Clinical Microbiology and Infection 22 (2016) 1003.e1e1003.

e8

Contents lists available at ScienceDirect

Clinical Microbiology and Infection

journal homepage: www.clinicalmicrobiologyandinfection.com

Original article

Outbreak investigation for toxigenic Corynebacterium diphtheriae

wound infections in refugees from Northeast Africa and Syria in
Switzerland and Germany by whole genome sequencing
D.M. Meinel 1, 2, 3, R. Kuehl 4, R. Zbinden 5, V. Boskova 6, C. Garzoni 7, D. Fadini 8,
M. Dolina 9, B. Blümel 10, T. Weibel 11, S. Tschudin-Sutter 4, A.F. Widmer 4, J.A. Bielicki 12,
A. Dierig 12, U. Heininger 12, R. Konrad 2, 13, A. Berger 2, 13, V. Hinic 1, D. Goldenberger 1,
A. Blaich 1, T. Stadler 6, M. Battegay 4, 14, A. Sing 2, 13, 14, A. Egli 1, 3, *, 14
1)
Clinical Microbiology, University Hospital Basel, Basel, Switzerland
2)
Bavarian Health and Food Safety Authority (LGL), Oberschleissheim, Germany
3)
Applied Microbiology Research, Department of Biomedicine, University Basel, Basel, Switzerland
4)
Infectious Diseases and Hospital Epidemiology, University Hospital Basel, Basel, Switzerland
5)
Institute for Medical Microbiology, University of Zurich, Zurich, Switzerland
6)
Computational Evolution, D-BSSE, ETH Zurich, Basel, Switzerland
7)
Department of Internal Medicine and Infectious Diseases, Clinica Luganese, Lugano, Switzerland
8)
Internal Medicine, Ospedale di Mendrisio, Mendrisio, Switzerland
9)
Clinical Microbiology, EOLAB, Bellinzona, Switzerland
10)
Institute of Medical Microbiology and Hygiene, University Medical Centre Freiburg, Freiburg, Germany
11)
Clinical Microbiology, Labor Team W, Saint Gallen, Switzerland
12)
Paediatric Infectious Diseases, University of Basel Children's Hospital, Basel, Switzerland
13)
German National Consiliary Laboratory on Diphtheria, Oberschleissheim, Germany

a r t i c l e i n f o a b s t r a c t

Article history: Toxigenic Corynebacterium diphtheriae is an important and potentially fatal threat to patients and public
Received 22 May 2016 health. During the current dramatic influx of refugees into Europe, our objective was to use whole
Received in revised form genome sequencing for the characterization of a suspected outbreak of C. diphtheriae wound infections
28 July 2016
among refugees. After conventional culture, we identified C. diphtheriae using matrix-assisted laser
Accepted 19 August 2016
Available online 30 August 2016
desorption/ionization time-of-flight (MALDI-TOF) and investigated toxigenicity by PCR. Whole genome
sequencing was performed on a MiSeq Illumina with >70
coverage, 2
250 bp read length, and
Editor: E. Bottieau mapping against a reference genome. Twenty cases of cutaneous C. diphtheriae in refugees from East
African countries and Syria identified between April and August 2015 were included. Patients presented
Keywords: with wound infections shortly after arrival in Switzerland and Germany. Toxin production was detected in
Corynebacterium diphtheriae 9/20 (45%) isolates. Whole genome sequencing-based typing revealed relatedness between isolates using
Emerging diseases neighbour-joining algorithms. We detected three separate clusters among epidemiologically related
Outbreak investigation refugees. Although the isolates within a cluster showed strong relatedness, isolates differed by >50
Refugee
nucleotide polymorphisms. Toxigenic C. diphtheriae associated wound infections are currently observed
Toxin-production
more frequently in Europe, due to refugees travelling under poor hygienic conditions. Close genetic
Typing
Whole genome sequencing relatedness of C. diphtheriae isolates from 20 refugees with wound infections indicates likely transmission
between patients. However, the diversity within each cluster and phylogenetic time-tree analysis suggest
that transmissions happened several months ago, most likely outside Europe. Whole genome sequencing
offers the potential to describe outbreaks at very high resolution and is a helpful tool in infection tracking
and identification of transmission routes. D.M. Meinel, CMI 2016;22:1003.e1e1003.e8
© 2016 Published by Elsevier Ltd on behalf of European Society of Clinical Microbiology and Infectious
Diseases.

* Corresponding author. A. Egli, Clinical Microbiology, University Hospital Basel,

Petersgraben 4, 4031 Basel, Switzerland.
E-mail address: adrian.egli@usb.ch (A. Egli).
14
M.B., A.S. and A.E. contributed equally to this study.

http://dx.doi.org/10.1016/j.cmi.2016.08.010
1198-743X/© 2016 Published by Elsevier Ltd on behalf of European Society of Clinical Microbiology and Infectious Diseases.
1003.e2 D.M. Meinel et al. / Clinical Microbiology and Infection 22 (2016) 1003.e1e1003.e8
Introduction
The current migration of refugees from Africa and the Middle
East is an important challenge for primary-care institutions and
public health systems in Europe [1e3]. Frequently, only a vague
medical history is available for refugees and immunization docu-
ments are lacking. Language barriers, cultural differences and
traumatic experiences may lead to missing details about the travel
route and further important aspects for medical management.
Tracking of potential infectious sources is made difficult as refugees
from the same route arrive in multiple, geographically disparate
refugee reception centres and are rapidly re-located on arrival.
In western countries, wound infections with Corynebacterium
diphtheriae are mainly a problem of travellers [4,5], associated with
poor hygiene conditions in some urban areas [6], and observed in
intravenous drug abusers [7]. The incubation period is short, usu-
ally 1 week [8]. In wound infections, C. diphtheriae is often
accompanied by other skin pathogens, such as Streptococcus pyo-
genes or Staphylococcus aureus [6]. In high-income countries, these
and other types of C. diphtheriae infections present an almost
forgotten disease, which is still endemic in low-income countries.
The ECDC warned in 2015 about the possibility of cutaneous
diphtheria cases in refugees from endemic countries and unvacci-
nated travellers returning from these regions [9].
The dominant symptoms of respiratory and cutaneous diph-
theria are caused by the diphtheria toxin, which is encoded on a
prophage [10,11]. Cutaneous diphtheria infections may serve as a
source for the toxigenic pathogens and cause the more severe
pharyngeal diphtheria in susceptible persons. Therefore, it poses a
potential threat to public health [12]. Additionally, the toxin gene
can be carried by the zoonotic Corynebacterium ulcerans and
Corynebacterium pseudotuberculosis [13,14]. Problematically there-
fore, wound infections with toxigenic zoonotic strains may pose an
additional public health threat if relevant strains spread within the
human population [15,16]. However, the clinical symptoms of
toxigenic C. ulcerans and C. diphtheriae infections are very similar
[17] and rising numbers of toxigenic C. ulcerans infections have
been reported, e.g. in the UK [18] and Germany [19].
After culture, C. diphtheriae, C. ulcerans and C. pseudotuberculosis
can be rapidly and reliably identified with matrix-assisted laser
desorption ionization time-of-flight mass spectrometry (MALDI-
TOF MS) [20e22]. Only specialized laboratories have the facilities to
readily detect the presence of the diphtheria toxin by PCR or
measure toxin production by modified Elek test [23,24].
We report on a series of 20 C. diphtheriae-associated wound
infections in refugees from East African countries and Syria
observed shortly after their arrival in reception centres in
Switzerland and Germany between April and August 2015. We
discuss the likely epidemiology of these cases using, for the first
time, whole genome sequencing technology for high-resolution
typing of the clinical isolates in the context of a suspected outbreak.
Materials and methods
Patients
Between April and August 2015, 20 refugees from Eritrea, Ethiopia,
Somalia and Syria presented with suspicious skin lesions to emer-
gency departments in Switzerland and Germany or were reported to
the National Consiliary Laboratory on Diphtheria in Germany. Cory-
nebacterium diphtheriae was identified as a causative agent in all
patients. In addition to isolates from affected refugees, we included
five clinical isolates from epidemiologically non-linked Swiss or
German patients also presenting with wound infections. These five
isolates were epidemiologically not associated with each other.
Microbiological identification and resistance testing
Swabs and biopsies were cultured on standard 5% sheep blood
and selective and differential tellurite agar plates (both Becton
Dickinson, Allschwil, Switzerland). Subsequent identification was
performed using MALDI-TOF MS (Microflex, Bruker Daltonics,
Bremen, Germany) using the database DB-5989 (May 2015; [22]).
Biochemical identification was used to confirm the MALDI-TOF MS
results (API Coryne, bioMe


rieux, Marcy-l'Etoile, France [25,26]).
Resistance testing was performed with Etests (bioMe
rieux) and
MICs of penicillin, ciprofloxacin and vancomycin were interpreted
according to EUCAST recommended breakpoints (v6, valid 1
January 2016) and MICs of ceftriaxone and imipenem according to
CLSI recommended breakpoints (M45, 3 October 2015).
Detection of the diphtheria toxin gene
The PCR was carried out in the German National Consiliary
Laboratory on Diphtheria and at the University Hospital Basel as
described earlier [23]. In addition, a modified Elek assay was per-
formed as described previously to detect toxin production [24].
Whole genome sequencing based typing
Seventeen of the twenty available microbiological isolates from
six different laboratories were used for whole genome sequencing.
Analysis was carried out as described previously [14] and
C. diphtheriae C7 (b), GenBank accession: CP003210.1 was used as
reference sequence. All sequencing data are available from the
Sequence Read Archive (http://www.ebi.ac.uk/ena) under experi-
ment accession number PRJEB14914.
Time-tree analysis, detection of virulence factors and prophages
are detailed in the Supplementary material (see Appendix S1).
Ethics
All patients were successfully treated, but due to re-location
could not be followed for a longer time period. Due to the impor-
tance for public health, the clinical features of the patients infected
by toxigenic C. diphtheriae have been in the public domain (Swiss
Ministry of Health, Diphtherie Note, June 2015).
Results
Case descriptions
From April to August 2015, 20 refugees from Eritrea, Ethiopia,
Syria and Somalia presented at emergency departments throughout
Switzerland and Germany with C. diphtheriae-associated wound
infections. Seventeen isolates were collected from refugees staying
at five different Swiss reception centres (Basel, Chiasso, Chur,
Kreuzlingen and Lausanne). Three more cases occurred in Germany
within the same time range and were included in the analysis. Be-
tween August and December 2015, to the best of our knowledge only
two more non-toxigenic cases of wound infections occurred in
Switzerland (cases and isolates were not available for analysis).
Most of the patients were male (88%) and the median age was 20
years (interquartile range 17e25) (Table 1). All patients presented
with wound infections, which occurred several days to few weeks
before the arrival at the reception centres. Most commonly, a soft-
tissue infection with superficial ulceration or abscess formation
was observed on the limbs or in the groin area. The past medical
history was largely unknown and all patients reported not having
received any prior treatment for their wound infection. The vacci-
nation status against C. diphtheriae toxin was either unknown or
 D.M. Meinel et al. / Clinical Microbiology and Infection 22 (2016) 1003.e1e1003.e8 1003.e3

Table 1
Clinical and travel related characteristics of patients

ID Age Gender Date of sampling Location of wound(s) Refugee centre Country of origin Travel route

1 32 M 12.06.15 Lower left leg Basel Eritrea Eritrea, Sudan, Libya, Italy, Switzerland
2 19 M 17.06.15 Multiple gluteal pustular lesions Basel Eritrea Eritrea, Sudan, Libya, Italy, Switzerland
3 16 M 29.05.15 Panaritium right toe, Dig I Basel Somalia Ethiopia, Sudan, Libya, Italy, Switzerland
4 16 F 05.05.15 Multiple wounds and ulcers lower leg Basel Eritrea Eritrea, Sudan, Libya, Italy, Switzerland
5 25 M 22.06.15 Abscess, left knee Basel Eritrea Eritrea, Sudan, Libya, Italy, Switzerland
6 17 M 24.06.15 Pustular lesions, abscess left leg Basel Eritrea Eritrea, Sudan, Libya, Italy, Switzerland
7 36 M 03.08.15 Ulcers, lower right limb and foot Basel Eritrea Eritrea, Sudan, Libya, Italy, Switzerland
8 20 M 22.06.15 Superficial wound, scrotum Chiasso Eritrea Unknown
9 25 F 22.06.15 Chronic wound, left foot Chiasso Somalia Unknown
10 25 M 13.05.15 Deep wound, right food and lower leg Chiasso Eritrea Eritrea, Sudan, Libya, Italy, Switzerland
11 17 M 30.04.15 Chronic wound, groin Chiasso Somalia Somalia, Libya, Italy, Switzerland
12 15 M 05.05.15 Skin ulcers, groin Chiasso Somalia Unknown
13 17 M 08.05.15 Superficial wound, right hand Chiasso unknown Unknown
14 22 M 26.06.15 Superficial wound, lower leg Chur unknown Unknown
15 23 M 29.04.15 Superficial wound, unknown location Kreuzlingen Eritrea Libya, Eritrea, Ethiopia, Sudan, Libya,
Italy, Switzerland
16 21 M 14.07.15 Wound, left knee Lausanne Eritrea Eritrea, Sudan, Libya, Italy, Switzerland
17 18 M 30.07.15 Wound, elbow Lausanne Eritrea Unknown
18 20 NA 02.06.15 Wound, right leg Stadtroda Eritrea Unknown
19 18 NA 02.06.15 Several wounds, skin ulcers Nurnberg Ethiopia Unknown
20 38 NA 15.06.15 Abscess, right foot Cham Syria Unknown
A 42 F 26.09.13 Pustular lesion d Switzerland Travel to South America
B 58 M 30.10.13 Skin ulcers d Switzerland Travel to Tanzania
C 18 M 24.12.14 Superficial skin lesion d Switzerland No travel history
D NA M 09.03.15 Wound infection d Germany Travel to Tanzania
E 35 M 30.06.15 Pharyngeal diphtheria d Germany Travel to Thailand

Patient ID is the same throughout the manuscript. Age in years at date of sampling. NA, not available. Diagnostic swabs were collected from locations of wounds. ID 1e20 were
collected from refugees. ID AeE were collected from Swiss or German residents. ID 3 and 5 had a documented surgical intervention.

reported as negative. Only 2/20 (10%) patients needed surgical
intervention due to abscess formationdin these cases the infection
was generally chronic and in one patient it was further complicated
by the presence of methicillin-resistant Staphylococcus aureus
(MRSA).
Most patients (70.6%) were started on a 1-week course of a b-
lactam antibiotic, most often amoxicillin/clavulanic acid. Patients
with chronic infection, abscesses or MRSA were treated for longer,
underwent surgical debridement and/or received combination
antibiotic therapy with clindamycin or sulfamethoxazole/

Table 2
Microbiological characteristics

ID Sub-species MALDI Toxin Elek Antibiotic resistance (S/I/R, MIC) Further pathogens
Score PCR test
Penicillin Ceftriaxone Imipenem Ciprofloxacin Vancomycin

1 gravis 2.312 Neg ND R, 0.25 I, 2.00 S, 0.064 S, 0.016 S, 0.50 MSSA, Streptococcus pyogenes
2 mitis 2.344 Neg ND R, 0.19 I, 1.50 S, 0.064) S, 0.032 S, 0.75 MSSA, Scabies
S. pyogenes
3 mitis 2.321 Pos þ R, 0.38 I, 1.50 S, 0.125 S, 0.023 S, 0.50 S. pyogenes
4 mitis 2.243 Neg ND R, 0.38 I, 1.50 S, 0.19 S, 0.032 S, 0.75 S. pyogenes
5 ND 2.377 Neg ND R, 0.50 S, 1.00 S, 0.032 S, 0.125 S, 2.00 MRSA, S. pyogenes
6 mitis 1.995 Pos (þ) R, 0.38 I, 1.50 S, 0.125 S, 0.023 S, 0.50 S. pyogenes, Scabies
7 ND ND Pos þ R, 0.25 I, 2.00 S, 0.19 S, 0.064 S, 0.50 MSSA, S. pyogenes
8 gravis 2.246 Neg ND R, 0.19 I, 1.50 S, 0.094 S, 0.064 S, 0.50 MSSA, S. pyogenes, F. nucleatum
9 mitis 2.376 Pos (þ) R, 0.19 I, 1.50 S, 0.094 S, 0.064 S, 0.50 MSSA, S. pyogenes
10 gravis 2.281 Neg ND R, 0.19 I, 2.00 S, 0.094 S, 0.064 S, 0.75 MSSA, S. pyogenes, Scabies
11 ND 2.347 Pos þ R, 0.25 I, 2.00 S, 0.19 S, 0.064 S, 0.50 MRSA
12 ND ND Pos ND ND ND ND ND ND MRSA, S. pyogenes
13 ND ND Neg ND ND ND ND ND ND MRSA
14 ND ND Neg ND R, 0.19 I, 1.50 S, 0.047 S, 0.094 S, 0.75 MSSA, S. pyogenes
15 gravis 2.368 Neg ND R, 0.25 I, 1.50 S, 0.125 S, 0.094 S, 0.50 MSSA
16 gravis 2.365 Neg ND S, 0.125 I, 1.50 S, 0.094 S, 0.094 S, 0.50 S. pyogenes
17 gravis 2.390 Neg ND S, 0.125 I, 1.50 S, 0.094 S, 0.064 S, 0.75 MSSA, S. pyogenes
18 mitis ND Pos þ R, 0.19 S, 1.00 S, 0.094 S, 0.047 S, 0.75 S. pyogenes
19 mitis ND Pos þ R, 0.19 S, 1.00 S, 0.094 S, 0.047 S, 0.75 none
20 mitis ND Pos þ R, 0.19 S, 1.00 S, 0.064 S, 0.047 S, 0.75 none
A ND 2.259 Neg ND S, 0.125 I, 1.5 S, 0.125 S, 0.125 S, 0.125 None
B ND 2.315 Neg ND R, 0.19 I, 2.0 S, 0.125 S, 0.25 S, 0.50 None
C ND 2.267 Neg ND R, 1.5 I, 1.5 S, 0.125 S, 0.064 S, 0.50 None
D gravis ND Neg ND S, 0.125 I, 1.5 S, 0.064 S, 0.047 S, 0.50 None
E mitis ND Pos þ R, 0.19 S, 1.0 S, 0.064 S, 0.047 S, 0.75 None

ND, not determined. MRSA, methicillin-resistant Staphylococcus aureus; MSSA, methicillin-sensitive Staphylococcus aureus; Pos, positive; Neg, negative; þ, positive; (þ), weak
positive.
Penicillin, ciprofloxacin and vancomycin MICs were interpreted according to EUCAST recommended breakpoints (v6, valid 01.01.2016) and ceftriaxone and imipenem
according to CLSI recommended breakpoints (M45, 3 October 2015). S, sensitive; I, intermediate; R, resistant.
ID 1e20 are from refugees and ID AeE are from Swiss and German residents.
1003.e4 D.M. Meinel et al. / Clinical Microbiology and Infection 22 (2016) 1003.e1e1003.e8

trimethoprim. All patients were lost during follow up, as they were
transferred to secondary refugee processing centres.
Microbiological investigation
The microbiological analyses demonstrated that C. diphtheriae
associated wound infection most often occurred in combination
with S. aureus and Streptococcus pyogenes (Table 2). The S. aureus
isolates showed methicillin-resistance in 4/20 cases. MALDI-TOF
MS was able to robustly identify all C. diphtheriae isolates with a
median score of 2.344 (interquartile range 2.281e2.368) and the
identification was confirmed with biochemical assay. Nine of the 20
isolates were tested toxin-gene-positive by PCR. For most isolates,
susceptibility to penicillin, ceftriaxone, ciprofloxacin, imipenem
and vancomycin was determined (Table 2).
Whole genome sequencing analysis for typing
We performed whole genome sequencing based typing on the
available 17/20 refugee isolates (isolates 5,12 and 13 were not
available for sequencing). In addition to the suspected outbreak-
related strains, we used five clearly non-outbreak-related isolates
from Switzerland and Germany (A to E). Fig. 1 shows the phy-
logeny based on the single nucleotide polymorphisms (SNPs)
detected in all isolates using the neighbour-joining algorithm for
all outbreak and non-outbreak-related isolates. Three clusters
could be observed. Importantly, those clusters were not formed
from perfectly identical isolates but from isolates that differed by
up to 50e150 SNPs, whereas isolates not found in those clusters
differ by up to approximately 30 000 SNPs. Although this finding
indicates extremely close relatedness, isolates within the clusters
were not identical. The additional isolates, originating from ref-
ugees as well as local persons with a relevant travel history, did
not form any dense clusters and so represent unrelated isolates.
Closer inspection of the patient history for the clustering isolates
showed that the dates of presentation/isolation diverged within
the clusters by up to several months: Cluster 1: 64 days; cluster
2: 22 days; and cluster 3: 61 days. Additionally, similar dates of
isolation did not form sub-clusters, but were completely unor-
dered within each cluster. Interestingly, clusters 1 and 2 were
completely comprised of toxigenic isolates. To verify our results
we carried out pulsed-field gel electrophoresis-based typing on
the Swiss isolates. Pulsed-field gel electrophoresis typing was
consistent with the typing by whole genome sequencing (see
Supplementary material, Fig. S1).
To get an estimate of the transmission times, we performed
phylogenetic analyses of the data. At the highest plausible substi-
tution rate for bacteria (1.82
105 substitutions/bp/year), the re-
sults of our analyses indicate that the shortest distance between a
transmission event and its closest corresponding tip (sample) is
greater than 1.5 years (Fig. 2). Moreover, even for bacteria at an
unrealistic substitution rate of 1.82
104 substitutions/bp/year,
every transmission event is at least 3 weeks apart from its closest
tip (see Supplementary material, Table S1). In brief, we did not

Fig. 1. Phylogenetic relationship of Corynebacterium diphtheriae isolates. Depicted are the phylogenetic relationships of the isolates among each other. The tree was calculated using
the neighbour-joining algorithm and C. diphtheriae C7 (b) as reference for single nucleotide polymorphism detection. The observed clustering isolates are marked by a box and
labelled with cluster 1, 2 and 3 on the rightmost panel. The date of isolation of the C. diphtheriae from the patient is given as well as the country of origin of the refugee and also
toxigenicity as determined by PCR. Bootstrapping values have been calculated from 1000 iterations and showed robust separation for all nodes (100 for all bootstrapping values).
The isolates from refugees are labelled with numbers, isolates from Swiss or German patients are labelled with letters for better visibility.
 D.M. Meinel et al. / Clinical Microbiology and Infection 22 (2016) 1003.e1e1003.e8 1003.e5

1 : 0.14
272.08
2 : 0.13
416.49
10 : 0.22

15 : 0.26
353.87
7.79
16 : 0.05
2.08
11.94
17 : 0.01

8 : 0.12

518.38 11 : 0.26
2.04

11.05 7 : 0.00

15.92 9 : 0.12

6
347.76 3 : 0.18

424.85
2
4 : 0.25

14 : 0.10
496.85
5

20 : 0.13
428.23
6 : 0.11
12.97
18 : 0.17
6.21
19 : 0.17

500.0 400.0 300.0 200.0 100.0 0.0

predicted years of pathogen evolution

Fig. 2. Time-tree analysis. We used the bifurcation times in the maximum clade credibility tree as a proxy for transmission events. The blue horizontal lines represent the 95%
highest posterior density (HPD) intervals of the node height estimates. The median node height estimate is indicated next to each internal node. Each tip is annotated with the
sample ID and the tip height (in years). In the phylogenetic tree reconstructed assuming the substitution rate of 1.82
105 substitutions/bp/year, the transmission closest to a
sampling event, i.e. the node with the shortest pendant branch, occurred 1.78 (95% HPD 1.57e2.03) years before its corresponding tip (sample number 11). Of note, the substitution
rate of 1.82
105 substitutions/bp/year is unlikely for this bacterium and must be considered the absolutely fastest mutation rate [46]. Using lower substitution rates for the tree
reconstruction, all bifurcation events get pushed further into the past (see Supplementary material, Table S1).

identify direct transmission events and the bifurcation times date
back to a time period before the refugees left Africa. Hence, our data
strongly suggest infection or colonization before the development
of acute wounds from at least three distinct reservoirs containing
highly similar C. diphtheriae.
Multiple virulence factors are present in the isolates
Interestingly, we found that the isolates associated with the
outbreak-like clusters 1 and 2 were toxigenic, whereas unrelated
isolates only very rarely carried the toxin gene. The results of the
toxin gene detection by whole genome sequencing were consistent
with the PCR-based detection method and for all available toxigenic
isolates expression of the diphtheria toxin was confirmed by the
modified Elek test.
In addition to the well-known toxin gene, we searched for other
previously described pathogenicity islands containing virulence
factors, which we grouped into adhesive pili, siderophore biogen-
esis, siderophore uptake gene clusters, and iron-uptake-related
genes [27,28]. Adhesive pili have been shown to be of great
importance for bacterial colonization, pathogenesis and biofilm
development. Notably, the minor pilin SpaA was shown to be suf-
ficient for the specific adhesion of C. diphtheriae to pharyngeal cells
and therefore of high importance for causing respiratory diphtheria
[29]. Interestingly, we did not detect the spaABC cluster in all iso-
lates, but eight isolates were carrying the spaABC gene cluster.
Cluster 1 harboured the toxin and the spaABC gene cluster. The
presence of several described adhesive pili gene clusters within
genomic islands is summarized in Fig. 3.
In addition to pilins, iron acquisition is of high importance for
the bacteria and the expression of the diphtheria toxin is regulated
by the iron concentration and connected to the iron homeostasis of
the bacterial cell. Iron uptake can be facilitated with the help of
siderophores, which are high-affinity iron chelators. We found the
siderophore biosynthesis gene clusters sidBA and ciuEFG, and the
siderophore-dependent iron uptake system genes ciuABCD and
irp6ABC. Furthermore, we were able to detect in several isolates
irp2 (GEI26), another siderophore biosynthesis and transport gene
cluster and the additional iron transport systems located in GEI36,
GEI37 and GEI42 (Fig. 3). In brief, we found several virulence factors
within the C. diphtheriae isolates, which were, as expected, identical
within the isolates corresponding to the three detected potential
outbreak clusters. In addition, we found that the isolates in clusters
more often carried the diphtheria toxin gene compared with non-
cluster forming isolates.
Toxigenic prophages are specific for the isolates of each outbreak
The most important virulence factor of C. diphtheriae is the toxin
gene, which is responsible for causing pharyngeal diphtheria and
cell necrosis. As it is encoded on a prophage, and so potentially
subjected to horizontal gene transfer, we next analysed the toxi-
genic prophages of the isolates. We found that within a cluster the
toxigenic prophages were almost perfectly identical, indicating a
very close relationship among them, consistent with acquisition of
the toxigenic prophage before C. diphtheriae spread among the
refugees within each cluster (Fig. 4a). Comparison of prophages
among the two toxigenic clusters yielded similar prophages, which
also differed significantly from the other cluster. Most strikingly, we
1003.e6 D.M. Meinel et al. / Clinical Microbiology and Infection 22 (2016) 1003.e1e1003.e8

Fig. 3. Genomic island with virulence factors. The colours indicate virulence factors found according to Trost et al. [28] and Cerdeno-Tarraga et al. [27]. The deduced similarities are
colour-coded in the figure legend. The outbreak-like clusters 1, 2 and 3 are marked as well as the non-refugee isolates. The pathogenicity islands are grouped according to Trost et al.
in adhesive pili, siderophore and iron-uptake-associated operons/islands.

detected several mutations and an approximately 1-kb sequence
block not conserved between the prophages of clusters 1 and 2
(Fig. 4b). Together with the phylogeny, this clearly indicates that the
prophages were not exchanged between the isolates of the two
outbreaks or with the non-outbreak-associated isolates, as those
again showed a clearly different architecture, especially in the 50
region of the prophage of isolate D (Fig. 4b). Overall, the prophages
are highly similar to the earlier described btoxþ and utoxþ prophages
from C7(b)toxþ and PW8 C. diphtheriae strains, respectively [28]. In
summary, the prophages confirm the relationships derived from
the whole genome sequencing analysis, and suggest that the toxin
gene was integrated into the genome of the bacteria before they
spread.
Discussion
The current arrival of refugees from East Africa and the Middle
East to Europe poses enormous challenges for clinicians and clinical
microbiologists, as they may be faced with neglected infectious
diseases that are rarely diagnosed in Europe. For example, cases of
Borrelia recurrentis have recently been diagnosed in refugees from
East Africa [30e32]. In addition, some of these serious infectious
diseases have the potential to cause local outbreaks in refugee
centres [33e35] and so pose a considerable risk to public health.
We faced a potential outbreak of C. diphtheriae wound infections in
refugee centres and present our outbreak investigation using whole
genome sequencing in the context of the ongoing migration of
refugees from East Africa and Syria.
Taking into account the patients' vague recollection of their travel
dates, a poorly definable incubation period or recallable time of
symptom onset, as well as a possibly underlying co-colonization or
super-infection of the skin by C. diphtheriae, it is difficult to estimate
when infections with C. diphtheriae occurred. We found three clus-
ters, which showed high isolate similarity. However, the relatively
high numbers of 50e150 SNPs within a related cluster do indicate
that most likely no direct transmission took place, but that several

Fig. 4. Toxigenic prophages are cluster specific. (a) Phylogenetic tree based on the sequences of the predicted toxigenic prophages in the isolates. The prophage sequences of the
isolates are similar within each outbreak-like cluster, but differ between the clusters and the unrelated isolate D, which was derived from a traveller returning from Asia. The
phylogeny calculated with the neighbour-joining algorithm is consistent with the tree based on whole genomic single nucleotide polymorphisms (SNPs). (b) Prophage conservation
within the isolates. For each cluster in A one representative prophage is depicted. For the first cluster isolate 3 (upper panel), for the second isolate 19 (middle panel) and the
unrelated isolate D (lower panel) are shown. Sequences were aligned using Mauve and the height of the lines within the boxes indicates the conservation of the local sequences
within the three aligned prophages. Importantly, the toxigenic prophage of isolate D shows unique sequences in the 50 region (depicted in white, for no homology). Attachment sites
could be detected 50 of unique sequences. Further, the prophage in the cluster of isolate 3 has an approximately 1 kb large, distinct different sequence block (marked with the black
box) from the prophages of the second cluster and the prophage from isolate D. Annotations for the modular structure of the corynephages as present in the highly similar
Corynebacterium diphtheriae NCTC13129.
 D.M. Meinel et al. / Clinical Microbiology and Infection 22 (2016) 1003.e1e1003.e8
reservoirs might exist (at least three) on the routes of the refugees
from East Africa to Europe. Given that refugees reported having
arrived within days or maximally a few weeks of clinical presentation,
we hypothesize that these reservoirs are outside Europe. This would
be compatible with (i) isolated bacteria having already had time to
accumulate the moderate genomic mutations found within each of
the three clusters, and (ii) the observation that the date of isolation
did not correspond to the order of strains within each cluster (Fig. 1).
As a caveat to our hypotheses, we do not know exactly when the
colonization of the patients took place. However, as we have
country-specific clusters and mixed clusters we assume that the
infection took place en route, after refugees had left their home
countries. Individual routes and durations are likely to vary
extremely, but refugees are reported to spend prolonged periods
(weeks to months) in camps near the Sahara and in Northern Africa
(e.g. Libya) waiting to cross the Mediterranean Sea, with a high
likelihood that the three reservoirs represented by the observed
clusters may be located at key nodes en route to Europe.
For the reasons outlined above and as mutation rates of
C. diphtheriae are unknown, precise estimation of where and when
the infection started is difficult. However, we do not expect muta-
tion rates for Corynebacteria spp. higher than 1.82
105 sub-
stitutions/bp/year. This would again be compatible with the
hypothesis that C. diphtheriae isolates were acquired before trav-
elling. Crossing the Mediterranean Sea and travelling to
Switzerland or Germany is the fastest part and thought to be
completed within days. In line with this hypothesis, we previously
found that the closely related toxigenic C. ulcerans is only very
slowly manifesting SNPs after transmission events [14], as we
found only 0e2 SNPs within transmission pairs. Additionally, high
numbers of SNPs separating the unrelated strains from each other
were similar in previously described unrelated C. ulcerans isolates.
Our data set, including the time-tree analysis, supports the hy-
pothesis that the isolates within each cluster are highly related, but
that the single isolates had time to moderately diverge from their
common ancestor strain.
In contrast to more conventional typing techniques, such as
pulsed-field gel electrophoresis, ribotyping and multilocus
sequence typing [36e40], whole genome sequencing allows the
determination of relatedness with a maximum of resolution [41]
and, in addition, provides further genetic information on resis-
tance genes and virulence factors [42,43], such as the diphtheria
toxin gene and the spaABC cluster, which might be important for
risk assessment as the combination of both is more efficient in
causing pharyngeal diphtheria [29]. This technique may therefore
provide crucial information on relatedness of C. diphtheriae strains,
and inform outbreak management in situations where clusters of
cases are observed. Especially in the case of refugees, where
classical epidemiological data are often not reliable, available or
re-constructible, classical methods reach their reliability limits.
This could be overcome in our situation by using a whole genome
sequencing approach, offering more resolution and giving addi-
tional important information, for example, on uniformity of a
single outbreak cluster and on potential virulence factors. The
relatedness of strains based on whole genome sequencing data
could be determined and matched with the geographic origin of
the patients and also the close time-frame of occurrence, similar to
approaches already performed with MRSA [44,45]. Hence, physi-
cians in countries currently hosting refugees need to be highly
aware of the diverse types of C. diphtheriae infections in migrants,
such as acute or chronic ulcers or even pharyngitis, as on travel
routes several reservoirs, established by refugees from countries
with high diphtheria incidence, might exist. This issue is intensi-
fied by the problematic living conditions of the refugees on their
travel routes through endemic countries and also by the crowded
1003.e7
living conditions in refugee camps in European countries. Finally,
the fraction of vaccinated refugees is not known and potentially
very low, so more cases of C. diphtheriae wound infections are
expected in the future and should be monitored by public health
institutions.
Transparency declaration
None of the authors have a conflict of interest to declare.
Financial disclosure
AE was supported by a Swiss National Foundation research
grant (PZ00P3_154709, Ambizione). VB and TS were supported in
part by the European Research Council under the Seventh Frame-
work Programme of the European Commission (PhyPD: grant
agreement number 335529). The study was partly supported by the
Bavarian State Ministry of Health and Care (Project 13-30) as well as
by the German Federal Ministry of Health via the Robert Koch
Institute (09-47 and FKZ 1369-359).
Contributions of authors
AE, RK, DMM, AS and MB wrote the manuscript; AE, VH, DG, AB,
DMM, MD, TW, RZ, AB, RK and BB performed laboratory in-
vestigations; VB, TS and DMM performed computational in-
vestigations; AE, DMM, DG, AB, VH, STS, AB, RK, MB, AS, AFW, JAB,
UH and BB revised the manuscript; and CG, RK, DF, AD and UH
managed the patients.
Acknowledgements
We wish to thank Rosa-Maria Vesco, Magdalena Schneider and
Elisabeth Schultheiss, Clinical Microbiology University Hospital
Basel, as well as Wolfgang Schmidt, Sabine Wolf, Katja Meindl and
Jasmin Fra€ssdorf, LGL Bayern, for excellent technical assistance.
Appendix A. Supplementary data
Additional Supporting Information may be found in the online
version of this article can be found at http://dx.doi.org/10.1016/j.
cmi.2016.08.010.
References
[1] Iserson KV. Tackling the global challenge: humanitarian catastrophes. West J
Emergency Med 2014;15:231e40.
[2] Leaning J, Guha-Sapir D. Natural disasters, armed conflict, and public health.
N Engl J Med 2014;370:783e4.
[3] Nicolai T, Fuchs O, von Mutius E. Caring for the wave of refugees in Munich.
N Engl J Med 2015;373:1593e5.
[4] Jakovljev A, Steinbakk M, Mengshoel AT, et al. Imported toxigenic cutaneous
diphtheria in a young male returning from Mozambique to Norway, March
2014. Eurosurveillance 2014;19(24).
[5] Nelson TG, Mitchell CD, Sega-Hall GM, Porter RJ. Cutaneous ulcers in a returning
traveller: a rare case of imported diphtheria in the UK. Clin Exp Dermatol 2016
Jan;41(1). http://dx.doi.org/10.1111/ced.12763 [Epub 2015 Oct 12].
[6] Lowe CF, Bernard KA, Romney MG. Cutaneous diphtheria in the urban poor
population of Vancouver, British Columbia, Canada: a 10-year review. J Clin
Microbiol 2011;49:2664e6.
[7] Gubler J, Huber-Schneider C, Gruner E, Altwegg M. An outbreak of non-
toxigenic Corynebacterium diphtheriae infection: single bacterial clone
causing invasive infection among Swiss drug users. Clin Infect Dis 1998;27:
1295e8.
[8] Hamborsky J, Kroger A, Wolfe C. Diphtheria. In: Hamborsky J, Kroger A,
Wolfe C, editors. Centers for Disease Control and Prevention Epidemiology
and Prevention of Vaccine-Preventable DiseasesVol. 13. Washington D.C.:
Public Health Foundation; 2015.
[9] ECDC. Cutaneous diphtheria among recently arrived refugees and asylum
seekers in the EU. Available at: http://ecdc.europa.eu/en/publications/
Publications/. Accessed 20 July 2016.
1003.e8 D.M. Meinel et al. / Clinical Microbiology and Infection 22 (2016) 1003.e1e1003.e8
[10] Arnold JW, Koudelka GB. The Trojan Horse of the microbiological arms race:
phage-encoded toxins as a defence against eukaryotic predators. Environ
Microbiol 2014;16:454e66.
[11] Mokrousov I. Corynebacterium diphtheriae: genome diversity, population
structure and genotyping perspectives. Infect Genet Evol 2009;9:1e15.
[12] de Benoist AC, White JM, Efstratiou A, et al. Imported cutaneous diphtheria,
United Kingdom. Emerg Infect Dis 2004;10:511e3.
[13] Meinel DM, Konrad R, Berger A, et al. Zoonotic transmission of toxigenic
Corynebacterium ulcerans strain, Germany, 2012. Emerg Infect Dis 2015;21:
356e8.
[14] Meinel DM, Margos G, Konrad R, Krebs S, Blum H, Sing A. Next generation
sequencing analysis of nine Corynebacterium ulcerans isolates reveals zoonotic
transmission and a novel putative diphtheria toxin-encoding pathogenicity
island. Genome Med 2014;6:113.
[15] Konrad R, Hormansdorfer S, Sing A. Possible human-to-human transmission
of toxigenic Corynebacterium ulcerans. Clin Microbiol Infect 2015;21:768e71.
[16] Zakikhany K, Efstratiou A. Diphtheria in Europe: current problems and new
challenges. Future Microbiol 2012;7:595e607.
[17] Moore LS, Leslie A, Meltzer M, Sandison A, Efstratiou A, Sriskandan S. Cory-
nebacterium ulcerans cutaneous diphtheria. Lancet Infect Dis 2015;15:1100e7.
[18] Wagner KS, White JM, Crowcroft NS, De Martin S, Mann G, Efstratiou A.
Diphtheria in the United Kingdom, 1986e2008: the increasing role of Cory-
nebacterium ulcerans. Epidemiol Infect 2010;138:1519e30.
[19] Sing A. Zur Charakterisierung von C.-diphtheriae-verd€ achtigen Isolaten. RKI:
Epidemiol Bull 2008;3:23e5.
[20] Alibi S, Ferjani A, Gaillot O, Marzouk M, Courcol R, Boukadida J. Identification
of clinically relevant Corynebacterium strains by Api Coryne, MALDI-TOF-mass
spectrometry and molecular approaches. Pathol-biol 2015;63:153e7.
[21] Barberis C, Almuzara M, Join-Lambert O, Ramirez MS, Famiglietti A, Vay C.
Comparison of the Bruker MALDI-TOF mass spectrometry system and con-
ventional phenotypic methods for identification of Gram-positive rods. PLoS
One 2014;9(9):e106303.
[22] Konrad R, Berger A, Huber I, et al. Matrix-assisted laser desorption/ionisation
time-of-flight (MALDI-TOF) mass spectrometry as a tool for rapid diagnosis of
potentially toxigenic Corynebacterium species in the laboratory management
of diphtheria-associated bacteria. Eurosurveillance 2010;15(43).
[23] Schuhegger R, Lindermayer M, Kugler R, Heesemann J, Busch U, Sing A.
Detection of toxigenic Corynebacterium diphtheriae and Corynebacterium
ulcerans strains by a novel real-time PCR. J Clin Microbiol 2008;46:2822e3.
[24] Engler KH, Glushkevich T, Mazurova IK, George RC, Efstratiou A. A modified
Elek test for detection of toxigenic corynebacteria in the diagnostic laboratory.
J Clin Microbiol 1997;35:495e8.
[25] Goldenberger D, Hinic V, Turan S, et al. Extended characterization of Cory-
nebacterium pyruviciproducens based on clinical strains from Canada and
Switzerland. J Clin Microbiol 2014;52:3180e3.
[26] Hinic V, Lang C, Weisser M, Straub C, Frei R, Goldenberger D. Corynebacterium
tuberculostearicum: a potentially misidentified and multiresistant Corynebac-
terium species isolated from clinical specimens. J Clin Microbiol 2012;50:
2561e7.
[27] Cerdeno-Tarraga AM, Efstratiou A, Dover LG, et al. The complete genome
sequence and analysis of Corynebacterium diphtheriae NCTC13129. Nucleic
Acids Res 2003;31:6516e23.
[28] Trost E, Blom J, Soares Sde C, et al. Pangenomic study of Corynebacterium
diphtheriae that provides insights into the genomic diversity of pathogenic
isolates from cases of classical diphtheria, endocarditis, and pneumonia.
J Bacteriol 2012;194:3199e215.
[29] Mandlik A, Swierczynski A, Das A, Ton-That H. Corynebacterium diphtheriae
employs specific minor pilins to target human pharyngeal epithelial cells. Mol
Microbiol 2007;64:111e24.
[30] Goldenberger D, Claas GJ, Bloch-Infanger C, et al. Louse-borne relapsing fever
(Borrelia recurrentis) in an Eritrean refugee arriving in Switzerland, August
2015. Eurosurveillance 2015;20(32):2e5.
[31] Wilting KR, Stienstra Y, Sinha B, Braks M, Cornish D, Grundmann H. Louse-
borne relapsing fever (Borrelia recurrentis) in asylum seekers from Eritrea, the
Netherlands, July 2015. Eurosurveillance 2015;20(30).
[32] Hoch M, Wieser A, Loscher T, et al. Louse-borne relapsing fever (Borrelia
recurrentis) diagnosed in 15 refugees from northeast Africa: epidemiology and
preventive control measures, Bavaria, Germany, July to October 2015. Euro-
surveillance 2015;20(42).
[33] Challa AA. Tuberculosis Incidence in Immigrants and Refugees. Ann Intern
Med 2015;163:149e50.
[34] Lederer I, Taus K, Allerberger F, et al. Shigellosis in refugees, Austria, July to
November 2015. Eurosurveillance 2015;20(48).
[35] Ope M, Ochieng SB, Tabu C, Marano N. Rotavirus enteritis in Dadaab refugee
camps: implications for immunization programs in Kenya and Resettlement
Countries. Clin Infect Dis 2014;59:vevi.
[36] De Zoysa A, Hawkey P, Charlett A, Efstratiou A. Comparison of four molecular
typing methods for characterization of Corynebacterium diphtheriae and
determination of transcontinental spread of C. diphtheriae based on BstEII
rRNA gene profiles. J Clin Microbiol 2008;46:3626e35.
[37] Zasada AA, Forminska K, Wolkowicz T, Badell E, Guiso N. The utility of the PCR
melting profile technique for typing Corynebacterium diphtheriae isolates. Lett
Appl Microbiol 2014;59:292e8.
[38] Bolt F, Cassiday P, Tondella ML, et al. Multilocus sequence typing identifies
evidence for recombination and two distinct lineages of Corynebacterium
diphtheriae. J Clin Microbiol 2010;48:4177e85.
[39] Grimont PA, Grimont F, Efstratiou A, et al. International nomenclature for
Corynebacterium diphtheriae ribotypes. Res Microbiol 2004;155:162e6.
[40] Santos LS, Sant'anna LO, Ramos JN, et al. Diphtheria outbreak in Maranhao,
Brazil: microbiological, clinical and epidemiological aspects. Epidemiol Infect
2015;143:791e8.
[41] Sabat AJ, Budimir A, Nashev D, et al. Overview of molecular typing methods
for outbreak detection and epidemiological surveillance. Eurosurveillance
2013;18(4):20380.
[42] Eyre DW, Cule ML, Wilson DJ, et al. Diverse sources of C. difficile infection
identified on whole-genome sequencing. N Engl J Med 2013;369:
1195e205.
[43] Gardy JL, Johnston JC, Ho Sui SJ, et al. Whole-genome sequencing and
social-network analysis of a tuberculosis outbreak. N Engl J Med 2011;364:
730e9.
[44] Harris SR, Feil EJ, Holden MT, et al. Evolution of MRSA during hospital
transmission and intercontinental spread. Science 2010;327(5964):469e74.
[45] Reuter S, Torok ME, Holden MT, et al. Building a genomic framework for
prospective MRSA surveillance in the United Kingdom and the Republic of
Ireland. Genome Res 2016;26:263e70.
[46] Kennemann L, Didelot X, Aebischer T, et al. Helicobacter pylori genome
evolution during human infection. Proc Natl Acad Sci U S A 2011;108:
5033e8.