Drug resistance and IS6110-RFLP patterns of Mycobacterium tuberculosis in patients with recurrent tuberculosis in northern Thailand



Srisin Khusmith, Department of Microbiology and Immunology, Faculty of Tropical Medicine, Mahidol University, 420/6 Rajvithi Road, Bangkok 10400, Thailand.

Tel: (66) 2 3549100, ext 1594; Fax: (66) 2 6435583; email: srisin.khu@mahidol.ac.th


The emergence of drug resistant Mycobacterium tuberculosis has become a global threat to tuberculosis (TB) prevention and control efforts. This study aimed to determine the drug resistance profiles and DNA fingerprints of M. tuberculosis strains isolated from patients with relapsed or retreatment pulmonary TB in Chiang Rai province in northern Thailand. Significant differences in multidrug resistance (MDR) (P = 0.025) and resistance to isoniazid (P = 0.025) and rifampin (P = 0.046) between first and second registrations of patients with retreatment TB were found. However, there were no significant differences in resistance to any drugs in patients with relapsed TB. The rate of MDR-TB strains was 12.2% among new patients at first registration, 22.5% among patients with recurrence who had previously undergone treatment at second registration and 12.5% at third registration. Two retreatment patients whose initial treatment had failed had developed MDR-TB with resistance to all TB drugs tested, including rifampin, isoniazid, streptomycin and ethambutol. IS6110-RFLP analysis revealed that 66.7% (10/15 isolates) of MDR-TB belonged to the Beijing family. In most cases, IS6110-RFLP patterns of isolates from the same patients were identical in relapse and retreatment groups. However, some pairs of isolates from retreatment patients after treatment failure had non-identical IS6110-RFLP patterns. These results suggest that, after failure and default treatment, patients with retreatment tuberculosis have a significantly greater risk of MDR-TB, isoniazid and rifampin resistance than do other patients.

List of Abbreviations:

bacillus Calmette–Guérin






insertion sequence 6110, M. tuberculosis, Mycobacterium tuberculosis


multi-drug resistance


multi-drug resistance tuberculosis




restriction fragment length polymorphism








variable number tandem repeat


extensive drug resistance

Tuberculosis remains a major public health problem worldwide. Almost two million people die of TB annually and an estimated one-third of the world's population has latent infection. The situation is worse in developing countries in South-East Asia and Africa where MDR is on the increase. In six Asian countries, drug resistance is reportedly present in an estimated 2.8% of new and 18.8% of previously treated TB patients [1]. Thailand, which is ranked 18th on a list of the 22 countries with the largest TB burdens, had a prevalence of approximately 192/100,000 people for all forms and an incidence rate of 62 new smear-positive cases per 100,000 in 2007 [2]. The first national drug-resistance survey conducted in 2002 reported 1% of MDR-TB in new TB cases and 20% in previously treated cases [3]. The rates had increased to 1.7% and 34.5%, respectively, by the second survey in 2006 [3]. Moreover, TB surveillance during 1996–1998 showed that a high proportion of TB cases with drug resistance occurred in Chiang Rai province in northern Thailand.

In general, patients who have undergone successful treatment with anti-TB drugs can develop active disease again subsequently, referred to as recurrent TB. However, the pathogenesis of recurrence and classification of recurrent cases are still unclear [4]. Recurrence of TB may be a result of regrowth of the treated bacterial strain in patients otherwise previously treated successfully, or of re-infection [4]. Relapse refers to a patient becoming culture-positive again, or evidencing clinical or radiographic deterioration consistent with active TB, sometime after completion of apparently successful anti-TB drug therapy that had resulted in culture-negativity [5]. The term retreatment refers to patients with recurrent TB who defaulted before completing their previous therapy or in whom initial treatments failed [4]. Whether recurrent TB represents exogenous re-infection by a new strain of M. tuberculosis or endogenous reactivation of the original strain is controversial and has been debated for decades [6], [7]. The importance of each of these possibilities likely varies according to the epidemiological context, the spread of MDR-TB, HIV infection [8], and the immigration of people from developing countries, which could modify disease transmission in areas at low risk of TB [8]. In patients infected with HIV and MDR-TB there is evidence for a greater risk of reactivation than of re-infection [9], [10].

The most-widely used method for typing M. tuberculosis to determine whether recurrent TB represents endogenous reactivation or exogenous re-infection has been a molecular method involving Southern blotting of PvuII-digested chromosomal DNA and hybridization with the insertion sequence (IS) 6110 [11]. In practice, it is generally accepted that two or more isolates with identical or near-identical (± one band) IS6110 fingerprints (known as clusters) represent a recent transmission event [12]. This technique is thus useful in distinguishing between recent epidemiological events (transmission) and distant epidemiological events (reactivation) [12]. In a single TB patient with a TB-free interval, it is assumed that isolates with identical IS6110- RFLP patterns denote endogenous reactivation of the previously infecting bacteria [13].

Apart from differentiating between endogenous reactivation and exogenous re-infections, IS6110-RFLP has allowed identification of different M. tuberculosis strains with varying degrees of virulence and drug resistance in different geographical areas [14]. Up to now, the largest family of M. tuberculosis strain has been the Beijing family. The highest prevalence of this family reportedly occurs in Asian patients [15] and it is associated with various phenotypes such as drug-resistance [14], treatment failure, relapse and febrile response to TB treatment [16]. In several Asian studies, the proportion of TB due to Beijing strains has been > 50% [14]. However, because the IS6110-RFLP patterns vary between different geographical areas, there is so far limited available data regarding recurrent TB in Thailand. Although there was a national anti-TB drug resistance survey during 1997–1998 as part of a global project to evaluate IS6110-RFLP patterns and the extent of clustering, this study did not assess linkage to TB treatment history and development of drug resistance. Other subsequent study in Chiang Rai assessed acquired drug resistance in patients who had become positive again after completion, default from or failure of a standardized treatment regimen [10]. These researchers commonly found non-identical IS6110-RFLPs in the first and subsequent episodes in TB-HIV patients. Successful treatment of TB depends upon selection of an effective drug regimen; however, drug resistance can evolve in originally drug-susceptible strains during anti-TB treatment. Therefore, this study aimed to evaluate the relationship between the quality of treatment and development of resistance by assessing drug resistant M. tuberculosis in relation to the molecular patterns in recurrent TB patients with either relapse or retreatment TB in Chiang Rai province in northern Thailand.



Two hundred and three M. tuberculosis isolates from 77 pulmonary TB patients who had registered twice or more, kindly provided by the Microbiology Laboratory, Chiang Rai provincial hospital and the National TB Reference Laboratory, Bureau of Tuberculosis, Thailand, were cultured. These isolates were selected from isolates of patients with recurrent TB that had been stored as part of a ten year analysis of TB by Chiang Rai provincial hospital from 1 January 1997 to 31 December 2006, as mentioned above. Among these 203 M. tuberculosis isolates, only 92 were successfully cultured and subjected to IS6110-RFLP analysis. These 92 isolates were from 42 patients with relapse or retreatment TB who had registered twice or more with pulmonary TB and had been treated with anti-TB drug regimens.


The patients were diagnosed by medical history, chest radiographic findings, microscopic examination for acid-fast bacilli in sputum and positive cultures of M. tuberculosis, followed by species identification by biochemical tests and gene probes (ACCUProbe, GenProbe, San Diego, CA, USA) at the National TB Reference Laboratory, Bureau of Tuberculosis, Thailand. The patients were categorized according to World Health Organization criteria [17], which include ascertaining whether or not the patients have previously received TB treatment. The TB drug regimens were based on the recommendations of the National Tuberculosis Program, Ministry of Public Health, Thailand. The standard TB treatment drugs were INH, RMP, PZA and EMB. Because immunocompromised patients are reportedly at greater risk of re-infection TB [10], [18], patients co-infected with HIV were excluded from this study by using particle agglutination assay (Serodia-HIV-1/2, Fujirebio, Tokyo, Japan) and micro particle enzyme immunoassay (AxSYM HIV Ag/Ab Combo, Abbott Laboratories, Abbott Park, IL, USA).

This study was approved by the Ethical Review Committee for Research in Human Subjects, Ministry of Public Health, Thailand (Reference number 3/2550).

Case definitions

According to World Health Organization definitions of cases and treatment outcomes [17], a patient whose sputum smear or culture is positive at the beginning of the treatment but negative in the last month of treatment and on at least one previous occasion is defined as “cured”. A patient who completes treatment but does not have negative sputum smears or cultures in the last month of treatment and on at least one previous occasion is defined as “completed”. “Treatment success” is the sum of cured and completed. “Failure” is patients whose sputum smears or cultures are positive 5 months or more after commencing treatment. A patient whose treatment is interrupted for two or more consecutive months is defined as “default”. “Died” refers to patient who have died for any reason during the course of treatment. “Relapses” refer to new TB episodes occurring after a period without TB. “Retreatments” (after failure or default) are continuations of TB episodes that require changes in treatment regimens: this traditionally requires re-registration.

In TB, drug resistance can be primary or acquired, primary resistance being defined as resistance in patients without a history of previous treatment. Acquired drug resistance is defined as resistance in those who have previously undergone TB treatment. Drug-resistant TB is classified as monodrug (resistance to a single first-line drug), polydrug (resistance to two or more first-line drugs) and multidrug resistance (MDR) (resistant to isoniazid and rifampicin, with or without resistance to any other drugs) [17].

Antibiotic sensitivity testing

Drug susceptibility testing was performed by the fully automated BACTEC MGIT 960 system (Becton Dickinson Biosciences, Sparks, MD, USA) for testing M. tuberculosis susceptibility to SM, INH, RMP and EMB at the following final drug concentrations: 1.0 μg/mL for SM, 0.1 μg/mL for INH, 1.0 μg/mL for RMP and 5.0 μg/mL for EMB.

Insertion sequence 6110-restriction fragment length polymorphism

Insertion sequence 6110-RFLP analysis was done by Southern blotting and DNA hybridization with an IS6110 probe ([10], [19], [20]). Briefly, chromosomal DNA of M. tuberculosis was extracted by chloroform-isoamyl alcohol. Three micrograms of DNA were digested with 10 U/μL of PvuII (Boehringer Mannheim, Mannheim, Germany) and electrophoresed in 0.8% agarose. The extracted DNA from M. tuberculosis MT14323 strain was used as control marker. DNA fragments were transferred to a nylon filter (Sigma Chemical, Saint Louis, MO, USA) by the capillary method [11] and hybridized with digoxiginin-labelled BamHI-SalI fragment of pDC73 [21]. The plasmid pDC73 contains a portion of the insertion sequence IS6110 on the right side of the PvuII restricted site. The IS6110 hybridization patterns were analyzed using Gelcompar II version 1.5 (Applied Maths, Kortrijk, Belgium). Based on 78% or more similarity as previously described, the isolates were classified as members of the Beijing family or Nonthaburi group, [20].

Data analyses

The data were statistically analyzed using SPSS version 17.0. Comparison of pair isolates within individuals was performed to assess similarity between patterns. Drug sensitivity test profiles between the first and the subsequent TB registrations were associated by X2 test. A P value < p; 0.05 was considered statistically significant.


Clinical characteristics of patients

The median ages of patients in the relapse and retreatment TB groups were 54 (range 25–74) and 47 (range 35–65) years, respectively. Thirty-four and eight patients had two and three registrations of pulmonary TB, respectively. The median interval between the first and second registrations was 13 months (range 3–65), with 21 months (range 8–46) for previous cure, 26 (range 7–63) for completed treatment, 8 (range 4–16) for failed treatment and 24 (range 15–42) for default treatment. The median interval between the second and third registrations was 9 months (range 2–38). Of the 42 patients, 28 were male (67.7%) and 14 female (33.3%). Twenty-two patients (52.4%) had relapsed TB after previously successful treatment (cure = 17, completed = 5), and the rest were retreatment TB, 16 (38.1%) after treatment failure and 4 (9.5%) after default treatment. The second treatments resulted in 23 patients (54.8%) having successful treatment (cure = 21, completed = 2), five (11.9%) treatment failures and 9 (21.4%) default treatment. Five patients (11.9%) died during the course of treatment. Eight patients received third treatments. However, treatments outcomes were available for only six of these patients: two with cure, two with failure and two with death (Table 1).

Table 1. Clinical characteristics of 42 patients with recurrent tuberculosis
No.Age/Sex (years)TB treatment outcomeTime interval between
FirstSecondThirdFirst–second (months)Second–third (months)
 262/Mcompletecurenot-available43 2
2542/Mdefaultcurenot-available6 7
2843/Ffailurefailurecure6 8
3665/Mfailurefailurefailure3 3

Resistance to anti-tuberculosis drugs

Table 2 summarizes anti-TB drug resistance among M. tuberculosis isolates from recurrent TB patients in relation to treatment outcomes. In the first, second and third registrations, monodrug resistant strains from successful and unsuccessful TB treatment comprised 7.3% (3/41), 5.0% (2/40) and 12.5% (1/8), respectively. Polydrug resistant strains after unsuccessful TB treatment comprised 4.9% (2/41), 2.5% (1/40) and 25.0% (2/8), respectively. MDR strains after unsuccessful TB treatment comprised 12.2% (5/41), 20.0% (8/40) and 12.5% (1/8), respectively. We found one MDR-TB strain in a case of successful TB treatment (2.5%). In patients with retreatment TB after unsuccessful treatment (failure and default), MDR (P = 0.025) and drug resistance to isoniazid (P = 0.025) and to rifampin (P = 0.046) occurred significantly less frequently in first registration than in second registration M. tuberculosis isolates. However, we found no significant differences in resistance to any TB drugs in isolates from the first and second registrations of patients with relapsed TB after successful treatment (completed and cure). We found no significant differences between first and second registrations in resistance of M. tuberculosis to ethambutol (P = 0.157) and to streptomycin (P = 0.564) in patients with retreatment after unsuccessful treatment (failure and default). Obviously, we more commonly found acquired MDR-TB strains in patients with failure (6/16) and default treatment (2/4), and rarely in those with successful treatment (1/22). Interestingly, two acquired MDR-TB strains from patients with retreatment after treatment failure developed resistance to all anti-TB drugs tested including RMP, INH, SM and EMB.

Table 2. Development of anti-TB drug resistance among M. tuberculosis isolates from recurrent TB patients in relation to treatment outcomes
DST resultsResistant isolates
First registration Number (%)Second registration Number (%)Third registration Number (%)
  1. DST, drug susceptibility testing.
Successful treatment
Monodrug resistance3 (7.3)2 (5.0)1 (12.5)
Isoniazid2 (4.9)1 (2.5)1 (12.5)
Ethambutol1 (2.4)1 (2.5)0
Polydrug resistance000
MDR01 (2.5)0
Unsuccessful treatment
Monodrug resistance3 (7.3)2 (5.0)1 (12.5)
Isoniazid1 (2.4)1 (2.5)0
Ethambutol2 (4.9)1 (2.5)1 (12.5)
Polydrug resistance2 (4.9)1 (2.5)2 (25.0)
MDR5 (12.2)8 (20.0)1 (12.5)

Insertion sequence 6110-restriction fragment length polymorphism patterns

The IS6110-RFLP patterns of M. tuberculosis isolates can be classified into five groups as previously described [20]. Figure 1 shows sampling examples of patients with identical IS6110-RFLP patterns and two patients with non-identical patterns and Table 3 shows the numbers of recurrent TB patients with various IS6110-RFLP patterns. We found identical patterns in 40/42 isolates (95%) from recurrent patients and non-identical patterns in 2 isolates (5%) in isolates from patients with retreatment after failure. Among 40 isolates with identical patterns, 21 (50.0%) belonged to the Beijing family with 15–20 copies and 3 (7.1%) to the Nonthaburi family with 11–15 copies patterns. We were unable to group other isolates from six patients (14.3%) with heterogeneous patterns with more than five bands as either Beijing or Nonthaburi families. Five patients (11.9%) had isolates hybridized at only one position and were either 1.45 kb or 1.3 kb long; these isolates were likely to contain only a single copy of IS6110 (defined as single band pattern). We defined the isolates from five patients (11.9%) as having 2–5 bands patterns with hybridization patterns of 2–5 copies. However, the isolates of two of these patients (5%) had non-identical patterns. One of these patient's isolates had a 2–5 bands pattern in the first and a Beijing pattern in the second registration, whereas the other's isolates had the Beijing pattern in the first and a 2–5 bands pattern in the second registration. In isolates from the patients with three TB registrations, recurrent IS6110-RFLP patterns were identical patterns of Beijing family type in six patients, Nonthaburi family in one and single band pattern in one.

Table 3. IS6110-RFLP patterns of M. tuberculosis in recurrent TB patients in the first, the second and the third registrations
IS6110-RFLP patternsNumber of patients
First registrationSecond registrationThird registration
Identical patterns
Single band551
2–5 bands550
Non- identical patterns
2–5 bands110
Figure 1.

Identical and non-identical IS6110-RFLP patterns of M. tuberculosis isolates from seventeen pulmonary TB patients. Identical patterns: S1–S12 are subjects 1–12 from two registrations and S13–S15 are subjects 13–15 from three registrations. Non-identical patterns: S16–S17 are subjects 16 and 17 from two registrations.

Insertion sequence 6110-restriction fragment length polymorphism patterns and drug resistance

When we analyzed anti-TB drug resistance in relation to IS6110- RFLP patterns (Table 4), we most commonly found monodrug resistance in isolates with the Beijing family pattern (8/12). The other monodrug resistant isolates had single-band patterns (two isolates), the Nonthaburi family pattern (one) and a heterogeneous pattern (one). We found polydrug resistance in isolates with Beijing family pattern (3/5) and single band pattern (2/5). We found MDR-TB in four, eight and one isolates with identical patterns in first, second and third registrations, respectively. One isolate with a non-identical pattern had MDR-TB at the first and second registrations. We commonly found MDR-TB strains in isolates with Beijing family patterns (10/15); the rest were heterogeneous (one isolate) and 2–5 bands (four). For all treatment outcomes, the strains that had been treated on the first registration were most often the cause of recurrent TB, as shown by their identical patterns, in which the Beijing family was predominant. We found acquired MDR (resistance in those who have previously undergone TB treatment) in M. tuberculosis of the Beijing family. Of these, two isolates had primary drug resistance (resistance without a history of previous treatment) in the first registration whereas there were six MDR-TB isolates in the second registration. Four of these isolates had acquired MDR. The time intervals between the first and subsequent registrations varied.

Table 4. The relationship between anti-TB drug resistance and IS6110- RFLP patterns
IS6110- RFLP patternsResistant isolates
Mono drugPoly drugsMDR
Single band110011000
2–5 bands000000220


In this study, we analyzed anti-TB drug resistance and IS6110-RFLP patterns of M. tuberculosis strains from recurrent patients in their first, second and third TB registrations. Resistance to INH, RMP, EMB and SM occurred more frequently in M. tuberculosis strains from the second registrations than in those from the first. The MDR rate was 12.2% among new cases at first registration, 22.5% among recurrent cases with previously treated TB at the second registration and 12.5% at third registration, indicating higher rate of drugs resistance in recurrent TB. In the present study, we found resistance to INH or RMP, with or without resistance to any other drugs (MDR-TB), more commonly in patients with failure and default treatment. Moreover, two acquired MDR-TB isolates from retreatment patients after treatment failure developed resistance to all anti-TB drugs tested (RMP, INH, Sm and EMB). These strains could well develop XDR-TB eventually. Recently, it has been recognized that MDR-TB and XDR-TB are serious problems for TB control program because XDR-TB strains are virtually untreatable [22]. There is evidence that inadequate treatment is the main cause of endogenous reactivation, which is likely to occur soon after completion of treatment for the first episode [23]. However, apparently successful treatment sometimes fails to totally eradicate the bacteria from patients; they can then re-activate later. TB patients who fail primary treatment have significantly greater risks of any drug resistance or MDR-TB than do those with successfully complete treatment (treatment completion or cure) [24]. There is evidence that irregular drug administration causes development of drug resistance during short-course therapy with multiple drugs, because the drugs are taken for only a few killing cycles and regrowth occurs when the drugs stop. During each cycle, it is possible that selection of mutants that are relatively resistant occurs [25]. We found significant differences in monodrug resistance to isoniazid or rifampin and MDR among M. tuberculosis strains from primary and secondary registrations, which could be attributable primarily due to poorly administered TB treatment [1]. Recurrent TB is common in patients who have failed to respond to first and second line drugs [26]. In line with this, our findings that MDR-TB occurs most commonly in patients with failure and default treatment and rarely in those with previously successful treatment (complete and cure) imply that failure of previous treatment is associated with drug resistance. Therefore, in areas with a high prevalence of drug resistance, we recommend use of alternative regimens, especially during the continuation phase. Drug resistant TB is a man-made phenomenon; inadequate or poorly administered treatment regimens can allow drug-resistant strains to become dominant in patients with TB [1].

In general, there is a very high rate of unexplained recurrent TB in areas with a high incidence of TB [27]. Endogenous reactivation is possibly the main cause of relapses after a period without TB and recurrent TB requiring retreatment in Chiang Rai province, an area with high prevalence of TB and a high proportion of drug resistance [10]. In this study, endogenous reactivation (30/42, 71.4%) was the major cause of recurrent TB either from relapse or retreatment as evidenced by identical IS6110-RFLP of Beijing (21/42, 50%), Nonthaburi (3/42, 7.1%) and heterogenous (6/42, 14.3%) patterns in a large proportion of isolated strains. The classical IS6110-RFLP method fails to adequately differentiate M. tuberculosis strains with identical patterns of low copy numbers of IS6110 with 2–5 band (5/42, 11.9%) and single band patterns (5/42, 11.9%) [28]. Other strain typing methods such as VNTR typing are required to infer epidemiological linkage between low-copy number isolates [28]. However, VNTR typing systems cannot define all unique isolates. If the primary genotype is IS6110-RFLP, VNTR typing is certainly useful as a secondary means of typing M. tuberculosis with small copy numbers of IS6110. [28]. It is possible that the identical IS6110-RFLP patterns that we found in first and subsequent TB registrations did not truly represent endogenous reactivation because these patterns appearances may have reflected the duration of the study, incidence of M. tuberculosis strains in the population and prevalence of dominant strains. Therefore, we recommend further comprehensive investigation of the prevalence of IS6110-RFLP patterns among M. tuberculosis strains from this set of patients during 1997 through 2006 to determine the proportion of these strains in recurrent TB.

The non-identical IS6110-RFLP patterns observed in two retreatment patients after failure could be caused by either exogenous re-infection (2/42, 5%) with new M. tuberculosis strains or mixed infection, which reportedly occurs after successful treatment [29], [30] and even during treatment [30]. Exogenous re-infection plays a dominant role in the pathogenesis of post-primary TB in areas with a high incidence of the disease such as in South Africa and China [30], [31]. Simultaneous infection with multiple strains of M. tuberculosis can cause exogenous reinfection of patients: this provides further evidence for occurrence of reinfection [32].

We found that the Beijing family IS6110-RFLP pattern predominated among strains causing recurrent TB (50%). The rest of the isolates belonged to the Nonthaburi family, were heterogeneous, or had 2–5 band or single band patterns. These findings are quite different from those of a previous study in which the Beijing family was not the predominant pattern in northern Thailand, comprising only 17.7% of isolates) [13]. The prevalence rates of Beijing strain in other geographical areas is diverse, for example being 42% in Bangkok, 27.9% in central and 31.3% in western Thailand [13]. However, such discrepancies may be attributable to previous studies including all forms of pulmonary TB without differentiating between different histories of anti-TB treatment. In this study, the major patterns in MDR-TB belonged to the Beijing family, the rest being heterogeneous and having 2–5 bands. Additionally, isolates with acquired MDR evidenced identical IS6110-RFLP patterns of the Beijing family, suggesting its strong association with response to treatment [14], [15]. In Germany, where the incidence of TB is steadily decreasing and the estimated overall percentage of MDR-TB is less than 3% of all TB cases, researchers have found similar evidence for Beijing genotypes among MDR-TB strains [33]. A study in Thailand between 1996 and 2007 proposed a classification of ancestral and modern Beijing sublineages based on the VNTR in which they identified 78.8% as modern Beijing strains and the remaining 21.2% as ancestral Beijing isolates [34]. Although researchers have further analyzed such data and combined it with those of previous studies to construct a comprehensive phylogenetic tree, they have not assessed linkage between TB treatment history and development of drug resistance [34]. Therefore, we recommend further comprehensive investigation of the genetic diversity and evolution of the Beijing genotype in M. tuberculosis isolates from patients with recurrent TB.


We thank the participating patients for their kind participation in this study and staff of the Chiang Rai provincial hospital and the National TB Reference Laboratory, Bureau of Tuberculosis, Bangkok for their technical support. This work was supported by Health and Labor Science Research Grants for Research on Emerging and Re-emerging Infectious Diseases (H17-shinko-021 and H20-shinko-014), Ministry of Health, Labor and Welfare, Japan; the TB/HIV Research Project, Thailand, a collaborative research project between the Research Institute of Tuberculosis and the Japan Anti-tuberculosis Association; and the Faculty of Tropical Medicine, Mahidol University.


There is no conflict of interest for any of the authors of the manuscript caused by financial, commercial or other affiliations.