Recent Transmission and Prevalent Characterization of the Beijing Family Mycobacterium tuberculosis in Jiangxi, China

Tuberculosis (TB) remains a health threat to humans. According to the World Health Organization (WHO) 2020 Global tuberculosis report (WHO 2020), there were 10.0 million new cases of TB and 1.5 million deaths from TB in 2019. China has the second largest number of patients with TB, multidrug-resistant Mycobacterium tuberculosis (MDR-TB), or extensively multidrug-resistant M. tuberculosis (XDR-TB), only behind India (WHO 2020). MDR-TB shows resistance to at least isoniazid and rifampin. XDR-TB exhibits additional resistance to all kinds of fluoroquinolone and a second-line injectable drug. Although aggressive TB control has resulted in a sharp decrease in cases in recent years, it remains a public health concern in China. In Jiangxi, a southeast province of China, the prevalence of TB is also high. According to the Chinese national baseline surveillance, the prevalence of tuberculosis in Jiangxi province was 230/100,000 in 2010. Therefore, there is an urgent need to reduce the prevalence of tuberculosis. Considering that the main epidemic tuberculosis strains were Beijing genotype strains in Jiangxi Province, it is necessary to investigate the infection status, epidemiological characteristics, and risk factors of Beijing genotype strains (Luo et al. 2019).

Molecular typing technologies have been used to determine the genotypic diversity of M. tuberculosis (MTB), which played an important role in TB control and understanding of recent transmission. Mycobacterial Interspersed Repetitive Units and Variable Number of Tandem Repeats (MIRU-VNTR) is a powerful tool to differentiate the M. tuberculosis complex into various clusters and define the recent transmission. As the most evolved biological lineage among seven human-adapted phylogenetic lineages, the Beijing lineage strains show a global distribution. According to the statistics, more than 50% of these strains were present in East and South East Asia (Brudey et al. 2006). It is very valuable to have a broad knowledge of the association of the characteristics of the Beijing genotype M. tuberculosis with social demography. One study (Singh et al. 2015) suggested that drug resistance was related to its genotype, which appeared quickly in Beijing genotype strains (Maeda et al. 2014). Some studies even considered that age, gender, and region have important effects on the recent transmission of Beijing family strains (Yang et al. 2012; Zanini et al. 2014; Mohajeri et al. 2016).

Jiangxi is facing the health challenge of TB prevention and control (Chen et al. 2019; Luo et al. 2019). However, the association between the prevalence of Beijing genotype strains and the related characterization is still unclear. To understand whether there is an association between Beijing genotype strains and patient characterization, we investigated sociodemographic factors and drug susceptibility tests (DST) between Beijing genotype and non-Beijing genotype strains in Jiangxi province.

The most popular MTB family was the Beijing genotype. In total, 8,351 suspected pulmonary TB patients visited Jiangxi Chest Hospital from January to December 2016, and 20.91% (1,746/8,351) were cultured positive. After excluding 96 non-tuberculous mycobacteria (NTM) and 217 contaminated MTB, 1,433 strains were adopted. Of all, 658 strains were clustered, and 775 were unique isolates. 1,220 MTB belonged to the Beijing family, including 612 strains that clustered and 508 had unique patterns. Compared to strains from the non-Beijing family, isolates from the Beijing family clustered more easily (p = 0.001, OR (95% CI): 6.99 (5.01–9.77). All results are shown in Fig. 1.

Fig. 1

Distribution of isolates. A total of 1,433 strains were acquired from patients diagnosed with tuberculosis during this period in Jiangxi province, China. Of all 1,433 isolates, 376 (26.24%, 376/1,433) were from Nanchang, 246 (17.17%, 246/1,433) from Shangrao, 141 (9.84%, 141/1,433) from Ji’an, 129 (9.00%, 129/1,433) from Yichun, 117 (8.16%, 117/1,433) from Jiujiang, 109 (7.61%, 109/1,433) from Fuzhou, 108 (7.54%, 108/1,433) from Ganzhou and 207 (14.44%, 207/1,433) from other districts. Consequently, for 1,120 strains of the Beijing family, 282 (282/1,120) of them were from Nanchang, 199 (199/1,120) from Shangrao, 118 (118/1,120) from Ji’an, 103 (103/1,120) from Yichun, 84 (84/1,120) from Jiujiang, 90 (90/1,120) from Fuzhou, 91 (91/1,120) from Ganzhou, and 153 (153/1,120) from the rest districts (Fig. 2, Table SI). According to the Jiangxi regional distribution, the whole province can be divided into three regions, including southern regions (Ganzhou and Ji’an districts), central regions (Nanchang, Fuzhou, Yichun, Xinyu, and Pingxiang districts), and northern regions (Jiujiang, Jingdezhen, Shangrao and Yingtan districts). Among 1,433 isolates, the majority come from the central regions (722, 50.38%), followed by the northern regions (462, 32.24%) and the southern regions (249, 17.38%). Fig. 2 indicates the different geographical sources of these isolates.

Fig. 2

Demographic characteristics and drug resistance patterns. Table I shows the demographic characteristics of these patients. Among 1,433 patients, 860 were men, and 573 were women. Age ranged from 3 to 89 years, with an average age of 43.61. Regarding the history of treatment, 1,174 patients received single treatment, and 257 patients were treated at least twice (including 168 recurrent cases and 91 treatment failure cases). There was a total of 1,120 (1,120/1,433, 78.16%) MTB belonging to the Beijing family. Compared to cases of M. tuberculosis of the non-Beijing lineage, we found a significant difference in the Beijing family isolates among young and middle-aged people (less than 50 years old), people with recurrent status, and personnel in the northern regions. More than half of the patients (722, 50.38%) came from the central regions. Among them, 992 isolates (992/1,433, 69.22%) were sensitive to the four first- and three second-line antituberculosis drugs, and 441 isolates (441/1,433, 30.78%) were resistant to at least one of these drugs. A total of 29.11% (326/1,120) of drug-resistant tuberculosis belonged to the Beijing genotype, and that of MDRTB was 40.80% (133/326). The drug-resistant rate of RIF/INH/SM/EMB/AK/LEV/CM in the Beijing genotype family strains (15.09%/17.23%/10.62%/6.25%/1.3 4%/2.68%, respectively) was higher than that in non-Beijing strains (13.10%/12.78%/7.99%/4.79%/1.28%/2. 56%, respectively). However, statistical analysis did not reveal such a statistical difference. We also compared the percentage of strains resistant to multidrug (two or more drugs) resistant strains between the Beijing and non-Beijing genotype strains. The only difference was found in MDR-TB (11.88% vs. 7.35%, p = 0.05, OR (95% CI): 0.64 (0.41–1.01)) (Table I).

Genotypes of M. tuberculosis strains. To investigate the genotypes of 1,433 M. tuberculosis, the MIRU-VNTR method was adopted. Of these strains, 78.2% (1,120/1,433) belonged to the Beijing genotype strains, and the rest were non-Beijing genotype strains (Table SII). Non-Beijing genotype families, including the S, Cameroon, and NEW-1 families, adopted the strain identification method (https://www miruvntrplus.org/MIRU/index.faces)

To investigate the allelic diversity of these MIRU-VNTR loci, we calculated the Hunter-Gaston discriminatory index (HGDI) for each locus. As previously reported, the MIRU-VNTR loci were considered highly discriminatory (> 0.6), moderately (0.3–0.6), or poorly (< 0.3) discriminatory loci based on HGDI scores (Chen et al. 2016). These loci had a significant discriminatory ability with various HGDI scores (Table SII). According to the situation described above, three loci were considered highly discriminatory, including MIRU26 (HGDI = 0.6580), Qub26 (HGDI = 0.6344), and ETRE (HGDI = 0.6320). Eight loci (Mtub04, MIRU40, MIRU10, Mtub21, Qub11b, Mtub30, Mtub39, and Qub4156) had a moderate discriminatory ability, and the biomarkers of the remains were poorly discriminatory loci. The 15-loci discriminatory power reached 0.9963. At the same time, 1,433 strains were classified into 878 genotypes by adopting MIRU-VNTR cluster analysis, including 103 clusters and 775 unique patterns. The largest cluster was made up of 67 strains, and 11 clusters were made up of two strains. As a result, the clustering rate was 38.7% (555/1,433), and the recent transmission rate was 31.5% (452/1,433).

Association with the recent transmission of Beijing genotype strains. To understand whether prevalent characteristics impact Beijing lineage tuberculosis’s clustering ability and composition, we performed a statistical analysis to reveal the relationship between clustering ability and Beijing family bacteria. The information is shown in Tables II and III. The results showed that the non-clustered bacteria showed a stronger tendency to infect women (non-Beijing family strains, p = 0.03, OR (95% CI): 1.50 (1.06–2.14); Beijing family strains, p = 0.001, OR (95% CI): 1.52 (1.20–1.93)). Furthermore, when the Beijing family bacteria infected people, patients between 30 and 50 years of age were more likely to be clustered (p = 0.003, OR (95% CI): 3.02 (1.53 to 5.97)) compared to those less than 30 years of age. Among patients who do not cluster, the amount of bacteria from the Beijing family is significantly higher than the Beijing family bacteria (p = 0.012, OR (95% CI): 0.68 (0.51–0.91)). Regarding treatment history, the recurrent infection of Beijing family bacteria is more often clustered (p = 0.05, OR (95% CI): 0.65 (0.43–0.99)). Meanwhile, among patients infected with clustered bacteria, the Beijing family is significantly different from non-Beijing family bacteria in recurrence and treatment failure cases (Table III). To understand the role of AIDS, sputum smear, and lung cavity in the transmission of Beijing family strains, we compared HIV-negative, sputum smear-negative, and non-pulmonary cavity patients with the corresponding positive ones. We identified that the clustering rate between HIV-positive cases was significantly lower than that of non-clustered positive cases (p = 0.001, OR (95% CI): 0.20 (0.09–0.45)), while the smear-negative non-clustered Beijing family bacteria negative to smears was significantly lower than that of smear-positive bacteria of non-Beijing family (p = 0.001, OR (95% CI): 0.57 (0.41–0.80)). However, the pulmonary cavity did not affect the clustering of M. tuberculosis and the composition of the Beijing genotype.

To better implement preventive measures in Jiangxi province, it is necessary to understand the association of Beijing genotype M. tuberculosis transmission and its prevalent characterization. Beijing genotype M. tuberculosis was considered one of the most successful lineages, and it was more transmissible than other families of M. tuberculosis in Peru children (Huang et al. 2020). To understand TB’s characteristics and seek methods to prevent a local TB epidemic, it is necessary to know the molecular prevalence of M. tuberculosis (Song et al. 2020). In this study, we collected 1,433 TB isolates and tried understanding the detailed population status between Beijing and non-Beijing family isolates. We also investigated the association of the prevalence of the Beijing M. tuberculosis genotype with normal characterization and drug resistance. The aim was to reveal the association between clustered ability and different MTB features in Jiangxi province. We found that men were more likely to be infected with tuberculosis, which is consistent with the reported literature (Liu et al. 2018), while some references have a converse viewpoint (Hertz and Schneider 2019; Yang et al. 2021). The reason might be attributed to geographical differences. We also found that women were more likely to be infected by non-clustered strains (Tables II and III). The reason might be that the women had fewer social activities and less interpersonal contact, leading to less exposure to TB patients (Lin et al 2021).

In addition to gender, age was another important factor for Beijing genotype strain infection. Two previous studies showed that younger people (less than 25 years old) were prone to be infected with Beijing genotype strains (Pang et al. 2012; Huang et al. 2020). Our analysis also had the same results that young people (less than 30 years old) were more likely to be infected with the Beijing genotype M. tuberculosis compared to older adults (more than 50 years old) (Table I). Furthermore, we found that middle-aged people (30–50 years old) infected with Beijing family bacteria were more likely to cluster compared to those less than 30 years old (Table IV) (Mathema et al. 2017). Moreover, we also found that the population in northern regions was more easily infected by Beijing family bacteria, probably because the north of Jiangxi is closer to the north of China, where there are more Beijing family strains (Chen et al. 2018). At the same time, we found that the recurrent population was more easily infected by the Beijing genotype M. tuberculosis, which could be carried for an extended period and spread easily. Molecular epidemiology has increased our understanding of the prevalence of tuberculosis.

We found that the Beijing genotype was significantly associated with clustering, suggesting that recent transmission was substantially different from non-Beijing genotype strains, consistent with a study of Shanghai (Zanini et al. 2014).

Evidence has shown that the drug-resistance ability of M. tuberculosis has no correlation with sublineages (Yuan et al. 2015). Our results also proved no association between drug-resistant patterns and the Beijing genotype. However, Beijing genotype strains are more likely to develop MDR-TB, consistent with the previous report (Zhou et al. 2017). AIDS is a significant risk factor for TB infection (Bell and Noursadeghi 2018; Khan et al. 2019). Our study found that the HIV-positive population had a lower clustering rate. We deduce that this is because patients with AIDS are more prone to endogenous recurrence due to low immune function (Jasenosky et al. 2015; Amelio et al. 2019).

Although we have demonstrated important findings, this study has some limitations. Firstly, we use traditional MIRU-VNTR methods to study all TB transmission, while some patients with the same MIRU-VNTR patterns did not have epidemiological links. As a result, recent transmission rates were overestimated (Chen et al. 2016). MIRU-VNTR is more convenient and cost-effective than whole-genome sequencing (WGS); it is also of great value in defining the recent transmission of tuberculosis (Rizvi et al. 2020). Second, concerns about risk factors for defining clusters and distinguishing Beijing genotype were arguable. We had no opportunity to overcome selection bias for incomplete data on tuberculosis in the local population. However, the individuals in our study were completely random. Therefore, our findings had a high level of feasibility. In conclusion, our results had a specific value in controlling tuberculosis spread, especially for the Beijing genotype M. tuberculosis.