Clin Chest Med 26 (2005) 217 – 231
Molecular Epidemiology: A Tool for Understanding Control
of Tuberculosis Transmission
Charles L. Daley, MD
Division of Mycobacterial and Respiratory Infections, National Jewish Medical and Research Center, 1400 Jackson Street,
Denver, CO 80206, USA
One of the primary goals of tuberculosis control
programs is to interrupt the transmission of Myco-
bacterium tuberculosis. The most effective way to
accomplish this goal is to identify and treat individu-
als who have active tuberculosis. Even in effective
tuberculosis control programs, however, M. tuber-
culosis continues to be transmitted to others, largely
because most transmission occurs before diagnosis
and initiation of therapy. The ability to track specific
strains of M. tuberculosis as they spread through a
community would greatly increase the understanding
of the transmission and pathogenesis of tuberculosis,
but, until recently, the only means of distinguishing
different strains of M. tuberculosis were drug-
resistance patterns and mycobacterial phage typing,
both of which have significant limitations. Fortu-
nately, several molecular genotyping techniques
available now allow differentiation of isolates of
M. tuberculosis for tracking strains in the community.
Epidemiologic investigations that incorporate geno-
typing of M. tuberculosis have provided important
information about the spread of tubercle bacilli by
identifying factors related to transmission and rapid
progression to disease. This article reviews how these
genotyping tools have increased the understanding of
the transmission and pathogenesis of M. tuberculosis.
E-mail address: daleyc@njc.org
0272-5231/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccm.2005.02.005
Genotyping methods and methodologic
considerations
Several nucleic acid – based genotyping methods
have been developed that allow different strains of
M. tuberculosis to be distinguished. The most widely
used method of genotyping, referred to as restriction
fragment-length polymorphism (RFLP) analysis, uses
restriction endonucleases to cleave the mycobacterial
DNA at the sites of specific repetitive sequences,
producing DNA restriction fragments of different
lengths that can be separated by gel electrophoresis
(Fig. 1) [1]. Only the genomic DNA restriction
fragments that are complementary to and hybridize
with specific probes are visible, resulting in an eas-
ily readable band pattern. Most laboratories use a
standardized protocol for RFLP genotyping of the
M. tuberculosis complex that takes advantage of a
specific, well-characterized, repetitive element, inser-
tion sequence 6110 (IS6110) [1].
Despite its widespread use, there are several
disadvantages with IS6110-based RFLP genotyping.
First, it can be done only on cultures of M. tuber-
culosis. Second, it is a slow, labor-intensive, and
technically demanding technique. Finally, it has rela-
tively poor discriminatory power for isolates that
have six or fewer copies of IS6110 and should be
supplemented with other methods such as polymor-
phic guanine-cytosine – rich genotyping or spoligo-
typing [2].
Spoligotyping is a polymerase chain reaction
(PCR)-based method that interrogates a direct repeat
sequence comprising a repetitive 36 – base-pair ele-
ment separated by short, unique, nonrepetitive se-
chestmed.theclinics.com
218 daley

M. tuberculosis Chromosomal DNA Digested DNA

1
1 2

Extract Digest
2 DNA DNA

IS6110 site Separate by gel 3

electrophoresis
1 2 1 2

Hybridization performed
with labeled IS6110

Fig. 1. IS6110-based restriction fragment length polymorphisms analysis. Depicted are two strains of Mycobacterium
tuberculosis, labeled 1 and 2. The location and number of the insertion sequence IS6110 is noted by the small black rectangles.
Step 1: Chromosomal DNA is extracted. Step 2: Extracted DNA is cleaved with a restriction endonuclease (Pvu-II). (In reality,
thousands of fragments are created.) Step 3: After digestion, the DNA fragments are separated according to molecular weight by
gel electrophoresis. (In reality, this process results in thousands of bands that are nearly confluent on the gel.) Step 4:
Hybridization with probe for IS6110 results in a gel with bands containing only the IS6110 element. The two strains can be seen
to differ in the number of bands (ie, the number of IS6110 copies in the genome) and the location of the bands.

quences [3]. By using one set of primers, all the
unique, nonrepetitive sequences, or spacers, between
the direct repeats can be amplified simultaneously.
Individual strains are differentiated by the number
and position of the spacers that are missing from
the complete spacer set. Spoligotyping has at least
two advantages over IS6110-based genotyping: (1)
smaller amounts of DNA are needed so the procedure
can be performed on clinical samples or on strains
of M. tuberculosis shortly after inoculation into liq-
uid culture, and (2) the spoligotyping results can be
expressed in a digital format [4]. Spoligotyping can
be used as either a secondary genotyping method or
as a primary genotyping method followed by another
genotyping method that has greater discriminatory
power [5,6].
A promising PCR-based method is a high-
resolution genotyping technique that characterizes
the number and size of the variable number tandem
repeats (VNTR) in each of 12 independent myco-
bacterial interspersed repetitive units (MIRUs) [7,8].
MIRU-VNTR profiling is appropriate for strains re-
gardless of their IS6110 RFLP copy number, can be
automated for large-scale genotyping, and permits
rapid comparison of results from independent labo-
ratories using a 12-digit classification system [9,10].
The Centers for Disease Control and Prevention
(CDC) will use this methodology, along with spoligo-
typing, for all initial isolates of M. tuberculosis in
the United States as part of a national genotyping
program. More studies are needed, however, to un-
derstand better the role of MIRU typing in the
molecular epidemiology of tuberculosis. In a recent
study from Quebec, Canada, 302 clinical isolates
were evaluated with three different genotyping meth-
ods: IS6110-based genotyping noted that 27% of the
isolates were clustered, MIRU noted clustering in
61%, and spoligotyping noted clustering in 77% [11].
When all three methods were used, only 14% were
clustered, closer to the percentage that would have
been expected in the population. This study provided
some insight into the evolution of genotypes in en-
demic areas and the potential for false clustering
when the wrong genotyping methodologies are used.
The genotyping method used depends on several
factors including technical capacity and the speed
with which results are needed. The genotyping
methods vary in the reproducibility of the tests and
in their ability to differentiate individual strains of
M. tuberculosis. An interlaboratory comparative
study compared several genotyping techniques [12].
Of the seven PCR-based assays, only mixed linker
 molecular epidemiology
Table 1
Reproducibility and differentiating capabilities of common
genotyping methods
No. of different
Method Reproducibility (%) genotypes (%)
IS6110 RFLP 100 84
IS6110 mixed 100 81
linker PCR
PGRS RFLP 100 70
Spoligotyping 94 61
Variable number 97 56
tandem repeats
Study analyzed 90 strains of Mycobacterium tuberculosis
and 10 nontuberculous strains.
Abbreviations: IS6110, insertion sequence 6110; PCR, poly-
merase chain reaction; PGRS, polymorphic guanine–cytosine-
rich sequence; RFLP, restriction fragment length polymorphisms.
Data from Kremer K, van Soolingen D, Frothingham R,
et al. Comparison of methods based on different molecular
epidemiological markers for typing of Mycobacterium
tuberculosis complex strains: interlaboratory study of dis-
criminatory power and reproducibility. J Clin Microbiol
1999;37:2607 – 18.
PCR and VNTR typing were highly reproducible
(Table 1). Only mixed linker PCR had discriminatory
power similar to IS6110-based RFLP analysis. Other
studies [10] showed VNTR-MIRU typing to be
slightly more discriminatory than spoligotyping.
The combination of MIRU-VNTR, IS6110 RFLP,
and spoligotyping has demonstrated the highest
specificity [9].
Regardless of the genotyping method used, in-
terpretation of the results is based on the assumption
that epidemiologically related strains will have the
same genotype pattern and epidemiologically unre-
lated strains will have different patterns. Clustering
has often been equated with recent or ongoing trans-
mission, and the factors associated with clustering
have been sought as a means to identify and target
subpopulations that have substantial ongoing trans-
mission [13]. In contrast, patients whose isolates of
M. tuberculosis have genotype patterns that do not
match any other isolates in the community are con-
sidered to be unique and likely represent disease
caused by reactivation of a latent tuberculosis infec-
tion (LTBI). Thus, genotyping allows tuberculosis
resulting from recent or ongoing infection to be dis-
tinguished from reactivation of LTBI and makes it
possible to estimate the proportion of ongoing tuber-
culosis transmission in a community.
There is not always an epidemiologic link be-
tween patients whose isolates have identical genotype
patterns. Some studies have demonstrated that clus-
219
tered cases often have no discernible contact or other
epidemiologic links among themselves, even in rela-
tively stable populations [14,15], whereas other
studies have shown that most patients do have epi-
demiologic links [16]. The amount of transmission
represented by genotypic clustering depends on the
sampling strategy and the duration of the study
[17,18]. Undersampling can bias the estimates of
the proportion of tuberculosis cases that were likely
caused by recent or ongoing transmission, and it can
bias the estimates of the risk factors associated with
clustering. Two population-based cohort studies in
San Francisco, California [19], and the Netherlands
[20] reported that the percentage of clustered strains
was high during the first 2 years and declined
thereafter. Thus, clustering based on less than 2 years
of sampling will probably underestimate the amount
of ongoing transmission. Despite its limitations, geno-
typing has provided investigators and tuberculosis
control programs new tools in which to uncover the
transmission of M. tuberculosis in our communities.
Lessons learned regarding the transmission and
pathogenesis of tuberculosis
Molecular genotyping has revolutionized the abil-
ity to track strains of M. tuberculosis as they spread
through a community. Studies using genotyping tech-
niques in combination with standard epidemiologic
investigations have provided insights into the trans-
mission and pathogenesis of M. tuberculosis and
in the process have provided important lessons for
tuberculosis control.
Infectiousness of patients
Several studies that have assessed tuberculin skin
test reactivity among contacts to cases of pulmonary
tuberculosis have documented the variation in infec-
tivity among source cases based on the bacteriologic
status of the source, the extent of disease, and the
frequency of cough [21]. These studies have docu-
mented that patients who have more extensive pulmo-
nary tuberculosis, as evidenced by cavitary changes
on the chest radiograph or the identification of
acid-fast bacilli on sputum smear examination, are
more likely to transmit M. tuberculosis to contacts.
Molecular epidemiology studies have confirmed the
variation in infectivity that exists among patients
who have tuberculosis and highlighted the infectivity
of patients who have smear-positive pulmonary
tuberculosis. For example, a single patient who had
smear-positive pulmonary tuberculosis was directly
220 daley
or indirectly responsible for 6% of the tuberculosis
cases in San Francisco during a 2-year period [22]. In
another report, investigators showed that a single
homeless tuberculosis patient who had highly infec-
tious pulmonary tuberculosis and was a regular pa-
tron of a neighborhood bar probably infected 42%
(41/97) of the contacts who were regular customers
and employees of the bar and caused disease in 14
(34%) of them. All 12 patients whose isolates of
M. tuberculosis were available had identical IS6110
RFLP band patterns [23].
Most infection-control policies and recommenda-
tions prioritize smear-positive pulmonary tuberculo-
sis over smear-negative cases, leading to the false
assumption that smear-negative cases are not infec-
tious. Several studies have demonstrated that patients
who have sputum smears that are negative for acid-
fast bacilli but culture-positive for M. tuberculosis
can transmit infection to others in the community.
Behr and colleagues [24] reported that patients who
have smear-negative culture-positive pulmonary
tuberculosis were probably responsible for 17% of
cases in San Francisco. In a recent study from Van-
couver, British Columbia [25], investigators reported
that a similar proportion of cases resulted from smear-
negative source cases. In this study, the authors also
included extrapulmonary cases of tuberculosis and
noted that the proportion of episodes of transmission
from smear-negative clustered cases increased to at
least 25%, suggesting that some transmission was
occurring from extrapulmonary cases. Pulmonary tu-
berculosis apparently was not ruled out in all of these
cases, so transmission probably occurred through
more traditional means of spread. As an illustration
of this point, investigators in San Francisco reported
that patients who have pleural tuberculosis combined
with negative sputum cultures were unlikely to gen-
erate secondary cases of tuberculosis [26].
Although the frequency of cough has been shown
to correlate with skin test reactivity among contacts
[21], genotyping has provided conflicting results
regarding the importance of symptoms in transmis-
sion. Investigators in Harris County, Texas, reported
no association between the duration of symptoms and
the size of molecularly defined clusters [27]. Cronin
and colleagues [28], however, reported that the time
from symptom onset to diagnosis was twice as long
for patients who were considered to be transmitters
than for nontransmitters. These latter data support the
belief that reducing diagnostic delays can prevent
transmission of M. tubeculosis.
The studies reviewed here have demonstrated that
the potential for transmitting tuberculosis should be
considered in all pulmonary tuberculosis patients/
suspects, particularly in settings and environments
that facilitate transmission, such as shelters, hospices,
health care facilities, prisons, and other institutional
or crowded settings [13]. It would be prudent to treat
smear-negative pulmonary tuberculosis suspects for
some period before removing them from isolation or
sending them into high-risk settings such as jails and
prisons. In addition, pulmonary tuberculosis should
be carefully ruled out in patients who have extra-
pulmonary diseases. Although international guide-
lines for the diagnosis and treatment of tuberculosis
prioritize the detection and treatment of infectious
sputum smear – positive patients [29], timely diag-
nosis and treatment of sputum smear – negative cases
should be considered when resources permit.
Exogenous reinfection and mixed infection
Molecular genotyping can determine whether a
patient who has a recurrent episode of tuberculosis
has a relapse with the previous strain of M. tuber-
culosis or exogenous reinfection with a new strain.
Although exogenous reinfection was reported before
the availability of genotyping [30], these techniques
have made the identification of reinfection easier and
more specific. Exogenous reinfection has been re-
ported in both immunocompromised and immuno-
competent persons (Table 2) [31 – 34]. In Cape Town,
South Africa, where there is a high incidence of
tuberculosis and ongoing transmission, 16 of 698 pa-
tients had more than one episode of tuberculosis.
Twelve of these 16 (75%) had pairs of isolates of
M. tuberculosis with different genotyping patterns
[34]. Exogenous reinfection is relatively uncommon
in areas that have a low incidence of tuberculosis,
such as Switzerland [35] and the Netherlands [36],
compared with high- to moderate-incidence regions
[37 – 45]. In Houston, Texas, among 100 patients
who have recurrent tuberculosis and have completed
therapy for a first episode of tuberculosis, exogenous
reinfection was reported to cause a surprisingly high
24% to 31% of the second episodes of tuberculo-
sis [46].
Some cases of suspected exogenous reinfection
might be caused by initial infections that include
more than one strain. These instances would repre-
sent repeated infections over time that lead to a mixed
infection with different strains of M. tuberculosis.
Multiple infections were demonstrated in a patient in
San Francisco [47], in two patients who worked in a
medical-waste processing plant in Washington State
[48], and among prisoners in Spain [49]. In South
Africa, a country that has a reportedly high frequency
of exogenous reinfection [34], mixed infections are
Table 2
The frequency of exogenous reinfection in selected studies by tuberculosis incidence rates
No. of patients who
have two episodes No. of patients who No. of patients who
First author/date [reference] Study location Study population TB or two isolates have genotyping have reinfection (%)
Low and moderate incidence areas (<100 per 100,000 population)
Small, 1993 [31] King’s County Hospital, AIDS patients with positive culture for 17 6 0 (0)
New York City, NY >1 y or increasing drug resistance 31 11 4 (36)

molecular epidemiology
Sudre, 1999 [32] Switzerland HIV cohort with two isolates 20 20 2 (10)
Chaves, 1999 [49] Madrid, Spain HIV-infected Spanish inmates who 11 9 2 (22)
remained culture positive for >4 mo
Bandera, 2001 [45] Lombardy, Italy TB recurrences separated >6 mo NA 32 5 (16)
Caminero, 2001 [38] Gran Canaria Island, Spain Two positive cultures >12 mo apart 23 18 8 (44)
Krüüner, 2002 [41] Tartu, Estonia Treatment failures 35 11 11 (100)
Garcia de Viedma, 2002 [40] Madrid, Spain HIV+ and HIV cases with two 172 43 14 (33)
isolates >100 d apart
El Sahly, 2004 [46] Houston, TX TB recurrences 100 100 . . . (24 – 31)
High incidence areas (100 per 100,000 population)
Godfrey-Faussett, 1994 [42] Nairobi, Kenya TB recurrences NA 4 1 (20)
Das, 1995 [37] Madras, India Recurrence or isolated positive culture 30 30 11 (37)
32 32 29 (91)
Van Rie, 1999 [34] Cape Town, South Africa Recurrent TB 48 16 12 (75)
Sonnenberg, 2000 [43] Gauteng Province, South Africa HIV+ and HIV gold miners 57 48 2 (4)
Lourenco, 2000 [39] Rio de Janeiro, Brazil HIV+ patients with multiple isolates 12 12 3 (25)
Fitzpatrick, 2002 [44] Kampala, Uganda HIV+ and HIV TB recurrences NA 40 9 (23)
Abbreviations: HIV , sero-negative for HIV; HIV+, seropositive for HIV; NA, not available; TB, tuberculosis.

221
222
common. Warren and colleagues [50], using a PCR-
based method of strain classification reported that
19% of all patients were simultaneously infected with
Beijing and non-Beijing strains and that 57% of
patients infected with Beijing strains also were in-
fected with a non-Beijing strain. These observations
indicate that simultaneous infections with multiple
strains of M. tuberculosis occur and may be re-
sponsible for conflicting drug-susceptibility results
[51] or episodes of recurrence caused by exoge-
nous reinfection.
Impact of drug resistance on transmission and
pathogenesis
Before the advent of genotyping, studies sug-
gested that isoniazid-resistant strains of M. tuber-
culosis were less likely to result in disease in animals
[52 – 54]. Mutations or deletions within the katG gene
of isoniazid-resistant strains of M. tuberculosis have
been associated with a decrease in the pathogenicity
in animal models [55]. Several molecular epidemio-
logic studies have reported that patients who have
drug-resistant strains were less likely to be in clusters,
suggesting that drug-resistant strains might be less
predisposed to being transmitted or to cause active
tuberculosis [20,56,57]. The spread of tuberculosis
involves a three-step process: transmission of bac-
teria, establishment of infection, and progression to
disease. Because genotyping studies require the de-
velopment of active tuberculosis, they cannot deter-
mine whether drug resistance influences only one,
two, or all three of the processes. Burgos and col-
leagues [58] recently reported that the number of
secondary cases generated by isoniazid-resistant
cases of tuberculosis was significantly less than the
drug-susceptible cases. This difference in the gen-
eration of secondary cases was noted regardless of
HIV status and place of birth.
These findings support the hypothesis that drug-
resistant strains are less likely than drug-susceptible
strains to result in disease. There are, however,
populations in which drug resistance is neither de-
tected nor treated effectively and where the longer
duration of infectiousness for patients who have drug-
resistant organisms treated with standard regimens
might offset the bacterium’s diminished capacity to
cause secondary cases [58]. In areas that have high
prevalence rates of HIV, the increased host suscep-
tibility, even to strains that have diminished viru-
lence, may offset bacterial differences. For example,
one multidrug-resistant strain of M. tuberculosis,
strain W, caused several nosocomial outbreaks in
New York City in the early 1990s [59]. Over the next
daley
few years, this strain was documented to have dis-
seminated widely in the community. Because poor
tuberculosis control and underlying HIV infection are
common in many areas, drug resistance may dissemi-
nate locally despite the diminished propensity of
drug-resistant strains to cause disease. In addition, it
is possible that some organisms could experience a
subsequent mutation that increases its virulence back
to its pre – drug-resistant state [60].
Contact and outbreak investigations
Conventional tuberculosis contact investigations
use the stone-in-the-pond or concentric circle ap-
proach to collect information and to screen household
contacts, coworkers, and increasingly distant contacts
for tuberculosis infection and disease [61]. Studies in
low-incidence areas such as San Francisco [22] and
Amsterdam [62] demonstrated that a relatively small
proportion (5% – 10%) of tuberculosis cases that had
identical IS6110 -based genotyping patterns were
named as a contact by the source case. One expla-
nation for these findings is that unsuspected trans-
mission of M. tuberculosis occurred and was not
easily detected by conventional contact tracing in-
vestigations. In a 5-year, population-based study in
the Netherlands, contact investigations of persons in
five of the largest clusters identified epidemiologic
links among them based on time, place, and risk
factors [20]. Tuberculosis transmission also occurred
through only short-term, casual contact that was not
easily identified in routine contact investigations.
In a more recent study [16] from the Netherlands,
patients were divided into one of five transmission
groups based on the results of contact investigations,
genotyping, and, in some cases, a second interview:
1. Clear epidemiologic links, confirmed by geno-
typing and contact tracing (24%)
2. Clear epidemiologic links, confirmed by geno-
typing and second interview but not by contact
tracing (6%)
3. Initially unclear epidemiologic links that be-
came likely after genotyping and second inter-
view (55%)
4. No epidemiologic links but genotyping indi-
cated clustering
5. Patients who were part of a different cluster
other than expected (1%)
Combining groups 1 and 2 would suggest that
at best contact investigations could identify about
30% of the clustered cases. Fifty-five percent of
 molecular epidemiology
the clustered cases had an epidemiologic link iden-
tified after the genotyping results became available
and a second interview was performed. These data
suggest that as newer, more rapid amplification-
based genotyping methods become available, this
approach might be able to improve contact inves-
tigations [63].
Genotyping has also demonstrated that even when
another case is identified through a contact inves-
tigation, the contact case may be unrelated to the
index case. For example, Marcel Behr and colleagues
[64] in San Francisco reported that 30% of case –
contact pairs had different strains of M. tuberculosis.
Unrelated strains were more common among foreign-
born, particularly Asian, contacts. Of 538 similar case
pairs in a study [65] involving seven sites in the
United States, 29% did not have matching genotype
patterns, similar to the finding in San Francisco. Case
pairs from the same household were no more likely to
have confirmed transmission than those linked else-
where. Among patients younger than 5 years of age,
15% of culture-confirmed cases and their suspected
source patient had different genotype patterns [66]. In
a recent study from South Africa, investigators
evaluated 129 households in which genotyping data
were available for more than one patient [67]. They
identified 313 patients of whom 145 (46%) had a
genotype pattern matching that of another member of
the household. These studies suggest that contact
investigations should not focus solely on the house-
hold but all settings frequented by the index case.
Even when the essential elements of tuberculosis
control are in place, ongoing transmission of
M. tuberculosis will continue until tuberculosis is
diagnosed and therapy is initiated. In a population-
based molecular epidemiologic study in an urban
community in the San Francisco Bay area, 75 (33%)
of 221 cases had the same strain of M. tuberculosis
[68]. Thirty-nine (53%) of the 73 patients developed
tuberculosis because they were not identified as con-
tacts of source-case patients; 20 case patients (27%)
developed tuberculosis because of delayed diagnosis
of their sources; 13 case patients (18%) developed
tuberculosis because of problems associated with the
evaluation or treatment of contacts; and one case
patient (1%) developed tuberculosis because of de-
lays identifying the person as a contact.
Contact tracing in the community can be ineffec-
tive in tuberculosis outbreaks if patients do not live in
stable settings and either do not know or are un-
willing to reveal the names and locations of contacts.
Fortunately, studies that incorporate genotyping are
able to provide information about the chains of
transmission in these groups [69 – 71]. A prospective
223
study of tuberculosis transmission in Los Angeles,
California, identified 162 patients who had culture-
positive tuberculosis and interviewed the patients to
identify their contacts and whereabouts [72]. Tradi-
tional contact investigations did not reliably identify
patients infected with the same strain of M. tuber-
culosis: only 2 of the 96 clustered cases named others
in the cluster as contacts. The degree of homelessness
and the use of daytime services at three shelters were
independently associated with clustering, however.
This study demonstrated that locations where the
homeless congregate are important sites of tuber-
culosis transmission.
Several studies support the idea that specific
locations can be associated with recent or ongoing
transmission of M. tuberculosis. In a 30-month
prospective, city-wide study of all tuberculosis cases
in Baltimore, Maryland, using traditional contact
investigations and IS6110-based genotyping, 46%
(84/182) of initial isolates were clustered, and 32%
(58/182) of the cases were considered to have
tuberculosis that was recently transmitted [73]. Only
24% (20/84) of clustered cases had an identifiable
epidemiologic link of recent contact with an infec-
tious tuberculosis patient. Using geographic informa-
tion system data, the 20 clustered cases, which have
epidemiologic links in geographic areas of the city
that have low socioeconomic status and high drug
use, were spatially aggregated. Therefore, in some
populations, location-based control efforts may be
more effective than traditional concentric circle –
based contact tracing for early identification of cases.
Genotyping has been particularly useful in
identifying otherwise unsuspected and undetected
transmission in the community [13]. Molecular epi-
demiologic studies have confirmed suspected and un-
suspected transmission of tuberculosis in places such
as residential care facilities [74], bars [23,75 – 77],
crack houses [78], sites of illegal floating card
games [79], schools [80,81], hospitals [82,83], and
jails and prisons [84 – 87]. Tuberculosis transmission
also has been demonstrated among groups such as
church choirs [88], interstate transgender social net-
works [89], renal transplant patients [90], from
patient to health care providers [91], and from health
care provider to patients [92,93]. Processing con-
taminated medical waste resulted in transmission of
M. tuberculosis to at least one worker in a medical
waste treatment facility [94]. Genotyping was also
used to document unsuspected bronchoscopy-related
transmission and the cross-contamination of patients
[95,96]. Without the availability of genotyping, it
would have been difficult to confirm that trans-
mission had occurred in such settings.
224
Community epidemiology and risk factors for
clustering
Tuberculosis develops by rapid progression from
a recently acquired infection, from LTBI, or from
exogenous reinfection. Most molecular epidemiology
studies have assumed that the proportion of clustered
isolates in a population estimates the amount of
recent or ongoing transmission of M. tuberculosis.
The frequency of clustering has ranged from 7%
to 32% in low-incidence areas such as Canada
[97 – 100] to 34% to 46% in urban areas in the
United States [22,28,73,101] and Europe [20,102 –
104]. Among gold miners in South Africa, 50% of
tuberculosis patients were in clusters [15], and in
Botswana 42% of the cases were clustered [105].
Whether or not clustered cases represent tuberculosis
caused by recent transmission has remained a con-
troversial point. A recent study from the Netherlands
suggests that clustered cases do, in fact, repre-
sent recent transmission and rapid progression. Of
481 patients who had tuberculosis, 29% were clus-
tered, suggesting recent transmission in 20% (using
the n-1 approach to calculate recent transmission).
The authors reported that 86% of the cases had
epidemiologic links consistent with recent trans-
mission [16]. In high-incidence areas, the frequency
of clustering has ranged from 25% in Hong Kong
[106], to 38% in India [107], to 42% to 72% in
various African populations [57,67,105].
Conventional epidemiologic methods can be used
in combination with molecular genotyping techniques
to identify the risk factors associated with recent
infection and rapid progression to disease (Table 3).
In studies in low-incidence areas, young age, being in
an ethnic minority group, homelessness, and sub-
stance abuse have been associated with recent
infection and rapid progression to disease [22,62,
73,101]. In New York City, birth outside the United
States, age of 60 years or older, and diagnosis after
1993 were factors independently associated with
having a unique strain; homelessness was associated
with clustering or recent transmission [108]. Tuber-
culosis among foreign-born persons was more likely
to result from recent transmission among those who
were HIV-infected and more likely to result from
LTBI among those who were not infected with HIV.
These data suggest that tuberculosis prevention and
control strategies need to be targeted to the large
number of foreign-born persons in New York City
who have latent tuberculosis infection. Among
foreign-born patients who have tuberculosis in Ham-
burg, Germany, risk factors for recent infection
included a history of contact tracing, intravenous
daley
drug use, alcohol dependence, asylum stay, and
unemployment [109]. Thus, the risk factors associ-
ated with recent infection and rapid progression to
disease have varied from study to study, partly
because of differences in populations, methodologies,
and definitions.
Unfortunately, there are few population-based
studies from high-incidence areas. In a study of
South African gold miners, tuberculosis patients who
had not responded to treatment at entry to the study
were more likely to be in clusters (adjusted odds ratio
[OR] = 3.41). Patients who have multidrug-resistant
tuberculosis were more likely not to have responded
to tuberculosis treatment but were less likely to be
clustered than those who have a drug-susceptible
strain (OR = 0.27) [15]. HIV infection, although
common (53.6%), was not associated with clustering.
Apparently, persistently infectious individuals who
had previously not responded to treatment were
responsible for one third of the tuberculosis cases in
this population. In a study from Cape Town, 72% of
cases were clustered, suggesting high rates of trans-
mission in the community [110].
Measuring the performance of a tuberculosis control
program
As noted previously, tuberculosis can develop
through three mechanisms: recent transmission and
rapid progression to disease, reactivation of latent
infection, or exogenous reinfection. Because cluster-
ing is considered a measure of recent transmission, a
decline in the rate of clustering could be used to
evaluate interventions aimed at reducing recent trans-
mission [111]. In an evaluation of tuberculosis trans-
mission over a seven-year period in San Francisco,
the number and proportion of clustered tuberculosis
cases declined, particularly among the native-born
population [19]. This decline was attributed to the
implementation of targeted tuberculosis prevention
and control programs such as screening high-risk
populations and implementing directly observed
therapy to ensure high cure rates. A recent study
in New York City showed that as tuberculosis case
rates fell from recent high levels, the proportion of
tuberculosis cases caused by recent transmission
dropped from 63.2% in 1993 to 31.4% in 1999
[108]. Tuberculosis was unlikely to result from recent
transmission in persons born outside the United
States. Investigators in Denver, Colorado [112], used
clustering to measure the impact of a skin testing
program among homeless persons and showed that
clustering decreased from 49% during the implemen-
 molecular epidemiology 225

Table 3
Frequency of clustering and risk factors for clustering in selected studies by tuberculosis incidence rate
First author/date N ever
[reference] Study location Study population clustered (%) Risk factors for clustering
Low and moderate incidence areas (<100 per 100,000 population)
Alland, 1994 [101] New York, NY Hospital-based 104 (38) HIV seropositive
Hispanic ethnicity
Younger age
Drug-resistant disease
Low income
Small, 1994 [22] San Francisco, CA Community-based 473 (40) AIDS
Born in the United States
Bishai, 1998 [73] Baltimore, MD Community-based 182 (46) Intravenous drug use
van Soolingen, 1999 [20] The Netherlands Country-based 4266 (46) Male gender
Urban residence
Dutch and Surinamese nationality
Long-term residence in
The Netherlands
Hernandez-Garduño, Vancouver, BC, Community-based 793 (17) Canadian-born aboriginals
2002 [97] Canada Canadian-born nonaboriginals
Injection drug users
Kulaga, 2002 [98] Montreal, QC, Community-based 243 (25) Haitian birth
Canada
Diel, 2002 [102] Hamburg, Community-based 423 (34) Alcohol abuse
Germany History of contact tracing
Unemployment
Fitzgerald, 2003 [99] Western Canada Regionally-based 944 (32) Younger age
Male gender
Pulmonary disease
Living in shelter
Drug-susceptible disease
Predisposing factors
Prior contact
Prior skin test
Blackwood, 2003 [100] MB, Canada Province-based 629 (7) Male gender
Younger age
Treaty aboriginals
Living on reserve land
Vokovic, 2003 [103] Belgrade, Random sample 176 (31) Multidrug-resistant disease
Central Serbia
Zolnir-Dove, 2003 [104] Slovenia Country-based 306 (38) Younger age
Alcohol abuse
Homelessness
High incidence areas (100 per 100,000 population)
Godfrey-Faussett, South Africa Gold miners 419 (50) Treatment failure
2000 [57] Time spent working in mines
Lockman, 2001 [105] Botswana Community-based 301 (42) Imprisonment
Narayanan, 2002 [107] Tiruvallur District, Community-based 378 (38) Identified by house-to-house
India survey
Chan-Yeung, 2003 [106] Hong Kong, China Community-based 702 (25) Permanent residents
Recent travel to mainland China
Verver, 2004 [110] Capetown, Community-based 797 (72) Smear positive
South Africa Defaulted retreatment cases
Specific community
226
tation of the program to 14% in the 4-year period
after the program.
By contrast, an 8-year study in Greenland showed
that the annual incidence of tuberculosis doubled
from 1990 to 1997, and the percentage of culture-
positive tuberculosis cases in RFLP-defined clusters
increased to 85%, reflecting microepidemics among
adults and young children in small, isolated set-
tlements [113]. Thus, genotyping was a useful in-
dicator of changes in the proportion of cases that
resulted from recent transmission and rapid progres-
sion to disease.
Geographic distribution and dissemination of
Mycobacterium tuberculosis
Genotyping has permitted the tracking of strains
of M. tuberculosis as they spread both locally and
globally. Population-based data from the San Fran-
cisco Bay area suggest that M. tuberculosis does not
rapidly transmit and spread across geographic bound-
aries and that tuberculosis control programs should
focus on transmission within well-defined areas
[114]. In fact, most clusters (66%) from the National
Tuberculosis Genotyping and Surveillance Network
in the United States were restricted to a single site
[115]. Some strains of M. tuberculosis are widely
dispersed both geographically and temporally, how-
ever, suggesting the strains are either older, more
transmissible, or are more likely than other strains to
cause disease. Data from the Genotyping Network
found that 25% of the clusters were in two sites, 5%
were in three, 2% were in four, and 1% each were in
five and six sites [115]. Further research is needed to
determine why some strains seem to be more dis-
seminated than others.
The Beijing family of strains has been detected in
high proportions among the strains in China [116],
other parts of Asia [117], the former Russian Fed-
eration [87], Estonia [118], Europe [119 – 121], and
South Africa [122] and has been associated with large
outbreaks, febrile responses [123], treatment failure
and relapse [124,125], and drug resistance [126]. The
W strain, a multidrug-resistant strain of M. tuber-
culosis that caused many cases of tuberculosis among
patients and health care workers in nosocomial
outbreaks and other institutional settings in New
York City [59,127 – 130], is a member of the Beijing
family [128]. It is unclear why the Beijing family
strains are so widely disseminated [131]. It is possible
that the Beijing genotype has a selective advantage
and is more readily aerosolized, can establish infec-
tion more effectively, or can progress more rapidly
from infection to disease [4,132]. On the other hand,
daley
it is also possible that the Beijing genotype was
introduced into multiple locations before other strains
and had more time to spread.
The future of molecular epidemiology
Molecular genotyping, in combination with con-
ventional epidemiologic investigations, has contrib-
uted greatly to the understanding of the transmission
and pathogenesis of tuberculosis and has identi-
fied inadequacies in tuberculosis control programs.
The development of new tools, such as real-time
amplification-based genotyping, should improve the
ability to genotype strains of M. tuberculosis and con-
duct effective, timely, contact and outbreak investiga-
tions. Other technologies based on the genome of
M. tuberculosis may eventually make it possible s to dif-
ferentiate strains based on important phenotypic char-
acteristics such as transmissibility and pathogenicity.
For these new tools to be used effectively, they
must be used in combination with conventional
epidemiologic investigations. With time, genotyping
should be integrated with new surveillance systems
that will allow more rapid responses to potential
outbreaks of tuberculosis. The integration of geno-
typing and new approaches to surveillance will be
particularly important in low-incidence areas of the
United States where resources are limited and early
detection of outbreaks is difficult. The CDC-funded
national genotyping network should help with the
integration of genotyping tools into standard tuber-
culosis control practices and pave the way for tu-
berculosis control in the future.
References
[1] van Embden JD, Cave MD, Crawford JT, et al.
Strain identification of Mycobacterium tuberculosis
by DNA fingerprinting: recommendations for a
standardized methodology. J Clin Microbiol 1993;
31:406 – 9.
[2] Yang ZH, Ijaz K, Bates JH, et al. Spoligotyping and
polymorphic GC-rich repetitive sequence finger-
printing of Mycobacterium tuberculosis strains hav-
ing few copies of IS6110. J Clin Microbiol 2000;38:
3572 – 6.
[3] Groenen PM, Bunschoten AE, van Soolingen D, et al.
Nature of DNA polymorphism in the direct repeat
cluster of Mycobacterium tuberculosis; application
for strain differentiation by a novel typing method.
Mol Microbiol 1993;10:1057 – 65.
[4] Barnes PF, Cave MD. Molecular epidemiology of
tuberculosis. N Engl J Med 2003;349:1149 – 56.
 molecular epidemiology
[5] Wilson SM, Goss S, Drobniewski F. Evaluation of
strategies for molecular fingerprinting for use in the
routine work of a Mycobacterium tuberculosis
reference unit. J Clin Microbiol 1998;36:3385 – 8.
[6] Soini H, Pan X, Amin A, et al. Characterization of
Mycobacterium tuberculosis isolates from patients in
Houston, Texas, by spoligotyping. J Clin Microbiol
2000;38:669 – 76.
[7] Supply P, Lesjean S, Savine E, et al. Automated high
throughput genotyping for study of global epidemi-
ology of Mycobacterium tuberculosis based on
mycobacterial interspersed repetitive units. J Clin
Microbiol 2001;39:3563 – 71.
[8] Mazars E, Lesjean S, Banuls AL, et al. High-
resolution minisatellite-based typing as a portable
approach to global analysis of Mycobacterium tuber-
culosis molecular epidemiology. Proc Natl Acad Sci
U S A 2001;98:1901 – 6.
[9] Cowan LS, Mosher L, Diem L, et al. Variable-number
tandem repeat typing of Mycobacterium tuberculosis
isolates with low copy numbers of IS6110 by using
mycobacterial interspersed repetitive units. J Clin
Microbiol 2002;40:1592 – 602.
[10] Kwara A, Schiro R, Cowan LS, et al. Evaluation of
the epidemiologic utility of secondary typing methods
for differentiation of Mycobacterium tuberculosis
isolates. J Clin Microbiol 2003;41:2683 – 5.
[11] Nguyen D, Brassard P, Menzies D, et al. Genomic
characterization of an endemic Mycobacterium tuber-
culosis strain: evolutionary and epidemiologic impli-
cations. J Clin Microbiol 2004;42:2573 – 80.
[12] Kremer K, van Soolingen D, Frothingham R, et al.
Comparison of methods based on different molecular
epidemiological markers for typing of Mycobacte-
rium tuberculosis complex strains: Interlaboratory
study of discriminatory power and reproducibility.
J Clin Microbiol 1999;37(8):2607 – 18.
[13] DeRiemer K, Daley CL. Tuberculosis transmission
based on molecular epidemiologic research. Seminars
in Respiratory and Critical Care Medicine 2004;
25:297 – 306.
[14] Braden CR, Templeton GL, Cave MD, et al.
Interpretation of restriction fragment length poly-
morphism analysis of Mycobacterium tuberculosis
isolates from a state with a large rural population.
J Infect Dis 1997;175:1446 – 52.
[15] Godfrey-Faussett P, Stoker NG. Aspects of tuber-
culosis in Africa. 3. Genetic ‘fingerprinting’ for clues
to the pathogenesis of tuberculosis. Trans R Soc Trop
Med Hyg 1992;86:472 – 5.
[16] van Deutekom H, Hoijng SP, de Haas PEW, et al.
Clustered tuberculosis cases: do they represent recent
transmission and can they be detected earlier? Am J
Respir Crit Care Med 2004;169:806 – 10.
[17] Murray M, Alland D. Methodological problems in the
molecular epidemiology of tuberculosis. Am J Epi-
demiol 2002;155:565 – 71.
[18] Glynn JR, Vynnycky E, Pine PE. Influence of
sampling on estimates of clustering and recent trans-
227
mission of Mycobacterium tuberculosis derived from
DNA fingerprinting techniques. Am J Epidemiol
1999;149:366 – 71.
[19] Jasmer RM, Hahn JA, Small PM, et al. A molecular
epidemiologic analysis of tuberculosis trends in San
Francisco, 1991 – 1997. Ann Intern Med 1999;130:
971 – 8.
[20] van Soolingen D, Borgdorff MW, de Haas PE, et al.
Molecular epidemiology of tuberculosis in the Neth-
erlands: a nationwide study from 1993 through 1997.
J Infect Dis 1999;180:726 – 36.
[21] Hopewell PC. Factors influencing transmission and
infectivity of Mycobacterium tuberculosis: implica-
tions for clinical and public health management.
In: Sande MA, Hudson LD, Root RK, editors.
Respiratory infections. New York7 Churchill Living-
stone; 1986. p. 191 – 216.
[22] Small PM, Hopewell PC, Singh SP, et al. The
epidemiology of tuberculosis in San Francisco. A
population-based study using conventional and
molecular methods. N Engl J Med 1994;330:1703 – 9.
[23] Kline SE, Hedemark LL, Davies SF. Outbreak of
tuberculosis among regular patrons of a neighbor-
hood bar. N Engl J Med 1995;333:222 – 7.
[24] Behr MA, Warren SA, Salamon H, et al. Trans-
mission of Mycobacterium tuberculosis from patients
smear-negative for acid-fast bacilli. Lancet 1999;353:
444 – 9.
[25] Hernandez-Garduno E, Cook V, Kunimoto D, et al.
Transmission of tuberculosis from smear negative
patients: a molecular epidemiology study. Thorax
2004;59:286 – 90.
[26] Ong A, Creasman J, Hopewell PC, et al. A molecular
epidemiological assessment of extrapulmonary tu-
berculosis in San Francisco. Clin Infect Dis 2004;
38:25 – 31.
[27] Giordano TP, Soini H, Teeter LD, et al. Relating the
size of molecularly defined clusters of tuberculosis to
the duration of symptoms. Clin Infect Dis 2004;38:
10 – 6.
[28] Cronin WA, Golub JE, Lathan MJ, et al. Molecular
epidemiology of tuberculosis in a low-to moderate-
incidence state: are contact investigations enough?
Emerg Infect Dis 2002;8(11):1271 – 9.
[29] World Health Organization Global Tuberculosis
Programme. An expanded DOTS framework for
effective tuberculosis control. WHO/CDS/TB/
2002297. Geneva (Switzerland)7 World Health Orga-
nization; 2002.
[30] Nardell E, McInnis B, Thomas B, et al. Exogenous
reinfection with tuberculosis in a shelter for the
homeless. N Engl J Med 1986;18:1570 – 5.
[31] Small PM, Shafer RW, Hopewell PC, et al. Exoge-
nous reinfection with multidrug-resistant Mycobacte-
rium tuberculosis in patients with advanced HIV
infection. N Engl J Med 1993;328:1137 – 44.
[32] Sudre P, Pfyffer GE, Bodmer T, et al. Molecular
epidemiology of tuberculosis among HIV-infected
persons in Switzerland: a country-wide 9-year cohort
228
study. Swiss HIV Cohort Study. Infection 1999;27:
323 – 30.
[33] Heldal E, Docker H, Caugant DA. Pulmonary tuber-
culosis in Norwegian patients. The role of reactiva-
tion, re-infection and primary infection assessed by
previous mass screening data and restriction fragment
length polymorphism analysis. Int J Tuberc Lung Dis
2000;4:300 – 7.
[34] van Rie A, Warren R, Richardson M, et al. Exogenous
reinfection as a cause of recurrent tuberculosis after
curative treatment. N Engl J Med 1999;341:1174 – 9.
[35] Lambert ML, Hasker E, van Deun A, et al. Re-
currence in tuberculosis: relapse or reinfection?
Lancet Infect Dis 2003;3:282 – 7.
[36] de Boer AS, Kremer K, Borgdorff MW. Genetic
heterogeneity in Mycobacterium tuberculosis isolates
reflected in IS6110 restriction fragment length poly-
morphism patterns as low-intensity bands. J Clin
Microbiol 2000;38:4478 – 84.
[37] Das S, Paramasivan CN, Lowrie DB, et al. IS6110
restriction fragment length polymorphism typing of
clinical isolates of Mycobacterium tuberculosis from
patients with pulmonary tuberculosis in Madras,
South India. Tuber Lung Dis 1995;76:550 – 4.
[38] Caminero JA, Pena MJ, Campos-Herrero MI, et al.
Exogenous reinfection with tuberculosis on a Euro-
pean island with a moderate incidence of disease.
Am J Respir Crit Care Med 2001;163:717 – 20.
[39] Lourenco MC, Grinsztejn B, Fadinho-Montes FC,
et al. Genotypic patterns of multiple isolates of
M. tuberculosis from tuberculous HIV patients. Trop
Med Int Health 2000;5:488 – 94.
[40] Garcia de Viedma DG, Marin M, Hernangomez S,
et al. Tuberculosis recurrences: reinfection plays a
role in a population whose clinical/epidemiological
characteristics do not favor reinfection. Arch Intern
Med 2002;162:1873 – 9.
[41] Krüüner A, Pehme L, Ghebremichael S, et al. Use of
molecular techniques to distinguish between treat-
ment failure and exogenous reinfection with Myco-
bacterium tuberculosis. Clin Infect Dis 2002;35:
146 – 55.
[42] Godfrey-Faussett P, Githui W, Batchelor B, et al.
Recurrence of HIV-related tuberculosis in an en-
demic area may be due to relapse or reinfection.
Tuber Lung Dis 1994;75(3):199 – 202.
[43] Sonnenberg P, Murray J, Shearer S, et al. Tuber-
culosis treatment failure and drug resistance – same
strain or reinfection? Trans R Soc Trop Med Hyg
2000;94(6):603 – 7.
[44] Fitzpatrick LK, Okwera A, Mugerwa R, et al. An
investigation of suspected exogenous reinfection in
tuberculosis patients in Kampala, Uganda. Int J
Tuberc Lung Dis 2002;6(6):550 – 2.
[45] Bandera A, Gori A, Catozzi L, et al. Molecular
epidemiology study of exogenous reinfection in an
area with a low incidence of tuberculosis. J Clin
Microbiol 2001;39(6):2213 – 8.
[46] El Sahly HM, Wright JA, Soini H, et al. Recurrent
daley
tuberculosis in Houston, Texas: a population-based
study. Int J Tuberc Lung Dis 2004;8(3):333 – 40.
[47] Yeh RW, Hopewell PC, Daley CL. Simultaneous
infection with two strains of Mycobacterium tuber-
culosis identified by restriction fragment length poly-
morphism analysis. Int J Tuberc Lung Dis 1999;3:
537 – 9.
[48] Braden CR, Morlock GP, Woodley CL, et al.
Simultaneous infection with multiple strains of
Mycobacterium tuberculosis. Clin Infect Dis 2001;
33:42 – 7.
[49] Chaves F, Dronda F, Alonso-Sanz M, et al. Evidence
of exogenous reinfection and mixed infection with
more than one strain of Mycobacterium tuberculosis
among Spanish HIV-infected inmates. AIDS 1999;
13:615 – 20.
[50] Warren RM, Victor TC, Streicher EM, et al. Patients
with active tuberculosis often have different strains in
the same sputum specimen. Am J Respir Crit Care
Med 2004;169:610 – 4.
[51] Niemann S, Richter E, Rusch-Gerdes S, et al. Double
infection with a resistant and a multidrug-resistant
strain of Mycobacterium tuberculosis. Emerg Infect
Dis 2000;6:548 – 51.
[52] Middlebrook G. Isoniazid resistance and catalase ac-
tivity of the tubercle bacilli. Am Rev Tuberc 1954;
69:471 – 2.
[53] Cohn M, Kovitz C, Oda U. Studies on isoniazid and
tubercle bacilli: the growth requirements, catalase
activities, and pathogenic properties of isoniazid
resistant mutants. Am Rev Tuberc 1954;70:641 – 64.
[54] Riley RL, Mills CC, et al. Infectiousness of air from
a tuberculosis ward: comparative infectiousness of
different patients. Am Rev Respir Dis 1962;85:
511 – 25.
[55] Pym AS, Domenech P, Honore N, et al. Regulation of
catalase-peroxidase (KatG) expression, isoniazid sen-
sitivity and virulence by furA of Mycobacterium
tuberculosis. Mol Microbiol 2001;40(4):879 – 89.
[56] Garcia-Garcia ML, Ponce-de-Leon A, Jimenez-
Corona ME, et al. Clinical consequences and trans-
missibility of drug-resistant tuberculosis in southern
Mexico. Arch Intern Med 2000;160:630 – 6.
[57] Godfrey-Faussett P, Sonnenberg P, Shearer SC, et al.
Tuberculosis control and molecular epidemiology
in a South African goldmining community. Lancet
2000;356:1066 – 71.
[58] Burgos M, DeRiemer K, Small PM, et al. Effect of
drug-resistance on the generation of secondary cases
of tuberculosis. J Infect Dis 2003;188:1878 – 84.
[59] Munsiff SS, Nivin B, Sacajiu G, et al. Persistence
of a highly resistant strain of tuberculosis in New
York City during 1990 – 1999. J Infect Dis 2003;188:
356 – 63.
[60] Cohen T, Sommers B, Murray M. The effect of drug
resistance on the fitness of Mycobacterium tuber-
culosis. Lancet Infect Dis 2003;3:13 – 21.
[61] Veen J. Microepidemics of tuberculosis: the stone-in-
the-pond principle. Tuber Lung Dis 1992;73:73 – 6.
 molecular epidemiology
[62] van Deutekom H, Gerritsen JJ, van Soolingen D,
et al. A molecular epidemiological approach to
studying the transmission of tuberculosis in Amster-
dam. Clin Infect Dis 1997;25:1071 – 7.
[63] Daley CL. Tuberculosis contact investigations. Please
don’t fail me now. Am J Respir Crit Care Med 2004;
169:779 – 81.
[64] Behr MA, Hopewell PC, Paz EA, et al. Predictive
value of contact investigation for identifying recent
transmission of Mycobacterium tuberculosis. Am J
Respir Crit Care Med 1998;158:465 – 9.
[65] Bennett DE, Onorato IM, Ellis BA, et al. DNA
fingerprinting of Mycobacterium tuberculosis isolates
from epidemiologically linked case pairs. Emerg
Infect Dis 2002;8:1224 – 9.
[66] Sun SJ, Bennett DE, Flood J, et al. Identifying the
sources of tuberculosis in young children: a multistate
investigation. Emerg Infect Dis 2002;8:1216 – 23.
[67] Verver S, Warren RM, Munch Z, et al. Proportion of
tuberculosis transmission that takes place in house-
holds in a high incidence area. Lancet 2004;363:
212 – 4.
[68] Chin DP, Crane CM, Diul MY, et al. Spread of
Mycobacterium tuberculosis in a community imple-
menting recommended elements of tuberculosis con-
trol. JAMA 2000;283:2968 – 74.
[69] Kimerling ME, Benjamin WH, Lok KH, et al.
Restriction fragment length polymorphism screening
of Mycobacterium tuberculosis isolates: population
surveillance for targeting disease transmission in a
community. Int J Tuberc Lung Dis 1998;2:655 – 62.
[70] Lemaitre N, Sougakoff W, Truffot-Pernot C, et al.
Use of DNA fingerprinting for primary surveillance
of nosocomial tuberculosis in a large urban hospital:
detection of outbreaks in homeless people and mi-
grant workers. Int J Tuberc Lung Dis 1998;2:390 – 6.
[71] McNabb SJN, Kammerer JS, Hickey AC, et al.
Added epidemiologic value to tuberculosis preven-
tion and control of the investigation of clustered
genotypes of Mycobacterium tuberculosis isolates.
Am J Epidemiol 2004;160:589 – 97.
[72] Barnes PF, Yang Z, Preston-Martin S, et al. Patterns
of tuberculosis transmission in central Los Angeles.
JAMA 1997;278:1159 – 63.
[73] Bishai WR, Fraham NM, Harrington S, et al.
Molecular and geographic patterns of tuberculosis
transmission after 15 years of directly observed
therapy. JAMA 1998;280:1679 – 84.
[74] Daley CL, Small PM, Schecter GF, et al. An outbreak
of tuberculosis with accelerated progression among
persons infected with the human immunodeficiency
virus. An analysis using restriction-fragment-length
polymorphisms. N Engl J Med 1992;326:231 – 5.
[75] Yaganehdoost A, Graviss EA, Ross MW, et al.
Complex transmission dynamics of clonally related
virulent Mycobacterium tuberculosis associated with
barhopping by predominantly human immunodefi-
ciency virus-positive gay men. J Infect Dis 1999;180:
1245 – 51.
229
[76] Garcia-Garcia M, Palacios-Martinez M, Ponce-de-
Leon A, et al. The role of core groups in transmitting
Mycobacterium tuberculosis in a high prevalence
community in Southern Mexico. Int J Tuberc Lung
Dis 2000;4:12 – 7.
[77] Diel R, Meywald-Walter K, Gottschalk R, et al.
Ongoing outbreak of tuberculosis in a low incidence
community: a molecular-epidemiologic evaluation.
Int J Tuberc Lung Dis 2004;8(7):855 – 61.
[78] Leonhardt KK, Gentile F, Gilbert BP, et al. A cluster
of tuberculosis among crack house contacts in San
Mateo County, California. Am J Public Health 1994;
84:1834 – 6.
[79] Bock NN, Mallory JP, Mobley N, et al. Outbreak
of tuberculosis associated with a floating card game
in the rural south: lessons for tuberculosis contact
investigations. Clin Infect Dis 1998;27:1221 – 6.
[80] Kim SJ, Bai GH, Lee H, et al. Transmission of
Mycobacterium tuberculosis among high school
students in Korea. Int J Tuberc Lung Dis 2001;5:
824 – 30.
[81] Bauer J, Kok-Jensen A, Faurschou P, et al. A pro-
spective evaluation of the clinical value of nation-
wide DNA fingerprinting of tuberculosis isolates in
Denmark. Int J Tuberc Lung Dis 2000;4:295 – 9.
[82] Edlin BR, Tokars JI, Grieco MH, et al. An outbreak
of multidrug-resistant tuberculosis among hospital-
ized patients with the acquired immunodeficiency
syndrome. N Engl J Med 1992;326:1514 – 21.
[83] Haas DW, Milton S, Kreiswirth BN, et al. Nosoco-
mial transmission of a drug-sensitive W-variant
Mycobacterium tuberculosis strain among patients
with acquired immunodeficiency syndrome in Ten-
nessee. Infect Control Hosp Epidemiol 1998;19:
635 – 9.
[84] Chaves F, Dronda F, Cave MD, et al. A longitudinal
study of transmission of tuberculosis in a large prison
population. Am J Respir Crit Care Med 1997;155:
719 – 25.
[85] Hanau-Bercot B, Gremy I, Raskine L, et al. A one-
year prospective study (1994 – 1995) for a first
evaluation of tuberculosis transmission in French
prisons. Int J Tuberc Lung Dis 2000;4:853 – 9.
[86] Mohle-Boetani JC, Miguelino V, Dewsnup D, et al.
Tuberculosis outbreak in a housing unit for human
immunodeficiency virus-infected patients in a correc-
tional facility: transmission risk factors and effective
outbreak control. Clin Infect Dis 2002;34:668 – 76.
[87] Toungoussova OS, Mariandyshev A, Bjune G, et al.
Molecular epidemiology and drug resistance of
Mycobacterium tuberculosis isolates in the Archangel
prison in Russia: predominance of the W-Beijing
clone family. Clin Infect Dis 2003;37:665 – 72.
[88] Mangura BT, Napolitano EC, Passannante MR, et al.
Mycobacterium tuberculosis miniepidemic in a
church gospel choir. Chest 1998;113:234 – 7.
[89] Centers for Disease Control and Prevention. HIV-
related tuberculosis in a transgender network –
Baltimore, Maryland and New York City area,
230 daley
1998 – 2000. MMWR Morb Mortal Wkly Rep 2000;
49:317 – 20.
[90] Jereb JA, Burwen DR, Dooley SW, et al. Nosocomial
outbreak of tuberculosis in a renal transplant unit:
application of a new technique for restriction frag-
ment length polymorphism analysis of Mycobacte-
rium tuberculosis isolates. J Infect Dis 1993;168:
1219 – 24.
[91] Wilkinson D, Crump J, Pillay M, et al. Nosocomial
transmission of tuberculosis in Africa documented
by restriction fragment length polymorphism. Trans
R Soc Trop Med Hyg 1997;91:318.
[92] Frieden TR, Woodley CL, Crawford JT, et al.
The molecular epidemiology of tuberculosis in New
York City: the importance of nosocomial transmis-
sion and laboratory error. Tuber Lung Dis 1996;77:
407 – 13.
[93] Ikeda RM, Birkhead GS, DiFerdinando Jr GT, et al.
Nosocomial tuberculosis: an outbreak of a strain
resistant to seven drugs. Infect Control Hosp Epi-
demiol 1995;16:152 – 9.
[94] Johnson KR, Braden CR, Cairns KL, et al. Trans-
mission of Mycobacterium tuberculosis from medical
waste. JAMA 2000;284:1683 – 8.
[95] Agerton T, Valway S, Gore B, et al. Transmission
of a highly drug-resistant strain (strain W1) of
Mycobacterium tuberculosis. Community outbreak
and nosocomial transmission via a contaminated
bronchoscope. JAMA 1997;278(13):1073 – 7.
[96] Michele TM, Cronin WA, Graham NM, et al.
Transmission of Mycobacterium tuberculosis by a
fiberoptic bronchoscope. Identification by DNA
fingerprinting. JAMA 1997;278:1093 – 5.
[97] Hernandez-Garduno E, Kunimoto D, Wang L, et al.
Predictors of clustering of tuberculosis in Greater
Vancouver: a molecular epidemiologic study. CMAJ
2002;167:349 – 52.
[98] Kulaga S, Behr M, Musana K, et al. Molecular epide-
miology of tuberculosis in Montreal. CMAJ 2002;
167(4):353 – 4.
[99] FitzGerald JM, Fanning A, Hoepnner V, et al,
Canadian Molecular Epidemiology of TB Study
Group. The molecular epidemiology of tuberculosis
in western Canada. Int J Tuberc Lung Dis 2003;7(2):
132 – 8.
[100] Blackwood KS, Al-Azem A, Elliott LJ, et al.
Conventional and molecular epidemiology of tuber-
culosis in Manitoba. BMC Infect Dis 2003;3:18.
[101] Alland D, Kalkut GE, Moss AR, et al. Transmission
of tuberculosis in New York City. An analysis by
DNA fingerprinting and conventional epidemiologic
methods. N Engl J Med 1994;330(24):1710 – 6.
[102] Diel R, Schneider S, Meywald-Walter K, et al.
Epidemiology of tuberculosis in Hamburg, Germany:
long-term population-based analysis applying classi-
cal and molecular epidemiologic techniques. J Clin
Microbiol 2002;40:532 – 9.
[103] Vukovic D, Rusch-Gerdes S, Savic B, et al. Molecu-
lar epidemiology of pulmonary tuberculosis in
Belgrade, Central Serbia. J Clin Microbiol 2003;
41(9):4372 – 7.
[104] Zolnir-Dovc M, Poljak M, Erzen D, et al. Molecular
epidemiology of tuberculosis in Slovenia: results of a
one-year (2001) nation-wide study. Scand J Infect Dis
2003;35(11 – 12):863 – 8.
[105] Lockman S, Sheppard JD, Braden CR, et al. Mo-
lecular and conventional epidemiology of Mycobac-
terium tuberculosis in Botswana: a population-based
prospective study of 301 pulmonary tuberculosis pa-
tients. J Clin Microbiol 2001;39:1042 – 7.
[106] Chan-Yeung M, Tam C-M, Wong H, et al. Molecular
and conventional epidemiology of tuberculosis in
Hong Kong: a population-based prospective study.
J Clin Microbiol 2003;41(6):2706 – 8.
[107] Narayanan S, Das S, Garg R, et al. Molecular
epidemiology of tuberculosis in a rural area of high
prevalence in south India: implications for disease
control and prevention. J Clin Microbiol 2002;
40(12):4785 – 8.
[108] Geng E, Kreiswirth B, Driver C, et al. Changes in the
transmission of tuberculosis in New York City from
1990 to 1999. N Engl J Med 2002;346:1453 – 8.
[109] Diel R, Rusch-Gerdes S, Niemann S. Molecular epi-
demiology of tuberculosis among immigrants in Ham-
burg, Germany. J Clin Microbiol 2004;42:2952 – 60.
[110] Verver S, Warren RM, Munch Z, et al. Transmission
of tuberculosis in a high incidence urban community
in South Africa. Int J Epidemiol 2004;33:351 – 7.
[111] McNabb SJN, Braden CR, Navin TR. DNA finger-
printing of Mycobacterium tuberculosis: lessons
learned and implications for the future. Emerg Infect
Dis 2002;8:1314 – 9.
[112] Kong PM, Tapy J, Calixto P, et al. Skin-test screening
and tuberculosis transmission among the homeless.
Emerg Infect Dis 2002;8:1280 – 4.
[113] Soborg C, Soborg B, Pouelsen S, et al. Doubling
of the tuberculosis incidence in Greenland over an
8-year period (1990 – 1997). Int J Tuberc Lung Dis
2001;5:257 – 65.
[114] Bradford WZ, Koehler J, El-Hajj H, et al. Dissemi-
nation of Mycobacterium tuberculosis across the San
Francisco Bay Area. J Infect Dis 1998;177:1104 – 7.
[115] Ellis BA, Crawford JT, Braden CR, et al. Molecu-
lar epidemiology of tuberculosis in a sentinel sur-
veillance population. Emerg Infect Dis 2002;8:
1197 – 209.
[116] van Soolingen D, Qian L, de Haas PE, et al.
Predominance of a single genotype of Mycobacterium
tuberculosis in countries of east Asia. J Clin Micro-
biol 1995;33:3234 – 8.
[117] Anh DD, Borgdorff MW, Van LN, et al. Mycobacte-
rium tuberculosis Beijing genotype emerging in
Vietnam. Emerg Infect Dis 2000;6:302 – 5.
[118] Krüüner A, Hoffner SE, Sillastu H, et al. Spread of
drug-resistant pulmonary tuberculosis in Estonia.
J Clin Microbiol 2001;39:3339 – 45.
[119] Niemann S, Rusch-Gerdes S, Richter E. IS6110
fingerprinting of drug-resistant Mycobacterium tuber-
 molecular epidemiology
culosis strains isolated in Germany during 1995.
J Clin Microbiol 1997;35:3015 – 20.
[120] Caminero JA, Pena MJ, Campos-Herrero MI, et al.
Epidemiological evidence of the spread of a Myco- [127]
bacterium tuberculosis strain of the Beijing genotype
on Gran Canaria Island. Am J Respir Crit Care Med
2002;164:1165 – 70.
[121] Borgdorff MW, de Haas P, Kremer K, et al. [128]
Mycobacterium tuberculosis Beijing genotype, the
Netherlands. Emerg Infect Dis 2003;9:1310 – 3.
[122] Richardson M, van Lill SW, van der Spuy GD, et al.
Historic and recent events contribute to the disease [129]
dynamics of Beijing-like Mycobacterium tuberculosis
isolates in a high incidence region. Int J Tuberc Lung
Dis 2002;6:1001 – 11.
[123] van Crevel R, Nelwan RH, de Lenne W, et al. [130]
Mycobacterium tuberculosis Beijing genotype strains
associated with febrile response to treatment. Emerg
Infect Dis 2001;7:880 – 3.
[124] Lan NT, Lien HT, Tungle B, et al. Mycobacterium [131]
tuberculosis Beijing genotype and risk for treatment
failure and relapse, Vietnam. Emerg Infect Dis 2003;
9(12):1633 – 5.
[125] Lillebaek T, Andersen AB, Dirksen A, et al. Myco- [132]
bacterium tuberculosis Beijing genotype. Emerg
Infect Dis 2003;9:1553 – 7.
[126] Tracevska T, Jansone I, Baumanis V, et al. Prevalence
231
of Beijing genotype in Latvian multidrug-resistant
Mycobacterium tuberculosis isolates. Int J Tuberc
Lung Dis 2003;7:1097 – 103.
Bifani PJ, Plikaytis BB, Kapur V, et al. Origin and
interstate spread of a New York City multidrug-
resistant Mycobacterium tuberculosis clone family.
JAMA 1996;275:452 – 7.
Bifani PJ, Mathema B, Liu Z, et al. Identification of
a W variant outbreak of Mycobacterium tuberculosis
via population-based molecular epidemiology. JAMA
1999;282:2321 – 7.
Agerton TB, Valway SE, Blinkhorn RJ, et al. Spread
of strain W, a highly drug-resistant strain of Myco-
bacterium tuberculosis, across the United States. Clin
Infect Dis 1999;29:85 – 92.
Glynn JR, Whiteley J, Bifani PJ, et al. Worldwide
occurrence of Beijing/W strains of Mycobacterium
tuberculosis: a systematic review. Emerg Infect Dis
2002;8:843 – 9.
Bifani PJ, Mathema B, Kurepina NE, et al. Global
dissemination of the Mycobacterium tuberculosis
W-Beijing family strains. Trends Microbiol 2002;10:
45 – 52.
Lopez B, Aguilar D, Orozco H, et al. A marked
difference in pathogenesis and immune response
induced by different Mycobacterium tuberculosis
genotypes. Clin Exp Immunol 2003;133:30 – 7.