molecular epidemiology of tuberculosis in viet nam (2003-2009)

MINISTRY OF EDUCATION
AND TRAINING
MINISTRY OF HEALTH
NATIONAL INSTITUTE OF HYGIENE AND EPIDEMIOLOGY



NGUYEN THI VAN ANH



MOLECULAR EPIDEMIOLOGY OF
TUBERCULOSIS IN VIETNAM (2003-2009)



Major :Public Health
Code : 62.72.03.01


SUMARY OF PhD THESIS IN PUBLIC HEALH





HANOI - 2012

The thesis has been done at the National Institute of
Hygiene and Epidemiology

Scientific supervisors:
1. Assoc. Prof. Dang DucAnh,PhD.
2. Prof. Tran Van Sang, PhD.

Referee 1: Assoc. Prof. Doan HuyHau, PhD., Military Medical
Institute

Referee 2: Assoc. Prof. Bui Duc Duong, PhD., Ministry of Health

Referee 3: Assoc. Prof. Nguyen Thi Tuyen, PhD., Hanoi Medical
University

The thesis will be defended at the InstitutionalScientific Committee
held at the National Institute of Hygiene and Epidemiology.
At 9 am, September 19, 2012



The thesis can be found at:
1. The National Library
2. The National Institute of Hygiene and Epidemiology
3.


ABBREVIATION

AFB Acid Fast Bacillus
AIDS Arquired Immuno-deficiency Syndrome
CAS Central Asian

CR Clustering rate
DR Direct Repeat
EAI East African Indian
EMB Ethambutol
INH Isoniazid
IS Insertion Sequence
LAM Latin American and Mediterranean
LSP large-Sequence Polymorphisms
MIRUs-
VNTR
Mycobacterial Interspered Repetitive Units - Variable
Number of Tandem Repead
MPTR Major Polymorphic Tandem Repeat
SNP Synonymous Single Nucleotide polymorphisms
PAUP Phylogenetic Analysis Using Parsimony
PCR Polymerase Chain Reaction
PGRS Polymorphic GC-rich Repetitive Sequence
RFLP Restriction Fragment Length Polymorphism
RMP Rifampicin
SM Streptomycin
SpolDB4 Spoligotyping database 4
ST Spoligotype
VNM Vietnam
WHO World Health Organization
1
INTRODUCTION

Tuberculosis (TB) is still the second leading cause of death among
infectious diseases, causing about 2 millions deaths per year globally.
According to the 2010 report from WHO, Vietnam ranked the 12

th
in the 22
TB highest burden countries in the world. In the Western Pacific Region,
Vietnam ranked the 3
rd
after China and the Philippines in the number of
prevalence cases as well as the annual new TB cases.
Effective standard regimens rapidly reduced TB incidence in
developed countries. Therefore,85% of the total number of TB cases in the
world is now in developing countries. In fact, due to the increase of
populations, political upheavals, migrations, the increase of drug resistant
TB and the development ofHIV/AIDS epidemic,the number TB cases at the
presence is higher than at any time in the human history.
However,it is profoundly disturbing that the means for combating TB
are a 100-year-old diagnostic test, a vaccine that was developed 80 years
ago, and drugs that have remained fundamentally unchanged for the past 50
years. These facts underlie the call for new tools.
Molecular epidemiology of TB is important in this circumstance
because it can providescientific fundaments for the development of new
tools and the establishment of more appropriate and effective control
strategies. The scientific fundaments include how the differences in M.
tuberculosispopulation structures in different regions can effect the TB
control and prevention measures including TB treatment,development of
new vaccines, anti-TB drugs and diagnostic tools;if there are differences
betweenM. tuberculosis isolates in children and those in adults; whether
these differences can affect vaccine developments; the role of different M.
tuberculosis lineages and genotypes in TB epidemiology in different
geographic regions; The transmission dynamic of TB in populations; Risk
factors for TB transmission, which ones can be prevented.
Studies of molecular epidemiology of TB have been carried out in

many parts of the world. In Vietnam, there have been a number of studies
using molecular techniques for classification and determination of
molecular characteristics of M. tuberculosis to investigate TB situation, the
distribution of Beijing genotypes and its association with drug resistance.
In our study,different aspects of molecular epidemiology of TB in Vietnam
wereinvestigated based on the analyses of the causative bacteria and
epidemiological data. The most up to date molecular tools and powerful
statistics were used in order to obtain the following aims:
2
SPECIFIC OBJECTIVES
1. Characterize the lineage, sub-lineage distribution of M. tuberculosisin
some provinces in the Northern plain and of M. tuberculosis isolates
from national hospitals, during 2003-2009.
2. Determine the drug resistance levelsof these different M.
tuberculosislineages; determine lineage distribution in M. tuberculosis
populations of different drug resistance levels,from national hospitals,
during 2005-2008.
3. Investigate molecular characteristics of M. tuberculosis isolates from
children, during 2005-2009.
STUDIED ISSUES
- Molecular biodiversity of M. tuberculosis insome provinces in the
Northern plain.
- Discovery of TB re-infected cases.
- Differences in molecular epidemiology of TB at the 3 studied sites in
the North, the Centre and the South.
- Drug resistance levels of different M. tuberculosis lineages at the 3
studied sites in the North, the Centre and the South.
- Lineage and genotypic distribution of M. tuberculosis population of
different drug resistance levels.
- Association of different drug resistancesin M. tuberculosis isolates in

Vietnam.
- Characteristics of M. tuberculosis isolates from children, compared to
those from adults.
NEW CONTRIBUTIONS IN THE THESIS
- The description of the lineage/ sub-lineage distribution of M.
tuberculosisthat were circulating in the Northern plain of Vietnam and M.
tuberculosis isolates from hospitals at national level, during 2003-2009.
- The effects of location (urban vs. rural), region (the Northern plain, the
North, the Centre and the South) and sampling (hospital vs.
population) on the distribution of M. tuberculosis.
- The discovery of TB re-infection in TB treatment and control.
- The drug resistance levels of different M. tuberculosis lineages in the 3
studied sites in the North, the Centre and the South; The differences in
population structuresof M. tuberculosis of different drug resistance
levels; The specific association of different drug resistances in M.
tuberculosis isolates in Vietnam.
- The level of drug resistance and molecular characteristics of M.
tuberculosis isolates from Children,the differences between them and
those from adults.
3
THESIS STRUCTURE

Introduction: 3 pages
Chapter 1. Literature review: 58 pages
Chapter 2. Materials and methods: 24pages
Chapter 3. Results: 39 pages
Chapter 4. Discussion: 40 pages
Conclusion: 2 pages
Proposition: 1 page

Chapter 1 - LITERATURE REVIEW
1.1.TB epidemiology
TB is caused by Mycobacterium tuberculosis. It is highly aerobic,
requires high levels of oxygen and can grow intra-cellular.M. tuberculosis
divides every 12-18 hours, which is extremely slow compared to other
bacteria. The diseaseis transmittedmainly through respiratory tract,rarely
through oral or skin contacts. The main sources of transmission are TB
patients with AFB smear positive.
Every year, there are about 9 million new TB cases and 2 million of
deaths caused by TB in the world, of which 0.4 million are HIV positive.
TB presents in all countries but is most prevalent in Africa (30%) and Asia
(55%). Eighty percent of the total cases are in the 22 TB highest burdern
countries. The number of annual new MDR-TB is about 0.4-0.5 million,the
prevalent numberis 2-3 times higher.Up toJanuary2010,XDR-TB cases were
reported from 58 countries.
Vietnam ranked the 12
th
in the 22 TB highest burden countries in the
world and ranked the 3
rd
after China and the Philippines in the Western
Pacific Region. In 2010,per 100.000 populations, there were36 deaths
caused by TB, 333 TB cases of all forms and 200 new TB cases. The
detection rate was 54%. The rate of HIV positive among TB patients was
3.2% in 2002.In 2007, it was 4.4%. In 2002, the rate of resistance to any
drug was 32.5%,the rateof MDR was 3%in new TB patientsand 23.5% in
previously treated patients.In 2009, the rate of MDR was 2.7% in new
patients and 19% in previously treated patients.
1.2. Population structure of M. tuberculosis
1.2.1. Genetic structure of M. tuberculosis chromosome

Repeat elements, single nucleotid polymorphisms, large sequence
polimorphisms in M. tuberculosisare important markersfor classification of
M. tuberculosis. Insertion sequences (IS)are small fragments of less than
2.5 kb commonly found in M. tuberculosischromosome.Polymorphic GC-
4
rich repetitive sequences (PGRS) are sequences carry a number of
tandem repeat 96 bp GC-rich sequences, presentin at leat 26 locations.
Direct repeat (DR) region consists of variable number of 36 bprepetitive
sequences separated by non-repetitiveunique spacer sequences with length
varies from27 to 41 bp.Major polymorphic tandem repeats(MPTR) consist
of 10 bp repetitive sequences separated by 5bp spacer sequences. The number
of MPTR can be up to 80 copies. Mycobacterial interspersed repetitive units of
variable number of tandem repeats (MIRU- VNTR)consist of variable number
of 40-100 bptandem repeat units,interspersed in M. tuberculosis chromosome.
Single nucleotide polymorphism (SNP)is a variation occurring when a single
nucleotide differs between different individuals. Large sequence
polymorphism (LSP)is a variation created by insertion, deletion or re-
arrangement of a large sequence.
1.2.2. Lineage distribution in M. tuberculosispopulation structures
All research on M. tuberculosis population structureshad the same
result that the bacteria havelineage distribution in population structures.
1.2.3. M. tuberculosishasa global geographicalstructure
Lineage distributions of M. tuberculosis are different in different
geographic areas.
1.3. Molecular epidemiology of TB – Applications and achievements
1.3.1. Definition of molecular epidemiology of TB
Molecular epidemiology is a field that has emerged from the
integration of molecular biology, clinical medicine, statistics and
epidemiology. In essence, molecular epidemiology is the use of a
multidisciplinary approach for identifying factors that determine the

causation, propagation, dissemination, spatial and chronological
development of the disease. It focuses on the role of genetic and
environmental risk factors, on disease etiology and distribution among host
populations at molecular level.
1.3.2. The development history of molecular epidemiology of TB
M. tuberculosisgenotyping has been applied since the 90s of the last
century. At the beginning it was used for investigations of TB outbreaks
and institutional TB transmission. After that applicationswereexpanded to
large populations and inter-populations for the determination of recent
transmission and risk factors.
Molecular epidemiology of TB was then applied for investigations of
laboratory cross contaminations, rates of relapses and re-infections among
previously treated patients, infectiousness of AFB smear (-) patients,chains
of transmission,and transmission of drug resistant or highly virulent strains.
5
It was also used forcontact tracing,studying population structure of M.
tuberculosis, the association between population structure changes
andenvironmental, climate anddemographic changes as well as applications
of control measures.
1.3.3. Applications and achievements in molecular epidemiology of TB
1.3.3.1. ControlTB bymolecular epidemiology of TB:Molecular
epidemiology of TB provides information on TB situation, speculatesits
future development and determines risk factors for TB transmission.
1.3.3.2.Determination of TB transmission indexes:Discoverchains of
transmission by moleculartyping, then determine TB transmission indexes.
1.3.3.3. Investigation of contacts,sources and routes of
transmission:Identical or similar genotypes are the clues for identifying
cases in transmission chains. This can help investigate contacts, sources
and routes of transmission.
1.3.3.4. Investigation of transmission capacity of drug resistant strains:

Transmission capacity of drug resistant strains can be determined by the
frequency thatthe strains occur in molecular clusters.
1.3.3.5. Tracing particular strains or strains causing outbreaks:Particular
strains or strains causing outbreak can be tracted by molecular markers.
1.3.3.6. Investigation of multiple infections:A TB patients can be infected
with more than one M. tuberculosis strainsat the same time. This is multiple
infection. Multiple infection can be determined by molecular typing.
1.3.3.7. Estimation of TB burdencaused by re-infection: Re-infection is
the casethat is re-infected with a new M. tuberculosis strain after being
infected with another strain previously. Re-infection can be determined by
molecular typing.
1.3.3.8. Investigation of M. tuberculosispopulation structuresand the
importanceof different lineages:Molecular typing could identify main M.
tuberculosis lineages that are circulating in the world such asBeijing,
Haarlem, CAS, EAI, LAM, Manu, S, T, X,and sub-lineages that are
specific for countries or regions such as: Manila, Delhi, (VNM) Việt Nam,
Northaburi, SOM (Somali), BDG (Bangladesh) etc Different lineages
and sub-lineages affect TB situationin each countryorregion
differently.Their importance can be investigated by their ablility to spread
in the populations which is expressed by their proportions and clustering
rates.
1.3.3.9. Development of diagnosis tools:The differences in individual
genetic structuresneed to be considered when selecting molecular markers
or antigens for the development of diagnosis tools. Sensitivity or specificity
6
can be affected by mutations or molecular diversity of M. tuberculosis
population.
1.3.3.10. Development of anti-TB drugs:M. tuberculosis population
structures also needto be taken into account when looking for new and
potential anti-TB drugs. It is necessary to discover lineages, sub-lineage

that may contain SNPs, LSPsthat cause natural resistancesto the drugs in
M. tuberculosis.
1.3.3.11. Development of new vaccines:Vaccine efficacy in
differentgeographic areas can be different because local M. tuberculosis
populations and immunogenicity of the local host populations are not the
same.
1.4. Genotyping techniques used forstudies ofmolecular epidemiology
of TB
1.4.1. IS6110- RFLPtyping
Restriction enzyme PvuII is used to cut M. tuberculosis chromosome at
a site onall IS6110. The restriction fragments are then hybridized with
probes. Hybridized fragments are visualized by exposedto an optic film.
Theresulted RFLP pattern used for classification is specific for each M.
tuberculosis strain. RFLP-IS6110 typing has been considered the gold
standard for M. tuberculosis classification. However, the technique is
complicated, labour intensive and time consuming. It can not be used in
quick respond. In additon, it is not efficient for the classification of strains
with low copy number of IS6110.
1.4.2. Spoligotyping
Spoligotyping is based on the detection of 43 specific spacers in the
DR region. The spacers are amplified by PCR using Dra and Drb primers,
of which Dra is biotinylated.PCR products are then hybridized on a
membrane coated with 43 oligonucleotides that are supplementary with the
43 spacers. The spacers presenting in M. tuberculosis strains hybridized
with the oligos on the membrane can be visualized by exposing the
membrane with an optic film. Spoligo pattern (spoligotype) is specific for
each strainand can be used for classification. Spoligotyping is fast, simple
and particularly effective foridentifying Beijing strains. However, its
discriminatroty power is lower than IS6110-RFLP.
Internationalpoligotypingdatabases and softwares for ananlysis include

SpolDB4,SITVIT WEB,SPOTCLUST,SpolTools.
1.4.3. MIRU- VNTRtyping
Repetitive sequence of each MIRU-VNTR is amplified and detected by
agarose electrophoresis.Based on the size of the PCR product, the repeat at
each MIRU-VNTR can be determined. MIRU-VNTR type is presented bya
7
number of 12, 15 or 24digit-format if using 12, 15 or 24 MIRUs-VNTR for
typing.Each digit of the MIRU-VNTR type corresponds to the repeat at each
MIRU-VNTR.Sometimes, the repeat at one MIRU-VNTR can be equal to or
more than 10, leading to 2 digits for one MIRU-VNTR. MIRU-VNTR typing
has high discriminatory power. MIRU-VNTR combined with spoligotyping
can replace the gold standard IS6110-RFLP typing. Softwares for ananlysis
include: MIRU-VNTRplus,eBURST,Bionumerics and PAUP.

Chapter 2 - MATERIALS AND METHODS
2.1. Studied subjects.place and time
The studied subjects are 5 collections of M. tuberculosis isolates. The
first collection includes221 isolates from 8 district TB prevention centres of
HaTay andHaNoi provinces and from a general population of Hung Yen
province. The second collection includes 300 isolates randomly selected
(usingcontinuous patients’ hospitalizedregistration number) from the three
national hospitals for TB treatment in the three main regions: the North, the
Centre and the South of Vietnam.The third collection includes400 isolates
selected by drug resistance criteria from the two national hospitals for TB
treatment in the North and the South. The fourth collection includes 19
isolates from children, and the fifth collection includes7 paired isolates
collected before and after 6 months of treatment fromthe same patients in
Chuong My, Ha Tay. All the studied isolates were collected during 2003-
2009.
2.2. Methodologies

2.2.1. Study design
The study is a descriptive study using specialized softwares for the
analyses of molecular genotyping results. Genotyping techniques used for
this study include spoligotyping and MIRU-VNTR typing.
2.2.2. Sample size
The formula for calculating the sample size for the estimation of the
proportion of isolates belonging to Beijing lineage(an important lineagein
TB epidemiologyin Vietnam). The same sample size was applied for
objective 1 and 2 of the study.

n: Minimum sample size
ε = 10% (Relative error)
Z (1-/2) = 1.96 (Confidence coefficient withlevel
of confidence of 95%)
P = 54% (The proportion of Beijing isolates in
Vietnam. Anh D. D.Emerg Infect Dis, 2000)
8
According to the formula, the minimum sample size is 328 for each
studied objective.
2.2.4. Studied variables and outcome measures
2.2.4.1. Studied variables
- M. tuberculosis lineages and spoligotypes based on spoligotyping.
- MIRU-VNTR types based on MIRU-VNTR typing.
- Drug susceptibility profiles based on drug susceptibility testing for 4
anti-TB drugs used as first line drugs: RMP,INH, SM and EMB.
2.2.4.2. Outcome measures
* To meet the fist objective:“Characterize the lineage, sub-lineage
distribution of M. tuberculosis in some provinces in the Northern plain and
of M. tuberculosis isolates from national hospitals, during 2003-2009”
- Lineage, sub-lineage distribution in M. tuberculosis populations.

- The association between the proportion of different lineages and
epidemiological factors to identify the factors that associate with the
transmission of different M. tuberculosis lineages.
- The association between molecular clustering rates and epidemiological
factors to identify the risk factors for the transmission of TB.
- Molecular clustering rates of different M. tuberculosis lineages, sub-
lineages in function of epidemiological factors to investigatetransmission
capacity ofdifferent M. tuberculosis lineages, sub-lineages.
- Cases of TB re-infection.
* To meet the second objective: “Determine the drug resistance levels of
these different M. tuberculosis lineages, determine lineage distribution in
M. tuberculosis populations of different drug resistance levels, from
national hospitals, during 2005-2008.”
- The level of drug resistances of different M. tuberculosis lineages in
function of epidemiological factors to identify the most resistant
lineages and the risk factors for drug resistances.
- Lineage distribution in M. tuberculosis populations of different drug
resistance levels in function of epidemiological factors to identify the
evolution of population structures under drug presence pressure.
- The association of different drug resistances with each other in M.
tuberculosis population in Vietnam in function of location and M.
tuberculosis lineages.
* To meet the third objective: “Investigate molecular characteristics of M.
tuberculosis isolates from children, during 2005-2009.”
- Molecular and drug resistance characteristics in M. tuberculosis
isolates from children.
- Differences between M. tuberculosis isolates from children and M.
tuberculosis isolates from adults.
9
2.2.5. Techniques used in the study

2.2.5.1.Data collection
Information onM. tuberculosis isolates was extracted from patients’ clinical
records and laboratory test records.
2.2.5.2. Culture of M. tuberculosis isolates
M. tuberculosis isolates were cultured from sputum specimens using
modifiedPetroffmethod.
2.2.5.3. Drug susceptibility testing
Drug susceptibility of M. tuberculosis isolates was determined by
Canetti-Grosset proportion test.
2.2.5.4. Spoligotyping
Spoligotyping was done following the international standard
procedure.

2.2.5.5.MIRU-VNTR typing
12- and 15-MIRU-VNTR typing was done following the international
standard procedures.
2.2.6. Data analyses
NJ phylogenic tree andminimum-spanning tree were build by MIRU-
VNTRplus based on MIRU-VNTR typing. A molecular cluster was defined
as a cluster of two or more isolates having identical genotypes by both
spoligotyping and MIRU-VNTR typing.
Recent transmission can be approximated by the clustering rate (CR).
CR = (n
c
-c)/n where n
c
is the total number of clustered isolates, c is the
total number of clusters and n is the total number of isolates.
M. tuberculosis isolates were classified into three groupscorrespond to 3
values of lineage variable: “Beijing”,“EAI4-VNM” and “Others” (i.e. other

than Beijing and EAI4-VNM)or “Beijing”, “EAI” and “Others” (i.e. other
than Beijing and EAI).
Multivariate logistic regression (with logit links) and variance analyses
were used to investigate how (i) age, (ii) sex of the patient, (iii) location
(urban vs. rural), and (iv) sampling strategy (population-based vs. hospital-
based) can affect (i) the proportion of isolates of different genotypes and
(ii) the proportion of clustered isolates of different genotypes.
Association of different drug resistances in M. tuberculosis was
analysedby sorting the isolates based on their drug resistance profiles to 4 first
line drugsRMP,INH, SM and EMB. The sorting was conducted usingthe
method applied for community ecology (Atmar& Patterson, 1993 and
Almeida-Netoet al 2008). The statistics of the resulting pattern is
computed and its significativity is calculated by random permutations of
10
the resistances for each individual. One thousand of such permutations
proved to be enough to reach stationarity in the statistics distribution, from
which we could calculate the p-value of our observed pattern.
All the statistical analyses were performed with the R software (R
Development Core Team, 2010).
2.3. Study limitation
The subjects of the study wereM. tuberculosis isolates. Because the
capacity for culture of M. tuberculosis isolates is still limited in Vietnam, the
studied isolates mainly came from hospitals.
2.4. Ethics
The study was based on the analyses of M. tuberculosis isolates from
TB patients.It thus did no harm to the patients. All patients’ information
was kept confidential. The study was conducted at highest-level research
and specialized institutions in Vietnam and France, by qualified researchers
and based on international standardized techniques. Thus, the quality of the
study can be assured.

Chapter 3 - RESULTS
3.1. Molecular biodiversity of M. tuberculosis isolates circulating in
some provinces in Northern plain and of M. tuberculosis isolates from
national hospitals, during 2003-2009
3.1.1.Molecular biodiversity of M. tuberculosis isolates from some
provinces in Northern plain
Table3.2. Lineage/sub-lineage distribution based on spoligotyingof M.
tuberculosisisolates from some provinces in Northern plain


Hospitals

Population

Lineage

Sub
-
lineage

Urban


# (%)

Rural

# (%)

Rural

# (%)

T
otal

# (%)


Beijing

33 (58
.
9)

30 (28
.
8)

19 (31
.
1)

82 (37
.
1)


Beijing
-

Like

2 (
3
.
6)


1 (1
.
6)

3 (1
.
4)

Beijing


35 (62
.
5)

30 (28
.
8)

20 (32
.
8)

85 (38
.
5)


EAI1_SOM


3 (2
.
9)


3 (1
.
4)


EAI4_VNM

9 (16
.
1)

32 (30
.
8)

15 (24

.
6)

56 (25
.
3)


EAI5

3 (5
.
4)

11 (10
.
6)

12 (19
.
7)

26 (11
.
7)

EAI

12 (21
.
4)

46 (44
.
3)

27 (44
.
3)

85 (38
.
5)

H

H3

2 (3
.
6)


1 (1
.
6)

3 (1

.
4)

LAM

LAM9


3 (2
.
9)

1
(1
.
6)

4 (1
.
8)


MANU1


2 (1
.
9)

1 (1

.
6)

3 (1
.
4)


MANU2


2 (1
.
9)


2 (0
.
9)

MANU



4 (3
.
8)

1 (1
.

6)

5 (2
.
3)

11

T1

4 (7
.
1)

9 (8
.
6)

1 (1
.
6)

14 (6
.
3)


T2

2 (3
.
3)

2 (0
.
9)


T2
-
T3

1 (1
.
8)

1 (1)


2 (0
.
9)

T

5 (8
.
9)

9 (8
.
6)

3 (4
.
9)

17 (8
.
1)

X

X3




2 (3
.
3)

2 (0
.
9)

ZERO




1 (1
.
6)

1 (0
.
5)

U



2 (3
.
6)

12 (11
.
5)

5 (8
.
2)

19 (8
.
6)

T
otal


56 (100)

104 (100)

61 (100)

221 (100)

A total of 221 M. tuberculosis isolates were analysed by different
typing schemes. The spoligotyping generated 53 spoligo patterns, Ofwhich
38 patterns were found in the SITVIT WEB and 15 patterns, designated as
spoligotype nvn1 to nvn15, were not found in the database. The upper clades of
the nvn spoligotypes were defined using SPOTCLUST and revised by MIRU-
VNTRplus. We found a cluster of 6 isolates in the same district of Chuong My
(Ha Tay province) shared the same new spoligo pattern nvn7.
There were 8 lineages and 13 sub-lineages circulating in the studied
region. Beijing and EAI lineages were the most predominant. Twenty-one
clusters and 32 unique isolates were identified by spoligotyping. The most
frequent spoligotypes were SIT1 (Beijing) and SIT139 (EAI4-VNM).
The 12 MIRU-VNTR typing further divided the 21 spoligo-clusters
into 113 MIRU-VNTR profiles including 21 clusters and 92 unique
profiles. The clustering rate was 34.4%.

The 21 clusters determined by 12 MIRU-VNTR typing were then
typed by 15 MIRU-VNTR typing generating 61 MIRU-VNTR profiles
including 16 clusters (52 isolates) and 45 unique profiles. The clustering
rate reduced to 16.3%. The sizes of the large Beijingand large EAI4-VNM
clusters reduced. The isolates in each cluster of all 14 small clusters of 2-3
isolates identified after 15-MIRU-VNTR typing, were consistently found to
be from patients living in the same district. The 6 isolates from Chuong My
district, sharing spoligotype nvn7 remained in cluster.Their 15-MIRU-
VNTR type and was different from the MIRU-VNTR types of all other
strains that were typed and named as M_CMY.
MIRU-VNTR types of the isolates belonging to the 2 main Beijing and
EAI lineages were very different. The two lineages can be differentiated
based on MIRU4 (2 repeat for Beijing/ 6 repeat for EAI) or MIRU24 (1
repeat for Beijing/ 2 repeat for EAI) or MIRU26 (4-10 repeat for Beijing/
2-3 repeat for EAI). The isolates of Beijing and EAI lineages aggregated in
two main branches of the minimum-spanning tree created by MIRU-
VNTRplus,based on 12 MIRU-VNTR typing.
3.1.2. Effects of sampling and location on molecular diversity of M.
tuberculosis isolates from the Northern plain
12
Analyses showed that lineage/ sub-lineage distributions of M.
tuberculosis isolatesin hospital samples and population samples were not
significantly different. However,clustering rates (CRs) were higher in the
hospital isolates (urban: 17.9%, rural: 19.2%) than in the populationisolates
(0%) (p<0.001). The CR of Beijing isolates(25.9%) was higher than that of
EAI4-VNM isolates (10.7%) andof the isolates other than Beijing and
EAI4-VNM(10.0%) (p<0.001).
The proportion of Beijing isolateswas significantly higher in urban
areas than in rural areas (OR: 3.8 [95%CI: 1.9 -7.6]),withp<0.001. The
proportion of Beijing isolates increased when the patients’ age

decreased,the proportion of EAI4-VNM isolates increased when the
patients’ age increased withp<0.05 (figure 3.5).






Figure3.5.Proportion ofBeijing (A), EAI4-VNM (C) andOther (B)
isolates in function of patients’ age and location (urban: red, rural: blue).
Female patients (mean age: 60.7 [95%CI: 58.7- 62.7]) were older than
male ones (mean age: 45.5 [95%CI 43.5- 47.5]),with p<0.0001.Patients
from rural areas (mean age: 52.3 [95%CI: 50.3- 54.3])were older than
patients from urban areas (42.8 [95%CI: 40.8- 44.8]), withp<0.05. Patients
infected with EAI4-VNM isolates (mean age: 55.6 [95% CI: 53.6- 57.6])
were older than patients infected with Beijing isolates (51.1 [95% CI: 49.1-
53.1]),withp<0.05.
3.1.3. Molecular characteristics of paired isolates collected before and
after 6 months of treatment from same patients
Except the paired CMY-T0-14 and CMY-T6-14having the same
spoligotype,all other 6 pairs had different spoligotypes.In 5 out ofthe above
6 pairs, the isolates collected at 6 months after treatment had the same
genotype nvn7.
3.1.4. Differences in molecular epidemiology of TB in the 3 studied
sites in the North, the Centre and the South
There were 6lineages and 13 sub-lineages of M. tuberculosisin the
three studiednational hospitals in the North, the Centre and the South.
Beijing(55.4%) and EAI(27.5%) lineageswere the two most predominant
lineages. The other 4 lineages accounted for 17.1%,including Haarlem
13

(1%), LAM (1.3%), T (6.4%) and Zero (0.3%).The undesignated U strains
accounted for 8%. The sub-lineage EAI4-VNM accounted for 14.8% of the
total samples.
Theproportion of Beijinglineagewas higher in the North (70.4%) and in
the South (68%), lower in the Centre (28%).The proportion of EAI lineage
and EAI4_VNMsub-lineage higher in the Centre (59% and 27%), lower in
the North (7.1% and 6.1%) and in the South (17% and 11%). The
differences among regions were significant (p<0.01).The proportion of
Beijing isolates increased when the patients’ age decreased (p<0.001).The
proportion of EAI and EAI4-VNM isolates increased when the patients’
age increased (p<0.001).
15MIRU-VNTR typingshowed that isolates from the same regionswere
more similar than the ones from different regions and tend to aggregate
together in the minimum spanning tree (figure 3.8). MIRU-VNTR types of
isolates from the 3 regions were relatively different.
Isolates of Beijing and EAI lineages aggregated in two main branches
of the minimum-spanning tree created by MIRU-VNTRplus based on 15
MIRU-VNTR typing. There were many Ustrains embedded in the EAI branch.
Among the 3 most predominant lineages,the repeats at each MIRU-
VNTR for Beijing and EAI were less diverse and the frequency for the
presence of a certain repeat often accounted for a major proportionof
isolates compared to other repeats.For the T lineage,the repeats at each
MIRU-VNTR were more diverse and there wereoften 3-4 repeats
occurredpredominantly.

Figure3.8.Minimum-spaning tree created by MIRU-VNTRplusbased
on 15 MIRU-VNTR typing for the isolates from the North, the Centre and
14
the South. Notes:The smal circles: MIRU-VNTR types.The colors of the circles
correspond to isolates from the North (blue),the South (red) and the Centre

(yellow).The elips: the location in the minimum spaning tree, where the isolates
from the same regions aggregated. Thecolors of the elips correspond to regions:
North (blue), Centre (brown), South (red).
3.2. Molecular epidemiology of drug resistant TB
3.2.1.The levels of drug resistances of different M. tuberculosis lineages
in the North, the Centre and the South
In general, M. tuberculosisof all lineages were most resistant to drugs
in the South, and most susceptible to drugs in the Centre. Beijing lineage
was most resistant and EAI lineage was mot susceptibleexcept in the South
In the South,the level of drug resistance of EAI lineage was very high, even
higher than Beijing lineage.
Multidrug resistance (MDR) isolates from the Centre were found only
in Beijing lineage (21.4%). The proportion of MDR in Beijing isolates from
the Centre was significantly lower than that from the North (34.8%) and from
the South (36.7%),withp<0.0001.
3.2.2. Lineage distributions in M. tuberculosisisolates of different drug
resistance levels from the North and the South

Figure 3.10.Proportion of Beijing, EAI lineages andEAI4-VNM sub-
lineage in function of drug resistance levels and regions.Notes: XDR:
extensively drug resistance, MDR: multidrug resistance, SDR: single drug
resistance, SST: drug sensitive,North (red) andSouth (blue).
Results showed that, the higher the levels of drug resistances, the
higher the proportion of Beijing isolates and the lower the proportion of
EAI and EAI4-VNM isolates in the populations. The proportions of these
different isolates were, however, not significantly different between MDR
and XDR populations. Variance analysis showed that lineage distributions
inM. tuberculosispopulations of different drug resistance levels were not
significantly different between the North and the South (figure 3.10).
15

3.2.3. Genotype distributions of M. tuberculosis isolates of different
drug resistance levels
Drug supceptible strains are more genotypically diverse compared to drug
resistant strains. In the minimum spanning tree,the strains with higher levels of
drug resistance are more aggregated in the centre, the strains with lower levels
of drug resistance are more scattered and far away from the centre.
3.2.4. The association of different drug resistances with each other in
M. tuberculosis populationin Vietnam
M.tuberculosisisolates resistant to EMB are also resistant to RMP,INH
and SM.Isolates resistant to RMP are also resistant to INH and SM.Isolates
resistant to INH are also resistant to SM(figure 3.12). This pattern was
consistently found in M. tuberculosisisolates belonging to different lineages
and from different regions of Vietnam (p<0.01).


Figure 3.12.Distribution of M. tuberculosis isolates belonging to
diferent lineages, from different regionsof Vietnam in function of drug
resistance profiles. Notes:each line corresponds to one isolate. each collumn
coressponds to one drug: drug resistance (red), drug supceptible (grey).
3.3. Characteristics of M. tuberculosis isolates from children
3.3.1. Characteristics of M. tuberculosis isolates from children
All 19 M. tuberculosis isolates from children were drug susceptible.
The isolates belonged mainly to Beijing lineage (13 isolates, 68.4%) and T
lineage (5 isolates, 26.3%). A total of 4 molecular clusters were found in
this small number of isolates, including 1 cluster from Hung Yen
province,3 clusters from the National hospital for Lung diseases. The
clustering rate for the studied isolates from children was 22%.
16

Figure 3.14. UPMAphylogenic tree created by MIRU-VNTRplus forM.

tuberculosis isolates from childrenbased on spoligotying and 15 MIRU-
VNTRtyping.
3.3.1. Comparison of M. tuberculosis isolates from children to those
from adults
M. tuberculosis isolates from childrenwere not different to M.
tuberculosis isolates from adults.

Chapter 4 - DISCUSSION

4.1. Molecular diversity of M. tuberculosisisolates circulating in some
provinces in the Northern plain and M. tuberculosis isolates from
national hospitals, during 2003-2009
4.1.1.Molecular epidemiology of TB in some provinces in the Northern
plain
 Beijing lineage and EAI4-VNM sub-lineage predominant in the
Northern plain
Results showed that there were a number of different M. tuberculosis
lineages circulating in the studiedregion. The two most predominant lineages
were Beijing (38.5%) and EAI (38.5%). Other lineages accounted only for a
small proportion H (1.4%), LAM (1.8%), T (8.1%), X (0.9%) and MANU
(2.3%). The result coincided with the lineage distribution of M. tuberculosis in
East Asia as showed in theinternational spoligotyping database.
Among the sub-lineages of EAI lineage, EAI4-VNM is a special sub-
lineage for Vietnam,accounted for 67% of the total isolates belonging to EAI
lineageor 25.3% of the total studied isolates.EAI4-VNM isolates were found
with a high proportion in Vietnam (50.7% in the Mekong river delta,BuuTN
2009), but found with very small proportions or not found in other countries.
This indicated that EAI4-VNM sub-lineage is specific for Vietnam.
17
The proportion of Beijing was higher in urban areas than in rural areas.

(62.5% vs. 28.8%,p<0.0001). But this location effect was not observed on
other lineages including EAI lineage, EAI4-VNM sub-lineage and the
Other group (isolates other than Beijing and EAI4-VNM).
 Sampling effect on the rate of recent transmission
Molecular clustering rates (the rate of recent transmission) in the
isolates of different sampling strategies were significantly different, with
p<0,0001. The molecular clustering rate in the hospital isolates was higher
(urban: 17.9% and rural: 19.2%) than in the population isolates (0%).
The finding suggested that the hospital isolates maybe more virulent
than population isolates. This suggestion is supported by the hypothesis of
Anderson, 1982, build based on a prediction model that the strains with the
highest transmission capacity would also be the ones producing the most
severe symptoms. Another possible cause for the higher clustering rate in
the hospital samples could be thatthe majority of TB patients only came to
hospitals when the disease became much advanced after long delaysand
after transmitting the disease toother people. In a previous study, the average
time delay for TB diagnosis in TB patients in Vietnam was12 weeks while the
recommended time delay should not be more than 3 weeks.
 Beijing lineage was emerging in the population and
replacingEAI4-VNM sub-lineage
The Beijing lineage was found to be associated with younger age.The
proportion of Beijing isolates increased when the patients’ age decreased.
This finding coincided with the findings of the previous studies in Vietnam
including a study conducted in the Mekong river delta.This confirmed that
Beijinglineage was emerging in the population in both the North and the
South of Vietnam.
Beijinglineage was emerging through modern migrations and travels.
In the study on the rural areas of Mekong river delta, Buu TN et al., 2009
suggested that Beijing isolates were spread from Ho Chi Minh city to rural
areas. They found that the proportion of Beijing isolates in the population

living along the main road leading to Ho Chi Minh city was higher than the
overall proportion of Beijing isolates in the whole studied areas.
In the contrary,the proportion of isolates belonging to EAI4-VNM sub-
lineage wasfound decreased in younger patients. This sub-lineage seems to be
disappearing in the population and replaced by Beijing lineage. EAI4-VNM
isolates were found the most susceptibleto anti-TB drugs, while in the contrary,
Beijing isolates were found the most drug resistant, the replacement of EAI4-
VNM isolates by Beijing isolates predicts a more serious TB situation in
Vietnam in the future.The association between the proportion of Beijing
18
isolates with patients’ age was not found in other Asian countries except
Bangladesh, suggested that TB epidemiology in Vietnam was very different
from the TB epidemiology in other countries in the region.
The hypothesis that Beijing lineage was invading and spreading into
Vietnam through modern migration and travel was supported by the fact
that MIRU-VNTRtypes of the 5 biggest Beijing clusters in the studied
samples were also found predominant in many other Asian countries. M11,
a previously suggested ancestral Beijing MIRU-VNTR typehighly
prevalent in many other countries was also found to be one of the most
prevalent Beijing genotype in the Northern plain, accounting for 19% of
the total Beijing strains.
 The spread of new genotype in Chuong My, Ha Tay
The fact that all 6nvn7 isolates in Chuong My, Ha Tayhad the same
MIRU-VNTR type suggested that they could have originated from a single
isolate.Genotypeof these isolates was different to that of all other studied
isolates and their spoligotype was not found in the international database.
This suggested that this genotype could be a new one. The high number of
cases caused by this genotype in the same district may reflect a possible TB
outbreak caused by this strain in the area.
 Discovery of TB re-infection

Genotyping of the paired isolates collected before and after 6 months
of treatment showed that only one pair having the same genotype,the paired
isolates of the other 6 cases were all different. Therefore, 6 out of 7 cases
having culture positive at 6 months of treatment were infected with a new
M. tuberculosis strain during treatment period (re-infection).
The above discovery may be a warning on the use of the successand
failure treatment rate for theevaluation of the effectiveness of TB control
programme and TB treatment regimens. Due to the differences in
pathogenesis between TB relapse and re-infection, if re-infection accounts
for large proportion of TB burden, this problem need to be considered
during the establishment of clinical trials and TB control and prevention
strategies. In addition, it is worth to note that infection control should not only
prevent TB transmission from TB patients to populations but also limit the
exposure of TB patients to other sources of TB infection.
 Other risk factors
In this study, TB affected male patients more than female ones (71%
vs. 29%). Male patients(mean age: 45.5) were significantly younger than
female ones (mean age: 60.7).The majority male patients contracted TB
during labour ages. This suggested that TB was often spread at working
places. The proportion of female patients in population samples (39%) was
19
higher than in hospital samples (23%). This would mean that female
patients tendto have less access to health services than male patients.
4.1.2. Differences in molecular epidemiology of the three studied sites
in the North, the Centre and the South
 Molecular distribution characteristics of M. tuberculosis in the three
studied sites in the North, the Centre and the South
In general, M. tuberculosislineages circulating inthe 3 regions were not
different. A few sub-lineages were found in one regionnot in the others.
However, the number of isolates belong to these sub-lineages were only

1or 2 and therefore not significant. The two most predominant lineages
were Beijing and EAI, but the proportions of each lineage were significant
different in the different regions.
The proportion of isolates belonging to Beijinglineage also increased
significantly when the patients’ age decreased and the proportion of
isolates belonging to EAI lineage and EAI4_VNM sub-lineage increased
significantlywhen the patients’ age increase, withp<0.001. This tendency
was observed in all three regions, suggesting that this is the tendency of the
whole country.
The proportion of Beijing isolates was high in the North (70.4%) and
the South (68%), lower in in the Centre (28%). The proportion of Beijing
isolates from the national hospital in theCentre in this study was as the
same as the proportion of Beijing isolates fromthe Centre (28%) in the
samples collected from the national survey for drug resistance. The
proportion of Beijing isolates in the samples from the national hospital in the
North was alsoalmost as the same as the proportion of Beijing isolates in the
samples from the urban areas in the Northern plain (62.5%). Thus, there was
no difference in lineage distribution between samples collected in hospitals at
national level and in hospitals at lower levels.
 The hypothesis of evolution of TB molecular epidemiology in Vietnam
Hue is still a city with not so many international activities and
exchanges. This feature would be the reason that the proportion of Beijing
in this region was still relatively low.Therefore, the collection of M.
tuberculosisstrains in Hue may bethe collection of local strains. The
proportion of isolates belonging to EAI4_VNM sub-lineage was not higher
than the proportion of EAI5 sub-lineage and not as predominant as showed
in the North and the South. In other the hand, the EAI5 lineage more
predominated in this region. This predominant distribution of EAI5 lineage
was similar to that in other countries in the region.
The proportion of isolates belong the ancestral EAIlineage increased

from India through Southeast Asia andpredominatedin Southeast Asia.In
20
each country, different EAI sub-lineages were circulating at different
levels.The spread of the EAI lineage was concomitant with ancient human
migration from Africa throughIndia to Southeast Asia.Therefore, a
hypothesis wasraised that the evolution of EAI lineage generated different
EAI sub-lineages, of which EAI5 sub-lineage became more and more
predominated during the human migration. This evolution resulted in the
increasing proportion of EAI5 sub-lineage from Africa through Indiato
Southeast Asia.That was whythe highest proportion of EAI5 sub-lineage
was found in Southeast Asia.
Beijing lineage was originated from Centre Asia. It thenspread to
North and East Asia. Thus, it seems that the spreadsof the two lineages EAI
and Beijing are meeting in Vietnam resulting in the replacement of EAI
lineage with Beijing lineage.The replacement would have happened firstly to
EAI5 sub-lineage then to EAI4-VNM sub-lineage.This explains why EAI4-
VNM sub-lineagewas still predominant in areas where Beijing isolates
accounted for a high proportion, while EAI5 sub-lineage predominated only in
the areas where the proportion of Beijing isolates was still low.
 Molecular characteristics of M. tuberculosis of the same lineages in the
three regions
M. tuberculosis of the same lineages in the different regions were
relatively different, resulting in low molecular clustering rate (4%) and
higher aggregation of isolates from the same regions compared to isolates
from different regions.
 M. tuberculosishaving lineage-specific molecular characteristics
M. tuberculosis isolates belonging to the same lineage had similar
genotypes and were very different to genotypes of isolates belonging to
other lineages. ForBeijing and EAI lineage, the frequency for the presence
of a certain repeat at each MIRU-VNTR often accounted for a major

proportion of isolates belonging to these lineages compared to other
repeats, and specific for the lineages. For the T lineage, there were several
certain repeats concomitantly predominatedat each MIRU-VNTR. The T
isolates were often scattered at the end of EAI and Beijing branches in the
minimum-spanning tree. This distribution indicates that thislineage
consisted of strains withdiverse genotypes.
4.2. Molecular epidemiology of drug resistant TB in Vietnam
 The levels of drug resistance of different M. tuberculosis lineages in the
three studied sites in the North, the Centre and the South
21
In general, Beijing lineage was mot resistant to drugs. In the contrary,
EAI lineage was most sensitive to drugs. The lowest level of drug resistance of
Beijing isolates found in the Centre indicated that big cities (Hanoi and Ho Chi
Minh city) were the sites for the invasion of not only Beijing strains but also
drug resistant Beijing strains.
The high proportion of drug resistance in the South found not only in
Beijing lineage (36.7% MDR) but also in other lineages, including EAI4-
VNM sub-lineage (45.5% MDR) even though this sub-lineage was very
sensitive in other regions.This finding suggests that isolates belonging to this
sub-lineage developed resistance to drugs while circulating in the region. In the
South, the proportion of drug resistant isolates of EAI4-VNM sub-lineage
higher than that of any other lineages.But in the North, EAI4-VNM sub-
lineagewas the most sensitive one. Thus, the high drug resistance level of this
sub-lineage in the South may be the factor leading to higher proportion of
EAI4-VNM isolates in the South compared to that in the North.
 The proportion of MDR higher in male patients than in female patients
The proportion of MDR significantly higher in male patients than in
female ones in all three regions. Thus, gender should be considered as a
risk factor in TB control. Studies in many developing countries showed that
male TB patients tend to have poorer treatment compliance compared to

female ones due to many reasons. The poorer compliance could be the
reason leading to higherlevel of drug resistance. Thus, measures should be
taken to assure treatment compliance of this subject group.
 Lineage distribution of M. tuberculosis of different drug resistance
levels in the North and the South
In general, the proportion of Beijing isolates increased steadilyin the
M. tuberculosispopulations of increasing levels of drug resistance, from
sensitive to single drug resistant and to multidrug resistant.For EAI
isolates,the difference was observed only between multidrug resistant and
drug sensitive groups, with lower proportion of EAI isolates in multidrug
resistant group. The difference between single drug resistant and sensitive
groups was not significant. However, the result could be affected due to the
low number of EAI isolates in the studied samples, compared to the
number of Beijing isolates.Significant difference would possibly be
observed if the sample size increased.
The proportion of isolates belonging to Other group decreased in
populations of increasinglevels of drug resistance, suggesting that M.
tuberculosispopulation became genotypically less diverse when the
population was more resistant to drugs. This indicated that drug resistance
occurred more frequent in certain M. tuberculosis lineages, such asBeijing
22
one. There was only a few numbersof isolates of Other group in the
multidrug resistant population.
 Changes in population structure of M. tuberculosis under the pressure
fromthe presence of anti-TB drugs
Genotypic homogeneity between MDR and XDR isolateswas higher
than that between MDR or XDR and sensitive ones. The highgenotypic
homogeneity between MDR and XDR isolates corresponded to lineage
distributionsof MDR and XDR isolates, which were also not significant
different. Result showed that the lower the levels of drug resistance, the

more genetically diverse of the M. tuberculosis populations.
The above findings guggested that M. tuberculosis evolves to adapt to
the pressure from the presence of anti-TB drugs. Population structures
evolved in a way to select M. tuberculosis isolates that can resist to drugs
generatingnewpopulations less genetically diverse but more resistant to
drugs. Due to the use of second line drugs was only recommended by WHO
since 1998, the effect of the second line drugs to the studied M. tuberculosis
population structure had not been expressed.
 Association of different drug resistance in M. tuberculosis isolates in
Vietnam.
M. tuberculosisisolates resistant to EMB are also resistant to RMP, INH
and SM. Isolates resistant to RMP are also resistant to INH and SM. Isolates
resistant to INH are also resistant to SM. This pattern was consistently found in
all M. tuberculosis lineages and in all three studied regions.
The discovery of efflux pumps suggested that in the presence of drugs
at inadequate concentration,M. tuberculosiscouldstill survive. For
adaptation, they initiate the forming of drug efflux pumps toget rid of anti-
TB drugs then develop mutations in drug resistance genes leading to high
rate of drug resistance. This ability maybe the cause for the association of
drug resistances in M. tuberculosis.Different drugs can be eliminated by the
same efflux pump, for example EMB efflux pump can eliminateRMP, INH
and SM.Similarity, RMP efflux pump can eliminate INH and SM
etc Thismechanism leads to latter mutations at drug resistance genes.
4.3.Characteristics of M. tuberculosis isolates from children
Spoligotyping showed that Beijing isolates accounted for a high
proportion (68.4%) in M. tuberculosis from children. The high proportion
ofBeijing isolates in children coincided with the finding that Beijing strains
are emerging in Vietnam. M. tuberculosis from children were still very
sensitive to anti-TB drugs. All 19 isolates from children were sensitive to
the first line anti-TB drugs RMP,INH, SM and EMB.

The study found that 8 out of the 19 studied isolates from children
were in molecular clusters, generating the clustering rate of 22%,higher
23
than the clustering rate of 4% in adult samples when applying the same
sampling strategy. This suggests that recent transmission plays an
important role in causing TB in children. The proportion of isolates having
the same genotypes was relatively high in this small number of M.
tuberculosis from children, the difference in drug susceptibility suggested
that selection pressure could have a role in causing TB in children.
Studied evidences suggested that transmission capacity of drug
resistant M. tuberculosisis lower than that of drug susceptible strains. In the
study, all isolates from children were sensitive to drugs. Among the
sensitive ones, there were certain predominant lineages, certain
predominant genotypes. This suggested that not all M. tuberculosis strain
can caused TB in children. This observation leaded to two hypotheses. One
is that immunity in children under 15 years old is better than in adults. The
other is thatBCG vaccine can protect children.Only strains with high
transmission capacity can cause disease in children. If the first one is true,
new TB vaccine should target not only children but also adults. If the
second one is true, BCG vaccinationhas to be maintained.
4.4. Develop MIRU-VNTR typing as routine molecular typing
Results showed that MIRU-VNTR typing was congruent with
spoligotyping, especially for the identification of Beijing and EAI lineages.
Thus,MIRU-VNTR typing can be used for the classification of Beijing and
EAI lineages even without the help of spoligotyping. The simplification of
genotyping procedure would promote routine molecular typing for M.
tuberculosis strains in low-income countries where Beijing and EAI
lineages are predominant like Vietnam.
Results of MIRU-VNTR typing presented very diverse M.
tuberculosispopulation in Vietnam. Thus, the typing method has high

discriminatory power and can be suitable for epidemiological investigations.

CONCLUSION

1. Lineage, sub-lineage distribution of M. tuberculosis in some
provinces in the Northern plain and M. tuberculosis isolates from
national hospitals, during 2003-2009.
- There were 8 lineages and 13 sub-lineages of M. tuberculosis
circulating in Vietnam. Beijing and EAI lineages were the two most
predominant lineages. Beijing lineage accounted for 28% to 70%. EAI
lineage varied largely from 7%to 59% in different regions. In the Northern
plain, the proportion of Beijing isolates in urban areas (62.5%) higher than
in rural areas (28.8%). EAI4-VNM sub-lineage was most predominant
24
among EAI sub-lineages in the North and the South. In the Centre, it was
not more predominant than EAI5 sub-lineage.
- There was no difference in lineage distribution betweenM. tuberculosis
isolates from hospitals and from population. However, the clustering
rates of the hospital samples (17.9% in urban and 19.2% in rural areas)
were higher than in population samples (0%).
- 6/7 studied cases having culture positive at 6 months of treatment were
re-infected.
2. Drug resistance levels of different M. tuberculosis lineages, lineage
distribution in M. tuberculosis populations of different drug resistance
levels from national hospitals, during 2005-2008.
- The levels of drug resistance were different in different M.
tuberculosislineages and in different regions (North, Centre and
South). In general, in M. tuberculosis isolates were most resistant in
the South, least resistant in the Centre, and most resistant in Beijing
lineage, least resistant in EAI lineage. In the South the proportion of

MDR varied from 33% to 45% among different lineages. In the Centre,
except Beijing lineage (21.4% MDR), other lineages were still highly
sensitive to drugs (no MDR). In the North, the level of drug resistance
in Beijing lineage was as high as that in the South. In other lineages it
was lower than that in the South but higher than that in the Centre.
- Statistic models showed that M. tuberculosis isolates resistant to EMB
are also resistant to RMP, INH and SM. Isolates resistant to RMP are
also resistant to INH and SM. Isolates resistant to INH are also
resistant to SM. This pattern was consistently found in all M.
tuberculosis lineages and in all the three studied regions.
- M. tuberculosis population became genotypically less diverse when the
population was more resistant to drugs.Population structures evolved in a
way to select M. tuberculosis isolates that can resist to drugs generating
new populations less genetically diverse but more resistant to drugs.
3. Molecular characteristics of M. tuberculosis isolates from children,
during 2005-2009.
- All 19 studied M. tuberculosis isolates from children were sensitive to
anti-TB drugs.
- Beijing isolates accounted for a high proportion (68%) in M.
tuberculosis from children. T isolates accounted for almost all the
remaining proportion (26%).
- The clustering rate in isolates from children (22%) higher than that in
isolates from adults (4%)when applying the same sampling strategy.

25
PROPOSITIONS

- Determination of TB burden should be supported by molecular
epidemiology of TB.
- Molecular epidemiology investigation should be integrated in the

national surveys for drug resistance.
- Molecular epidemiology investigation should be one of the component
activities of the National TB programme. Results should be used for
the establishment of intervention and control measures.
- Studies of molecular epidemiology of TB need to be carried out
continuously to follow the evolution of M. tuberculosispopulation
andTB transmission in Vietnam, especially drug resistant and highly
transmissible and virulent M. tuberculosis, in order to respond
spontaneously.
- Early detection of high risk groups and risk factors for TB transmission
by molecular epidemiology of TB can helpscreen TB patients and
control risk factors actively and effectively.
26
LIST OF PUBLICATIONS INVOLVED TO THE THESIS

1. Investigation of spoligotypes of Mycobacterium tuberculosis isolated
from tuberculosis patients in some areas of Ha Tay and Ha
Noi.Journal of Preventive Medicine.Vol. XVIII.No 7(99). 2008.Page.
27-34.
2. Molecular characteristics of Mycobacterium tuberculosis strains
isolated from children in the Northern plain. Journal of Preventive
Medicine. Vol. XIX.No 7(106). 2009.Page. 81-87.
3. High Prevalence of Beijing and EAI4-VNM Genotypes among M.
tuberculosis Isolates in Northern Vietnam: Sampling Effect, Rural
and Urban Disparities.PloS ONE. Accepted August 21, 2012.