Journal of Infection (2014) 68, 332e337

www.elsevierhealth.com/journals/jinf

Molecular epidemiology of Mycobacterium tuberculosis in aboriginal peoples of Taiwan, 2006e2011 Yih-Yuan Chen a,j, Jia-Ru Chang a,j, Wei-Feng Huang a, Shu-Chen Kuo a, Jun-Jun Yeh b,c,d,e,f, Jen-Jyh Lee g, Chang-Sheng Jang h, Jun-Ren Sun i, Tzong-Shi Chiueh i, Ih-Jen Su a,*, Horng-Yunn Dou a,* a

National Institute of Infectious Diseases and Vaccinology, National Health Research Institutes, Zhunan, Miaoli, Taiwan b Ditmanson Medical Foundation Chia-Yi Christian Hospital, Chiayi, Taiwan c China Nan University of Pharmacy and Science, Taian, Taiwan d Meiho University, Pingtung, Taiwan e Department of Medical Research, Pingtung Christian Hospital, Pingtung, Taiwan f Chung Shan Medical University, Taichung, Taiwan g Department of Microbiology, Hualien Tzu Chi Medical Center, Taiwan h Pau-Le Christian Hospital, Taiwan i Division of Clinical Pathology, Department of Pathology, Tri-Service General Hospital and National Defense Medical Center, Taipei, Taiwan Accepted 4 December 2013 Available online 24 December 2013

KEYWORDS Mycobacterium tuberculosis; Aborigines; Haarlem strain; EAI strain; Multiple-drug resistant

Summary Previous research revealed a 6-fold higher incidence of tuberculosis (TB) amongst aborigines compared to Han Chinese in Taiwan. To investigate the reasons for this disparity, we genotyped Mycobacterium tuberculosis (MTB) strains obtained from members of different aboriginal tribes in different geographical regions of Taiwan by using molecular methods. In total, 177 isolates of MTB collected from patients at four hospitals in Taiwan from January 2006 to December 2011 were analysed by spoligotyping, mycobacterial interspersed repetitive unit-variable number tandem-repeat (MIRU-VNTR) typing. The most prevalent strains in the eastern and central regions of Taiwan were Beijing (45.7% in eastern) and Haarlem (39.1% in eastern, 37.1% in central) lineages, whereas in southern regions the

* Corresponding authors. National Institute of Infectious Diseases and Vaccinology, National Health Research Institutes, No. 35, Keyan Road, Zhunan Town, Miaoli County 350, Taiwan. Tel.: þ886 37 246166x35529; fax: þ886 37 583009. E-mail addresses: [email protected] (I.-J. Su), [email protected] (H.-Y. Dou). j These authors contributed equally to this work. 0163-4453/$36 ª 2013 The British Infection Association. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jinf.2013.12.004

Molecular epidemiology of M. tuberculosis

333

most prevalent strains were EAI (47.7%) and Haarlem (20.5%) lineages. The high prevalence of EAI in southern Taiwan aborigines may be closely associated with Austronesian culture. This study provides a first overview of the M. tuberculosis strains circulating in aboriginal populations in Taiwan. The high prevalences of certain MTB lineages within aboriginal sub-populations suggest that transmission of MTB may have been restricted to close contacts. ª 2013 The British Infection Association. Published by Elsevier Ltd. All rights reserved.

Introduction Tuberculosis (TB) remains a major health problem worldwide. In Taiwan, pulmonary TB is the No. 1 infectious disease, according to Department of Health statistics. Notably, Taiwan aboriginals have a 6-fold higher incidence and prevalence of TB compared to the general and Han populations, and aboriginal children under the age of 15 years experience tuberculous meningitis at rates up to 6e12 times higher than Han children.1 Several reasons may explain the high prevalence of TB in aboriginals: host factors such as polymorphism of the NRAMP-1 gene,2 environmental factors such as the close interpersonal contacts and poor hygiene in aboriginal communities, and microbial factors such as the virulence and drug resistance inherent to the Mycobacterium tuberculosis (MTB) strains in aboriginals. Clarifying the factors which underlie the high prevalence of MTB in aboriginals is therefore important for the control of TB in this population. In an earlier preliminary study, we analysed (by spoligotyping and MIRU typing) 200 random MTB isolates from patients of three hospitals representing different ethnic populations in Taiwan. That study demonstrated that the Beijing ancient strain and the Haarlem strain were the predominant MTB strains in aborigines in Hualien County in eastern Taiwan, whereas the Beijing modern strain was the predominant genotype in Han Chinese.3 The availability of molecular typing and high-resolution genotyping tools has led to a profusion of studies comparing mycobacterial isolates from different populations.4 The methods commonly used for MTB genotyping include restriction fragment length polymorphism analysis,5 spoligotyping,6 and variable-number tandem-repeat (VNTR) typing.7e9 We have undertaken to establish systematic surveillance of MTB strains in different aboriginal communities in Taiwan, which may provide important information to clarify transmission patterns and hostepathogen relationships, leading to more effective management and control. The purpose of the present study was twofold: 1) to determine, by using spoligotyping and 24-loci MIRU-VNTR technology, the genotypes of MTB isolates obtained from aborigines in different regions (eastern, central, and southern) or belonging to different tribes of Taiwan; and 2) to determine the prevalence of drug-resistant MTB in those same isolates. We selected four hospitals serving different aboriginal communities as our sampling sites. The proportional method for drug susceptibility testing (DST) of MTB was used, and the drugs tested were isoniazid, streptomycin, rifampin and ethambutol. The data were analysed with respect to prevalence of geno-

types, cluster pattern, and drug resistance of MTB isolates.

Methods Study setting This retrospective study was conducted at the National Health Research Institute in Taiwan. Most of the aboriginal peoples live in the central, eastern and southern parts of Taiwan. Thus, four hospitals including Mennonite Christian Hospital (eastern), Hualien Tzu Chi Hospital (eastern), Puli Christian Hospital (central), and Pingtung Christian Hospital (southern) in different regions of Taiwan were selected as the sampling sites for the study. From January 2006 to December 2011, 171 consecutive aboriginal patients with symptoms compatible with pulmonary TB and with sputum cultures positive for M. tuberculosis complex were included. One sputum isolate per patient was analysed in this study. MTB isolates were confirmed by conventional methods, including routine microscopy, culture, and positive nitrate and niacin tests. This study was approved by the Human Ethics Committee of the National Health Research Institutes, Taiwan (Code: EC1010804-E). Because of the retrospective nature, routine collection of clinical data in daily practice, and dislinkage of personal information, the requirement to obtain informed consent was waived by our institutional review board.

Mycobacterial genomic DNA Mycobacterial genomic DNA was extracted from cultured cells as described previously.10 Briefly, mycobacterial colonies were resuspended in 100e200 ml of distilled H2O and incubated at 85  C for 30 min. After centrifugation of the suspension, the supernatant containing the DNA was removed and stored at 20  C until further use.

Spoligotyping and spoligotype analysis Spoligotyping was carried out according to the manufacturer’s instructions (Isogen Bioscience B.V., Maarsen, The Netherlands). The SITVITWEB database11 and a web-based computer algorithm, Spotclust,12 were used to assign new isolates to families, subfamilies and variants. SITVITWEBassigned names (shared types) were used whenever a spoligopattern was found in the database. Patterns not found in SITVITWEB were assigned to families and subfamilies by Spotclust. Spoligotypes described only once (non-

334

Y.-Y. Chen et al.

Table 1 Distribution of strain lineages amongst MTB isolates from aboriginal patients. Genotype Region Eastern (%) Central (%) Southern (%) Total (%) Haarlem Beijing EAI T LAM Others

36 42 1 8 4 1

(39.1) (45.7) (1.1) (8.7) (4.3) (1.1)

13 6 4 8 0 4

(37.1) (17.1) (11.4) (22.9) (0) (11.4)

9 8 21 5 1 0

(20.5) (18.2) (47.7) (11.4) (2.3) (0)

58 56 26 21 5 5

(33.9) (32.8) (15.2) (12.3) (2.9) (2.9)

clustered) in this study and in SITVITWEB were designated as ‘orphan’. A cluster was defined as two or more isolates from different patients with identical spoligotype patterns.

PCR and MIRU analysis PCRs were carried out using a PCR reagent system (GibcoBRL). Sequences of the primers used for amplification of the 24 MIRU loci were selected according to descriptions in another study.5 The primer and MgCl2 concentrations used were as described by Mazars et al.9 The MIRU copy number per locus was calculated by using the conventions described by Supply et al.13

NTF locus analysis Multiplex PCR was used to determine possible IS6110 insertions in the NTF region of M. tuberculosis strains. The method, including primers as well as the IS6110 sequence and PCR parameters, was adapted from Plikaytis et al.14

Drug resistance testing The proportional method for drug susceptibility testing (DST) of MTB was performed as described previously.17 Briefly, for each drug, a 1:10 dilution of standardised suspension was inoculated onto the control and drugcontaining media. The extent of growth in the presence of drug was compared to the control and expressed as a percentage. If growth at the critical concentration of a drug was >1%, the isolate was considered to be clinically resistant. 7H10 agar with 0.2 or 1 mg/L isoniazid, 1 or 5 mg/L rifampin, 5 or 10 mg/L ethambutol, and 5 or 10 mg/L streptomycin was used.

Statistical analysis The strain-clustering rate (R) was calculated as (C  N)/S, where C was the total number of clustered isolates, N was the number of clusters, and S was the total number of isolates. Age of patients in different areas was compared with ANOVA and Scheffe’s post-hoc test. Categorical variables were compared using the Chi-square test.

Results Characteristics of aboriginal TB patients The average age (in years) of aboriginal TB patients in different areas of Taiwan was as follows: central: 51.3; southern: 64.6; eastern: 38.3 (Supplementary Table 2); eastern aborigines were significantly younger than southern aborigines (p Z 0.013). In total, 171 isolates from 171 patients (median age Z 51.4 years; n Z 108 male, 63.2%) with cultureconfirmed TB were subjected to spoligotyping and MIRUVNTR typing.

Detection of RD deletions Analysis of spoligotype patterns Previous studies have shown that certain RD deletions correlate with the relative age of the MTB strain.4,15,16 Beijing strains in our collection were further classified by using PCR to determine the presence or absence of RD105, RD181, RD150, RD142, and RD207. Details of this methodology have been described.4,15,16

Molecular analysis showed that all of the TB cases were caused by M. tuberculosis, except one by Mycobacterium bovis. Of the 171 isolates analysed, spoligotypes from 159 isolates (90.9%) were classified according to SITVITWEB into one of 33 shared international types (SITs)

Table 2 Genotype and cluster distribution based on combined spoligotyping and MIRU-VNTR typing amongst culture-positive aboriginal TB patients. Region

Eastern

Central

Strain

N

C (%)

U (%)

N

C (%)

U (%)

N

C (%)

U (%)

Beijing EAI H T LAM others Total

8 0 3 0 1 0 12

24 0 12 0 3 0 39

18 1 24 8 1 1 53

0 1 2 0 0 0 3

0 2 4 0 0 0 6

6 2 9 8 0 4 29

0 2 0 0 0 0 2

0 5 0 0 0 0 5

8 16 9 5 1 0 39

(57) (0) (33) (0) (75) (0) (42)

(43) (100) (67) (100) (25) (100) (58)

N: Number of clusters; C: Cluster of isolates; U: unique isolates.

Southern

(0) (50) (31) (0) (0) (0) (17)

(100) (50) (69) (100) (0) (100) (83)

(0) (24) (0) (0) (0) (0) (11)

(100) (76) (100) (100) (100) (0) (89)

Molecular epidemiology of M. tuberculosis Table 3 Region

Eastern Central Southern

335

Molecular cluster rates of major strain lineages. Age 65 >65 65 >65 65 >65

Beijing

EAI

Haarlem

U

N

C

R (%)

U

N

C

R (%)

U

N

C

R (%)

17 7 5 1 6 2

6 0 0 0 0 0

18 0 0 0 0 0

34.29 0 0 0 0 0

1 0 2 2 4 13

1 0 0 0 2 0

1 0 0 0 4 0

0 0 0 0 25.00 0

19 5 8 3 6 3

3 1 1 0 0 0

10 2 2 0 0 0

24.14 14.2 10.00 0 0 0

N: Number of clusters; C: Cluster of isolates; U: unique isolates; R: strain-clustering rate.

(Supplementary Table 1). Based on these defined spoligotypes, the most frequent strains belong to the Beijing lineage (35.2%) and the Haarlem lineage (33.9%) (Supplementary Table 1). Of the remaining 12 isolates that were not represented in SITVITWEB, 4 were Haarlem, 1 was EAI, 1 was LAM, 4 were T, 1 was X, and 1 was family-33, based on Spotclust (Supplementary Table 3). The prevalence of Beijing strains in the different geographic areas was 17.1% (central), 18.2% (southern), and 45.7% (eastern) (p < 0.001), whereas that of Haarlem was 37.1% (central), 20.5% (southern), and 39.1% (eastern) (p Z 0.089). Furthermore, the prevalence of EAI was 11.4% (central), 47.7% (southern) and 1% (eastern) (Table 1, p < 0.001). MTB-infected aborigines in central and eastern Taiwan were more likely to be infected with a Haarlem strain than with any other strain. In contrast, the EAI lineage was more prevalent in the aboriginal population in southern Taiwan.

Clustering analysis by 24-loci MIRU-VNTR All 171 isolates were also categorised by MIRU-VNTR typing and cluster analysis, which detected 4 major clusters in three MTB lineages (Tables 2 and 3). The cluster proportions in the 3 different sampled areas of Taiwan were as follows in patients younger than 65 years: eastern region, 34% in the Beijing lineage and 24% in the Haarlem lineage; central region, 10% in the Haarlem lineage; southern region, 25% in the EAI lineage (Table 3). To further sub-divide the isolates in the Beijing family, analysis of both the NTF and RD regions was performed. The distribution of the ancient N branch was 0% in central aborigines, 50% in eastern aborigines, and 50% in southern aborigines. The relatively modern NTF::IS6110 branch was

Table 4 Distribution of sublineages of the M. tuberculosis Beijing genotype in the general population, veterans, and aborigines. No. of isolates(%)

Central Eastern

Southern Total

W branch 3 (50%) 9 (21%) 1 (12%) NTF::IS6110 branch 3 (50%) 12 (29%) 3 (38%) N branch 0 21 (50%) 4 (50%)

13 (23%) 18 (32%) 25 (45%)

*N branch: No IS6110 insertion in the NTF region; NTF::IS6110 branch: a single IS6110 insertion upstream of the NTF region; W branch: two IS6110 insertions in the NTF region.

50% in central aborigines, 29% in eastern aborigines, and 38% in southern aborigines. The modern W branch was 50% in central aborigines, 21% in eastern aborigines, and 12% in southern aborigines (Table 4). Results based on RD determination showed a similar trend to the NTF analysis (Supplementary Table 4). Beijing isolates can be classified into one of the following five subgroups: subgroup 1 with the RD105 deletion; subgroup 2 with concurrent deletions of RD105 and RD207; subgroup 3 with concurrent deletions of RD105, RD207 and RD181; subgroup 4 with concurrent deletions of RD105, RD207, RD181 and RD150; subgroup 5 with concurrent deletions of RD105, RD207, RD181, RD150 and RD142. Our sample demonstrates a predominance of subgroups 2, 3 and 4 in eastern aborigines, subgroup 4 in southern aborigines, and subgroups 3 and 7 in central aborigines.

Drug susceptibility testing Of the 92 strains isolated from eastern aborigines, 68.5% were sensitive to all four of the first-line agents tested, 17.4% were resistant to streptomycin, 22.8% were resistant to isoniazid, 6.5% were resistant to ethambutol, 22.8% were resistant to rifampin, and 17.4% were multidrug resistant (MDR). In central aborigines, 84.5% were sensitive to all four of the first-line agents tested, 6.3% were resistant to streptomycin, and 12.5% were resistant to isoniazid; in southern aborigines, 84.7% were sensitive to all four of the first-line agents tested, 6.8% were resistant to streptomycin, 4.6% were resistant to isoniazid, 2.3% were resistant to ethambutol and 2.3% were resistant to rifampicin (Table 5). Analysis of the

Table 5 Drug susceptibility and resistance to first-line anti-tuberculosis drugs. Region

Fully sensitive Streptomycin Isoniazid Ethambutol Rifampicin MDR

No. (%) of drug-resistant TB cases Eastern

Central

Southern

63 16 21 6 21 16

27 2 4 0 0 0

39 3 2 1 1 0

(68.48) (17.39) (22.83) (6.52) (22.83) (17.4)

(84.38) (6.25) (12.50) (0) (0) (0)

MDR: resistant to at least isoniazid and rifampicin.

(88.64) (6.82) (4.55) (2.27) (2.27) (0)

336 association between MDR and genotypes (as determined by spoligotyping) showed that the Beijing genotype is more likely to be MDR than all other genotypes (33.3% vs. 4.0%, p < 0.001).

Discussion Aborigines are estimated to constitute about 2% of the population of the island of Taiwan. Taiwanese aborigines are Austronesian peoples with linguistic and genetic ties to peoples of the Philippines, Malaysia, Indonesia, Madagascar, Polynesia, and Oceania.18,19 he results of linguistic studies suggest that Taiwan is the origin of Austronesians.20 The original Austronesians lived in southern Taiwan, and then migrated to central, northern and eastern Taiwan. Because eastern and central Taiwan are fairly remote and isolated, the aboriginal people residing there have retained relatively intact village lifestyles. Aborigines in the south also live in aboriginal settlements, but have closer interactions with the Han Chinese. In northern Taiwan, the aboriginal people are fully integrated with the Han Chinese. We demonstrated several years ago that the Haarlem strain is highly prevalent (47%) in eastern TaiwaneHualien aborigines, whereas the Beijing strain is the most common genotype in Han Chinese military veterans.3 In the present study, we found that the EAI strain is prevalent in southern Taiwan aborigines, perhaps closely associated with the Austronesian culture.21 This may explain why EAI is also predominant in southern Han Chinese people,22 due to transfer from aborigines to the latter. The high prevalence of the EAI lineage in southern Asia and southern Taiwan, which share related Austronesian languages, suggests a possible relationship between EAI spreading and migration of Austronesian-speaking people, but this hypothesis needs further investigation. Also in the present study, we found Haarlem lineages to comprise 39.1% and 37.1% of MTB strains in aborigines of eastern and central Taiwan, respectively (Table 1). In our earlier survey, the Haarlem family was present in 27% of aborigines and 13% of the general population.3 During Dutch colonisation of the island in the 17th century, transmission of Haarlem strains from the Dutch to the aborigines of Taiwan may have occurred. However, the lifestyle of the aboriginal people may be the reason for the current high percentage of this MTB family in the eastern and central aboriginal populations of Taiwan. Most patients with TB can be cured with treatment. Standard treatment worldwide for confirmed or suspected TB disease is based on the four first-line bactericidal drugs, of which isoniazid and rifampin are the most effective. In order to optimise standard anti-TB drug therapy and to increase the success rate of control programmes, it is important to understand the drug-resistance patterns in each region. Therefore, in this study we investigated the prevalence of genotype and drug resistance of MTB isolates in Taiwan. Trend analysis showed a significantly higher frequency of MDR-TB in eastern aborigines. Hsu et al. reported in 2010 that the prevalence of MDR-TB amongst new TB cases in eastern Taiwan was 2.1% in 2001e2002, and increased to 4.0% in 2004e2008.23 In eastern Taiwan

Y.-Y. Chen et al. (Hualien and Taitung Counties), the aboriginal population comprises about 29.4% of the total population, and 77% of MDR-TB patients were aboriginals,23 indicating that MDRTB is a major problem in the aboriginal population in eastern Taiwan. Our study had some limitations. The study participants came from those regions of Taiwan with the most aboriginal settlements (eastern, central, southern). The aboriginal people in the northern part of Taiwan have much closer interactions with Han Chinese and are not easily traced. Furthermore, this study was a retrospective study. Thus, the effect of missing data on the results is unknown. In conclusion, this study provides a first overview of the M. tuberculosis strains circulating in aboriginal populations in Taiwan. Based on a combination of spoligotyping and MIRU-VNTR analysis, our data show that Haarlem and Beijing are the prevalent lineages amongst aborigines. The presence of highly transmissible lineages such as the Beijing strain is significantly associated with MDR-TB. These findings indicate that TB is not optimally controlled in aborigines, and that efforts for control strategies should be reinforced. Strain analysis, together with virulence studies, will also help in pinpointing isolates associated with higher morbidity and mortality, with the aim of directing efforts to limit the spread of those strains within aboriginal communities. The high prevalence of the Beijing genotype in the aboriginal population warrants close attention to control policy and vaccination strategy.

Acknowledgements This project was supported by grants from National Health Research Institutes and National Science Council (NSC 1022320-B-400-006), Taiwan. We thank the mycobacteriology laboratories of Hualien Mennonite Christian Hospital, Hualien Tzu Chi Medical Centre, Pingtung Christian Hospital, and Pau-Le Christian Hospital for providing bacterial isolates. We also thank Dr Daryl Henderson for his kind help in improving the English of this manuscript.

Appendix A. Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.jinf.2013.12.004.

References 1. Tuberculosis annual report. Taipei, Taiwan: Center for Disease Control, Department of Health; 2002. 2. Hsu YH, Chen CW, Sun HS, Jou R, Lee JJ, Su IJ. Association of NRAMP 1 gene polymorphism with susceptibility to tuberculosis in Taiwanese aboriginals. J Formos Med Assoc 2006;105(5): 363e9. 3. Dou HY, Tseng FC, Lu JJ, Jou R, Tsai SF, Chang JR, et al. Associations of Mycobacterium tuberculosis genotypes with different ethnic and migratory populations in Taiwan. Infect Genet Evol 2008;8(3):323e30. 4. Tsolaki AG, Hirsh AE, DeRiemer K, Enciso JA, Wong MZ, Hannan M, et al. Functional and evolutionary genomics of

Molecular epidemiology of M. tuberculosis

5.

6.

7.

8.

9.

10. 11.

12.

Mycobacterium tuberculosis: insights from genomic deletions in 100 strains. Proc Natl Acad Sci U S A 2004;101(14): 4865e70. Supply P, Allix C, Lesjean S, Cardoso-Oelemann M, RuschGerdes S, Willery E, et al. Proposal for standardization of optimized mycobacterial interspersed repetitive unit-variablenumber tandem repeat typing of Mycobacterium tuberculosis. J Clin Microbiol 2006;44(12):4498e510. van Embden JD, Cave MD, Crawford JT, Dale JW, Eisenach KD, Gicquel B, et al. Strain identification of Mycobacterium tuberculosis by DNA fingerprinting: recommendations for a standardized methodology. J Clin Microbiol 1993;31(2):406e9. Brudey K, Driscoll JR, Rigouts L, Prodinger WM, Gori A, AlHajoj SA, et al. Mycobacterium tuberculosis complex genetic diversity: mining the fourth international spoligotyping database (SpolDB4) for classification, population genetics and epidemiology. BMC Microbiol 2006;6:23. Kamerbeek J, Schouls L, Kolk A, van Agterveld M, van Soolingen D, Kuijper S, et al. Simultaneous detection and strain differentiation of Mycobacterium tuberculosis for diagnosis and epidemiology. J Clin Microbiol 1997;35(4):907e14. Mazars E, Lesjean S, Banuls AL, Gilbert M, Vincent V, Gicquel B, et al. High-resolution minisatellite-based typing as a portable approach to global analysis of Mycobacterium tuberculosis molecular epidemiology. Proc Natl Acad Sci U S A 2001;98(4): 1901e6. Yamaoka Y. Helicobacter pylori typing as a tool for tracking human migration. Clin Microbiol Infect 2009;15(9):829e34. Kolk AH, Schuitema AR, Kuijper S, van Leeuwen J, Hermans PW, van Embden JD, et al. Detection of Mycobacterium tuberculosis in clinical samples by using polymerase chain reaction and a nonradioactive detection system. J Clin Microbiol 1992;30(10):2567e75. Demay C, Liens B, Burguiere T, Hill V, Couvin D, Millet J, et al. SITVITWEB e a publicly available international multimarker database for studying Mycobacterium tuberculosis genetic diversity and molecular epidemiology. Infect Genet Evol 2012; 12(4):755e66.

337 13. Supply P, Mazars E, Lesjean S, Vincent V, Gicquel B, Locht C. Variable human minisatellite-like regions in the Mycobacterium tuberculosis genome. Mol Microbiol 2000;36(3):762e71. 14. Plikaytis BB, Marden JL, Crawford JT, Woodley CL, Butler WR, Shinnick TM. Multiplex PCR assay specific for the multidrugresistant strain W of Mycobacterium tuberculosis. J Clin Microbiol 1994;32(6):1542e6. 15. Hirsh AE, Tsolaki AG, DeRiemer K, Feldman MW, Small PM. Stable association between strains of Mycobacterium tuberculosis and their human host populations. Proc Natl Acad Sci U S A 2004;101(14):4871e6. 16. Tsolaki AG, Gagneux S, Pym AS, Goguet de la Salmoniere YO, Kreiswirth BN, Van Soolingen D, et al. Genomic deletions classify the Beijing/W strains as a distinct genetic lineage of Mycobacterium tuberculosis. J Clin Microbiol 2005;43(7):3185e91. 17. PA W. Susceptibility testing of mycobacteria, nocardia and other aerobic actinomycetes. National Committee for Clinical Laboratroy Standards; 2000. 18. Hill C, Soares P, Mormina M, Macaulay V, Clarke D, Blumbach PB, et al. A mitochondrial stratigraphy for island southeast Asia. Am J Hum Genet 2007;80(1):29e43. 19. Michael I, Bird GH, Taylor David. Populating PEPII: the dispersal of humans and agriculture through Austral-Asia and Oceania. Quat Int 2007;118e119:145e63. 20. Gray RD, Jordan FM. Language trees support the express-train sequence of Austronesian expansion. Nature 2000;405(6790): 1052e5. 21. Dou HY, Huang SC, Su IJ. Prevalence of Mycobacterium tuberculosis in Taiwan: a model for strain evolution linked to population migration. Int J Evol Biol 2011;2011:937434. 22. Chang JR, Chen YY, Huang TS, Huang WF, Kuo SC, Tseng FC, et al. Clonal expansion of both modern and ancient genotypes of Mycobacterium tuberculosis in southern Taiwan. PLoS One 2012;7(8):e43018. 23. Hsu AH, Lin CB, Lee YS, Chiang CY, Chen LK, Tsai YS, et al. Molecular epidemiology of multidrug-resistant Mycobacterium tuberculosis in Eastern Taiwan. Int J Tuberc Lung Dis 2010; 14(7):924e6.