Recent transmission of Mycobacterium tuberculosis in China: the implication of molecular epidemiology for tuberculosis control

Abstract

Tuberculosis (TB) has remained an ongoing concern in China. The national scale-up of the Directly Observed Treatment, Short Course (DOTS) program has accelerated the fight against TB in China. Nevertheless, many challenges still remain, including the spread of drug-resistant strains, high disease burden in rural areas, and enormous rural-to-urban migrations. Whether incident active TB represents recent transmission or endogenous reactivation has helped to prioritize the strategies for TB control. Evidence from molecular epidemiology studies has delineated the recent transmission of Mycobacterium tuberculosis (M. tuberculosis) strains in many settings. However, the transmission patterns of TB in most areas of China are still not clear. Studies carried out to date could not capture the real burden of recent transmission of the disease in China because of the retrospective study design, incomplete sampling, and use of low-resolution genotyping methods. We reviewed the implementations of molecular epidemiology of TB in China, the estimated disease burden due to recent transmission of M. tuberculosis strains, the primary transmission of drug-resistant TB, and the evaluation of a feasible genotyping method of M. tuberculosis strains in circulation.

Keywords

tuberculosis / recent transmission / molecular epidemiology / China

Introduction

Tuberculosis (TB) has remained a significant public health concern in China, accounting for one million new cases of TB and 130 000 deaths every year [1,2]. Around 8.3% of the culture-positive cases have been found to be multidrug-resistant TB (MDR-TB) [3]. During the last two decades, China accomplished significant success in tracking the TB epidemic, reducing the active incident TB cases by more than 50% from 2000 to 2010 [1]. Despite this achievement, only slight decrease in the overall prevalence of disease could be observed from 2000 to 2010 (466/100 000 to 459/100 000 population) [4]. The high prevalence of MDR-TB, high disease burden in rural areas, and enormous migrations from rural to urban settings are still significant challenges which pose a major hurdle for TB control in the country. A recent mathematical modeling study has shown that the control of tuberculosis in China will still not be optimistic in near future under present strategies [5].

The hallmark of the disease is that most of the people remain asymptomatic after initial infection, i.e., the latent phase. Only 5%–10% of the infected individuals develop active disease within few months to several decades. Given the complex incubation period, from the time of infection to development of active disease, incident TB is sometimes also represented by a combination of recent transmission or reactivation (remote transmission). No separate treatment regimens are there to deal with these two forms of active TB cases, but different control strategies may be adopted, which can offer varying public health significance. If the number of cases from recent transmission is high, case-finding strategies must focus on active detection of sources of infection, their treatment and completion of treatment to render the source case non-infectious, and timely screening of close contacts for active TB cases. Meanwhile, if the case comes from endogenous reactivation, then screening of individuals/population at high risk to develop active TB and preventive measures for latent TB must be given full consideration. Therefore, having a clear distinction on the proportion of incident TB due to recent transmission and endogenous reactivation is important in TB control.

Given the complex nature of the disease, traditional epidemiological methods are not sufficient to decode the exact transmission in populations, which makes the current TB control strategy less effective. Conventionally, most of the cases in a setting are thought to be due to reactivation. However, molecular epidemiological studies have revealed that recent transmission can account for 35%–40% or even up to 70% of the notified TB cases [6–9]. Based on these observations, targeted control strategies are implemented to reduce significantly the incidence of disease in some areas. Targeted interventions, for example, were taken to address newly identified transmission high-risk groups, including human immunodeficiency virus (HIV)-positive individuals, drug addicts, and homeless people in San Francisco, United States. A significant decline in overall incidence of TB and recent transmission from 51.2/100 000 and 10.4/100 000 in 1992 to 29.8/100 000 and 3.8/10 million in 1997, respectively, was observed [10].

The control and prevention strategies of one region cannot be necessarily and effectively implemented in another region. Along with this, the key drivers for TB transmission may also vary. Therefore, molecular epidemiological studies that illustrate the transmission patterns of TB in China are particularly important for controlling the disease. In the present article, we review the recent achievements in molecular epidemiology of TB and estimate the role of recent transmission of M. tuberculosis on the disease burden in China.

Genotyping methods of Mycobacterium tuberculosis in China

Genotyping of M. tuberculosis strains has enhanced the understanding of TB epidemiology. Strains resulting from the recent transmission should belong to the same genotype (clustered), whereas those from reactivation will have a unique signature in the population. The most common genotyping methods used are IS6110-RFLP, followed by mycobacterial interspersed repetitive unit-variable number tandem repeat (MIRU-VNTRs) typing and spoligotyping [11].

IS6110-RFLP has been considered as the standard method for genotyping of M. tuberculosis in the 1990s. It works on the principle of copy number variation of insertion sequence IS6110 in different clinical isolates. Despite its high discriminatory power, it has technical difficulties in data production and comparison of results or sharing the data among different laboratories, limiting its large-scale use.

Spoligotyping is considered the simplest, rapid, and cost-effective technique that has been used to define the predominant and growing number of important clinical M. tuberculosis clones or strains. It works on the presence or absence of specific spacers between the direct repeat (DR) loci and requires small amount of DNA. The results are converted into an octal code and are easy to share among laboratories. However, spoligotyping also has some limitations in differentiating the dominant Beijing strains in China, which reduces its ability to explore transmission dynamics of M. tuberculosis in China.

MIRU-VNTR typing is currently widely used method, which consists of variable sizes of repeat regions in the genome. The discriminatory power of this method, as well as IS6110-RFLP, has been reported by using a combination of 15/24-loci spectrum [12]. MIRU-VNTR typing offers high-throughput typing, is convenient, and allows easy sharing of digital data for comparison among different laboratories. The standard 24-VNTR loci spectrum has been recommended as the routine genotyping method in the United States and many countries in Europe. However, many of the VNTR loci in the standard 15-/24-loci spectrum have limited discriminatory power among the Beijing strains [13]. The resolving power of same VNTR combinations may also vary among different mycobacterial lineages in various regions or settings. Recently, Allix-Beguec et al. proposed the addition of four hypervariable loci (VNTR-1982, 3820, 3232, and 4120) into the 24-VNTR set to increase their resolution in the regions mainly dominated by Beijing strains [14]. However, this 24+ 4 loci combination again includes many loci that have low discriminatory power, which could limit its application in high-disease-burden countries such as China [13].

Currently, no widely accepted standard VNTR set is available in China. Most of the studies used the 15- or 24-VNTR loci spectrum. A few studies tried to evaluate the potential combinations (ranging from 3 to 24 VNR loci) relevant to the local circulating strains (supplemental data) [13,15–19]. One of the studies proposed a new 15-VNTR pattern, but the sample size was very small (54 strains), which limited its application in other parts of the country [16]. Our group collected 1362 isolates from six research sites across China and proposed an optimized 9+ 3 VNTR loci (9 loci plus 3 hypervariable loci, VNTR-3232, 3820, and 4120) to study the transmission of M. tuberculosis in China [13]. Chen et al. evaluated the discriminatory power of standard 15-VNTR loci spectrum on 3966 samples collected from the national drug survey (31 provinces) in 2007 [3] and proposed a combination of 8 to 10 VNTR sets for different provinces [17]. These sets shared 7 VNTR loci with the 9+ 3 VNTR set (supplemental data). However, this study did not include the evaluation of the hypervariable loci, which are needed to discriminate the most common circulated Beijing strains. In addition, the cluster-randomized sampling method of the national drug-resistance survey might not be feasible to assess the ability of the VNTR combination in investigating the transmission of M. tuberculosis.

Further evaluation study systematically compared the 9+ 3 VNTR set with the standard 15/24 loci plus 4 hypervariable loci and other combinations used in China. By using a population-based collection of 891 clinical isolates from five provinces across the country, this study revealed that the 9+ 3 VNTR set has comparable discriminatory power to the 24-VNTR set plus 4 hypervariable loci and requires less labor demand [15]. Therefore, we suggested that the 9 VNTR loci might serve as the initial genotyping tool for comparative analysis of circulation isolates among different regions in China and that at least these three hypervariable loci should be added to study the recent transmission or outbreaks of M. tuberculosis.

Implementation of whole-genome sequencing analysis

The development of whole-genome sequencing (WGS) and its cost effectiveness utility has revolutionarily improved the understanding of TB transmission. WGS identifies sequence variations in the whole-genome level (or almost whole-genome) and helps in identifying recent transmission with more accuracy [20–23]. Genetic distance below 5–12 SNPs can be used as threshold to explain the recent transmission events [24]. However, these studies were from low-TB-burden countries. At present, there are lacunae of evidence on the comparison of WGS to traditional genotyping methods in different populations. We compared both methods with two large (>10 cases) clusters from Shanghai and Heilongjiang provinces and observed that WGS helps in further differentiation of VNTR-defined clusters [22]. Similar finding was also observed in a retrospective study on MDR-TB strains in Shanghai [23]. Most importantly, based on precise identification of transmission events, WGS can help in better allocating public health resources to conduct more effective investigation of transmission (e.g., the identification of high-risk population of recent transmission and the infectious source). In addition, WGS has also proved worthy in unraveling the long-term evolution of M. tuberculosis and clinical applications of drug-susceptible testing [25,26].

Assessing the burden of recent transmission of M. tuberculosis in China

The proportion of genotypic clusters in a population represents the recent transmission rate of M. tuberculosis. This proportion can be ascertained by analyzing all sputum culture-positive patients in a region for a certain period of time (at least three years). Studies carried out to date in China used either convenient sampling procedure from hospitals or samples obtained from epidemiological surveys. Such studies were not population based, and the cluster rate obtained could not reflect the real situation of recent transmission [27]. To the best of our knowledge, only a few population-based studies were conducted. Our group carried out a four-year prospective molecular epidemiology in five different fields and observed that, on average, 31% of TB patients were genotypically clustered, indicating the recent transmission of M. tuberculosis [28,29]. One of the studies revealed 23% cluster rate in one-year period in two provinces of Eastern China [30], whereas another study from the same group in four rural areas using combined genotyping methods observed cluster rate of 14% in a three-year period [31]. Similarly, Wang et al. observed 15% cluster rate from six provinces of Eastern China in six months [32]. A study in 30 townships of Jiangsu Province for a time span of three months observed 27% transmission rate [33]. Two studies without the statement of sampling strategies reported clustering rates of 25% and 55% in Hebei Province and Gansu Province, respectively [34,35] (Table 1). Unfortunately, the above observations could not make good comparison due to variation in proper research design and time and the use of different genotyping methods.

Molecular epidemiological studies require detailed epidemiological information and the use of high-resolution genotyping methods to elucidate recent transmission events. The VNTR genotyping combinations being used at present in China could not provide satisfactory resolution. Moreover, two patients without any transmission link may possibly harbor the same “genotype.” Such cases are more common among genetic conserved strains, such as Beijing family strains. The misclassification of real clusters or transmission can further hamper the efficiency of epidemiological investigation. Numerous studies have revealed that genotypic clustered strains have low proportion (<30%) of epidemiological links. A study in a rural area of China did not reveal any contacts among VNTR-defined clustered cases, suggesting that occasional exposure can also contribute to transmission [32]. Recently, One of our studies. suggested that spatial proximity (i.e., clustered cases most commonly living in the same community or connected streets) can also be the link for recent transmission of MDR-TB in Shanghai [23]. However, we cannot exclude the possibility of unidentified samples within the transmission chain. The development of high-resolution genotyping methods could enhance the accuracy of clustered isolates, hence increasing the possibility of identifying epidemiological links.

The identification of high-risk population groups and places is one of the major purposes of molecular epidemiology study of TB transmission. However, identifying high-risk factors of recent transmission of TB in China is difficult [22,23,28,32,33]. One of the common reasons could be the limited discrimination power of genotyping methods being used and the many barriers in epidemiological investigations to identify transmission links. In addition, being a high-disease-burden country, the infectious sources in the community are common in the general population and difficult to track. Despite some reports of suspicious outbreaks in schools [36–38], public places were considered the high-risk place of TB transmission [22,23,28,32]. In other words, implementing targeted interventions without a high-risk group or hotspot is difficult. The circumstances of risk factors for TB transmission in China was different from those reported in other countries, where recent transmission is found mostly among specific population groups, such as HIV-positive individuals, homeless people, and drug abusers. Thus, we suggest that for TB control in China, active case finding, especially the identification of high-risk groups or places, needs to be strengthened.

Indirect evidence of ongoing transmission-exogenous reinfection

Reinfection of M. tuberculosis among patients with recurrence or treatment history also suggests the importance of ongoing/recent transmission [39]. Patients with recurrent TB are usually considered as endogenous reactivation. However, comparison of the genotype of strains before and after recurrence can determine whether the patient has a true endogenous relapse or an exogenous reinfection. Data on the prevalence of reinfection of recurrent TB in China are limited. Two of our studies separately using VNTR genotyping method revealed that 42% (59/141) and 68% (25/37) of recurrent TB cases were due to exogenous reinfection [40,41]. A study on two-year follow-up of 249 TB patients from Anhui Province revealed that out of five recurrent patients, two had exogenous reinfection [42].

Similarly, development of drug resistance among patients with treatment history was believed to be caused by reactivation of the original strains in previous episodes. However, we observed that 84% (32/36) of patients with new resistance profiles were in fact due to reinfection of a new VNTR genotype with drug resistance [43]. Recently, we expanded the sample size and found the proportion of exogenous reinfection to be 59% (48/81) among retreated drug-resistance patients using both VNTR genotyping and WGS analysis [44]. Thus, these findings suggest a continued transmission force of M. tuberculosis in the community, which always tends to create new infections.

Transmission of MDR-TB in China

The prevalence of MDR-TB has posed a serious challenge in China as well as other high-TB-burden countries. In 2007, the National Drug Surveillance in China estimated that 37.7% of TB cases were resistant to any single drug (DR), and 8.3% were MDR-TB, of which only less than 10% could be diagnosed and treated [2]. In general, poor compliance to chemotherapy or inappropriate treatment during anti-TB therapy might lead to high risk of acquired drug resistance. The national survey reported that patients with multiple treatment history significantly increase the possibility of developing drug resistance. One study has also reported the treatment history as one of the risk factors in developing drug resistance, suggesting that inappropriate treatment of TB patients in China can be more common than expected, particularly in general hospitals [2,45]. Meanwhile, a recent study has shown that the genetic heterogeneity in M. tuberculosis population within host is much higher than we previously expected as large numbers of low-frequency mutations were found [46]. These low-frequency mutations, although neglected in most analyses, can give rise to the emergence of drug-resistant clones in cases of drug pressure weakening (e.g., inappropriate treatment) [46,47].

The extensive scale-up of the DOTS strategy has reduced acquired resistance due to inappropriate treatment, but the prevalence of MDR-TB in China still could not be curtailed. This paradoxical situation may be due to the initial spread of MDR strains in the population. Disease surveillance systems usually provide the proportion of DR/MDR-TB among new or retreated patients. However, the high proportion of drug resistance among treated patients somehow misleads the priority of controlling MDR-TB infection. In fact, screening for the proportion of new and retreated patients among MDR-TB cases revealed that most cases were from new patients rather than retreated patients, suggesting direct transmission of MDR strains [48]. In 2007–2008, the national drug-resistance tuberculosis surveys in China reported that more than 70% of DR patients actually had no treatment history [3]. Similarly, another study from urban and rural settings also reported 71% and 59% of patients with DR-TB and MDR-TB, respectively, had no treatment history [28]. Furthermore, molecular epidemiology study using WGS and epidemiological investigation delineated the evidence of recent transmission of MDR M. tuberculosis strains [23] and also revealed that MDR strains were more likely to be genotypically clustered [28]. Extensive transmission of isoniazid-resistant M. tuberculosis strains is reported to be associated with increased MDR-TB in rural areas of the country [49]. Mathematical modeling approach has suggested that if necessity measures and control strategies are not considered in the current situation, then there is a likely possibility that 90% of the MDR-TB in China will be due to the transmission of drug-resistant strains [50].

Current deficiencies and future challenges

Molecular epidemiological investigations in TB have been carried out for more than a decade in China. However, only a few population-based studies systematically evaluated the contribution of recent transmission of TB to the disease burden. Current evidence suggests that recent transmission of M. tuberculosis plays an important role in the generation of new TB cases and high MDR-TB burden. These findings may help devise targeted and effective control strategies for TB epidemics in China. However, due to the limitations and challenges in study design, case finding, and genotyping methods, the impact of recent transmission on the disease epidemic may be underestimated.

(1) Lack of prospective molecular epidemiological studies and incomplete sampling scheme. Ideally, molecular epidemiological investigation needs to incorporate all clinical isolates in a population over a certain period of time. Various factors, such as length of the study, case notification rate, and proportion of culture-positive cases, may affect the sample collection and lead to underestimation of recent transmission rate [27]. In most regions in China, due to the limited resource, the current case-finding strategy is a passive scheme, which mainly relies on symptomatic individuals voluntarily seeking medical care and treatment. One of the active case-finding studies from a rural area in Sichuan Province revealed that the passive strategy could capture only 25% of total TB cases compared with active case-finding strategies [51]. Our study in both rural and urban settings observed that Shanghai was the only region with intensive case finding and the combination of active case finding and routine case-recommendation strategy for many years, which resulted in the reliable case notification rate. Thus, the recent transmission rate of 32% in Shanghai is much more convincing than that observed in rural areas with high TB burden but low case notification rate in the study [28]. In high-TB-burden countries, such as China, elevated cost is one of the main barriers in implementing active case-finding interventions. However, mathematical modeling study suggests that even with a cost up to $4000 per case finding in China, it still has a high cost-effective impact in a long-time frame [52]. Therefore, the exploration of novel active case-finding strategies in China needs to be continued.

(2) Another challenging situation is the low proportion of bacteriologically confirmed TB cases in China. A study from our group in five provinces revealed that 30% of active TB cases were culture positive. The reasons behind this finding are not clear; they could be due to the low sensitivity of the diagnostic method and/or the low quality of diagnostic procedures. Meanwhile, such low proportion of culture-positive TB cases can slow down the effectiveness of molecular epidemiology study in delineating the real TB burden caused by recent transmission. Thus, strategies to augment the capacity building and improve the quality of sputum sampling procedures, microscopy testing, and bacterial culture are urgently needed to increase the proportion of bacteriologically confirmed TB cases in China. A long-term time frame (i.e., more than three years) is also important to assess the extent of recent transmission, since recently infected individuals could not be included due to the short duration of the study period.

(3) Non-availability of a standard genotyping method throughout the country. Given the absence of uniform genotyping methods, genotypic data from different studies and various laboratories are not comparable. To the best of our knowledge, we recommend the national scale-up of the “9+ 3” loci pattern or an optimized pattern based on this scheme. Moreover, establishment of a data center can effectively promote the nationwide sharing of genotypic data. This capacity building may benefit the comparison of various studies and further strengthen attempts to investigate the transmission of M. tuberculosis nationwide, especially the enormous internal migrant populations from rural areas with high TB prevalence to the urban cities with low TB prevalence. Meanwhile, with its reduced cost, WGS can become a routine method and a good alternative to distinguish clinical isolates and establish recent transmission events.

Summary

Endogenous reactivation can be treated as the pool reservoir of TB with the recent transmission of M. tuberculosis as the source, which is continually being added to the reservoir. If recent transmission cannot be controlled in time, it may deteriorate the current achievements and even increase disease burden. Therefore, large-scale prospective molecular epidemiological studies on TB based on intensive case finding need to be established. Capacity building in the diagnosis of culture-positive TB cases is another important aspect for the precise assessment of recent TB transmission. Along with this, an epidemiological meta-database linked to the genomic data needs to be developed as it will allow molecular epidemiology findings to be transformed into the corresponding prevention and control measures. We suggest that targeted interventions emphasized in controlling the recent transmission of M. tuberculosis based on molecular epidemiology studies will be an effective means for rapidly decreasing incident TB cases and can serve as an example to control TB in China and other high-disease-burden courtiers.

References

[1] Wang L, Zhang  H, Ruan Y,  Chin DP,  Xia Y, Cheng  S, Chen M,  Zhao Y, Jiang  S, Du X,  He G, Li  J, Wang S,  Chen W, Xu  C, Huang F,  Liu X, Wang  Y. Tuberculosis prevalence in China, 1990–2010; a longitudinal analysis of national survey data. Lancet 2014; 383(9934): 2057–2064

[2] WHO. Global tuberculosis report 2015. Geneva: World Health Organization, 2015. 

[3] Zhao Y, Xu  S, Wang L,  Chin DP,  Wang S, Jiang  G, Xia H,  Zhou Y, Li  Q, Ou X,  Pang Y, Song  Y, Zhao B,  Zhang H,  He G, Guo  J, Wang Y. National survey of drug-resistant tuberculosis in China. N Engl J Med 2012; 366(23): 2161–2170

[4] Tang S, Ping  H. New features and controlling strategies of tuberculosis in the new century in China. Chin J Pract Intern Med (Zhongguo Shi Yong Nei Ke Za Zhi) 2011; 6:403–5 (in Chinese)

[5] Lin HH, Wang  L, Zhang H,  Ruan Y, Chin  DP, Dye C. Tuberculosis control in China: use of modelling to develop targets and policies. Bull World Health Organ 2015; 93(11): 790–798

[6] Small PM, Hopewell  PC, Singh SP,  Paz A, Parsonnet  J, Ruston DC,  Schecter GF,  Daley CL,  Schoolnik GK. The epidemiology of tuberculosis in San Francisco. A population-based study using conventional and molecular methods. N Engl J Med 1994; 330(24): 1703–1709

[7] Alland D, Kalkut  GE, Moss AR,  McAdam RA,  Hahn JA,  Bosworth W,  Drucker E,  Bloom BR. Transmission of tuberculosis in New York City. An analysis by DNA fingerprinting and conventional epidemiologic methods. N Engl J Med 1994; 330(24): 1710–1716

[8] Clark CM, Driver  CR, Munsiff SS,  Driscoll JR,  Kreiswirth BN,  Zhao B, Ebrahimzadeh  A, Salfinger M,  Piatek AS,  Abdelwahab J. Universal genotyping in tuberculosis control program, New York City, 2001–2003. Emerg Infect Dis 2006; 12(5): 719–724

[9] Glynn JR, Crampin  AC, Yates MD,  Traore H,  Mwaungulu FD,  Ngwira BM,  Ndlovu R,  Drobniewski F,  Fine PE. The importance of recent infection with Mycobacterium tuberculosis in an area with high HIV prevalence: a long-term molecular epidemiological study in Northern Malawi. J Infect Dis 2005; 192(3): 480–487

[10] Jasmer RM, Hahn  JA, Small PM,  Daley CL,  Behr MA,  Moss AR,  Creasman JM,  Schecter GF,  Paz EA, Hopewell  PC. A molecular epidemiologic analysis of tuberculosis trends in San Francisco, 1991–1997. Ann Intern Med 1999; 130(12): 971–978

[11] Barnes PF, Cave  MD. Molecular epidemiology of tuberculosis. N Engl J Med 2003; 349(12): 1149–1156

[12] Supply P, Allix  C, Lesjean S,  Cardoso-Oelemann M,  Rüsch-Gerdes S,  Willery E,  Savine E,  de Haas P,  van Deutekom H,  Roring S,  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 Soolingen D. Proposal for standardization of optimized mycobacterial interspersed repetitive unit-variable-number tandem repeat typing of Mycobacterium tuberculosis. J Clin Microbiol 2006; 44(12): 4498–4510

[13] Luo T, Yang  C, Pang Y,  Zhao Y, Mei  J, Gao Q. Development of a hierarchical variable-number tandem repeat typing scheme for Mycobacterium tuberculosis in China. PLoS One 2014; 9(2): e89726

[14] Allix-Béguec C,  Wahl C, Hanekom  M, Nikolayevskyy V,  Drobniewski F,  Maeda S,  Campos-Herrero I,  Mokrousov I,  Niemann S,  Kontsevaya I,  Rastogi N,  Samper S,  Sng LH, Warren  RM, Supply P. Proposal of a consensus set of hypervariable mycobacterial interspersed repetitive-unit-variable-number tandem-repeat loci for subtyping of Mycobacterium tuberculosis Beijing isolates. J Clin Microbiol 2014; 52(1): 164–172

[15] Liu M, Luo  T, Yang C,  Liu Q, Gao  Q. Evaluation of different sets of variable number of tandem repeats loci for genotyping Mycobacterium tuberculosis isolates in China. Chin J Tuberc Respir Dis (Zhonghua Jie He He Hu Xi Za Zhi)  2015; 38(10): 746–750 (in Chinese)

[16] Lu B, Pourcel  C, Liu JH,  Dong HY,  Zhang YY,  Jiang Y,  Liu ZG, Zhao  XQ, Wan KL. Establishment of the standard operation program on multiple loci variable numbers tandem repeats analysis typing on Mycobacterium tuberculosis. Chin J Epidemiol  (Zhong-hua Liu Xing Bing Xue Za Zhi)  2008; 29(9): 919–924 (in Chinese)

[17] Chen HX, Cai  C, Liu JY,  Zhang ZG,  Yuan M, Jia  JN, Sun ZG,  Huang HR,  Gao JM, Li  WM. Discriminatory power of variable number on tandem repeats loci for genotyping Mycobacterium tuberculosis strains in China. Chin J Epidemiol (Zhonghua Liu Xing Bing Xue Za Zhi)  2017; 38(6): 794–799 (in Chinese)

[18] Shi L, Dong  H. BA Sang,  Xi Luo, Suo  L, Zhao X,  Wan K. Evaluation and analysis on genetic diversity of clinical Mycobacterium tuberculos isisolates from Tibetan patients by variable-number tandem repeats of mycobacterial interspersed repetitive units. Chin J Ioonoses (Zhongguo Ren Shou Gong Huan Bing Xue Bao) 2014; 30: 4 (in Chinese)

[19] Zhang L, Chen  J, Shen X,  Gui X, Mei  J, Deriemer K,  Gao Q. Highly polymorphic variable-number tandem repeats loci for differentiating Beijing genotype strains of Mycobacterium tuberculosis in Shanghai, China. FEMS Microbiol Lett 2008; 282(1): 22–31

[20] Gardy JL, Johnston  JC, Ho Sui SJ,  Cook VJ,  Shah L, Brodkin  E, Rempel S,  Moore R,  Zhao Y, Holt  R, Varhol R,  Birol I,  Lem M, Sharma  MK, Elwood K,  Jones SJ,  Brinkman FS,  Brunham RC,  Tang P. Whole-genome sequencing and social-network analysis of a tuberculosis outbreak. N Engl J Med 2011; 364(8): 730–739

[21] Walker TM, Ip  CL, Harrell RH,  Evans JT,  Kapatai G,  Dedicoat MJ,  Eyre DW,  Wilson DJ,  Hawkey PM,  Crook DW,  Parkhill J,  Harris D,  Walker AS,  Bowden R,  Monk P, Smith  EG, Peto TE. Whole-genome sequencing to delineate Mycobacterium tuberculosis outbreaks: a retrospective observational study. Lancet Infect Dis 2013; 13(2): 137–146

[22] Luo T, Yang  C, Peng Y,  Lu L, Sun  G, Wu J,  Jin X, Hong  J, Li F,  Mei J, DeRiemer  K, Gao Q. Whole-genome sequencing to detect recent transmission of Mycobacterium tuberculosis in settings with a high burden of tuberculosis. Tuberculosis (Edinb) 2014; 94(4): 434–440

[23] Yang C, Luo  T, Shen X,  Wu J, Gan  M, Xu P,  Wu Z, Lin  S, Tian J,  Liu Q, Yuan  Z, Mei J,  DeRiemer K,  Gao Q. Transmission of multidrug-resistant Mycobacterium tuberculosis in Shanghai, China: a retrospective observational study using whole-genome sequencing and epidemiological investigation. Lancet Infect Dis 2017; 17(3): 275–284

[24] Nikolayevskyy V, Kranzer  K, Niemann S,  Drobniewski F. Whole genome sequencing of Mycobacterium tuberculosis for detection of recent transmission and tracing outbreaks: a systematic review. Tuberculosis (Edinb) 2016; 98: 77–85

[25] Walker TM, Kohl  TA, Omar SV,  Hedge J,  Del Ojo Elias C,  Bradley P,  Iqbal Z,  Feuerriegel S,  Niehaus KE,  Wilson DJ,  Clifton DA,  Kapatai G,  Ip CL, Bowden  R, Drobniewski FA,  Allix-Béguec C,  Gaudin C,  Parkhill J,  Diel R, Supply  P, Crook DW,  Smith EG,  Walker AS,  Ismail N,  Niemann S,  Peto TE; Modernizing Medical Microbiology (MMM) Informatics Group. Whole-genome sequencing for prediction of Mycobacterium tuberculosis drug susceptibility and resistance: a retrospective cohort study. Lancet Infect Dis 2015; 15(10): 1193–1202

[26] Luo T, Comas  I, Luo D,  Lu B, Wu  J, Wei L,  Yang C, Liu  Q, Gan M,  Sun G, Shen  X, Liu F,  Gagneux S,  Mei J, Lan  R, Wan K,  Gao Q. Southern East Asian origin and coexpansion of Mycobacterium tuberculosis Beijing family with Han Chinese. Proc Natl Acad Sci USA 2015; 112(26): 8136–8141

[27] Glynn JR, Vynnycky  E, Fine PEM. Influence of sampling on estimates of clustering and recent transmission of Mycobacterium tuberculosis derived from DNA fingerprinting techniques. Am J Epidemiol 1999; 149(4): 366–371

[28] Yang C, Shen  X, Peng Y,  Lan R, Zhao  Y, Long B,  Luo T, Sun  G, Li X,  Qiao K, Gui  X, Wu J,  Xu J, Li  F, Li D,  Liu F, Shen  M, Hong J,  Mei J, DeRiemer  K, Gao Q. Transmission of Mycobacterium tuberculosis in China: a population-based molecular epidemiologic study. Clin Infect Dis 2015; 61(2): 219–227

[29] Yang C, Luo  T, Sun G,  Qiao K, Sun  G, DeRiemer K,  Mei J, Gao  Q. Mycobacterium tuberculosis Beijing strains favor transmission but not drug resistance in China. Clin Infect Dis 2012; 55(9): 1179–1187

[30] Hu Y, Jiang  WL, Zhao Q,  Wang WB,  Xu B. The combined application of multiple genotyping methods in identifying genotypes of Mycobacterium tuberculosis strain circulating in rural China. Chin J Tuberc Respir Dis (Zhonghua Jie He He Hu Xi Za Zhi) 2009; 32(8): 576–580 (in Chinese)

[31] Hu Y, Mathema  B, Zhao Q,  Chen L, Lu  W, Wang W,  Kreiswirth B,  Xu B. Acquisition of second-line drug resistance and extensive drug resistance during recent transmission of Mycobacterium tuberculosis in rural China. Clin Microbiol Infect 2015; 21(12): 1093 e9–e18

[32] Wang W, Mathema  B, Hu Y,  Zhao Q, Jiang  W, Xu B. Role of casual contacts in the recent transmission of tuberculosis in settings with high disease burden. Clin Microbiol Infect 2014; 20(11): 1140–1145

[33] Lu W, Lu  B, Liu Q,  Dong H, Shao  Y, Jiang Y,  Song H, Chen  C, Li G,  Xu W, Zhao  X, Wan K,  Zhu L. Genotypes of Mycobacterium tuberculosis isolates in rural China: using MIRU-VNTR and spoligotyping methods. Scand J Infect Dis 2014; 46(2): 98–106

[34] Tian LL, Si  HY, Mu TJ,  Fan WB, Wang  J, Jiang WM,  Li Q, Yang  B, Zhang Y,  Zhu BD. Molecular epidemiology of Mycobacterium tuberculosis in Gansu Province of China. Chin Med J (Engl) 2012; 125(19): 3458–3464

[35] Li Y, Cao  X, Li S,  Wang H, Wei  J, Liu P,  Wang J, Zhang  Z, Gao H,  Li M, Wan  K, Dai E. Characterization of Mycobacterium tuberculosis isolates from Hebei, China: genotypes and drug susceptibility phenotypes. BMC Infect Dis 2016; 16(1): 107

[36] Huang Y, Zhong  J, Zhang Y,  Chen S, Wang  X. Outbreak of tuberculosis in a high school in Zhejiang, China. Chin J Epidemiol (Zhonghua Liu Xing Bing Xue Za Zhi) 2015; 36(2): 172–175 (in Chinese)

[37] Ma MJ, Yang  Y, Wang HB,  Zhu YF, Fang  LQ, An XP,  Wan KL, Whalen  CC, Yang XX,  Lauzardo M,  Zhang ZY,  Cao JF, Tong  YG, Dai EH,  Cao WC. Transmissibility of tuberculosis among school contacts: an outbreak investigation in a boarding middle school, China. Infect Genet Evol 2015; 32: 148–155

[38] Chen W, Xia  Y, Li X,  Zhou L, Li  C, Wan K,  Cheng S. A tuberculosis outbreak among senior high school students in China in 2011. J Int Med Res 2012; 40(5): 1830–1839

[39] Nardell E, Churchyard  G. What is thwarting tuberculosis prevention in high-burden settings? N Engl J Med 2011; 365(1): 79–81

[40] Shen X, Yang  C, Wu J,  Lin S, Gao  X, Wu Z,  Tian J, Gan  M, Luo T,  Wang L, Yu  C, Mei J,  Pan Q, DeRiemer  K, Yuan Z,  Gao Q. Recurrent tuberculosis in an urban area in China: relapse or exogenous reinfection? Tuberculosis (Edinb) 2017; 103: 97–104

[41] Shen G, Xue  Z, Shen X,  Sun B, Gui  X, Shen M,  Mei J, Gao  Q. The study recurrent tuberculosis and exogenous reinfection, Shanghai, China. Emerg Infect Dis 2006; 12(11): 1776–1778

[42] Kan X, Wang  K, Yang JA,  Liu Z, Chen  G, Zhang Y,  Wang J, Liu  X, Wang Q. Molecular epidemiology study on relapse after cure of smear positive cases. Anhui Med J (Anhui Yi Xue) 2011; 32(1): 89–91 (in Chinese)

[43] Li X, Zhang  Y, Shen X,  Shen G, Gui  X, Sun B,  Mei J, Deriemer  K, Small PM,  Gao Q. Transmission of drug-resistant tuberculosis among treated patients in Shanghai, China. J Infect Dis 2007; 195(6): 864–869

[44] Nsofor CA, Jiang  Q, Wu J,  Gan MY, Liu  QY, Zuo TY,  Zhu GF, Gao Q. Transmission is a noticeable cause of resistance among treated tuberculosis patients in Shanghai, China. Sci Rep 2017; 7(1): 7691

[45] Xue He G, van den Hof  S, van der Werf MJ, Guo H,  Hu YL, Fan  JH, Zhang WM,  Tostado CP,  Borgdorff MW. Inappropriate tuberculosis treatment regimens in Chinese tuberculosis hospitals. Clin Infect Dis 2011; 52(7): e153–e156

[46] Liu Q, Via  LE, Luo T,  Liang L,  Liu X, Wu  S, Shen Q,  Wei W, Ruan  X, Yuan X,  Zhang G,  Barry CE 3rd,  Gao Q. Within patient microevolution of Mycobacterium tuberculosis correlates with heterogeneous responses to treatment. Sci Rep 2015; 5(1): 17507

[47] Sun G, Luo  T, Yang C,  Dong X, Li  J, Zhu Y,  Zheng H,  Tian W, Wang  S, Barry CE 3rd,  Mei J, Gao  Q. Dynamic population changes in Mycobacterium tuberculosis during acquisition and fixation of drug resistance in patients. J Infect Dis 2012; 206(11): 1724–1733

[48] Gao Q, Li  X. Transmission of MDR tuberculosis. Drug Discov Today Dis Mech 2010; 7(1): e61–e65

[49] Hu Y, Hoffner  S, Jiang W,  Wang W, Xu  B. Extensive transmission of isoniazid resistant M. tuberculosis and its association with increased multidrug-resistant TB in two rural counties of eastern China: a molecular epidemiological study. BMC Infect Dis 2010; 10(1): 43

[50] Mehra M, Cossrow  N, Kambili C,  Underwood R,  Makkar R,  Potluri R. Assessment of tuberculosis burden in China using a dynamic disease simulation model. Int J Tuberc Lung Dis 2013; 17(9): 1186–1194

[51] Liao P, Guo  M, Tang F,  Wang G, Siyuan  LI, Yang Z,  Cao B. Analysis of patients from active case finding in rural areas of Wusheng, Sichuan. Chin J Antituberculosis (Zhongguo Fang Lao Za Zhi) 2016; 38(7): 576–581 (in Chinese)

[52] Azman AS, Golub  JE, Dowdy DW. How much is tuberculosis screening worth? Estimating the value of active case finding for tuberculosis in South Africa, China, and India. BMC Med 2014; 12(216): 1–9

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag GmbH Germany, part of Springer Nature