Unraveling tuberculosis patient cluster transmission chains: integrating WGS-based network with clinical and epidemiological insights

Abstract

Background: Tuberculosis remains a global health threat, and the World Health Organization reports a limited reduction in disease incidence rates, including both new and relapse cases. Therefore, studies targeting tuberculosis transmission chains and recurrent episodes are crucial for developing the most effective control measures. Herein, multiple tuberculosis clusters were retrospectively investigated by integrating patients' epidemiological and clinical information with median-joining networks recreated based on whole genome sequencing (WGS) data of Mycobacterium tuberculosis isolates.

Methods: Epidemiologically linked tuberculosis patient clusters were identified during the source case investigation for pediatric tuberculosis patients. Only M. tuberculosis isolate DNA samples with previously determined spoligotypes identical within clusters were subjected to WGS and further median-joining network recreation. Relevant clinical and epidemiological data were obtained from patient medical records.

Results: We investigated 18 clusters comprising 100 active tuberculosis patients 29 of whom were children at the time of diagnosis; nine patients experienced recurrent episodes. M. tuberculosis isolates of studied clusters belonged to Lineages 2 (sub-lineage 2.2.1) and 4 (sub-lineages 4.3.3, 4.1.2.1, 4.8, and 4.2.1), while sub-lineage 4.3.3 (LAM) was the most abundant. Isolates of six clusters were drug-resistant. Within clusters, the maximum genetic distance between closely related isolates was only 5-11 single nucleotide variants (SNVs). Recreated median-joining networks, integrated with patients' diagnoses, specimen collection dates, sputum smear microscopy, and epidemiological investigation results indicated transmission directions within clusters and long periods of latent infection. It also facilitated the identification of potential infection sources for pediatric patients and recurrent active tuberculosis episodes refuting the reactivation possibility despite the small genetic distance of ≤5 SNVs between isolates. However, unidentified active tuberculosis cases within the cluster, the variable mycobacterial mutation rate in dormant and active states, and low M. tuberculosis genetic variability inferred precise transmission chain delineation. In some cases, heterozygous SNVs with an allelic frequency of 10-73% proved valuable in identifying direct transmission events.

Conclusion: The complex approach of integrating tuberculosis cluster WGS-data-based median-joining networks with relevant epidemiological and clinical data proved valuable in delineating epidemiologically linked patient transmission chains and deciphering causes of recurrent tuberculosis episodes within clusters.

Keywords: WGS; network; reactivation; recurrence; reinfection; transmission; tuberculosis.

Figure 1

Geospatial distribution of studied tuberculosis clusters. The map displays geographical locations (i.e., cities, municipalities, or parishes in Latvia) representing the residence addresses of clustered patients, or other documented locations where physical contact occurred between these individuals and TB infection might be transmitted. Each dot on the map corresponds to one of these locations. The dots are annotated with symbols corresponding to each studied TB cluster (labeled A-R) and colored according to the determined sub-lineages of clusters’ Mtb isolates.

Figure 2

Comparison of phenotypic and WGS-based drug susceptibility testing results of drug-resistant M. tuberculosis isolates. AMK, amikacin; BDQ, bedaquiline; CFZ, clofazimine; DCS, D-cycloserine; DLM, delamanid; DST, drug susceptibility testing; EMB, ethambutol; ETO, ethionamide; FQ, fluoroquinolones; INH, isoniazid; LZD, linezolid; N, drug susceptibility testing results are not available; PAS, para-aminosalicylic acid; pDST, phenotypic drug susceptibility testing; PZA, pyrazinamide; R, resistant; RIF, rifampicin; S, sensitive; STM, streptomycin.

Figure 3

Maximum-likelihood phylogeny of the studied dataset. Hr, isoniazid-resistant; MDR, multidrug-resistant; Pre-XDR, pre-extensively drug-resistant; S, sensitive.

Figure 4

Median-joining networks of studied tuberculosis clusters. Networks illustrate genetic distances between Mtb isolates obtained from clustered patients with active TB and belonging to sub-lineages (A) 4.3.3 (LAM; clusters A–F), (B) 4.1.2.1 (Haarlem; clusters G–I), (C) 4.8 (T; clusters J–M), (D) 4.2.1 (Ural; clusters N and O), and (E) 2.2.1 (Beijing; clusters P–R). Using the Mtb ancestral genome as a reference point, networks demonstrate the direction of genetic changes between isolates. Each network node represents TB episodes of one or more patients infected with identical Mtb strains (zero SNV distance). Symbol combinations (i.e., A1) within nodes, and for recurrent TB cases—next to nodes (i.e., A2.1 and A2.2), represent each active TB episode. Network nodes are scaled according to the number of identical Mtb isolates, while numbers in brackets along network branches indicate the number of differing SNVs between nodes.