A predominant VNTR cluster of Mycobacterium tuberculosis isolates among. asylum seekers in the Netherlands and Denmark deciphered by whole genome

Transcription

1 JCM Accepted Manuscript Posted Online 22 November 2017 J. Clin. Microbiol. doi: /jcm Copyright 2017 Jajou et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license A predominant VNTR cluster of Mycobacterium tuberculosis isolates among asylum seekers in the Netherlands and Denmark deciphered by whole genome sequencing Rana Jajou 1#, Albert de Neeling 1, Erik Michael Rasmussen 2, Anders Norman 2, Arnout Mulder 1, Rianne van Hunen 1,4, Gerard de Vries 1,4, Walid Haddad 3, Richard Anthony 1, Troels Lillebaek 2, Wim van der Hoek 1, Dick van Soolingen 1, National Institute for Public Health and the Environment, Bilthoven, The Netherlands; 2 International Reference Laboratory of Mycobacteriology, Statens Serum Institut, Copenhagen, Denmark; 3 Municipal Health Service, GGD Hart voor Brabant, s-hertogenbosch, The Netherlands; 4 KNCV Tuberculosis Foundation, Den Haag, the Netherlands; 5 Radboud University Medical Center, Department of Medical Microbiology, Nijmegen, The Netherlands # Corresponding author: Rana Jajou, National Tuberculosis Reference Laboratory, Centre for Infectious Disease Control, National Institute for Public Health and the Environment (RIVM), P.O. Box 1, 3720 BA Bilthoven, The Netherlands. T: rana.jajou@rivm.nl

2 Abstract Background: In many countries, Mycobacterium tuberculosis (Mtb) isolates are routinely subjected to VNTR typing to investigate Mtb transmission. Unexpectedly, cross border clusters have been identified among African refugees in the Netherlands and Denmark, although transmission in these countries was unlikely. Whole genome sequencing (WGS) was applied to analyse in depth transmission and assess the precision of VNTR typing. Methods: WGS was applied to 40 Mtb isolates from refugees in the Netherlands and Denmark, mostly from the Horn of Africa, sharing the exact same VNTR profile. Cluster investigations were undertaken to identify in-country epidemiological links. Results: Combining WGS results of isolates, all members of the CAS/Delhi genotype, from both European countries, an average genetic distance of 80 SNPs (maximum 153 SNPs) was observed. Conversely, the few pairs of isolates with confirmed epidemiological links, except for one pair, had a maximum of 12 SNPs distance. WGS divided this refugee cluster into several sub-clusters of patients from the same country of origin. Conclusion: Although the Mtb cases, mainly originating from African countries, shared the exact same VNTR profile, most were clearly distinguished by WGS. The average genetic distance in this specific VNTR cluster was two times greater than in other VNTR clusters. Thus, identical VNTR profiles did not represent recent and direct Mtb transmission for this group of patients. It appears that either these strains from Africa are genetically extremely conserved or there is ongoing transmission of this genotype among refugees on their long migration routes from Africa to Europe. 39

3 40 41 Keywords: Mycobacterium tuberculosis, asylum seekers, The Netherlands, Denmark, VNTR/ WGS Introduction Tuberculosis (TB) is caused by Mycobacterium tuberculosis (Mtb) and is the leading cause of death worldwide from a single pathogen [1]. The Netherlands is a low incidence country, between 1994 and 2014 the incidence of TB declined from 11.7 to 4.9 cases per inhabitants. However, due to the increasing number of immigrants and asylum seekers from high TB incidence countries, in 2015 the TB incidence increased to 5.1 cases per inhabitants. Of all 867 patients diagnosed in 2015, 625 (72.1%) were born outside the Netherlands, with the majority of TB patients originating from Eritrea, Somalia, and Morocco [2]. The increase in immigrants from higher TB incidence countries is also an important issue in other low incidence European countries. In the Nordic countries for instance, a considerable influx of immigrants has been reported that has contributed significantly to the TB incidence in these countries [3-6]. In Denmark for example, in 2015 the overall TB incidence was 6.3 cases per inhabitants of which 75% was of non-danish origin, 19% of these cases were from the Horn of Africa [7]. DNA fingerprinting of all Mtb isolates has been structurally applied in the Netherlands since 1993 and in Denmark since 1992 to monitor TB transmission and is performed by the (inter) National Tuberculosis Reference Laboratories in both countries. IS6110 restriction fragment length polymorphism (RFLP) typing was the first gold standard in typing and was applied from 1993 to 2008 in the Netherlands and from 1992 to 2006 in Denmark. In 2009, 24-loci variable

4 number tandem repeat (VNTR) typing was introduced in the Netherlands as the new standard, in line with international developments [8], with a retrospective re-typing of all isolates from 2004 to For the isolates typed by both RFLP and VNTR typing, the concordance between both methods was described earlier [9]. In Denmark the switch to VNTR typing took place in 2007, with retrospective re-typing of the isolates from 2003 to The Municipal Health Services (MHSs) in the Netherlands receive information on clustering of DNA fingerprints of Mtb isolates on a weekly basis to guide investigations of epidemiological relations and transmission patterns. In Denmark, investigations based on epidemiological links are not as systematically registered and only limited linkage information is available. Already in 1993, the RFLP pattern associated with cluster number 15 was detected in the northern part of the Netherlands and concerned a 20-year-old patient from Somalia who was assumed to suffer from an endogenous reactivation of a latent infection contracted before arriving in the Netherlands. Over the years, this cluster expanded with at least two new cases per year throughout the Netherlands. The change in typing methodology transformed RFLP cluster 15 into VNTR cluster , which has also been identified in Denmark. In the Netherlands, 45 isolates from the period were assigned to cluster Full concordance between the two DNA fingerprinting methods was found in 14/45 (31%) isolates, meaning these isolates had exactly the most prevalent profile of RFLP cluster 15 and VNTR cluster Another 14/45 (31%) isolates, however, presented with a slightly different VNTR, but the same RFLP pattern 15. The remaining 17/45 (38%) isolates revealed the most common VNTR profile , but slightly different RFLP patterns.

5 As both in the Netherlands and Denmark an unexpected high degree of VNTR clustering among refugees and immigrants occurred, WGS was applied to isolates belonging to the asylum seekers VNTR cluster from both countries. The mainly non-infectious character of the Mtb in asylum seekers from several countries and the low degree of confirmed epidemiological links makes transmission in the Netherlands unlikely, and brings into question the specificity of clustering on the basis of RFLP/VNTR in this specific cluster. Based on two previous studies [5, 10], a similar conclusion was made for the majority of cases in the refugee cluster in Denmark. This suggests that 1) transmission occurred on the migration route, or 2) there is a predominant genotype of Mtb with a high degree of genetic conservation circulating in the countries of origin. In this study, the high resolution of WGS and detailed information on patient characteristics was combined to disclose the confusion in the molecular epidemiology of TB in refugees. If a part of the clustering on basis of VNTR typing in this group of patients is resolved in WGS analysis, recent transmission may be overestimated Methods Patient selection Using the most common VNTR profile of cluster RFLP 15/VNTR (24-loci VNTR pattern in the order of loci VNTR154 - VNTR424 - VNTR577 - VNTR VNTR802 - VNTR960 - VNTR VNTR VNTR VNTR2163b - VNTR VNTR VNTR VNTR VNTR VNTR VNTR VNTR VNTR VNTR VNTR VNTR VNTR VNTR4348: ), twelve of the 205 patient isolates from the Netherlands were selected to obtain variation in country of birth, duration of

6 living in the Netherlands (< 6 versus 6 months), multi-drug resistant TB (MDR-TB) diagnosis, and in epidemiological linkage with another patient in the cluster. Patient characteristics were obtained from the Netherlands Tuberculosis Register (NTR) and information on epidemiological links was obtained from the respective MHS. From Denmark, isolates from 31 patients shared the VNTR profile and were all subjected to WGS. The samples of these patients were collected in the period from 2001 through DNA purification The Mtb isolates from the Netherlands were re-grown in Mycobacteria Growth Indicator Tube (MGIT) tubes from the -70 C freezer stocks, except for the isolates from 2016 that were directly available. DNA was isolated and purified with the QIAamp DNA mini kit method (QIAGEN GmbH, Hilden, Germany). Prior to this, the samples were incubated with lysis buffer and proteinase K to release the DNA [11]. Mtb isolates from Denmark were re-grown in MGIT tubes from the -80 C freezer stocks. One ml was taken from a positive MGIT and centrifuged. DNA was isolated according to a previous study [12] Read mapping and variant calling The isolates from the Netherlands were sequenced on an Illumina HiSEQ2500 sequencer and the isolates from Denmark on an Illumina NextSeq following the Nextera XT protocol according to manufacturer s instructions. Breseq software version was used for mapping the reads

7 to the reference genome H37Rv version 3.0 (GenBank Accession AL or Refseq Accession NC_00962) and for calling SNPs supported by 80% of the mapped reads in regions of minimum five reads coverage. Fastq-files of the isolates from the Netherlands were uploaded to PhyResSe ( for assigning a lineage to each isolate. For the isolates from Denmark, the SNP-barcode system proposed by Coll et al [13] was used to assign the isolates to a lineage. This method detected the G/A SNP at position 1,084,911 in all isolates, which identifies them as lineage 3.1.1, i.e. CAS-Delhi. Called variants were compared to a set of global reference strains belonging to the same lineage (ENA accession no.: ERP001731) Data analysis WGS data analysis was performed in R statistics version 3.2. SNPs annotated as PE/PPE, PGRS, esx, repeat, polyketide, pks, or transposase were excluded during WGS data analysis. Squared Euclidean distance matrices were generated to analyse the genetic distance in SNPs between isolates. A minimum spanning tree was built using Ridom SeqSphere+ software version 4.0 (Ridom GmbH, Münster, Germany) to visualize how isolates from the Netherlands and Denmark were clustered according to WGS using a cut-off of maximum 12 SNPs [14]. All isolates from the Netherlands cultured between January 1, 2016 and December 31, 2016 were subjected to both VNTR typing and WGS. Therefore, it was possible to compare the average genetic distance between isolates in cluster with isolates belonging to other VNTR clusters. The phenotypic resistance profile of the MDR-TB patients was verified on the basis of detected

8 mutations in resistance genes using the PhyResSe pipeline. Phenotypic drug susceptibility testing (DST) was performed on a MGIT 960 system Results Description of the clusters Cluster in the Netherlands comprised of 205 patients in 2016, of which 156 (76.1%) were born in Somalia, 29 (14.1%) in Eritrea /Ethiopia, eight (3.9%) in the Netherlands, two (1%) in Sudan, two (1%) in Burundi, two (1%) in Surinam, one (0.5%) in the Netherlands Antilles, one (0.5%) in Pakistan, one (0.5%) in Zambia, one (0.5%) in Uganda, one (0.5%) in Yemen, and one (0.5%) patient whose origin was not registered (Figure 1). Among those 205 patients, 102 were asylum seekers of which 63.7% (65/102) of them had TB diagnosed within 2.5 years after arriving in the Netherlands. Epidemiological links were confirmed for only 3.9% (8/205) of the cases, two cases were identified by RFLP typing and six by VNTR typing. Fifty-four percent (111/205) of the patients were males. Seventy (34.1%) patients had pulmonary tuberculosis (PTB) only, 33 (16.1%) had a combination of PTB/extra-pulmonary tuberculosis (EPTB), and 102 (49.8%) had EPTB only. Data on DST for first line antibiotics showed that 66.8% (137/205) of the isolates were fully sensitive, fifteen (7.3%) were mono-resistant to isoniazid (INH), four (2%) MDR-TB (i.e. INH and rifampicin [RIF] resistant), two (1%) were mono-resistant to RIF, and for the remaining 47 (22.9%) isolates DST results were missing in the NTR Twelve of the 205 isolates from patients in the Netherlands were included in this study consisting of eight patients from Somalia, three from Eritrea (two of them were

9 epidemiologically linked), and one patient from the Netherlands. Four Somalian patients from 2016 were from one family of which the source case, NL16-10, was suspected to have infected the three family members, NL16-09, NL16-11, and NL16-12 based on MHS data (Table 1). Of the twelve included isolates, two were MDR-TB, two streptomycin (STR) resistant, four susceptible, and for the remaining four isolates DST data was not available in the NTR A cluster with the same VNTR pattern in Denmark comprised of 31 cases, of which three cases were excluded due to low coverage of the sequenced isolates, leaving 28 strains available for inclusion in this study (Table 1). Of these 28 cases, 15 (53.6%) originated from Somalia, nine (32.1%) from Eritrea, two (7.1%) from Denmark, one (3.6%) from Uganda, and one (3.6%) case was of unknown country of origin (Figure 2). One of the two Danish-born cases had Somalian parents. Sixty-four percent (18/28) of the patients were males. Ten (35.7%) patients had PTB only, two (7.1%) had a combination of PTB/EPTB, and 16 (57.1%) patients had EPTB only. Isolates from 26 (92.9%) patients were fully sensitive to all first line antibiotics and the remaining two were RIF resistant, of which one was a MDR-TB isolate with additional resistance to EMB, pyrazinamide (PZA), roxithromycin (ROX), cycloserin (CYA), and capreomycin (CPM) WGS clustering of asylum seekers in the Netherlands and Denmark The genetic distance between the isolates from the Netherlands and Denmark, all belonging to the CAS/Delhi genotype, ranged from SNPs; the degree of genetic conservation was comparable among the isolates from the Netherlands as for the isolates from Denmark. The

10 average genetic distance between all 40 isolates in cluster was 80 SNPs compared to 43 SNPs between isolates in other VNTR clusters in the Netherlands. WGS divided the isolates from cluster into several sub-clusters, where all sub-clusters consisted of patients from the same country of origin, except for sub-cluster D, where a native Danish case, DK15-25, clustered with two patients from Eritrea, DK15-04 and DK16-12 (Figure 3). The six isolates from patients in the Netherlands with a confirmed epidemiological link had a genetic distance of 12 SNPs and formed 2/5 WGS sub-clusters (Figure 3: sub-cluster A and B). The only identified epidemiological link for the cases from Denmark was between DK06-29, father, and DK06-14, son, with a genetic distance of 40 SNPs (Figure 3). The genetic distance between isolates from patients without a confirmed epidemiological link ranged from SNPs, meaning some patients for whom epidemiological links could not be identified were clustered on basis of WGS (Figure 3: sub-cluster C, D, and E). Five isolates from Denmark, DK14-07, DK07-16, DK05-18, DK04-20, and DK14-26, had much longer genetic distances than almost all other isolates in the cluster. DK14-26, a Danish born patient with Somalian parents, and DK04-20 who was Somalian, had a distance of 57 SNPs from each other, but varied by 105 SNPs from all other isolates. A similar situation was observed for DK14-07, originally from Somalia, and DK07-16, the only patient originally from Uganda who were most related to each other, but 90 SNPs distant from all other isolates studied. DK05-18 was 100 SNPs distant from all other included isolates.

11 The genetic distance between the three MDR-TB isolates was maximum 59 SNPs, where NL11-03 from Eritrean origin was more related to the Danish case from Somalian origin than to the MDR-TB case in the Netherlands from Somalian origin (Figure 3) Resistance profiles of MDR-TB patients The two MDR-TB isolates from the Netherlands were both resistant to INH, RIF, and EMB. NL11-03 was additionally resistant to STR and PZA. Mutations in associated resistance genes were identified for all drugs except STR. The MDR-TB isolate from Denmark was resistant to RIF, INH, PZA, EMB, ROX, CYA, and CPM. This isolate had mutations in the resistance-associated rpob S450L, katg S315T, pnca W119R, pnca A-11G, and embb M497R genes (Table 2) Discussion This study showed that a large cluster of Mtb isolates, mostly from asylum seekers, shared the same VNTR profile, but were not identical based on WGS analysis. WGS divided this VNTR cluster into several sub-clusters, where almost all sub-clusters included patients from the same country of origin. Moreover, the average genetic distance in this specific VNTR cluster was twice as large as compared to other VNTR clusters. This supports the findings from epidemiological investigation that transmission did not occur in the destination countries. Refugees originate from countries with wars and hunger/financial crises resulting in sustained migration to European countries. The respective patients share migration routes to Europe including long-term stays in large refugee camps where TB transmission most likely occurred. 231

12 This hypothesis is supported by a recent multi-country cluster investigation on 28 migrant MDR- TB cases in Europe, in which it is suggested that cases were part of a chain of recent transmission that has likely occurred in the country of origin or during the migration route [15]. A similar study was performed in Switzerland, a low TB incidence country like the Netherlands. This study re-analysed VNTR clusters using WGS to assess TB transmission among native and foreign-born TB cases. The outcome of this study was comparable to our results; an overestimation of TB transmission based on VNTR typing compared to WGS, especially among immigrants from high TB incidence countries [16]. Several other findings from this study support our hypotheses. In the Netherlands, 19% of the overall TB cases that were clustered by VNTR typing within a two year interval were found to be epidemiologically linked by MHSs in 2014 [2]. However, in the VNTR cluster described in this study TB nurses interviewing the patients could hardly confirm any epidemiological links between cases. The low degree of confirmed epidemiological links identified by MHSs seems to be logical, as in the majority of cases, only non-transmissible EPTB was observed. Also, the average genetic distance between the isolates in cluster was much higher than in other clusters in the Netherlands, in which the degree of confirmation of epidemiological links is in general higher [17]. Therefore, the benefits of WGS for epidemiological research among asylum seekers is even higher than for regular patients in these low prevalence settings. Screening programs in the Netherlands targeting asylum seekers/immigrants from high TB incidence countries, aim to detect TB disease in these populations. However, these screening programs only focus on active disease and not on migrants latently infected. Also in Denmark, asylum seekers are offered health interviews, and in case of symptoms, additional testing can

13 be performed. It is assumed that most asylum seekers who develop disease after arrival are experiencing an endogenous reactivation rather than a new infection due to recent transmission after arriving in the destination country [3, 18]. Nevertheless, the possibility of transmission in the destination country remains. In WGS sub-cluster D for instance, a native Danish patient clustered with patients from Eritrean origin. No information was available about the potential source of infection of this native Danish patient. Although previous studies suggested that transmission of Mtb from immigrants to the native population is limited [5, 10, 19] [20], it is likely that such transmission did occur, since TB was diagnosed in 2015/2016 for all involved patients. Also among the five native Dutch patients in cluster , for two an epidemiological link was confirmed with a patient from Zambia, and one patient had worked in Africa. For the remaining two native Dutch patients an epidemiological link with another patient in the cluster could not be found. Identifying epidemiological links among immigrants remains a challenge due to cultural/language barriers or there is a reluctance to reveal families/friends real identity or actual travel history. In the six pairs of isolates from patients in the Netherlands that had confirmed epidemiological links, the genetic distance was limited to 12 SNPs. In line with the criteria described by Walker et al [14] to rule in a possible epidemiological link, these cases were correctly clustered by VNTR typing. However, the two epidemiologically linked cases in Denmark were 40 SNPs distant from each other. The pace of genetic turnover may be for unknown reasons higher in this strain. For the other non-linked cases in both countries, a genetic distance was found of up to only 153 SNPs. This is highly remarkable, as this suggests an extremely high degree of genetic conservation in this predominant genotype of Mtb circulating in such a large geographic area.

14 One way to investigate the reason for this high degree of conservation would be to sample Mtb in a representative way from TB patients in both Somalia and Eritrea and possibly other countries, to study the actual population structure of Mtb and the degree of genetic conservation within the countries of origin. Giving the low discriminative power of the current DNA typing methods in Western countries when analysing cases from refugees, this seems a sensible investment that may be supported by larger health organizations such as the WHO or the ECDC. In conclusion, VNTR typing of Mtb isolates from African asylum seekers is misleading. When WGS was applied, several sub-clusters were formed consisting of mainly patients from the same country of origin. Despite the fact that the average genetic distance within this cluster was twice as high compared to other VNTR clusters, the degree of genetic conservation within this diverse study population remains remarkable, suggesting these cases are linked and transmission occurred during the migration route or a highly conserved genotype is circulating in the refugees countries of origin Acknowledgement We would like to thank laboratory technician Pia Kristiansen from the International Reference Laboratory of Mycobacteriology, Statens Serum Institut, Copenhagen, Denmark for the excellent laboratory work in Denmark. The Danish part of the study was supported by the Novo Nordisk Foundation (grant number NNF7651) and the Lundbeck Foundation (grant number R ). The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

15 References 1. World Health Organization Global tuberculosis report Slump E, Erkens CGM, van Hunen R, Schimmel HJ, van Soolingen D, de Vries G Tuberculose in Nederland Retrieved from f175a9d18d98&type=pdf&disposition=inline. 3. Herrador BRG, Rønnig K, Borgen K, Mannsåker T, Dahle UR Description of the largest cluster of tuberculosis notified in Norway : is the Norwegian tuberculosis control programme serving its purpose for high risk groups? BMC Public Health. 15: Søborg B EPIET report - Summary of work activities Lillebaek T, Andersen AB, Bauer J, Dirksen A, Glismann S, de Haas P, Kok-Jensen A Risk of Mycobacterium tuberculosis transmission in a low-incidence country due to immigration from high-incidence areas. J Clin Microbiol. 39(3): Lillebaek T, Andersen AB, Dirksen, A. Smith, E. and L.T.K.-J. Skovgaard, A, Kok-Jensen A Persistent high incidence of tb in immigrants in a low-incidence country. Emerg Infect Dis. 8(7): Statens Serum Institut Tuberculosis Retrieved from 8. Supply P, Allix C, Lesjean S, Cardoso-Oelemann M, Rüsch-Gerdes S, Willery E, Savine E, de Haas P, van Deutekom H, Roring R, Bifani P, Kurepina N, Kreiswirth B, Sola C, Rastogi N, Vatin V, Gutierrez MC, Fauville M, Niemann S, Skuce R, Kremer K, Locht C, van

16 Soolingen D Proposal for standardization of optimized mycobacterial interspersed repetitive unit-variable-number tandem repeat typing of Mycobacterium tuberculosis. J Clin Microbiol. 44(12): de Beer JL, van Ingen J, de Vries G, Erkens C, Sebek M, Mulder A, Sloot R, van den Brandt AM, Enaimi M, Kremer K, Supply P, van Soolingen D Comparative study of IS6110 restriction fragment length polymorphism and variable-number tandem-repeat typing of Mycobacterium tuberculosis isolates in the Netherlands, based on a 5-year nationwide survey. J Clin Microbiol. 51(4): Kamper-Jørgensen Z, Andersen AB, Kok-Jensen A, Kamper-Jørgensen M, Bygbjerg C, Andersen PH, Thomsen VØ, Lillebaek T Migrant tuberculosis: the extent of transmission in a low burden country. BMC Infect Dis. 12: Qiagen QIAamp DNA mini and blood mini handbook Bjorn-Mortensen K, Zallet J, Lillebaek T, Andersen AB, Niemann S, Rasmussen EM, Kohl TA Direct DNA Extraction from Mycobacterium tuberculosis Frozen Stocks as a Reculture-Independent Approach to Whole-Genome Sequencing. J Clin Microbiol. 53(8): Coll F, McNerney R, Guerra-Assun JA, Glynn JR, Perdigão J, Viveiros M, Portugal I, Pain A, Martin N, Clark TG A robust SNP barcode for typing Mycobacterium tuberculosis complex strains. Nat Commun. 5: Walker TM, Camilla LC, Harrell RH, Evans JT, Kapatai G, Dedicoat, MJ, Eyre, DW, Wilson DJ, Hawkey PM, Crook DW, Parkhill J, Harris D, Walker S, Bowden R, Philip M, Smith EG,

17 Petoa TEA Whole-genome sequencing to delineate Mycobacterium tuberculosis outbreaks: a retrospective observational study. Lancet Infect Dis. 13(2): European Centre for Disease Prevention and Control Multidrug-resistant tuberculosis in migrants, multi-country cluster 12 April Stucki D, Ballif M, Egger M, Furrer H, Altpeter E, Battegay M, Droz S, Bruderer T, Coscolla M, Borrell S, Zürcher K, Janssens JP, Calmy A, Stalder JM, Jaton K, Rieder HL, Pfyffer GE, Siegrist HH, Hoffmann M, Fehr J, Dolina M, Frei R, Schrenzel J, Böttger EC, Gagneux S, Fenner L Standard Genotyping Overestimates Transmission of Mycobacterium tuberculosis among Immigrants in a Low-Incidence Country. J Clin Microbiol. 54(7): Sloot R, Borgdorff MW, de Beer J, van Ingen J, Supply P, van Soolingen D Clustering of tuberculosis cases based on variable-number tandem-repeat typing in relation to the population structure of Mycobacterium tuberculosis in the Netherlands. J Clin Microbiol. 51(7): Pareek M, Greenaway C, Noori T, Munoz J, Zenner D The impact of migration on tuberculosis epidemiology and control in high-income countries: a review. BMC Med. 14: Dahle UR, Eldholm V, Winje BA, Mannsåker T, Heldal, E Impact of immigration on the molecular epidemiology of Mycobacterium tuberculosis in a low-incidence country. Am J Respir Crit Care Med. 176(9):

18 Sandgren A, Schepisi MS, Sotgiu G, Huitric E, Migliori GB, Manissero D, van der Werf M, Girardi E Tuberculosis transmission between foreign- and native-born populations in the EU/EEA: a systematic review. Eur Respir.43(4): Figure 1 Cluster RFLP 15 / VNTR in the Netherlands according to origin Figure 2 Cluster VNTR in Denmark according to origin Figure 3 Minimum spanning tree of 40 M. tuberculosis isolates from asylum seekers in the Netherlands (orange) and Denmark (blue) belonging to the same VNTR cluster. No unknown epidemiological links were allowed. The numbers on the branches indicate the genetic distance in SNPs to the nearest isolate. SNP differences between distant strains cannot be reconstituted by summing the number of SNPs on the branches. The red dotted line represents cases of Somalian origin and the red solid line represents cases of Eritrean origin; the first two digits in the case number indicate the year of TB diagnosis. Sub-clustering on the basis of WGS ( 12 SNPs distance between isolates) is shown in grey. 377

19 Somalia Eritrea/Ethiopia Netherlands Sudan Burundi Surinam Netherlands Antilles Pakistan Zambia Uganda Yemen n.a Somalia Eritrea Uganda Denmark n.a.

20 382

21 Table 1 Characteristics of the 40 included M. tuberculosis patients from the Netherlands and Denmark Patient Gender Age at Year of Country of Epidemiological link with Comments clinical diagnosis birth sample taken NL05-01 M Somalia - 6 months in the Netherlands NL06-02 M Somalia - < 6 months in the Netherlands; RFLP 276 NL11-03 M Eritrea - MDR NL13-04 F Somalia - MDR NL14-05 M Eritrea NL NL14-06 M Eritrea NL NL15-07 M Netherlands - - NL15-08 F Somalia - < 6 months in the Netherlands NL16-09 M Somalia NL NL16-10 F Somalia NL16-09,NL16-11,NL NL16-11 M Somalia NL NL16-12 F Somalia NL

22 DK15-01 M Unknown - - DK16-03 M Eritrea - - DK15-04 M Eritrea - - DK01-05 F Somalia - MDR DK16-06 M Eritrea - - DK14-07 M Somalia - Rifampicin monoresistant DK14-08 F Somalia - - DK03-09 M Somalia - - DK05-11 M Somalia - - DK16-12 M Eritrea - - DK06-14 M Somalia DK DK06-15 F Somalia - - DK07-16 F Uganda - - DK07-17 F Somalia - - DK05-18 F Somalia - - DK05-19 M Eritrea - - DK04-20 M Somalia - - DK09-21 F Somalia - - DK16-22 F Eritrea - - DK15-23 M Somalia - -

23 DK16-24 F Eritrea - - DK15-25 M Denmark - - DK14-26 M Denmark - Somalian-born parents DK16-27 F Eritrea - - DK06-29 M Somalia DK DK06-30 M Somalia - - DK13-31 M Somalia - - DK15-32 M Eritrea Table 2 Comparison of phenotypic drug susceptibility testing and WGS of the three MDR patients RIF INH PZA EMB STR Patient DST rpob varia DST katg vari DST pnca vari DST embb vari DST rpsl/ nt ant ant ant gidb/rr s/tlya NL11-03 R S450L C R S315T G R W119R T R Q497R A R S T C C G NL13-04 R S450L C R S315T G S S S R M306I G S S T C T DK01-05 R S450L C R S315T G R A-11G A R Q497R A S S T C G G

24 RIF=Rifampicin; INH=Isoniazid; PZA=Pyrazinamide; EMB=Ethambutol; STR=Streptomycin; DST=Drug susceptibility testing; R=Resistant; S=Susceptible