Tuberculosis transmission and resistance among Ukrainian migrants and its impact on multidrug-resistant tuberculosis dynamics in high-influx host countries

Abstract

Background

The war in Ukraine triggered unprecedented migration into Central Europe, raising concerns regarding tuberculosis (TB) spread.

Methods

We conducted a molecular-epidemiological study including all Ukrainian migrants with culture-confirmed TB in Slovakia and the Czech Republic between September 2021 and December 2024 (n = 229), together with all other multidrug-resistant (MDR) TB cases reported in these countries between 2023 and 2024 (n = 28). Whole-genome sequencing (WGS) was performed on cultured Mycobacterium tuberculosis complex (MTBC) isolates, and the resulting data were used for phylogenetic reconstruction, lineage classification, genomic resistance prediction, and cluster analysis.

Findings

Since the war began in 2022, the share of TB cases from Ukrainian migrants in the Czech Republic rose significantly (P < 0.001), along with an evident increase in total MDR-TB cases. Out of 199 MTBC strains from Ukrainian migrants with good-quality WGS data, 129 (64.8%) were predicted to be susceptible, 25 MDR (12.6%), 8 pre-extensively drug-resistant (pre-XDR; 4%), and 37 (18.6%) had other resistance patterns including bedaquiline/clofazimine monoresistance. Among MDR/pre-XDR strains (n = 59) lineage 2 predominated (63.5%), mainly belonging to the Central Asia clades (45.0%) and the Europe/Russian W148 outbreak clone (52.9%), followed by lineage 4 (28.9%). Cluster analysis (5 allele threshold) identified 10 MDR/pre-XDR TB clusters, each with two isolates. Only one of those clusters included a Ukrainian migrant and non-Ukrainian patient from the host country.

Conclusion

Migration increased both, the TB and MDR-TB burden especially in the Czech Republic. Transmission of imported MTBC strains from migrants to host populations appears to be limited, but needs to be closely monitered prospectively.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12866-026-04874-3.

Introduction

Following Russia’s full-scale invasion of Ukraine on 24 February 2022, millions of people were displaced. By 31 December 2024, nearly 4.3 million people fleeing Ukraine had temporary protection status in the European Union [1].​ European countries, particularly those close to Ukraine – including Slovakia, Poland and the Czech Republic – reported a significant influx of migrants [2]. As of 31 December 2024, Czech Republic ranked as the third-largest host country in the European Union, accommodating 388,625 beneficiaries of temporary protection from Ukraine (35.7 per 1,000 population; approximately 3.6%). Slovakia hosted 131,525 beneficiaries (24.2 per 1,000; approximately 2.4%), placing it among the EU member states with the highest relative proportion of Ukrainian migrants in the total population [3]. This growing influx of war refugees has been accompanied by a marked rise in rifampicin-resistant (RR) and multidrug-resistant tuberculosis (MDR-TB) cases in several countries, with the number of MDR-TB cases doubling in 2022 [4, 5]. In the context of a deteriorating TB situation, transmission dynamics within the migrant group and between migrants and host populations have not yet been adequately explored [6].

According to the WHO report, Ukraine ranks among the countries with the highest incidence of TB in Europe with 112 cases per 100,000 population [7]. Especially the emergence of MDR-TB strains represents a major public health threat, with serious implications for disease control and treatment effectiveness. In Ukraine, MDR-TB accounts for 29% of new cases and 43% of previously treated cases [8]. A recent study demonstrated a renewed increase in MDR-TB strains across various regions of Ukraine, associated with population displacement within the country [9]. As civilians relocate from regions with higher prevalence of MDR-TB to areas with lower prevalence, the risk of transmission may increase. Similar dynamics may be observed in transmission between Ukrainian migrants and host country populations [10].

The high incidence of MDR-TB in Ukraine has been strongly associated with Mycobacterium tuberculosis complex (MTBC) strains of the Beijing lineage (lineage 2) [8]. Compared with strains of other MTBC lineages, Beijing strains have been associated with higher transmissibility combined with higher rates of acquired resistance, potentially facilitating cross-border spread of MDR-TB in the context of migration [11]. Clustering of closely related MDR-TB strains further indicates ongoing transmission and highlights the phenomenon of transmitted drug resistance [8, 11]. The successful global spread of the Beijing lineage has been linked to several historical events, including Chinese migration to North and South America, as well as to the Russian Empire following periods of national unrest [12]. Multiple investigations have identified genetic characteristics associated with the enhanced pathogenicity and resistance of the Beijing lineage [13, 14]. Mutations in mycobacterial DNA repair and replication genes have been implicated in increased resistance rates, the development of MDR, and potentially treatment failure in patients infected with the Beijing lineage of MTBC [15].

While the high drug resistance rates in Ukraine in combination with war and migration are alarming, in depth studies tracing the transmission of particular MDR-MTBC strains to Western European low incidence settings are scarce. Accordingly, this study addressed three research questions: (i) How did TB notifications and the contribution of Ukrainian citizens to the notified TB caseload evolve in Slovakia and the Czech Republic before and after 2022? (ii) What are the phylogenetic lineages and drug-resistance profiles (including resistance to new/repurposed drugs) among MTBC isolates from Ukrainian TB patients in both countries? and (iii) Is there genomic evidence of recent transmission within Ukrainian migrants and, in particular, of MDR/pre-XDR transmission between Ukrainian migrants and the host populations?

Materials and methods

Data sources and study design

Aggregated epidemiological data were obtained from the national TB registers of the Czech Republic (2004–2024) and Slovakia (2007–2024) [16, 17]. The datasets encompassed all notified (both culture positive and negative) TB paediatric (< 15 years) and adult cases and included annual case counts, demographic characteristics, and available phenotypic drug-resistance profiles. Additional patient-level information, not routinely captured in the national registers, was acquired through collaboration with the National TB Surveillance Unit in the respective country.

The overall study design is based on preliminary research that analyzed TB cases in patients with Ukrainian citizenship in the early stages after the outbreak of the war in Ukraine [6]. The current study extends this framework by incorporating a longer observation period (9/2021-12/2024), additional epidemiological indicators, and by assessing potential transmission events involving MDR-TB between Ukrainian migrants and local populations in both countries.

For isolate-based molecular analyses, the analytic cohort comprised all TB patients with Ukrainian citizenship for whom an MTBC isolate was available. Patient-level demographic and clinical variables were compiled only for culture-confirmed cases to maintain consistent denominators across the isolate-based molecular analyses. If multiple MTBC isolates were obtained from the same patient during the study period (September 2021–December 2024), only the earliest isolate was included for WGS-based analyses to avoid duplicate sampling; subsequent isolates from the same patient were not treated as independent cases. Species identification within the M. tuberculosis complex was performed according to standard diagnostic procedures used in each country at the time of diagnosis. For each patient, a standard set of routinely collected epidemiological and clinical variables was extracted, including sex, age, form of TB (pulmonary or extrapulmonary), the administrative region of diagnosis (available only for patients diagnosed in the Czech Republic), treatment history (new case or retreatment, defined as a documented previous TB episode prior to the current notification), death during TB treatment, HIV status, and the method of case detection (symptom-based investigation, screening of high-risk groups, incidental finding, contact tracing, or post-mortem examination).

In addition, we included all non-Ukrainian MDR-TB (resistant to at least isoniazid and rifampicin) and pre-XDR-TB (MDR/RR-TB with additional resistance to any fluoroquinolone) cases diagnosed in the Czech Republic and Slovakia in 2023–2024 with an available MTBC isolate. As the largest migration waves occurred in 2022, we assumed that any measurable impact on MDR-TB in the host population would be reflected in 2023 and 2024. For this reason, non-UA MDR-TB cases from 2021 to 2022 were not included. Extensively drug-resistant (XDR) TB strains, defined as pre-XDR with additional resistance to at least one additional Group A drug (bedaquiline or linezolid), were also not available in our cohort [18].

DNA extraction and whole-genome sequencing (WGS) adhered to the protocol described in our prior publication [19]. The genomic DNA from inactivated bacterial cultures was extracted according to the manufacturer’s protocol using QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany). WGS was performed on the Illumina MiSeq platform (Illumina, San Diego, USA).

WGS data analysis

Quality of the sequencing reads was assessed with FastQC v0.11.4 and multiqc v1.13.dev0 while contamination was assessed with kraken2 v2.1.2 [20–22]. Samples with > 5% contaminating reads (n = 3) or a coverage less than 30x (n = 5) were removed from further analyses [20–22].

MTBseq with default settings was used to map FastQ files to the M. tuberculosis H37Rv genome as described previously [23]. The resulting concatenated single nucleotide polymorphism (SNP) alignment was processed with snpsites to remove invariant sites. A maximum likelihood (ML) tree was calculated from the filtered SNP alignment by IQ-TREE v1.6.12 with automatic model selection and including ascertainment bias correction (ASC) [24]. Branch support was assessed via Ultrafast bootstrap approximation (UFBoot) with 1000 replicates and using the -bbni parameter to further optimize each bootstrap tree using a hill-climbing nearest neighbor interchange.

Phylogenetic lineage and sublineage classification was done with specific signature SNPs [25, 26]. Genome-based resistance prediction was performed by identifying resistance-associated SNPs in 27 genes from the WGS data, using a previously published interpretation catalogue [27, 28]. Non-MDR was defined as resistance to one or more drugs but not meeting MDR criteria.

Potential transmission clusters were identified using a previously designed core genome multilocus sequence typing (cgMLST) scheme comprising 2,891 loci [29]. Strains were grouped into clusters by applying thresholds of 5 and 12 allelic differences (d5 and d12 clusters, respectively) to identify tuberculosis infections potentially associated with direct transmission events, as previously described [29].

Statistical analysis

Data on MDR-TB/pre-XDR cases reported annually in the Czech Republic and Slovakia between 2014 and 2024 were compiled from national surveillance records. Temporal trends on MDR-TB cases reported annually were visualized using line graphs with fitted trend lines to illustrate changes over time. Graphs were generated in Python (version 3.13.2) using the matplotlib and seaborn libraries and GraphPad (version 8.0.1). To compare the proportions of Ukrainian citizens among all notified TB cases between different years, chi-square test was performed using GraphPad (version 8.0.1). Trends in antibiotic resistance over time were evaluated using the CochranArmitageTest function from the DescTools package in R, with two-tailed tests and a significance threshold of P < 0.05 [30].

Patient and public involvement

Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.

Results

TB notifications and the contribution of Ukrainian citizens

Following 2022, TB notifications (including migrants) increased with distinct age-specific patterns in Slovakia and the Czech Republic. In Slovakia (Fig. 1A), both adult and paediatric (< 15 years) notifications rose to a peak in 2023 and then declined in 2024. In the Czech Republic (Fig. 1B), paediatric notifications peaked in 2023 and decreased in 2024, whereas adult notifications increased markedly in 2023 and remained stable in 2024. Although this year coincided with the onset of the war in Ukraine and the subsequent mass migration, the observed increase in paediatric cases was not primarily related to migration [31]. In the Czech Republic, Ukrainian citizens accounted for only 3 of 29 paediatric TB cases in 2023 and 4 of 17 in 2024. Instead, the rising number of paediatric cases was mainly associated with large-scale outbreaks within Roma communities in both countries.

Fig. 1.

Overal number of tuberculosis cases in adults and children under 15 years of age in Slovakia (A) and the Czech Republic (B), and the distribution of nationalities among TB patients in Slovakia (C) and the Czech Republic (D) notified during 2007–2024 and 2004–2024, respectively

In the adult population, a significant increase in the proportion of TB cases caused by Ukrainian citizens was observed in the Czech Republic between the years 2021 and 2022 (13.2%, P < 0.001; Fig. 1D). In Slovakia, despite a modest rise in the proportion of Ukrainian TB patients (approximately 10%), the situation has remained relatively stable (Fig. 1C).

Regarding MDR-TB, a sharp increase was observed during 2022 and 2023, especially in Czech Republic, temporally coinciding with the large-scale migration following the onset of the conflict in Ukraine (Supplementary Fig. 1). Although a decline was noted in 2024, the overall trend indicates higher case numbers in the following years.

WGS analysis

Characteristics of Ukrainian TB patients

In total, 381 Ukrainian patients were diagnosed with TB in the Czech Republic between September 2021 and December 2024. However, for only 206 of them (54.1%), culture isolates were available and included in the WGS analysis (Table 1 and Supplementary Fig. 2). In contrast, the number of Ukrainian patients from Slovakia (n = 23) included in the study represents all cases diagnosed during the study period. Most Ukrainian patients were male (69.9%) with a median age of 40 years (IQR 33–50) (Table 1). The majority presented with pulmonary TB (96.1%), and HIV co-infection was detected in 9.6% of cases. TB-related deaths occured in 1.7% of patients, while 3.1% died of other causes. Most cases were identified through symptom-based diagnosis (80.3%), whereas screening and contact tracing contributed only a small proportion (Table 1).

Table 1.

Demographic and clinical characteristics of culture-confirmed Ukrainian TB patients with an available MTBC isolate and thus eligible for WGS-based analyses

N	229
Characteristic	Category	n (%)
Country of diagnosis	Czech Republic	206 (89.96)
	Slovakia	23 (10.0)
Sex	Male	160 (69.9)
	Female	69 (30.1)
Age, years	Median (range/IQR)	40 (8–88/33–50)
Form of TB	Pulmonary	220 (96.1)
	Extrapulmonary	6 (2.6)
	Acute miliary	3 (1.3)
HIV status	Positive	22 (9.6)
Death	TB-related	4 (1.7)
	Other cause	7 (3.1)
Treatment history	Retreatment	9 (3.9)
Case detection	Symptom-based	184 (80.3)
	Screening of high-risk population	19 (8.3)
	Acidential finding	17 (7.4)
	Contact tracing	6 (2.6)
	Post-mortem examination	3 (1.3)

To evaluate possible MDR-TB transmission to the host populations, we additionally included all non-Ukrainian (non-UA) patients that were diagnosed with MDR-TB in the Czech Republic (n = 20) and Slovakia (n = 9) between January 2023 and December 2024 (Supplementary Fig. 1).

Phylogenetic lineages and genotypic resistance among isolates from Ukrainian migrants

Of the 229 MTBC strains obtained from Ukrainian TB patients, 199 (86.9%) met the predefined criteria for minimum DNA quantity and quality, as well as WGS data quality, and were included in the subsequent analyses. Among these, strains of Beijing sublineage 2.2.1 were predominant (n = 66; 33.2%), with the most frequent subclade being Central Asian strains (n = 45; 68.2%), followed by the Europe/Russian W148 outbreak strains (n = 13; 19.7%) and the Central Asia outbreak strains (n = 6; 9.14%) (Supplementary Table 1). Other frequently detected strains types were lineage 4, most notably 4.1.2.1 (n = 40; 20.1%), 4.8 (n = 32; 16.1%), and 4.3.3 (n = 24; 12.1%).

Genotypic drug-resistance prediction identified 129 MTBC strains (64.8%) from Ukrainian patients as fully susceptible, 37 (18.6%) as non-MDR drug-resistant, 25 (12.6%) as MDR, and eight (4.0%) as pre-XDR (Supplementary Table 1.). Among non-MDR isolates, resistance to isoniazid was most common (21/37, 56.8%). Interestingly, while no BDQ resistance was detected in MDR strains, three of the non-MDR strains were predicted to be monoresistant to bedaquiline/clofazimine, based on mutations in the Rv0678 gene: S53L in two clustered strains (with allele frequencies of 20.0% and 71.43%) and S63G in one strain (allele frequency 100.0%).

The distribution of resistances varied across lineages and time (Fig. 2A, B). Strains of Beijing sublineage 2.2.1 accounted for the majority of MDR (22 out of 25; 88%) and pre-XDR (6 out of 8; 75%) isolates, whereas strains of other common lineages, such as 4.8 and 4.7, were mainly drug-susceptible (Fig. 2A; Supplementary Table 1). Non-MDR drug-resistant isolates were frequently observed within all sublineages, although their overall prevalence remained low. None of the strains were predicted to be XDR.

Fig. 2.

A Relative frequency of resistance profiles among phylogenetic lineages of Mycobacterium tuberculosis strains isolated from 199 Ukrainian patients. Lineages with fewer than 5 isolates were grouped as “Other lineages”, B Percentage of resistance profiles by year of diagnosis (2021–2024). S – susceptible, nonMDR – resistant to one or more drugs but not meeting MDR criteria, MDR – multidrug-resistant (resistant against at least isoniazid and rifampicin), preXDR – pre-extensively drug-resistant (resistant against at least isoniazid, rifampicin and a fluoroquinolone)

On a subtype level, the 22 L2.2.1 MDR strains are composed of strains from three main previously described Beijing clades namely Central Asia (n = 8), Central Asia outbreak (n = 5), and Europe/Russian W148 Outbreak (n = 9) (Supplementary Table 1). Out of 8 Central asia strains, four belonged to the Ukrainian MDR outbreak cluster, which was previously described and characterised by the mutations embA − 12c > t, fabG1 -15c > t; katG S315T, rpoB S450L and rpsL K88R (Supplementary Table 1) [8] Most of the pre-XDR strains belonged to the Europe/Russian W148 Outbreak clone (4 out of 6) (Supplementary Table 1.). In fact, all L2.2.1 Central Asia outbreak, L2.2.1 Ukraine outbreak (part of Central Asia Beijing lineage) and L2.2.1 Europe/Russian W148 Outbreak strains were MDR/pre-XDR while the majority of other L2.2.1 Central Asia Beijing lineage strains were susceptible (24/45, 53.3%) or non-MDR (12/45, 26.7%).

Over time, the percentage of susceptible isolates decreased in the Ukrainian migrant population, but this trend was not statistically significant (Fig. 2B).

Transmission of TB within Ukrainian migrants and between migrants and host populations

Using a 12 allele threshold (d12, a commonly used cut-off for recent transmission in TB), we identified 25 genomic clusters among strains from Ukrainian migrants involving a total of 61 patients (cluster rate of 30.6%). The largest cluster comprised eight strains, including the two isolates with the bedaquiline/clofazimine monoresistance mutation S53L in Rv0678, while all other clusters comprised three or less isolates. However, with a stricter genomic distance threshold of 5 alleles (d5) only 13 clusters were identified with a maximum of two isolates per cluster (Fig. 3).

Fig. 3.

Maximum likelihood phylogenetic tree of 225 Mycobacterium tuberculosis isolates, including 199 from Ukrainian refugees diagnosed in the Czech Republic (CZ) and Slovakia (SVK), as well as multidrug-resistant isolates from 19 non-Ukrainian patients from CZ and 7 from SVK. The tree was calculated from the concatenated single nucleotide polymorphism alignment of 15,709 informative sites. Clusters were calculated based on core genome multilocus sequence typing (cgMLST) with a distance threshold of 5 alleles (d5) and 12 alleles (d12). Branches with ultrafast bootstrap support values less than 95% are indicated with blue dots

Clusters were often geographically dispersed: 12 d12 clusters of Ukranian patient isolates included patients diagnosed in different administrative districts/regions of the Czech Republic, indicating either cross-district transmission linked to patients‘ movement between regions or transmission during migration or already in their home country. Moreover, six clusters involved Ukrainian migrants diagnosed in both Slovakia and the Czech Republic. Recent transmission however, could be clearly confirmed only in cluster 1 by epidemiological data, which consisted of family relatives. In the other cross-border clusters, transmission most likely occurred either prior to migration or during transit (Fig. 4). Given that Slovakia directly borders Ukraine, a large proportion of migrants had to pass through its territory; thus, it is plausible that some patients remained in Slovakia while others moved to the Czech Republic, potentially after sharing temporary accommodation facilities. Unfortunately, detailed data on individual migration routes, timing of movements, and district-level information for patients diagnosed in Slovakia were not available, which limits further epidemiological interpretation.

Fig. 4.

Sankey diagram depicting the distribution of patients of Ukrainian nationality assigned to individual genetic clusters (n = 61; Cluster, based on a distance threshold of 12 alleles) and their corresponding districts (District) of diagnosis. The flows are color-coded by the country (Country) where diagnosis was originally made. The diagram effectively visualizes the spatial spread of Mycobacterium tuberculosis complex strains by linking genetic clustering with precise geographic locations, providing insights into transmission patterns across different regions

To elucidate resistant TB transmission between Ukrainian migrants and non-UA patients, we performed a cluster analysis of 59 MDR-TB and pre-XDR strains from both groups (19 citizens of the Czech Republic, 7 of Slovakia and 33 migrants from Ukraine). A total of 13 d12 clusters comprising 32 strains with a maximum of four strains per cluster and 10 d5 clusters comprising 20 strains with a maximum of two strains per cluster were detected (Figs. 3 and 5). Two d12 clusters included isolates from patients of both Czech and Ukrainian nationality, while another d12 cluster contained isolates from patients of Czech and Slovak nationality (Figs. 3 and 5). Epidemiological data revealed that these non-UA patients had recently shared the same accommodation facility with UA-patients in respective cluster, supporting the likelihood of direct transmission. Only one d5 cluster contained an isolate from a Ukrainian migrant and a non-UA patient (Czech nationality). Of all identified clusters, only one, comprising two Ukrainian isolates, contained both a pre-XDR and an MDR strain; all other clusters consisted of isolates with the same resistance category. Notably, 12 of the 26 non-UA MDR-TB patients (46.1%) were infected with the Beijing lineage, of which eight belonged to the Europe/Russian W148 outbreak clade and 3 to Central Asia (11.1%). This finding indicates that the impact of migration from former Soviet Union countries on the MDR-TB epidemic in the region likely predates the recent influx of Ukrainian migrants. All seven MDR-TB cases from patients with Slovak nationality belonged to lineage 4 (Supplementary Table 1).

Fig. 5.

Transmission of MDR and pre-XDR M. tuberculosis isolates among Ukrainian (UA; n = 16) and non-Ukrainian patients in Slovakia (n = 3) and the Czech Republic (n = 13). Clusters were determined using a 12 allele threshold: members of a cluster all had a maximum core genome distance of 12 alleles towards at least one of the other cluster member(s)

Discussion

This study provides novel insights into the epidemiology and transmission dynamics of TB among Ukrainian migrants in the Czech Republic and Slovakia, as well as the impact of high migration influx on MDR-TB in the host countries following the onset of the war in Ukraine. Our findings indicate that Ukrainian patients currently account for approximately 30% of all TB cases in the Czech Republic, compared to only about 10% prior to the war. The more pronounced increase in Czech Republic compared with Slovakia is likely multifactorial. A key driver is the substantially larger population of Ukrainian citizens residing in Czechia during 2022–2024, which would be expected to translate into higher absolute TB notifications even under comparable underlying incidence. Beyond population size, between-country differences may reflect heterogeneity in the demographic and social composition of Ukrainian citizens captured by surveillance (including long-standing labour migrants), differential geographic concentration in high-density urban settings, and local outbreak dynamics. Finally, differences in routine case ascertainment (access to care, diagnostic capacity and contact investigation intensity) may contribute to observed notification patterns, even in the absence of systematic screening. It is plausible that the true burden is underestimated, given that up to 50% of pulmonary TB cases may be asymptomatic or subclinical and therefore missed without radiological screening [32]. Importantly, the post-2021 changes in TB notifications observed in our settings should be interpreted in the broader European context, as several countries with comparatively lower levels of displacement from Ukraine than Czech Republic and Slovakia (e.g., Spain and France) also reported increases after the COVID-19 period [33]. This may reflect a post-pandemic rebound in case detection and restoration of TB services/surveillance, in addition to migration-related changes in population structure and TB burden. The predominance of men among Ukrainian TB patients contrasts with the sex distribution among beneficiaries of temporary protection in the EU [3]. This is nevertheless consistent with global TB epidemiology, where adult men typically account for a disproportionate share of notified TB and likely reflects a combination of higher exposure and prevalence of key risk factors (e.g., smoking and harmful alcohol use) as well as gender-related differences in health-seeking and access to TB services [34–36]. In addition, “Ukrainian citizenship” in national registers does not exclusively capture recent war refugees and may include longer-term migrants whose sex distribution differs from that of temporary protection beneficiaries. In our study, only 8.3% of Ukrainian patients were diagnosed through screening, in contrast to other countries like Austria, Belgium, Finland, France and United Kingdom that implemented systematic or targeted screening of refugees admitted to community facilities, where 40–90% of TB cases were detected through this approach [2]. In the Czech Republic, screening has not been introduced until the influx of refugees had already placed substantial strain on the healthcare system, which likely contributed to underdiagnosis and unnoticed transmission events during the peak migration period [5]. Additionally, only 9 patients (3.9%) were reported as previously treated or undergoing retreatment, however, this proportion is probably underreported. Migrants may hide their prior TB treatment history due to concerns about stigmatization, potential barriers to employment, or fear of limited access to healthcare in the host countries. Such underreporting has important public health implications, as previous TB treatment is a well-recognized risk factor for drug resistance and relapse. Failure to capture accurate treatment histories may therefore obscure the true burden of MDR-TB and complicate surveillance, highlighting the need for improved cross-border data sharing and more systematic documentation of patient treatment histories in host settings. In patients under 15 years of age, we observed an increase in number of TB cases since 2022, exceeding the 10% rise reported by WHO/ECDC for the European region [33]. This increase in both host countries can only partly be attributed to cross-border movement related to the war in Ukraine, as the majority of pediatric cases originated within Roma communities.

The number of MDR-TB patients in Czech Republic in 2022 was already nearly twice as high as the annual averages observed in the years preceding the war, and in 2023 it was almost threefold higher. Still, these numbers remain lower than estimates provided by the WHO, with similar trends being reported in neighboring countries [4]. Our data showed that the vast majority (77%) of MDR and pre-XDR isolates from Ukrainian patients, as well as from non-Ukrainian patients in the Czech Republic, belonged to the Beijing sublineage 2.2.1, predominantly represented by strains of the Europe/Russian W148 and Central Asia clades. The presence of these clades in the non-UA population and low clustering rate with Ukrainian patients indicates their evolutionary success in the country long before the recent migration influx, suggesting that their expansion cannot be solely attributed to the war-related refugee crisis. Importantly, these strain types have been documented in several Eastern European countries, including Ukraine, and the transcontinental spread of certain strains, such as the W148 clade has also been reported [8, 11]. The potential evolutionary advantage of the Beijing sublineage has been widely documented and is likely multifactorial. Reported factors include higher transmissibility e.g. via compensatory mutations and virulence compared to other M. tuberculosis lineages, increased adaptability to diverse host populations, and a higher propensity for acquiring drug resistance due to specific mutations in DNA repair and replication pathways [10, 12, 37–39]. In particular, the Europe/Russian W148 clade has been associated with large MDR-TB outbreaks across Eastern Europe and Central Asia, while the Central Asia outbreak clade is increasingly recognized as an emerging contributor to resistance-driven transmission in cross-border settings e.g. Central Asia [11, 12, 40]. The predominance of strains of these clades in both migrant and non-migrant populations with MDR/pre-XDR TB underscores their public health importance and the necessity of targeted surveillance to monitor their spread, and drug resistance evolution. In Slovakia, none of the non-UA MDR isolates belonged to the Beijing sublineage 2.2.1, instead, all fell within Euro-American sublineages (L4). This pattern suggests that the current MDR burden among non-UA patients in Slovakia is driven primarily by long-standing L4 strains circulating regionally rather than by recent high migrant-influx. Two non-exclusive mechanisms are plausible: (i) ongoing transmission of established Euro-American MDR clones and/or (ii) de novo acquisition of resistance under treatment [41]. Notably, data from 2018 to 2019 already reported predominance of L4.7 among MDR/XDR isolates, supporting a pre-existing local reservoir of Euro-American resistant strains [42].

In our study, three isolates carried Rv0678 mutations (two isolates carried S53L and one isolate carried S63G) that have been previously associated with resistance to bedaquiline and/or cross-resistance to bedaquiline/clofazimine in clinical MTBC isolates [43, 44]. According to the WHO catalogue, the S63G mutation is presently listed as of uncertain significance, while S53L has not been classified [45]. The isolates carrying the S53L mutation were phenotypically resistant to bedaquiline, delamanid, and pretomanid, while remaining fully susceptible to other anti-TB drugs both phenotypically and genotypically. Phenotypic drug susceptibility testing for bedaquiline, delamanid and pretomanid was repeated twice to confirm these results. Notably, no established resistance-associated variants for delamanid or pretomanid were detected in ddn, fbiA/B/C, fgd1 or Rv2983, which underscores current gaps in the genomic catalogue for these newer agents and highlights the need to identify and curate additional resistance determinants. Taken together, these findings suggest that S53L may warrant prioritisation for further evaluation and potential future re-assessment in the WHO catalogue, although confirmation in larger datasets and functional/phenotypic validation will be required. Also, phenotypic drug susceptibility testing for second-line agents is routinely performed only for MDR isolates, which further underscores the clinical importance of genomic resistance prediction, particularly in the context of the Beijing lineage, known for its higher propensity to develop MDR. Notably, previous studies have reported bedaquiline and delamanid resistance in treatment-naïve MTBC isolates, suggesting that certain strains might possess intrinsic resistance to these agents [46–48]. The emergence of such resistance in patients without prior documented exposure to bedaquiline is concerning and may be explained by spontaneous mutations under natural selection pressure as reported in an MDR strain from Eswatini which carried bedaquiline resistance before the drug was used in country [48, 49]. Although the clinical significance of low-frequency mutations (e.g., allele frequency 20%) remains uncertain, their presence raises concerns about the potential for amplification and fixation of resistant subpopulations during therapy. Given the increasing global reliance on bedaquiline as a cornerstone of MDR-TB treatment regimens, the emergence of monoresistance may pose substantial challenges to treatment success in the future [50, 51].

Several limitations should be considered. First, whereas the Slovak dataset included MTBC isolates from all notified Ukrainian TB patients, the Czech dataset was constrained by isolate availability, with isolates retrievable for only 206/381 (54.1%) cases reported during the study period. Nevertheless, the relatively large and representative cohort, together with high-quality genomic data, provides sufficient power to draw meaningful conclusions regarding transmission patterns and the molecular epidemiology of TB in the studied populations. Secondly, the absence of data on the timing and routes of migration limited our ability to interpret transmission dynamics and to distinguish infections acquired prior to migration from those contracted within the host countries.

Conclusions

This study represents the first large-scale molecular epidemiological investigation of TB in the context of migration driven by the war in Ukraine, providing novel evidence on its longer-term impact on TB/MDR-TB transmission in countries with high migration influx. Extending the observation period beyond preliminary analysis allowed for more robust assessment of temporal trends in drug resistance and lineage distribution, while increasing the representativeness of the dataset across diverse epidemiological contexts. Importantly, the inclusion of additional cases enabled improving our understanding of TB transmission dynamics within this high-risk population [6]. We demonstrate a substantial rise in the proportion of TB cases among Ukrainian patients, accompanied by an increase in MDR-TB, particularly in the Czech Republic. Genomic analyses confirmed the predominance of strains of Beijing lineage 2.2.1, including strains of the Europe/Russian W148 and Central Asia outbreak clades, highlighting their role in cross-border transmission and persistence of drug resistance. Our findings underscore the neccesity of use of WGS as a powerful tool for real-time surveillance of TB transmission. In future large-scale displacement scenarios, our findings support the use of integrated register-based monitoring and routinely performed WGS to enable timely identification of outbreaks and early detection of transmission links and emerging resistance (including new/repurposed drugs), thereby facilitating targeted interventions and reducing the overall burden on healthcare systems. The observation of rare monoresistance to bedaquiline underlines the necessity for careful monitoring of novel drug resistance, even in the absence of prior treatment with these agents. Finally, strengthening cross-border data linkage (e.g., migration timing and movement history) would further improve incidence estimation and operational planning.

In conclusion, strengthening regional cooperation, ensuring equitable access to diagnostic and treatment services, and integrating genomic surveillance into routine TB monitoring will be critical to mitigating the future impact of MDR-TB across Europe.

Abbreviations

TB

Tuberculosis

MDR

Multidrug-resistant

WGS

Whole genome sequencing

MTBC

Mycobacterium tuberculosis complex

Pre-XDR

Pre-extensively drug-resistant

Authors’ contributions

All authors made substantial contributions to conception and design; and/or acquisition, and/or analysis or interpretation of the data. The study was conceived by MD, JM, and VD, and designed by MD, JM, VD, SN, MER, PK, IP and JH. MH, JW, IP, MŠ, SM, IS and VD contributed to data acquisition. Data were analysed by MD and MDi, with input from KD, SM, IB, MŠ and MK. The manuscript was drafted by MD. All authors contributed to the interpretation of findings and critically reviewed the manuscript. All authors have read and approved the final version, and take responsibility for the decision to submit the manuscript.

Funding

The study was supported by APVV-18-0084, APVV-22-0342, VEGA-1/0093/22, VEGA- 1/0049/25, NIPH-75010330 and European Union within the Slovakia Programme for the project entitled Innovative trends in the treatment of respiratory diseases, ITMS21+: 401101C648.

Data availability

All newly sequenced datasets were submitted to European Nucleotide Archive under accession number PRJEB96899, while previously sequenced datasets are publicly available under accession number PRJNA994428. Individual accession numbers for all isolates are available in Supplementary Table 1.

Declarations

Ethics approval and consent to participate

The study was approved by the Ethical Committee of Jessenius Faculty of Medicine in Martin, Comenius University Bratislava, Slovakia (approval No. EK65/2021), in accordance with the Declaration of Helsinki. This study used Mtb isolates obtained as part of routine diagnostic procedures, and no additional biological material was collected specifically for this research. All patient-related data were fully anonymized prior to analysis, and no identifiable information was accessible to the researchers. Therefore, informed consent to participate was not required based on decision by Ethical Committee of Jessenius Faculty of Medicine in Martin, Comenius University Bratislava, Slovakia.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.