreak strain,
6 probable cases. Striped boxes
represent cases without isolates
available for genotyping (probable
cases).

Microbiological Investigations tested isolates from staff samples were not genetically related to
Between 08/04/2015 and 15/11/2015, E. cloacae ssp. dissol- this outbreak strain.
vens was detected in 92 samples from 41 patients and 18 samples NGS analysis was additionally performed for 26 isolates
from medical staff (17 individuals). E. cloacae were detected in 4 that were PFGE-typed. This analysis confirmed the 14 isolates
environmental samples. (from 10 patients, 1 environmental sample) as the outbreak strain.
Inferred MLST grouped these 14 isolates into sequence type (ST)
Molecular Analyses 815 forming the outbreak clade (Fig. 3). All nonoutbreak isolates
Thirty-five isolates from 28 patients, 18 from staff (17 indi- were assigned to 11 different STs including 2 isolates belonging to
viduals) and the environmental isolates were available when typ- the same ST (ST145) and 2 new STs (ST816 and ST907).
ing was started and analyzed with AP-PCR. Twenty-three isolates SNP analysis revealed high homology between the 14 out-
from patients, 4 from medical staff and 3 environmental isolates break isolates; only 15 SNPs were detected. Single isolates diver-
shared identical banding patterns when AP-PCR was applied. sified in direct comparisons in only zero up to 8 SNP positions.
For subsequent PFGE analyses at the RKI, only 26 isolates (all Comparisons to nonoutbreak isolates showed a more diverse SNP
with prior positive AP-PCR results) were available: 22 from 19 pattern (Min 1043 SNPs − Max 1529 SNPs difference). Figure 3
patients, 2 from medical staff on the NICU, and 2 from siphons shows the phylogenetic tree and enables the identification of the
on the NICU and maternity ward. Macro restriction patterns were outbreak clade, concordant with PFGE results (Fig. 2).
visible for all these isolates, when thiourea (39 µM) was added to Nine of the outbreak isolates and 2 nonoutbreak isolates were
TBE running buffer. Thirteen isolates from 10 neonates and one resistant to cefotaxime and ceftazidime (Table 2), but NGS analysis
isolate of a siphon in the NICU grouped into 1 clade (Fig. 2). All using ResFinder 2.1 and PlasmidFinder 1.3 (https://cge.cbs.dtu.dk/
services/) revealed no plasmid-mediated resistance mechanisms.
However, multiple sequence variants of ampC, ampR and ampD
were detected (see figures, Supplemental Digital Content 1–3,
TABLE 1.  Clinical Data of Outbreak Cases (n = 16; http://links.lww.com/INF/D391; http://links.lww.com/INF/D392;
PFGE Confirmed n = 10); From 01/01/2015 to 31/12/2015 http://links.lww.com/INF/D393). Outbreak isolates 768/15 and
775/15 presented enhanced MIC values for cefotaxime (4 mg/L,
Total (n = 16) Median [Min–Max] resistant) and ceftazidime (4 mg/L, intermediate) and revealed a 26
Gestational age (weeks) 34 + 3 [30 + 3 to 39 + 1] bp-deletion at the 3'-end of ampC, leading to an altered amino acid
 Female 34 + 3 [30 + 3 to 36 + 4] sequence. Furthermore, both highly resistant isolates (770/15 and
 Male 34 + 4 [31 + 4 to 39 + 1] 771/15) of 1 patient (case) contained a base substitution at posi-
Birth weight (g) 1.810 [1210 to 4040] tion 93 in ampD. This substitution leads to a premature stop and
 Female 1.795 [1210 to 2780] a truncated ampD. However, for the remaining outbreak isolates
 Male 1.885 [1700 to 4040]
Age at first detection (d) 10 [1 to 86]
with resistance (n = 7) and susceptibility (n = 3) to cefotaxime/
 Female 10 [2 to 31] ceftazidime and the resistant nonoutbreak isolate 904/15 no altera-
 Male 9.5 [1 to 86] tion of ampC, ampR, ampD or ompF and ompC and the attendant
Days from admission to first detection (d) 6.5 [0 to 22] promotor regions could be observed. Disc tests for AmpC produc-
 Female 10 [0 to 22] tion (D68C ESBL/AmpC-ID, MAST Group) were positive except
 Male 5 [0 to 13]
for the outbreak isolate 905/15 and 2 nonoutbreak isolates. These

© 2019 Wolters Kluwer Health, Inc. All rights reserved. www.pidj.com | 633

Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Steffen et al The Pediatric Infectious Disease Journal  •  Volume 38, Number 6, June 2019

FIGURE 2.  UPGMA clustering of genetic relationships of E. cloacae isolates (XbaI macrorestriction and PFGE). In gray ST815-
isolates of the 10 confirmed cases and one environmental sample. The dendrogram was calculated using BioNumerics
software (version 7.6.2, parameters tolerance 1.0; optimization 0.5). Displayed results of all 26 isolates were available for
analyses at RKI.

3 isolates were susceptible to cefotaxime and ceftazidime and
showed positive results in an ampC induction disc test using imipe-
nem/cefotaxime and imipeneme/ceftazidime as inducer/substrate
combinations as previously described.25
DISCUSSION
We investigated an E. cloacae outbreak in a German NICU
and found 16 cases (13 colonizations, 3 infections). Beside 10 con-
firmed cases, 6 probable cases without isolates available for geno-
typing but with a clear epidemiological link in place and time could
be attributed to the outbreak.
Three of the affected infants developed severe infections, 2
of them with brain abscesses and potential neurological sequelae.
Similar neurological E. cloacae infections were already described
in literature.5,6 This highlights the seriousness of Gram-negative
commensals as outbreak pathogens in neonatal settings and the
importance of early outbreak control.
The pathogen spread was most likely maintained by continu-
ous person-to-person transmission where colonized patients were
the sources for transient contamination of HCWs’ hands. Micro-
biological screening of the NICU staff had been conducted but did
not reveal colonization of staff members with the outbreak strain.
Extensive environmental investigations did not disclose outbreak
source or vehicle.
The following results point toward person-to-person trans-
mission: 1. the epidemiological link in place and time between the
16 cases; 2. the environmental screening led to the identification
of the outbreak strain in only 1 sample (siphon in the NICU, likely
having been contaminated by a case and no relevance for trans-
mission); 3. our outbreak investigation did not reveal alternative
hypotheses and 4. the outbreak could be stopped by enhanced bar-
rier precautions.
In accordance with our findings, several reports on E. cloa-
cae outbreak among neonates indicated patient-to-patient trans-
mission because of contaminated environment or cross-contami-
nation via HCWs’ hands8,26–28, while colonized HCWs as a source
of outbreaks with Gram-negative-bacteria have only rarely been
described in the literature.9,10 Talon et al29 described not only an
extensive spread of sporadic E. cloacae colonization in premature
neonates but also with a sequential spread of epidemic strains. This
is also similar to our observations. In a study by Stoesser et al,30
an outbreak setting of multidrug-resistant E. cloacae in a neona­
tal unit in Nepal was investigated on a genomic level. This led
to the detailed description of several genetic clusters of E. cloa-
cae outbreak strains of the outbreak situation and, like the present
study, also identified discrepancies in the susceptibility of outbreak
isolates, which were, nevertheless, genetically highly related. The
following challenges occurred when case search was performed.
On the one hand, the characteristics of E. cloacae initially led to a
concealment of the outbreak: owing to the facultative pathogenicity
of E. cloacae, only a small proportion of affected patients devel-
oped symptoms. Thus, microbiological screening was conducted
and gradually extended to other potentially affected wards leading
to detection of a high number of E. cloacae carriers. On the other

634 | www.pidj.com © 2019 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
The Pediatric Infectious Disease Journal  •  Volume 38, Number 6, June 2019 NICU Outbreak E. cloacae, Germany

FIGURE 3.  Phylogenetic tree (RAxML 7,2,8, GTR GAMMA model, Rapid hill-climbing, 100 starting trees) of E. cloacae isolates
calculated based on 3161 SNP positions (150 bp exclusion distance) after mapping NGS data to the reference sequence
(NZ_CP008823.1). Isolates involved in the outbreak are highlighted in gray. Displayed results of all 26 isolates were available
for analyses at RKI.

hand, isolates of the outbreak strain showed differing antibiotic
resistance patterns, which complicated the exclusion of nonout-
break-related E. cloacae carriers.
Macro restriction pattern of the E. cloacae outbreak strain
could not be resolved using a standard PFGE protocol because of
DNA degradation. This resulted in a delay of case confirmation.
Other molecular typing methods for rapid case confirmation in the
hospital (eg, the used AP-PCR) were not specific enough to differen-
tiate between outbreak cases and noncases and led to an overestima-
tion of the actual outbreak size. Consequence was the unnecessary
isolation of patients. The occurrence of epidemiologically remote
cases led to altering hypotheses regarding the cause of the outbreak.
Finally, the difficulties in identifying epidemiological links
between cases resulted in a search for more specific typing meth-
ods. By using a modified PFGE protocol, we identified the outbreak
strain. NGS-based MLST and NGS-based SNP-phylogeny also
clearly differentiated the outbreak strain from nonoutbreak strains.
Therefore, the outbreak was considerably reduced in number of
cases and affected wards. All cases had exposure to the NICU. The
last case was found to be colonized with the outbreak strain when
screened at readmission to the pediatric ward in October 2015.
Transmission had likely occurred during previous admission on the
NICU in July 2015.
The E. cloacae outbreak strain belonged to a new ST, the
ST815, a singleton that does not belong to any known clonal com-
plex. Nonoutbreak isolates were assigned to 11 different STs,
which is concordant with the reported phylogenetic diversity of
E. cloacae in general.31 Two nonoutbreak isolates belonged to the
same ST (ST145), but lacked a clear epidemiological link. Hence,
all nonoutbreak-related cases were considered as sporadic.
Interestingly, the isolates belonging to the outbreak strain
exhibited different MICs for ceftazidime and cefotaxime. NGS
analysis detected a truncated ampD for 2 of the ST815-isolates
(770/15, 771/15). Based on the ampC-mediated β-lactam resist-
ance model, the inactivation of ampD leads to an accumulation
of the ampD substrate, which acts as signaling effector of ampR
and consequently ampR effects the induction of ampC constantly.32
The effect of the 26 bp deletion at the 3'-end of the ampC gene
in ST815-isolates 768/15 and 775/15 is not yet clear and needs
further investigations. However, it was not possible to clarify the
underlying resistance mechanism in 6 of the 9 cefotaxime/ceftazi-
dime-resistant outbreak isolates with the present genome data. In
numerous studies, the source for ampC-mediated β-lactam resist-
ance was investigated; as cause of resistance, the involvement of
ampD, ampR and conclusively ampC expression or porine altera-
tions (ompC, ompF) are known.33,34 Detailed analysis of the gene
expression of the possible involved genes might clarify the resist-
ance mechanisms in our isolates. The bigger part of identified
sequence alterations found with the SNPfilter (3161 SNP positions)
between outbreak clade and nonoutbreak isolates were identified as
ST-specific SNP profiles. Using selected parameters of the analysis
(150 bp exclusion distance), the outbreak clade presents only 15
SNPs in total among the 14 sequenced ST815-isolates. These SNPs
cannot explain the phenotypic differences of the isolates.
Generalizability of our results may be reduced because of
limited number of available samples and the fact that we report on
one particular outbreak investigation. But the challenges described
here are mostly due to particularities of E. cloacae colonization,
infection and diagnostic methods, which are less affected by out-
break, setting and number of isolates.
Summarizing, the difficulties in discriminating outbreak
cases from noncases resulted in weeks of intense E. cloacae out-
break investigation activities in November 2015, even though the
transmission of the E. cloacae outbreak strain had been stopped
successfully beforehand. For early recognition of E. cloacae
clusters and prompt identification of all related cases, a close

© 2019 Wolters Kluwer Health, Inc. All rights reserved. www.pidj.com | 635

Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Steffen et al The Pediatric Infectious Disease Journal  •  Volume 38, Number 6, June 2019

TABLE 2.  Characteristics of Enterobacter cloacae Isolates From a Neonatal Ward in Germany, From 10/08/2015 to
06/11/2015

Patient Date of PFGE- Sequence MIC CTX MIC CAZ

Isolate No. No. Material Isolation Type type (ST) Resistances (mg/L) (mg/L)

765/15 2 Brain abscess 10/08/2015 A ST815 AMP, CTX, CAZ >16 >32
773/15 2 Nasopharyngeal swab 10/08/2015 A ST815 AMP, CTX, CAZ >16 >32
774/15 9 Perianal swab 10/08/2015 A ST815 AMP ≤1 ≤2
770/15 6 Blood culture 13/08/2015 A ST815 AMP, CTX, CAZ >16 >32
771/15 6 Blood culture 13/08/2015 A ST815 AMP, CTX, CAZ >16 >32
764/15 1 Brain abscess 17/08/2015 A ST815 AMP, CTX, CAZ >16 >32
766/15 3 Perianal swab 17/08/2015 A ST815 AMP, CTX, CAZ >16 >32
767/15 4 Perianal swab 17/08/2015 A ST815 AMP, CTX, CAZ >16 >32
769/15 1 Perianal swab 17/08/2015 A ST815 AMP, CTX, CAZ >16 >32
768/15 5 Perianal swab 24/08/2015 A ST815 AMP, CTX 4 ≤2
772/15 7 Perianal swab 24/08/2015 A ST815 AMP, CTX, CAZ >16 >32
775/15 10 Perianal swab 31/08/2015 A ST815 AMP, CTX, CAZ (I) 4 4
776/15 11 Perianal swab 03/09/2015 B ST104 AMP, CTX, CAZ, ETP >16 >32
777/15 Siphon 1 Siphon NICU 19/09/2015 A ST815 AMP ≤1 ≤2
900/15 14 Perianal swab 21/09/2015 H ST97 AMP ≤1 ≤2
896/15 13 Nasopharyngeal swab 23/09/2015 D ST190 – ≤1 ≤2
897/15 Siphon 2 Siphon maternity ward 24/09/2015 E ST907 AMP, CIP ≤1 ≤2
901/15 15 Nasopharyngeal swab 03/10/2015 I ST145 – ≤1 ≤2
899/15 Staff Stool sample 06/10/2015 G ST816 AMP ≤1 ≤2
898/15 Staff Stool sample 09/10/2015 F ST128 AMP ≤1 ≤2
902/15 16 Perianal swab 12/10/2015 J ST278 – ≤1 ≤2
903/15 17 Nasopharyngeal swab 12/10/2015 K ST254 AMP ≤1 ≤2
904/15 18 Perianal swab 15/10/062015 L ST114 AMP, CTX, CAZ >16 >32
905/15 19 Perianal swab 22/10/2015 A ST815 AMP ≤1 ≤2
778/15 12 Perianal swab 29/10/2015 C ST118 AMP 2 ≤2
906/15 20 Nasopharyngeal swab. 06/11/2015 M ST145 – ≤1 ≤2
Displayed results of all 26 isolates were available for analyses at RKI. Sequence types were identified by MLST (https://pubmlst.org/ecloacae/).
AMP indicates ampicillin; CTX, cefotaxime; CAZ, ceftazidime; CIIP, ciprofloxacin; ETP, ertapenem; I, intermediate-resistant.

collaboration between microbiologists, clinicians and epidemiolo-
gists is essential.
Highly discriminatory genotyping methods such as PFGE
and NGS are needed for a clear differentiation of outbreak and non-
outbreak isolates in time and for avoiding unnecessary isolation of
patients.
ACKNOWLEDGMENTS
The authors thank Sibylle Müller-Bertling and Kirstin Gan-
ske for excellent technical assistance. The authors are grateful to
Dr. Birgit Strommenger for supporting the PFGE gel image analy-
sis using BioNumerics software. They thank Christian Winter and
Andrea Sanchini for commenting earlier versions of this manuscript
and the staff of the hospital for their support in the outbreak inves-
tigation. G.S., W.S., H.A., T.E. and S.H. were part of the outbreak
team and conducted the epidemiological outbreak investigations.
M.P., M.K., S.G., G.W., S.F. and Y.P. conducted the microbiologi-
cal investigations and designed the microbiological experiments.
G.S., G.W., Y.P., T.E. and S.H. conceptualized the studies. G.S., Y.P.
and S.H. drafted the manuscript. All authors critically revised the
manuscript and approved the final version. G.S. is corresponding
author and guarantor.
REFERENCES
1. Mezzatesta ML, Gona F, Stefani S. Enterobacter cloacae complex:
clinical impact and emerging antibiotic resistance. Future Microbiol.
2012;7(7):887–902.
2. European Center for Disease Prevention and Control. Point prevalence sur-
vey of healthcare-associated infections and antimicrobial use in European
acute care hospitals. Stockholm: European Center for Disease Prevention
and Control; 2013.
3. Parm U, Metsvaht T, Sepp E, et al. Risk factors associated with gut and
nasopharyngeal colonization by common Gram-negative species and yeasts
in neonatal intensive care units patients. Early Hum Dev. 2011;87:391–399.
4. Fanaro S, Chierici R, Guerrini P, et al. Intestinal microflora in early infancy:
composition and development. Acta Paediatr Suppl. 2003;91(441):48–55.
5. Haller S, Cassini PDA, Suetens C, et al. Neurological sequelae of health-
care-associated sepsis in very-low-birthweight infants: umbrella review and
evidence-based outcome tree. Eurosurveillance 2016;21(8):1–10.
6. Chen HN, Lee ML, Yu WK, et al. Late-onset Enterobacter cloacae sepsis
in very-low-birth-weight neonates: experience in a medical center. Pediatr
Neonatol. 2009;50(1):3–7.
7. Dalben M, Varkulja G, Basso M, et al. Investigation of an outbreak of
Enterobacter cloacae in a neonatal unit and review of the literature. J Hosp
Infect. 2008;70(1):7–14.
8. Pestourie N, Garnier F, Barraud O, et al. Outbreak of AmpC beta-lactamase-
hyper-producing Enterobacter cloacae in a neonatal intensive care unit in a
French teaching hospital. Am J Infect Control. 2014;42(4):456–458.
9. Waters V, Larson E, Wu F, et al. Molecular epidemiology of gram-negative
bacilli from infected neonates and health care workers’ hands in neonatal
intensive care units. Clin Infect Dis. 2004;38(12):1682–1687.
10. Ulrich N, Gastmeier P, Vonberg RP. Effectiveness of healthcare worker
screening in hospital outbreaks with gram-negative pathogens: a systematic
review. Antimicrob Resist Infect Control. 2018;7:36.
11. Flacking R, Lehtonen L, Thomson G, et al. Closeness and separation in
neonatal intensive care. Acta Paediatr. 2012;101(10):1032–1037.
12. Antwort zu der Anfrage der Abgeordneten Astrid Schramm (DIE LINKE),
Betreff: Keime auf der Kinderintensivstation des Marienhaus Klinikums
Saarlouis: Landtag des Saarlandes, 15th legislative period; 2016.
13. Haller S, Eckmanns T, Benzler J, et al. Results from the first 12 months
of the national surveillance of healthcare associated outbreaks in Germany,
2011/2012. PloS One 2014;9(5):e98100.
14. Menard C, Brousseau R, Mouton C. Application of polymerase chain
reaction with arbitrary primer (AP-PCR) to strain identification of
Porphyromonas (Bacteroides) gingivalis. FEMS Microbiol Lett. 1992;74(2–
3):163–168.
15. Tenover FC, Arbeit RD, Goering RV, et al. Interpreting chromosomal DNA
restriction patterns produced by pulsed-field gel electrophoresis: criteria for
bacterial strain typing. J Clin Microbiol. 1995;33(9):2233–2239.
16. Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for
Illumina sequence data. Bioinformatics 2014;30(15):2114–2120.

636 | www.pidj.com © 2019 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
The Pediatric Infectious Disease Journal  •  Volume 38, Number 6, June 2019 NICU Outbreak E. cloacae, Germany
17. Coil D, Jospin G, Darling AE. A5-miseq: an updated pipeline to assem-
ble microbial genomes from Illumina MiSeq data. Bioinformatics
2015;31(4):587–589.
18. Zankari E, Hasman H, Cosentino S, et al. Identification of acquired antimi-
crobial resistance genes. J Antimicrob Chemother. 2012;67(11):2640–2644.
19. Larsen MV, Cosentino S, Rasmussen S, et al. Multilocus sequence typing of
total-genome-sequenced bacteria. J Clin Microbiol. 2012;50(4):1355–1361.
20. Miyoshi-Akiyama T, Hayakawa K, Ohmagari N, et al. Multilocus sequence
typing (MLST) for characterization of Enterobacter cloacae. PLoS One
2013;8:e66358.
21. Li H, Durbin R. Fast and accurate long-read alignment with Burrows-
Wheeler transform. Bioinformatics 2010;26(5):589–595.
22. Conlan S, Thomas PJ, Deming C, et al.; NISC Comparative Sequencing
Program. Single-molecule sequencing to track plasmid diversity of hospital-
associated carbapenemase-producing Enterobacteriaceae. Sci Transl Med.
2014;6:254ra126.
23. Becker L, Fuchs S, Pfeifer Y, et al. Whole genome sequence analysis of
CTX-M-15 producing Klebsiella isolates allowed dissecting a polyclonal
outbreak scenario. Front Microbiol. 2018;9:322.
24. Stamatakis A. RAxML version 8: a tool for phylogenetic analysis and post-
analysis of large phylogenies. Bioinformatics 2014;30:1312–1313.
25. Dunne WM Jr, Hardin DJ. Use of several inducer and substrate antibiotic
combinations in a disk approximation assay format to screen for AmpC
induction in patient isolates of Pseudomonas aeruginosa, Enterobacter spp.,
Citrobacter spp., and Serratia spp. J Clin Microbiol. 2005;43:5945–5949.
26. Adler A, Solter E, Masarwa S, et al. Epidemiological and microbio-
logical characteristics of an outbreak caused by OXA-48-producing
© 2019 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Enterobacteriaceae in a neonatal intensive care unit in Jerusalem, Israel. J
Clin Microbiol. 2013;51:2926–2930.
27. Liu SC, Leu HS, Yen MY, et al. Study of an outbreak of Enterobacter cloa-
cae sepsis in a neonatal intensive care unit: the application of epidemio-
logic chromosome profiling by pulsed-field gel electrophoresis. Am J Infect
Control. 2002;30:381–385.
28. Fernández-Baca V, Ballesteros F, Hervás JA, et al. Molecular epidemiologi-
cal typing of Enterobacter cloacae isolates from a neonatal intensive care
unit: three-year prospective study. J Hosp Infect. 2001;49:173–182.
29. Talon D, Menget P, Thouverez M, et al. Emergence of Enterobacter cloacae
as a common pathogen in neonatal units: pulsed-field gel electrophoresis
analysis. J Hosp Infect. 2004;57(2):119–125.
30. Stoesser N, Sheppard AE, Shakya M, et al. Dynamics of MDR Enterobacter
cloacae outbreaks in a neonatal unit in Nepal: insights using wider sam-
pling frames and next-generation sequencing. J Antimicrob Chemother.
2015;70(4):1008–1015.
31. Izdebski R, Baraniak A, Herda M, et al. MLST reveals potentially high-
risk international clones of Enterobacter cloacae. J Antimicrob Chemother.
2015;70(1):48–56.
32. Jacobs C, Frère JM, Normark S. Cytosolic intermediates for cell wall bio-
synthesis and degradation control inducible beta-lactam resistance in gram-
negative bacteria. Cell 1997;88:823–832.
33. Guérin F, Isnard C, Cattoir V, et al. Complex regulation pathways of AmpC-
mediated β-lactam resistance in Enterobacter cloacae complex. Antimicrob
Agents Chemother. 2015;59:7753–7761.
34. Babouee Flury B, Ellington MJ, Hopkins KL, et al. The differential importance
of mutations within AmpD in cephalosporin resistance of Enterobacter aero-
genes and Enterobacter cloacae. Int J Antimicrob Agents. 2016;48:555–558.
www.pidj.com | 637