Documente Academic
Documente Profesional
Documente Cultură
CONTENTS
Preface ix
Neil W. Schluger
The DOTS Strategy for Controlling the Global Tuberculosis Epidemic 197
Thomas R. Frieden and Sonal S. Munsiff
This article reviews the principles, scientific basis, and experience with implementation
of the directly observed treatment, short-course (DOTS) strategy for tuberculosis. The rel-
evance of DOTS in the context of multidrug-resistant tuberculosis and the HIV epidemic
also is discussed.
vi CONTENTS
issues posed by the use of combination antiretroviral therapy. The author reviews these con-
troversies and makes recommendations for the management of HIV-related tuberculosis.
New Drugs for Tuberculosis: Current Status and Future Prospects 327
Richard J. O’Brien and Mel Spigelman
This article reviews two classes of compounds that have advanced into phase II and III
clinical trials, long-acting rifamycins and fluoroquinolones, and a number of other drugs
that have entered or may enter clinical development in the near future.
Index 349
CONTENTS vii
Clin Chest Med 26 (2005) 349 – 353
Index
B Culture(s)
Bacille Calmette-Guerin vaccination in pulmonary tuberculosis diagnosis,
effect on tuberculin skin test, 306 257 – 258
Bronchoscopy
fiberoptic D
in pulmonary tuberculosis diagnosis, Diarylquinolones (R207910)
256 – 257 for tuberculosis, 333
Dihydroimidazo-oxazoles (OPC-67683)
for tuberculosis, 335
C 1,25-Dihydroxyvitamin D3
Chemotherapy in genetic susceptibility to tuberculosis, 238
for active tuberculosis
axioms of, 273 – 274 Directly observed treatment, short-course
standardized short-course (DOTS) strategy
in DOTS strategy for controlling global for controlling global tuberculosis epidemic,
tuberculosis epidemic, 198 – 199 197 – 205
administrative commitment to, 197
Children drug quality in, 199
latent Mycobacterium tuberculosis infection in for multi-drug resistant TB, 200 – 201
treatment of, 322 HIV infection and, 201 – 202
0272-5231/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/S0272-5231(05)00042-0 chestmed.theclinics.com
350 INDEX
political commitment to, 197 Global Alliance for Tuberculosis Drug Development,
results of, 200 341 – 347
sputum microscopy of patients attending described, 341 – 342
health facilities, 198 future directions for, 346
standardized short-course chemotherapy, global product development public – private
198 – 199 partnerships, 346
systemic monitoring and strategy of, 342 – 345
accountability, 200 Global product development public – private
DOTS. See Directly observed treatment, short-course partnerships
(DOTS) strategy. in Global Alliance for Tuberculosis Drug
Development, 345 – 346
Drug(s)
for tuberculosis, 327 – 340
new agents. See also specific drug and H
Tuberculosis, treatment of, drugs in. HIV. See Human immunodeficiency virus
tuberculosis resistant to, 175 – 176 (HIV) infection.
Drug resistance HLA-DR2
in pulmonary tuberculosis diagnosis in genetic susceptibility to tuberculosis, 238
rapid detection of, 261 – 262
tuberculosis effects of Human immunodeficiency virus (HIV) infection
transmission- and pathogenesis-related, 222 tuberculosis and. See Tuberculosis, HIV-related.
I
E Interferon gamma signaling pathway
Ethambutol in genetic susceptibility to tuberculosis, 237 – 238
for active tuberculosis, 276 Isoniazid
Ethnicity for active tuberculosis, 275
as factor in tuberculosis, 186 for latent Mycobacterium tuberculosis infection,
316 – 319
rifampin with
for latent Mycobacterium tuberculosis
F infection, 320 – 321
Fiberoptic bronchoscopy
in pulmonary tuberculosis diagnosis, 256 – 257
L
Fingerprinting patterns Large-sequence polymorphisms
in Mycobacterium tuberculosis complex in Mycobacterium tuberculosis complex study, 210
study, 208
Line probe assays
Fluoroquinolones in pulmonary tuberculosis diagnosis, 261
for active tuberculosis, 276 – 277
LL3858
for tuberculosis, 335
Luciferase reporter phages
G in pulmonary tuberculosis diagnosis, 262
Genetic susceptibility
to tuberculosis, 233 – 246. See also Tuberculosis,
genetic susceptibility to. M
Macrolide(s)
Genome(s)
for tuberculosis, 335 – 336
sequenced
in Mycobacterium tuberculosis complex Mannan-binding lectin (MBL)
study, 208 in genetic susceptibility to tuberculosis, 238
INDEX 351
Historical background data collection was a priority for the National Tu-
berculosis Association. As the mortality rate con-
The epidemiology of tuberculosis (TB) in the tinued to decrease, attention focused on TB case
United States has changed remarkably over the last finding. Armed with a new diagnostic tool, the chest
2 centuries. In the nineteenth century, TB was the roentgenogram, mass chest radiograph screenings
leading cause of death. As the nineteenth century prog- were conducted beginning in the early 1930s and
ressed, TB mortality decreased, partly because of continuing into the 1950s, enabling the diagnosis of
improved socioeconomic conditions [1,2], especially TB patients before they became symptomatic [2].
in urban settings, and partly owing to the natural The need to expand data collection to include TB
behavior of epidemics [3]. After the tubercle bacillus morbidity in addition to TB mortality was acknowl-
was identified as the causative agent of TB by Robert edged [1,2]. Reliable and complete morbidity data
Koch in 1882, the approach to TB control changed would allow TB experts to measure more accurately
greatly, and the concepts of public health, prevention, the magnitude of the TB problem and the effec-
and segregation of TB patients gained more accep- tiveness of control efforts. In 1920, the National Tu-
tance. As a result, in industrialized countries, the berculosis Association published its first Diagnostic
prescribed treatment of rest, isolation, nutrition, and Standards and Classifications of TB to assist health
fresh air for TB patients was achieved with long stays care providers and standardize diagnostic criteria
in sanatoria [1,2]. [5,6]. National TB mortality and morbidity data,
By the late 1800s, TB was more than ever con- coordinated by the National Tuberculosis Associa-
sidered a public health issue, even though there were tion, became available in 1933. In 1944, a United
few well established local or state public health States Public Health Service Act mandated the crea-
departments [1,2,4]. More resources became avail- tion of a national TB control program [1]. With
able, and public health programs dedicated to TB con- the introduction of the therapeutic agents streptomy-
trol were established. In 1904, the first voluntary cin (1947), p-aminosalicylic acid (1949), isoniazid
health agency dedicated to TB, the National Tuber- (1952), and pyrazinamide (1952) TB mortality rates
culosis Association (now the American Lung Asso- decreased dramatically. Between 1930 and 1960, the
ciation), was organized [1,5]. TB surveillance and mortality rate decreased by 92%, from 71 to 6 deaths
per 100,000 population.
Because of the widespread use of chemotherapy,
This work was funded by the Division of Tuberculosis long hospitalizations for TB were no longer needed,
and Elimination, Centers for Disease Control and Prevention. and TB sanatoria and hospitals began to close [1,2,5].
* Corresponding author. Having a standard definition for a reportable case of
E-mail address: ees2@cdc.gov (E. Schneider). TB for surveillance purposes became paramount [7],
and in 1951, a committee consisting of state TB con- Reporting of RVCT data to the CDC also will
trol officers and sanatoria directors published recom- be modified with the transitioning of the TIMS to
mendations for TB case reporting and counting the Web-based National Electronic Disease Surveil-
procedures [8]. In 1952, the United States Public lance System.
Health Service (USPHS) Tuberculosis Control Pro-
gram instituted procedures to report new cases of
TB. Not until 1953, through the cooperation of the Tuberculosis resurgence
states, did the USPHS receive reports from the entire
United States, heralding the birth of the national TB Noting that extraordinary strides against TB have
surveillance system [1,2]. been made both in treatment and surveillance since
the 1950s, many TB experts have believed that TB
National tuberculosis surveillance system elimination in the United States is within reach [1,2].
In 1959, the historic Arden House Conference, spon-
Since 1953, the national TB surveillance system sored by the National Tuberculosis Association and
has been modified several times to monitor and re- the USPHS Tuberculosis Control Program, brought
spond better to changes in TB morbidity. Data are together TB experts to formulate a plan on how to
collected on TB cases that have been verified and eliminate TB; this plan served as a basis for future TB
have met the Centers for Disease Control and Pre- control efforts [12]. TB incidence continued to
vention (CDC) public health surveillance case defi- decrease. From 1953 through 1985, TB case numbers
nition for TB [9,10]. TB is a reportable disease in decreased by 74%, from 84,304 to 22,201 cases,
each state [11]. In 1985, the national TB surveillance and the case rate decreased by 82%, from 53.0 to
system changed: originally collecting aggregate data, 9.3 cases per 100,000 population. As a result, many
the CDC began collecting individual case reports no longer considered TB to be a major problem.
on a form called the Report of Verified Case of In the early 1970s, federal funding allocated for
Tuberculosis (RVCT). Currently, data are collected TB control began to decrease, and, as a result, many
by reporting areas (the 50 states, the District of Co- TB control services were dismantled [13,14]. Al-
lumbia, New York City, Puerto Rico, and jurisdic- though TB funds were decreasing, the cost of treating
tions in the Pacific and Caribbean) using the RVCT. TB was increasing. In 1981, only $3.7 million was
An RVCT is completed for each reported new TB appropriated to the CDC to fight TB nationally.
disease case and contains patient demographic, clini- In 1987, the Advisory Committee (now Council)
cal, and laboratory information. An RVCT is com- for Elimination of Tuberculosis (ACET) was estab-
pleted by the health department for each confirmed lished, and its membership was directed to develop a
TB case and transmitted to the CDC to be included strategic plan for TB elimination [15]. The ACET and
in the national TB surveillance database. The CDC the CDC published this plan, proposing a TB in-
annually publishes a report summarizing national cidence interim goal for the year 2000 of 3.5 or fewer
TB statistics [10]. Also included in this annual report TB cases per 100,000 population and an elimination
are the ‘‘Recommendations for Counting Reported target of less than 1 TB case per million population
TB Cases,’’ which were last revised in 1997. by 2010. In the mid-to-late 1980s, however, the
The CDC has maintained a computer database longstanding downward trend in TB incidence was
on TB surveillance data since 1985. State and local interrupted. In 1986, a 2.6% annual increase in the
TB programs have been able to collect, manage, and case number was documented, signaling the begin-
transfer TB surveillance data (i.e., RVCT) electroni- ning of the TB resurgence (Fig. 1). In the late 1980s,
cally to the CDC first through software for expanded after decades of decreasing TB incidence, TB once
TB surveillance (SURVS-TB, 1993 – 1997) and cur- again became a major threat.
rently through the Tuberculosis Information Manage- The resurgence had a significant impact on TB
ment System (TIMS, 1998 – present). In 1993, the control strategies in the United States. Because of
RVCT was expanded to collect additional information newly identified risk groups, the focus of many TB
(eg, drug resistance, HIV infection) in response to the control strategies had to be shifted, and many pro-
TB epidemic of the mid-to-late 1980s and early grams needed to be overhauled. CDC researchers
1990s. The most recent modification was imple- concluded that the resurgence had resulted in an es-
mented in January 2003 to meet federal standards for timated 52,100 excess TB cases from 1985 through
the classification of race and ethnicity. Additional 1992 [16]. Several factors have been linked to the
changes for the national TB surveillance system are resurgence, including the deterioration of the TB
on the horizon with a revision of the RVCT. program infrastructure, the HIV/AIDS epidemic,
epidemiology of tuberculosis in the united states 185
28,000
Number of TB Cases
26,000
24,000
22,000
20,000
18,000
16,000
14,000
12,000
1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003
Year
Fig. 1. Reported tuberculosis cases in the United States from 1981 to 2003.
drug-resistant TB, TB among foreign-born persons, veillance, augment case finding and contact inves-
and an increase in transmission, especially in con- tigations, advance laboratory capacity (eg, drug-
gregate and institutional settings [16 – 19]. susceptibility testing and new diagnostic tools), and
The degree to which each of these factors affected ensure each patient completed therapy through the
TB control at the local level varied, but two of these use of directly observed therapy (DOT).
factors, the HIV/AIDS epidemic and TB among
foreign-born persons, strongly influenced the TB re-
surgence in the United States. HIV infection is con-
sidered to be the greatest risk factor known today After the tuberculosis resurgence, 1993 – 2003
for TB. Several large outbreaks of multidrug-resistant
TB (MDR-TB) (ie, TB resistant to at least isoniazid During the resurgence, the national TB incidence
and rifampin) among persons infected with HIV were peaked in 1992 at 26,673 cases (10.5 cases per
documented in Florida and New York City [20 – 22]. 100,000 population). The aggressive attack on TB in
In 1991, 41% of culture-positive TB patients in New the United States resulted in the annual TB case num-
York City were also infected with HIV, and 19% had ber and case rate decreasing in 1993 to 25,108 cases,
MDR-TB [23]. Early diagnosis of TB among persons 9.7 cases per 100,000 population. Tuberculosis be-
infected with HIV was difficult because of the lack of came more localized to well-defined risk groups and
specific clinical findings, such as a positive tuberculin geographic areas [34,35]. In response, strategic plans
skin test result and an abnormal chest radiograph. were revised to help prioritize efforts and outline
Ineffective isolation precautions also contributed to updated recommendations for TB elimination in the
nosocomial transmission of MDR-TB among patients United States [36,37].
and health care providers [24 – 26]. HIV-related TB From 1993 to 2002, the average year-to-year de-
outbreaks were also documented in other congregate crease in TB rate was 6.9%. In 2003, however, the
settings [27] such as correctional facilities [28] and CDC reported the smallest annual decrease in the TB
homeless shelters [29,30]. Another important factor rate (1.9%) and TB case numbers (184) since the re-
fueling the TB resurgence was the immigration of surgence, raising concern about a possible slowing
persons from countries that have high rates of TB of the progress against TB. For 2003, 14,874 TB
[19]. The proportion of reported TB cases among cases were reported in the United States, with a rate
foreign-born persons increased from 22% in 1986 of 5.1 per 100,000 population that remains higher
(the first year birthplace data were collected by the than the national interim goal of 3.5 cases per
national TB surveillance system) to 30% in 1993. 100,000 population set for 2000. Moreover, despite
In the early 1990s, the newly established Federal the decline in TB nationwide, rates have increased
Tuberculosis Task Force revaluated existing TB in certain states, and elevated TB rates continue to be
strategies and formulated the National Action Plan reported in certain populations (eg, foreign-born per-
to Combat MDR-TB [31]. In the United States, a sons and racial/ethnic minorities). In 2003, 12 states
monumental public health effort to control TB was and the District of Columbia reported case rates
initiated [32,33]. Federal funding was increased and above the national average, and 20 states reported
used to rebuild the TB infrastructure, strengthen sur- increases in case number compared with 2002 [10].
186 schneider et al
187
188 schneider et al
8,000 50
40
6,000
30
4,000
20
2,000 10
0 0
1986 1988 1990 1992 1994 1996 1998 2000 2002
Year
Number of Foreign-born TB Cases Percentage of Foreign-born TB Cases
Fig. 2. Trends in tuberculosis cases in foreign-born persons in the United States from 1986 to 2003.
Most TB cases among foreign-born persons are requirements for persons seeking permanent resi-
caused by Mycobacterium tuberculosis complex in- dency in the United States [43].
fections acquired abroad [41]. Among foreign-born TB among foreign-born persons is a major com-
children, aged younger than 15 years, who had TB, ponent of TB morbidity in the United States [40]
60% were diagnosed within 18 months of arrival in and reflects the global TB situation, defined in 1993
the United States [38]. Prompt evaluation of foreign- by the World Health Organization (WHO) as a
born persons for TB following their arrival in the global emergency [44,45]. The WHO estimated that
United States can help identify persons who have in 2002 there were 8.8 million new cases of TB
LTBI and are eligible for preventive therapy; prompt (141 cases per 100,000 population) [46]. Among the
evaluation can prevent development of active TB 22 high-burden countries, India and China ac-
disease [41,42]. Foreign-born TB patients are also counted for 46% of the total. Among the 15 coun-
more likely to have drug resistance and are less likely tries that have the highest TB rates (>400 cases per
to be HIV infected than US-born TB patients [40]. 100,000 population), 13 are in Africa, and 12 of
The lower proportion of foreign-born TB patients these had high TB/HIV incidence rates (>100 cases
infected with HIV results in part from HIV screening per 100,000 population) among adults 15 to 49 years
Fig. 3. Trends in tuberculosis cases in persons born in the United States from 1986 to 2003.
epidemiology of tuberculosis in the united states 189
in foreign-born persons
25
23
22 22
20
15 15
14
10
10 9
6
5
4 4
3 3
2 2 2
0
1986 1988 1990 1992 1994 1996 1998 2000 2002
Year
Fig. 4. Number of states with 50% or more of tuberculosis cases in foreign-born persons in the United States from 1986 to 2003.
incidence of resistance to at least one drug, and the at least isoniazid and 0.9% had MDR-TB. Addition-
incidence of MDR-TB was 2.2% [56]. Higher drug- ally, drug resistance (MDR-TB and resistance to at
resistance rates were seen among TB patients who least isoniazid) has been seen more commonly in
have had a previous episode of TB, foreign-born foreign-born TB patients (2003: MDR-TB, 1.2%;
persons, HIV-infected persons, and persons residing isoniazid, 10.6%) than in US-born TB patients (2003:
in specific geographic areas (eg, New York City). MDR-TB, 0.6%; isoniazid, 4.6%).
In the mid-to-late 1990s, several outbreaks involv- Knowledge of drug-resistance rates worldwide
ing highly drug-resistant strains of M. tuberculosis is critical to controlling the global epidemic and has
(ie, strain W) were investigated [57 – 59]. These direct implications for TB control in the United States
strains share a common drug resistance to first-line [60,61]. A more comprehensive understanding of
antituberculosis medications (eg, isoniazid, rifampin, global drug resistance was made possible with the
ethambutol, and, at that time, streptomycin) as well formation of the Supranational Reference Labora-
as resistance to some second-line medications, mak- tory Network in 1994 and the WHO/IUATLD Global
ing treatment difficult and costly. The majority of Project on Anti-Tuberculosis Drug Resistance Sur-
strain W TB cases were reported by New York City veillance. Newly released data reveal that TB patients
[57,59], although outbreaks have occurred elsewhere, in parts of Eastern Europe and Central Asia are
including one that was attributed to bronchoscope 10 times more likely to have MDR-TB than patients
contamination in South Carolina [58]. To facilitate in the rest of the world, with some MDR-TB inci-
early detection of strain W isolates, the CDC began dence rates higher than 10% (Israel, 14.2%; Kazakh-
recommending that health departments notify the stan,14.2%; Tomsk Oblast [Russian Federation],
CDC of all M. tuberculosis isolates that have 13.7%; Uzbekistan, 13.2%; Estonia, 12.2%; and
strain W – resistance patterns [59]. Liaoning [China], 10.4%) [61].
Since 1998, overall multidrug resistance among
culture-positive TB patients, who do not have a prior Tuberculosis/HIV coinfection
history of TB, has been relatively stable (~1%)
(Fig. 5), although outbreaks and regional differ- Today, any discussion about TB is incomplete
ences continue to occur. Historically, overall drug- without a discussion about HIV/AIDS. Knowing a
resistance rates among those who have a previous TB patient’s HIV status is critical to management,
history of TB have been higher than for those who treatment, contact investigation, and prevention
do not have a previous history of TB. In 2003, [62 – 67]. The CDC recommends that all TB patients,
among TB patients who had a prior history of TB, independent of risk factors, should undergo voluntary
12.6% had resistance to at least isoniazid and 3.6% HIV counseling, testing, and referral [64,65,67].
had MDR-TB, whereas 7.9% of TB patients who Nonetheless, HIV status is not reported nationally
did not have a prior history of TB had resistance to for many TB patients in the United States. This in-
Fig. 5. MDR-TB among persons without a history of tuberculosis in the United States from 1993 to 2003. MDR-TB is defined
as resistance to at least isoniazid and rifampin.
epidemiology of tuberculosis in the united states 191
complete reporting of HIV status probably reflects mented in HIV-infected patients who have low CD4+
several factors including concerns about confiden- T-lymphocyte counts, extrapulmonary disease, and
tiality, interpretation of laws and regulations in cer- concomitant antifungal therapy [74,75]. Clinicians
tain states and local jurisdictions, and reluctance by treating TB-HIV – coinfected persons should be fa-
health care providers to report HIV test results to the miliar with current diagnostic, management (eg, DOT),
TB surveillance program staff [10]. Information on and treatment modalities to maximize therapeutic
HIV status was added to the national TB surveillance success and minimize TB transmission, drug resis-
system in 1993, in response to the TB resurgence. tance, adverse effects, and treatment failures [64,67].
HIV test results (ie, negative, positive, or indeter- Globally, the HIV/AIDS epidemic has had an
minate) were reported for 45.7% of TB patients aged immense impact on TB control, especially in sub-
25 to 44 years in 1993 and for 65.3% in 2002. In Saharan Africa, where an estimated two thirds of
this group, positive HIV test results were reported for persons who have HIV/AIDS live, and has contrib-
29.1% in 1993 and for 15.9% in 2002. Historically, uted significantly to TB morbidity and mortality
reported TB/HIV coinfection rates and case numbers [76 – 78]. In these countries, TB incidence and case
have been relatively high in a few states and urban fatality are strongly associated with HIV prevalence.
areas. In 2002, 60% of the positive HIV test results The prevalence of drug-resistant TB is expected to
among TB patients aged 25 to 44 years were reported increase greatly as the HIV epidemic spreads to areas
from five areas: California, Florida, Georgia, New of the world where drug-resistant TB is more
York City, and Texas. Crossmatching of state TB prevalent (eg, Asia, Eastern Europe) [61,76]. The
registries and HIV/AIDS registries in 1993 and 1994 scaling up of treatment programs providing anti-
revealed that 14% (range, 0% – 31%) of persons retroviral therapy will require patient and health care
reported to have TB in the United States were also provider education and close monitoring to opti-
listed in HIV/AIDS registries [68]. TB-AIDS cases mize therapy, reduce transmission, and reduce drug-
were more likely to be in persons aged 25 to 44 years, resistant TB [79].
male, culture-positive for M. tuberculosis, and US-
born. In geographic areas where the prevalence rates Development of new tools
of HIV-infected persons were high, drug resistance,
especially MDR-TB (6%) and rifampin monoresis- An important component of disease control is the
tance (3%), was reported among TB-AIDS patients. development of new diagnostic tests, pharmacologic
HIV coinfection has several key implications for agents, and vaccines. The resurgence of TB in the
the overall treatment and management of TB. HIV mid-to-late 1980s to 1992 was associated with delays
infection increases the risk of (1) TB disease pro- in the diagnosis and identification of drug resistance.
gression among persons who have LTBI, (2) rapid This situation generated renewed interest in the de-
progression of those newly infected with M. tuber- velopment of several new diagnostic tools and the
culosis to active TB disease, and (3) reinfection with subsequent genomic sequencing of M. tuberculosis.
M. tuberculosis [67,69]. Many of the TB outbreaks During the past few years, TB diagnostic capabili-
among persons infected with HIV that occurred ties have improved through new techniques for
during the resurgence were complicated by high the rapid detection of M. tuberculosis complex (eg,
drug-resistance rates and resulted in mortality rates nucleic acid amplification tests) [80], identification
reaching 70% [21 – 23]. TB outbreaks among HIV- of M. tuberculosis (eg, nucleic acid probe), rapid de-
infected persons have illustrated the continued need tection of latent TB infection (eg, whole-blood inter-
for appropriate treatment and monitoring of this feron gamma assay [QuantiFERON (Cellestis Inc.,
population [70 – 73]. The use of antiretroviral ther- Valencia, California)]) [81,82], the investigational
apy has significantly decreased mortality and mor- enzyme-linked immunospot test (ELISPOT) [83],
bidity, including the development of opportunistic and differentiation of M. tuberculosis strains (eg,
infections (eg, TB) among HIV-infected persons. DNA fingerprinting) [84,85].
New concerns have developed, however, concern- In the 1990s, molecular genetic typing (genotyp-
ing the potential for drug – drug interactions, develop- ing) of M. tuberculosis strains became a commonly
ment of resistance to rifamycin, and paradoxical used tool to understand outbreaks and transmission
reactions. Drug – drug interactions, primarily between dynamics. In 1996, the CDC established the National
rifamycin and protease inhibitors and nonnucleoside TB Genotyping and Surveillance Network to deter-
reverse transcriptase inhibitors, have resulted in new mine the usefulness of molecular genotyping in more
treatment guidelines and recommendations [66,66a]. routine TB control settings using the IS6110-based
Acquired rifampin monoresistance has been docu- restriction fragment length polymorphism (RFLP)
192 schneider et al
technique supplemented with spacer oligonucleotide 1992, with case numbers increasing by 20%. Follow-
typing (spoligotyping) on M. tuberculosis isolates ing an intensive campaign and mobilization of new
[86,87]. Genotyping, in conjunction with epidemio- resources, TB cases once again began to decline.
logic investigation, has proven a useful adjunct to Remarkable gains have been made since the early
epidemiologic investigations in tracing the chain of 1990s, with efforts being concentrated on maintain-
transmission [88]. The techniques are particularly ing control of TB, speeding the decline of TB, and
useful in outbreaks and institutional settings, identi- developing new tools [37]. Key TB epidemiologic
fying groups at risk for TB (eg, homeless persons), features that have been identified include an increas-
identifying contacts and social networks, under- ing proportion of TB cases among persons born in
standing exogenous reinfection, and confirming labo- countries where TB is endemic, racial and ethnic
ratory cross-contamination [89,90]. To refine the disparities, and localized unique epidemiologic pro-
understanding of TB transmission and epidemiol- files in areas throughout the United States. Develop-
ogy and to advance TB control, the CDC has ment of new tools, such as vaccines, antituberculosis
launched the National TB Genotyping Program, drugs, and rapid diagnostic tests have also been
which provides the capacity to genotype M. tuber- identified as vital measures needed to eliminate TB
culosis isolates from all culture-positive TB patients in the United States.
in the United States. Two polymerase chain reaction – The smallest decline since the resurgence was
based genotyping tests (spoligotyping, mycobacterial seen in 2003, raising the concern about a possible
interspersed repetitive units analysis) will be supple- slowing of the progress against TB or even a reversal
mented with IS6110 RFLP testing for selected speci- of the decline. Despite increasing health care costs
mens [91]. The goal of this program is to improve and demands for increased programmatic and opera-
the characterization of TB transmission dynamics and tional efforts, funding for TB control has not in-
to use the results to improve the efficiency of public creased [95]. The elimination of TB in the United
health interventions. States will require sustained efforts such as identify-
In 1995, following a several-year hiatus in the ing and targeting populations at high risk for TB,
USPHS-sponsored clinical trials, the CDC reinstated remaining actively involved in the global effort
clinical TB research, creating the TB Trials Con- against TB, and maintaining adequate resources.
sortium (TBTC). The TBTC currently is coordinat-
ing several studies, including efficacy trials for the
use of moxifloxcin as a first-line drug in the treat- Acknowledgments
ment of TB disease. Information gained from the
earliest of these studies contributed to the Food and The authors thank the state and local tuberculosis
Drug Administration licensure of rifapentine, a long- control officials in health departments throughout
acting rifampin and the first anti-TB drug approved the United States who collected and reported the
in 25 years [92]. Additional studies include a com- national surveillance data presented in this article,
parison of several generations of QuantiFERON the surveillance staff at the Division of TB Elimi-
with the tuberculin skin test in the diagnosis of LTBI nation, Centers for Disease Control and Prevention,
[81]. In 2001, the CDC established the TB Epide- who maintain the database, Ann H. Lanner for her
miologic Studies Consortium to conduct multicenter editorial review of the manuscript, and Dr. Thomas
epidemiologic, behavioral, and operational research Navin and Dr. Michael Iademarco for their critical
studies. Furthermore, recognizing the need for TB review of the manuscript.
prevention globally, a renewed interest, fueled by
generous funding has resulted in actively revisiting
vaccine development [93,94]. Numerous organiza- References
tions, both in the public and private sectors, are con-
ducting major research efforts to develop a safe and [1] Lowell AM. Tuberculosis morbidity and mortality and
effective TB vaccine. its control. In: Lowell AM, Edwards LB, Palmer CE,
editors. Tuberculosis. Cambridge (MA)7 Harvard Uni-
versity Press; 1969. p. 1 – 65.
[2] Jacobs PP. The control of tuberculosis in the United
Elimination of tuberculosis in the United States: States. New York7 National Tuberculosis Association;
remaining challenges 1940. p. 1 – 387.
[3] Blower SM, McLean AR, Porco TC, et al. The intrinsic
After decades of decline, an unprecedented re- transmission dynamics of tuberculosis epidemics. Nat
surgence in TB began in 1986 and continued through Med 1995;1(8):815 – 21.
epidemiology of tuberculosis in the united states 193
[4] Paterson RG. The evolution of official tuberculosis 1988 – 1991. MMWR Morb Mortal Wkly Rep 1991;
control in the United States. Public Health Rep 1947; 40(34):585 – 91.
62(10):336 – 41. [22] Edlin BR, Tokars JI, Grieco MH, et al. An outbreak of
[5] Shryock RH. National Tuberculosis Association, multidrug-resistant tuberculosis among hospitalized
1904 – 1954. New York7 National Tuberculosis Asso- patients with the acquired immunodeficiency syn-
ciation; 1957. p. 1 – 342. drome. N Engl J Med 1992;326:1514 – 21.
[6] Long ER, Hopkins FD. History of diagnostic stan- [23] Frieden TR, Sterling T, Pablos-Mendez A, et al. The
dards and classification of tuberculosis of the Na- emergence of drug-resistant tuberculosis in New York
tional Tuberculosis Association. Am Rev Tuberc 1952; City. N Engl J Med 1993;328:521 – 6.
65(4):494 – 504. [24] Dooley SW, Villarino ME, Lawrence M, et al. Noso-
[7] Northrop C, Anderson RJ, Sauer HI. What is a re- comial transmission of tuberculosis in a hospital unit
portable case of tuberculosis? Public Health Rep 1949; for HIV-infected patients. JAMA 1992;267:2632 – 5.
64(31):961 – 8. [25] Jereb JA, Klevens M, Privett TD, et al. Tuberculosis
[8] United States Public Health Service. What is a re- in health care workers at a hospital with an outbreak
portable case of tuberculosis? Public Health Rep 1951; of multidrug-resistant Mycobacterium tuberculosis.
66(40):1291 – 4. Arch Intern Med 1995;155:854 – 9.
[9] Centers for Disease Control and Prevention. Case [26] Pearson ML, Jereb JA, Frieden TR, et al. Nosocomial
definitions for infectious conditions under public transmission of multidrug-resistant Mycobacterium
health surveillance. MMWR Morb Mortal Wkly Rep tuberculosis. Ann Intern Med 1992;117:191 – 6.
1997;46(RR-10):40 – 1. [27] Daley CL, Small PM, Schecter GF, et al. An outbreak
[10] Centers for Disease Control and Prevention. of tuberculosis with accelerated progression among
Reported tuberculosis in the United States, 2003. persons infected with the human immunodeficiency
Atlanta (GA)7 US Department of Health and Human virus. N Engl J Med 1992;326:231 – 5.
Services; 2004. [28] Valway SE, Greifinger RB, Papania M, et al. Multidrug-
[11] Centers for Disease Control and Prevention. Tuber- resistant tuberculosis in the New York State prison
culosis control laws – United States, 1993. MMWR system, 1990 – 1991. J Infec Dis 1994;170:151 – 6.
Morb Mortal Wkly Rep 1993;42(RR-15):1 – 28. [29] Centers for Disease Control and Prevention. Drug-
[12] The Arden House Conference on Tuberculosis. Wash- resistant tuberculosis among the homeless – Boston.
ington, DC; 1960. US Department of Health, Educa- MMWR Morb Mortal Wkly Rep 1985;34(28):429 – 31.
tion, and Welfare, Public Health Service Publication [30] Centers for Disease Control and Prevention. Tuber-
No. 784, p. 1 – 68. culosis among residents of shelters for the homeless –
[13] Reichman LB. The U-shaped curve of concern. Am Ohio, 1990. MMWR Morb Mortal Wkly Rep 1991;
Rev Respir Dis 1991;144:741 – 2. 40(50):869 – 71, 877.
[14] Brown RE, Miller B, Taylor WR, et al. Health-care [31] Centers for Disease Control and Prevention. National
expenditures for tuberculosis in the United States. action plan to combat multidrug-resistant tuberculosis.
Arch Intern Med 1995;155:1595 – 600. MMWR Morb Mortal Wkly Rep 1992;41(RR-11):
[15] Centers for Disease Control and Prevention. A stra- 1 – 30.
tegic plan for the elimination of tuberculosis in the [32] Frieden TR, Fujiwara PI, Washko RM, et al. Tuber-
United States. MMWR Morb Mortal Wkly Rep 1989; culosis in New York City – turning the tide. N Engl
38(S-3):1 – 25. J Med 1995;333:229 – 33.
[16] Cantwell MF, Snider DE, Cauthen GM, et al. Epi- [33] McKenna MT, McCray E, Jones JL, et al. The fall af-
demiology of tuberculosis in the United States, 1985 ter the rise: tuberculosis in the United States, 1991
through 1992. JAMA 1994;272(7):535 – 9. through 1994. Am J Pub Health 1998;88:1059 – 63.
[17] Brudney K, Dobkin J. Resurgent tuberculosis in New [34] Fujiwara PI. Tide pools: what will be left after the
York City: human immunodeficiency virus, home- tide has turned? Int J Tuberc Lung Dis 2000;4(12):
lessness, and the decline of tuberculosis control pro- S111 – 6.
grams. Am Rev Respir Dis 1991;144:745 – 9. [35] Centers for Disease Control and Prevention. Progress-
[18] Snider DE, Roper WL. The new tuberculosis. JAMA ing toward tuberculosis elimination in low-incidence
1992;326(10):703 – 5. areas of the United States. MMWR Morb Mortal Wkly
[19] McKenna MT, McCray E, Onorato I. The epidemiol- Rep 2002;51(RR-5):1 – 14.
ogy of tuberculosis among foreign-born persons in the [36] Centers for Disease Control and Prevention. Tuber-
United States, 1986 to 1993. N Engl J Med 1995;332: culosis elimination revisited: obstacles, opportunities,
1071 – 6. and a renewed commitment. MMWR Morb Mortal
[20] Beck-Sague C, Dooley SW, Hutton MD, et al. Hospital Wkly Rep 1999;48(RR-9):1 – 13.
outbreak of multidrug-resistant Mycobacterium tuber- [37] Centers for Disease Control and Prevention. Federal
culosis infections. JAMA 1992;268(10):1280 – 6. tuberculosis task force plan in response to the Insti-
[21] Centers for Disease Control and Prevention. Nosoco- tute of Medicine report, ending neglect: the elimination
mial transmission of multidrug-resistant tuberculosis of tuberculosis in the United States. Atlanta (GA)7
among HIV-infected persons – Florida and New York, Centers for Disease Control and Prevention, US
194 schneider et al
Department of Health and Human Services; 2003. New York7 Kluwer Academic/Plenum Publishers;
p. 1 – 76. 2003. p. 245 – 93.
[38] Nelson LJ, Schneider E, Wells CD, et al. Epidemiology [53] Rajbhandary SS, Marks SM, Bock NN. Costs of
of childhood tuberculosis in the United States, 1993 – patients hospitalized for multidrug-resistant tuber-
2001: the need for continued vigilance. Pediatrics 2004; culosis. Int J Tuberc Lung Dis 2004;8(8):1012 – 6.
114:333 – 41. [54] Bloch AB, Simone PM, McCray E, et al. Preventing
[39] Centers for Disease Control and Prevention. Racial multidrug-resistant tuberculosis. JAMA 1996;275(6):
disparities in tuberculosis – selected southeastern states, 487 – 9.
1991 – 2002. MMWR Morb Mortal Wkly Rep 2004; [55] Bloch AB, Cauthen GM, Onorato IM, et al. Nation-
53(25):556 – 9. wide survey of drug-resistant tuberculosis in the
[40] Talbot EA, Moore M, McCray E, et al. Tuberculosis United States. JAMA 1994;271(9):665 – 71.
among foreign-born persons in the United States, [56] Moore M, Onorato IM, McCray E, et al. Trends in
1993 – 1998. JAMA 2000;284:2894 – 900. drug-resistant tuberculosis in the United States,
[41] Zuber PL, McKenna MT, Binkin NJ, et al. Long-term 1993 – 1996. JAMA 1997;278(10):833 – 7.
risk of tuberculosis among foreign-born persons in [57] Frieden TR, Fine Sherman L, Lay Maw K, et al. A
the United States. JAMA 1997;278:304 – 7. multi-institutional outbreak of highly drug-resistant
[42] Wells CD, Zuber PL, Nolan CM, et al. Tuberculosis tuberculosis. JAMA 1996;276(15):1229 – 35.
prevention among foreign-born persons in Seattle-King [58] Agerton T, Valway S, Gore B, et al. Transmission of a
County, Washington. Am J Respir Crit Care Med 1997; highly drug-resistant strain (Strain W1) of Mycobac-
156(2):573 – 7. terium tuberculosis. JAMA 1997;278(13):1073 – 7.
[43] Centers for Disease Control and Prevention. Division [59] Agerton TB, Valway SE, Blinkhorn RJ, et al. Spread
of Global Migration and Quarantine. Technical in- of strain W, a highly drug-resistant strain of Myco-
structions for panel physicians. Available at: http:// bacterium tuberculosis, across the United States. Clin
www.cdc.gov/ncidod/dq/panel.htm. Accessed August Infect Dis 1999;29:85 – 92.
5, 2004. [60] Dye C, Espinal MA, Watt CJ, et al. Worldwide
[44] World Health Organization. TB—a global emergency. incidence of multidrug-resistant tuberculosis. J Infec
WHO TB programme WHO/TB/94.177. Geneva Dis 2002;185:1197 – 202.
(Switzerland)7 World Health Organization; 1994. [61] World Health Organization. Anti-tuberculosis drug re-
p. 1 – 28. sistance in the world, Report No. 3. Geneva (Swit-
[45] Dye C, Scheele S, Dolin P, et al, for the WHO Global zerland)7 World Health Organization; 2004. p. 1 – 92.
Surveillance and Monitoring Project. Global burden [62] Centers for Disease Control and Prevention. Missed
of tuberculosis. JAMA 1999;282:677 – 86. opportunities for prevention of tuberculosis among
[46] World Health Organization. Global tuberculosis con- persons with HIV infection – selected locations,
trol: surveillance, planning, financing. WHO report United States, 1996 – 1997. MMWR Morb Mortal
WHO/HTM/TB2004.331. Geneva (Switzerland)7 Wkly Rep 2000;49(30):685 – 7.
World Health Organization; 2004. [63] Centers for Disease Control and Prevention. Acquired
[47] Laserson KF, Yen NTN, Thornton C, et al. Improve- rifamycin resistance in persons with advanced HIV
ment i acid-fast bacilli sputum smear sensitivity disease being treated for active tuberculosis with
following processing with C18-carboxypropylbetaine: intermittent rifamycin-based regimens. MMWR Morb
a study in US-bound immigrants from Viet Nam. J Clin Mortal Wkly Rep 2002;51(10):214 – 5.
Microbiol, in press. [64] American Thoracic Society. Centers for Disease
[48] Centers for Disease Control and Prevention. Prevent- Control and Prevention, Infectious Diseases Society
ing and controlling tuberculosis along the US – Mexico of America. ATS/CDC/IDSA Statement on the treat-
border. MMWR Morb Mortal Wkly Rep 2002; ment of tuberculosis. Am J Respir Crit Care Med
50(RR-1):1 – 28. 2003;167:603 – 62.
[49] Schneider E, Laserson KF, Wells CD, et al. Tuberculo- [65] Centers for Disease Control and Prevention. Guide-
sis along the United States – Mexico border, 1993 – lines for preventing opportunistic infections among
2001. Rev Panam Salud Publica 2004;16(1):23 – 34. HIV-infected persons – 2002. MMWR Morb Mortal
[50] Wells CD, Nelson LJ. New international efforts in Wkly Rep 2002;51(RR-8):1 – 52.
childhood tuberculosis: proceedings from the 2002 [66] Centers for Disease Control and Prevention. Updated
workshop on childhood tuberculosis, Montreal, Can- guidelines for the use of rifabutin or rifampin for the
ada, 6 – 7 October 2002. Int J Tuberc Lung Dis 2004; treatment and prevention of tuberculosis among HIV-
8(5):630 – 5. infected patients taking protease inhibitors or non-
[51] Nelson LJ, Wells CD. Global epidemiology of child- nucleoside reverse transcriptase inhibitors. MMWR
hood tuberculosis. Int J Tuberc Lung Dis 2004;8(5): 2000;49(9):185 – 9.
636 – 47. [66a] Centers for Disease Control and Prevention. Guide-
[52] Spradling P, Ridzon R. Multidrug-resistant tuberculosis lines updated January 20, 2004. Updated guidelines
(MDR TB). In: Fong IW, Drlica K, editors. Reemer- for the use of rifamycins for the treatment of
gence of established pathogens in the 21st century. tuberculosis among HIV-infected patients taking
epidemiology of tuberculosis in the united states 195
protease inhibitors or nonnucleoside reverse tran- diagnostic tests for tuberculosis. Am J Respir Crit Care
scriptase inhibitors. Available at: http://www.cdc.gov/ Med 1997;155:1804 – 14.
nchstp/tb/TB_HIV_Drugs/TOC.htm. Accessed Sep- [81] Mazurek GH, LoBue PA, Daley CL, et al. Comparison
tember 1, 2004. of a whole-blood interferon-gamma assay with tuber-
[67] Centers for Disease Control and Prevention. Pre- culin skin testing for detecting latent Mycobacterium
vention and treatment of tuberculosis among pa- tuberculosis infection. JAMA 2001;286(14):1740 – 7.
tients infected with human immunodeficiency virus: [82] Centers for Disease Control and Prevention. Guidelines
principles of therapy and revised recommendations. for using QuantiFERON1-TB Test for diagnosing
MMWR Morb Mortal Wkly Rep 1998;47(RR-20): latent Mycobacterium tuberculosis infection. MMWR
1 – 58. Morb Mortal Wkly Rep 2003;52(RR-2):15 – 8.
[68] Moore M, McCray E, Onorato IM. Cross-matching [83] Lalvani A, Pathan AA, McShane H, et al. Rapid
TB and AIDS registries: TB patients with HIV co- detection of Mycobacterium tuberculosis infection by
infection, United States, 1993 – 1994. Public Health enumeration of antigen-specific T cells. Am J Respir
Rep 1999;114:269 – 77. Crit Care Med 2001;163:824 – 8.
[69] Shafer RW, Singh SP, Larkin C, Small PM. Exoge- [84] Barnes PF, Cave MD. Molecular epidemiology of
nous reinfection with multidrug-resistant Mycobacte- tuberculosis. N Engl J Med 2003;349(12):1149 – 56.
rium tuberculosis in an immunocompetent person. [85] Woods GL. The mycobacteriology laboratory and new
Tuber Lung Dis 1995;76:575 – 7. diagnostic techniques. Infect Dis Clin North Am 2002;
[70] Centers for Disease Control and Prevention. Drug- 16(1):127 – 45.
susceptible tuberculosis outbreak in a state correctional [86] Castro KG, Jaffe HW. Rationale and methods for the
facility housing HIV-infected inmates – South Caro- national tuberculosis genotyping and surveillance net-
lina, 1999 – 2000. MMWR Morb Mortal Wkly Rep work. Emerg Infect Dis 2002;8(11):1188 – 91.
2000;49(46):1041 – 4. [87] Crawford JT, Braden CR, Schable BA, et al. National
[71] Centers for Disease Control and Prevention. Tuber- tuberculosis genotyping and surveillance network:
culosis outbreaks in prison housing units for HIV- design and methods. Emerg Infect Dis 2002;8(11):
infected inmates – California, 1995 – 1996. MMWR 1192 – 6.
Morb Mortal Wkly Rep 1999;48(4):79 – 82. [88] Small PM, Hopewell PC, Singh SP, et al. The epi-
[72] Spradling P, Drociuk D, McLaughlin S, et al. Drug- demiology of tuberculosis in San Francisco. N Engl
drug interactions in inmates treated for human immu- J Med 1994;330(24):1703 – 9.
nodeficiency virus and Mycobacterium tuberculosis [89] Small PM, McClenny NB, Singh SP, et al. Molecular
infection or disease: an institutional tuberculosis out- strain typing of Mycobacterium tuberculosis to con-
break. Clin Infect Dis 2002;35:1106 – 12. firm cross-contamination in the mycobacteriology
[73] McElroy PD, Southwick KL, Fortenberry ER, et al. laboratory and modification of procedures to minimize
Outbreak of tuberculosis among homeless persons occurrence of false-positive cultures. J Clin Microbiol
coinfected with human immunodeficiency virus. Clin 1993;31(7):1677 – 82.
Infect Dis 2003;36:1305 – 12. [90] Braden CR, Templeton GL, Stead WW, et al. Retro-
[74] Vernon A, Burman W, Benator D, et al for the TB Trials spective detection of laboratory cross-contamination of
Consortium. Acquired rifamycin monoresistance in Mycobacterium tuberculosis cultures with use of DNA
patients with HIV-related tuberculosis treated with fingerprint analysis. Clin Infec Dis 1997;24:35 – 40.
once-weekly rifapentine and isoniazid. Lancet 1999; [91] Rosenblum LS, Navin TR, Crawford JT. Molecular
353:1843 – 7. epidemiology of tuberculosis. N Engl J Med 2003;
[75] Sandman L, Schluger NW, Davidow AL, et al. Risk 349(24):2364.
factors for rifampin-monoresistant tuberculosis. Am J [92] The TB Trials Consortium. Rifapentine and isoniazid
Respir Crit Care Med 1999;159:468 – 72. once a week versus rifampicin and isoniazid twice a
[76] Corbett EL, Watt CJ, Walker N, et al. The growing week for treatment of drug-susceptible pulmonary
burden of tuberculosis. Arch Intern Med 2003;163: tuberculosis in HIV-negative patients: a randomised
1009 – 21. clinical trial. Lancet 2002;360:528 – 34.
[77] Corbett EL, Steketee R, ter Kuile FO, et al. HIV-1/ [93] Centers for Disease Control and Prevention. Develop-
AIDS and the control of other infectious diseases in ment of new vaccines for tuberculosis. MMWR Morb
Africa. Lancet 2002;359:2177 – 87. Mortal Wkly Rep 1998;47(RR-13):1 – 6.
[78] Joint United Nations Programme on HIV/AIDS. 2004 [94] Reed SG, Alderson MR, Dalemans W, et al. Prospects
Report on the global AIDS epidemic. UNAIDS/04. for a better vaccine against tuberculosis. Tuberculosis
16E. Geneva (Switzerland)7 UNAIDS; 2004. 2003;83:213 – 9.
[79] Gupta R, Irwin A, Ravigilone MC, et al. Scaling-up [95] National Coalition for Elimination of Tuberculosis.
treatment for HIV/AIDS: lessons learned from multi- Tuberculosis elimination: the federal funding gap.
drug-resistant tuberculosis. Lancet 2004;363:320 – 4. National Coalition for Elimination of Tuberculosis;
[80] American Thoracic Society Board of Directors. Rapid 2004. p. 1 – 19.
Clin Chest Med 26 (2005) 233 – 246
Genetic factors are important contributors to the considering the contribution of genetic factors to ob-
development of a wide range of complex or multi- served familial clustering of disease. In a study of
factorial diseases. For example, if one has a close almost 1000 adoptees, Sorensen et al [1] found that
relative who died of ischemic heart disease, one will the host genetic component of susceptibility to pre-
be at increased risk of developing ischemic heart dis- mature death from infection was greater than for
ease oneself. It is not inevitable that one will develop cancer or cardiovascular disease. This finding does
the same condition, because other factors such as not mean that if one’s father dies of tuberculosis, one
cigarette smoking, diet, exercise, and ‘‘bad luck’’ also inevitably will succumb to the same disease. Rather,
contribute to the risk of developing ischemic heart it indicates that a person who has such a family
disease. A positive family history simply indicates history has a higher probability of dying from tuber-
that one is at increased risk of developing heart dis- culosis than a person who does not have a positive
ease compared with a person of the same age and sex family history. Thus the statement that host genetic
who does not have a positive family history. factors make a person susceptible to a particular in-
Patients and their doctors recognize the impor- fectious disease simply means that the risk of devel-
tance of family history for a wide range of multi- oping the disease is higher than for someone who has
factorial, noncommunicable diseases such as cancer, not inherited the genetic risk factor (ie, who can be
cardiovascular disease, and diabetes mellitus. Much described as resistant). Someone whose genetic
of the clustering in families that occurs in these con- constitution makes him or her susceptible to Myco-
ditions, however, probably results from shared en- bacterium tuberculosis will not necessarily develop
vironmental risk factors. Although there is much clinical disease after exposure to the microorganism.
debate about the importance of nature versus nurture, Conversely someone whose genetic constitution
few scientists would dismiss the importance of host makes him or her resistant to M. tuberculosis may
genetics in the development of chronic noncommu- still develop clinical disease after exposure. The term
nicable diseases. In contrast, host genetic factors ‘‘susceptible’’ simply indicates a genetic make-up
are often given little attention in infectious disease. with a higher risk of developing tuberculosis than
The familial clustering that is frequently observed for a resistant one.
infectious diseases is commonly dismissed as the In 1949, Haldane [2] suggested that microorgan-
result of transmission of infection among household isms have been the main agents of natural selection
members. This assumption is unfair, because host among human populations for the past 5000 years.
genetic factors are probably at least as important in He believed that the recent evolution of the human
determining the outcome of infection as they are in genome was driven primarily by the need to resist
other complex diseases. pathogens because infectious diseases were the most
Studies on adoptees are sometimes used to help important cause of premature death. This hypothesis
eliminate the effects of shared environment when proposes that most of the genetic diversity found
within human populations has been selected for
and maintained by pathogenic microorganisms. Until
E-mail address: bellamyrj2000@yahoo.co.uk recently, progress in identifying the host genes con-
0272-5231/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccm.2005.02.006 chestmed.theclinics.com
234 bellamy
ferring resistance to infectious diseases was limited received direct inoculation with the same dose of
to malaria. Geographic variations in the prevalence the same strain of M. tuberculosis. Therefore the dif-
of malaria facilitated the identification of several gene ference in outcome of the infection is probably ex-
variants conferring resistance to malaria, including plained by host genetic factors.
sickle cell hemoglobin, a- and b-hemoglobin, and It has been suggested that a population’s resis-
glucose-6-phosphate dehydrogenase deficiency [3,4]. tance to M. tuberculosis is determined by its history
Genetic mutations conferring malaria resistance are of previous exposure [8]. Resistance is built up over a
common in populations originating in areas where number of generations by selection pressure in favor
malaria is endemic and are rare among populations of mycobacteria-resistant gene variants. When the
where malaria does not occur. selection pressure is strong, gene variants conferring
Historically, tuberculosis has not shown the disease-resistance can rise to high population fre-
marked geographic variation shown by malaria. This quencies in relatively short periods of time. The most
lack of variation has made the task of identifying the striking example of a population acquiring resistance
host genes conferring tuberculosis susceptibility and to tuberculosis has been described by Motulsky [9].
resistance more difficult to identify. It has been esti- At the end of the nineteenth century the Qu’Appelle
mated that M. tuberculosis was responsible for more Indians suffered their first cases of tuberculosis. The
than one fifth of all deaths in Western Europe after disease spread rapidly through the population and
the industrial revolution [5]. It is therefore logical to caused the deaths of 10% of the total population
expect that M. tuberculosis would have been among annually. After two generations (40 years), half of the
the microorganisms that have had the most important families had been lost, but the annual death rate from
effects on the evolution of the human genome. It tuberculosis had fallen to 0.2% [9]. This striking
has been estimated that approximately one third of reduction in tuberculosis-specific mortality is be-
the world’s population is infected with mycobacteria, lieved to have resulted from the strong selection
but among those infected only about 10% will ever pressure in favor of M. tuberculosis – resistant gene
develop clinical disease [6]. Much scientific research variants [9].
has focused on proving that host genetic factors Tuberculosis has been a major cause of premature
are important in determining why only a minority of death in Europe since the Industrial Revolution. In
those infected by M. tuberculosis develops clinical Africa it is believed that tuberculosis did not become
disease. This article provides a brief summary of this widespread until the start of the twentieth century
research and describes the strategies that have been [10 – 12], when the building of densely populated
used to identify the host genes involved in tuber- towns and cities facilitated the spread of M. tuber-
culosis resistance. culosis. Haldane’s theory would therefore predict
that present-day people of European origin would
have greater resistance to tuberculosis than people
Evidence showing the importance of host genetics of African origin. A study of 25,000 nursing home
in tuberculosis susceptibility residents in Arkansas confirmed that black resi-
dents were twice as likely to become infected with
In 1926, in Lubeck, Germany, a tragic accident M. tuberculosis as white residents of the same nurs-
occurred during the bacille Calmette-Guerin (BCG) ing homes [13]. This difference could not be ex-
vaccination program: 249 babies were injected with plained by environmental or social factors, indicating
the same live dose of virulent M. tuberculosis bac- that it was most likely caused by host genetics [13].
teria. Seventy-six babies died; 173 babies survived Twin studies are the most reliable method of
[7]. This experience showed that there is wide determining whether genetic factors are involved
variation in the degree of innate immunity against in determining who will develop a particular dis-
M. tuberculosis. The babies who died presumably had ease. Monozygotic twins inherit identical genomes,
weaker resistance to M. tuberculosis than the babies whereas dizygotic twins share only 50% of the genes
who managed to clear the infection. This difference in they inherit. If monozygotic twins have higher
host immunity could not have resulted from variation concordance for a disease than dizygotic twins, this
in acquired resistance to M. tuberculosis, because the finding indicates that genetic factors are important
babies were too young to have had significant prior in the etiology of the disease. Several studies have
exposure to mycobacteria, and they had not pre- compared the concordance for tuberculosis in mono-
viously been vaccinated with BCG. The difference zygotic twins with that in dizygotic twins. These
could also not be explained by variation in virulence studies have shown concordance rates among mono-
or infectivity of the bacteria, because all the infants zygotic twins to be approximately twice as high as
genetic susceptibility to tuberculosis 235
concordance rates among dizygotic twins [14 – 16]. complex, family-based designs are also occasionally
This evidence that host genes are important in de- used. Association studies generally involve typing
termining susceptibility/resistance to tuberculosis has individual gene variants that are believed to be poten-
led to the development of several strategies to attempt tially important because of previous work on animal
to identify the genes involved. models of the disease or for theoretical reasons. As-
sociation between a genetic marker and a disease can
be caused by either the genetic variant itself confer-
ring disease susceptibility/resistance or the genetic
Strategies to identify host genes determining variant being in linkage disequilibrium with the true
tuberculosis susceptibility disease-susceptibility locus. Linkage disequilibrium
occurs only when two gene variants are located close
Identifying the genes influencing susceptibility together (within 1 centi-Morgan) on the same chro-
to multifactorial diseases is a complex process. No mosome. A genome-wide association study would
single strategy could identify all of the genes of require typing more than 3000 genetic markers; this
interest, because each study design has advantages approach would be time-consuming and expensive
and limitations. and therefore generally is not feasible. Because as-
Genome-wide linkage studies are used to identify sociation studies are usually restricted to candidate
regions of the genome that contain major disease- genes, this approach cannot be used alone. Regard-
susceptibility loci. This approach involves typing a less of how many candidate genes are typed, the
large number of genetic markers (usually more than genes that exert the largest effects on population-wide
300) covering the whole human genome. Sibling-pair disease risk could be overlooked. Candidate gene
families that contain two or more siblings affected studies and genome-wide linkage studies are there-
by the disease of interest are usually used in studies fore complementary, and both are generally required
of multifactorial diseases. These families are pre- when investigating a multifactorial disease.
ferred to large, extended families (which are generally Animal models have several advantages for the
used for single-gene Mendelian disorders), because study of multifactorial diseases. Animal models can
they are believed to be more representative of the be used in breeding experiments, targeted gene dis-
disease in the general population. Large numbers ruption can be performed, and the animals can be
of sibling-pair families must be typed to provide suf- challenged with pathogens. Identifying genes that
ficient power to detect genes exerting a large effect confer susceptibility to individual pathogens in animal
on population-wide disease risk. The approach is species is a useful strategy for identifying candidate
systematic and comprehensive and in theory should genes for case-control studies in human populations.
detect any genomic region that contains a major Human patients who suffer from extreme suscep-
disease-susceptibility locus. This approach has rela- tibility to specific opportunistic infections can also
tively low power, however, and it cannot detect provide insight into the host immune response to
gene loci that confer only a moderate effect on common infectious disease. For example, individuals
population-wide disease risk. For example if a who have developed overwhelming infection follow-
disease-susceptibility allele has a frequency of 0.5 ing BCG infection or who have developed dissemi-
(ie, 50% of the alleles in the population are of nated atypical mycobacterial infections have provided
this type) and exerts a twofold increased risk of dis- insight into host resistance to M. tuberculosis. This
ease compared with the resistant wild-type allele, article describes how several of these methods have
2498 sibling-pair families are required to provide been used to identify some of the host genes involved
an 80% probability of identifying this effect by in susceptibility to tuberculosis.
linkage analysis [17]. It would be difficult to col-
lect and type this number of families in a genome-
wide screen.
Association-based candidate gene studies have Natural resistance – associated macrophage
much greater power. Only 340 cases are required to protein
have the power to detect the effect of the disease-
susceptibility locus described previously [17]. Asso- More than 20 years ago, innate susceptibility to
ciation-based candidate gene studies can therefore infection with M. bovis BCG was shown to be de-
detect genetic effects that would be overlooked in termined by a single genetic factor in inbred strains
a genome-wide linkage study. The simplest form of of mice [18]. Gros and colleagues [18] named this
association study is the case-control study, but more putative gene Bcg. Following a single inoculation
236 bellamy
with BCG-Montreal, mice carrying the resistant allele Nramp1D169/ , and Nramp1/ should have identical
(Bcgr ) had between 100- and 1000-fold fewer splenic phenotypes. Gros’ group found that mice with these
colony-forming units than mice carrying the suscep- three Nramp1 genotypes had indistinguishable resis-
tible gene variant (Bcgs ) [19]. The Bcgr allele was tance to mycobacteria, confirming that Nramp1D169
found to be dominant over the Bcgs allele [18]. The is a null allele [38]. In the third experiment the
Bcg locus was mapped to murine chromosome 1 by normal Nramp1G169 allele was transferred onto the
linkage analysis [20]. Linkage analysis also showed background of a mouse with the homozygous
that Bcg was the same gene that determined resis- Nramp1D169/D169 genotype [39]. Macrophages from
tance to Leishmania, salmonella, and other myco- this transgenic mouse expressed the Nramp1 protein,
bacterial species such as M. lepraemurium and and the BCG-resistant phenotype was restored [39].
M. intracellulare [21 – 28]. In vitro studies showed These experiments proved beyond doubt that Nramp1
that macrophages from mice carrying the Bcgs allele is the putative Bcg gene.
had decreased ability to restrict the growth of myco- The Nramp1 gene consists of 15 exons spanning
bacteria, salmonella, and Leishmania compared with 11.5 kb. The gene encodes a 90- to 100-kiloDalton
Bcgr macrophages [29 – 32]. As a result, Bcgs mice membrane-bound protein containing 12 hydrophobic
had impaired ability to control the initial stage of transmembrane domains [36,40,41]. Macrophages
mycobacterial infection, but the genetic suscepti- from Nramp1D169/D169 mice do not produce detect-
bility does not affect the later stages of an infection, able Nramp1 protein [42]. The human homolog of
which are determined by the acquired immune re- Nramp1 was originally designated NRAMP1 but
sponse [19]. has now been renamed Slc11a1 (solute carrier family
A high-resolution linkage map of murine chro- 11 member 1). In this article the name NRAMP1 is
mosome 1 enabled the Bcg locus to be pin-pointed retained to avoid confusion. NRAMP1 maps to chro-
to a 0.3 – centi-Morgan region [33,34]. The search mosome 2q35, contains 15 exons spanning 12 kb
for the Bcg gene itself was then vigorously pursued of DNA, and encodes a polypeptide consisting of
by Gros’ group. They produced a 400 – kg-base (kb) 550 amino acids [43,44]. There are several Nramp-
bacteriophage and cosmid contig of the region related proteins in humans, mice, and many other
surrounding the Bcg gene [35] and used a molecular species. In mice, Nramp2 and Nramp-rs have been
method called exon trapping to isolate potential mapped to chromosomes 15 and 17, respectively [45,
candidate genes for Bcg [36]. One gene was ex- 46]. The human homologue of Nramp2 (NRAMP2)
pressed solely in macrophage populations and en- has been cloned and mapped to chromosome 12q13
coded a polypeptide with characteristics suggestive [47,48].
of a transport protein. This gene was believed to be Evidence is now accumulating to suggest that
the Bcg gene, because it contained a nonconservative the NRAMP1 protein is a transmembrane iron trans-
base change (glycine to aspartic acid at position 169: porter. It was first suggested in 1996 that Nramp1
designated G169D) in the Bcgs strains studied [36]. may be a metal cation transporter because of its
The group named this gene the natural resistance- structural similarity to a known manganese trans-
associated macrophage protein gene (Nramp, later porter in Saccharomyces cerevisiae [49]. A missense
re-named Nramp1) [36]. Three experiments were then mutation in the Nramp2 gene (glycine to arginine
performed to demonstrate that Nramp1 is the putative at position 185; G185D) was subsequently found to
Bcg gene. Twenty inbred strains of mice with the be responsible for microcytic anemia in a murine
Bcgr phenotype and seven inbred strains of mice with model [50]. The same Nramp2G185D mutation is also
the Bcgs phenotype were typed for the Nramp1 responsible for microcytic anemia in the Belgrade
G169D variant [37]. All 20 strains with the Bcgr rat [51]. Murine Nramp2 and human NRAMP2
phenotype were found to be Nramp1G169/G169 wild- are integral membrane glycoproteins located at the
type homozygotes, and all seven Bcgs strains were plasma membrane and at recycling endosomes
Nramp1D169/D169 gene-variant homozygotes [37]. The [52,53]. In contrast, Nramp1 is localized to the late
second experiment involved the production of an endosomal compartment of resting macrophages and
Nramp1 gene-disrupted knock-out mouse [38]. is recruited to the phagosome on phagocytosis [54].
This knock-out mouse, designated Nramp1/ , was In the presence of excess iron, macrophages from
mated with the homozygous gene-variant mouse Nramp1D169/D169 mice have the same ability to limit
Nramp1 D169/D169 . The resulting offspring all had intracellular M. avium replication as macrophages
the compound genotype Nramp1D169/- [38]. If the from wild-type Nramp1G169/G169 mice [55]. This
Nramp1D169 allele is a nonfunctional (or null) finding suggests that Nramp1 may be an iron trans-
allele, mice with the genotypes Nramp1D169/D169 , porter that becomes saturated at high concentration.
genetic susceptibility to tuberculosis 237
Iron Iron
Nramp2
H+ Iron H+
Nramp2 Iron
Nramp1
Iron
Iron
Nramp2
H+ Nramp1 H+
Fig. 1. Nramp1 and Nramp2 function as hypothesized by Gruenheid et al [52]. The ubiquitous Nramp2 is a membrane
glycoprotein that becomes internalized to endosomes with iron. Acidification of the endosome activates Nramp2 resulting in the
transport of iron and protons into the cytoplasm. Nramp1 is recruited to the macrophage phagosome after bacteria and iron are
ingested. Following acidification, Nramp1 transports iron and protons out of the phagosome into the macrophage cytoplasm.
Presumably mycobacterial Mramp competes with Nramp1 for the available intraphagosomal iron.
M. tuberculosis possesses a member of the Nramp culosis. In a case-control study of more than 800 per-
protein family called Mramp [56]. Mramp is a trans- sons from Gambia, West Africa, individuals that
porter of iron and other divalent metal cations [56]. had tuberculosis had more than four times the odds of
Mycobacterial Mramp is probably competing with carrying the disease-associated NRAMP1 genotype
host Nramp for control of the iron concentration than ethnically matched controls [61]. The associa-
within the host phagosome (Fig. 1). tion between NRAMP1 polymorphisms and tuber-
The identification of several polymorphisms in culosis has since been confirmed in further patient
the human NRAMP1 gene [57] has enabled linkage populations from Gambia [62], Japan [63], Korea [64,
and association studies to be used to study the im- 65], and Guinea-Conakry [66]. One of the NRAMP1
portance of the NRAMP1 gene in explaining individ- polymorphisms, a (GT)n repeat in the 50-untranslated
ual variability in susceptibility to tuberculosis in region, influences gene expression following lipo-
human populations. Weak evidence of linkage was polysaccharide or interferon gamma (IFNg) stimula-
found between NRAMP1 polymorphisms and tuber- tion [67]. Persons who possess the less actively
culosis in 98 Brazilian families [58] and 173 African expressed allele are more likely to develop tubercu-
sibling-pairs [59]. Strong evidence of linkage between losis. Although these studies confirm that NRAMP1
NRAMP1 and tuberculosis was found in a single influences susceptibility to tuberculosis in human
large aboriginal Canadian family [60]. These results populations, it is likely that NRAMP1 accounts for
suggested that NRAMP1 is involved in susceptibility only a small proportion of the total genetic compo-
to tuberculosis in humans, but that the effect is too nent of tuberculosis susceptibility.
small for NRAMP1 to be the major gene involved.
Association studies have greater power than linkage
studies to evaluate the effects of a candidate gene on
disease susceptibility. Association-based case-control Interferon gamma signaling pathway
studies are therefore more useful than linkage-based
family studies for assessing the effects of NRAMP1 In 1995, Levin et al [68] described four children
on the population-wide variability in risk of tuber- from a village in Malta who had suffered from severe
238 bellamy
and recurrent infections with the atypical myco- chain (IFNgR2) because of IFNcR2 gene mutations
bacteria M. fortuitum, M. chelonae, and M. avium- [83,84]. Deficiency of signal transducer and activator
intracellulare. Three of the children were related, of transcription-1 (STAT-1) protein can lead to de-
suggesting that they had inherited the same genetic creased response to IFNg stimulation [85]. A patient
immune defect [68]. Leukocytes from the affected was described who had a dominant mutation in the
children had impaired IFNg production in response to STAT1 gene causing partial STAT1 deficiency result-
mycobacterial antigens [68]. A genome-wide screen ing in impaired immunity to atypical mycobacteria
identified a single region of homozygosity on chro- [85]. Some patients who have disseminated atypical
mosome 6q in the three related children [69]. The mycobacterial infections have been found to have
region identified contains the gene encoding the IFNg abnormal IFNg production in response to IL-12
receptor ligand binding chain (IFNgR1). Sequencing stimulation but normal cellular responses to IFNg.
the IFNcR1 gene identified a single-base transver- This result can be caused by complete deficiency of
sion at position 395 of the coding sequence, which the IL-12 receptor b1 chain (IL12Rb1) caused by
produced a premature stop codon. The three related recessive mutations in the IL12Rb1 gene [86 – 91] or
children were all homozygous for this single-base by deficiency of the IL-12 p40 subunit caused by
transversions, and the IFNgR protein was absent from mutations in the IL12B gene [92,93].
the children’s leukocytes. This finding established Patients inheriting one of the gene mutations
that IFNgR1 deficiency could be a cause of increased causing abnormal IL-12 – IFNg signaling suffer re-
susceptibility to mycobacterial infections. current infections with mycobacteria that are usually
Independently of Levin’s group, Casanova and nonpathogenic, such as M. bovis BCG, M. avium-
colleagues [70] identified IFNgR1 deficiency as a intracellulare, M. kansasii, M. chelonae, and M. smeg-
cause of disseminated BCG infection. They identified matis [94]. The immune deficiency is relatively
121 French children who developed disseminated specific, because, although recurrent salmonella in-
BCG infection following vaccination. Sixty-one of fections occur, there does not seem to be increased
the children had an identifiable underlying immune susceptibility to a wider range of pathogens [94].
defect such as severe combined immune deficiency, Mendelian susceptibility to mycobacterial infections
chronic granulomatous disease, Di George syndrome, is rare. Most persons who develop tuberculosis do
or HIV infection. Casanova and colleagues attempted not have a recognized IL-12 – IFNg signaling path-
to find an underlying immune deficiency in the re- way defect. Although common gene variants in the
maining 60 children. They screened a number of IFNcR1 gene were not found to be associated with
candidate genes and found that autosomal recessively tuberculosis [82], mutations in other IL-12 – IFNg
inherited mutations in the IFNcR1 gene could lead signaling pathway genes might partly explain popu-
to disseminated BCG infection following vaccina- lation variation in susceptibility to tuberculosis [95].
tion [71]. Complete IFNgR1 deficiency is caused by Lio et al [96] genotyped 45 Sicilian patients who had
gene mutations that abolish receptor expression [69, tuberculosis and 97 controls for a single nucleotide
71 – 76] or binding of the receptor to IFNg [77,78]. polymorphism in the INFc gene and found the
There are two reports of dominant IFNcR1 gene mu- genotype associated with high IFNg production was
tations in patients suffering from disseminated atypi- under-represented among the tuberculosis patients.
cal mycobacterial infections [79,80]. There is also a Case-control studies of the same gene variant in tu-
report of partial IFNgR1 deficiency in two siblings, berculosis patients from Spain [97] and South Africa
one of whom developed disseminated BCG infection [98] found results consistent with this finding,
and the other tuberculosis [81]. This report led to indicating it is unlikely to result from chance. The
speculation that the existence of other common gene population-wide effect of the IFNc gene polymor-
variants of the IFNcR1 gene might explain the popu- phism on susceptibility to tuberculosis is comparable
lation variability in susceptibility to tuberculosis. to that of NRAMP1 [99]. Therefore these gene
When six common IFNcR1 gene polymorphisms were variants can explain only a small percentage of the
typed in a case-control study of 640 persons from total genetic component of the host variability in
Gambia, however, no association with tuberculosis susceptibility to tuberculosis. A provisional associa-
was found [82]. tion has also been described between common
Mutations in four other genes in the interleukin-12 IL-12Rb1 gene variants and tuberculosis in Japanese
(IL-12) – IFNg signaling pathway have been found patients [100]. It is therefore possible that several
to lead to disseminated mycobacterial infections. genes in the IL-12 – IFNg signaling pathway may
Patients have been identified who have complete or contribute to the population-wide variability in tuber-
partial deficiency of the IFNgR signal transduction culosis susceptibility.
genetic susceptibility to tuberculosis 239
bellamy
region associated with disease or other mycobacterial infections or other mycobacterial infections
NRAMP1 Several polymorphisms described Studies of inbred strains of mice with increased Common gene polymorphisms are associated [61 – 66]
function of these remains uncertain susceptibility to mycobacterial infections with susceptibility to tuberculosis
IFNcR1 Gene mutations result in severe Maltese family with Mendelian susceptibility The rare condition of partial or complete [69,71 – 81]
deficiency of the protein to mycobacterial infections and rare individuals deficiency leads to disseminated infections with
with disseminated BCG infections BCG and atypical mycobacteria
IFNcR2 Gene mutations result in severe Rare individuals with disseminated BCG and The rare condition of partial or complete [83,84]
deficiency of the protein atypical mycobacterial infections deficiency leads to disseminated infections with
BCG and atypical mycobacteria
STAT1 Gene mutations result in severe Single individual with disseminated atypical The rare condition of partial deficiency leads to [85]
deficiency of the protein mycobacterial infection disseminated infections with atypical mycobacteria
IL12Rb1 Gene mutations result in severe Rare individuals with disseminated BCG and The rare condition of partial or complete [86 – 91]
deficiency of the protein atypical mycobacterial infections deficiency leads to disseminated infections with
BCG and atypical mycobacteria
IL12B Gene mutations result in severe Rare individuals with disseminated BCG and The rare condition of partial or complete [92,93]
deficiency of the protein atypical mycobacterial infections deficiency leads to disseminated infections with
BCG and atypical mycobacteria
IFNc Gene polymorphism described, Rare individuals with IFNgR1 deficiency A common gene polymorphism is associated with [96 – 98]
function of this remains uncertain suggested IFNg is important in tuberculosis susceptibility to tuberculosis
Mbl2 Gene mutations result in severe Children with recurrent infections and inability Carriers of MBL-variant alleles have been found [112 – 117]
deficiency of the protein to opsonize Baker’s yeast to have increased resistance to tuberculosis,
leprosy and M. avium
VDR Several polymorphisms described Epidemiologic and in vitro data suggested Common gene polymorphisms are associated with [131 – 133]
function of these remains uncertain vitamin D is important in immunity to tuberculosis susceptibility to tuberculosis
HLA class II Large number of different Case-control studies performed because the HLA HLA-DR2 associated with tuberculosis and leprosy [134 – 137]
HLA types system is known to be a key component of the in several populations
acquired immune system
X chromosome Minisatellite markers on Genome-wide linkage analysis carried out on Suggestive evidence of linkage to tuberculosis for [59]
chromosome Xq26 sibling-pair families with tuberculosis Xq26 in Africans
Chromosome 15 Minisatellite markers on Genome-wide linkage analysis carried out on Suggestive evidence of linkage to tuberculosis for [59]
chromosome 15q11-13 sibling-pair families with tuberculosis 15q11-13 in Africans
Abbreviations: BCG, bacille Calmette-Guerin; IFNgR1, interferon-g receptor ligand binding chain; MBL, mannan-binding lectin.
a
Studies assessing the relevance of the gene to susceptibility to tuberculosis or to other mycobacterial infections.
241
242 bellamy
[32] Denis M, Forget A, Pelletier M, et al. Killing of and characterization of a second human NRAMP gene
Mycobacterium smegmatis by macrophages from ge- on chromosome 12q13. Mamm Genome 1995;6(4):
netically susceptible and resistant mice. J Leuk Biol 224 – 30.
1990;47(1):25 – 30. [48] Kishi F, Tabuchi M. Human natural resistance-
[33] Malo D, Vidal SM, Hu J, Skamene E, et al. High associated macrophage protein 2: gene cloning and
resolution linkage map in the vicinity of the host protein identification. Biochem Biophys Res Com-
resistance locus Bcg. Genomics 1993;16(3):655 – 63. mun 1998;251(3):775 – 83.
[34] Gros P, Malo D. A reverse genetics approach to [49] Supek F, Supekova L, Nelson H, et al. A yeast
Bcg/Ity/Lsh gene cloning. Res Immunol 1989;140(8): manganese transporter related to the macrophage
774 – 7. protein involved in conferring resistance to mycobac-
[35] Malo D, Vidal S, Lieman JH, et al. Physical teria. Proc Natl Acad Sci U S A 1996;93(10):5105 – 10.
delineation of the minimal chromosomal segment [50] Fleming MD, Trenor III CC, Su MA, et al. Microcytic
encompassing the murine host resistance locus Bcg. anaemia mice have a mutation in Nramp2, a candidate
Genomics 1993;17(3):667 – 75. iron transporter gene. Nat Genet 1997;16(4):383 – 6.
[36] Vidal SM, Malo D, Vogan K, et al. Natural resistance [51] Fleming MD, Romano MA, Su MA, et al. Nramp2 is
to infection with intracellular parasites: isolation of mutated in the anaemic Belgrade (b) rat: evidence of
a candidate gene for Bcg. Cell 1993;73(3):469 – 85. a role for Nramp2 in endosomal iron transport. Proc
[37] Malo D, Vogan K, Vidal S, et al. Haplotype mapping Natl Acad Sci U S A 1998;95(3):1148 – 53.
and sequence analysis of the mouse Nramp gene [52] Gruenheid S, Canonne-Hergaux F, Gauthier S, et al.
predict susceptibility to infection with intracellular The iron transport protein Nramp2 is an integral
parasites. Genomics 1994;23(1):51 – 61. membrane glycoprotein that colocalizes with transfer-
[38] Vidal S, Tremblay ML, Govoni G, et al. The Ity/Lsh/ rin in recycling endosomes. J Exp Med 1999;189(5):
Bcg locus: natural resistance to infection with intra- 831 – 41.
cellular parasites is abrogated by disruption of the [53] Tabuchi M, Yoshimori T, Yamaguchi K, et al. Human
Nramp1 gene. J Exp Med 1995;182(3):655 – 66. NRAMP2/DMT1, which mediates iron transport
[39] Govoni G, Vidal S, Gauthier S, et al. The Bcg/Ity/Lsh across membranes, is localized to late endosomes and
locus: genetic transfer of resistance to infections in lysosomes in Hep-2 cells. J Biol Chem 2000;275(29):
C57BL/6J mice transgenic for the Nramp1G169 allele. 22220 – 8.
Infect Immun 1996;64(8):2923 – 9. [54] Gruenheid S, Pinner E, Desjardins M, et al. Natural
[40] Barton CH, White JK, Roach TIA, et al. NH2- resistance to infection with intracellular pathogens:
terminal sequence of macrophage-expressed natural the Nramp1 protein is recruited to the membrane
resistance-associated macrophage protein (Nramp) of the phagosome. J Exp Med 1997;185(4):717 – 30.
encodes a proline/ serine-rich putative Src homology [55] Gomes MS, Appelberg R. Evidence for a link
3-binding domain. J Exp Med 1994;179(5):1683 – 7. between iron metabolism and Nramp1 gene function
[41] Govoni G, Vidal S, Cellier M, et al. Genomic in innate resistance against Mycobacterium avium.
structure, promoter sequence, and induction of ex- Immunology 1998;95(2):165 – 8.
pression of the mouse Nramp1 gene in macrophages. [56] Agranoff D, Monahan IM, Mangan JA, et al.
Genomics 1995;27(1):9 – 19. Mycobacterium tuberculosis expresses a novel pH-
[42] Vidal SM, Pinner E, Lepage P, et al. Natural dependent divalent cation transporter belonging to the
resistance to intracellular infections. Nramp1 encodes Nramp family. J Exp Med 1999;190(5):717 – 24.
a membrane phosphoglycoprotein absent in macro- [57] Liu J, Fujiwara M, Buu NT, et al. Identification of
phages from susceptible (Nramp1D169 ) mouse strains. polymorphisms and sequence variants in the human
J Immunol 1996;157(8):3559 – 68. homologue of the mouse natural resistance-associated
[43] Cellier M, Govoni G, Vidal S, et al. Human natu- macrophage protein gene. Am J Hum Genet 1995;
ral resistance-associated macrophage protein: cDNA 56(4):845 – 53.
cloning, chromosomal mapping, genomic organiza- [58] Shaw MA, Collins A, Peacock CS, et al. Evidence
tion, and tissue-specific expression. J Exp Med 1994; that genetic susceptibility to Mycobacterium tuber-
180(5):1741 – 52. culosis in a Brazilian population is under oligo-
[44] Blackwell JM, Barton H, White JK, et al. Genomic genic control: linkage study of the candidate genes
organization and sequence of the human NRAMP NRAMP1 and TNFA. Tuber Lung Dis 1997;78(1):
gene: identification and mapping of a promoter region 35 – 45.
polymorphism. Mol Med 1995;1(2):194 – 205. [59] Bellamy R, Beyers N, McAdam KPWJ, et al. Genetic
[45] Gruenheid S, Cellier M, Vidal S, et al. Identification susceptibility to tuberculosis in Africans: a genome-
and characterization of a second mouse Nramp gene. wide scan. Proc Natl Acad Sci U S A 2000;97(14):
Genomics 1995;25(2):514 – 25. 8005 – 9.
[46] Dosik JK, Barton CH, Holiday DL, et al. An Nramp- [60] Greenwood CM, Fujiwara TM, Boothroyd LJ, et al.
related sequence maps to mouse chromosome 17. Linkage of tuberculosis to chromosome 2q35 loci,
Mamm Genome 1994;5(7):458 – 60. including NRAMP1, in a large aboriginal Canadian
[47] Vidal S, Belouchi A-M, Cellier M, et al. Cloning family. Am J Hum Genet 2000;67(2):405 – 16.
244 bellamy
[61] Bellamy R, Ruwende C, Corrah T, et al. Variations (IFNgR1) deficiency: mutational analysis and evalua-
in the NRAMP1 gene and susceptibility to tubercu- tion of therapeutic options. Exp Haematol 1999;
losis in West Africans. N Engl J Med 1998;338(10): 27(9):1368 – 74.
640 – 4. [76] Cunningham JA, Kellner JD, Bridge PJ, et al.
[62] Awomoyi AA, Marchant A, Howson JM, et al. Disseminated bacille Calmette-Guerin infection in
Interleukin-10, polymorphism in SLC11A1 (formerly an infant with a novel deletion in the interferon-g-
NRAMP1), and susceptibility to tuberculosis. J Infect receptor gene. Int J Tuberc Lung Dis 2000;4(8):
Dis 2002;186(12):1808 – 14. 791 – 7.
[63] Gao P-S, Fujishima S, Mao X-Q, et al. Genetic [77] Jouanguy E, Dupuis S, Pallier A, et al. In a novel
variants of NRAMP1 and active tuberculosis in form of IFNc receptor 1 deficiency, cell surface re-
Japanese populations. Clin Genet 2000;58(1):74 – 6. ceptors fail to bind IFN-g. J Clin Invest 2000;
[64] Ryu S, Park YK, Bai GH, et al. 30UTR poly- 105(10):1429 – 36.
morphisms in the NRAMP1 gene are associated with [78] Allende LM, Lopez-Goyenes A, Paz-Artal E, et al. A
susceptibility to tuberculosis in Koreans. Int J Tuberc point mutation in a domain of gamma interferon
Lung Dis 2000;4(6):577 – 80. receptor 1 provokes severe immunodeficiency. Clin
[65] Kim JH, Lee SY, Lee SH, et al. NRAMP1 genetic Diagn Lab Immunol 2001;8(1):133 – 7.
polymorphisms as a risk factor of tuberculous pleu- [79] Jouanguy E, Lamhamedi-Cherradi S, Lammas D,
risy. Int J Tuberc Lung Dis 2003;7(4):370 – 5. et al. A human IFNGR1 small deletion hotspot as-
[66] Cervino AC, Lakiss S, Sow O, et al. Allelic sociated with dominant susceptibility to mycobacterial
association between the NRAMP1 gene and suscep- infection. Nat Genet 1999;21(4):370 – 8.
tibility to tuberculosis in Guinea-Conakry. Ann Hum [80] Villela A, Picard C, Jouanguy E, et al. Recurrent
Genet 2000;64(6):507 – 12. Mycobacterium avium osteomyelitis associated with
[67] Searle S, Blackwell JM. Evidence for a functional a novel dominant interferon gamma receptor muta-
repeat polymorphism in the promoter of the human tion. Pediatrics 2001;107(4):e47.
NRAMP1 gene that correlates with autoimmune [81] Jouanguy E, Lamhamedi-Cherradi S, Altare F, et al.
versus infectious disease susceptibility. J Med Genet Partial interferon-g receptor 1 deficiency in a child
1999;36(4):295 – 9. with tuberculoid bacillus Calmette-Guerin infection
[68] Levin M, Newport MJ, D’Souza S, et al. Famil- and a sibling with clinical tuberculosis. J Clin Invest
ial disseminated atypical mycobacterial infection 1997;100(11):2658 – 64.
in childhood: a human mycobacterial susceptibility [82] Awomoyi AA, Nejentsev S, Richardson A, et al.
gene? Lancet 1995;345(8942):79 – 83. No association between interferon-g receptor-1 gene
[69] Newport MJ, Huxley CM, Huston S, et al. A mutation polymorphisms and pulmonary tuberculosis in a Gam-
in the interferon-g-receptor gene and susceptibility to bian population sample. Thorax 2004;59(4):291 – 4.
mycobacterial infection. N Engl J Med 1996;335(26): [83] Dorman SE, Holland SM. Mutation in the signal-
1941 – 9. transducing chain of the interferon-g receptor and
[70] Casanova J-L, Blanche S, Emile J-F, et al. Idiopathic susceptibility to mycobacterial infection. J Clin Invest
disseminated bacillus Calmette-Guerin infection: 1998;101(11):2364 – 9.
a French national retrospective study. Pediatrics 1996; [84] Doffinger R, Jouanguy E, Dupuis S, et al. Partial
98(4):774 – 8. interferon-g receptor signaling chain deficiency in a
[71] Jouanguy E, Altare F, Lamhamedi S, et al. Interferon- patient with bacille Calmette-Guerin and Mycobacte-
g-receptor deficiency in an infant with fatal bacille rium abscessus infection. J Infect Dis 2000;181(1):
Calmette-Guerin infection. N Engl J Med 1996; 379 – 84.
335(26):1956 – 61. [85] Dupuis S, Dargemont C, Fieschi C, et al. Impair-
[72] Pierre-Audigier C, Jouanguy E, Lamhamedi S, et al. ment of mycobacterial but not viral immunity by a
Fatal disseminated Mycobacterium smegmatis infec- germline STAT1 mutation. Science 2000;293(5528):
tion in a child with inherited interferon g receptor 300 – 3.
deficiency. Clin Infect Dis 1997;24(5):982 – 4. [86] Altare F, Durandy A, Lammas D, et al. Impairment
[73] Altare F, Jouanguy E, Lamhamedi-Cherradi S, et al. of mycobacterial immunity in human interleukin-12
A causative relationship between mutant IFNcR1 receptor deficiency. Science 1998;280(5368):1432 – 5.
alleles and impaired cellular response to IFNg in [87] de Jong R, Altare F, Haagen I-A, et al. Severe myco-
a compound heterozygous child. Am J Hum Genet bacterial and Salmonella infections in interleukin-12
1998;62(3):723 – 6. receptor-deficient patients. Science 1998;280(5368):
[74] Holland SM, Dorman SE, Kwon A, et al. Abnormal 1435 – 8.
regulation of interferon-g, interleukin-12, and tumour [88] Verhagen CE, de Boer T, Smits HH, et al. Residual
necrosis factor-a in human interferon-g receptor 1 type 1 immunity in patients genetically deficient for
deficiency. J Infect Dis 1998;178(4):1095 – 104. interleukin 12 receptor b1 (IL12Rb1): evidence for
[75] Roesler J, Kofink B, Wendisch J, et al. Listeria an IL12Rb1-independent pathway of IL-12 respon-
monocytogenes and recurrent mycobacterial infec- siveness in human T cells. J Exp Med 2000;192(4):
tions in a child with complete interferon-g-receptor 517 – 28.
genetic susceptibility to tuberculosis 245
[89] Aksu G, Trpan C, Cavusoglu C, et al. Mycobacterium frequencies in African and non-African populations
fortuitum-chelonae complex infection in a child with of independent mutations in the mannose binding
complete interleukin-12 receptor b1 deficiency. protein gene. Hum Mol Genet 1992;1(9):709 – 15.
Pediatr Infect Dis J 2001;20(5):551 – 3. [106] Madsen HO, Garred P, Kurtzhals JAL, et al. A new
[90] Altare F, Ensser A, Breiman A, et al. IL-12 receptor frequent allele is the missing link in the structural
b1 deficiency in a patient with abdominal tuber- polymorphism of the human mannan-binding protein.
culosis. J Infect 2001;184(2):231 – 6. Immunogenetics 1994;40(1):37 – 44.
[91] Sakai T, Matsuoka M, Aoki M, et al. Missense [107] Jack DL, Klein NJ, Turner MW. Mannose-binding
mutation of the interleukin-12 receptor b1 chain- lectin deficiency and susceptibility to infectious dis-
encoding gene is associated with impaired immunity ease. In: Bellamy R, editor. Susceptibility to infec-
against Mycobacterium avium complex infection. tious diseases: the importance of host genetics.
Blood 2001;97(9):2688 – 94. Cambridge (UK)7 Cambridge University Press; 2004.
[92] Altare F, Lammas D, Revy P, et al. Inherited p. 279 – 307.
interleukin 12 deficiency in a child with bacille [108] Summerfield JA, Sumiya M, Levin M, et al.
Calmette-Guerin and Salmonella enteritidis dissemi- Association of mutations in mannose binding protein
nated infection. J Clin Invest 1998;102(12):2035 – 40. gene with childhood infection in consecutive hospital
[93] Picard C, Fieschi C, Altare F, et al. Inherited series. BMJ 1997;314(7089):1229 – 32.
interleukin-12 deficiency: IL12B genotype and clini- [109] Koch A, Melbye M, Sorensen P, et al. Acute re-
cal phenotype of 13 patients from six kindreds. Am J spiratory tract infections and mannose binding lectin
Hum Genet 2002;70(2):336 – 48. insufficiency during early childhood. JAMA 2001;
[94] Casanova JL, Abel L. Genetic dissection of immu- 285(10):1316 – 21.
nity to mycobacteria: the human model. Annu Rev [110] Summerfield JA, Ryder S, Sumiya M, et al. Mannose-
Immunol 2002;20:581 – 620. binding protein gene-mutations associated with un-
[95] Dupuis S, Doffinger R, Picard C, et al. Human usual and severe infections in adults. Lancet 1995;
interferon-g-mediated immunity is a genetically con- 345:886 – 9.
trolled continuous trait that determines the outcome of [111] Garred P, Madsen HO, Hoffmann B, et al. Increased
mycobacterial invasion. Immunol Rev 2000;178: frequency of homozygosity of abnormal mannan-
129 – 37. binding protein alleles in patients with suspected
[96] Lio D, Marino V, Serauto A, et al. Genotype immunodeficiency. Lancet 1995;346(8980):941 – 3.
frequencies of the + 874T!A single nucleotide poly- [112] Garred P, Harboe M, Oettinger T, et al. Dual role of
morphism in the first intron of the interferon-gamma mannan-binding protein in infections: another case
gene in a sample of Sicilian patients affected by of heterosis? Eur J Hum Genet 1994;21(2):125 – 31.
tuberculosis. Eur J Immunogenet 2002;29(5):371 – 4. [113] Garred P, Richter C, Andersen AB, et al. Mannan-
[97] Lopez-Maderuelo D, Arnalich F, Serantes R, et al. binding lectin in the sub-Saharan HIV and tuberculo-
Interferon gamma and interleukin-10 gene polymor- sis epidemics. Scand J Immunol 1997;46(2):204 – 8.
phisms in pulmonary tuberculosis. Am J Respir Crit [114] Bellamy R, Ruwende C, McAdam KPWJ, et al.
Care Med 2003;167(7):970 – 5. Mannose binding protein deficiency is not associated
[98] Rossouw M, Nel HJ, Cooke GS, et al. Association with increased susceptibility to malaria, hepatitis B
between tuberculosis and a polymorphic NFkB carriage nor tuberculosis in Africans. QJM 1998;
binding site in the interferon g gene. Lancet 2003; 91(1):13 – 8.
361(9372):1871 – 2. [115] Hoal-Van Helden EG, Epstein J, Victor TC, et al.
[99] Bellamy R. Interferon gamma and host susceptibility Mannose-binding protein B allele confers protection
to tuberculosis. J Resp Crit Care Med 2003;167(7): against tuberculosis meningitis. Pediatr Res 1999;
946 – 7. 45(4):459 – 64.
[100] Akahoshi M, Nakashima H, Miyake K, et al. Influ- [116] Soborg C, Madsen HO, Andersen AB, et al.
ence of interleukin-12 receptor beta1 polymorphisms Mannose-binding lectin polymorphisms in clinical
on tuberculosis. Hum Genet 2003;112(3):237 – 43. tuberculosis. J Infect Dis 2003;188(5):777 – 82.
[101] Kawasaki T. Structure and biology of mannan- [117] Polotsky VY, Belisle BT, Mikusova K, et al. Inter-
binding protein, MBP, an important component of in- action of human mannose-binding lectin with Myco-
nate immunity. Biochim Biophys Acta 1999;1473(1): bacterium avium. J Infect Dis 1997;175(5):1159 – 68.
186 – 95. [118] Selvaraj P, Narayanan PR, Reetha AM. Association
[102] Turner MW. The role of mannose-binding lectin in of functional mutant homozygotes of the mannose
health and disease. Mol Immunol 2003;40(7):423 – 9. binding protein gene with susceptibility to pulmonary
[103] Thompson C. Protein proves to be a key link in innate tuberculosis in India. Tuber Lung Dis 1999;79(4):
immunity. Science 1995;269(5222):301 – 2. 221 – 7.
[104] Sumiya M, Super M, Tabona P, et al. Molecular [119] El Sahly HM, Reich RA, Dou SJ, et al. The effect
basis of opsonic defect in immunodeficient children. of mannose binding lectin gene polymorphisms on
Lancet 1991;337(8757):1569 – 70. susceptibility to tuberculosis in different ethnic
[105] Lipscombe RJ, Sumiya M, Hill AV, et al. High groups. Scand J Infect Dis 2004;36(2):106 – 8.
246 bellamy
[120] Rook GAW, Steele J, Fraher L, et al. Vitamin D3, and variation in the vitamin D receptor gene. J Infect
gamma interferon, and control of Mycobacterium tu- Dis 1999;179(3):722 – 4.
berculosis by human monocytes. Immunology 1986; [132] Wilkinson RJ, Llewelyn M, Toossi Z, et al. Influence
57(1):159 – 63. of vitamin D deficiency and vitamin D receptor poly-
[121] Tsoukas CD, Provvedini DM, Manolagas SC. 1,25- morphisms on tuberculosis amongst Gujarati Asians
dihydroxyvitamin D3: a novel immunoregulatory hor- in west London. Lancet 2000;355(9204):618 – 21.
mone. Science 1984;224(4656):1438 – 40. [133] Roy S, Frodsham A, Saha B, et al. Association
[122] Lemire JM, Adams JS, Sakai R, et al. 1a, 25-dihy- of vitamin D receptor genotype with leprosy type.
droxyvitamin D3 suppresses proliferation and immu- J Infect Dis 1999;179(1):187 – 91.
noglobulin production by normal human peripheral [134] Singh SPN, Mehra NK, Dingley HB, et al. Human
blood mononuclear cells. J Clin Invest 1984;74(2): leukocyte antigen (HLA)-linked control of suscepti-
657 – 61. bility to tuberculosis and association with HLA-DR
[123] Rook GAW, Taverne J, Leveton C, et al. The role types. J Infect Dis 1983;148(4):676 – 81.
of gamma-interferon, vitamin D3 metabolites and [135] Bothamley GH, Beck JS, Schreuder GMT, et al. As-
tumour necrosis factor in the pathogenesis of tuber- sociation of tuberculosis and tuberculosis-specific
culosis. Immunology 1987;62(2):229 – 34. antibody levels with HLA. J Infect Dis 1989;159(3):
[124] Denis M. Killing of Mycobacterium tuberculosis 549 – 55.
within human monocytes: activation by cytokines [136] Todd JR, West BC, McDonald JC. Human leukocyte
and calcitriol. Clin Exp Immunol 1991;84(2):200 – 6. antigen and leprosy: study in Northern Louisiana
[125] Strachan DP, Powell KJ, Thaker A, et al. Vegetarian and review. Rev Infect Dis 1990;12(1):63 – 74.
diet as a risk factor for tuberculosis in immigrant [137] Brahmajothi V, Pitchappan RM, Kakkanaiah VN,
south London Asians. Thorax 1995;50(2):175 – 80. et al. Association of pulmonary tuberculosis and HLA
[126] Davies PDO, Brown RC, Woodhead JS. Serum in South India. Tubercle 1991;72(2):123 – 32.
concentrations of vitamin D metabolites in untreated [138] Cox RA, Downs M, Neimes RE, et al. Immuno-
tuberculosis. Thorax 1985;40(3):187 – 90. genetic analysis of human tuberculosis. J Infect Dis
[127] Grange JM, Davies PDO, Brown RC, et al. A study 1988;158(6):1302 – 8.
of vitamin D levels in Indonesian patients with un- [139] Goldfeld AE, Delgado JC, Thim S, et al. Associa-
treated pulmonary tuberculosis. Tubercle 1985;66(3): tion of an HLA-DQ allele with clinical tuberculosis.
187 – 91. JAMA 1998;279(3):226 – 8.
[128] Reichel H, Koeffler HP, Tobler A, et al. 1a, [140] Sharma SK, Turaga KK, Balamurugan A, et al.
25- dihydroxyvitamin D3 inhabits gamma-interferon Clinical and genetic risk factors for the development
synthesis by normal human peripheral blood lym- of multi-drug resistant tuberculosis in non-HIV in-
phocytes. Proc Natl Acad Sci U S A 1987;84(10): fected patients at a tertiary care center in India: a case-
3385 – 9. control study. Infect Genet Evol 2003;3(3):183 – 8.
[129] Provvedini DM, Tsoukas CD, Deftos LJ, et al. 1,25 [141] Balamurugan A, Sharma SK, Mehra NK. Human
dihydroxyvitamin D3 receptors in human leukocytes. leukocyte antigen supertypes influence susceptibility
Science 1983;221(4616):1181 – 3. and severity of tuberculosis. J Infect Dis 2004;189(5):
[130] Morrison NA, Qi JC, Tokita A, et al. Prediction of 805 – 11.
bone density from vitamin D receptor alleles. Nature [142] Campbell SJ, Sabeti P, Fielding K, et al. Variants of
1994;367(6460):284 – 7. the CD40 ligand gene are not associated with in-
[131] Bellamy R, Ruwende C, Corrah T, et al. Tuberculosis creased susceptibility to tuberculosis in West Africa.
and chronic hepatitis B virus infection in Africans Immunogenet 2003;55(7):502 – 7.
Clin Chest Med 26 (2005) 167 – 182
In 1993, the World Health Organization (WHO) The size of the burden of tuberculosis
declared tuberculosis a global emergency because of
the scale of the epidemic and the urgent need to Tuberculosis case notifications and reported deaths
improve global tuberculosis control [1,1a]. Since Tuberculosis notification data are important and
then, WHO has promoted the strategy for global are routinely reported by WHO [6]. At the country
tuberculosis control known as DOTS (a name derived level, a system of recording and reporting tuber-
originally from directly observed treatment, short- culosis cases and their treatment outcomes (including
course) [2,3] and its adaptations (eg, as part of a death) is an intrinsic part of the DOTS strategy
strategy of expanded scope where HIV prevalence (Box 1). Therefore as the number of countries
is high [4] and as DOTS-Plus in areas where the implementing the DOTS strategy increases, routine
prevalence of multidrug-resistant [MDR] tuberculosis national tuberculosis program (NTP) data on tuber-
is high) [5]. This article provides an overview of the culosis cases and deaths are becoming more widely
current scale of the global tuberculosis epidemic. It available [6]. Notification data reflect health ser-
describes the global tuberculosis situation as mea- vice coverage and the efficiency of case-finding and
sured by reported and estimated cases and deaths. reporting activities of NTPs. Thus, in the developing
The increasing threats of HIV-related tuberculosis countries where tuberculosis incidence is generally
and drug-resistant tuberculosis receive particular high, where access to health services may be limited,
attention. There is a brief review of the extent of im- and where NTP performance may be suboptimal,
plementation of effective tuberculosis control using notification data often represent only a fraction of
the DOTS strategy. The article ends with a summary the true incident cases. In addition, because case
of the approaches needed to accelerate progress in definitions vary among countries (eg, when some
global tuberculosis control. countries’ notification data include all cases, both
new and re-treatment cases), comparisons of case
notification data from different countries are difficult.
Review of the global tuberculosis epidemic In industrialized countries, however, where tuber-
culosis incidence is generally low, where health ser-
As part of the description of the global tuber- vice coverage is generally universal, and where NTPs
culosis epidemic, the size of the burden of tuber- are effective, notifications of cases often closely ap-
culosis indicates progress in tuberculosis control and proximate to the true incidence of tuberculosis. In
draws attention to the scale of the problem, thereby any country, under stable program conditions, case
helping to mobilize resources for tuberculosis control. notifications may provide useful data on the trend of
incidence and a means for obtaining rates by age, sex,
and risk group.
Despite the limitations of tuberculosis case noti-
fications, WHO has since 1997 published worldwide
* Corresponding author. data provided by its member states, most recently
E-mail address: raviglionem@who.ch (M. Raviglione). referring to the 4.1 million cases reported in 2002 [6].
0272-5231/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccm.2005.02.009 chestmed.theclinics.com
168 maher & raviglione
Fig. 1. Tuberculosis case notification rates by country in 2002. The designations employed and the presentation of material on
this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the
legal status of any country, territory, city or area, or of its authorities, or concerning the delimitation of its frontiers or boundaries.
White lines on maps represent approximate border lines for which there may not yet be full agreement. [From World Health
Organization. Global tuberculosis control: surveillance, planning, financing. WHO report 2004. Document WHO/HTM/TB/
2004.331. Geneva (Switzerland): World Health Organization; 2004. p. 218, fig. 4; with permission.]
are declining [12]. In other industrialized countries, living conditions (resulting in malnutrition, crowding,
including Australia, New Zealand, and Canada, rates and stress) and in some cases civil conflicts and
have leveled off during the past few years below wars, deteriorating health services, and lack of drugs,
10 per 100,000. The proportion of cases in the resulting in decreased rates of cure of tuberculosis
foreign-born is about 70% in Australia and about patients and continued transmission in the commu-
50% in Canada [12]. The implication of the high nity. The spread of HIV in some countries, particu-
proportion of cases in the foreign-born in most indus- larly the Russian Federation and Ukraine, has the
trialized countries is that tuberculosis control in these potential, if unchecked, to fuel the tuberculosis epi-
settings depends on tuberculosis control globally. demic further.
Tuberculosis case notification rates are still high Data on tuberculosis deaths are reported through
in the countries of the former Soviet Union [6]. In national vital registration systems and through the
many countries the previous continued decline in case routine standard NTP recording and reporting system.
notifications stopped or reversed from the early 1990s Few developing countries have comprehensive vital
onwards. For example, annual notification rates registration systems for the accurate reporting of
doubled in Russia from 1990 to 2002, with an in- deaths. Routine NTP data on tuberculosis deaths are
creased proportion of cases in young adults [6]. becoming more widely available in developing
Dramatic social changes following the end of the countries [6]. NTPs report these tuberculosis-cohort
Soviet Union engendered a combination of factors deaths (the number and proportion of tuberculosis
responsible for the reversal of the previous trend, patients dying during treatment) without specifying
probably through increased susceptibility to infection cause, because the cause of death can rarely be
and increased breakdown to disease after infection. determined in countries where income is low and the
These factors include increased poverty and poor prevalence of tuberculosis is high [13]. Inaccurate
170 maher & raviglione
routine NTP reporting of cohort deaths and incom- [6]. Table 2 summarizes tuberculosis incidence and
plete NTP coverage of all incident cases in many mortality estimates in 2002 by WHO regions [6,14].
countries limit the extent to which tuberculosis cohort Fig. 2 shows estimated tuberculosis incidence by
deaths reflect tuberculosis mortality. country for 2002 [6]. The ranking of countries by
number of tuberculosis cases draws attention to the
22 countries that account for roughly 80% of the
Estimated tuberculosis cases and deaths world’s tuberculosis burden (Table 3). Developing
Because of the limitations of tuberculosis notifi- countries suffer the brunt of the tuberculosis epi-
cations and the difficulties in directly measuring the demic. Overall, it is estimated that 95% of the world’s
numbers of cases and deaths, the size of the tuber- tuberculosis cases and 98% of the tuberculosis deaths
culosis disease burden must be estimated. WHO esti- occur in the developing world [12], and that tuber-
mates of tuberculosis incidence and deaths are based culosis causes more than 25% of avoidable adult
on a variety of inputs, including surveys of preva- deaths in the developing world [16]. The importance
lence of tuberculosis infection and disease, vital of the tuberculosis problem for individual countries is
registration data, and independent assessments of expressed as the annual incidence (absolute number
quality of surveillance systems [14,15]. In 2002 there of cases occurring yearly) and as the incidence rate
were an estimated 8.8 million new cases of tuber- (cases per 100,000 population). Fig. 3 shows esti-
culosis worldwide, with an incidence rate of 141 per mated tuberculosis incidence rates by country in 2002
100,000 population [6]. The global incidence rate of [6]. Tuberculosis incidence rates are generally much
tuberculosis is growing at approximately 1.1% per lower in industrialized countries than in developing
year, although this overall global trend is fueled by countries. Among the 15 countries with the highest
and hides much faster increases in sub-Saharan estimated tuberculosis incidence rates, 13 are in sub-
Africa and in countries of the former Soviet Union Saharan Africa, and in most of these countries the
Table 2
Summary of tuberculosis estimates by World Health Organization region in 2002
WHO region
AFRa AMR EMR EUR SEAR WPR Global
Population (millions) 672 857 507 877 1591 1718 6222
New cases of TB (all forms)
No. of incident cases (thousands) 2354 370 622 472 2890 2090 8798
Incidence rate (per 100,000) 350 43 123 54 182 122 141
Change in incidence rate 1997 – 2000 (%/y) 5.9 3.6 0.7 1.9 2.1 0.2 1.1
HIV prevalence in new adult cases (%) 37.0 5.5 2.8 3.6 3.5 1.2 12.0
Attributable to HIV (thousands) 506.0 11.0 9.8 10.0 56.0 14.0 656.0
Attributable to HIV (% of adult cases) 31.0 5.0 2.5 3.3 2.9 1.1 11.0
New SS+ cases of TB
No. of incident cases (thousands) 1000 165 279 211 1294 939 3888
Prevalence rate of SS+ TB (per 100,000) 224 25 102 34 166 104 112
Prevalent SS+ cases HIV+ (%) 6.9 1.0 0.4 0.7 0.5 0.2 1.8
Deaths from TB
No. of deaths from TB (thousands) 556 53 143 73 625 373 1823
Deaths from TB (per 100,000) 83.0 6.2 28.0 8.3 39.0 22.0 29.0
Deaths from TB in HIV-positive adults (thousands) 208.0 3.7 4.8 3.0 26.0 5.5 251.0
Adult AIDS deaths caused by TB (%) 15.0 5.4 20.0 13.0 7.6 14.0 13.0
TB deaths attributable to HIV (%) 34.0 6.5 3.2 3.9 3.8 1.4 13.0
Abbreviations: adult, 15 – 49 years old; AFR, African; AMR, Americas; EMR, Eastern Mediterranean; EUR, European; SEAR,
Southeast Asia; SS+, sputum smear-positive; TB, tuberculosis; WPR, Western Pacific.
a
WHO African region comprises sub-Saharan Africa and Algeria. The remaining North African countries are included in
the WHO Eastern Mediterranean region.
Data from World Health Organization. Global tuberculosis control: surveillance, planning, financing. WHO report 2004.
Document WHO/HTM/TB/2004.331. Geneva (Switzerland): World Health Organization; 2004; and Corbett EL, Watt CJ,
Walker N, et al. The growing burden of tuberculosis: global trends and interactions with the HIV epidemic. Arch Intern Med
2003;163:1009 – 21.
global epidemiology of tuberculosis 171
Fig. 2. Estimated tuberculosis incidence by country in 2002. The designations employed and the presentation of material on this
map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal
status of any country, territory, city or area, or of its authorities, or concerning the delimitation of its frontiers or boundaries.
White lines on maps represent approximate border lines for which there may not yet be full agreement. [From World Health
Organization. Global tuberculosis control: surveillance, planning, financing. WHO report 2004. Document WHO/HTM/TB/
2004.331. Geneva (Switzerland): World Health Organization; 2004; with permission.]
prevalence of HIV infection among tuberculosis factor for progression of dormant M. tuberculosis
patients is high [6]. infection to clinical tuberculosis disease [18]. A short
In conclusion, worldwide notification data and overview of HIV epidemiology is useful because HIV
estimates suggest a steady decline in the tuberculosis is such an important force driving the tuberculosis
burden in many regions except in sub-Saharan Africa epidemic in sub-Saharan Africa and has the potential
and the former Soviet Union. The reasons for the to drive the tuberculosis epidemic in other regions
persisting global tuberculosis burden include (1) pov- wherever HIV transmission spreads unchecked. HIV
erty and the widening gap between rich and poor surveillance systems in most countries with gener-
in various populations (eg, developing countries, alized epidemics rely on tracking HIV prevalence
inner city populations in developed countries); among pregnant women attending antenatal clinics.
(2) previous neglect of tuberculosis control (inad- These antenatal clinic data, supplemented by data
equate case detection, diagnosis, and cure); (3) from other sources such as blood donors and sex
changing demography (increasing world population workers, are used to obtain national estimates of HIV
and changing age structure); and (4) the impact of the prevalence among men and women and to assess
HIV pandemic [17]. trends. By the end of 2003, an estimated 38 million
adults and children worldwide had HIV infection or
AIDS [19]. Of these, 25 million (66%) were in sub-
HIV-related tuberculosis Saharan Africa, and 6.5 million (17%) were in South
and South-East Asia. In 2003, 4.8 million adults and
Through potent immunocompromise of infected children were newly infected with HIV. An estimated
hosts, HIV has emerged as the most important risk 2.9 million adults and children died from HIV/AIDS
172 maher & raviglione
Table 3
Ranking of countries by estimated number of tuberculosis cases
Number estimated
All cases Smear-positive cases
Rate Rate
Population No. (per 100,000 No. (per 100,000 Cumulative
Rank Country (thousands) (thousands) population) (thousands) population) incidence (%)
1 India 1,049,549 1761 168 787 75 20
2 China 1,294,867 1459 113 656 51 37
3 Indonesia 217,131 557 256 250 115 43
4 Nigeria 120,911 368 304 159 132 47
5 Bangladesh 143,809 318 221 143 99 51
6 Pakistan 149,911 272 181 122 81 54
7 Ethiopia 68,961 255 370 110 159 57
8 Philippines 78,580 251 320 113 144 60
9 South Africa 44,759 250 558 102 227 62
10 Democratic Republic 51,201 196 383 85 167 65
of the Congo
11 Russian Federation 144,082 182 126 81 56 67
12 Kenya 31,540 170 540 70 223 69
13 Vietnam 80,278 155 192 69 86 70
14 United Republic 36,276 132 363 56 155 72
of Tanzania
15 Brazil 176,257 110 62 49 28 73
16 Uganda 25,004 94 377 41 164 74
17 Zimbabwe 12,835 88 683 35 271 75
18 Mozambique 18,537 81 436 34 182 76
19 Thailand 62,193 80 128 35 57 77
20 Afghanistan 22,930 76 333 34 150 78
21 Cambodia 13,810 76 549 33 242 79
22 Myanmar 48,852 75 154 33 68 80
High-burden countries 3,892,274 7005 180 3100 80 80
during 2003. Roughly 2.2 million (76%) of these some other large populations, for example in the
deaths occurred in sub-Saharan Africa. Sub-Saharan Russian Federation.
Africa is the region with the highest overall HIV
prevalence rate in the general adult (15 – 49 years)
population, 7.5% at the end of 2003. Tuberculosis cases
Of 20 countries in the world with an adult HIV The HIV pandemic has dramatically fuelled
prevalence rate in 2003 above 5%, 19 are in sub- tuberculosis in populations where there is overlap
Saharan Africa (the other is Haiti). In seven countries between those infected with M. tuberculosis and
in southern Africa, adult HIV prevalence is 15% or those infected with HIV. Table 4 shows the number
above. Although the countries that have the highest of M. tuberculosis- and HIV-coinfected adults (15 –
rates of HIV infection are in Africa, certain countries 49 years) in WHO regions and globally by the end of
in South-East Asia and Latin America are also badly 2000 [14]. Of the 11.4 million adults coinfected with
affected, with an adult HIV prevalence of 1% to 5%. M. tuberculosis and HIV worldwide by the end of
Although the rise in HIV prevalence seems now to be 2000, 70% were in sub-Saharan Africa (Table 4) [14].
decelerating or even decreasing in parts of Eastern The estimated national HIV prevalence in tuber-
and Southern Africa, it is still increasing rapidly in culosis patients reflects the extent of the overlap
global epidemiology of tuberculosis 173
Fig. 3. Estimated tuberculosis incidence rates by country in 2002. The designations employed and the presentation of material on
this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the
legal status of any country, territory, city or area, or of its authorities, or concerning the delimitation of its frontiers or boundaries.
White lines on maps represent approximate border lines for which there may not yet be full agreement. [From World Health
Organization. Global tuberculosis control: surveillance, planning, financing. WHO report 2004. Document WHO/HTM/TB/
2004.331. Geneva (Switzerland): World Health Organization; 2004. p. 215, fig. 1; with permission.]
Fig. 4. Estimated HIV prevalence in tuberculosis cases by country in 2002. The designations employed and the presentation of
material on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization
concerning the legal status of any country, territory, city or area, or of its authorities, or concerning the delimitation of its frontiers
or boundaries. White lines on maps represent approximate border lines for which there may not yet be full agreement. [From
World Health Organization. Global tuberculosis control: surveillance, planning, financing. WHO report 2004. Document WHO/
HTM/TB/2004.331. Geneva (Switzerland): World Health Organization; 2004. p. 216, fig. 2; with permission.]
primary resistance) and resistance among previ- 0.8%. Prevalence of MDR ranged from 0% in eight
ously treated cases (previously known as acquired countries to 14.2% in Kazakhstan (51/359) and Israel
resistance) is important because of their different im- (36/253) (median, 1.1%). Fig. 6 shows by participat-
plications for NTPs. Three rounds of surveys coor- ing country the prevalence of MDR-tuberculosis
dinated by WHO and the International Union Against among new tuberculosis cases. Other high prevalen-
Tuberculosis and Lung Disease (IUATLD) between ces of MDR were observed in Tomsk Oblast (Russian
1996 and 2002 have yielded data on antituberculosis Federation) (13.7%), Karakalpakstan (Uzbekistan)
drug resistance among new and previously treated (13.2%), Estonia (12.2%), Liaoning Province (China)
cases. The third round of surveys included new data (10.4%), Lithuania (9.4%), Latvia (9.3%), Henan
from 77 settings or countries collected between 1999 Province (China) (7.8%), and Ecuador (6.6% on
and 2002 and gave the following results for resistance preliminary data).
among new and previously treated cases [29]. Trends in drug resistance in new cases were
determined in 46 settings (20 with two data points
New cases and 26 with at least three). Significant increases in
Data on new cases were available for 75 settings. prevalence of any resistance were found in Peru,
In total, 55,779 patients were surveyed. The preva- Botswana, New Zealand, Poland, and Tomsk Oblast,
lence of resistance to at least one antituberculosis (Russian Federation). Cuba, Hong Kong SAR, and
drug (any resistance) ranged from 0% in some Thailand reported significant decreases over time.
Western European countries to 57.1% in Kazakhstan Tomsk Oblast (Russian Federation) and Poland
(median, 10.2%). Median prevalences of resistance to showed significantly increased prevalences of MDR.
specific drugs were as follows: streptomycin, 6.3%; Decreasing trends in MDR were observed in Hong
isoniazid, 5.9%; rifampicin, 1.4%; and ethambutol, Kong SAR, Thailand, and the USA.
Fig. 6. Prevalence of MDR-tuberculosis among new tuberculosis cases, 1994 – 2002. The designations employed and the
presentation of material on this map do not imply the expression of any opinion whatsoever on the part of the World Health
Organization concerning the legal status of any country, territory, city or area, or of its authorities, or concerning the delimitation
of its frontiers or boundaries. Dashed lines represent approximate border lines for which there may not yet be full agreement.
[From World Health Organization. Anti-tuberculosis drug resistance in the world. Report no. 3. The WHO/IUATLD Global
Project on Anti-tuberculosis Drug Resistance Surveillance 1999 – 2002. Document WHO/HTM/TB/2004.343. Geneva
(Switzerland): World Health Organization; 2004. p. 47; with permission.]
global epidemiology of tuberculosis 177
Previously treated cases points and 24 with at least three data points). A
Data on previously treated cases were available significant increase in the prevalence of any resis-
for 66 settings. In total, 8405 patients were surveyed. tance was observed in Botswana. Cuba, Switzerland,
The median prevalence of resistance to at least one and the United States showed significant decreases.
drug (any resistance) was 18.4%, with the highest The prevalence of MDR significantly increased in
prevalence, 82.1%, in Kazakhstan (262/319). Median Estonia, Lithuania, and Tomsk Oblast (Russian
prevalences of resistance to specific drugs were as Federation). Decreasing trends were significant in
follows: isoniazid, 14.4%; streptomycin, 11.4%; Slovakia and the United States. More representative
rifampicin, 8.7%; and ethambutol, 3.5%. The median geographic coverage of global antituberculosis drug
prevalence of MDR was 7.0%. Fig. 7 shows by par- resistance surveillance, with further data from longi-
ticipating country the prevalence of MDR tuber- tudinal studies, will enable more accurate and com-
culosis among previously treated tuberculosis cases. prehensive monitoring of global trends in the spread
The highest prevalences of MDR were reported in of MDR tuberculosis. Increases in prevalence of
Oman (58.3%; 7/12) and Kazakhstan (56.4%; 180/ resistance can be caused by poor or worsening tuber-
319). Among countries of the former Soviet Union, culosis control, immigration of patients from areas of
the median prevalence of resistance to the four drugs higher resistance, outbreaks of drug-resistant disease,
was 30%, compared with a median of 1.3% in all and variations in surveillance methodologies.
other settings. Given the small number of subjects In conclusion, although drug-resistant tuberculosis
tested in some settings, prevalence of resistance is present in all settings surveyed, the prevalence of
among previously treated cases should be interpreted MDR is high only in certain settings. Because good
with caution. tuberculosis control practices are generally associ-
Drug-resistance trends in previously treated cases ated with lower or decreasing levels of resistance,
were determined in 43 settings (19 with two data the findings of the WHO/IUATLD Global Project
Fig. 7. Prevalence of MDR-tuberculosis among previously treated tuberculosis cases, 1994 – 2002. The designations employed
and the presentation of material on this map do not imply the expression of any opinion whatsoever on the part of the World
Health Organization concerning the legal status of any country, territory, city or area, or of its authorities, or concerning the
delimitation of its frontiers or boundaries. Dashed lines represent approximate border lines for which there may not yet be full
agreement. [From World Health Organization. Anti-tuberculosis drug resistance in the world. Report no. 3. The WHO/IUATLD
Global Project on Anti-tuberculosis Drug Resistance Surveillance 1999 – 2002. Document WHO/HTM/TB/2004.343. Geneva
(Switzerland): World Health Organization; 2004. p. 53; with permission.]
178 maher & raviglione
emphasize the vital importance of strengthening tuberculosis incidence rate of 6% to 7% per year [37].
tuberculosis control worldwide, by expanding and The epidemiologic impact on the global tuberculosis
improving the quality of implementation of the epidemic of sustained achievement of these targets is
DOTS strategy, to prevent the emergence of further expressed in the MDG relevant to tuberculosis (Goal
drug resistance. National programs need to manage 6, Target 8), ‘‘to have halted and begun to reverse
MDR tuberculosis cases, regardless of prevalence, incidence by 2015’’ [22]. The epidemiologic inter-
through application of the DOTS-Plus strategy [30]. pretation of this goal set by politicians is to decrease
tuberculosis prevalence and deaths by half by 2015.
The following section summarizes the most recent
assessment of progress in implementation of the
Status of global tuberculosis control DOTS strategy toward achieving the cure rate and
case detection targets as set out in the 2004 WHO
The scale of the tuberculosis epidemic, as de- Report, which reports on the cases detected in 2002
scribed previously, and the human rights approach and the outcomes of treatment of patients detected in
to tuberculosis demand effective and urgent action 2001 [6].
[31]. WHO has promoted the DOTS strategy to
control tuberculosis primarily by the interruption of Cases detected and notified
transmission through the rapid identification and cure
of infectious cases. By 2002, the number of countries Through the global tuberculosis monitoring and
and territories implementing the DOTS strategy was evaluation project coordinated by WHO, countries
180 (of 210), with an estimated 69% of the world’s report annually the number and type of tuberculosis
population living in administrative areas of countries cases detected, reported, and treated under DOTS and
where the DOTS strategy was being implemented [6]. non-DOTS programs [6]. In 2002, approximately
In practice, however, the proportion of the population 3 million patients who were newly diagnosed with
with access to the DOTS strategy is less than this tuberculosis, 1.4 million of whom were smear-
administrative figure because of several possible positive, were reported in DOTS programs. A total
barriers to access, including geographic, financial, of 13.3 million tuberculosis patients and 6.8 million
and cultural impediments, within the administrative smear-positive patients were treated in DOTS pro-
area. Relying on currently available methods of diag- grams between 1995 and 2002. Regarding new cases
nosis and treatment, the DOTS strategy is effective, of sputum smear – positive pulmonary tuberculosis,
affordable, and adaptable in different settings (eg, as for the calculation of case detection rate in each
part of a strategy of expanded scope where HIV country, the numerator is the number of annual cases
prevalence is high [4], as DOTS-Plus in areas where detected and reported under the DOTS strategy, and
the prevalence of MDR tuberculosis is high [5], and the denominator is the estimated annual incidence
as public-private mix [PPM] DOTS where the ma- of cases in that country. The numerator is derived
jority of tuberculosis suspects consult private practi- annually from country reports of registered cases (ie,
tioners) [32]. cases detected and reported under the DOTS strat-
WHO coordinates a global tuberculosis monitor- egy). The denominator is an estimate based on a
ing and evaluation project in which countries report variety of inputs, as outlined earlier. One of the
annual progress in implementation of the DOTS challenges in improving the accuracy of measurement
strategy [33]. The World Health Assembly (WHA) of the case detection rate is ensuring that all cases
has set global targets for tuberculosis control through detected by different care providers (eg, private
the implementation of the DOTS strategy [34]. The practitioners) and treated in line with the DOTS strat-
choice of these global targets reflected the need to egy are reported through the NTP. The 1.4 million
achieve a significant epidemiologic impact by reach- smear-positive cases reported globally by DOTS
ing targets that field experience had demonstrated programs in 2002 represent 37% of the estimated
were feasible in countries with a high incidence of incidence, a little more than half of the 70% target.
tuberculosis. These targets are to detect at least 70%
of all new infectious cases and to cure at least 85% of Treatment success
those detected by 2005 [35]. A 70% case detection
rate and an 85% cure rate eventually would reduce The cure rate is reported by each country through
both the prevalence of infectious tuberculosis cases cohort analysis of standard treatment outcomes of
and the number of infected contacts by about 40% registered patients (Table 5) [13]. Because practice
[36] and would lead to an expected decline in annual varies considerably among countries in documenting
global epidemiology of tuberculosis 179
groups [45], private practitioners [32], and employers were unaffordable in poor settings. As a result, prices
[46]). In practice, all health providers should refer of the most expensive regimens have dropped
patients to public health facilities delivering tuber- by 95%.
culosis care under the DOTS strategy or deliver PPM-DOTS is the means of engaging private
tuberculosis care consistent with the DOTS strategy practitioners in collaboration with the NTP in the
in collaboration with the NTP. The failure of delivery of tuberculosis care consistent with the
providers to deliver care consistent with the DOTS DOTS strategy. This approach is necessary where
strategy compromises the achievements of NTPs and large numbers of tuberculosis suspects seek care from
the chances of successful tuberculosis control. Gov- private practitioners rather than from public health
ernments should consider reform of legislative and services. Recent studies indicate the success of the
regulatory frameworks to engage the full range of PPM approach in achieving high rates of case detec-
health providers and will need to invest in developing tion, notification, and cure [48]. A global subgroup of
human resource capacity (for strengthened NTP the DEWG concerned with PPM-DOTS is promoting
stewardship and service delivery) [47]. the scaling up of this approach, accompanied by the
Three of the main adaptations of the DOTS necessary strengthening of the NTP stewardship and
strategy are as part of a strategy of expanded scope leadership roles. Lessons learned from PPM-DOTS
where HIV prevalence is high [4], as DOTS-Plus in are applicable to engaging the contributions of a wide
areas where the prevalence of MDR tuberculosis is range of public providers who in many countries are
high [5], and as PPM DOTS where the majority of providing tuberculosis care independently of the NTP
tuberculosis suspects consult private practitioners (eg, in prisons and social security programs).
[32]. Until recently, the efforts to control tuberculosis Accelerating progress in global tuberculosis con-
among HIV-infected people have focused mainly on trol depends on developments in the specific field of
identifying and curing infectious tuberculosis cases tuberculosis control and on strengthening health
among patients presenting to general health services. systems. In 2003, the Stop TB Partnership convened
This approach targets the final step in the sequence of a second ad hoc committee on the tuberculosis epi-
events by which HIV fuels tuberculosis, namely the demic to seek solutions to the health system con-
transmission of M. tuberculosis infection by infec- straints to more rapid progress in global tuberculosis
tious tuberculosis cases. The strategy of expanded control and to make recommendations to overcome
scope for tuberculosis control in populations with those constraints [49,50]. The committee made
high HIV prevalence comprises interventions against recommendations under seven headings (many of
tuberculosis (the DOTS strategy and tuberculosis which cut across the different aspects of tuberculosis
preventive treatment) and interventions against HIV control) [49]:
(and therefore indirectly against tuberculosis) (eg,
condoms, treatment of sexually transmitted infec- 1. Consolidate, sustain, and advance achievements
tions, safe injecting drug use, and highly active 2. Enhance political commitment (and its trans-
antiretroviral treatment) [4]. lation into policy and action)
DOTS-Plus is the programmatic approach to the 3. Address the health workforce crisis
diagnosis and treatment of MDR tuberculosis within 4. Strengthen health care systems, particularly
the context of DOTS programs. Management primary care delivery
involves the diagnosis of MDR tuberculosis through 5. Accelerate the response to the TB/HIV emergency
quality-assured culture and drug-susceptibility testing 6. Mobilize communities and the corporate sector
and treatment with second-line drugs under proper 7. Invest in research and development to shape
case management conditions. In response to the the future.
seriousness of MDR tuberculosis as a global public
health problem, the DOTS-Plus Working Group was Implementation of these recommendations de-
established in 1999 to promote improved manage- pends on coordination between the health care sector
ment of MDR tuberculosis in resource-limited and other sectors to deliver effective tuberculosis
countries. The Working Group aims to assess the control covering all populations in need.
feasibility and cost effectiveness of the use of second-
line antituberculosis drugs in DOTS-Plus projects.
Since 2000, the Working Group’s Green Light Summary
Committee has successfully negotiated with the
pharmaceutical industry to obtain substantial conces- In 2002 there were an estimated 8.8 million new
sionary prices for second-line drugs that otherwise cases of tuberculosis worldwide, and the global
global epidemiology of tuberculosis 181
incidence rate was growing at approximately 1.1% [9] Rieder HL, Zellwegger J-P, Raviglione MC, et al.
per year. The scale of the global tuberculosis epi- Tuberculosis control in Europe and international
demic indicates the huge challenge for tuberculosis migration. Report of European Task Force. Eur Respir
J 1994;7:1545 – 53.
control, which is complicated by the impact of HIV
[10] Raviglione MC, Sudre P, Esteves K, et al. Tuber-
and drug-resistant tuberculosis. Global efforts to
culosis – Western Europe, 1974 – 1991. MMWR Morb
implement the DOTS strategy widely and effectively Mortal Wkly Rep 1993;42:628 – 31.
have resulted in a global case detection rate of 37% [11] European Centre for the Epidemiological Monitoring
in 2002, more than half of the target of 70%, and of AIDS. HIV/AIDS surveillance in Europe: quarterly
treatment success for the 2001 cohort of 82%, on report no. 46, 30 June 1995.
average, close to the target of 85%. Faster progress [12] Raviglione MC, Snider D, Kochi A. Global epidemi-
toward global targets depends on future development ology of tuberculosis: morbidity and mortality of a
of new drugs, diagnostics, and vaccines, meanwhile worldwide epidemic. JAMA 1995;273(3):220 – 6.
overcoming health system constraints to intensified [13] World Health Organization. Treatment of tuberculosis:
guidelines for national programmes. 3rd edition. Docu-
implementation of the DOTS strategy and to its
ment WHO/CDS/TB/2003.313. Geneva (Switzerland)7
adaptations in areas with high prevalence of HIV or
World Health Organization; 2003.
MDR tuberculosis or where not all care providers [14] Corbett EL, Watt CJ, Walker N, et al. The growing
deliver the internationally recommended standard burden of tuberculosis: global trends and interactions
of care. with the HIV epidemic. Arch Intern Med 2003;163:
1009 – 21.
[15] Dye C, Scheele S, Dolin P, et al for the WHO Global
Surveillance and Monitoring Project. Global burden of
References tuberculosis: estimated incidence, prevalence and
mortality by country. JAMA 1999;282:677 – 86.
[1] World Health Organization. 47th World Health Assem- [16] Murray CJL, Styblo K, Rouillon A. Tuberculosis in
bly: provisional agenda item 19. Tuberculosis pro- developing countries: burden, intervention and cost.
gramme—Progress report by the Director-General. Bull Int Union Tuberc Lung Dis 1990;65:6 – 24.
Document WHA47/1994/A47/12. Geneva (Swit- [17] Raviglione MC, Luelmo F. Update on the global epi-
zerland)7 World Health Organization; 1994. demiology of tuberculosis. Curr Issues Public Health
[1a] World Health Organization. WHO declares tuberculo- 1996;2:192 – 7.
sis a global emergency [press release]. Geneva (Swit- [18] Rieder HL, Cauthen GM, Comstock GW, et al.
zerland)7 World Health Organization; April 23, 1993. Epidemiology of tuberculosis in the United States.
[2] World Health Organization. Stop TB. An expanded Epidemiol Rev 1989;11:79 – 98.
DOTS framework for effective tuberculosis control. [19] Joint United Nations Programme on HIV/AIDS
Document WHO/CDS/TB/2002.297. Geneva (Swit- (UNAIDS). 2004 report on the global AIDS epidemic.
zerland)7 World Health Organization; 2002. Geneva (Switzerland)7 UNAIDS; 2004.
[3] Maher D, Mikulencak M. What is DOTS? A guide to [20] Kenyon TA, Mwasekaga MJ, Huebner R, et al. Low
understanding the WHO-recommended TB control levels of drug-resistance amidst rapidly increasing
strategy known as DOTS. Document WHO/CDS/ tuberculosis and human immunodeficiency virus co-
CPC/TB/99.270. Geneva (Switzerland)7 World Health epidemics in Botswana. Int J Tuberc Lung Dis 1999;3:
Organization; 1999. 4 – 11.
[4] Maher D, Floyd K, Raviglione M. A strategic frame- [21] Harries AD, Nyong’Onya Mbewe L, Salaniponi FM,
work to decrease the burden of TB/HIV. Geneva et al. Tuberculosis programme changes and treatment
(Switzerland)7 World Health Organization; 2002. outcomes in patients with smear-positive pulmonary
[5] World Health Organization. DOTS-plus and the Green tuberculosis in Blantyre, Malawi. Lancet 1996;347:
Light Committee. Document WHO/CDS/TB/2000. 807 – 9.
283. Geneva (Switzerland)7 World Health Organiza- [22] United Nations Statistics Division. Millennium Indi-
tion; 2000. cators Database. Available at: http://unstats.un.org/
[6] World Health Organization. Global tuberculosis con- unsd/mi/mi_goals.asp. Accessed August 4, 2004.
trol: surveillance, planning, financing. WHO report [23] Buhl K, Nyboe J. Epidemiological basis of tuber-
2004. Document WHO/HTM/TB/2004.331. Geneva culosis eradication. Changes in the mortality of Danish
(Switzerland)7 World Health Organization; 2004. tuberculosis patients since 1925. Bull World Health
[7] Cantwell MF, Snider Jr DE, Cauthen GM, et al. Organ 1967;37:907 – 25.
Epidemiology of tuberculosis in the United States, [24] Thompson BC. Survival rates in pulmonary tubercu-
1985 through 1992. JAMA 1994;272:535 – 9. losis. BMJ 1943;2:721.
[8] Raviglione MC, Sudre P, Rieder HL, et al. Secular [25] Second East African/British Medical Research Coun-
trends of tuberculosis in Western Europe. Bull World cil Kenya Tuberculosis Survey follow-up 1979. Tu-
Health Organ 1993;71:297 – 306. berculosis in Kenya: follow-up of the second (1974)
182 maher & raviglione
national sampling survey and a comparison with the [38] Stop TB Partnership. Annual Report 2001. Document
follow-up data from the first (1964) national sampling WHO/CDS/STB/2002.17. Geneva (Switzerland)7
survey. An East African and British Medical Research World Health Organization; 2001.
Council co-operative investigation. Tubercle 1979;60: [39] Young DB. Current tuberculosis vaccine development.
125 – 49. Clin Infect Dis 2000;30(Suppl 3):S254 – 6.
[26] Mukadi YD, Maher D, Harries AD. Tuberculosis case [40] Perkins MD. New diagnostics for tuberculosis. Int J
fatality rates in high HIV prevalence populations in Tuberc Lung Dis 2000;4(12):S182 – 8.
sub-Saharan Africa. AIDS 2001;15:143 – 52. [41] Centers for Disease Control and Prevention. Preven-
[27] Harries AD, Hargreaves NJ, Kemp J, et al. Deaths from tion and treatment of tuberculosis among patients
tuberculosis in sub-Saharan African countries with a infected with human immunodeficiency virus: princi-
high prevalence of HIV-1. Lancet 2001;357:1519 – 23. ples of therapy and revised recommendations. MMWR
[28] Crofton J, Chaulet P, Maher D. Guidelines for the Morb Mortal Wkly Rep 1998;47(No. RR-20):1 – 58.
management of drug-resistant tuberculosis. Document [42] Barry III CE, Slayden RA, Sampson AE, et al. Use of
WHO/TB/1996.210. Geneva (Switzerland)7 World genomics and combinatorial chemistry in the develop-
Health Organization; 1997. ment of new antimycobacterial drugs. Biochem Phar-
[29] World Health Organization. Anti-tuberculosis drug macol 2000;59:221 – 31.
resistance in the world. Report number 3. The WHO/ [43] World Health Organization. Report on DOTS Expan-
IUATLD Global project on anti-tuberculosis drug re- sion Working Group meeting, Montreal, Canada, in
sistance surveillance 1999 – 2002. Geneva (Switzer- October 2002 [internal document]. Geneva (Switzer-
land)7 World Health Organization; 2004. land)7 World Health Organization; 2002.
[30] World Health Organization. Guidelines for establishing [44] World Health Organization. Global DOTS expansion
DOTS-plus pilot projects for the management of plan—progress in TB control in high-burden countries
multidrug-resistant tuberculosis. Document WHO/ 2001, 1 year after the Amsterdam Ministerial Confer-
CDS/TB/2000.278. Geneva (Switzerland)7 World ence. Document WHO/CDS/TB/2003.312. Geneva
Health Organization; 2000. (Switzerland)7 World Health Organization; 2001.
[31] World Health Organization. Guidelines for social [45] World Health Organization. Community contribution
mobilization: a human rights approach to tuberculosis. to TB care: practice and policy. Document WHO/CDS/
Document WHO/CDS/TB/2001.9. Geneva (Switzer- TB/2003.312. Geneva (Switzerland)7 World Health Or-
land)7 World Health Organization; 2001. ganization; 2003.
[32] Murthy KJR, Frieden TR, Yazdani A, et al. Public- [46] World Health Organization and International Labour
private partnership in tuberculosis control: experience Office. Guidelines for workplace TB control activities.
in Hyderabad, India. Int J Tuberc Lung Dis 2001;5: Document WHO/CDS/TB/2003.323. Geneva (Swit-
354 – 9. zerland)7 World Health Organization; 2003.
[33] Raviglione MC, Dye C, Schmidt S, et al. Assessment [47] World Health Organization. Good practice in legisla-
of worldwide tuberculosis control. Lancet 1997;350: tion and regulations for TB control: an indicator of
624 – 9. political will. Document WHO/CDS/TB/2001.290. Ge-
[34] World Health Organization. Forty-fourth World Health neva (Switzerland)7 World Health Organization; 2001.
Assembly. Resolutions and decisions. Resolution WHA [48] Lönnroth K, Uplekar M, Arora VK, et al. Public-
44.8. WHA44/1991/REC/1. Geneva (Switzerland)7 private mix for improved TB control—what makes it
World Health Organization; 1991. work? Bull World Health Organ 2004;82(8):580 – 6.
[35] World Health Organization. Fifty-third World Health [49] World Health Organization. Report on the meeting of
Assembly. Resolutions and decisions. Resolution WHA the second ad hoc committee on the TB epidemic.
53.1. Geneva (Switzerland)7 World Health Organiza- Montreux, Switzerland, 18 – 19, 2003. Recommenda-
tion; 2000. tions to Stop TB partners. Document WHO/HTM/
[36] Styblo K., Bumgarner J.R. Tuberculosis can be con- STB/2004.28. Geneva (Switzerland)7 World Health
trolled with existing technologies: evidence. Tuber- Organization; 2004.
culosis Surveillance Research Unit Progress Report [50] World Health Organization. Background document
(The Hague, The Netherlands) 1991;2:60 – 72. prepared for the meeting of the second ad hoc com-
[37] Dye C, Williams BG. Criteria for the control of drug- mittee on the TB epidemic. Document WHO/HTM/
resistant tuberculosis. Proc Natl Acad Sci USA 2000; STB/2004.27. Geneva (Switzerland)7 World Health Or-
97:8180 – 5. ganization; 2004.
Clin Chest Med 26 (2005) 283 – 294
The HIV pandemic poses major problems for the rifabutin, pyrazinamide, and ethambutol—followed
tuberculosis control program and for the individual by a 4-month continuation phase of isoniazid and
clinician treating HIV-related tuberculosis. HIV- rifampin or rifabutin. A remaining challenge for the
related immunosuppression is the single most potent treatment of active tuberculosis among HIV-infected
risk factor for progression from latent tuberculosis and uninfected persons is finding efficient and pro-
infection to active tuberculosis [1]. As a result, HIV is grammatically relevant ways to identify patients at
major factor driving the global resurgence of tuber- increased risk for relapse and targeting them for pro-
culosis; incidence rates of tuberculosis in countries longed or otherwise altered treatment [4,5]. Ad-
with high prevalence of HIV infection have increased herence to multidrug therapy is difficult, particularly
up to fivefold [2]. Severe immunosuppression also when it must be sustained for at least 6 months.
results in marked changes in the clinical, radiographic, Directly observed therapy, the most effective way of
and histopathologic presentation of tuberculosis [3]. promoting adherence to tuberculosis treatment [6], is
This article does not review all aspects of HIV-related all the more important in the management of HIV-
tuberculosis; the dramatic effects of HIV on the related tuberculosis [7]. There is little margin for er-
epidemiology, presentation, and diagnosis of tuber- ror in the treatment of tuberculosis in a severely
culosis have been reviewed elsewhere recently [2,3]. immunocompromised person.
This article focuses on the ways in which HIV infec-
tion and the associated immunodeficiency affect the
management of active tuberculosis. Controversial Issues in the treatment of HIV-related tuberculosis
topics are highlighted, followed by a suggested strat-
egy for management while awaiting additional data. When these basic principles are observed, the
There are a number of unique challenges in the outcomes of treatment of active tuberculosis among
treatment of HIV-related tuberculosis, but the basic persons with HIV infection are similar to those of
principles of tuberculosis treatment developed over HIV-negative patients with tuberculosis [5,8 – 14].
the past 50 years are applicable to HIV-related The rates of treatment failure (a positive culture at or
tuberculosis. Drug-susceptible tuberculosis is treated beyond 4 months of treatment) and relapse are low.
most efficiently with regimens including an initial Whether the rates of treatment failure and relapse are
intensive phase—2 months of isoniazid, rifampin or somewhat higher among patients with HIV coinfec-
tion and how the rate of treatment failure should
affect the management of HIV-related tuberculosis
This work was supported in part by the Tuberculosis
remain subjects of controversy. One key difference is
Trials Consortium, Centers for Disease Control and Pre-
vention, Atlanta, GA. The author has had research contracts
that the risk of death during tuberculosis treatment is
with Roche Laboratories, Merck, Glaxo-Smith Kline, and much higher among persons with HIV-related tuber-
Bristol Myers-Squibb. culosis. For example, in a study from South Africa
* 605 Bannock Street, Denver, CO 80204. the risk of death during tuberculosis treatment was
E-mail address: bburman@dhha.org 13.7% among HIV-infected miners versus 0.5%
0272-5231/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccm.2005.02.002 chestmed.theclinics.com
284 burman
among HIV-negative miners [11]. After the first few patients with advanced HIV disease (in the era before
weeks of tuberculosis treatment, nearly all the excess combination antiretroviral therapy) required lifelong
mortality is related to complications of AIDS other therapy to prevent recurrent meningitis [21]. Such
than tuberculosis [10,15]. Combination antiretrovi- experience suggested that treatment of tuberculosis
ral therapy markedly reduces new opportunistic in- might also have to be longer among patients with
fections and death among persons with AIDS [16] advanced HIV disease.
and seems to do so among persons with HIV-related Despite the importance of the question, there have
tuberculosis [17 – 19]. The use of combination anti- been no definitive studies of the optimal duration of
retroviral therapy in persons being treated with tuber- treatment for HIV-related tuberculosis. A large study
culosis poses a number of challenges for the patient performed in Zaire randomly assigned patients to 6 or
and clinician, however. 12 months of therapy and found a higher rate of
To summarize, the controversies in the manage- recurrent tuberculosis among patients treated with
ment of HIV-related tuberculosis can be grouped into 6 months of therapy [22]. High losses to follow-up
issues about tuberculosis treatment itself and issues and the inability to distinguish infection with a new
posed by the use of combination antiretroviral ther- strain of Mycobacterium tuberculosis (re-infection)
apy. The issues related to tuberculosis treatment are from relapse of the initial infecting strain make
the uncertainties about the optimal duration of this study difficult to interpret. In an area with high
therapy and the adequacy of intermittent dosing of tuberculosis case rates, re-infection can be a com-
tuberculosis therapy (dosing less frequently than mon cause of recurrent tuberculosis, particularly
daily). Use of combination antiretroviral therapy dur- among HIV-infected persons [23,24]. A trial in the
ing tuberculosis treatment is complicated by (1) the United States comparing 6 versus 9 months of ther-
adherence challenge of polypharmacy, (2) overlapping apy showed low relapse rates in both arms (<3%),
side-effect profiles of the antituberculosis drugs, but the study was inadequately powered to be defini-
antiretroviral therapy, and drugs used to prevent or tive [25].
treat other opportunistic infections, (3) drug – drug The results of published observational cohort
interactions, and (4) the occurrence of immune studies of standard 6-month regimens (isoniazid,
reconstitution inflammatory syndromes following rifampin, pyrazinamide, and ethambutol for 2 months,
the institution of effective antiretroviral therapy. followed by isoniazid and rifampin for 4 months)
These four issues lead to uncertainties about the op- given by directly observed therapy to patients with
timal timing of antiretroviral therapy during tuber- and without HIV coinfection are shown in Table 1.
culosis treatment. Most studies have shown similar rates of treatment
failure and relapse among HIV-positive and HIV-
negative persons [5,13,14,23]. The one study that
Issues related to tuberculosis treatment found a much higher risk of recurrent tuberculosis
among HIV-infected patients was done in a setting of
Optimal duration of therapy high rates of tuberculosis (gold mines in South
Africa) [23]. Notably, this study used DNA finger-
Early in the HIV pandemic, it became clear that printing of initial and relapse isolates and showed that
some opportunistic infections required prolonged, the higher rate of recurrent tuberculosis in HIV-
if not lifelong, treatment in persons with advanced infected patients resulted entirely from an increased
HIV disease. For example, cryptococcal meningitis risk of re-infection.
can be cured in a high percentage of immunocompe- These cohort studies do not offer a definitive
tent patients with 6 weeks of treatment [20], but answer to the question whether patients with HIV-
Table 1
Comparison of the outcomes of tuberculosis treatment by HIV serostatus in cohort studies using directly-observed, 6-month,
rifampin-containing regimens
HIV positive HIV negative
Study [reference] Treatment failure (%) Relapse (%) Treatment failure (%) Relapse (%)
Haiti (n = 427) [14] 2.0 5.4 3.0 2.8
South Africa (n = 403) [13] 3.0 5.0 7.0 5.0
Baltimore (n = 407) [5] 0 8.3 0 1.7
South Africa (n = 385) [23] 5.3 21.5 8.1 13.0
management of hiv-related tuberculosis 285
related tuberculosis might require longer or more in- directly observed therapy does not prevent acquired
tensive therapy. The size of the cohorts evaluated drug resistance among patients with advanced HIV
(approximately 400) is not been adequate to eval- disease. The clearest demonstration of the association
uate the optimal duration of therapy, particularly in between HIV infection and acquired drug resis-
key subgroups, such as patients with advanced HIV tance despite use of directly observed therapy comes
disease (CD4 cell count <200/mm3). Cohort studies from a clinical trial evaluating once-weekly rifapen-
and clinical trials suggest that patients with advanced tine and isoniazid during the last 4 months of therapy
HIV disease are at increased risk of treatment failure for drug-susceptible tuberculosis. Among 30 HIV-
and relapse [5,18,25,26]. These studies also demon- infected patients treated with this regimen, 5 relapsed,
strated that a surprisingly high percentage of M. tu- 4 of whom had isolates with acquired drug resistance
berculosis isolates from cases of treatment failure [30], whereas there were no cases of acquired drug
and relapse among patients with advanced HIV resistance among the 502 HIV-negative patients
disease and tuberculosis had acquired rifamycin treated with the same regimen [4]. In all cases, the
resistance. Before making provisional recommenda- drug class to which resistance was acquired was the
tions about duration of treatment for HIV-related rifamycin class, and all patients with acquired rifa-
tuberculosis, it is important to review the factors mycin resistance had advanced HIV disease (CD4
associated with acquired rifamycin resistance. cell counts <200/mm3) [30].
One of the key advantages of directly observed Subsequent clinical trials and observational cohort
therapy is that it prevents selective drug taking (ie, a studies have confirmed and extended the findings from
patient takes some, but not all, medications in a the trial evaluating once-weekly rifapentine-based
multidrug regimen). Among HIV-negative patients, therapy (Table 2). Acquired rifamycin resistance is
this feature of directly observed therapy almost clearly associated with HIV coinfection. Among
completely prevents acquired drug resistance among patients with HIV-related tuberculosis, the consistent
the 2% to 5% of patients whose tuberculosis relapses associations are advanced HIV disease (CD4 counts
[27 – 29]. There is, however, increasing evidence that in cases of acquired rifamycin resistance have
Table 2
Summary of studies of acquired rifamycin resistance in studies using directly observed therapy treatment of HIV-related
tuberculosis
Rate of Rate of CD4 cell
Dosing frequency
treatment acquired counts of
Study Rifamycin Intensive phase Continuation failure/ rifamycin cases of acquired
[reference] used (first 8 weeks) phase relapse resistance rifamycin resistance
TBTC 23 [18] Rifabutin Daily for 2 weeks, Twice weekly 5.3% (9/169) 4.7% (8/169) 4 – 55
then daily or
intermittent
TBTC 22 [30] Rifapentine Daily for 2 weeks, Once weekly 16.7% (5/30) 13.3% (4/30) 8, 8, 23, 188
then daily or
intermittent
Rifampin Daily for 2 weeks, Twice weekly 10.0% (3/30) 0% Not applicable
then daily or
intermittent
CPCRA/ Rifampin Daily for 2 weeks, Twice weekly 3.0% (3/101) 2.0% (2/101)a 17, 26
ACTG [25] then thrice weekly
Baltimore [5] Rifampin 5 d/wk for 3 weeks, Twice weekly 11.1% (9/81) 3.7% (3/81) Median, 61
then twice weekly
Rifabutin 5 d/wk for 3 weeks, Twice weekly 0% (0/27) 0% Not applicable
then twice weekly
South Rifampin 5 d/wk 5 d/wk 0% (0/151)b Not applicable
Africa [23]
Abbreviations: CPCRA/ACTG, Community Programs for Clinical Research on AIDS/AIDS Clinical Trials Group; TBTC,
Tuberculosis Trials Consortium.
a
One patient was thought to have re-infection with rifamycin-monoresistant tuberculosis, based on DNA fingerprinting
of the relapse isolate.
b
Cases of treatment failure and relapse occurred in the cohort study, but none had acquired rifamycin resistance.
286 burman
all been <200 cells/mm3 and usually <50 cells/mm3) the current recommendation that patients with
and the use of highly intermittent therapy (once or advanced HIV disease (CD4 cell count <200/mm3)
twice weekly) [5,18,23,25,26,30]. The timing of the should receive daily (5 days/week) tuberculosis treat-
use of intermittent therapy may also be important. ment, at least for the first 2 months of treatment. The
The greatest risk of acquired rifamycin resistance choice of rifamycin (rifampin or rifabutin) should be
seems to be among patients who received twice- based on the drug interactions with the planned
weekly therapy during the intensive phase (the first antiretroviral treatment regimen. The available data
2 months of treatment) [18,26]. Despite the appeal of suggest that rifampin and rifabutin are equally effec-
the hypothesis that acquired rifamycin resistance tive for tuberculosis treatment [34 – 36].
might be caused by the pharmacokinetic mismatch Whether caused by re-infection or relapse, the
between isoniazid (half-life of 1 – 3 hours) and the high rates of recurrent tuberculosis among HIV-
long-acting rifamycins, rifabutin and rifapentine coinfected patients in areas of the world with endemic
(half-lives of 33 and 15 hours, respectively) [31,32], tuberculosis are worrisome. Possible interventions to
it is clear that acquired rifamycin resistance can occur decrease the risk of recurrent tuberculosis in these
with twice-weekly rifampin-based therapy (half-life areas include changes in tuberculosis treatment and
of rifamycin, 2 – 3 hours) [5,25,26]. the use of combination antiretroviral therapy. A
longer duration of multidrug tuberculosis treatment
Recommendations for tuberculosis treatment [22] or the use of isoniazid after completion of the
standard 6-month multidrug regimen may decrease
The current recommendation for the duration of the risk of recurrent tuberculosis [37,38], probably
therapy for HIV-related tuberculosis is to use stan- by decreasing rates of re-infection. Use of potent
dard 6-month regimens with extension of therapy to antiretroviral therapy seems to decrease the rate of
9 months for patients with a delayed clinical or mi- initial progression to active tuberculosis [39,40] and
crobiologic response (continued symptoms or sputum may decrease the risk of recurrent tuberculosis
culture-positive at 2 months) [33]. Given the lack of after completion of tuberculosis treatment. Random-
data that patients with HIV-related tuberculosis re- ized trials are underway to evaluate some of these
quire longer durations of treatment, the recommen- possibilities. If available, the most promising inter-
dation for 6-month regimens is reasonable and has the vention seems to be use of potent antiretroviral ther-
programmatic advantage of having consistent recom- apy, because it should protect against the many
mendations for HIV-infected and HIV-uninfected complications of AIDS, not only prevent recur-
patients. If a subgroup is to be targeted for longer rent tuberculosis.
therapy, it seems that criteria derived from studies of
HIV-uninfected patients, such as 2-month sputum
culture positivity and pulmonary cavitation [4], may Challenges of using antiretroviral therapy during
not be appropriate for HIV-related tuberculosis. The tuberculosis therapy
key patient-related risk factor for treatment failure
and relapse among patients with HIV coinfection is Combination antiretroviral therapy has revolution-
advanced immunodeficiency, not 2-month sputum ized the treatment of advanced HIV disease, dramati-
culture results [5,18,25,30]. cally decreasing rates of death and opportunistic
A high percentage of cases of treatment failure infection. Because patients with HIV-related tuber-
and relapse among patients with HIV-related tuber- culosis have relatively high rates of HIV disease pro-
culosis are cases of acquired rifamycin resistance gression (death or a new opportunistic illness) during
(Table 2). Therefore, the question whether to treat the 6-month period of tuberculosis treatment [15,25],
HIV-related tuberculosis differently from tuberculosis antiretroviral therapy should markedly improve the
among immunocompetent patients devolves to con- outcome of patients with HIV-related tuberculosis.
siderations of how to avoid acquired rifamycin Despite its promise, the use of antiretroviral ther-
resistance. There have been no randomized trials of apy during tuberculosis treatment is complex for
different dosing frequencies of tuberculosis therapy both patient and clinician. Therefore, it is important
among HIV-coinfected patients, so all recommenda- that use of antiretroviral therapy be targeted to those
tions for avoiding or reducing acquired rifamycin at substantial risk of HIV disease progression during
resistance are derived from comparisons of cohort tuberculosis treatment. As is true for all persons
studies and expert opinion. Given the consistency of with HIV infection, the short-term (6 – 12 month) risk
the association between acquired rifamycin resistance of HIV disease progression among persons with
and highly intermittent dosing, it is prudent to follow HIV-related tuberculosis is related closely to the de-
management of hiv-related tuberculosis 287
gree of immunodeficiency at the time of tuberculosis importance of adherence and behavioral techniques
diagnosis (as measured by CD4 cell count or CD4 to promote adherence are associated with better
cell percentage) [10,11,25]. Therefore, it is reason- response to antiretroviral therapy [46]. Finally, it
able to use the standards for initiating combination may be possible to use the team that provides directly
antiretroviral therapy among patients with concomi- observed treatment for tuberculosis to promote
tant tuberculosis that are used for all patients with adherence to antiretroviral therapy [47]. This coop-
HIV disease: start combination antiretroviral therapy eration is one of many benefits that can come from
for those with clinical or laboratory evidence of ad- close collaboration between HIV and tuberculosis
vanced HIV disease (presence of an opportunistic care providers.
illnesses or a CD4 cell count <200/mm3) [41].
Given the lack of clear evidence that, in general, Overlapping adverse event profiles
patients having CD4 cell counts higher than 200 to
250/mm3 benefit from earlier initiation of antiretro- Adverse clinical events are common among
viral therapy, it seems prudent to avoid the sub- patients with HIV-related tuberculosis [18,48], and
stantial difficulties of using antiretroviral therapy it may be difficult to ascribe a specific cause to many
during tuberculosis treatment for the subset with adverse events (Table 3). The severe immunosup-
concomitant active tuberculosis. Those with higher pression of advanced HIV disease may lead to
CD4 cell counts can be re-evaluated during and additional opportunistic illnesses or direct complica-
after tuberculosis treatment. It is also noteworthy tions of HIV itself (eg, HIV-related thrombocytope-
that tuberculosis treatment itself results in substan- nia). Although tuberculosis treatment is generally
tial increases in CD4 cell counts among coinfected well tolerated, there are a host of common bother-
patients [42]. some side effects (eg, nausea from pyrazinamide) and
occasional serious side effects (eg, hepatitis from
Adherence to treatment of HIV and tuberculosis isoniazid or pyrazinamide). Antiretroviral drugs have
many of the same side effects as drugs used to treat
Combination antiretroviral treatment regimens tuberculosis or other opportunistic infections (eg,
have become simpler during the past few years. Most efavirenz, pyrazinamide, and cotrimoxacole all com-
regimens can now be given twice daily with food, monly cause skin rash). Finally, the immune recon-
and a number of potent combinations can be given stitution following successful antiretroviral therapy
once daily without regard to food. Even so, adherence may cause clinical events that may mimic drug side
to any long-term therapy is a tremendous challenge, effects (eg, rising hepatic transaminases in a patient
and adherence remains one of the most important with chronic viral hepatitis).
determinants of survival among patients with ad- Even an experienced clinician may have great
vanced AIDS [43]. Therefore, it is important to difficulty determining the cause of a specific clinical
consider how the adherence challenge of concomitant event, such as skin rash in a patients with HIV-related
multidrug therapy for tuberculosis treatment might tuberculosis who has recently started multidrug tu-
affect the success of antiretroviral therapy. berculosis treatment, cotrimoxazole, and efavirenz-
Adherence to antiretroviral therapy seems to based antiretroviral therapy. Laboratory testing is
decrease with increasing number of pills per day seldom helpful in these situations; the temporal
and increasing overall complexity of the regimen sequence of events is often the best clinical tool for
(number of different medications, number of different ascribing a cause to the many clinical events in
doses per day) [44]. The patient’s readiness to start patients with HIV-related tuberculosis. Therefore, the
combination antiretroviral therapy is also a determi- implication of overlapping adverse event profiles is
nant of outcomes [45]. There are no published studies that, as much as possible, the clinician should initiate
of how tuberculosis treatment affects adherence to one intervention at a time, allowing time to gauge the
antiretroviral therapy, but studies of adherence in success and tolerability of that intervention. The
the general HIV-infected population suggest that, attempt to start tuberculosis treatment, opportunistic
to promote readiness to begin antiretroviral treat- infection prophylaxis, and antiretroviral therapy in a
ment, it is important to take the time to deal with the short period may result in adverse events that lead to
psychologic reactions to the diagnoses of tubercu- the discontinuation of all drug therapy while awaiting
losis and advanced HIV disease, as well as the social resolution or stabilization of the adverse event. The
issues—poverty, homelessness, substance abuse— first-line tuberculosis drugs should not be discon-
that are commonly present among patients with tinued permanently without clear evidence that they
HIV-related tuberculosis. Programs focusing on the are causing a serious side effect. Despite the fre-
288 burman
Table 3
The challenge of attributing adverse events to a specific cause among patients with HIV-related tuberculosis
Possible causes
Medications other
Antituberculosis than antiretroviral Antiretroviral Immune reconstitution
Adverse event drugs HIV disease drugs drugs inflammatory syndrome
Skin rash Pyrazinamide, rifampin, Folliculitis, Cotrimoxazole Nevirapine,
rifabutin, isoniazid severe delavirdine,
asteatosis efavirenz,
abacavir
Nausea, Pyrazinamide, rifampin, Other Cotrimoxazole Zidovudine, Immune reconstitution
vomiting rifabutin, isoniazid opportunistic ritonavir, intraabdominal adenitis
infections amprenavir, or pancreatitis
indinavir
Hepatitis Pyrazinamide, rifampin, Cotrimoxazole Nevirapine, Immune reconstitution
rifabutin, isoniazid HIV-1 in patients with chronic
protease viral hepatitis
inhibitors
Leukopenia, Rifabutin, rifampin HIV-related Cotrimoxazole, Zidovudine
anemia, bone marrow valganciclovir
thrombocytopenia dysplasia
quency of adverse events, experienced clinicians can can be given with the non-nucleoside reverse-
complete tuberculosis treatment with the first-line transcriptase inhibitor efavirenz and probably with
drugs in a high percentage of cases [18,49]. nevirapine, as well.
Antiretroviral drugs can also affect the concen-
Drug – drug interactions trations of the rifamycins, particularly rifabutin. Con-
centrations of rifabutin and its metabolites are
Clinically significant interactions are common markedly increased by the protease inhibitors [52],
between drugs for tuberculosis and HIV treatment and this increased concentration can lead to increased
[50]. Absorption interactions have been described, rifabutin toxicity [53]. Conversely, concentrations of
such as the marked decrease in fluoroquinolone rifabutin are substantially decreased by efavirenz
exposure if given with the buffered formulation of [54]. Therefore, the clinician caring for the patient
didanosine [51], but these interactions are uncommon with HIV-related tuberculosis must be aware of both
and are easily managed by spacing the ingestion of sides of the possible interactions between HIV and
these two medications several hours apart. The more tuberculosis drugs: the effect of the rifamycins on
common and problematic interactions between drugs antiretroviral drugs and the effect of antiretroviral
used to treat tuberculosis and HIV disease occur in drugs on rifabutin.
the process of hepatic and gut-wall metabolism. Recommendations for dosing of the rifamycins
The rifamycins increase the synthesis of a number and antiretroviral therapy are provided in Tables 4
of hepatic enzyme systems involved in drug metabo- and 5 [55]. Because this field is rapidly changing, the
lism. Particularly important in drug interactions clinician should use a regularly updated source of
with the antiretroviral drugs is the effect of the rifa- information on the interactions between tuberculosis
mycins on the cytochrome P450 3A (CYP3A) enzyme drugs and antiretroviral drugs, such as the website
system. Rifampin administration results in marked sponsored by the Centers for Disease Control and
decreases in the concentrations of antiretroviral Prevention (http://www.cdc.gov/nchstp/tb/TB_HIV_
drugs predominantly metabolized by CYP3A. The Drugs/TOC.htm). Other ways to avoid major drug –
magnitude of this interaction is such that rifampin drug interactions (Box 1) include ongoing communi-
cannot be used with the HIV-1 protease inhibitors cation between HIV and tuberculosis care providers
other than relatively high-dose ritonavir (400 – 600 mg and awareness of the need to readjust doses of
twice daily) [50]. Rifabutin also increases the expres- drugs that have been altered because of an interaction
sion of CYP3A, but the effect is much less marked, when the use of an interacting drug has been stopped
so rifabutin can be given with the protease inhibitors (eg, decreasing the dose of methadone back to
(other than unboosted saquinavir). Both rifamycins baseline when rifampin is discontinued).
management of hiv-related tuberculosis 289
Table 4
Recommendations for coadministering protease inhibitors and non-nucleoside reverse transcriptase inhibitors with rifabutin –
United States, 2003
Antiretroviral
Drug dose change Rifabutin dose change Comments
Non-nucleoside reverse-transcriptase inhibitors
Efavirenz None " to 450 – 600 mg/d or Rifabutin AUC # by 38%; effect of efavirenz +
600 mg 2 – 3 /wk protease inhibitor(s) on rifabutin concentration
has not been studied
Nevirapine None 300 mg/d or 3 /wk Rifabutin and nevirapine AUC not
significantly changed
Delavirdine Rifabutin and delavirdine should not be Delavirdine AUC # by 80%;
used together rifabutin AUC " by 100%
Single protease inhibitors
Amprenavir None # to 150 mg/d or Rifabutin AUC " by 193%; no change in
300 mg 3 /wk amprenavir concentration
fos-Amprenavir None # to 150 mg/d or
300 mg 3 /wk
Atazanavir None # to 150 mg every Rifabutin AUC " by 250%
other day or 3 /wk
Indinavir " to 1000 every # to 150 mg/d or Rifabutin AUC " by 204%; indinavir AUC
mg q 8 h 300 mg 3 /wk # by 32%
Nelfinavir None # to 150 mg/d or Rifabutin AUC " by 207%; insignificant
300 mg 3 /wk change in nelfinavir concentration
Saquinavir Rifabutin and saquinavir should not be Saquinavir AUC # by 43%
used together
Dual-protease inhibitor combinations
Lopinavir/ritonavir (Kaletra) None # to 150 mg every Rifabutin AUC " by 303%; 25-O-des-acetyl
other day or 3 /wk rifabutin AUC " by 47.5 fold
Ritonavir (any dose) None # to 150 mg every
with saquinavir, indinavir, other day or 3 /wk
amprenavir, fos-amprenavir,
or atazanavir
Abbreviation: AUC, area under the curve.
Table 5
Recommendations for coadministering protease inhibitors and non-nucleoside reverse transcriptase inhibitors with rifampin –
United States, 2003
Rifampin
Drug Antiretroviral dose change dose change Comments
Single protease inhibitors
Ritonavir None None (600 mg/d) Use with caution. Ritonavir
AUC # by 35%; no change
in rifampin concentration.
Monitor for antiretroviral
activity of ritonavir
Amprenavir Rifampin and amprenavir should not be used together Amprenavir AUC # by 82%,
Cmin # by 92%
fos-Amprenavir Rifampin and fos-amprenavir should not be used together See amprenavir
Atazanavir Rifampin and atazanavir should not be used together Interaction studies not
performed, but marked
decrease in atazanavir
concentrations predicted
Indinavir Rifampin and indinavir should not be used together Indinavir AUC # by 89%
Nelfinavir Rifampin and nelfinavir should not be used together Nelfinavir AUC # 82%
Saquinavir Rifampin and saquinavir should not be used together Saquinavir AUC # by 84%
Dual protease-inhibitor combinations
Saquinavir/ritonavira Saquinavir 400 mg + None (600 mg/d) Limited clinical experience
ritonavir 400 mg twice daily
Boosted lopinavir/ Lopinavir/ritonavir (Kaletra) 3 capsules + None (600 mg/d) Limited clinical experience
ritonavir (Kaletra) 300 mg ritonavir twice daily
Lopinavir/ritonavir Rifampin and unboosted lopinavir/ritonavir (Kaletra) should not Lopinavir AUC # by 75%
(Kaletra) be used together and Cmin # by 99%
Non-nucleoside reverse transcriptase inhibitors
Efavirenz None or " to 800 mg/d None (600 mg/d) Efavirenz AUC # by 22%;
no change in rifampin
concentration
Nevirapine Consider " to 300 mg twice dailyb None (600 mg/d) Nevirapine AUC #
37% – 58% and Cmin # 68%
with 200 mg 2 /d dose
Delavirdine Rifampin and delavirdine should not be used together Delavirdine AUC # by 95%
Abbreviation: AUC, area under the curve.
a
In a recent drug interaction study among healthy volunteers, there was a high and unacceptable rate of hepatoxicity during
treatment with saquinavir, ritonavir, and rifampin. The FDA and Roche Laboratories advise that this combination not be used.
b
No safety data available; limited data on antiretroviral efficacy.
event in a large series of patients with HIV-related drug therapy but are common in the era of potent
tuberculosis was 11 days [61], with some events combination antiretroviral therapy (being reported in
beginning within 2 days [57]. On the other hand, 11% to 35% of patients starting potent antiretroviral
other patients can have the onset of an immune therapy) [18,57,61,63]. Finally, the timing of anti-
reconstitution event months after starting antiretrovi- retroviral therapy may also be important; earlier
ral therapy [62]. The most consistent risk factor for initiation of antiretroviral therapy may increase the
immune reconstitution inflammatory events seems risk and severity of immune reconstitution inflam-
to be severity of immunosuppression; patients with matory event [61,64].
very low CD4 cell counts, and hence those in greatest
need of effective antiretroviral therapy, seem to be at Summary of issues involved in the use of combination
highest risk of an immune reconstitution inflamma- antiretroviral therapy
tory event [57,58,63,64]. The potency of the anti-
retroviral regimen seems to be related to the risk of The use of antiretroviral therapy during tuber-
an immune reconstitution inflammatory event; such culosis treatment seems to reduce markedly the risk
events were infrequently reported in the era of single- of death and opportunistic infections among patients
management of hiv-related tuberculosis 291
Table 6
Possible advantages and disadvantages of early versus delayed initiation of antiretroviral therapy during tuberculosis treatment
Early antiretroviral therapy Delayed antiretroviral therapy
(before 8 weeks of tuberculosis treatment) (after 8 weeks of tuberculosis treatment)
Adherence demands Problematic with use of four-drug therapy Less problematic because fewer drugs
for TB and multidrug therapy for HIV necessary for TB treatment
Ability to determine the cause Complex because of the large number Simpler because the number of drugs
of adverse events of medications started in a short time for TB treatment is less and there has
period and overlapping side effect profiles been more time to evaluate response to
tuberculosis treatment
Drug – drug interactions Problematic Problematic
Severe immune reconstitution Risk may be increased Risk may be decreased
inflammatory events
HIV disease progression Risk may be decreased Risk may be increased
(new opportunistic infection
or death)
Abbreviation: TB, tuberculosis.
292 burman
[14] Chaisson RE, Clermont HC, Holt EA, et al. Six-month (IUATLD) World Conference on Lung Health. Paris,
supervised intermittent tuberculosis therapy in Haitian October 30 – November 2, 2003.
patients with and without HIV infection. Am J Respir [27] Weis SE, Slocum PC, Blais FX, et al. The effect
Crit Care Med 1996;154:1034 – 8. of directly observed therapy on the rates of drug re-
[15] Chaisson RE, Schecter GF, Theuer CP, et al. Tuber- sistance and relapse in tuberculosis. N Engl J Med
culosis in patients with acquired immunodeficiency 1994;330:1179 – 84.
syndrome: clinical features, response to therapy, and [28] Burman WJ, Cohn DL, Rietmeijer CA, et al. Non-
survival. Am Rev Respir Dis 1987;136:570 – 4. compliance with directly observed therapy for tuber-
[16] Mocroft A, Vella S, Benfield TL, et al. Changing culosis: epidemiology and effect on the outcome of
patterns of mortality across Europe in patients infected treatment. Chest 1997;111:1168 – 73.
with HIV-1. EuroSIDA Study Group. Lancet 1998; [29] Mitchison DA, Nunn AJ. Influence of initial drug
352(9142):1725 – 30. resistance on the response to short-course chemother-
[17] Garcia de Olalla P, Martinez-Gonzalez MA, Cayla JA, apy of pulmonary tuberculosis. Am Rev Respir Dis
et al. Influence of highly active anti-retroviral therapy 1986;133:423 – 30.
(HAART) on the natural history of extra-pulmonary [30] Vernon A, Burman W, Benator D, et al. Relapse
tuberculosis in HIV patients. Int J Tuberc Lung Dis with rifamycin mono-resistant tuberculosis in HIV-
2002;6(12):1051 – 7. infected patients treated with supervised once-
[18] Burman W, Benator D, Vernon A, et al. Use of anti- weekly rifapentine and isoniazid. Lancet 1999;353:
retroviral therapy during treatment of active tuber- 1843 – 7.
culosis with a rifabutin-based regimen [abstract 136]. [31] Skinner MH, Hseih M, Torseth J, et al. Pharmacoki-
Program and abstracts of the 10th Conference on netics of rifabutin. Antimicrob Agents Chemother
Retroviruses and Opportunistic Infections. Boston7 1989;33:1237 – 41.
Foundation for Human Retrovirology; 2003. p. 106. [32] Weiner M, Bock N, Peloquin CA, et al. Pharmacoki-
[19] Hung CC, Chen MY, Hsiao CF, et al. Improved netics of rifapentine 600, 900 and 1,200 mg during
outcomes of HIV-1-infected adults with tuberculosis in once-weekly tuberculosis therapy. Am J Respir Crit
the era of highly active antiretroviral therapy. AIDS Care Med 2004;169:1191 – 7.
2003;17(18):2615 – 22. [33] Blumberg HM, Burman WJ, Chaisson RE, et al.
[20] Bennett JE, Dismukes WE, Duma RJ, et al. A American Thoracic Society/Centers for Disease Con-
comparison of amphotericin B alone and combined trol and Prevention/Infectious Diseases Society of
with flucytosine in the treatment of cryptoccal menin- America: treatment of tuberculosis. Am J Respir Crit
gitis. N Engl J Med 1979;301(3):126 – 31. Care Med 2003;167(4):603 – 62.
[21] Bozzette SA, Larsen RA, Chiu J, et al. A placebo- [34] Gonzalez-Montaner LJ, Natal S, Yonchaiyud P, et al.
controlled trial of maintenance therapy with flucona- Rifabutin for the treatment of newly-diagnosed pulmo-
zole after treatment of cryptococcal meningitis in nary tuberculosis: a multinational, randomized, com-
the acquired immunodeficiency syndrome. California parative study versus rifampicin. Tuber Lung Dis
Collaborative Treatment Group. N Engl J Med 1991; 1994;75:341 – 7.
324(9):580 – 4. [35] Schwander S, Rusch-Gerdes S, Mateega A, et al. A
[22] Perriens JH, St. Louis ME, Mukadi YB, et al. pilot study of antituberculosis combinations comparing
Pulmonary tuberculosis in HIV-infected patients in rifabutin with rifampicin in the treatment of HIV-1
Zaire: a controlled trial of treatment for either 6 or associated tuberculosis: a single-blind randomized
12 months. N Engl J Med 1995;332:779 – 84. evaluation in Ugandan patients with HIV-1 infection
[23] Sonnenberg P, Murray J, Glynn JR, et al. HIV-1 and and pulmonary tuberculosis. Tubercle Lung Dis 1995;
recurrence, relapse, and reinfection of tuberculosis 76:210 – 8.
after cure: a cohort study in South African mine- [36] McGregor MM, Olliaro P, Wolmarans L, et al. Efficacy
workers. Lancet 2001;358(9294):1687 – 93. and safety of rifabutin in the treatment of patients with
[24] van Rie A, Warren R, Richardson M, et al. Exoge- newly diagnosed pulmonary tuberculosis. Am J Respir
nous reinfection as a cause of recurrent tuberculosis Crit Care Med 1996;154:1462 – 7.
after curative treatment. N Engl J Med 1999;341(16): [37] Fitzgerald DW, Desvarieux M, Severe P, et al. Effect of
1174 – 9. post-treatment isoniazid on prevention of recurrent
[25] El-Sadr W, Perlman DC, Matts JP, et al. Evaluation tuberculosis in HIV-1-infected individuals: a random-
of an intensive intermittent-induction regimen and ised trial. Lancet 2000;356(9240):1470 – 4.
short course duration of treatment for HIV-related [38] Churchyard GJ, Fielding K, Charalambous S, et al.
pulmonary tuberculosis. Clin Infect Dis 1998;26: Efficacy of secondary isoniazid preventive therapy
1148 – 58. among HIV-infected Southern Africans: time to change
[26] Li J, Munsiff SS, Driver CR, et al. Outcome of HIV- policy? AIDS 2003;17(14):2063 – 70.
infected tuberculosis patients treated with rifabutin or [39] Kirk O, Gatell JM, Mocroft A, et al. Infections with
rifampin-based regimens, New York City. 1997 – 2000 Mycobacterium tuberculosis and Mycobacterium
[late-breaker abstract]. Program and abstracts of the avium among HIV-infected patients after the introduc-
International Union Against TB and Lung Disease tion of highly active antiretroviral therapy. EuroSIDA
294 burman
Study Group JD. Am J Respir Crit Care Med 2000; kinetic interaction between multiple doses of efavirenz
162(3 Pt 1):865 – 72. and rifabutin in healthy volunteers [abstract 461].
[40] Badri M, Wilson D, Wood R. Effect of highly active In: Program and abstracts of the 36th Annual Meeting
antiretroviral therapy on incidence of tuberculosis in of the Infectious Diseases Society of America. Denver
South Africa: a cohort study. Lancet 2002;359(9323): (CO): Infectious Diseases Society of America; 1998.
2059 – 64. [55] Centers for Disease Control and Prevention. TB/HIV
[41] Yeni PG, Hammer SM, Hirsch MS, et al. Treatment for Drug Interactions. Updated guidelines for the use of
adult HIV infection: 2004 recommendations of the rifamycins for the treatment of tuberculosis among
International AIDS Society-USA Panel. JAMA 2004; HIV-infected patients taking protease inhibitors or
292(2):251 – 65. nonnucleoside reverse transcriptase Inhibitors. Up-
[42] Elliott AM, Luzze H, Quigley MA, et al. A random- dated Jan. 20, 2004. Available at: http://www.cdc.
ized, double-blind, placebo-controlled trial of the use gov/nchstp/tb/TB_HIV_Drugs/TOC.htm. Accessed
of prednisolone as an adjunct to treatment in HIV-1- September 6, 2004.
associated pleural tuberculosis. J Infect Dis 2004; [56] French MA, Price P, Stone SF. Immune restoration
190(5):869 – 78. disease after antiretroviral therapy. AIDS 2004;18(12):
[43] Wood E, Hogg RS, Yip B, et al. Effect of medication 1615 – 27.
adherence on survival of HIV-infected adults who start [57] Narita M, Ashkin D, Hollender ES, et al. Paradoxical
highly active antiretroviral therapy when the CD4 + worsening of tuberculosis following antiretroviral
cell count is 0.200 to 0.350 x 10(9) cells/L. Ann Intern therapy in patients with AIDS. Am J Respir Crit Care
Med 2003;139(10):810 – 6. Med 1998;158:157 – 61.
[44] Bartlett JA, DeMasi R, Quinn J, et al. Overview [58] Hollender E, Narita M, Ashkin D, et al. CNS
of the effectiveness of triple combination therapy in manifestations of paradoxical reactions in HIV + TB
antiretroviral-naive HIV-1 infected adults. AIDS 2001; patients on HAART [abstract 258]. In: Program and
15(11):1369 – 77. abstracts of the 7th Conference on Retroviruses
[45] Singh N, Berman SM, Swindells S, et al. Adherence and Opportunistic Infections. San Francisco (CA):
of human immunodeficiency virus-infected patients Foundation for Human Retrovirology; 2000. p. 124.
to antiretroviral therapy. Clin Infect Dis 1999;29(4): [59] Fishman JE, Saraf-Levi E, Narita M, et al. Pulmonary
824 – 30. tuberculosis in AIDS patients: transient chest radio-
[46] Pradier C, Bentz L, Spire B, et al. Efficacy of an graphic worsening after initiation of antiretroviral
educational and counseling intervention on adherence therapy. AJR Am J Roentgenol 2000;174:43 – 9.
to highly active antiretroviral therapy: French pro- [60] Shelburne III SA, Hamill RJ, Rodriguez-Barradas
spective controlled study. HIV Clin Trials 2003;4(2): MC, et al. Immune reconstitution inflammatory syn-
121 – 31. drome: emergence of a unique syndrome during highly
[47] Jones B, Otaya M, Rayos E, et al. Directly observed active antiretroviral therapy. Medicine (Baltimore)
therapy of highly active antiretroviral therapy in 2002;81(3):213 – 27.
patients with HIV and tuberculosis [abstract 60582]. [61] Breen RA, Smith CJ, Bettinson H, et al. Paradoxical
In: Program and abstracts of the 12th World AIDS reactions during tuberculosis treatment in patients with
Conference. Geneva (Switzerland): Marathon Multi- and without HIV co-infection. Thorax 2004;59(8):
media; 1998. p. 1107. 704 – 7.
[48] Dean GL, Edwards SG, Ives NJ, et al. Treatment of [62] Olalla J, Pulido F, Rubio R, et al. Paradoxical
tuberculosis in HIV-infected persons in the era of responses in a cohort of HIV-1-infected patients with
highly active antiretroviral therapy. AIDS 2002;16(1): mycobacterial disease. Int J Tuberc Lung Dis 2002;
75 – 83. 6(1):71 – 5.
[49] Cook SV, Fujiwara PI, Frieden TR. Rates and risk [63] Wendel KA, Alwood KS, Gachuhi R, et al. Paradoxical
factors for discontinuation of rifampicin. Int J Tuberc worsening of tuberculosis in HIV-infected persons.
Lung Dis 2000;4(2):118 – 22. Chest 2001;120(1):193 – 7.
[50] Burman WJ, Gallicano K, Peloquin C. Therapeutic im- [64] Navas E, Martin-Davila P, Moreno L, et al. Paradoxical
plications of drug interactions in the treatment of HIV- reactions of tuberculosis in patients with the acquired
related tuberculosis. Clin Infect Dis 1999;28:419 – 30. immunodeficiency syndrome who are treated with
[51] Sahai J, Gallicano K, Oliveras L, et al. Cations in highly active antiretroviral therapy. Arch Intern Med
didanosine tablet reduce ciprofloxacin bioavailability. 2002;162(1):97 – 9.
Clin Pharmacol Ther 1993;53:292 – 7. [65] Burman WJ, Jones BE. Treatment of HIV-related
[52] Cato A, Cavanaugh J, Shi H, et al. The effect of tuberculosis in the era of effective antiretroviral ther-
multiple doses of ritonavir on the pharmacokinetics of apy. Am J Respir Crit Care Med 2001;164(1):7 – 12.
rifabutin. Clin Pharmacol Ther 1998;63:414 – 21. [66] Morris L, Martin DJ, Bredell H, et al. Human immu-
[53] Hsu A, Granneman GR, Bertz RJ. Ritonavir. Clinical nodeficiency virus-1 RNA levels and CD4 lymphocyte
pharmacokinetics and interactions with other anti- HIV counts, during treatment for active tuberculosis, in
agents. Clin Pharmacokinet 1998;35(4):275 – 91. South African patients. J Infect Dis 2003;187(12):
[54] Benedeck IH, Fiske WD, White SJ, et al. Pharmaco- 1967 – 71.
Clin Chest Med 26 (2005) 217 – 231
One of the primary goals of tuberculosis control Genotyping methods and methodologic
programs is to interrupt the transmission of Myco- considerations
bacterium tuberculosis. The most effective way to
accomplish this goal is to identify and treat individu- Several nucleic acid – based genotyping methods
als who have active tuberculosis. Even in effective have been developed that allow different strains of
tuberculosis control programs, however, M. tuber- M. tuberculosis to be distinguished. The most widely
culosis continues to be transmitted to others, largely used method of genotyping, referred to as restriction
because most transmission occurs before diagnosis fragment-length polymorphism (RFLP) analysis, uses
and initiation of therapy. The ability to track specific restriction endonucleases to cleave the mycobacterial
strains of M. tuberculosis as they spread through a DNA at the sites of specific repetitive sequences,
community would greatly increase the understanding producing DNA restriction fragments of different
of the transmission and pathogenesis of tuberculosis, lengths that can be separated by gel electrophoresis
but, until recently, the only means of distinguishing (Fig. 1) [1]. Only the genomic DNA restriction
different strains of M. tuberculosis were drug- fragments that are complementary to and hybridize
resistance patterns and mycobacterial phage typing, with specific probes are visible, resulting in an eas-
both of which have significant limitations. Fortu- ily readable band pattern. Most laboratories use a
nately, several molecular genotyping techniques standardized protocol for RFLP genotyping of the
available now allow differentiation of isolates of M. tuberculosis complex that takes advantage of a
M. tuberculosis for tracking strains in the community. specific, well-characterized, repetitive element, inser-
Epidemiologic investigations that incorporate geno- tion sequence 6110 (IS6110) [1].
typing of M. tuberculosis have provided important Despite its widespread use, there are several
information about the spread of tubercle bacilli by disadvantages with IS6110-based RFLP genotyping.
identifying factors related to transmission and rapid First, it can be done only on cultures of M. tuber-
progression to disease. This article reviews how these culosis. Second, it is a slow, labor-intensive, and
genotyping tools have increased the understanding of technically demanding technique. Finally, it has rela-
the transmission and pathogenesis of M. tuberculosis. 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-
E-mail address: daleyc@njc.org ment separated by short, unique, nonrepetitive se-
0272-5231/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccm.2005.02.005 chestmed.theclinics.com
218 daley
1
1 2
Extract Digest
2 DNA DNA
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 The Centers for Disease Control and Prevention
unique, nonrepetitive sequences, or spacers, between (CDC) will use this methodology, along with spoligo-
the direct repeats can be amplified simultaneously. typing, for all initial isolates of M. tuberculosis in
Individual strains are differentiated by the number the United States as part of a national genotyping
and position of the spacers that are missing from program. More studies are needed, however, to un-
the complete spacer set. Spoligotyping has at least derstand better the role of MIRU typing in the
two advantages over IS6110-based genotyping: (1) molecular epidemiology of tuberculosis. In a recent
smaller amounts of DNA are needed so the procedure study from Quebec, Canada, 302 clinical isolates
can be performed on clinical samples or on strains were evaluated with three different genotyping meth-
of M. tuberculosis shortly after inoculation into liq- ods: IS6110-based genotyping noted that 27% of the
uid culture, and (2) the spoligotyping results can be isolates were clustered, MIRU noted clustering in
expressed in a digital format [4]. Spoligotyping can 61%, and spoligotyping noted clustering in 77% [11].
be used as either a secondary genotyping method or When all three methods were used, only 14% were
as a primary genotyping method followed by another clustered, closer to the percentage that would have
genotyping method that has greater discriminatory been expected in the population. This study provided
power [5,6]. some insight into the evolution of genotypes in en-
A promising PCR-based method is a high- demic areas and the potential for false clustering
resolution genotyping technique that characterizes when the wrong genotyping methodologies are used.
the number and size of the variable number tandem The genotyping method used depends on several
repeats (VNTR) in each of 12 independent myco- factors including technical capacity and the speed
bacterial interspersed repetitive units (MIRUs) [7,8]. with which results are needed. The genotyping
MIRU-VNTR profiling is appropriate for strains re- methods vary in the reproducibility of the tests and
gardless of their IS6110 RFLP copy number, can be in their ability to differentiate individual strains of
automated for large-scale genotyping, and permits M. tuberculosis. An interlaboratory comparative
rapid comparison of results from independent labo- study compared several genotyping techniques [12].
ratories using a 12-digit classification system [9,10]. Of the seven PCR-based assays, only mixed linker
molecular epidemiology 219
or indirectly responsible for 6% of the tuberculosis suspects, particularly in settings and environments
cases in San Francisco during a 2-year period [22]. In that facilitate transmission, such as shelters, hospices,
another report, investigators showed that a single health care facilities, prisons, and other institutional
homeless tuberculosis patient who had highly infec- or crowded settings [13]. It would be prudent to treat
tious pulmonary tuberculosis and was a regular pa- smear-negative pulmonary tuberculosis suspects for
tron of a neighborhood bar probably infected 42% some period before removing them from isolation or
(41/97) of the contacts who were regular customers sending them into high-risk settings such as jails and
and employees of the bar and caused disease in 14 prisons. In addition, pulmonary tuberculosis should
(34%) of them. All 12 patients whose isolates of be carefully ruled out in patients who have extra-
M. tuberculosis were available had identical IS6110 pulmonary diseases. Although international guide-
RFLP band patterns [23]. lines for the diagnosis and treatment of tuberculosis
Most infection-control policies and recommenda- prioritize the detection and treatment of infectious
tions prioritize smear-positive pulmonary tuberculo- sputum smear – positive patients [29], timely diag-
sis over smear-negative cases, leading to the false nosis and treatment of sputum smear – negative cases
assumption that smear-negative cases are not infec- should be considered when resources permit.
tious. Several studies have demonstrated that patients
who have sputum smears that are negative for acid- Exogenous reinfection and mixed infection
fast bacilli but culture-positive for M. tuberculosis
can transmit infection to others in the community. Molecular genotyping can determine whether a
Behr and colleagues [24] reported that patients who patient who has a recurrent episode of tuberculosis
have smear-negative culture-positive pulmonary has a relapse with the previous strain of M. tuber-
tuberculosis were probably responsible for 17% of culosis or exogenous reinfection with a new strain.
cases in San Francisco. In a recent study from Van- Although exogenous reinfection was reported before
couver, British Columbia [25], investigators reported the availability of genotyping [30], these techniques
that a similar proportion of cases resulted from smear- have made the identification of reinfection easier and
negative source cases. In this study, the authors also more specific. Exogenous reinfection has been re-
included extrapulmonary cases of tuberculosis and ported in both immunocompromised and immuno-
noted that the proportion of episodes of transmission competent persons (Table 2) [31 – 34]. In Cape Town,
from smear-negative clustered cases increased to at South Africa, where there is a high incidence of
least 25%, suggesting that some transmission was tuberculosis and ongoing transmission, 16 of 698 pa-
occurring from extrapulmonary cases. Pulmonary tu- tients had more than one episode of tuberculosis.
berculosis apparently was not ruled out in all of these Twelve of these 16 (75%) had pairs of isolates of
cases, so transmission probably occurred through M. tuberculosis with different genotyping patterns
more traditional means of spread. As an illustration [34]. Exogenous reinfection is relatively uncommon
of this point, investigators in San Francisco reported in areas that have a low incidence of tuberculosis,
that patients who have pleural tuberculosis combined such as Switzerland [35] and the Netherlands [36],
with negative sputum cultures were unlikely to gen- compared with high- to moderate-incidence regions
erate secondary cases of tuberculosis [26]. [37 – 45]. In Houston, Texas, among 100 patients
Although the frequency of cough has been shown who have recurrent tuberculosis and have completed
to correlate with skin test reactivity among contacts therapy for a first episode of tuberculosis, exogenous
[21], genotyping has provided conflicting results reinfection was reported to cause a surprisingly high
regarding the importance of symptoms in transmis- 24% to 31% of the second episodes of tuberculo-
sion. Investigators in Harris County, Texas, reported sis [46].
no association between the duration of symptoms and Some cases of suspected exogenous reinfection
the size of molecularly defined clusters [27]. Cronin might be caused by initial infections that include
and colleagues [28], however, reported that the time more than one strain. These instances would repre-
from symptom onset to diagnosis was twice as long sent repeated infections over time that lead to a mixed
for patients who were considered to be transmitters infection with different strains of M. tuberculosis.
than for nontransmitters. These latter data support the Multiple infections were demonstrated in a patient in
belief that reducing diagnostic delays can prevent San Francisco [47], in two patients who worked in a
transmission of M. tubeculosis. medical-waste processing plant in Washington State
The studies reviewed here have demonstrated that [48], and among prisoners in Spain [49]. In South
the potential for transmitting tuberculosis should be Africa, a country that has a reportedly high frequency
considered in all pulmonary tuberculosis patients/ 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 daley
common. Warren and colleagues [50], using a PCR- few years, this strain was documented to have dis-
based method of strain classification reported that seminated widely in the community. Because poor
19% of all patients were simultaneously infected with tuberculosis control and underlying HIV infection are
Beijing and non-Beijing strains and that 57% of common in many areas, drug resistance may dissemi-
patients infected with Beijing strains also were in- nate locally despite the diminished propensity of
fected with a non-Beijing strain. These observations drug-resistant strains to cause disease. In addition, it
indicate that simultaneous infections with multiple is possible that some organisms could experience a
strains of M. tuberculosis occur and may be re- subsequent mutation that increases its virulence back
sponsible for conflicting drug-susceptibility results to its pre – drug-resistant state [60].
[51] or episodes of recurrence caused by exoge-
nous reinfection.
Contact and outbreak investigations
Impact of drug resistance on transmission and
pathogenesis Conventional tuberculosis contact investigations
use the stone-in-the-pond or concentric circle ap-
Before the advent of genotyping, studies sug- proach to collect information and to screen household
gested that isoniazid-resistant strains of M. tuber- contacts, coworkers, and increasingly distant contacts
culosis were less likely to result in disease in animals for tuberculosis infection and disease [61]. Studies in
[52 – 54]. Mutations or deletions within the katG gene low-incidence areas such as San Francisco [22] and
of isoniazid-resistant strains of M. tuberculosis have Amsterdam [62] demonstrated that a relatively small
been associated with a decrease in the pathogenicity proportion (5% – 10%) of tuberculosis cases that had
in animal models [55]. Several molecular epidemio- identical IS6110 -based genotyping patterns were
logic studies have reported that patients who have named as a contact by the source case. One expla-
drug-resistant strains were less likely to be in clusters, nation for these findings is that unsuspected trans-
suggesting that drug-resistant strains might be less mission of M. tuberculosis occurred and was not
predisposed to being transmitted or to cause active easily detected by conventional contact tracing in-
tuberculosis [20,56,57]. The spread of tuberculosis vestigations. In a 5-year, population-based study in
involves a three-step process: transmission of bac- the Netherlands, contact investigations of persons in
teria, establishment of infection, and progression to five of the largest clusters identified epidemiologic
disease. Because genotyping studies require the de- links among them based on time, place, and risk
velopment of active tuberculosis, they cannot deter- factors [20]. Tuberculosis transmission also occurred
mine whether drug resistance influences only one, through only short-term, casual contact that was not
two, or all three of the processes. Burgos and col- easily identified in routine contact investigations.
leagues [58] recently reported that the number of In a more recent study [16] from the Netherlands,
secondary cases generated by isoniazid-resistant patients were divided into one of five transmission
cases of tuberculosis was significantly less than the groups based on the results of contact investigations,
drug-susceptible cases. This difference in the gen- genotyping, and, in some cases, a second interview:
eration of secondary cases was noted regardless of
HIV status and place of birth. 1. Clear epidemiologic links, confirmed by geno-
These findings support the hypothesis that drug- typing and contact tracing (24%)
resistant strains are less likely than drug-susceptible 2. Clear epidemiologic links, confirmed by geno-
strains to result in disease. There are, however, typing and second interview but not by contact
populations in which drug resistance is neither de- tracing (6%)
tected nor treated effectively and where the longer 3. Initially unclear epidemiologic links that be-
duration of infectiousness for patients who have drug- came likely after genotyping and second inter-
resistant organisms treated with standard regimens view (55%)
might offset the bacterium’s diminished capacity to 4. No epidemiologic links but genotyping indi-
cause secondary cases [58]. In areas that have high cated clustering
prevalence rates of HIV, the increased host suscep- 5. Patients who were part of a different cluster
tibility, even to strains that have diminished viru- other than expected (1%)
lence, may offset bacterial differences. For example,
one multidrug-resistant strain of M. tuberculosis, Combining groups 1 and 2 would suggest that
strain W, caused several nosocomial outbreaks in at best contact investigations could identify about
New York City in the early 1990s [59]. Over the next 30% of the clustered cases. Fifty-five percent of
molecular epidemiology 223
the clustered cases had an epidemiologic link iden- study of tuberculosis transmission in Los Angeles,
tified after the genotyping results became available California, identified 162 patients who had culture-
and a second interview was performed. These data positive tuberculosis and interviewed the patients to
suggest that as newer, more rapid amplification- identify their contacts and whereabouts [72]. Tradi-
based genotyping methods become available, this tional contact investigations did not reliably identify
approach might be able to improve contact inves- patients infected with the same strain of M. tuber-
tigations [63]. culosis: only 2 of the 96 clustered cases named others
Genotyping has also demonstrated that even when in the cluster as contacts. The degree of homelessness
another case is identified through a contact inves- and the use of daytime services at three shelters were
tigation, the contact case may be unrelated to the independently associated with clustering, however.
index case. For example, Marcel Behr and colleagues This study demonstrated that locations where the
[64] in San Francisco reported that 30% of case – homeless congregate are important sites of tuber-
contact pairs had different strains of M. tuberculosis. culosis transmission.
Unrelated strains were more common among foreign- Several studies support the idea that specific
born, particularly Asian, contacts. Of 538 similar case locations can be associated with recent or ongoing
pairs in a study [65] involving seven sites in the transmission of M. tuberculosis. In a 30-month
United States, 29% did not have matching genotype prospective, city-wide study of all tuberculosis cases
patterns, similar to the finding in San Francisco. Case in Baltimore, Maryland, using traditional contact
pairs from the same household were no more likely to investigations and IS6110-based genotyping, 46%
have confirmed transmission than those linked else- (84/182) of initial isolates were clustered, and 32%
where. Among patients younger than 5 years of age, (58/182) of the cases were considered to have
15% of culture-confirmed cases and their suspected tuberculosis that was recently transmitted [73]. Only
source patient had different genotype patterns [66]. In 24% (20/84) of clustered cases had an identifiable
a recent study from South Africa, investigators epidemiologic link of recent contact with an infec-
evaluated 129 households in which genotyping data tious tuberculosis patient. Using geographic informa-
were available for more than one patient [67]. They tion system data, the 20 clustered cases, which have
identified 313 patients of whom 145 (46%) had a epidemiologic links in geographic areas of the city
genotype pattern matching that of another member of that have low socioeconomic status and high drug
the household. These studies suggest that contact use, were spatially aggregated. Therefore, in some
investigations should not focus solely on the house- populations, location-based control efforts may be
hold but all settings frequented by the index case. more effective than traditional concentric circle –
Even when the essential elements of tuberculosis based contact tracing for early identification of cases.
control are in place, ongoing transmission of Genotyping has been particularly useful in
M. tuberculosis will continue until tuberculosis is identifying otherwise unsuspected and undetected
diagnosed and therapy is initiated. In a population- transmission in the community [13]. Molecular epi-
based molecular epidemiologic study in an urban demiologic studies have confirmed suspected and un-
community in the San Francisco Bay area, 75 (33%) suspected transmission of tuberculosis in places such
of 221 cases had the same strain of M. tuberculosis as residential care facilities [74], bars [23,75 – 77],
[68]. Thirty-nine (53%) of the 73 patients developed crack houses [78], sites of illegal floating card
tuberculosis because they were not identified as con- games [79], schools [80,81], hospitals [82,83], and
tacts of source-case patients; 20 case patients (27%) jails and prisons [84 – 87]. Tuberculosis transmission
developed tuberculosis because of delayed diagnosis also has been demonstrated among groups such as
of their sources; 13 case patients (18%) developed church choirs [88], interstate transgender social net-
tuberculosis because of problems associated with the works [89], renal transplant patients [90], from
evaluation or treatment of contacts; and one case patient to health care providers [91], and from health
patient (1%) developed tuberculosis because of de- care provider to patients [92,93]. Processing con-
lays identifying the person as a contact. taminated medical waste resulted in transmission of
Contact tracing in the community can be ineffec- M. tuberculosis to at least one worker in a medical
tive in tuberculosis outbreaks if patients do not live in waste treatment facility [94]. Genotyping was also
stable settings and either do not know or are un- used to document unsuspected bronchoscopy-related
willing to reveal the names and locations of contacts. transmission and the cross-contamination of patients
Fortunately, studies that incorporate genotyping are [95,96]. Without the availability of genotyping, it
able to provide information about the chains of would have been difficult to confirm that trans-
transmission in these groups [69 – 71]. A prospective mission had occurred in such settings.
224 daley
Community epidemiology and risk factors for drug use, alcohol dependence, asylum stay, and
clustering unemployment [109]. Thus, the risk factors associ-
ated with recent infection and rapid progression to
Tuberculosis develops by rapid progression from disease have varied from study to study, partly
a recently acquired infection, from LTBI, or from because of differences in populations, methodologies,
exogenous reinfection. Most molecular epidemiology and definitions.
studies have assumed that the proportion of clustered Unfortunately, there are few population-based
isolates in a population estimates the amount of studies from high-incidence areas. In a study of
recent or ongoing transmission of M. tuberculosis. South African gold miners, tuberculosis patients who
The frequency of clustering has ranged from 7% had not responded to treatment at entry to the study
to 32% in low-incidence areas such as Canada were more likely to be in clusters (adjusted odds ratio
[97 – 100] to 34% to 46% in urban areas in the [OR] = 3.41). Patients who have multidrug-resistant
United States [22,28,73,101] and Europe [20,102 – tuberculosis were more likely not to have responded
104]. Among gold miners in South Africa, 50% of to tuberculosis treatment but were less likely to be
tuberculosis patients were in clusters [15], and in clustered than those who have a drug-susceptible
Botswana 42% of the cases were clustered [105]. strain (OR = 0.27) [15]. HIV infection, although
Whether or not clustered cases represent tuberculosis common (53.6%), was not associated with clustering.
caused by recent transmission has remained a con- Apparently, persistently infectious individuals who
troversial point. A recent study from the Netherlands had previously not responded to treatment were
suggests that clustered cases do, in fact, repre- responsible for one third of the tuberculosis cases in
sent recent transmission and rapid progression. Of this population. In a study from Cape Town, 72% of
481 patients who had tuberculosis, 29% were clus- cases were clustered, suggesting high rates of trans-
tered, suggesting recent transmission in 20% (using mission in the community [110].
the n-1 approach to calculate recent transmission).
The authors reported that 86% of the cases had
epidemiologic links consistent with recent trans- Measuring the performance of a tuberculosis control
mission [16]. In high-incidence areas, the frequency program
of clustering has ranged from 25% in Hong Kong
[106], to 38% in India [107], to 42% to 72% in As noted previously, tuberculosis can develop
various African populations [57,67,105]. through three mechanisms: recent transmission and
Conventional epidemiologic methods can be used rapid progression to disease, reactivation of latent
in combination with molecular genotyping techniques infection, or exogenous reinfection. Because cluster-
to identify the risk factors associated with recent ing is considered a measure of recent transmission, a
infection and rapid progression to disease (Table 3). decline in the rate of clustering could be used to
In studies in low-incidence areas, young age, being in evaluate interventions aimed at reducing recent trans-
an ethnic minority group, homelessness, and sub- mission [111]. In an evaluation of tuberculosis trans-
stance abuse have been associated with recent mission over a seven-year period in San Francisco,
infection and rapid progression to disease [22,62, the number and proportion of clustered tuberculosis
73,101]. In New York City, birth outside the United cases declined, particularly among the native-born
States, age of 60 years or older, and diagnosis after population [19]. This decline was attributed to the
1993 were factors independently associated with implementation of targeted tuberculosis prevention
having a unique strain; homelessness was associated and control programs such as screening high-risk
with clustering or recent transmission [108]. Tuber- populations and implementing directly observed
culosis among foreign-born persons was more likely therapy to ensure high cure rates. A recent study
to result from recent transmission among those who in New York City showed that as tuberculosis case
were HIV-infected and more likely to result from rates fell from recent high levels, the proportion of
LTBI among those who were not infected with HIV. tuberculosis cases caused by recent transmission
These data suggest that tuberculosis prevention and dropped from 63.2% in 1993 to 31.4% in 1999
control strategies need to be targeted to the large [108]. Tuberculosis was unlikely to result from recent
number of foreign-born persons in New York City transmission in persons born outside the United
who have latent tuberculosis infection. Among States. Investigators in Denver, Colorado [112], used
foreign-born patients who have tuberculosis in Ham- clustering to measure the impact of a skin testing
burg, Germany, risk factors for recent infection program among homeless persons and showed that
included a history of contact tracing, intravenous 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 daley
tation of the program to 14% in the 4-year period it is also possible that the Beijing genotype was
after the program. introduced into multiple locations before other strains
By contrast, an 8-year study in Greenland showed and had more time to spread.
that the annual incidence of tuberculosis doubled
from 1990 to 1997, and the percentage of culture-
positive tuberculosis cases in RFLP-defined clusters The future of molecular epidemiology
increased to 85%, reflecting microepidemics among
adults and young children in small, isolated set- Molecular genotyping, in combination with con-
tlements [113]. Thus, genotyping was a useful in- ventional epidemiologic investigations, has contrib-
dicator of changes in the proportion of cases that uted greatly to the understanding of the transmission
resulted from recent transmission and rapid progres- and pathogenesis of tuberculosis and has identi-
sion to disease. fied inadequacies in tuberculosis control programs.
The development of new tools, such as real-time
Geographic distribution and dissemination of amplification-based genotyping, should improve the
Mycobacterium tuberculosis ability to genotype strains of M. tuberculosis and con-
duct effective, timely, contact and outbreak investiga-
Genotyping has permitted the tracking of strains tions. Other technologies based on the genome of
of M. tuberculosis as they spread both locally and M. tuberculosis may eventually make it possible s to dif-
globally. Population-based data from the San Fran- ferentiate strains based on important phenotypic char-
cisco Bay area suggest that M. tuberculosis does not acteristics such as transmissibility and pathogenicity.
rapidly transmit and spread across geographic bound- For these new tools to be used effectively, they
aries and that tuberculosis control programs should must be used in combination with conventional
focus on transmission within well-defined areas epidemiologic investigations. With time, genotyping
[114]. In fact, most clusters (66%) from the National should be integrated with new surveillance systems
Tuberculosis Genotyping and Surveillance Network that will allow more rapid responses to potential
in the United States were restricted to a single site outbreaks of tuberculosis. The integration of geno-
[115]. Some strains of M. tuberculosis are widely typing and new approaches to surveillance will be
dispersed both geographically and temporally, how- particularly important in low-incidence areas of the
ever, suggesting the strains are either older, more United States where resources are limited and early
transmissible, or are more likely than other strains to detection of outbreaks is difficult. The CDC-funded
cause disease. Data from the Genotyping Network national genotyping network should help with the
found that 25% of the clusters were in two sites, 5% integration of genotyping tools into standard tuber-
were in three, 2% were in four, and 1% each were in culosis control practices and pave the way for tu-
five and six sites [115]. Further research is needed to berculosis control in the future.
determine why some strains seem to be more dis-
seminated than others.
The Beijing family of strains has been detected in
References
high proportions among the strains in China [116],
other parts of Asia [117], the former Russian Fed-
[1] van Embden JD, Cave MD, Crawford JT, et al.
eration [87], Estonia [118], Europe [119 – 121], and Strain identification of Mycobacterium tuberculosis
South Africa [122] and has been associated with large by DNA fingerprinting: recommendations for a
outbreaks, febrile responses [123], treatment failure standardized methodology. J Clin Microbiol 1993;
and relapse [124,125], and drug resistance [126]. The 31:406 – 9.
W strain, a multidrug-resistant strain of M. tuber- [2] Yang ZH, Ijaz K, Bates JH, et al. Spoligotyping and
culosis that caused many cases of tuberculosis among polymorphic GC-rich repetitive sequence finger-
patients and health care workers in nosocomial printing of Mycobacterium tuberculosis strains hav-
outbreaks and other institutional settings in New ing few copies of IS6110. J Clin Microbiol 2000;38:
York City [59,127 – 130], is a member of the Beijing 3572 – 6.
[3] Groenen PM, Bunschoten AE, van Soolingen D, et al.
family [128]. It is unclear why the Beijing family
Nature of DNA polymorphism in the direct repeat
strains are so widely disseminated [131]. It is possible cluster of Mycobacterium tuberculosis; application
that the Beijing genotype has a selective advantage for strain differentiation by a novel typing method.
and is more readily aerosolized, can establish infec- Mol Microbiol 1993;10:1057 – 65.
tion more effectively, or can progress more rapidly [4] Barnes PF, Cave MD. Molecular epidemiology of
from infection to disease [4,132]. On the other hand, tuberculosis. N Engl J Med 2003;349:1149 – 56.
molecular epidemiology 227
[5] Wilson SM, Goss S, Drobniewski F. Evaluation of mission of Mycobacterium tuberculosis derived from
strategies for molecular fingerprinting for use in the DNA fingerprinting techniques. Am J Epidemiol
routine work of a Mycobacterium tuberculosis 1999;149:366 – 71.
reference unit. J Clin Microbiol 1998;36:3385 – 8. [19] Jasmer RM, Hahn JA, Small PM, et al. A molecular
[6] Soini H, Pan X, Amin A, et al. Characterization of epidemiologic analysis of tuberculosis trends in San
Mycobacterium tuberculosis isolates from patients in Francisco, 1991 – 1997. Ann Intern Med 1999;130:
Houston, Texas, by spoligotyping. J Clin Microbiol 971 – 8.
2000;38:669 – 76. [20] van Soolingen D, Borgdorff MW, de Haas PE, et al.
[7] Supply P, Lesjean S, Savine E, et al. Automated high Molecular epidemiology of tuberculosis in the Neth-
throughput genotyping for study of global epidemi- erlands: a nationwide study from 1993 through 1997.
ology of Mycobacterium tuberculosis based on J Infect Dis 1999;180:726 – 36.
mycobacterial interspersed repetitive units. J Clin [21] Hopewell PC. Factors influencing transmission and
Microbiol 2001;39:3563 – 71. infectivity of Mycobacterium tuberculosis: implica-
[8] Mazars E, Lesjean S, Banuls AL, et al. High- tions for clinical and public health management.
resolution minisatellite-based typing as a portable In: Sande MA, Hudson LD, Root RK, editors.
approach to global analysis of Mycobacterium tuber- Respiratory infections. New York7 Churchill Living-
culosis molecular epidemiology. Proc Natl Acad Sci stone; 1986. p. 191 – 216.
U S A 2001;98:1901 – 6. [22] Small PM, Hopewell PC, Singh SP, et al. The
[9] Cowan LS, Mosher L, Diem L, et al. Variable-number epidemiology of tuberculosis in San Francisco. A
tandem repeat typing of Mycobacterium tuberculosis population-based study using conventional and
isolates with low copy numbers of IS6110 by using molecular methods. N Engl J Med 1994;330:1703 – 9.
mycobacterial interspersed repetitive units. J Clin [23] Kline SE, Hedemark LL, Davies SF. Outbreak of
Microbiol 2002;40:1592 – 602. tuberculosis among regular patrons of a neighbor-
[10] Kwara A, Schiro R, Cowan LS, et al. Evaluation of hood bar. N Engl J Med 1995;333:222 – 7.
the epidemiologic utility of secondary typing methods [24] Behr MA, Warren SA, Salamon H, et al. Trans-
for differentiation of Mycobacterium tuberculosis mission of Mycobacterium tuberculosis from patients
isolates. J Clin Microbiol 2003;41:2683 – 5. smear-negative for acid-fast bacilli. Lancet 1999;353:
[11] Nguyen D, Brassard P, Menzies D, et al. Genomic 444 – 9.
characterization of an endemic Mycobacterium tuber- [25] Hernandez-Garduno E, Cook V, Kunimoto D, et al.
culosis strain: evolutionary and epidemiologic impli- Transmission of tuberculosis from smear negative
cations. J Clin Microbiol 2004;42:2573 – 80. patients: a molecular epidemiology study. Thorax
[12] Kremer K, van Soolingen D, Frothingham R, et al. 2004;59:286 – 90.
Comparison of methods based on different molecular [26] Ong A, Creasman J, Hopewell PC, et al. A molecular
epidemiological markers for typing of Mycobacte- epidemiological assessment of extrapulmonary tu-
rium tuberculosis complex strains: Interlaboratory berculosis in San Francisco. Clin Infect Dis 2004;
study of discriminatory power and reproducibility. 38:25 – 31.
J Clin Microbiol 1999;37(8):2607 – 18. [27] Giordano TP, Soini H, Teeter LD, et al. Relating the
[13] DeRiemer K, Daley CL. Tuberculosis transmission size of molecularly defined clusters of tuberculosis to
based on molecular epidemiologic research. Seminars the duration of symptoms. Clin Infect Dis 2004;38:
in Respiratory and Critical Care Medicine 2004; 10 – 6.
25:297 – 306. [28] Cronin WA, Golub JE, Lathan MJ, et al. Molecular
[14] Braden CR, Templeton GL, Cave MD, et al. epidemiology of tuberculosis in a low-to moderate-
Interpretation of restriction fragment length poly- incidence state: are contact investigations enough?
morphism analysis of Mycobacterium tuberculosis Emerg Infect Dis 2002;8(11):1271 – 9.
isolates from a state with a large rural population. [29] World Health Organization Global Tuberculosis
J Infect Dis 1997;175:1446 – 52. Programme. An expanded DOTS framework for
[15] Godfrey-Faussett P, Stoker NG. Aspects of tuber- effective tuberculosis control. WHO/CDS/TB/
culosis in Africa. 3. Genetic ‘fingerprinting’ for clues 2002297. Geneva (Switzerland)7 World Health Orga-
to the pathogenesis of tuberculosis. Trans R Soc Trop nization; 2002.
Med Hyg 1992;86:472 – 5. [30] Nardell E, McInnis B, Thomas B, et al. Exogenous
[16] van Deutekom H, Hoijng SP, de Haas PEW, et al. reinfection with tuberculosis in a shelter for the
Clustered tuberculosis cases: do they represent recent homeless. N Engl J Med 1986;18:1570 – 5.
transmission and can they be detected earlier? Am J [31] Small PM, Shafer RW, Hopewell PC, et al. Exoge-
Respir Crit Care Med 2004;169:806 – 10. nous reinfection with multidrug-resistant Mycobacte-
[17] Murray M, Alland D. Methodological problems in the rium tuberculosis in patients with advanced HIV
molecular epidemiology of tuberculosis. Am J Epi- infection. N Engl J Med 1993;328:1137 – 44.
demiol 2002;155:565 – 71. [32] Sudre P, Pfyffer GE, Bodmer T, et al. Molecular
[18] Glynn JR, Vynnycky E, Pine PE. Influence of epidemiology of tuberculosis among HIV-infected
sampling on estimates of clustering and recent trans- persons in Switzerland: a country-wide 9-year cohort
228 daley
study. Swiss HIV Cohort Study. Infection 1999;27: tuberculosis in Houston, Texas: a population-based
323 – 30. study. Int J Tuberc Lung Dis 2004;8(3):333 – 40.
[33] Heldal E, Docker H, Caugant DA. Pulmonary tuber- [47] Yeh RW, Hopewell PC, Daley CL. Simultaneous
culosis in Norwegian patients. The role of reactiva- infection with two strains of Mycobacterium tuber-
tion, re-infection and primary infection assessed by culosis identified by restriction fragment length poly-
previous mass screening data and restriction fragment morphism analysis. Int J Tuberc Lung Dis 1999;3:
length polymorphism analysis. Int J Tuberc Lung Dis 537 – 9.
2000;4:300 – 7. [48] Braden CR, Morlock GP, Woodley CL, et al.
[34] van Rie A, Warren R, Richardson M, et al. Exogenous Simultaneous infection with multiple strains of
reinfection as a cause of recurrent tuberculosis after Mycobacterium tuberculosis. Clin Infect Dis 2001;
curative treatment. N Engl J Med 1999;341:1174 – 9. 33:42 – 7.
[35] Lambert ML, Hasker E, van Deun A, et al. Re- [49] Chaves F, Dronda F, Alonso-Sanz M, et al. Evidence
currence in tuberculosis: relapse or reinfection? of exogenous reinfection and mixed infection with
Lancet Infect Dis 2003;3:282 – 7. more than one strain of Mycobacterium tuberculosis
[36] de Boer AS, Kremer K, Borgdorff MW. Genetic among Spanish HIV-infected inmates. AIDS 1999;
heterogeneity in Mycobacterium tuberculosis isolates 13:615 – 20.
reflected in IS6110 restriction fragment length poly- [50] Warren RM, Victor TC, Streicher EM, et al. Patients
morphism patterns as low-intensity bands. J Clin with active tuberculosis often have different strains in
Microbiol 2000;38:4478 – 84. the same sputum specimen. Am J Respir Crit Care
[37] Das S, Paramasivan CN, Lowrie DB, et al. IS6110 Med 2004;169:610 – 4.
restriction fragment length polymorphism typing of [51] Niemann S, Richter E, Rusch-Gerdes S, et al. Double
clinical isolates of Mycobacterium tuberculosis from infection with a resistant and a multidrug-resistant
patients with pulmonary tuberculosis in Madras, strain of Mycobacterium tuberculosis. Emerg Infect
South India. Tuber Lung Dis 1995;76:550 – 4. Dis 2000;6:548 – 51.
[38] Caminero JA, Pena MJ, Campos-Herrero MI, et al. [52] Middlebrook G. Isoniazid resistance and catalase ac-
Exogenous reinfection with tuberculosis on a Euro- tivity of the tubercle bacilli. Am Rev Tuberc 1954;
pean island with a moderate incidence of disease. 69:471 – 2.
Am J Respir Crit Care Med 2001;163:717 – 20. [53] Cohn M, Kovitz C, Oda U. Studies on isoniazid and
[39] Lourenco MC, Grinsztejn B, Fadinho-Montes FC, tubercle bacilli: the growth requirements, catalase
et al. Genotypic patterns of multiple isolates of activities, and pathogenic properties of isoniazid
M. tuberculosis from tuberculous HIV patients. Trop resistant mutants. Am Rev Tuberc 1954;70:641 – 64.
Med Int Health 2000;5:488 – 94. [54] Riley RL, Mills CC, et al. Infectiousness of air from
[40] Garcia de Viedma DG, Marin M, Hernangomez S, a tuberculosis ward: comparative infectiousness of
et al. Tuberculosis recurrences: reinfection plays a different patients. Am Rev Respir Dis 1962;85:
role in a population whose clinical/epidemiological 511 – 25.
characteristics do not favor reinfection. Arch Intern [55] Pym AS, Domenech P, Honore N, et al. Regulation of
Med 2002;162:1873 – 9. catalase-peroxidase (KatG) expression, isoniazid sen-
[41] Krüüner A, Pehme L, Ghebremichael S, et al. Use of sitivity and virulence by furA of Mycobacterium
molecular techniques to distinguish between treat- tuberculosis. Mol Microbiol 2001;40(4):879 – 89.
ment failure and exogenous reinfection with Myco- [56] Garcia-Garcia ML, Ponce-de-Leon A, Jimenez-
bacterium tuberculosis. Clin Infect Dis 2002;35: Corona ME, et al. Clinical consequences and trans-
146 – 55. missibility of drug-resistant tuberculosis in southern
[42] Godfrey-Faussett P, Githui W, Batchelor B, et al. Mexico. Arch Intern Med 2000;160:630 – 6.
Recurrence of HIV-related tuberculosis in an en- [57] Godfrey-Faussett P, Sonnenberg P, Shearer SC, et al.
demic area may be due to relapse or reinfection. Tuberculosis control and molecular epidemiology
Tuber Lung Dis 1994;75(3):199 – 202. in a South African goldmining community. Lancet
[43] Sonnenberg P, Murray J, Shearer S, et al. Tuber- 2000;356:1066 – 71.
culosis treatment failure and drug resistance – same [58] Burgos M, DeRiemer K, Small PM, et al. Effect of
strain or reinfection? Trans R Soc Trop Med Hyg drug-resistance on the generation of secondary cases
2000;94(6):603 – 7. of tuberculosis. J Infect Dis 2003;188:1878 – 84.
[44] Fitzpatrick LK, Okwera A, Mugerwa R, et al. An [59] Munsiff SS, Nivin B, Sacajiu G, et al. Persistence
investigation of suspected exogenous reinfection in of a highly resistant strain of tuberculosis in New
tuberculosis patients in Kampala, Uganda. Int J York City during 1990 – 1999. J Infect Dis 2003;188:
Tuberc Lung Dis 2002;6(6):550 – 2. 356 – 63.
[45] Bandera A, Gori A, Catozzi L, et al. Molecular [60] Cohen T, Sommers B, Murray M. The effect of drug
epidemiology study of exogenous reinfection in an resistance on the fitness of Mycobacterium tuber-
area with a low incidence of tuberculosis. J Clin culosis. Lancet Infect Dis 2003;3:13 – 21.
Microbiol 2001;39(6):2213 – 8. [61] Veen J. Microepidemics of tuberculosis: the stone-in-
[46] El Sahly HM, Wright JA, Soini H, et al. Recurrent the-pond principle. Tuber Lung Dis 1992;73:73 – 6.
molecular epidemiology 229
[62] van Deutekom H, Gerritsen JJ, van Soolingen D, [76] Garcia-Garcia M, Palacios-Martinez M, Ponce-de-
et al. A molecular epidemiological approach to Leon A, et al. The role of core groups in transmitting
studying the transmission of tuberculosis in Amster- Mycobacterium tuberculosis in a high prevalence
dam. Clin Infect Dis 1997;25:1071 – 7. community in Southern Mexico. Int J Tuberc Lung
[63] Daley CL. Tuberculosis contact investigations. Please Dis 2000;4:12 – 7.
don’t fail me now. Am J Respir Crit Care Med 2004; [77] Diel R, Meywald-Walter K, Gottschalk R, et al.
169:779 – 81. Ongoing outbreak of tuberculosis in a low incidence
[64] Behr MA, Hopewell PC, Paz EA, et al. Predictive community: a molecular-epidemiologic evaluation.
value of contact investigation for identifying recent Int J Tuberc Lung Dis 2004;8(7):855 – 61.
transmission of Mycobacterium tuberculosis. Am J [78] Leonhardt KK, Gentile F, Gilbert BP, et al. A cluster
Respir Crit Care Med 1998;158:465 – 9. of tuberculosis among crack house contacts in San
[65] Bennett DE, Onorato IM, Ellis BA, et al. DNA Mateo County, California. Am J Public Health 1994;
fingerprinting of Mycobacterium tuberculosis isolates 84:1834 – 6.
from epidemiologically linked case pairs. Emerg [79] Bock NN, Mallory JP, Mobley N, et al. Outbreak
Infect Dis 2002;8:1224 – 9. of tuberculosis associated with a floating card game
[66] Sun SJ, Bennett DE, Flood J, et al. Identifying the in the rural south: lessons for tuberculosis contact
sources of tuberculosis in young children: a multistate investigations. Clin Infect Dis 1998;27:1221 – 6.
investigation. Emerg Infect Dis 2002;8:1216 – 23. [80] Kim SJ, Bai GH, Lee H, et al. Transmission of
[67] Verver S, Warren RM, Munch Z, et al. Proportion of Mycobacterium tuberculosis among high school
tuberculosis transmission that takes place in house- students in Korea. Int J Tuberc Lung Dis 2001;5:
holds in a high incidence area. Lancet 2004;363: 824 – 30.
212 – 4. [81] Bauer J, Kok-Jensen A, Faurschou P, et al. A pro-
[68] Chin DP, Crane CM, Diul MY, et al. Spread of spective evaluation of the clinical value of nation-
Mycobacterium tuberculosis in a community imple- wide DNA fingerprinting of tuberculosis isolates in
menting recommended elements of tuberculosis con- Denmark. Int J Tuberc Lung Dis 2000;4:295 – 9.
trol. JAMA 2000;283:2968 – 74. [82] Edlin BR, Tokars JI, Grieco MH, et al. An outbreak
[69] Kimerling ME, Benjamin WH, Lok KH, et al. of multidrug-resistant tuberculosis among hospital-
Restriction fragment length polymorphism screening ized patients with the acquired immunodeficiency
of Mycobacterium tuberculosis isolates: population syndrome. N Engl J Med 1992;326:1514 – 21.
surveillance for targeting disease transmission in a [83] Haas DW, Milton S, Kreiswirth BN, et al. Nosoco-
community. Int J Tuberc Lung Dis 1998;2:655 – 62. mial transmission of a drug-sensitive W-variant
[70] Lemaitre N, Sougakoff W, Truffot-Pernot C, et al. Mycobacterium tuberculosis strain among patients
Use of DNA fingerprinting for primary surveillance with acquired immunodeficiency syndrome in Ten-
of nosocomial tuberculosis in a large urban hospital: nessee. Infect Control Hosp Epidemiol 1998;19:
detection of outbreaks in homeless people and mi- 635 – 9.
grant workers. Int J Tuberc Lung Dis 1998;2:390 – 6. [84] Chaves F, Dronda F, Cave MD, et al. A longitudinal
[71] McNabb SJN, Kammerer JS, Hickey AC, et al. study of transmission of tuberculosis in a large prison
Added epidemiologic value to tuberculosis preven- population. Am J Respir Crit Care Med 1997;155:
tion and control of the investigation of clustered 719 – 25.
genotypes of Mycobacterium tuberculosis isolates. [85] Hanau-Bercot B, Gremy I, Raskine L, et al. A one-
Am J Epidemiol 2004;160:589 – 97. year prospective study (1994 – 1995) for a first
[72] Barnes PF, Yang Z, Preston-Martin S, et al. Patterns evaluation of tuberculosis transmission in French
of tuberculosis transmission in central Los Angeles. prisons. Int J Tuberc Lung Dis 2000;4:853 – 9.
JAMA 1997;278:1159 – 63. [86] Mohle-Boetani JC, Miguelino V, Dewsnup D, et al.
[73] Bishai WR, Fraham NM, Harrington S, et al. Tuberculosis outbreak in a housing unit for human
Molecular and geographic patterns of tuberculosis immunodeficiency virus-infected patients in a correc-
transmission after 15 years of directly observed tional facility: transmission risk factors and effective
therapy. JAMA 1998;280:1679 – 84. outbreak control. Clin Infect Dis 2002;34:668 – 76.
[74] Daley CL, Small PM, Schecter GF, et al. An outbreak [87] Toungoussova OS, Mariandyshev A, Bjune G, et al.
of tuberculosis with accelerated progression among Molecular epidemiology and drug resistance of
persons infected with the human immunodeficiency Mycobacterium tuberculosis isolates in the Archangel
virus. An analysis using restriction-fragment-length prison in Russia: predominance of the W-Beijing
polymorphisms. N Engl J Med 1992;326:231 – 5. clone family. Clin Infect Dis 2003;37:665 – 72.
[75] Yaganehdoost A, Graviss EA, Ross MW, et al. [88] Mangura BT, Napolitano EC, Passannante MR, et al.
Complex transmission dynamics of clonally related Mycobacterium tuberculosis miniepidemic in a
virulent Mycobacterium tuberculosis associated with church gospel choir. Chest 1998;113:234 – 7.
barhopping by predominantly human immunodefi- [89] Centers for Disease Control and Prevention. HIV-
ciency virus-positive gay men. J Infect Dis 1999;180: related tuberculosis in a transgender network –
1245 – 51. Baltimore, Maryland and New York City area,
230 daley
1998 – 2000. MMWR Morb Mortal Wkly Rep 2000; Belgrade, Central Serbia. J Clin Microbiol 2003;
49:317 – 20. 41(9):4372 – 7.
[90] Jereb JA, Burwen DR, Dooley SW, et al. Nosocomial [104] Zolnir-Dovc M, Poljak M, Erzen D, et al. Molecular
outbreak of tuberculosis in a renal transplant unit: epidemiology of tuberculosis in Slovenia: results of a
application of a new technique for restriction frag- one-year (2001) nation-wide study. Scand J Infect Dis
ment length polymorphism analysis of Mycobacte- 2003;35(11 – 12):863 – 8.
rium tuberculosis isolates. J Infect Dis 1993;168: [105] Lockman S, Sheppard JD, Braden CR, et al. Mo-
1219 – 24. lecular and conventional epidemiology of Mycobac-
[91] Wilkinson D, Crump J, Pillay M, et al. Nosocomial terium tuberculosis in Botswana: a population-based
transmission of tuberculosis in Africa documented prospective study of 301 pulmonary tuberculosis pa-
by restriction fragment length polymorphism. Trans tients. J Clin Microbiol 2001;39:1042 – 7.
R Soc Trop Med Hyg 1997;91:318. [106] Chan-Yeung M, Tam C-M, Wong H, et al. Molecular
[92] Frieden TR, Woodley CL, Crawford JT, et al. and conventional epidemiology of tuberculosis in
The molecular epidemiology of tuberculosis in New Hong Kong: a population-based prospective study.
York City: the importance of nosocomial transmis- J Clin Microbiol 2003;41(6):2706 – 8.
sion and laboratory error. Tuber Lung Dis 1996;77: [107] Narayanan S, Das S, Garg R, et al. Molecular
407 – 13. epidemiology of tuberculosis in a rural area of high
[93] Ikeda RM, Birkhead GS, DiFerdinando Jr GT, et al. prevalence in south India: implications for disease
Nosocomial tuberculosis: an outbreak of a strain control and prevention. J Clin Microbiol 2002;
resistant to seven drugs. Infect Control Hosp Epi- 40(12):4785 – 8.
demiol 1995;16:152 – 9. [108] Geng E, Kreiswirth B, Driver C, et al. Changes in the
[94] Johnson KR, Braden CR, Cairns KL, et al. Trans- transmission of tuberculosis in New York City from
mission of Mycobacterium tuberculosis from medical 1990 to 1999. N Engl J Med 2002;346:1453 – 8.
waste. JAMA 2000;284:1683 – 8. [109] Diel R, Rusch-Gerdes S, Niemann S. Molecular epi-
[95] Agerton T, Valway S, Gore B, et al. Transmission demiology of tuberculosis among immigrants in Ham-
of a highly drug-resistant strain (strain W1) of burg, Germany. J Clin Microbiol 2004;42:2952 – 60.
Mycobacterium tuberculosis. Community outbreak [110] Verver S, Warren RM, Munch Z, et al. Transmission
and nosocomial transmission via a contaminated of tuberculosis in a high incidence urban community
bronchoscope. JAMA 1997;278(13):1073 – 7. in South Africa. Int J Epidemiol 2004;33:351 – 7.
[96] Michele TM, Cronin WA, Graham NM, et al. [111] McNabb SJN, Braden CR, Navin TR. DNA finger-
Transmission of Mycobacterium tuberculosis by a printing of Mycobacterium tuberculosis: lessons
fiberoptic bronchoscope. Identification by DNA learned and implications for the future. Emerg Infect
fingerprinting. JAMA 1997;278:1093 – 5. Dis 2002;8:1314 – 9.
[97] Hernandez-Garduno E, Kunimoto D, Wang L, et al. [112] Kong PM, Tapy J, Calixto P, et al. Skin-test screening
Predictors of clustering of tuberculosis in Greater and tuberculosis transmission among the homeless.
Vancouver: a molecular epidemiologic study. CMAJ Emerg Infect Dis 2002;8:1280 – 4.
2002;167:349 – 52. [113] Soborg C, Soborg B, Pouelsen S, et al. Doubling
[98] Kulaga S, Behr M, Musana K, et al. Molecular epide- of the tuberculosis incidence in Greenland over an
miology of tuberculosis in Montreal. CMAJ 2002; 8-year period (1990 – 1997). Int J Tuberc Lung Dis
167(4):353 – 4. 2001;5:257 – 65.
[99] FitzGerald JM, Fanning A, Hoepnner V, et al, [114] Bradford WZ, Koehler J, El-Hajj H, et al. Dissemi-
Canadian Molecular Epidemiology of TB Study nation of Mycobacterium tuberculosis across the San
Group. The molecular epidemiology of tuberculosis Francisco Bay Area. J Infect Dis 1998;177:1104 – 7.
in western Canada. Int J Tuberc Lung Dis 2003;7(2): [115] Ellis BA, Crawford JT, Braden CR, et al. Molecu-
132 – 8. lar epidemiology of tuberculosis in a sentinel sur-
[100] Blackwood KS, Al-Azem A, Elliott LJ, et al. veillance population. Emerg Infect Dis 2002;8:
Conventional and molecular epidemiology of tuber- 1197 – 209.
culosis in Manitoba. BMC Infect Dis 2003;3:18. [116] van Soolingen D, Qian L, de Haas PE, et al.
[101] Alland D, Kalkut GE, Moss AR, et al. Transmission Predominance of a single genotype of Mycobacterium
of tuberculosis in New York City. An analysis by tuberculosis in countries of east Asia. J Clin Micro-
DNA fingerprinting and conventional epidemiologic biol 1995;33:3234 – 8.
methods. N Engl J Med 1994;330(24):1710 – 6. [117] Anh DD, Borgdorff MW, Van LN, et al. Mycobacte-
[102] Diel R, Schneider S, Meywald-Walter K, et al. rium tuberculosis Beijing genotype emerging in
Epidemiology of tuberculosis in Hamburg, Germany: Vietnam. Emerg Infect Dis 2000;6:302 – 5.
long-term population-based analysis applying classi- [118] Krüüner A, Hoffner SE, Sillastu H, et al. Spread of
cal and molecular epidemiologic techniques. J Clin drug-resistant pulmonary tuberculosis in Estonia.
Microbiol 2002;40:532 – 9. J Clin Microbiol 2001;39:3339 – 45.
[103] Vukovic D, Rusch-Gerdes S, Savic B, et al. Molecu- [119] Niemann S, Rusch-Gerdes S, Richter E. IS6110
lar epidemiology of pulmonary tuberculosis in fingerprinting of drug-resistant Mycobacterium tuber-
molecular epidemiology 231
culosis strains isolated in Germany during 1995. of Beijing genotype in Latvian multidrug-resistant
J Clin Microbiol 1997;35:3015 – 20. Mycobacterium tuberculosis isolates. Int J Tuberc
[120] Caminero JA, Pena MJ, Campos-Herrero MI, et al. Lung Dis 2003;7:1097 – 103.
Epidemiological evidence of the spread of a Myco- [127] Bifani PJ, Plikaytis BB, Kapur V, et al. Origin and
bacterium tuberculosis strain of the Beijing genotype interstate spread of a New York City multidrug-
on Gran Canaria Island. Am J Respir Crit Care Med resistant Mycobacterium tuberculosis clone family.
2002;164:1165 – 70. JAMA 1996;275:452 – 7.
[121] Borgdorff MW, de Haas P, Kremer K, et al. [128] Bifani PJ, Mathema B, Liu Z, et al. Identification of
Mycobacterium tuberculosis Beijing genotype, the a W variant outbreak of Mycobacterium tuberculosis
Netherlands. Emerg Infect Dis 2003;9:1310 – 3. via population-based molecular epidemiology. JAMA
[122] Richardson M, van Lill SW, van der Spuy GD, et al. 1999;282:2321 – 7.
Historic and recent events contribute to the disease [129] Agerton TB, Valway SE, Blinkhorn RJ, et al. Spread
dynamics of Beijing-like Mycobacterium tuberculosis of strain W, a highly drug-resistant strain of Myco-
isolates in a high incidence region. Int J Tuberc Lung bacterium tuberculosis, across the United States. Clin
Dis 2002;6:1001 – 11. Infect Dis 1999;29:85 – 92.
[123] van Crevel R, Nelwan RH, de Lenne W, et al. [130] Glynn JR, Whiteley J, Bifani PJ, et al. Worldwide
Mycobacterium tuberculosis Beijing genotype strains occurrence of Beijing/W strains of Mycobacterium
associated with febrile response to treatment. Emerg tuberculosis: a systematic review. Emerg Infect Dis
Infect Dis 2001;7:880 – 3. 2002;8:843 – 9.
[124] Lan NT, Lien HT, Tungle B, et al. Mycobacterium [131] Bifani PJ, Mathema B, Kurepina NE, et al. Global
tuberculosis Beijing genotype and risk for treatment dissemination of the Mycobacterium tuberculosis
failure and relapse, Vietnam. Emerg Infect Dis 2003; W-Beijing family strains. Trends Microbiol 2002;10:
9(12):1633 – 5. 45 – 52.
[125] Lillebaek T, Andersen AB, Dirksen A, et al. Myco- [132] Lopez B, Aguilar D, Orozco H, et al. A marked
bacterium tuberculosis Beijing genotype. Emerg difference in pathogenesis and immune response
Infect Dis 2003;9:1553 – 7. induced by different Mycobacterium tuberculosis
[126] Tracevska T, Jansone I, Baumanis V, et al. Prevalence genotypes. Clin Exp Immunol 2003;133:30 – 7.
Clin Chest Med 26 (2005) 327 – 340
0272-5231/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccm.2005.02.013 chestmed.theclinics.com
328 o’brien & spigelman
TB and would improve the effectiveness of the Clinical Trials Program of the International Union
WHO-supported MDR-TB treatment programs Against Tuberculosis and Lung Disease. With sup-
known as DOTS-Plus [7]. port from the European Community, the European
The resurgence of tuberculosis in the United and Developing Countries Clinical Trials Partnership
States beginning in the late 1980s, coupled with the aims to provide A600 million over 5 years to per-
outbreaks of MDR-TB largely associated with HIV form clinical trials and to establish capacity for the
infection, led to increased federal support for both conduct of high-quality clinical trials, including
domestic and global tuberculosis control [8]. That those for tuberculosis, throughout Africa [12]. Under-
support has resulted in continued declines in tuber- pinning all this effort is the Global Alliance for TB
culosis in the United States beginning in 1993 and a Drug Development (TB Alliance), a recently estab-
renewed call for the elimination of tuberculosis as a lished organization that is forging public – private part-
public health problem [9]. An important component nerships with the objective of building a portfolio of
of the tuberculosis elimination strategy in the United new tuberculosis drugs and bringing a major new
States is the treatment of individuals who have LTBI tuberculosis drug to market in the next decade [13].
and are at increased risk of developing active TB This article reviews two classes of compounds
[10]. The most widely used LTBI treatment regimen, that have advanced into phase II and III clinical trials,
9 months of isoniazid, is associated with significant long-acting rifamycins and fluoroquinolones, and a
nonadherence, however. Thus, a more easily admin- number of other drugs that have entered or it is hoped
istered LTBI treatment regimen is a priority in a will enter clinical development in the near future.
number of low-incidence countries.
Tuberculosis drug development—a changing Rifapentine: the search for widely spaced
environment intermittent treatment
Increased resources directed toward tuberculosis Rifampin is the cornerstone of modern short-
drug development are now being marshaled from course tuberculosis treatment, but rifampin-based
both the public and private sectors. Governmental regimens must be administered for at least 6 months
organizations, such as the United States National In- for optimal effectiveness. Although this treatment is
stitutes of Health (NIH), are investing in basic re- also highly effective when given three times per week
search aimed at the identification of new drug targets throughout the course of treatment [14], more widely
and a better understanding of the phenomena of spaced regimens are less effective and may be as-
mycobacterial latency. Foundations, such as the Bill sociated with acquired drug resistance in HIV-
and Melinda Gates Foundation, are supporting re- infected patients, even when properly taken.
search and development to enhance the understand- A number of rifamycin derivatives with much
ing of the basic biology of tuberculosis and to longer serum half-lives than that of rifampin (2 –
develop new tuberculosis drugs. A number of small 4 hours) have been evaluated in regimens given in-
biotech companies have programs focused on the termittently. The first of these compounds to undergo
identification of new chemical entities with anti- clinical investigation was rifabutin [15]. The initial
mycobacterial activities that could become lead com- clinical trials of the drug focused on the prevention
pounds in the drug-development process. Several of Mycobacterium avium complex (MAC) infection
large pharmaceutical companies, such as GlaxoSmith- in HIV-infected patients [16]. Although the drug was
Kline (Brentford, United Kingdom), AstraZeneca approved for MAC prophylaxis in the United States
(London, United Kingdom), and Novartis (Basel, and for the treatment of tuberculosis in several other
Switzerland), have launched programs directed at the countries, it now is used primarily as a substitute for
discovery and development of new tuberculosis rifampin in patients who cannot use that drug be-
drugs. Other companies, notably Aventis and Bayer, cause of drug – drug interactions [17]. A TBTC trial
have made compounds available for clinical studies. of a rifabutin-containing regimen given twice weekly
At the same time, the clinical trials infrastructure, in HIV-infected patients found high rates of acquired
which had been greatly eroded beginning in the early rifamycin resistance among patients who had more
1980s, is being reestablished with the formation of advanced immunosuppression, leading to CDC recom-
groups such as the United States Tuberculosis Trials mendations against the use of widely spaced treat-
Consortium (TBTC) [11] sponsored by the Centers ment of tuberculosis with rifamycin regimens in such
for Disease Control and Prevention (CDC) and the patients [18].
new drugs for tuberculosis 329
Another long-acting rifamycin derivative, rifalazil, dard therapy given daily for 6 months [24]. These
has an even longer half-life and potent activity in studies provided the scientific underpinning for the
animal models suggesting that it might be used in large phase III trial that was begun by CDC in
ultrashort treatment regimens [19]. One attractive 1995 and subsequently became known as TBTC
feature of the compound is its rather low potential for Study 22.
enzyme induction and drug interactions [20]. Initial Study 22 was an unmasked clinical trial that
phase I tolerability studies, however, found relatively randomly assigned adults who had newly diagnosed,
high rates of side effects manifesting as a flulike drug-susceptible pulmonary tuberculosis to a
syndrome when the drug was administered as a single 4-month (16-week) continuation-phase regimen of
50-mg dose [21]. The hypothesized mechanism either once-weekly rifapentine-isoniazid or twice-
causing the dose limiting side effect is release of weekly rifampin-isoniazid following successful com-
cytokines with evidence for increased interleukin-6 pletion of a standard 2-month induction phase [25].
levels in the serum. Following an early bactericidal The primary study end points were treatment failure
activity (EBA) study that did not demonstrate drug and relapse and safety and tolerability of rifapentine.
activity of once-weekly rifalazil (at 10- and 25-mg The rifamycins were dosed at 600 mg and isoniazid at
doses) plus isoniazid given for 2 weeks [22], further 900 mg. Although the trial focused on HIV-negative
clinical development stopped. It is believed that patients, HIV-positive patients were also enrolled
closely related compounds can be identified that are initially to gain experience with this important subset
better tolerated and lack the propensity for enzyme of patients. Enrollment of HIV-positive patients was
induction. Currently, there is significant interest in the stopped early in the trial, however, following the
use of rifalazil for the treatment of chlamydia finding of a high rate of relapse with acquired ri-
infections [23]. fampin monoresistance among HIV-positive patients
The greatest interest and investment in long-acting assigned to the rifapentine arm [26].
rifamycins has been in rifapentine, a cyclopentyl- A total of 1003 HIV-negative patients were
substituted rifampin with a half-life of 14 to 18 hours enrolled into the completed study. The crude rate of
in normal adults. Following a 600-mg dose, serum failure and relapse was significantly higher in the
levels in excess of the minimum inhibitory concen- rifapentine arm (9.2% versus 5.6%, P = 0.04). In a
tration (MIC) persist beyond 72 hours, suggesting multivariate analysis, the factors statistically associ-
that the drug might be useful in intermittent regimens ated with an adverse outcome were the presence of
(Fig. 1). A series of experimental studies in mice cavitary disease on chest radiograph, sputum culture
found that a once-weekly continuation phase of rifa- positivity at study entry (ie, at the end of the intensive
pentine and isoniazid for 4 months following a stan- phase of therapy), white race, and weight less than
dard 2-month induction phase with daily isoniazid, 90% of ideal body weight at time of the diagnosis of
rifampin, and pyrazinamide was as effective as stan- tuberculosis. The treatment regimen was not associ-
12.00
10.00 Rifaentine
Plasma Concentration
Rifampin
8.00
(µg/mL)
6.00
4.00
2.00
0.00
0 12 24 36 48 60 72
Time (h)
Fig. 1. Rifampin and rifapentine time-concentration curves following 600-mg dose in normal adults.
330 o’brien & spigelman
20
Once weekly
15 rifapentine/isoniazid
10 Twice weekly
rifampin/isoniazid
5
0
Culture+ Culture+ Culture- Culture-
Cavity+ Cavity- Cavity+ Cavity-
Fig. 2. Tuberculosis Trials Consortium study 22. Relapse rate by arm, cavitary chest radiograph, and 2-month culture. (Adapted
from Tuberculosis Trials Consortium. Rifapentine and isoniazid once a week versus rifampicin and isoniazid twice a week for
treatment of drug-susceptible pulmonary tuberculosis in HIV-negative patients: a randomised clinical trial. Lancet
2002;360:528 – 34; with permission.)
ated with an adverse outcome. Cavitary disease and Following the completion of Study 22, the TBTC
culture positivity after 2 months were also predictors undertook a large phase II trial of higher rifapentine
of an adverse outcome among patients in the rifampin doses. In Study 25, 150 HIV-negative patients who
arm (Fig. 2). Among patients who had noncavitary had drug-susceptible pulmonary tuberculosis and
tuberculosis and negative 2-month sputum cultures, completed initial-phase treatment were randomly as-
the relapse rate was low in both arms. Rifapentine signed to 600, 900, and 1200 mg rifapentine given
was well tolerated, and rates of adverse events were once weekly with isoniazid for 16 weeks. The ri-
similar in both treatment groups, with 3% of pa- fapentine dose was masked with the use of dummy
tients in both groups discontinuing treatment be- tablets of rifapentine. The primary study end points
cause of a drug-related adverse event. These results were adverse events and drug discontinuation. All
were similar to those from a study in Hong-Kong that regimens were well tolerated, and only one patient
used Chinese-manufactured rifapentine of inferior assigned to the 1200-mg dose stopped treatment be-
bioavailability [27] and with those from a company- cause of a possible drug-related adverse event [32].
sponsored trial that enrolled patients largely from Because the results of Study 22 were known when
Africa [28]. this study began, the protocol was modified to pro-
The TBTC study results led to new recommen- vide extended treatment for an additional 3 months
dations for the use of the rifapentine-isoniazid (or 12 weeks) for patients who had cavitary disease
continuation-phase regimen for HIV-negative adults and had positive sputum cultures at entry (ie, at
who have drug-susceptible, noncavitary tuberculo- 2 months). Twenty such patients were enrolled,
sis and negative acid-fast bacillus (AFB) smears at received extended treatment, and were followed
2 months [29]. This category includes approxi- prospectively for relapse. Only one patient who was
mately 40% of patients in the United States who assigned to the 600-mg dose relapsed. The relapse
have newly diagnosed pulmonary tuberculosis. The rate of 5%, when compared with historical data from
regimen provides substantial cost savings for these Study 22 (22%), suggests that extended treatment and
patients, because encounters for directly observed higher rifapentine doses may provide more effective
treatment during the continuation phase are reduced treatment for patients who are at increased risk of
by 50% [30]. relapse [33]. The results also suggest that the 900-mg
Rifapentine-based treatment is not recommended rifapentine dose would be appropriate to use in
for patients who have more advanced tuberculosis or subsequent trials.
patients who have HIV infection. Pharmacokinetic Experimental studies have also suggested that
studies undertaken as part of Study 22 indicated that once-weekly rifapentine and isoniazid for as short a
low levels of isoniazid and rapid isoniazid acetylation period as 3 months may provide effective treatment
were associated with relapse, suggesting that a more for LTBI, comparable to that conferred by 6 months
effective companion drug might improve once- of daily isoniazid or by 2 months of daily rifampin
weekly treatment [31]. Experimental studies have and pyrazinamide [34]. Based on these findings,
also suggested that, in addition to a better companion the TBTC has embarked on an ambitious study of
drug, higher doses of rifapentine might also result in rifapentine/isoniazid for LTBI treatment, intending
more effective treatment [24]. to enroll and randomly assign 8000 patients to either
new drugs for tuberculosis 331
9
Mean CFU/Spleen (log 10)
8
7
6
5
4
3
2
1
0
Day 1 Day 30 INH, 100 25 50 100 25 50 100
25 (1/7)
Controls Sparfloxacin Moxifloxacin
Fig. 3. Thirty-day experimental study of isoniazid (INH), sparfloxacin, and moxifloxacin in a mouse model of acute tuberculosis.
Drug doses in mg/kg. (Adapted from Ji B, Lounis N, Maslo C, et al. In vitro and in vivo activities of moxifloxacin and
clinafloxacin against Mycobacterium tuberculosis. Antimicrob Agents Chemother 1998;42:2066 – 9; with permission.)
332 o’brien & spigelman
10
Untreated
8
4 2HRM/4HR
2HZM/4HM
2
2RZM/4RM
0
0 1 2 3 4 5 6
Duration of Treatment, Months
*H=isoniazid, R=rifampin, Z=pyrazinamide, M=moxifloxacin
Fig. 4. Experimental study of moxifloxacin-containing regimens in murine tuberculosis. (Adapted from Nuermberger EL,
Yoshimatsu T, Tyagi S, et al. Moxifloxacin-containing regimen greatly reduces time to culture conversion in murine tuberculosis.
Am J Respir Crit Care Med 2004;169:334 – 5; with permission.)
moxifloxacin also has potent sterilizing activity and one of four 2-month intensive-phase regimens: two
might substantially improve the efficacy of once- standard-treatment regimens given either daily or
weekly rifapentine treatment, replacing isoniazid that three times weekly or similar regimens in which
has been shown in clinical studies to be a poor com- moxifloxacin replaces ethambutol, with assignment
panion drug [43]. The most recent study found that masked by placebo moxifloxacin and ethambutol.
the combination of rifampin, pyrazinamide, and The primary study end points are 2 month sputum
moxifloxacin had substantially greater sterilizing culture conversion and withdrawal because of ad-
activity than the standard regimen, again suggesting verse events. Investigators from Johns Hopkins
the possibility that the drug would permit significant University are working with colleagues from Rio de
shortening of treatment (Fig. 4) [44]. Janeiro on a similar study that is supported by the
The results of two small EBA studies have dem- United States Food and Drug Administration Office
onstrated that moxifloxacin has bactericidal activity of Orphan Products Development (R. Chaison, per-
superior to that of rifampin and perhaps comparable sonal communication, 2004).
to that of isoniazid, the most potent bactericidal drug A product development team supported by the
in EBA studies [45,46]. The only other published United Nations Childrens Fund/United Nations De-
experience with moxifloxacin treatment of tubercu- velopment Program/World Bank/WHO Special Pro-
losis is a small case series that indicated good toler- gram for Research and Training in Tropical Diseases
ability to chronic administration of the drug [47]. The and the European Commission is embarking on
next step in the clinical development of moxifloxa- several studies of a gatifloxacin fixed-dose com-
cin for TB is the conduct of a series of phase II bination product for the treatment of drug-susceptible
clinical trials in which moxifloxacin replaces various tuberculosis. These efforts include preclinical phar-
drugs in the initial 2-month phase of TB treatment macology and toxicology studies and a phase I study
and where sputum culture conversion at 2 months designed to compare the drug – drug/pharmacokinetic
is the primary study end point [48]. Data from such interactions of gatifloxacin and isoniazid, rifampin,
studies, which have historically taken 2 years to and pyrazinamide. A phase II study is being con-
complete, are usually required to proceed to the larger ducted in Durban, South African, randomly assigning
and more costly phase III trials that commonly take newly diagnosed patients to one of three fluoroqui-
much longer to complete. nolone-containing regimens (ofloxacin, moxifloxacin,
To develop clinical data that would justify larger and gatifloxacin) in combination with isoniazid,
phase III efficacy trials of moxifloxacin, the TBTC rifampin, and pyrazinamide during the first 2 months
has embarked on a phase II trial of the drug, Study of treatment. A variety of bacteriologic markers are
27. This study randomly assigns newly diagnosed, being evaluated as potential surrogate markers of
AFB-positive, HIV-positive and -negative patients treatment response. A large phase III trial of gati-
who have suspected pulmonary tuberculosis to floxacin included in a 4-month regimen that intends
new drugs for tuberculosis 333
to enroll over 2000 patients at centers in five coun- spleens at 12.5 mg/kg to the same extent as isoniazid
tries in sub-Saharan Africa was expected to begin (25 mg/kg). When therapy was started on day 14 after
in late 2004 (C. Lienhardt, personal communica- inoculation and continued until day 70 (established
tion, 2004). infection model), 12.5 mg/kg of R207910 was at least
as active in decreasing CFU count in lung as was
isoniazid (25 mg/kg) or rifampin (10 mg/kg). At a
dose of 25 mg/kg, R207910 was even more active
The emerging tuberculosis drug pipeline than at 12.5 mg/kg, reducing lung CFU count from
6 to 0.4 log. The combination of R207910 with any
In addition to the rifamycin derivatives and fluo- two of the three commonly used drugs (isoniazid,
roquinolones, a variety of other compounds or classes rifampin, and pyrazinamide) was more effective than
of compounds are under investigation as potential the standard regimen of isoniazid, rifampin, and
antimycobacterial drugs. These include a diarylquino- pyrazinamide. In fact, the combination of R207910,
line (R207910), a nitroimidazopyran (PA-824), a isoniazid, and pyrazinamide and the combination of
nitro-dihydroimidazo-oxazole (OPC 67,683), a pyr- R207910, rifampin, and pyrazinamide both resulted
role (LL3858), macrolides, oxazolidinones, and a in negative spleen and lung cultures after 8 weeks
diamine (SQ109). of therapy.
Pending results of the phase I studies, the ability
of R207910 to shorten the therapy of active TB will
Diarylquinolines (R207910) be tested.
sensitive and drug-resistant tuberculosis. In vitro tion models are below the acute and chronic toxic
studies demonstrate that the MIC of PA-824 against thresholds observed for PA-824 in mice.
a variety of drug-sensitive tuberculosis isolates More recent studies by Grosset et al [51] have
(0.015 – 0.25 mg/mL) is similar to that of isoniazid indicated that, in a murine model, the minimum
(0.03 – 0.06 mg/mL). PA-824 is highly selective, with effective dose (defined as the minimum dose which
potent activity only against bacille Calmette-Guerin prevents the development of gross lung lesions and
(BCG) and M. tuberculosis among the mycobacterial splenomegaly) of PA-824 is 12.5 mg/kg/day, that the
species tested, and without significant activity against absence of lung lesions on gross inspection correlates
a broad range of gram-positive and gram-negative well with bacteriostatic activity measured by CFU
bacteria (with the exception of H. pylori and some count, that the minimum bactericidal dose (defined as
anaerobes). In vitro studies using anaerobic culture the minimum dose which reduces the long colony
models indicate that PA-824 has activity against forming unit counts by 99%) is 100 mg/kg/day, and
nonreplicating bacilli, whereas isoniazid does not that the activity of PA-824 at 100 mg/kg is com-
have activity in these models. Finally, PA-824 has parable to the activity of isoniazid at 25 mg/kg.
been shown to have activity against strains of tu- The potential genotoxicity of PA-824 was exam-
berculosis with known resistance to standard anti- ined further with chromosomal aberration, mouse
tuberculosis therapies, indicating a novel mechanism micronucleus, and mouse lymphoma tests. The
of action. results indicate that PA-824 is not genotoxic. Fur-
To evaluate in vivo activity, PathoGenesis tested thermore, in vitro studies indicate that PA-824 nei-
PA-824 in a mouse model of tuberculosis, employing ther inhibits nor is metabolized by major P450
an M. tuberculosis reporter strain expressing firefly enzyme isoforms.
luciferase. PA-824 was administered orally at 25, 50, Pharmacokinetic studies have been performed in
and 100 mg/kg/day in mice for 10 days, with the rat, dog, and monkey, because the systemic
isoniazid used in the control arm. Administration of exposure in dogs is low for both males and females
PA-824 at all doses significantly reduced M. tuber- secondary to poor absorption and rapid metabolism.
culosis levels in both spleen and lung compared with Results of the single-dose studies indicate that the
controls and demonstrated a linear dose response. In half-life of PA-824 is approximately 2 to 5 hours in
longer-term studies, PA-824 at 50 mg/kg/day dem- male rats and monkeys and trends toward a longer
onstrated reductions in bacillary burden similar to half-life in female rats (8 – 9 hours). The half-life in
isoniazid at 25 mg/kg/day in murine lungs, and all dogs is shorter (1 – 2 hours). In monkeys, single
mice treated with PA-824 survived infection, doses of PA-824 are rapidly absorbed with a time
whereas all untreated control animals died by day 35. to maximal concentration (Tmax) of 3.33 hours or
Daily oral administration of PA-824 at 37 mg/kg/day less, whereas Tmax in the rat ranges up to 8 hours.
for 35 days in a guinea pig aerosol infection model There was no significant effect of sex on rate of
also caused statistically significant reductions of absorption in any species. There was not a sig-
M. tuberculosis in lungs and spleens compared nificant food-effect on PA-824 pharmacokinetics in
with controls, reductions comparable to those caused the rat.
by isoniazid. The pharmacokinetics of PA-824 was determined
The activity of PA-824 against MDR-TB isolates in plasma, heart, liver, kidney, spleen, and lung
and against both replicating (aerobic) and nonrep- following a single 100-mg/kg oral dose of PA-824
licating (anaerobic) M. tuberculosis bacilli indicates in rats. The time to reach maximal concentrations of
this compound has a novel mechanism of action. PA-824 in these tissues was 4 hours as compared
PA-824 seems to inhibit significantly both protein with 6 hours in plasma. Exposure (area under the
and lipid synthesis but does not affect nucleic acid curve) in tissues was approximately three- to eight-
synthesis. PA-824 produces an accumulation of fold higher than that in plasma. These data suggest
hydroxymycolic acid with a concomitant reduction that, in the rat model, penetration of PA-824 into
in ketomycolic acids, suggesting inhibition of an lung, spleen, and other tissues is extensive. In re-
enzyme responsible for the oxidation of hydroxy- peated dose studies, there was no evidence of accu-
mycolate to ketomycolate. mulation in the rat or monkey.
Unlike the Ciba-Geigy lead compound, CGI- Two 14-day good – laboratory practice toxicology
17341, PA-824 has not demonstrated mutagenicity studies, one in the rat and one in the monkey, have
in the Ames test (with or without S9 activation), and been completed. The results of these studies indicate
initial toxicity studies indicated the doses needed for that toxicity is observed when exposures at or above
therapeutic activity in murine and guinea pig infec- approximately 500 mg/hour/mL are achieved. Phase I
new drugs for tuberculosis 335
studies of PA-824 are planned for the first quarter antituberculous drugs, LL3858 sterilizes lungs and
of 2005. spleens of infected animals in a shorter timeframe
than conventional therapy.
Dihydroimidazo-oxazoles (OPC-67683)
Macrolides
OPC-67683 is a newly synthesized nitro-dihydro-
imidazo-oxazole derivative under development by The Institute for Tuberculosis Research, College
Otsuka Pharmaceutical Company (Tokyo, Japan) of Pharmacy at the University of Illinois at Chicago,
for the treatment of tuberculosis and is currently in in conjunction with the TB Alliance, is currently
phase I study in normal volunteers (Otsuka Pharma- studying the potential for macrolide antibiotics in the
ceutical Company, personal communication, 2004). treatment of tuberculosis. Among approved antimi-
The compound has potent in vitro antimicrobial crobial agents that do not include tuberculosis as an
activity against M. tuberculosis, with MICs against indication, the macrolides are one of the more
H37Rv and 67 clinically isolated strains ranging from promising to yield a clinically useful tuberculosis
0.006 to 0.024 mg/mL. Furthermore, OPC-67683 drug. This potential is based on their oral bioavail-
shows no cross-resistance with any of the currently ability and distribution to the lungs, low toxicity,
used first-line tuberculosis drugs, most likely indi- infrequent adverse reactions, extensive intracellular
cating a novel mechanism of action. Therefore the concentration and activity, anti-inflammatory activity,
compound may be of benefit both in shortening and, perhaps most importantly, demonstrated clinical
duration of therapy in the treatment of active disease utility and bactericidal activity in infections caused
and in the treatment of MDR-TB. by several pathogenic and opportunistic mycobac-
In vivo studies using a chronic mouse model of teria, including M. avium, M. leprae, M chelonei, and
tuberculosis have demonstrated the efficacy of OPC- M. fortuitum.
67683 to be superior to that of the currently used Erythromycin, the first-generation prototypical
tuberculosis drugs. In the mouse model, the dose that macrolide, is a natural product derived from Strepto-
provided the effective plasma concentration of myces erythreus. The compound interferes with
0.100 mg/mL was 0.625 mg/kg, confirming the re- protein synthesis and possesses most of the favorable
markable in vivo potency of OPC-67683. properties mentioned previously but suffers from a
In other nonclinical in vitro and in vivo studies, short serum half-life and acid lability, which results in
OPC-67683 does not have any antagonistic activity gastric motility – based discomfort. In addition, activ-
with other first-line tuberculosis drugs when used in ity is restricted to gram-positive bacteria.
combination. Combinations with other first-line thera- Therefore, second-generation macrolides with
peutic drugs reveal synergistic, additive, or no appre- superior acid stability and serum half-life were devel-
ciable interactions. oped. Clarithromycin, roxithromycin, and azithromy-
cin represent the most successful second-generation
Pyrrole (LL3858) macrolides. It quickly became apparent that the
second-generation macrolides were, along with rifa-
Pyrrole derivatives were first described by Deidda butin, the most active clinical agents against the
et al [52] as having fairly potent antimycobacterial MAC. With the exception of azithromycin (an azalide
activities against several strains of M. tuberculosis. that possesses a spectrum of activity different from
The MICs were between 0.7 and 1.5 mg/mL for that of other macrolides), these compounds also were
the most potent derivative, 1,5-diaryl-2-methyl-3- found to possess potent activity against M. leprae in
(4-methylpiperazin-1-yl) methyl-pyrrole (BM212). macrophages and mice and were shown to be effec-
The activity of BM212 against various drug-resistant tive in clinical trials. Clarithromycin is currently
strains of M. tuberculosis was similar to its activity recommended by the WHO for treatment of leprosy
against sensitive strains, probably indicating a novel in cases of rifampin resistance or intolerance. Other
mechanism of action. Although some nontuberculosis studies demonstrated low MICs or clinical utility of
mycobacterial strains seemed to be sensitive, the second-generation macrolides against M. kansasii,
MICs were higher than for M. tuberculosis. M. marinum, M. xenopi, and other opportunistic
A novel pyrrole compound, LL3858, is currently mycobacterial pathogens. The impressive activity of
in development for tuberculosis by Lupin Limited second-generation macrolides unfortunately did not
(Mumbai, India). This compound has submicromolar include M. tuberculosis.
MICs and seems to be very active in a mouse model The third-generation macrolides, represented
of tuberculosis. In combination with currently used largely by the ketolides, were developed with the
336 o’brien & spigelman
HO R2 Me Me
H H
N R1 N N
N N R3 Me N
H H H
R4
OH
Ethambutol Library of 63,238 diamines SQ109
Fig. 5. Chemical structures of ethambutol, compounds in the original combinatorial library, and SQ109.
new drugs for tuberculosis 337
1,2-diamine pharmacophore of ethambutol and tested The mechanism of action of SQ109 seems to be
them for activity against M. tuberculosis using an that of a cell wall inhibitor because, like the cell
MIC- and target-based (cell wall) reporter high- wall – targeting antibiotics (ethambutol, isoniazid,
throughput screening assay [62]. These efforts found ethionamide, and thiacetazone), it induces a pro-
2796 mostly lipophilic compounds to be active moter, Rv0341, that was employed in the original
against M. tuberculosis in vitro, and 26 demonstrated luciferase high-throughput screening assay. Because
in vitro activity equal to or greater than (up to 14-fold) the Rv0341 luciferase reporter responds with light
ethambutol. Sixty-nine of the most potent hit com- production to inhibition of a wide variety of enzyme
pounds were later studied in a sequential set of in targets involved in cell wall construction, the specific
vitro and in vivo tests: MIC followed by cytotoxicity target of SQ109 is not known. To address the issue,
screen, followed by activity in infected macrophages, a proteomic study was initiated to identify proteins
followed by permeability evaluation, followed by in in H37Rv M. tuberculosis that are affected by the
vivo efficacy testing, followed by pharmacokinetic drug in comparison with ethambutol and isoniazid.
studies. SQ109 was identified as the most potent The results of this study suggest that most of the
compound in the series and was then subjected to in- 44 distinct proteins whose expression is increased
tensive pharmacokinetic/ pharmacodynamic testing. (ESAT-6 and others) or decreased (MPT64 and
SQ109 is a lipophilic, nonsymmetric derivative of others) by SQ109 were similar to those affected by
1,2-ethylenediamine with unsaturated geranyl and ethambutol. Only two gene products whose functions
bulky adamantane fragments present. SQ109 has are unknown were regulated differently by ethambu-
been synthesized as a stable dihydrochloride salt on tol and SQ109. Similarly, two different genes were
a multikilogram scale with high chemical purity affected, but in opposite directions, by exposure of
(99.7%). The formulation to be used in clinical deve- M. tuberculosis to SQ109 or ethambutol [63].
lopment, hard gelatin capsules, has been developed. The pharmacokinetic/pharmacodynamic profiles
SQ109 has an MIC against M. tuberculosis in the of SQ109 were evaluated in three species (mice,
range of 0.1 to 0.63 mg/mL (broth microdilution, rats, and dogs). Single-dose pharmacokinetic studies
Alamar blue, BACTEC [Becton Dickinson, Franklin in mice indicate that SQ109 has 4% oral bio-
Lakes, New Jersey]). The compound is bactericidal availability as measured by drug concentration in
with 99% inhibition of M. tuberculosis growth in plasma. The high potency of SQ109 in vivo at low
macrophages at its MIC. When tested in vivo (in doses (1 mg/kg) combined with tissue distribution
mice), SQ109 is able to reduce infection in lungs data argue, however, that, despite low bioavailabil-
and spleen by 2 to 2.5 log. It is active against MDR ity, SQ109 antimicrobial effects can be attributed
strains of M. tuberculosis in vitro. SQ109 has a low to effective concentrations achieved at the sites of
mutational frequency in M. tuberculosis in vitro bacterial infection. Although blood concentrations
(2.18 109) and demonstrates enhanced antimyco- remain low, SQ109 distributes into lungs and spleen
bacterial activity in vitro and in vivo when used (target sites of the bacterial infection), greatly
in combination with rifampicin and isoniazid (rapid exceeding the MIC (Fig. 6). Oral administration
mouse model and chronic infection model). of SQ109, 30 to 75 mg/m2 (10 – 25 mg/kg in mice)
A B
25000 25000
Concentrations (ng/g)
Concentrations (ng/g)
20000 20000
1h
1h
15000 15000 4h
4h
10 h
10 h
10000 10000
5000 5000
0 0
Ki r
ey
ng
en
M n
e
Ki r
ey
ng
en
rt
M in
e
a
te e
ar
Fa
Fa
ve
ve
in
in
cl
in
cl
m
ac
ai
ac
a
tin
a
dn
dn
le
le
Lu
Lu
He
He
us
us
rg est
st
Br
st
Br
as
as
Li
Li
Sm tom
m
Sp
Sp
rg es
te
Pl
Pl
Sm to
La Int
La nt
in
in
S
lI
l
e
al
al
Fig. 6. Tissue levels of SQ109 after intravenous administration of 3 mg/kg (A) and oral administration of 25 mg/kg
(B) to mice.
338 o’brien & spigelman
one time per day maintains drug levels above the with the HIV epidemic. Arch Intern Med 2003;163:
MIC without accumulation of the drug in the 1009 – 21.
target tissues. [3] Raviglione MC, Pio A. Evolution of WHO policies for
tuberculosis control, 1948 – 2001. Lancet 2002;359:
The liver may have a first-pass effect on SQ109
775 – 80.
metabolism, resulting in low content of the drug in
[4] Murray CJ, Styblo K, Rouillon A. Tuberculosis in
plasma after oral dosing. P450 reaction phenotyping developing countries: burden, intervention and cost.
suggests exclusive involvement of CYP2D6 and Bull Int Union Tuberc Lung Dis 1990;65:6 – 24.
CYP2C19 in SQ109 metabolism; analysis of metabo- [5] Datta M, Radhamani MP, Selvaraj R, et al. Critical
lites formed upon incubation of SQ109 with human, assessment of smear-positive pulmonary tuberculosis
mouse, dog, and rat microsomes suggest similar patients after chemotherapy under the district tuber-
metabolism of the drug in all tested species. SQ109 is culosis programme. Tuber Lung Dis 1993;74:180 – 6.
undergoing formal preclinical 90-day pharmacology [6] Dye C, Espinal MA, Watt CJ, et al. Worldwide in-
and toxicology studies in preparation for human cidence of multidrug-resistant tuberculosis. J Infect Dis
2002;185:1197 – 202.
clinical trials.
[7] Stop TB. Working Group on DOTS-Plus for MDR-TB.
In summary, SQ109 is a novel 1,2-diamine-based
A prioritised research agenda for DOTS-Plus for
drug candidate with in vitro and in vivo activity multidrug-resistant tuberculosis (MDR-TB). Int J
against M. tuberculosis. It has pharmacokinetic/ Tuberc Lung Dis 2003;7:410 – 4.
pharmacodynamic properties that are characterized [8] Schneider E, Castro KG. Tuberculosis trends in the
by a rapid and broad distribution into various tis- United States, 1992 – 2001. Tuberculosis (Edinb) 2003;
sues (ie, lungs) that is advantageous for tuberculo- 83:21 – 9.
sis infection. [9] Centers for Disease Control and Prevention. Tuber-
culosis elimination revisited: obstacles, opportunities,
and a renewed commitment. Advisory Council for
the Elimination of Tuberculosis (ACET). MMWR
Summary
Recomm Rep 1999;48(RR-9):1 – 13.
[10] American Thoracic Society (ATS) and the Centers for
During the recent decade, significant progress has Disease Control and Prevention (CDC). Targeted tuber-
been made in reinvigorating the almost nonexistent culin testing and treatment of latent tuberculosis infec-
pipeline of novel agents for the treatment of tuber- tion. Am J Respir Crit Care Med 2000;161:S221 – 47.
culosis and in reestablishing the infrastructure for the [11] Tuberculosis Trials Consortium. The Tuberculosis
conduct of clinical trials of new tuberculosis drugs Trials Consortium: a model for clinical trials collabo-
and treatment regimens. Recent studies of long-acting rations. Public Health Rep 2001;116(Suppl 1):41 – 9.
rifamycin derivatives and potent fluoroquinolone an- [12] Medaglini D, Hoeveler A. The European research
tibiotics are leading to improved regimens for the effort for HIV/AIDS, malaria and tuberculosis. Vaccine
2003;21(Suppl 2):S116 – 20.
treatment of active and latent tuberculosis. A number
[13] Global Alliance for TB Drug Development. Available
of other compounds in late preclinical and early
at: http://www.tballiance.org.
clinical development show great promise. The rapid [14] Hong Kong Chest Service/British Medical Research
increase in knowledge of mycobacterial pathogenesis Council. Controlled trial of 2, 4, and 6 months of pyr-
is leading to the identification of new drug targets, azinamide in 6-month, three-times-weekly regimens
including those believed to play a role in latent in- for smear-positive pulmonary tuberculosis, including
fection or in the phenomenon of persistence. A major an assessment of a combined preparation of isoniazid,
challenge will be to sustain and increase funding for rifampin, and pyrazinamide. Results at 30 months. Am
continued developmental and clinical work if the Rev Respir Dis 1991;143:700 – 6.
promise of tuberculosis elimination, or at least sig- [15] O’Brien RJ, Lyle MA, Snider Jr DE. Rifabutin
(ansamycin LM 427): a new rifamycin-S derivative
nificant lessening of the global tuberculosis epi-
for the treatment of mycobacterial diseases. Rev Infect
demic, is to be achieved.
Dis 1987;9:519 – 30.
[16] Nightingale SD, Cameron DW, Gordin FM, et al. Two
controlled trials of rifabutin prophylaxis against
References Mycobacterium avium complex infection in AIDS.
N Engl J Med 1993;329:828 – 33.
[1] O’Brien RJ, Nunn PP. The need for new drugs against [17] Centers for Disease Control and Prevention. Updated
tuberculosis. Obstacles, opportunities, and next steps. guidelines for the use of rifabutin or rifampin for the
Am J Respir Crit Care Med 2001;163:1055 – 8. treatment and prevention of tuberculosis among HIV-
[2] Corbett EL, Watt CJ, Walker N, et al. The growing infected patients taking protease inhibitors or non-
burden of tuberculosis: global trends and interactions nucleoside reverse transcriptase inhibitors. Centers for
new drugs for tuberculosis 339
Disease Control and Prevention. MMWR Morb Mortal with once-weekly isoniazid and rifapentine. Am J
Wkly Rep 2000;49:185 – 9. Respir Crit Care Med 2003;167:1341 – 7.
[18] Centers for Disease Control and Prevention. Notice [32] Bock NN, Sterling TR, Hamilton CD, et al. A
to readers: acquired rifamycin resistance in persons prospective, randomized, double-blind study of the
with advanced HIV disease being treated for ac- tolerability of rifapentine 600, 900, and 1,200 mg plus
tive tuberculosis with intermittent rifamycin-based isoniazid in the continuation phase of tuberculosis
regimens. MMWR Morb Mort Wkly Rep 2002;51: treatment. Am J Respir Crit Care Med 2002;165:
214 – 5. 1526 – 30.
[19] Klemens SP, Cynamon MH. Activity of KRM-1648 [33] Bock NN, Sterling TR, Khan A, et al. Extension
in combination with isoniazid against Mycobacterium of continuation phase therapy to reduce relapse
tuberculosis in a murine model. Antimicrob Agents rates among HIV-negative TB patients at high risk
Chemother 1996;40:298 – 301. for relapse. Am J Respir Crit Care Med 2003;
[20] Mae T, Hosoe K, Yamamoto T, et al. Effect of a new 167(Suppl):A433.
rifamycin derivative, rifalazil, on liver microsomal en- [34] Chapuis L, Ji B, Truffot-Pernot C, et al. Preventive
zyme induction in rat and dog. Xenobiotica 1998;28: therapy of tuberculosis with rifapentine in immuno-
759 – 66. competent and nude mice. Am J Respir Crit Care Med
[21] Rose L, Vasiljev-K M, Adams P, et al. Safety and 1994;150:1355 – 62.
pharmacokinetics of PA-1648, a new rifamycins in [35] Tuberculosis Research Centre. Shortening short course
normal volunteers. Am J Respir Crit Care Med 1999; chemotherapy: a randomized clinical trial for treatment
159(Suppl):A495. of smear positive pulmonary tuberculosis with regi-
[22] Dietze R, Teixeira L, Rocha LM, et al. Safety and mens using ofloxacin in the intensive phase. Ind J Tub
bactericidal activity of rifalazil in patients with pul- 2002;49:27 – 38.
monary tuberculosis. Antimicrob Agents Chemother [36] Mitchison DA. Assessment of new sterilizing drugs for
2001;45:1972 – 6. treating pulmonary tuberculosis by culture at 2 months.
[23] Rothstein DM, Hartman AD, Cynamon MH, et al. Am Rev Respir Dis 1993;147:1062 – 3.
Development potential of rifalazil. Expert Opin Inves- [37] Alvirez-Freites EJ, Carter JL, Cynamon MH. In vitro
tig Drugs 2003;12:255 – 71. and in vivo activities of gatifloxacin against Myco-
[24] Daniel N, Lounis N, Ji B, et al. Antituberculosis bacterium tuberculosis. Antimicrob Agents Chemother
activity of once-weekly rifapentine-containing regi- 2002;46:1022 – 5.
mens in mice. Long-term effectiveness with 6- and [38] Rodriguez JC, Ruiz M, Climent A, et al. In vitro
8-month treatment regimens. Am J Respir Crit Care activity of four fluoroquinolones against Mycobacte-
Med 2000;161:1572 – 7. rium tuberculosis. Int J Antimicrob Agents 2001;17:
[25] Tuberculosis Trials Consortium. Rifapentine and iso- 229 – 31.
niazid once a week versus rifampicin and isoniazid [39] Ji B, Lounis N, Maslo C, et al. In vitro and in vivo
twice a week for treatment of drug-susceptible pulmo- activities of moxifloxacin and clinafloxacin against
nary tuberculosis in HIV-negative patients: a random- Mycobacterium tuberculosis. Antimicrob Agents Che-
ised clinical trial. Lancet 2002;360:528 – 34. mother 1998;42:2066 – 9.
[26] Vernon A, Burman W, Benator D, et al. Acquired [40] Miyazaki E, Miyazaki M, Chen JM, et al. Moxiflox-
rifamycin monoresistance in patients with HIV-related acin (BAY12 – 8039), a new 8-methoxyquinolone, is
tuberculosis treated with once-weekly rifapentine and active in a mouse model of tuberculosis. Antimicrob
isoniazid. Lancet 1999;353:1843 – 7. Agents Chemother 1999;43:85 – 9.
[27] Tam CM, Chan SL, Lam CW, et al. Rifapentine and [41] Hu Y, Coates AR, Mitchison DA. Sterilizing activities
isoniazid in the continuation phase of treating pulmo- of fluoroquinolones against rifampin-tolerant popula-
nary tuberculosis. Initial report. Am J Respir Crit Care tions of Mycobacterium tuberculosis. Antimicrob
Med 1998;157:1726 – 33. Agents Chemother 2003;47:653 – 7.
[28] US Food and Drug Administration. Review of new [42] Wright DH, Brown GH, Peterson ML, et al. Applica-
drug application from Aventis for Priftin. Available tion of fluoroquinolone pharmacodynamics. J Anti-
at: http://www.fda.gov/cder/foi/nda/98/21024.htm. microb Chemother 2000;46:669 – 83.
Accessed March 10, 2005. [43] Lounis N, Bentoucha A, Truffot-Pernot C, et al.
[29] American Thoracic Society/Centers for Disease Con- Effectiveness of once-weekly rifapentine and moxi-
trol and Prevention/Infectious Diseases Society of floxacin regimens against Mycobacterium tuberculo-
America. Treatment of tuberculosis. Am J Respir Crit sis in mice. Antimicrob Agents Chemother 2001;45:
Care Med 2003;167:603 – 62. 3482 – 6.
[30] Taylor Z, Qualls N, Vernon A, et al. A prevention [44] Nuermberger EL, Yoshimatsu T, Tyagi S, et al.
effectiveness study of rifapentine in the continuation Moxifloxacin-containing regimen greatly reduces time
phase of therapy for active pulmonary tuberculosis. to culture conversion in murine tuberculosis. Am J
Am J Respir Crit Care Med 2000;161(Suppl):A524. Respir Crit Care Med 2004;169:334 – 5.
[31] Weiner M, Burman W, Vernon A, et al. Low isoniazid [45] Gosling RD, Uiso LO, Sam NE, et al. The bactericidal
concentrations and outcome of tuberculosis treatment activity of moxifloxacin in patients with pulmonary
340 o’brien & spigelman
tuberculosis. Am J Respir Crit Care Med 2003;168: [55] Gregory WA, Brittelli DR, Wang C-LJ, et al. Anti-
1342 – 5. bacterials, synthesis and structure-activity studies of
[46] Pletz MW, De Roux A, Roth A, et al. Early bactericidal 3-Aryl-2-oxooxazolidinones. 1. The ‘‘B’’ group. J Med
activity of moxifloxacin in treatment of pulmonary Chem 1989;32:1673 – 81.
tuberculosis: a prospective, randomized study. Anti- [56] Birmingham MC, Rayner CR, Meagher AK, et al.
microb Agents Chemother 2004;48:780 – 2. Linezolid for the treatment of multidrug-resistant,
[47] Valerio G, Bracciale P, Manisco V, et al. Long-term gram-positive infections: experience from a com-
tolerance and effectiveness of moxifloxacin therapy for passionate use program. Clin Infect Dis 2003;36c:
tuberculosis: preliminary results. J Chemother 2003; 159 – 68.
15:66 – 70. [57] Eustice DC, Feldman PA, Zajac PA, et al. Mechanism
[48] Global Alliance for TB Drug Development. Scientific of action of DuP 721: inhibition of an early event
blueprint for tuberculosis drug development. Tuber- during initiation of protein synthesis. Antimicrob
culosis (Edinb) 2001;81(Suppl 1):1 – 52. Agents Chemother 1988;32:1218 – 22.
[49] Andries K, Verhasselt P, Guillemont J, et al. A diaryl- [58] Slee AM, Wuonola MA, McRipley RJ, et al. Oxazo-
quinoline drug active on the ATP synthase of Myco- lidinones, a new class of synthetic antibacterial agents:
bacterium tuberculosis. Science 2005;307:223 – 7. in vitro and in vivo activities of DuP 105 and DuP 721.
[50] Stover CK, Warrener P, VanDevanter DR, et al. A Antimicrob Agents Chemother 1987;31:1791 – 7.
small-molecule nitroimidazopyran drug candidate for [59] Cynamon MH, Kelmens SP, Sharpe CA, et al.
the treatment of tuberculosis. Nature 2000;405:962 – 6. Activities of several novel oxazolidinones against
[51] Grosset J, Nuermberger E, Yoshimatsu T, et al. The Mycobacterium tuberculosis in a murine model.
nitroimidazopyran PA-824 has promising activity in Antimicrob Agents Chemother 1999;43:1189 – 91.
the mouse model of TB. Am J Respir Crit Care Med [60] Dworkin F, Winters SS, Munsiff C, et al. Use of
2004;169(Suppl):A260. linezolid in treating multidrug-resistant tuberculosis
[52] Diedda D, Lampis G, Fioravanti R, et al. Bactericidal in New York City. Am J Respir Crit Care Med 2004;
activities of the pyrrole derivative BM212 against 169(Suppl):A233.
multidrug-resistant and intramacrophagic Mycobacte- [61] Bressler AM, Zimmer SM, Gilmore JL, et al.
rium tuberculosis strains. Antimicrob Agents Chemo- Peripheral neuropathy associated with prolonged use
ther 1998;42:3035 – 7. of linezolid. Lancet Infect Dis 2004;4:528 – 31.
[53] Rastogi N, Goh KS, Berchel M, et al. In vitro activities [62] Lee R, Protopopova M, Crooks E, et al. Combinatorial
of the ketolides telithromycin (HMR 3647) and HMR lead optimization of [1,2]-diamines based on etham-
3004 compared to those of clarithromycin against butol as potential antituberculosis preclinical candi-
slowly growing mycobacteria at pHs 6.8 and 7.4. dates. J Comb Chem 2003;5:172 – 87.
Antimicrob Agents Chemother 2000;44:2848 – 52. [63] Boshoff HI, Myers TG, Copp BR, et al. The transcrip-
[54] Gregory WA, Brittelli DR, Wang C-LU, et al. Anti- tional responses of, M. tuberculosis to inhibitors of
bacterials, synthesis and structure—activity studies of metabolism: novel insights into drug mechanisms of
3-Aryl-2-oxooxazolidines. 2. The ‘‘A’’ Group. J Med action. J Biol Chem 2004;279:40174 – 84.
Chem 1990;33:2569 – 78.
Clin Chest Med 26 (2005) 247 – 271
Tuberculosis is transmitted from person to per- and specificity. Newer tests for LTBI offer the prom-
son by respiratory droplets. Although some people ise of greatly improved diagnostic accuracy.
develop active tuberculosis disease after infection, Tools for the diagnosis of active disease include
almost all tuberculosis infections are asymptomatic clinical suspicion, response to treatment, chest radio-
and remain latent. Latent tuberculosis infection graphs, staining for acid-fast bacilli (AFB), culture
(LTBI) itself progresses to active disease in approx- for mycobacteria, and, more recently, nucleic acid
imately 5% to 10% of infected persons. The rate of amplification (NAA) assays. AFB smears lack both
progression is much greater in immunocompromised sensitivity and specificity, and culture is very slow to
individuals. The estimated 2 billion people living with produce results, limiting the ability to diagnose active
LTBI represent a vast reservoir of potential cases of disease effectively. NAA assays and several other
tuberculosis around the world. This reservoir of LTBI experimental diagnostic tools can add significantly
is therefore a major barrier to the ultimate control and to the active disease diagnostic armamentarium. The
elimination of tuberculosis. suitability of newer diagnostic tests in a given popu-
Strategies to combat tuberculosis in regions that lation varies according to the resources available to
are resource-rich aim, first, to identify and treat per- pay for and implement those tests, however [2].
sons who have active disease; second, to find and In resource-poor countries, where options are lim-
treat contacts of cases of active disease who develop ited, current approaches, such as relying almost ex-
LTBI, and, third, to screen high-risk populations and clusively on the sputum smear for the diagnosis of
treat LTBI [1]. Diagnosis and treatment of LTBI are active disease, leave a significant number of cases
crucial in this effort. In most of the world, however, undetected [3 – 6]. This approach may be the only
resources are devoted exclusively to the highest pri- economically feasible strategy given the initial costs
orities of tuberculosis control: identification and treat- involved in the widespread use of other diagnostic
ment of active disease [1]; for lack of resources, LTBI modalities. Although smear-positive cases are the
is neither diagnosed nor treated. most infectious, neglecting smear-negative disease
Diagnostic testing for both LTBI and active dis- (approximately half of cases overall) increases the
ease has changed little during the last century. Be- morbidity and mortality of the disease in those pa-
cause of limitations in available tests, there has long tients and does not account for the significant burden
been a clear need for better diagnostic tests. LTBI, of transmission attributable to these smear-negative
until very recently, has been diagnosed exclusively by cases (17% of all transmission in one study using
the tuberculin skin test (TST). The TST is fraught molecular epidemiology techniques) [6]. The in-
with problems including relatively poor sensitivity creased likelihood of smear-negative tuberculosis in
HIV patients, particularly those who have advanced
immunosuppression [7], makes this diagnostic
approach especially problematic, because the regions
* Corresponding author. most afflicted by tuberculosis are similarly inundated
E-mail address: ns311@columbia.edu (N.W. Schluger). with HIV infection.
0272-5231/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccm.2005.02.012 chestmed.theclinics.com
248 brodie & schluger
5 TU PPD-S
100
Tuberculosis patients
Per cent reacting
80
60
Healthy children
40
20
0
10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1 100
Dose of tuberculin (mg)
Fig. 1. Cumulative frequency of reactors responding to increasing doses of tuberculin among healthy children and patients
with tuberculosis. TU, tuberculin units. (From Reider H. The epidemiologic basis of tuberculosis control. Paris: International
Union Against Tuberculosis and Lung Disease; 1999. p. 32; with permission.)
the diagnosis of tuberculosis 249
Fig. 2. Frequency distribution of skin test reaction with 5-TU PPD-S (solid line) and 2-TU PPD RT 23 (dotted line) among
Eskimo children and United States Navy recruits. TU, tuberculin units. (From Reider H. The epidemiologic basis of tuberculo-
sis control. Paris: International Union Against Tuberculosis and Lung Disease; 1999. p. 33; with permission.)
tuberculosis disease, those at extremely low like- using standardized preparations of PPD made by the
lihood of having latent infection, and those likely Statenseruminstitut in Amsterdam and testing them in
to be close contacts of persons who have active groups of Eskimo children (a group that has a high
tuberculosis. Reider [10] has described these trials likelihood of latent infection acquired from close
in detail. contact with active cases and very little exposure to
The dosing of tuberculin for use in skin testing environmental mycobacteria) and US Navy recruits
was determined in studies such as one done in Ohio, who have little chance of contact with active tuber-
in which skin testing was performed on tuberculosis culosis but have frequent exposure to environmental
patients and a group of children in orphanages who mycobacteria (Fig. 2). Testing of 5440 tuberculosis
had little chance of tuberculosis exposure. At a dose patients revealed a normal distribution of extent of
of 10 4 mg of tuberculin, a clear distinction could be induration, with a mean of 16 to 17 mm (Fig. 3).
made between the two groups (Fig. 1). Refinements Finally, a massive survey of more than 700,000 US
in dosing and criteria for positivity were achieved by military recruits, of whom 400,000 had no known
Fig. 3. Frequency distribution of tuberculin skin test results (5 – tuberculin unit PPD S) among 5544 tuberculosis patients in
the United States. (From Reider H. The epidemiologic basis of tuberculosis control. Paris: International Union Against
Tuberculosis and Lung Disease; 1999. p. 35; with permission.)
250 brodie & schluger
Fig. 4. Frequency distribution of tuberculin skin test results in United States Navy recruits with (dashed line) or without (dotted
line) tuberculosis. The solid line shows the difference between the two groups. (From Reider H. The epidemiologic basis of
tuberculosis control. Paris: International Union Against Tuberculosis and Lung Disease; 1999. p. 36; with permission.)
contact with tuberculosis patients and 10,000 had risk for developing tuberculosis. Anergy caused by
definite contacts, provided meaningful data on which an immunocompromised state (especially with HIV
to base recommendations for interpretation of skin infection or medication-induced immunosuppression)
tests that are still useful today (Fig. 4). may lead to false-negative results [24]. False-negative
Although the TST has been in widespread use for results also may occur up to about 10 weeks after
a century and is the only universally accepted test for infection with M. tuberculosis [25 – 27]. False nega-
the diagnosis of LTBI, it suffers from significant tives, particularly in the HIV population where the
inherent limitations. To understand these limitations, implications of active disease are most pressing
it is useful to review the mechanism of the TST. [3,28,29], greatly limit the utility of this test.
Infection with M. tuberculosis results in a cell- The exact sensitivity and specificity of the TST
mediated immune response giving rise to sensitized for LTBI is impossible to know with certainty, given
T lymphocytes (both CD4+ and CD8+ [11]) targeted the lack of a reference standard for diagnosis. In that
to M. tuberculosis antigens. Stimulation by M. tu- context, estimates of the global burden of LTBI are
berculosis antigens causes these T cells to release especially problematic. Estimates indicate that the
interferon-gamma (IFN-g). The TST functions by problem is enormous, but these estimates are based
eliciting this response in previously sensitized indi- on the performance of the TST, and such estimates,
viduals. In such individuals, an intradermal injection particularly in developing countries, are notoriously
of PPD evokes a delayed-type hypersensitivity re- unreliable [30,31]. Studies of the prevalence of LTBI
sponse mediated by sensitized T cells and results in in India, for instance, have yielded prevalence rates
cutaneous induration. PPD, however, is a precipitate ranging from 9% [32] to more than 80% in various
of M. tuberculosis culture supernatant which contains populations [33]. A more accurate epidemiologic tool
roughly 200 antigens, many of which are shared by would greatly facilitate a better estimation of the true
other mycobacteria including many NTM and scope of the problem.
M. bovis BCG [12,13]. A response to PPD may sig- The TST is limited further by the subjectivity of
nify infection with M. tuberculosis or, just as readily, its interpretation [34], in particular, by problems with
infection with NTM [14 – 18] or vaccination with interreader and intrareader reliability [25,35 – 37] and
BCG [18 – 22]. This cross-reactivity seriously limits digit preference [38,39]. Also, the existence of the
the specificity of the TST in many populations [23]. booster phenomenon [24,25,39 – 41], poor standardi-
Given that one quarter to one half of the burden zation of PPD preparations [31], and, logistically, the
of tuberculosis in developed countries is found in need for a return visit to have the test read make the
foreign-born immigrants from high-prevalence coun- TST a highly imperfect diagnostic tool. That it does
tries, and this population is made up precisely of not distinguish between LTBI and active disease also
those who are likely to be BCG-vaccinated and to limits its usefulness.
have been exposed to NTM, the TST is least reliable Whether the extent of induration resulting from
in those most in need of screening. Specificity is a tuberculin skin testing can predict the development of
major shortcoming of the TST. In addition, sensitiv- tuberculosis in a linear (or at least dependent fashion)
ity of the TST may also be poorest in patients at high has also been the subject of considerable discussion
the diagnosis of tuberculosis 251
and investigation. Recently, Horsburgh [42] has pro- contact, the risk of LTBI may be expressed as an
vided a well-reasoned and -supported data set that increasing likelihood of infection with increasing ex-
gives guidance in this area. posure to the source case or increasing amounts of
Alternatives to the TST are lacking. Serologic high-risk behavior. It is against this standard that the
tests for the diagnosis of M. tuberculosis have been sensitivity of any new diagnostic test must be
disappointing [43,44]. Although an antibody re- compared with the TST itself. A test that is superior
sponse to M. tuberculosis antigens occurs, there is to the TST in sensitivity would be more likely to be
great individual variability in the number and type positive given a greater degree of contact with the
of serologically reactive antibodies [44], making this source case. Specificity must similarly be gleaned
diagnostic tool too unreliable. Because no serologic from the expectation that only infection with
tests for tuberculosis are remotely good enough to be M. tuberculosis and not with NTM or M. bovis BCG
used clinically at present, they are not discussed fur- would give a positive result. A test would be superior
ther in this article. in specificity if its results seemed to be independent
Despite its many limitations, the TST by neces- of NTM exposure and BCG vaccination status.
sity remains in widespread use. In 2000, the Centers Recently, a new generation of tests for LTBI has
for Disease Control and Prevention (CDC), the been developed. They are the QuantiFERON-TB and
American Thoracic Society (ATS), and the Infectious QuantiFERON-TB Gold (QFN-Gold) tests (Cellestis
Disease Society of America (IDSA) issued updated Limited, St. Kilda, Australia) and the T SPOT-TB test
guidelines for the use of the TST in screening for (Oxford Immunotec, Oxford, UK) The basis of these
LTBI [24]. These guidelines stress that in general one tests is the detection in serum of either the release of
should not place a TST unless treatment would be IFN-g on stimulation of sensitized T cells by
offered in the event of a positive test. In addition, M. tuberculosis antigens in vitro (QuantiFERON) or
cut-off points of induration (5, 10, or 15 mm) for detection of the T cells themselves (T SPOT-TB). In
determining a positive test vary by the pretest risk 1990, Wood and colleagues [45] developed a whole-
category into which a patient falls. This approach blood assay for the detection of IFN-g in response to
may further decrease the specificity of the test, but a specific antigen, PPD, intended for diagnosing
it increases the sensitivity for capturing those at bovine tuberculosis. Later that year, Rothel and col-
highest risk for developing active disease in the short leagues [46] introduced a sandwich enzyme immuno-
term. Continued focus on this century-old test high- assay for bovine IFN-g that streamlined the assay,
lights its continued importance, but the need for a making it more practical for widespread testing. This
more accurate diagnostic tool is evident. assay was shown to be both sensitive and specific in
field comparisons with the intradermal tuberculin test
[46 – 48] and later was accredited in Australia for use
Beyond the tuberculin skin test in the diagnosis of bovine tuberculosis [49]. In 1995,
the test was successfully used in the diagnosis of
Development of novel diagnostic tests for LTBI M. tuberculosis and M. avium complex infection in
is hampered by the lack of a true reference standard humans [50,51].
for diagnosis. Without such a standard, the best ap- Similar to the TST, the QuantiFERON-TB test
proach might be to apply a new test to a population in detects cell-mediated immunity to tuberculin. In con-
a controlled study, observe all patients positive by the trast to intradermal injection of PPD, however, whole
novel and reference tests, and determine which test blood is incubated overnight with PPD from
more accurately predicts the development of active M. tuberculosis, and the IFN-g that is released from
disease. This approach, however, is limited by ethical sensitized lymphocytes is subsequently quantified by
and practical considerations. ELISA [38]. As discussed previously, PPD antigens
Demonstrating that any test is better than the TST are shared across mycobacterial species, including
is therefore difficult. Studies of the magnitude of M. bovis BCG [13,52]. A positive response to the
those cited by Reider [10] for the development of whole-blood IFN-g assay for PPD therefore, like the
tuberculin skin testing are unlikely to be repeated. TST itself, lacks specificity for M. tuberculosis infec-
What approach, then, might be taken? It is well do- tion and may reflect infection with NTM or vacci-
cumented that the greater the proximity to the nation with BCG [18]. Early studies were nonetheless
source case of active disease and the greater the encouraging [38,49,53 – 56], demonstrating decreased
duration of exposure to that case, the more likely it is false-positive results relative to the TST in BCG-
that a person will develop LTBI. Although the risk of vaccinated individuals [38] and those exposed to
LTBI cannot be precisely quantified for all degrees of NTM [38], with equal or better apparent sensitivity
252 brodie & schluger
and specificity than the TST in multiple studies that identifies ESAT-6 – or CFP-10 – specific IFN-g –
[38,49,53,56], including in populations of intrave- secreting CD4+ T cells.
nous drug users and HIV-positive patients [54,55]. Studies of the IFN-g release assays employing
Few studies purported to demonstrate the superiority ESAT-6 or CFP10 in humans have been promising
of the TST for LTBI [57,58]. [79 – 88]. ESAT-6 is a major target of the cellular
The discovery of M. tuberculosis – specific anti- immune response in humans [62,63]. Early on, assays
gens opened the way to improving the specificity of employing ESAT-6 were shown to be more specific
the assay. In 1986, Harboe and colleagues [59] re- although less sensitive than PPD-based assays for
ported the first M. tuberculosis – specific antigen, active disease [79,82,89 – 91]. This improved speci-
MPB-64 (later known as MPT-64). In 1995, Ander- ficity over PPD with loss of sensitivity was later
sen and colleagues [60] reported the highly immuno- shown by Arend and colleagues [80,81] to hold true
genic antigen target of the cellular immune response in LTBI as well. More importantly, they also dem-
to tuberculosis in mice, known as the early secreted onstrated improved specificity over the TST. They
antigenic target 6 (ESAT-6). Subsequently, in 1998, reported loss of sensitivity with respect to the TST,
Berthet and colleagues [61] described culture filtrate but, because the TST was used as the reference
protein (CFP-10) [61], another highly immuno- standard for LTBI, this result probably reflected the
genic antigen. MPT-64 has been studied extensively improved specificity rather than poorer sensitivity
[62 – 66], but, because it is present in some strains of [80]. The improved specificity over the Quanti-
BCG and is a less potent target of the immune re- FERON-TB assay for PPD and over the TST was
sponse, it has limited utility [62,63,67]. On the other confirmed in a study by Johnson et al [79] of 60 Aus-
hand, ESAT-6 [61 – 64,68 – 70] and CFP-10 [61,71] tralian medical students who did not have BCG
have demonstrated great potential. vaccination or known exposure to M. tuberculosis or
In 1998, the complete genome sequence of NTM. The specificity of both the QuantiFERON-TB
M. tuberculosis was determined [72]. An earlier com- for PPD and the TST were reduced after BCG vac-
parison of the M. tuberculosis genome with the ge- cination was administered to the students, but the
nomic composition of M. bovis and M. bovis BCG in QuantiFERON-TB for ESAT-6 was unaffected [79].
1996 by subtractive genomic hybridization [68] and, Brock and colleagues [85] also demonstrated im-
subsequently, in 1999, by comparative hybridiza- proved specificity for both the ESAT-6 and CFP-10
tion experiments on a DNA microarray [73] led to over PPD in BCG-vaccinated versus nonvaccinated
the identification of a genomic region known as subjects. Recently, Brock and colleagues [88] pub-
RD1. The gene products of RD1 are found only in lished an outbreak study based on a case of active
M. tuberculosis, in pathogenic M. bovis strains [64, disease in a Danish high school student and the
68,73], and in four NTM (M. kansasii, M. szulgai, student’s mostly non – BCG-vaccinated contacts. The
M. flavescens, and M. marinum) [69,70]. Because, of TST was used as the reference standard for LTBI in
these, only M. kansasii overlaps clinically with this population, and there was excellent agreement
M. tuberculosis, and because M. kansasii infection between the TST and the QuantiFERON-TB ESAT-6/
is uncommon, the RD1 region encodes antigens that CFP-10 assay (94%) [88]. Superiority of the assay
are essentially specific to M. tuberculosis. Among could not be established in this study, because the
these antigens are, of course, ESAT-6 and CFP10, as TST was itself the reference standard. Nonetheless,
well as MPT-64. ESAT-6 and CFP-10 are secreted by significant specificity with regard to BCG status was
M. tuberculosis into the extracellular environment suggested by the findings in subjects who were BCG-
and are potent targets of the cell-mediated immune vaccinated. Of these, 50% in the high-exposure
response [61 – 63,71,74]. ESAT-6, which has been group had a positive assay, compared with 53% of
shown to be highly immunogenic in animals [66, high-exposure subjects who did not have BCG vacci-
75 – 78], readily discriminated between bovine tuber- nation. In the low-exposure group, 5% of BCG-
culosis and cattle sensitized to environmental myco- vaccinated persons were assay positive, similar to the
bacteria [75]. CFP-10 also has demonstrated utility in 6% of those who did not have BCG vaccination [88].
diagnosing bovine tuberculosis because of its sig- Finally, Mori and colleagues [87] studied a group of
nificant specificity [78]. 216 Japanese student nurses who had no identified
The melding of the ELISA-based QuantiFERON- risk for M. tuberculosis exposure, all of whom had
TB test and the RD1 antigens, ESAT-6 and CFP-10, been vaccinated with BCG [87]. In this group 64.6%
led to a more specific test, now known as the QFN- of the subjects had a TST response measuring 10 mm
Gold. Similarly, the T SPOT-TB test employs the or more, yielding a specificity of 35.4% for the TST if
RD1 antigens but links them to an ELISPOT assay it is assumed that none had true LTBI using exposure
the diagnosis of tuberculosis 253
history as the reference standard. The QuantiFERON- Five hundred thirty-five representative students
TB ESAT-6/CFP-10 assay, on the other hand, yielded were enrolled in the study and underwent ELISPOT
a specificity of 98.1% in this group, far superior to testing. The significance of the study derives from
the TST. The sensitivity of the assay, 89.0%, was the detailed contact information available for each
determined in a separate group of patients who had student by virtue of their mandatory, scheduled daily
culture-proven active disease. Extrapolating the sen- activities. Degree of exposure to the source case
sitivity for LTBI from the sensitivity for active dis- could be readily quantified and grouped as (1) same
ease is problematic, however, because both IFN-g class, same year, with regularly shared lessons;
activity and TST reactivity are reduced in active (2) same year with only weekly shared events; and
disease; extrapolation probably underestimates the (3) other years. Using ORs, the authors provide an es-
sensitivity for LTBI [12,92,93]. timate of the increase in odds of a positive ELISPOT
Ajit Lalvani and colleagues [94] have adapted the for each increase in level of exposure. The ELISPOT
ELISPOT technique, an ex vivo T-cell – based assay correlated significantly better than the TST with
for the detection of cell-mediated immunity, for use in increasing exposure across each group. The relative
detecting M. tuberculosis. The technique detects and risk (RR) of direct exposure to the index case if one
enumerates peripheral blood IFN-g – secreting T cells was both TST and ELISPOT positive was 17.6. If one
that respond to ESAT-6 or CFP-10. In 2001, Lalvani was ELISPOT positive but TST negative, it was 11.7.
and colleagues [94] established a sensitivity of 96% If one was ELISPOT negative and TST positive, the
in active disease (100% in the subpopulation that had RR was only 2.97. Also, TST positivity was sig-
extrapulmonary tuberculosis) as compared with 69% nificantly associated with BCG vaccination status
sensitivity for the TST. In healthy BCG-vaccinated and with birth in a region of high prevalence for
controls, the ELISPOT was not confounded by BCG. NTM, whereas no significant association was found
In TST-positive household contacts of a case of ac- for the ELISPOT. The superior correlation with de-
tive disease (expected cases of LTBI), 85% were gree of exposure strongly suggests improved sensi-
ELISPOT positive, suggesting a sensitivity for LTBI tivity with the ELISPOT. The lack of confounding
of 85% if the TST is taken to be the reference by BCG or NTM suggests improved specificity with
standard [94]. the ELISPOT.
In 2001, Pathan and colleagues [92] also exam- The role of IFN-g – or T-cell – based assays in the
ined a low-exposure population of mostly BCG- diagnosis of LTBI is being defined. These tests show
vaccinated subjects. None of the 32 healthy controls promise as replacements for the TST in diagnosing
were positive on ELISPOT. That year, Lalvani and LTBI among persons at risk for infection in the
colleagues [95] also reported on an outbreak study developed world. Both the QFN-Gold and T SPOT-
[95]. The odds ratio (OR) of a positive ELISPOT with TB tests are approved for diagnostic use throughout
increasing proximity and duration of exposure to the the European Union. In addition, the QFN-GOLD
index case was 9.0, whereas the OR of a positive TST was approved by the United States Food and Drug
(by Heaf test, a less well-standardized approach Administration (FDA) in December 2004. Clinical
to skin testing than the tuberculin test) was only experience with these tests should accumulate rap-
1.9. Another study published in 2001 by Lalvani idly. The less accurate original QuantiFERON-TB
and colleagues [12] looked at 40 healthy controls in is approved for use in the United States by the FDA,
the United Kingdom, 82% of whom were BCG- but guidelines for its use are confusing, and the test
vaccinated and all of whom were ELISPOT negative. has not been widely adopted in clinical settings.
Another study in the United Kingdom similarly found The improved sensitivity of these tests over the
that none of 40 healthy controls were ELISPOT TST would capture a cohort of patients who other-
positive [96]. wise would go without treatment of LTBI. Within that
In 2003, Ewer and colleagues [97] published a cohort, those that would have progressed to active
meticulously investigated outbreak study that evalu- disease would be spared the attendant morbidity and
ated the ESAT-6/CFP-10 – based ELISPOT assay for mortality. The future contacts of those destined to
LTBI. Two years earlier, in the United Kingdom, a progress to active disease would likewise be spared.
secondary school student had been diagnosed with This cohort would be overrepresented by those most
sputum smear – positive cavitary pulmonary tuber- likely to have false-negative TST results, namely im-
culosis. The health authority screened 1128 students munosuppressed individuals. This population is
at the school with the TST (HEAF test). Screening precisely the one in which it is most important to
detected 69 cases of active disease and 254 cases of identify LTBI because of their increased risk for
LTBI, 87% of whom had been vaccinated with BCG. developing active disease [3]. Unanswered is whether
254 brodie & schluger
there is something different about this cohort that improved sensitivity. Operator bias and inter- and
gives positive results on these assays and negative intrareader variability are significantly reduced.
results on the TST and whether their risk for de- Only a single patient visit is required. There is no
veloping active disease is less than that of those who booster effect. The results are obtained rapidly, within
are also TST positive. The TST has predictive value 24 hours. There may well be cost savings to the
for the subsequent development of active disease in health care system. The assays have proven to be
both HIV-negative [9,25] and HIV-positive patients robust in different populations and in different
[98]: a stronger skin test response indicates an in- settings, in both the developed and developing
creased risk of developing active disease [99,100]. worlds. The technology for running the assays has
Longitudinal studies linking positive assays with risk improved so that requirements for equipment and
for development of active disease are ongoing and technical expertise are reasonable.
are crucial to demonstrating the true role of these Further study is needed. Longitudinal data, as
tests. If they demonstrate a high degree of accuracy, mentioned previously, are critical. To date there are
treatment of LTBI might, under the right conditions, no large-scale trials of these assays, and few studies
become a viable strategic component of tuberculosis have employed the Mantoux test while using current
control efforts in high- and low-prevalence countries ATS/CDC/IDSA criteria for a positive TST. Study of
[97]. One small study of 24 healthy household con- the effect of treatment on assay results in LTBI may
tacts of persons who had smear-positive pulmonary provide the data necessary to monitor therapy for
tuberculosis in Ethiopia looked at QuantiFERON-TB LTBI, allowing the clinician, for example, to distin-
ESAT-6 responses at the initial visit and approxi- guish response to therapy from lack of response
mately 2 years later. The subjects were not treated for caused by noncompliance or isoniazid resistance.
LTBI. Seven of 24 patients went on to develop pul- Another area of interest would be the ability to
monary active disease. The subjects who responded distinguish past exposure to M. tuberculosis without
to ESAT-6 at study entry were significantly more ongoing infection from true ongoing LTBI. It may
likely to develop active disease than those who were be that detection of T cells specific for ESAT-6 or
not responsive [101]. CFP-10 suggests that tubercle bacilli continue to
The improved specificity would decrease unnec- secrete these antigens [12]. Identification of an anti-
essary treatment in those who are not truly infected, gen that is expressed either during LTBI or during
thereby avoiding the costs to the health care system active disease, but not during both, and that could
for medication, follow-up, and management of com- distinguish these two states would greatly enhance
plications. It would also spare the individual patient the role of the assays in active disease and would
these same costs. increase the specificity for LTBI. Finally, the point
Costs would also likely be reduced by the in- after infection at which the RD1 antigens become
creased capture of cases of LTBI, because their detectable has yet to be defined precisely and has
identification would eliminate the future—much significant implications for testing in contact inves-
greater—costs of treating an outbreak of active tigations and for reducing false-negative results soon
disease. Costs also might be reduced by the decreased after infection with M. tuberculosis.
number of clinic visits required, because the TST
requires a follow-up visit for reading the TST and
may require a second skin test to overcome the Tests for tuberculosis disease
booster phenomenon.
At present, LTBI is neither diagnosed nor treated The reference standard for diagnosing active dis-
in most high-burden, resource-poor countries, except ease remains largely clinical: documented response
in certain situations, such as young children who are to appropriate therapy. Of course, establishing a mi-
close contacts of active cases. The expense of a novel crobiologic diagnosis is preferable. AFB smear,
diagnostic test must be justified in terms of the cost mycobacterial culture, and NAA assays may all be
savings realized from treating LTBI. Such savings used in confirming a diagnosis of active disease (both
could accrue by reducing the number of cases of pulmonary and extrapulmonary). In the case of pul-
active disease that develop from the vast reservoir of monary tuberculosis, the method of obtaining a
LTBI that exists in the developing world. sample greatly affects the sensitivity of testing. Ex-
Overall, the potential advantages of the IFN-g – trapulmonary tuberculosis frequently poses a diag-
release and T-cell – based assays over the TST in nostic challenge because specimens may be difficult
diagnosing LTBI seem to include improved specific- to obtain. After identifying M. tuberculosis, the most
ity (lack of confounding by BCG and NTM) and pressing issue is drug-susceptibility testing, in which
the diagnosis of tuberculosis 255
traditional culture techniques are giving way to more of expectorated sputum ranges from 34% to 80%
advanced technologies that produce rapid results. [3 – 5,104,107 – 116]; the sensitivity tends to be high-
est in patients who have cavitary disease and lowest
in patients who have weak cough or less advanced
disease. In no way does a negative sputum smear
Mode of diagnosis of pulmonary tuberculosis eliminate the diagnosis of active tuberculosis, particu-
larly if the clinical suspicion is high. Instituting ther-
Is there a role for the tuberculin skin test in the apy in such cases often is warranted while awaiting
diagnosis of active disease? culture results. If a patient is suspected of having
pulmonary tuberculosis but is smear negative on ex-
The TST was originally a test for active disease pectorated sputum or is unable to produce sputum for
[9]. It is unsuited for that purpose because its testing (30% of patients in one series [117]), further
specificity is limited by cross reactions with NTM diagnostic testing may be warranted. The options
and M. bovis BCG, and, more importantly, by its include sputum induction (SI), fiberoptic bronchos-
detection of LTBI itself, and by alterations in general copy (FOB), and perhaps gastric washings (GW).
immune responsiveness that may occur in cases of The following discussion refers specifically to patients
active tuberculosis. The sensitivity of the TST for who are expectorated sputum smear negative or who
active disease varies considerably, from 65% to 94% cannot produce an expectorated sputum sample.
[31,34,58,87,89,94]. A study of 3600 hospitalized
patients done by the World Health Organization in
the 1950s found a sensitivity of 93% for reactions of Sputum induction
10 mm or more and a sensitivity of 78% when a
cut-off of 14 mm or more was used [31]. The sen- SI in the diagnosis of active disease was first de-
sitivity is decreased in certain populations (eg, to less scribed in 1961 by Hensler and colleagues [117].
than 50% in critically ill patients who have dissemi- They adapted an earlier technique used to obtain
nated tuberculosis) [25]. Lacking both specificity sputum for cytology in diagnosing lung cancer. Early
and sensitivity for active disease, the TST is not par- studies compared SI with the well-established method
ticularly useful in this setting. of gastric aspiration [117 – 119]. In patients unable to
expectorate or who had smear-negative sputum
samples, SI was superior to GW in obtaining a suit-
Sputum-based diagnosis able sample for culture, although the two techniques
were noted to be complementary [119]. GW probably
To establish a diagnosis of pulmonary tuber- adds to overall diagnosis, and, according to one au-
culosis, respiratory samples are submitted to the thor, its value has been underestimated in recent years
laboratory for microscopy (AFB smear) and for my- [120]. The role of GW in adults is probably quite
cobacterial culture. NAA assays may also be used in limited, however. SI, on the other hand, has proven
the diagnostic algorithm, as discussed later. effective in patients clinically suspected of having
The technique used to obtain the respiratory pulmonary tuberculosis who are either unable to pro-
sample strongly influences the ability to detect pul- duce sputum or are sputum smear negative.
monary tuberculosis. Expectorated sputum is gener- SI has performed well in resource-poor countries
ally the starting point. Three samples are collected on with little added cost [121 – 123]. In South Africa, SI
three separate days and stained for AFB [102,103]. performed on 51 patients yielded a suitable sample in
Although, the utility of collecting three samples has 36 [123]. Fifteen of the 36 patients (42%) were smear
been questioned [104], the overall yield for smear and positive, 12 of whom were ultimately culture positive
culture is superior to collecting fewer specimens as well. In Malawi, Parry and colleagues [122] were
[105,106]. Samples are generally sent simultaneously able to obtain SI specimens in 73 of 82 patients.
for smear and culture, because culture data are es- Eighteen of the 73 (25%) were smear positive, and
sential for confirmation of the diagnosis. In resource- 30 of 73 (42%) were culture positive. Similarly, of
poor countries, the cost of culture is often too great, 1648 patients in China, 558 (34%) were smear posi-
resulting in reliance solely on AFB smears. tive on SI samples. The direct cost per SI in that study
The sensitivity of sputum AFB smears for de- was 37 cents [121]. In these studies, SI provided
tecting pulmonary tuberculosis is limited by the need appropriate samples for diagnosis and increased the
for 5000 to 10,000 bacilli per milliliter to be present early diagnostic yield significantly. SI also seems to
in a specimen to allow detection [3]. The sensitivity be cost-effective in the resource-poor setting.
256 brodie & schluger
Conversely, in a retrospective review of 114 pa- worthwhile in resource-rich settings but may be less
tients who had culture-positive M. tuberculosis in- applicable in resource-poor settings where repeated
fection at an urban hospital in New York, SI added SI alone would diagnose most of the cases at a
little to overall diagnosis and was deemed costly by substantially reduced cost.
the investigators [124]. In 1 year, they performed Anderson [128] prospectively compared SI and
1566 SIs yielding only 16 positive smears in 10 pa- FOB with bronchoalveolar lavage (BAL) in 101 pa-
tients. At a cost of $28.65 per SI, the annual cost of tients who had suspected pulmonary tuberculosis
$45,000 would indeed be difficult to justify [124]. A in Montreal. SI yielded a positive smear in 19% of
study in the United Kingdom confirmed a low yield cases; the yield of FOB smear was 12%. The yield
but suggested there might be a role for SI [125]. was much higher in obtaining culture-positive sam-
Is there a role for SI in resource-rich countries? ples: 87% with SI, as compared with 73% for FOB.
A large, prospective study from Montreal, Canada, Overall, SI performed better than FOB, and direct
assessed 500 patients who were either smear negative costs of FOB were more than eight times those of
(5%) or could not produce sputum (95%) with re- SI [128].
peated SI [126]. An adequate sample was obtained A Brazilian study compared SI with FOB in
in 99.8% of patients. The cumulative yield of SI for HIV-positive and HIV-negative patients [129]. One
smear-positive samples with successive attempts was hundred forty-three patients were diagnosed with pul-
64%, 81%, 91%, and, after four inductions, 98%. The monary tuberculosis, 17% of whom were HIV posi-
culture yield also increased with each attempt from tive. The sensitivity of SI smear was 33.8%, and
70% to 91% to 99% to 100%. This study suggests that of FOB was 38.1% in HIV-negative patients. In
that the use of repeated SI has a high yield in this HIV-positive patients, the sensitivities were similar:
setting and that repeated SI should be considered 36% for SI smear and 40% for FOB smear. SI
seriously in this subset of patients [126]. produced an adequate sample in 97% of patients in
this study [129].
Sputum induction versus fiberoptic bronchoscopy SI performs well in both resource-poor and
resource-rich countries, is useful in HIV-positive
How does SI compare with FOB in the diagnosis and -negative patients and compares favorably with
of pulmonary tuberculosis in expectorated-sputum FOB in diagnostic yield and cost. Some authors argue
smear-negative patients or patients unable to produce that neither SI nor FOB should be performed unless
sputum? A study by McWilliams and colleagues absolutely necessary, given the risk of exposure of
[127] from New Zealand compared repeated SI with health care workers and other patients to the aerosol-
FOB, which was performed if at least two SIs were generating procedures [130]. This warning, however,
smear negative. They prospectively studied 129 pa- applies mostly to environments where proper respi-
tients who underwent both procedures. Each succes- ratory protective equipment and exhaust ventilation
sive SI, up to three in total, increased the yield for devices or appropriate isolation rooms are in short
culture-positive samples significantly. SI was per- supply [130].
formed without difficulty in 96% of patients and had
an overall yield of 96.3% after three tests, confirm-
ing the utility of repeated SIs. By contrast, the yield The role of fiberoptic bronchoscopy
of FOB was only 51.9%, making SI significantly
more sensitive in this population. The authors also FOB encompasses BAL, bronchial washings
noted that the overall cost of FOB was three times (BW), bronchial brushings (BB), transbronchial
that of doing three SIs. They offered several strategies biopsy (TBB), and postbronchoscopy sputum collec-
for diagnosis: FOB alone was too insensitive, whereas tion (PBS). FOB has been studied by several inves-
SI alone was sensitive (missed only one case) and cost tigators (although usually in relatively small studies)
effective. Although the combination of SI and FOB in pulmonary tuberculosis suspects who are smear
would have captured all culture-confirmed cases of negative or unable to produce a sputum sample. The
pulmonary tuberculosis, it would have done so at four utility of FOB (or SI) in this setting is twofold. First,
times the cost. The preferred strategy, according to generating a sample in patients who do not have
the authors, would employ SI followed by FOB only spontaneous sputum creates the potential for making
in patients who were negative on SI but had features a diagnosis. Second, rapid diagnosis (by positive
of pulmonary tuberculosis on chest radiograph. This smear or histopathology) in either subset of patients
strategy missed no cases and was only 2.5 times provides the potential for earlier intervention and
the cost of SI alone [127]. This strategy may be treatment while awaiting culture results.
the diagnosis of tuberculosis 257
In 1988, Chawla and colleagues [131] at the HIV-negative patient. TBB yielded a rapid diagnosis
University of Delhi in India prospectively studied in 16% of HIV-positive and 42% of HIV-negative
50 pulmonary tuberculosis suspects who were smear patients. Overall, TBB provided the exclusive early
negative or unable to produce sputum. Overall, cul- diagnosis in 10% of patients [136].
tures of M. tuberculosis from FOB were positive in Although not all studies report such high yields
90%. More significantly, a rapid diagnosis was made from FOB [137 – 142], it definitely has utility [103].
in fully 72% of cases. Smear-positive samples were The ability to achieve rapid diagnosis—a crucial step
obtained in 28% of PBS specimens, 24% of BW in the management of pulmonary tuberculosis—with
specimens, and 56% of BB specimens. In the case of FOB generally ranges from around 30% to 70%, and
BB specimens, 10 patients (20% of those studied) the overall yield of culture from FOB specimens is
were rapidly diagnosed exclusively by this means. much higher [112,113,131,132,134 – 136,143 – 147].
PBS and BW each provided the exclusive diagnosis Although the yield of the different techniques varied
for 6% of patients. TBB was performed in 30 pa- significantly among studies, each one clearly con-
tients, and histopathology was positive in 9 (3 were tributed to the overall yield of FOB.
exclusively diagnosed on biopsy). The authors com- The most productive use of FOB is in pulmonary
ment that the high yield from the BB smears was a tuberculosis suspects who produce no sputum or
result of brushing caseous material in the bronchi who are smear negative and in patients in whom there
when visible [131]. is considerable diagnostic uncertainty, where lung
In a study from Hong Kong in 1982, So and biopsy may produce an alternative diagnosis. These
colleagues [132] also prospectively examined the benefits must always be weighed against the costs of
capability of FOB for rapid diagnosis. They per- the procedure, concerns regarding infection control,
formed FOB in 65 pulmonary tuberculosis suspects. and the risk of TBB in any given patient.
Overall, rapid diagnosis was achieved in 42 of 65
(65%). TBB gave a rapid diagnosis in 33 of the Cultures
57 patients in whom it was performed (58%) and was
the exclusive means of rapid diagnosis in 12% [132]. Cultures of mycobacteria require only 10 to
Willcox and colleagues [133] conducted a study 100 organisms to detect M. tuberculosis. As a result,
in Cape Town, South Africa, in 1982 that looked at the sensitivity of culture is excellent, ranging from
275 pulmonary tuberculosis suspects. Seventy-nine 80% to 93% [3,107]. Moreover, the specificity is quite
were diagnosed with pulmonary tuberculosis. FOB high, at 98% [3]. Cultures increase the sensitivity for
made the culture diagnosis in 60 of 79 (76%). BB diagnosis of M. tuberculosis and allow speciation,
gave a rapid diagnosis in 33%, and TBB did so in drug-susceptibility testing, and, if needed, genotyping
43%. Similarly, Sarkar and colleagues [134] prospec- for epidemiologic purposes [3]. Therefore, all speci-
tively performed FOB in 30 pulmonary tuberculosis mens should be cultured.
suspects in Rajasthan, India. Rapid diagnosis was There are three types of culture media: solid
made in 22 of 30 persons (73%). media, which includes egg-based media (Lowenstein-
In a retrospective review of 41 patients who had Jensen) and agar-based media (Middlebrook 7H10
culture-proven pulmonary tuberculosis and under- and 7H11), and liquid media (Middlebrook 7H12 and
went FOB, a rapid diagnosis was obtained in 34% of other broths). Solid media, long the standard for
patients [135]. Finally, Mehta and colleagues [112] culturing mycobacteria, are slower than liquid media,
looked retrospectively at 30 patients who had culture- which now are widely used alongside solid media to
positive pulmonary tuberculosis and a negative spu- increase sensitivity and decrease recovery time [148,
tum smear or no sample. FOB (BW and BB) made a 149]. In fact, Lowenstein-Jensen 7H10 and 7H11
rapid diagnosis in 18 of 30 patients (60%). media may detect mycobacteria in less than 4 weeks
The potential utility of BAL and TBB for rapid [148,150,151], but they require incubation for as
diagnosis in HIV-positive and HIV-negative patients long as 6 to 8 weeks before they can be classified as
was demonstrated in a study by Kennedy and col- negative. In contrast, broth media combined with
leagues [136]. They retrospectively reviewed 67 HIV- DNA probes for rapid species identification typically
positive and 45 HIV-negative patients who had provide results in less than 2 weeks with smear-
culture-proven pulmonary tuberculosis. Of those positive samples and somewhat longer with smear-
who had smear-negative sputum, BAL provided a negative samples [148,151,152]. Broth media
rapid diagnosis in 24% of HIV-positive and 8% of formulations include both manual and automated
HIV-negative patients. BAL was the exclusive means systems using radiometric or colorimetric methods for
of diagnosis in seven HIV-positive patients and in one detection of mycobacteria. Examples of broth media
258 brodie & schluger
include the BACTEC 460TB and BACTEC MB9000 Inc., Branchburg, NJ), and the Amplified Mycobate-
radiometric methods, the Mycobacterial Growth rium Tuberculosis Direct (MTD) Test (Gen-Probe,
Indicator Tube or MGIT nonradiometric method, and Inc., San Diego, CA).
the manual Septi-Chek AFB system (all from Becton The AMPLICOR assay uses DNA polymerase
Dickinson Microbiology Systems, Franklin Lakes, chain reaction (PCR) to amplify nucleic acid targets.
NJ), the MB/Bac T (Biomerieux, Durham, NC), Extra The FDA approved its use in smear-positive respira-
Sensing Power (ESP) and Myco-ESPculture System tory specimens in December 1996. The COBAS
II (Trek Diagnostic Systems, Cleveland, OH), and AMPLICOR is an automated version of the AMPLI-
BacT/ALERT MB Susceptibility Kit (Organon COR MTB. The MTD assay is an isothermal strategy
Teknika, Durham, NC). for detection of M. tuberculosis rRNA. The FDA
Broth media also may allow more rapid determi- approved its use for use with smear-positive respira-
nation of drug susceptibilities, particularly if direct tory specimens in December 1995. A reformulated
susceptibility testing is used. Direct susceptibility MTD (AMTDII or E-MTD, for enhanced MTD) was
testing may be done with smear-positive samples that approved by the FDA in May 1998 for smear-positive
are simultaneously inoculated into bottles lacking and specimens and in September 1999 for detection of
containing antibiotics. With this technique, drug M. tuberculosis in both smear-positive and smear-
susceptibilities can be known at the same time as negative respiratory specimens.
culture results. In clinical and laboratory studies, the original
Newer culture technologies are in development. MTD assay ranged in sensitivity from 83% to 98% for
One such product is TK Medium (Salubris, Inc., smear-positive respiratory samples [107,153 – 160]
Cambridge, MA). TK Medium uses multiple-color and from 70% to 81% for smear-negative respiratory
dye indicators to identify M. tuberculosis rapidly. It samples. In a recent study in Zambia (one of rela-
can also be used for drug-susceptibility testing and tively few studies in a resource-poor country), the sen-
can differentiate a contaminated specimen. Informa- sitivity was only 64% [116]. The specificity in these
tion is available at www.salubrisinc.com. studies was 98% to 99%. The AMPLICOR assay
performed similarly. The sensitivity was 74% to 92%
for smear-positive respiratory samples [5,107,109,
Nucleic acid amplification assays 157,161 – 166] and 40% to 73% for smear-negative
samples [5,107,161,164 – 166]. Specificity ranged
NAA assays amplify M. tuberculosis – specific from 93% to 99%. In Switzerland, Laifer and col-
nucleic acid sequences using a nucleic acid probe. leagues [167] recently tested the AMPLICOR assay
NAA assays enable direct detection of M. tuber- in 3119 war refugees from Kosovo and found a sen-
culosis in clinical specimens. Such assays comple- sitivity of only 64% for pulmonary tuberculosis
ment the conventional laboratory approach to the [167]. They noted, however, that the negative pre-
diagnosis of active disease. Whereas AFB smears are dictive value of three consecutive PCRs (in two sputa
rapid but lack sensitivity and specificity, and culture and one BAL) was 100%. In studies where MTD
is both sensitive and specific but may take from 2 to and AMPLICOR have been compared directly,
8 weeks to produce results, NAA assays allow rapid, MTD has consistently had a small advantage [107,
sensitive, and specific detection of M. tuberculosis. 157,159].
The sensitivity of the NAA assays currently in The E-MTD brings with it an improved sensitivity
commercial use is at least 80% in most studies, and [4,108,153,168], especially in smear-negative speci-
these assays require as few as 10 bacilli from a given mens [4,108]. Bergmann and colleagues [4] inves-
sample under research conditions [3]. Although the tigated the E-MTD in a 1999 study of Texas prison
sensitivity of these assays in AFB smear-negative inmates [4]. One thousand four respiratory specimens
samples is lower than for smear-positive samples, from 489 inmates tested with E-MTD were compared
newer assays perform much better in this regard than with culture, smear, and clinical course. Twenty-two
earlier versions, increasing the sensitivity for smear- inmates were diagnosed with pulmonary tuberculosis
negative specimens as well as overall sensitivity (10 smear-positive and 12 smear-negative.) Overall,
[4,108]. NAA assays are also quite specific for the E-MTD had a sensitivity of 95.2% and a
M. tuberculosis, with specificities in the range of specificity of 99.1%. In smear-positive patients, the
98% to 99%. sensitivity and specificity were both 100%. In smear-
At present two FDA-approved NAA assays are negative patients, the sensitivity was 90.2%, and the
widely available for commercial use: the AMPLI- specificity was 99.1% [4]. A 1999 study from the
COR M. tuberculosis (Roche Diagnostic Systems, Central Public Health Laboratory in Etobicoke,
the diagnosis of tuberculosis 259
Ontario, looked at 823 specimens (616 respiratory) pared favorably with AMPLICOR [174,175] and
over a 1-year period [108]. Using clinical diagnosis E-MTD [173]. None of these tests has been approved
as the reference standard, the specificity approxi- for use in the United States.
mated 100%, and the sensitivity for either smear- In 2000, the CDC updated its recommendations
positive or smear-negative respiratory samples was for use of NAA tests for the diagnosis of active
100%, an exceptionally high value, especially for the disease [176]. The CDC now recommends that AFB
smear-negative specimens. Specimens that were smear and NAA be performed on the first sputum
smear negative were preselected for testing with the smear collected. If smear and NAA are both positive,
E-MTD based on a clinical determination that the pulmonary tuberculosis is diagnosed with near
patients were at high risk for tuberculosis. Preselec- certainty. If the smear is positive and the NAA is
tion no doubt contributed to the high sensitivity and negative, the statement recommends testing the
specificity in this study, but results indicate there is sputum for inhibitors by spiking the sputum sample
great utility in selecting appropriate patients for with an aliquot of lysed M. tuberculosis and repeating
testing [108]. the assay. If inhibitors are not detected, the process
An investigation of the E-MTD with particular is repeated on additional specimens. If the sputum
clinical relevance was undertaken by Catanzaro and remains smear positive without inhibitors and NAA
colleagues [169] who evaluated the performance of negative, the patient can be assumed to have NTM.
the E-MTD in a multicenter, prospective trial. In If a sputum sample is smear negative but E-MTD
this study, the E-MTD was evaluated against the positive (only the E-MTD is approved for smear-
backdrop of a patient’s clinical suspicion for pul- negative specimens), the CDC recommends testing
monary tuberculosis, which was stratified into low, additional samples. If further samples are E-MTD
intermediate, or high risk as determined by physi- positive, the patient can be assumed to have pul-
cians who had expertise in evaluating patients for monary tuberculosis. If both the smear and E-MTD
tuberculosis. Clinical investigators determined the are negative, an additional specimen should be tested
risk for 338 patients. The specificity of the E-MTD by E-MTD. If negative, the patient can be assumed
was high in all groups. The sensitivities were 83%, not to have infectious pulmonary tuberculosis. The
75%, and 87% respectively. The positive predictive recommendations conclude by noting that clinicians
value, however, was low in the low-risk group must always rely on clinical judgment and that, ulti-
(59%, as compared with 100% in the other two mately, definitive diagnosis rests on response to
groups). The negative predictive value was espe- therapy and culture results [176]. Although they have
cially high in the low-risk group (99%) and a certain logic, these recommendations are expensive
remained high (91%) in the intermediate- and and based on few published data.
high-risk groups. These results compared favorably Overall, a reasonable use of NAA assays for rapid
with the AFB smear, which had positive predictive diagnosis of pulmonary tuberculosis is as follows:
values of 36% (low), 30% (intermediate), and 94% NAA assays should be used to confirm that a positive
(high), respectively, and negative predictive values AFB smear does indeed represent M. tuberculosis. If
of 96% (low), 71% (intermediate), and 37% (high), both smear and NAA are positive, pulmonary tuber-
respectively. This study demonstrates the clear culosis is diagnosed with near certainty. If the smear
utility of the E-MTD test and suggests that it may is positive and the NAA is negative, testing the
be particularly helpful for confirming disease in sputum for inhibitors and repeating the assay is war-
intermediate- and high-risk patients and for exclud- ranted [177]. If inhibitors are not detected, and the
ing cases in low-risk patients [169]. process is repeated on additional specimens and is
Other NAA assays have been tested, such as a negative, the patient can be presumed to have NTM.
ligase chain-reaction – based test (LCx test; Abbott If smears are negative, but clinical suspicion is inter-
Diagnostics Division, Abbott Park, IL), and the mediate or high (based on the impression of expe-
strand displacement amplification (SDA) test known rienced observers [169,178,179]), NAA should be
as the BDProbeTec ET Mycobacterium tuberculo- performed on a sputum sample, and a presumptive
sis Complex Direct Detection Assay (DTB) (Becton diagnosis of tuberculosis is made if the test is posi-
Dickinson Biosciences, Sparks, MD). DTB is a tive. NAA should not be performed on sputum sam-
1-hour assay that couples SDA to a fluorescent ples from cases in which the AFB smear is negative
energy-transfer detection system. DTB performs and the clinical index of suspicion is low [169,
similarly to the E-MTD [170,171]. A variety of less 179,180]. Testing should also be limited to those
standardized PCR assays have been developed and who have not been treated recently for active dis-
tested [172 – 175]. Real-time PCR assays have com- ease [177].
260 brodie & schluger
Cost is the main consideration limiting the use of AFB smear and culture are used but generally are
the NAA assays, particularly in the developing world. less sensitive in nonrespiratory samples. Respiratory
A study in Nairobi, Kenya, compared the cost- samples are sometimes of benefit in extrapulmonary
effectiveness of AMPLICOR and that of an AFB tuberculosis. In the case of pleural tuberculosis, the
smear [181]. The AFB smear was 1.8 times as cost- finding of M. tuberculosis in the sputum is diagnostic
effective. The authors, however, concluded that of tuberculosis in patients who have an effusion. Such
AMPLICOR could be cost-effective if ‘‘the largest patients may not easily give expectorated sputum
contributing component, the costs of the PCR-kit, can samples, however. In this setting, IS has been shown
be reduced substantially.’’ A cost-effectiveness analy- to have a sensitivity of 52% for M. tuberculosis
sis conducted in Finland in 2004 showed that the [188], compared with the 60% to 80% sensitivity of
addition of COBAS AMPLICOR PCR to smear and the more invasive pleural biopsy [189].
culture was not cost-effective unless limited to smear- In the case of miliary tuberculosis, sputum smears
positive specimens [182]. Extending this assay to are warranted, but if smears are negative, FOB may
smear-negative specimens may be possible when g play a significant role. FOB was performed in
the E-MTD is used, however, because of its superior 41 patients who had miliary tuberculosis and smear-
sensitivity in smear-negative patients who have pul- negative sputum [190]. Diagnosis was obtained in
monary tuberculosis. Furthermore, centralized labo- 34 patients (83%). BB captured 57% of cases, and
ratories offer the ability to invest in technology, TBB was diagnostic in 73% of cases. A rapid di-
conduct batch testing, develop expertise, and bene- agnosis was made in 27 of 34 patients [190]. In a
fit from economies of scale. In such settings, regu- separate study, 22 patients who had smear-negative
lar NAA testing may be economically feasible miliary tuberculosis underwent FOB with brushings,
[108,183]. aspirate, and TBB. Tuberculosis was diagnosed in
A major limitation of NAA tests is that they give 16 of the 22 patients (73%). A rapid diagnosis was
no drug-susceptibility information. In addition, they made in 14 of 16, from brush smears alone in
are able to detect nucleic acids from both living and 3 patients, aspirate alone in 1, and biopsy alone in
dead organisms and may be falsely positive for ac- 7 [191]. Sampling multiple sites may also be of
tive disease in patients who have a recent history benefit in miliary tuberculosis.
of infection and have been adequately treated [156, There is clearly a role for NAA assays in the
184 – 186]. In contrast to NAAs that employ DNA or diagnosis of extrapulmonary tuberculosis, although
rRNA, the use of an assay to detect M. tuberculosis this role needs to be better defined. The overall sen-
mRNA, with a half-life of only minutes, offers an sitivity in nonrespiratory specimens for the MTD or
indicator of the viability of M. tuberculosis. Assays E-MTD ranges from 67% to 100% [108,153 – 155,
that detect mRNA remain positive only while viable 160,168,170,192]. In smear-negative samples, the
mycobacteria persist and therefore are useful as sensitivity was 52% in one study [160] and 100%
sensitive indicators of adequate treatment and for in another [108]. The AMPLICOR had a similar
rapid determination of drug susceptibility [187]. This sensitivity [162,193], and the specificity of both
technology is under study. assays remains high in nonrespiratory samples. The
assays do not perform equally well in all sample
types; for example, they are much more sensitive in
Extrapulmonary tuberculosis cerebrospinal fluid [192,194] than in pleural fluid
[154]. The sensitivities vary significantly among
Diagnosing extrapulmonary tuberculosis presents studies, as shown in recent meta-analyses of the use
the clinician with many challenges. In most cases, the of NAA tests in tuberculous meningitis [195] and
samples are paucibacillary, decreasing the sensitivity tuberculous pleuritis [196]. In one study, the combi-
of diagnostic tests. Testing for extrapulmonary tuber- nation of AFB smear and MTD in cerebrospinal fluid
culosis follows the same principles as for pulmonary had a sensitivity of 64%, which increased to 83% by
tuberculosis, but, because accuracy of diagnosis is the third sample tested [197]. The DTB system
attenuated in extrapulmonary tuberculosis, clinicians delivers sensitivity similar to the E-MTD in non-
must rely more heavily on clinical judgment and respiratory samples [170,198,199].
response to treatment to diagnose extrapulmonary The use of adenosine deaminase (ADA) levels,
tuberculosis. Meanwhile, the increased incidence of especially in pleural fluid samples, to diagnose extra-
extrapulmonary tuberculosis in HIV patients makes it pulmonary tuberculosis has shown great promise. A
all the more urgent to improve diagnostic strategies recent meta-analysis of 40 studies investigating ADA
for this entity. for the diagnosis of tuberculous pleuritis yielded the
the diagnosis of tuberculosis 261
summary measure of test characteristics derived from specimens or from culture specimens [209,211 – 215].
the receiver operator characteristic curve where sen- Another approach detects actual phenotypic resis-
sitivity equaled specificity at 92.2% [200]. Similarly, tance seen as persistence of the organism in a
a meta-analysis of 31 studies on ADA in pleural rifamycin-containing medium (eg, luciferase reporter
tuberculosis yielded a joint sensitivity and specificity phage assays.)
of 93% [201]. The performance of ADA in diagnos-
ing pleural tuberculosis is inconsistent across studies, Line probe assays
however. In one study, the sensitivity and specificity
were both 55% [202]; in another, they were 88% and Line probe assays use PCR and reverse hybridi-
85.7%, respectively [203]. Some authors report the zation with specific oligonucleotide probes fixed to
need to combine ADA determination with PCR nitrocellulose strips in parallel lines. These assays
analysis, yielding a combined sensitivity of 87.5% may be used for the detection and identification
[204], but others argue that ADA alone is superior to of mycobacterial species or for rapid identification
ADA combined with PCR [205]. of mutations in the rpoB gene. The INNO-LiPA
ADA use outside the pleural space has been MYCOBACTERIA v2 (Innogenetics, Ghent, Bel-
explored as well. ADA may be of limited value gium) and GenoType Mycobacterium (Hain Diag-
in diagnosing tuberculous meningitis [206] but nostika, Nehren, Germany) are line probe assays for
was sensitive for tuberculous pericarditis in one the simultaneous detection and identification of
study [207]. mycobacteria; both are very sensitive [216]. The
Another test that has received some attention for INNO-LiPA Rif.TB assay detects M. tuberculosis and
the diagnosis of pleural and pericardial tuberculosis is very sensitive for detecting rifampin resistance
is pleural or pericardial fluid IFN-g, which has proven [213 – 215,217,218].
comparable to or even better than ADA in some
studies [201,203,207]. Finally, there may be a role for Molecular beacons
the serum IFN-g assays, discussed earlier, in the di-
agnosis of extrapulmonary tuberculosis [81,208]. Molecular beacons are nucleic acid hybridiza-
tion probes. They are designed to bind to target
DNA sequences in regions, such as the rpoB, where
Rapid detection of drug resistance resistance mutations are known to occur. Molecular
beacons fluoresce only when bound to their targets,
Multidrug-resistant (MDR) tuberculosis poses a so that a mutation—even a single-nucleotide sub-
major public health problem in many parts of the stitution—prevents fluorescence. A PCR assay using
world. Traditional methods of drug-susceptibility molecular beacons can identify drug resistance in
testing rely on cultures of M. tuberculosis inoculated sputum samples in less than 3 hours and is both
with antibiotics and can take weeks for results to be sensitive and specific [219]. Lin and colleagues [211]
known. The ability to detect drug resistance rapidly designed a set of molecular beacons for the detection
would be vitally important to tuberculosis-control ef- of isoniazid- and rifampin-resistant mutations in
forts, enabling expeditious administration of appro- M. tuberculosis organisms from both culture-
priate treatment and a decrease in transmission of the and smear-positive respiratory specimens [211]. The
MDR strain. The detection of rifampin resistance may sensitivity and specificity for detection of isoniazid
be used as a surrogate for uncovering multidrug resistance were 82.7% and 100%, respectively, and
resistance, because most rifampin-resistant isolates for rifampin resistance were 97.5% and 100%,
are also isoniazid-resistant [209,210]. Rifampin re- respectively. Piatek et al [220] previously reported
sistance signals the need for treatment with second- similar findings.
line drugs. It is currently feasible to detect rifampin
resistance rapidly. One approach takes advantage Phage amplification
of genotypic abnormalities by identifying mutations
primarily in the region of the M. tuberculosis rpoB Phage amplification uses a bacteriophage to detect
gene associated with most rifampin-resistant strains M. tuberculosis in a given sample within 48 hours.
of M. tuberculosis. Coupling a variety of assays that FASTPlaqueTB (Biotec, Ipswich, Suffolk, UK) uses
identify genetic mutations (line probe assays and phage amplification technology to detect viable M. tu-
molecular beacons, for instance) to PCR or related berculosis in sputum samples and has had mixed
technologies allows rapid detection of the drug- results with excellent specificity (96% – 99%) but
resistant mutations from smear-positive respiratory lesser overall sensitivity (70% – 87%) [116,221 – 223].
262 brodie & schluger
It detected 48.8% of smear-negative cases in one luciferase reporter phages in diagnosing tuberculosis
study [223]. The FASTPlaqueTB-MDRi or FAST- and rapidly detecting drug resistance is still being
PlaqueTB-RIF uses the phage amplification technol- defined. These tests are likely to play an ever-
ogy to determine rifampin resistance in culture or increasing role in the coming years. Ultimately, the
sputum specimens. Albert and colleagues [224] appropriate and affordable use of any of these tests
demonstrated a sensitivity of 100% and a specificity depends on the setting (low or high prevalence of
of 97% for identifying rifampicin-resistant strains in active disease, low or high clinical suspicion in a
solid culture media and in a separate study demon- given patient, available resources, and laboratory
strated similar results using a liquid culture system capabilities) in which they are employed.
[210]. A more recent study by Albert and colleagues
[225] showed a 100% sensitivity and specificity for
determining rifampin resistance directly from smear-
positive sputum, with results also available within References
48 hours.
[1] Centers for Disease Control and Prevention. Screen-
Luciferase reporter phages ing for tuberculosis and tuberculosis infection in
high-risk populations. Recommendations of the Ad-
Firefly luciferase catalyzes the reaction of lucif- visory Council for the Elimination of Tuberculosis.
MMWR Morb Mortal Wkly Rep 1995;44(RR-11):
erin with ATP to generate photons efficiently and
19 – 34.
thereby emit light. Mycobacteriophages expressing [2] Foulds J, O’Brien R. New tools for the diagnosis of
the firefly luciferase gene may be introduced into tuberculosis: the perspective of developing countries
viable mycobacteria [226]. The presence of cellular [see comments]. Int J Tuberc Lung Dis 1998;2(10):
ATP in viable mycobacteria causes visible light to 778 – 83.
be emitted when exogenous luciferin is added. The [3] Diagnostic Standards and Classification of Tuber-
emitted light is measured by a luminometer or on culosis in Adults and Children. This official statement
film (eg, with the Bronx box [227,228]). In the of the American Thoracic Society and the Centers for
presence of adequate antimycobacterial therapy, my- Disease Control and Prevention was adopted by the
cobacteria are rendered nonviable, and the light ATS Board of Directors, July 1999. This statement
was endorsed by the Council of the Infectious Disease
is extinguished. Drug-resistant strains of M. tuber-
Society of America, September 1999. Am J Respir
culosis continue to produce light in the presence of Crit Care Med 2000;161(4 Pt 1):1376 – 95.
antimycobacterial therapy, revealing their resistance. [4] Bergmann JS, Yuoh G, Fish G, et al. Clinical
This method can determine drug susceptibility in 1 to evaluation of the enhanced Gen-Probe Amplified
4 days, and it is also a sensitive and specific means Mycobacterium Tuberculosis Direct Test for rapid
for identifying M. tuberculosis [228 – 232]. diagnosis of tuberculosis in prison inmates. J Clin
Cost and the need for advanced technology and Microbiol 1999;37(5):1419 – 25.
laboratory skills limit the applicability of most of [5] Cohen RA, Muzaffar S, Schwartz D, et al. Diagnosis
these technologies. Efforts to reduce costs and sim- of pulmonary tuberculosis using PCR assays on
plify the technology may make these tests practical sputum collected within 24 hours of hospital ad-
mission. Am J Respir Crit Care Med 1998;157(1):
for widespread use in the near future in resource-rich
156 – 61.
and, perhaps, even in resource-poor countries. [6] Behr MA, Warren SA, Salamon H, et al. Trans-
mission of Mycobacterium tuberculosis from patients
smear-negative for acid-fast bacilli. Lancet 1999;
Summary 353(9151):444 – 9.
[7] Haramati LB, Jenny-Avital ER, Alterman DD. Effect
Diagnostic testing for tuberculosis remained un- of HIV status on chest radiographic and CT findings
changed for nearly a century, but newer technologies in patients with tuberculosis. Clin Radiol 1997;52(1):
hold the promise of a true revolution in tuberculo- 31 – 5.
sis diagnostics. The IFN-g release and T-cell – based [8] Griffiths RI, Hyman CL, McFarlane SI, et al.
Medical-resource use for suspected tuberculosis in
assays may well supplant the TST in diagnosing
a New York City hospital [see comments]. Infect
LTBI in much of the world. NAA assays are proving Control Hosp Epidemiol 1998;19(10):747 – 53.
their worth in more rapidly diagnosing both pulmo- [9] Edwards PQ, Edwards LB. Story of the tuberculin test
nary and extrapulmonary tuberculosis with great from an epidemiologic viewpoint. Am Rev Respir
sensitivity and specificity. The role of line probe as- Dis 1960;81(1)Pt 2:1 – 47.
says, molecular beacons, phage amplification, and [10] Reider H. Epidemiologic basis of tuberculosis con-
the diagnosis of tuberculosis 263
trol. Paris7 International Union Against Tuberculosis Thoracic Society (ATS) and the Centers for Disease
and Lung Disease; 1999. Control and Prevention (CDC) endorsed by the
[11] Lalvani A, Brookes R, Wilkinson RJ, et al. Human Council of the Infectious Diseases Society of
cytolytic and interferon gamma-secreting CD8 + T America. (IDSA), September 1999, and the sections
lymphocytes specific for Mycobacterium tuberculo- of this statement. Am J Respir Crit Care Med 2000;
sis. Proc Natl Acad Sci U S A 1998;95(1):270 – 5. 161(4 Pt 2):S221.
[12] Lalvani A, Nagvenkar P, Udwadia Z, et al. Enu- [25] Huebner RE, Schein MF, Bass Jr JB. The tuberculin
meration of T cells specific for RD1-encoded antigens skin test. Clin Infect Dis 1993;17(6):968 – 75.
suggests a high prevalence of latent Mycobacterium [26] Wallgren A. The time-table of tuberculosis. Tubercle
tuberculosis infection in healthy urban Indians. 1948;29:245 – 51.
J Infect Dis 2001;183(3):469 – 77. [27] Poulsen A. Some clinical features of tuberculosis. 1.
[13] Harboe M. Antigens of PPD, old tuberculin, and Incubation period. Acta Tuberc Scand 1950;24(3 – 4):
autoclaved Mycobacterium bovis BCG studied by 311 – 46.
crossed immunoelectrophoresis. Am Rev Respir Dis [28] Markowitz N, Hansen NI, Wilcosky TC, et al.
1981;124(1):80 – 7. Tuberculin and anergy testing in HIV-seropositive
[14] Judson FN, Feldman RA. Mycobacterial skin tests and HIV-seronegative persons. Pulmonary complica-
in humans 12 years after infection with Mycobac- tions of HIV infection study group. Ann Intern Med
terium marinum. Am Rev Respir Dis 1974;109(5): 1993;119(3):185 – 93.
544 – 7. [29] Graham NM, Nelson KE, Solomon L, et al. Preva-
[15] von Reyn CF, Williams DE, Horsburgh Jr CR, et al. lence of tuberculin positivity and skin test anergy in
Dual skin testing with Mycobacterium avium sensi- HIV-1-seropositive and -seronegative intravenous
tin and purified protein derivative to discriminate drug users. JAMA 1992;267(3):369 – 73.
pulmonary disease due to M. avium complex from [30] Dye C, Scheele S, Dolin P, et al. Consensus
pulmonary disease due to Mycobacterium tuber- statement. Global burden of tuberculosis: estimated
culosis. J Infect Dis 1998;177(3):730 – 6. incidence, prevalence, and mortality by country.
[16] Kwamanga DO, Swai OB, Agwanda R, et al. Effect WHO Global Surveillance and Monitoring Project.
of non-tuberculous Mycobacteria infection on tuber- JAMA 1999;282(7):677 – 86.
culin results among primary school children in [31] Rieder HL. Methodological issues in the estimation
Kenya. East Afr Med J 1995;72(4):222 – 7. of the tuberculosis problem from tuberculin surveys.
[17] Bosman MC, Swai OB, Kwamanga DO, et al. Tuber Lung Dis 1995;76(2):114 – 21.
National tuberculin survey of Kenya, 1986 – 1990. [32] Mayurnath S, Vallishayee RS, Radhamani MP, et al.
Int J Tuberc Lung Dis 1998;2(4):272 – 80. Prevalence study of tuberculous infection over fifteen
[18] Black GF, Weir RE, Floyd S, et al. BCG-induced years, in a rural population in Chingleput district
increase in interferon-gamma response to mycobacte- (south India). Indian J Med Res 1991;93:74 – 80.
rial antigens and efficacy of BCG vaccination in [33] Narain R, Kirishnamurthy MS, Anantharaman DS.
Malawi and the UK: two randomised controlled Prevalence of non-specific sensitivity in some parts
studies. Lancet 2002;359(9315):1393 – 401. of India. Indian J Med Res 1975;63(8):1098 – 109.
[19] Snider Jr DE. Bacille Calmette-Guerin vaccinations [34] Chaparas SD, Vandiviere HM, Melvin I, et al.
and tuberculin skin tests. JAMA 1985;253(23): Tuberculin test. Variability with the Mantoux proce-
3438 – 9. dure. Am Rev Respir Dis 1985;132(1):175 – 7.
[20] Menzies R, Vissandjee B. Effect of bacille Calmette- [35] Bearman JE, Kleinman H, Glyer VV, et al. A study of
Guerin vaccination on tuberculin reactivity. Am Rev variability in tuberculin test reading. Am Rev Respir
Respir Dis 1992;145(3):621 – 5. Dis 1964;90:913 – 9.
[21] Wang L, Turner MO, Elwood RK, et al. A meta- [36] Reichman LB. A scandalous incompetence contin-
analysis of the effect of bacille Calmette Guerin ued. Chest 1998;113(5):1153 – 4.
vaccination on tuberculin skin test measurements. [37] Pouchot J, Grasland A, Collet C, et al. Reliability of
Thorax 2002;57(9):804 – 9. tuberculin skin test measurement. Ann Intern Med
[22] Jasmer RM, Nahid P, Hopewell PC. Clinical practice. 1997;126(3):210 – 4.
Latent tuberculosis infection. N Engl J Med 2002; [38] Mazurek GH, LoBue PA, Daley CL, et al. Com-
347(23):1860 – 6. parison of a whole-blood interferon gamma assay
[23] Fine PE, Bruce J, Ponnighaus JM, et al. Tuberculin with tuberculin skin testing for detecting latent
sensitivity: conversions and reversions in a rural Afri- Mycobacterium tuberculosis infection. JAMA 2001;
can population. Int J Tuberc Lung Dis 1999;3(11): 286(14):1740 – 7.
962 – 75. [39] Comstock GW. False tuberculin test results. Chest
[24] American Thoracic Society. Targeted tuberculin test- 1975;68(3 Suppl):465 – 9.
ing and treatment of latent tuberculosis infection. [40] Lifson AR, Grant SM, Lorvick J, et al. Two-step
This official statement of the American Thoracic tuberculin skin testing of injection drug users
Society was adopted by the ATS Board of Direc- recruited from community-based settings. Int J
tors, July 1999. Joint statement of the American Tuberc Lung Dis 1997;1(2):128 – 34.
264 brodie & schluger
[41] Gordin FM, Perez-Stable EJ, Reid M, et al. Stability bacterium tuberculosis infection. Clin Diagn Lab
of positive tuberculin tests: are boosted reactions Immunol 1998;5(4):531 – 6.
valid? Am Rev Respir Dis 1991;144(3 Pt 1):560 – 3. [57] Bellete B, Coberly J, Barnes GL, et al. Evaluation
[42] Horsburgh Jr CR. Priorities for the treatment of latent of a whole-blood interferon-gamma release assay
tuberculosis infection in the United States. N Engl J for the detection of Mycobacterium tuberculosis
Med 2004;350(20):2060 – 7. infection in 2 study populations. Clin Infect Dis
[43] Andersen P, Munk ME, Pollock JM, et al. Specific 2002;34(11):1449 – 56.
immune-based diagnosis of tuberculosis. Lancet [58] Fietta A, Meloni F, Cascina A, et al. Comparison
2000;356(9235):1099 – 104. of a whole-blood interferon-gamma assay and tuber-
[44] Lyashchenko K, Colangeli R, Houde M, et al. culin skin testing in patients with active tuberculosis
Heterogeneous antibody responses in tuberculosis. and individuals at high or low risk of Mycobacte-
Infect Immun 1998;66(8):3936 – 40. rium tuberculosis infection. Am J Infect Control
[45] Wood PR, Corner LA, Plackett P. Development of 2003;31(6):347 – 53.
a simple, rapid in vitro cellular assay for bovine tu- [59] Harboe M, Nagai S, Patarroyo ME, et al. Properties
berculosis based on the production of gamma inter- of proteins MPB64, MPB70, and MPB80 of Myco-
feron. Res Vet Sci 1990;49(1):46 – 9. bacterium bovis BCG. Infect Immun 1986;52(1):
[46] Rothel JS, Jones SL, Corner LA, et al. A sandwich 293 – 302.
enzyme immunoassay for bovine interferon-gamma [60] Andersen P, Andersen AB, Sorensen AL, et al. Recall
and its use for the detection of tuberculosis in cattle. of long-lived immunity to Mycobacterium tuber-
Aust Vet J 1990;67(4):134 – 7. culosis infection in mice. J Immunol 1995;154(7):
[47] Wood PR, Corner LA, Rothel JS, et al. Field 3359 – 72.
comparison of the interferon-gamma assay and the [61] Berthet FX, Rasmussen PB, Rosenkrands I, et al.
intradermal tuberculin test for the diagnosis of bovine A Mycobacterium tuberculosis operon encoding
tuberculosis. Aust Vet J 1991;68(9):286 – 90. ESAT-6 and a novel low-molecular-mass culture
[48] Wood PR, Corner LA, Rothel JS, et al. A field filtrate protein (CFP-10). Microbiol 1998;144(Pt 11):
evaluation of serological and cellular diagnostic tests 3195 – 203.
for bovine tuberculosis. Vet Microbiol 1992;31(1): [62] Mustafa AS, Amoudy HA, Wiker HG, et al. Compar-
71 – 9. ison of antigen-specific T-cell responses of tuber-
[49] Streeton JA, Desem N, Jones SL. Sensitivity and culosis patients using complex or single antigens of
specificity of a gamma interferon blood test for tuber- Mycobacterium tuberculosis. Scand J Immunol
culosis infection [see comments]. Int J Tuberc Lung 1998;48(5):535 – 43.
Dis 1998;2(6):443 – 50. [63] Ulrichs T, Munk ME, Mollenkopf H, et al. Dif-
[50] Streeton J. QuantiFERON-TB: clinical applications ferential T cell responses to Mycobacterium tuber-
of a new laboratory test for tuberculosis [abstract]. culosis ESAT6 in tuberculosis patients and healthy
Aust N Z J Med 1995;25:435. donors. Eur J Immunol 1998;28(12):3949 – 58.
[51] Mukherjee S. Evaluation of a gamma interferon (IFN- [64] Philipp WJ, Nair S, Guglielmi G, et al. Physical
gamma) assay in human mycobacterial infection mapping of Mycobacterium bovis BCG pasteur
[abstract]. Aust N Z J Med 1995;25:436. reveals differences from the genome map of Myco-
[52] Barnes PF. Diagnosing latent tuberculosis infection: bacterium tuberculosis H37Rv and from M. bovis.
the 100-year upgrade. Am J Respir Crit Care Med Microbiol 1996;142(Pt 11):3135 – 45.
2001;163(4):807 – 8. [65] Roche PW, Feng CG, Britton WJ. Human T-cell epi-
[53] Pottumarthy S, Morris AJ, Harrison AC, et al. topes on the Mycobacterium tuberculosis secreted pro-
Evaluation of the tuberculin gamma interferon assay: tein MPT64. Scand J Immunol 1996;43(6):662 – 70.
potential to replace the Mantoux skin test. J Clin [66] Elhay MJ, Oettinger T, Andersen P. Delayed-type
Microbiol 1999;37(10):3229 – 32. hypersensitivity responses to ESAT-6 and MPT64
[54] Converse PJ, Jones SL, Astemborski J, et al. Com- from Mycobacterium tuberculosis in the guinea pig.
parison of a tuberculin interferon-gamma assay with Infect Immun 1998;66(7):3454 – 6.
the tuberculin skin test in high-risk adults: effect of [67] Roche PW, Winter N, Triccas JA, et al. Expression
human immunodeficiency virus infection. J Infect Dis of Mycobacterium tuberculosis MPT64 in recombi-
1997;176(1):144 – 50. nant Myco. smegmatis: purification, immunogenicity
[55] Kimura M, Converse PJ, Astemborski J, et al. and application to skin tests for tuberculosis. Clin Exp
Comparison between a whole blood interferon- Immunol 1996;103(2):226 – 32.
gamma release assay and tuberculin skin testing for [68] Mahairas GG, Sabo PJ, Hickey MJ, et al. Molecular
the detection of tuberculosis infection among patients analysis of genetic differences between Mycobacte-
at risk for tuberculosis exposure. J Infect Dis 1999; rium bovis BCG and virulent M. bovis. J Bacteriol
179(5):1297 – 300. 1996;178(5):1274 – 82.
[56] Desem N, Jones SL. Development of a human gam- [69] Harboe M, Oettinger T, Wiker HG, et al. Evidence for
ma interferon enzyme immunoassay and comparison occurrence of the ESAT-6 protein in Mycobacterium
with tuberculin skin testing for detection of Myco- tuberculosis and virulent Mycobacterium bovis and
the diagnosis of tuberculosis 265
for its absence in Mycobacterium bovis BCG. Infect Diagnosis of tuberculosis based on the two specific
Immun 1996;64(1):16 – 22. antigens ESAT-6 and CFP10. Clin Diagn Lab
[70] Sorensen AL, Nagai S, Houen G, et al. Purification Immunol 2000;7(2):155 – 60.
and characterization of a low-molecular-mass T-cell [83] Ulrichs T, Anding P, Porcelli S, et al. Increased
antigen secreted by Mycobacterium tuberculosis. numbers of ESAT-6- and purified protein derivative-
Infect Immun 1995;63(5):1710 – 7. specific gamma interferon-producing cells in subclini-
[71] Dillon DC, Alderson MR, Day CH, et al. Molecular cal and active tuberculosis infection. Infect Immun
and immunological characterization of Mycobacte- 2000;68(10):6073 – 6.
rium tuberculosis CFP-10, an immunodiagnostic [84] Arend SM, Geluk A, van Meijgaarden KE, et al.
antigen missing in Mycobacterium bovis BCG. Antigenic equivalence of human T-cell responses to
J Clin Microbiol 2000;38(9):3285 – 90. Mycobacterium tuberculosis-specific RD1-encoded
[72] Cole ST, Brosch R, Parkhill J, et al. Deciphering the protein antigens ESAT-6 and culture filtrate protein
biology of Mycobacterium tuberculosis from the 10 and to mixtures of synthetic peptides. Infect
complete genome sequence [see comments] [pub- Immun 2000;68(6):3314 – 21.
lished erratum appears in Nature 1998;396(6707): [85] Brock I, Munk ME, Kok-Jensen A, et al. Performance
190]. Nature 1998;393(6685):537 – 44. of whole blood IFN-gamma test for tuberculosis
[73] Behr MA, Wilson MA, Gill WP, et al. Comparative diagnosis based on PPD or the specific antigens
genomics of BCG vaccines by whole-genome DNA ESAT-6 and CFP-10. Int J Tuberc Lung Dis 2001;
microarray. Science 1999;284(5419):1520 – 3. 5(5):462 – 7.
[74] Skjot RL, Oettinger T, Rosenkrands I, et al. Com- [86] Wu-Hsieh BA, Chen CK, Chang JH, et al. Long-lived
parative evaluation of low-molecular-mass proteins immune response to early secretory antigenic target 6
from Mycobacterium tuberculosis identifies members in individuals who had recovered from tuberculosis.
of the ESAT-6 family as immunodominant T-cell Clin Infect Dis 2001;33(8):1336 – 40.
antigens. Infect Immun 2000;68(1):214 – 20. [87] Mori T, Sakatani M, Yamagishi F, et al. Specific
[75] Pollock JM, Andersen P. The potential of the ESAT-6 detection of tuberculosis infection: an interferon-
antigen secreted by virulent mycobacteria for specific gamma-based assay using new antigens. Am J Respir
diagnosis of tuberculosis. J Infect Dis 1997;175(5): Crit Care Med 2004;170(1):59 – 64.
1251 – 4. [88] Brock I, Weldingh K, Lillebaek T, et al. Comparison
[76] Brandt L, Oettinger T, Holm A, et al. Key epitopes on of tuberculin skin test and new specific blood test in
the ESAT-6 antigen recognized in mice during the tuberculosis contacts. Am J Respir Crit Care Med
recall of protective immunity to Mycobacterium 2004;170(1):65 – 9.
tuberculosis. J Immunol 1996;157(8):3527 – 33. [89] Vekemans J, Lienhardt C, Sillah JS, et al. Tuber-
[77] Pollock JM, Andersen P. Predominant recognition of culosis contacts but not patients have higher gamma
the ESAT-6 protein in the first phase of interferon interferon responses to ESAT-6 than do community
with Mycobacterium bovis in cattle. Infect Immun controls in The Gambia. Infect Immun 2001;69(10):
1997;65(7):2587 – 92. 6554 – 7.
[78] Vordermeier HM, Whelan A, Cockle PJ, et al. Use of [90] Ravn P, Demissie A, Eguale T, et al. Human T cell
synthetic peptides derived from the antigens ESAT-6 responses to the ESAT-6 antigen from Mycobacte-
and CFP-10 for differential diagnosis of bovine rium tuberculosis. J Infect Dis 1999;179(3):637 – 45.
tuberculosis in cattle. Clin Diagn Lab Immunol [91] Lein AD, von Reyn CF, Ravn P, et al. Cellular
2001;8(3):571 – 8. immune responses to ESAT-6 discriminate between
[79] Johnson PD, Stuart RL, Grayson ML, et al. Tuber- patients with pulmonary disease due to Mycobacte-
culin-purified protein derivative-, MPT-64-, and rium avium complex and those with pulmonary
ESAT-6-stimulated gamma interferon responses in disease due to Mycobacterium tuberculosis. Clin
medical students before and after Mycobacterium Diagn Lab Immunol 1999;6(4):606 – 9.
bovis BCG vaccination and in patients with tuber- [92] Pathan AA, Wilkinson KA, Klenerman P, et al. Direct
culosis. Clin Diagn Lab Immunol 1999;6(6):934 – 7. ex vivo analysis of antigen-specific IFN-gamma-
[80] Arend SM, Engelhard AC, Groot G, et al. Tuberculin secreting CD4 T cells in Mycobacterium tuberculosis-
skin testing compared with T-cell responses to infected individuals: associations with clinical disease
Mycobacterium tuberculosis-specific and nonspecific state and effect of treatment. J Immunol 2001;167(9):
antigens for detection of latent infection in persons 5217 – 25.
with recent tuberculosis contact. Clin Diagn Lab [93] Barnes PF. Diagnosing latent tuberculosis infection:
Immunol 2001;8(6):1089 – 96. turning glitter to gold. Am J Respir Crit Care Med
[81] Arend SM, Andersen P, van Meijgaarden KE, et al. 2004;170(1):5 – 6.
Detection of active tuberculosis infection by T cell [94] Lalvani A, Pathan AA, McShane H, et al. Rapid
responses to early-secreted antigenic target 6-kDa detection of Mycobacterium tuberculosis infection by
protein and culture filtrate protein 10. J Infect Dis enumeration of antigen-specific T cells. Am J Respir
2000;181(5):1850 – 4. Crit Care Med 2001;163(4):824 – 8.
[82] van Pinxteren LA, Ravn P, Agger EM, et al. [95] Lalvani A, Pathan AA, Durkan H, et al. Enhanced
266 brodie & schluger
contact tracing and spatial tracking of Mycobacte- a large public health mycobacteriology laboratory.
rium tuberculosis infection by enumeration of Diagn Microbiol Infect Dis 1999;35(3):185 – 91.
antigen-specific T cells. Lancet 2001;357(9273): [109] Rajalahti I, Vuorinen P, Nieminen MM, et al.
2017 – 21. Detection of Mycobacterium tuberculosis complex
[96] Chapman AL, Munkanta M, Wilkinson KA, et al. in sputum specimens by the automated Roche Cobas
Rapid detection of active and latent tuberculosis Amplicor Mycobacterium Tuberculosis Test. J Clin
infection in HIV-positive individuals by enumeration Microbiol 1998;36(4):975 – 8.
of Mycobacterium tuberculosis-specific T cells. AIDS [110] Murray PR, Elmore C, Krogstad DJ. The acid-fast
2002;16(17):2285 – 93. stain: a specific and predictive test for mycobacterial
[97] Ewer K, Deeks J, Alvarez L, et al. Comparison of disease. Ann Intern Med 1980;92(4):512 – 3.
T-cell-based assay with tuberculin skin test for [111] Strumpf IJ, Tsang AY, Sayre JW. Re-evaluation of
diagnosis of Mycobacterium tuberculosis infection sputum staining for the diagnosis of pulmonary tuber-
in a school tuberculosis outbreak. Lancet 2003; culosis. Am Rev Respir Dis 1979;119(4):599 – 602.
361(9364):1168 – 73. [112] Mehta J, Krish G, Berro E, et al. Fiberoptic bron-
[98] Antonucci G, Girardi E, Raviglione MC, et al. Risk choscopy in the diagnosis of pulmonary tuberculosis.
factors for tuberculosis in HIV-infected persons. A South Med J 1990;83(7):753 – 5.
prospective cohort study. The Gruppo Italiano di [113] Baughman RP, Dohn MN, Loudon RG, et al.
Studio Tubercolosi e AIDS (GISTA). JAMA 1995; Bronchoscopy with bronchoalveolar lavage in tuber-
274(2):143 – 8. culosis and fungal infections. Chest 1991;99(1):
[99] Comstock GW. Frost revisited: the modern epidemi- 92 – 7.
ology of tuberculosis. Am J Epidemiol 1975;101(5): [114] Kim TC, Blackman RS, Heatwole KM, et al. Acid-
363 – 82. fast bacilli in sputum smears of patients with
[100] Sepkowitz KA. Tuberculin skin testing and the health pulmonary tuberculosis. Prevalence and significance
care worker: lessons of the Prophit Survey. Tuber of negative smears pretreatment and positive smears
Lung Dis 1996;77(1):81 – 5. post-treatment. Am Rev Respir Dis 1984;129(2):
[101] Doherty TM, Demissie A, Olobo J, et al. Immune 264 – 8.
responses to the Mycobacterium tuberculosis-specific [115] British Medical Research Council. Sputum-smear-
antigen ESAT-6 signal subclinical infection among negative pulmonary tuberculosis: controlled trial of
contacts of tuberculosis patients. J Clin Microbiol 3-month and 2-month regimens of chemotherapy.
2002;40(2):704 – 6. Lancet 1979;1(8131):1361 – 3.
[102] Centers for Disease Control and Prevention. Guide- [116] Mbulo GM, Kambashi BS, Kinkese J, et al. Compari-
lines for preventing the transmission of Mycobacte- son of two bacteriophage tests and nucleic acid
rium tuberculosis in health-care facilities, 1994. amplification for the diagnosis of pulmonary tuber-
MMWR Recomm Rep 1994;43(RR-13):1 – 132. culosis in sub-Saharan Africa. Int J Tuberc Lung Dis
[103] Jacubowiak W. TB manual. National tuberculosis 2004;8(11):1342 – 7.
programme guidelines. Geneva (Switzerland)7 World [117] Hensler NM, Spivey Jr CG, Dees TM. The use of
Health Organization; 2001. hypertonic aerosol in production of sputum for
[104] Mathew P, Kuo YH, Vazirani B, et al. Are three diagnosis of tuberculosis. Comparison with gastric
sputum acid-fast bacillus smears necessary for dis- specimens. Dis Chest 1961;40:639 – 42.
continuing tuberculosis isolation? J Clin Microbiol [118] Jones Jr FL. The relative efficacy of spontaneous
2002;40(9):3482 – 4. sputa, aerosol-induced sputa, and gastric aspirates in
[105] Craft DW, Jones MC, Blanchet CN, et al. Value of the bacteriologic diagnosis of pulmonary tuberculosis.
examining three acid-fast bacillus sputum smears for Dis Chest 1966;50(4):403 – 8.
removal of patients suspected of having tuberculosis [119] Carr DT, Karlson AG, Stilwell GG. A comparison of
from the ‘‘airborne precautions’’ category. J Clin cultures of induced sputum and gastric washings in
Microbiol 2000;38(11):4285 – 7. the diagnosis of tuberculosis. Mayo Clin Proc 1967;
[106] Wu ZL, Wang AQ. Diagnostic yield of repeated 42(1):23 – 5.
smear microscopy examinations among patients [120] Dickson SJ, Brent A, Davidson RN, et al. Compari-
suspected of pulmonary TB in Shandong province son of bronchoscopy and gastric washings in the
of China. Int J Tuberc Lung Dis 2000;4(11):1086 – 7. investigation of smear-negative pulmonary tubercu-
[107] Dalovisio JR, Montenegro-James S, Kemmerly SA, losis. Clin Infect Dis 2003;37(12):1649 – 53.
et al. Comparison of the amplified Mycobacterium [121] Li LM, Bai LQ, Yang HL, et al. Sputum induction to
tuberculosis (MTB) direct test, Amplicor MTB PCR, improve the diagnostic yield in patients with sus-
and IS6110-PCR for detection of MTB in respiratory pected pulmonary tuberculosis. Int J Tuberc Lung Dis
specimens. Clin Infect Dis 1996;23(5):1099 – 106 1999;3(12):1137 – 9.
[discussion: 1107 – 8]. [122] Parry CM, Kamoto O, Harries AD, et al. The use of
[108] Chedore P, Jamieson FB. Routine use of the Gen- sputum induction for establishing a diagnosis in
Probe MTD2 amplification test for detection of patients with suspected pulmonary tuberculosis in
Mycobacterium tuberculosis in clinical specimens in Malawi. Tuber Lung Dis 1995;76(1):72 – 6.
the diagnosis of tuberculosis 267
[123] Hartung TK, Maulu A, Nash J, et al. Suspected pulmonary tuberculosis. Acta Tuberc Scand 1957;
pulmonary tuberculosis in rural South Africa— 33(3):241 – 64.
sputum induction as a simple diagnostic tool? S Afr [139] Ip M, Chau PY, So SY, et al. The value of routine
Med J 2002;92(6):455 – 8. bronchial aspirate culture at fibreoptic bronchoscopy
[124] Merrick ST, Sepkowitz KA, Walsh J, et al. Compari- for the diagnosis of tuberculosis. Tubercle 1989;
son of induced versus expectorated sputum for 70(4):281 – 5.
diagnosis of pulmonary tuberculosis by acid-fast [140] Chan HS, Sun AJ, Hoheisel GB. Bronchoscopic
smear. Am J Infect Control 1997;25(6):463 – 6. aspiration and bronchoalveolar lavage in the diag-
[125] Bell D, Leckie V, McKendrick M. The role of nosis of sputum smear-negative pulmonary tuber-
induced sputum in the diagnosis of pulmonary culosis. Lung 1990;168(4):215 – 20.
tuberculosis. J Infect 2003;47(4):317 – 21. [141] al-Kassimi FA, Azhar M, al-Majed S, et al. Diag-
[126] Al Zahrani K, Al Jahdali H, Poirier L, et al. Yield nostic role of fibreoptic bronchoscopy in tuberculosis
of smear, culture and amplification tests from in the presence of typical X-ray pictures and adequate
repeated sputum induction for the diagnosis of pul- sputum. Tubercle 1991;72(2):145 – 8.
monary tuberculosis. Int J Tuberc Lung Dis 2001; [142] Palenque E, Amor E, Bernaldo de Quiros JC.
5(9):855 – 60. Comparison of bronchial washing, brushing and
[127] McWilliams T, Wells AU, Harrison AC, et al. biopsy for diagnosis of pulmonary tuberculosis. Eur
Induced sputum and bronchoscopy in the diagnosis J Clin Microbiol 1987;6(2):191 – 2.
of pulmonary tuberculosis. Thorax 2002;57(12): [143] Uddenfeldt M, Lundgren R. Flexible fiberoptic
1010 – 4. bronchoscopy in the diagnosis of pulmonary tuber-
[128] Anderson C, Inhaber N, Menzies D. Comparison of culosis. Tubercle 1981;62(3):197 – 9.
sputum induction with fiber-optic bronchoscopy in [144] Wallace JM, Deutsch AL, Harrell JH, et al. Bronchos-
the diagnosis of tuberculosis. Am J Respir Crit Care copy and transbronchial biopsy in evaluation of pa-
Med 1995;152(5 Pt 1):1570 – 4. tients with suspected active tuberculosis. Am J Med
[129] Conde MB, Soares SL, Mello FC, et al. Comparison 1981;70(6):1189 – 94.
of sputum induction with fiberoptic bronchoscopy in [145] Fujii H, Ishihara J, Fukaura A, et al. Early diagnosis
the diagnosis of tuberculosis: experience at an of tuberculosis by fibreoptic bronchoscopy. Tuber
acquired immune deficiency syndrome reference Lung Dis 1992;73(3):167 – 9.
center in Rio de Janeiro, Brazil. Am J Respir Crit [146] Wongthim S, Udompanich V, Limthongkul S, et al.
Care Med 2000;162(6):2238 – 40. Fiberoptic bronchoscopy in diagnosis of patients with
[130] Larson JL, Ridzon R, Hannan MM. Sputum induction suspected active pulmonary tuberculosis. J Med
versus fiberoptic bronchoscopy in the diagnosis of Assoc Thai 1989;72(3):154 – 9.
tuberculosis. Am J Respir Crit Care Med 2001; [147] de Gracia J, Curull V, Vidal R, et al. Diagnostic value
163(5):1279 – 80. of bronchoalveolar lavage in suspected pulmonary
[131] Chawla R, Pant K, Jaggi OP, et al. Fibreoptic tuberculosis. Chest 1988;93(2):329 – 32.
bronchoscopy in smear-negative pulmonary tuber- [148] Morgan MA, Horstmeier CD, DeYoung DR, et al.
culosis. Eur Respir J 1988;1(9):804 – 6. Comparison of a radiometric method (BACTEC) and
[132] So SY, Lam WK, Yu DY. Rapid diagnosis of conventional culture media for recovery of mycobac-
suspected pulmonary tuberculosis by fiberoptic bron- teria from smear-negative specimens. J Clin Micro-
choscopy. Tubercle 1982;63(3):195 – 200. biol 1983;18(2):384 – 8.
[133] Willcox PA, Benatar SR, Potgieter PD. Use of the [149] Sharp SE, Lemes M, Sierra SG, et al. Lowenstein-
flexible fibreoptic bronchoscope in diagnosis of Jensen media. No longer necessary for mycobacterial
sputum-negative pulmonary tuberculosis. Thorax isolation. Am J Clin Pathol 2000;113(6):770 – 3.
1982;37(8):598 – 601. [150] Ichiyama S, Shimokata K, Takeuchi J. Comparative
[134] Sarkar SK, Sharma GS, Gupta PR, et al. Fiberoptic study of a biphasic culture system (Roche MB Check
bronchoscopy in the diagnosis of pulmonary tuber- system) with a conventional egg medium for recovery
culosis. Tubercle 1980;61(2):97 – 9. of mycobacteria. Aichi Mycobacteriosis Research
[135] Danek SJ, Bower JS. Diagnosis of pulmonary tuber- Group. Tuber Lung Dis 1993;74(5):338 – 41.
culosis by flexible fiberoptic bronchoscopy. Am Rev [151] Kanchana MV, Cheke D, Natyshak I, et al. Evaluation
Respir Dis 1979;119(4):677 – 9. of the BACTEC MGIT 960 system for the recovery
[136] Kennedy DJ, Lewis WP, Barnes PF. Yield of of mycobacteria. Diagn Microbiol Infect Dis
bronchoscopy for the diagnosis of tuberculosis in 2000;37(1):31 – 6.
patients with human immunodeficiency virus infec- [152] Sharp SE, Lemes M, Erlich SS, et al. A comparison
tion. Chest 1992;102(4):1040 – 4. of the Bactec 9000MB system and the Septi-Chek
[137] Russell MD, Torrington KG, Tenholder MF. A ten- AFB system for the detection of mycobacteria. Diagn
year experience with fiberoptic bronchoscopy for Microbiol Infect Dis 1997;28(2):69 – 74.
mycobacterial isolation. Impact of the Bactec system. [153] Gamboa F, Fernandez G, Padilla E, et al. Compara-
Am Rev Respir Dis 1986;133(6):1069 – 71. tive evaluation of initial and new versions of the Gen-
[138] Palva T, Elo R, Saloheimo M. Bronchoscopy in Probe Amplified Mycobacterium Tuberculosis Direct
268 brodie & schluger
Test for direct detection of Mycobacterium tuber- reaction for Mycobacterium tuberculosis: impact on
culosis in respiratory and nonrespiratory specimens. clinical management of refugees with pulmonary
J Clin Microbiol 1998;36(3):684 – 9. infiltrates. Chest 2004;125(3):981 – 6.
[154] Vlaspolder F, Singer P, Roggeveen C. Diagnostic [168] Piersimoni C, Callegaro A, Scarparo C, et al. Com-
value of an amplification method (Gen-Probe) com- parative evaluation of the new Gen-Probe Mycobac-
pared with that of culture for diagnosis of tuber- terium tuberculosis amplified direct test and the
culosis. J Clin Microbiol 1995;33(10):2699 – 703. semiautomated Abbott LCx Mycobacterium tuber-
[155] Pfyffer GE, Kissling P, Jahn EM, et al. Diagnostic culosis assay for direct detection of Mycobacterium
performance of amplified Mycobacterium tuberculo- tuberculosis complex in respiratory and extra-
sis direct test with cerebrospinal fluid, other non- pulmonary specimens [see comments]. J Clin Micro-
respiratory, and respiratory specimens. J Clin biol 1998;36(12):3601 – 4.
Microbiol 1996;34(4):834 – 41. [169] Catanzaro A, Perry S, Clarridge JE, et al. The role
[156] Bradley SP, Reed SL, Catanzaro A. Clinical efficacy of clinical suspicion in evaluating a new diagnostic
of the amplified Mycobacterium tuberculosis direct test for active tuberculosis: results of a multicenter
test for the diagnosis of pulmonary tuberculosis. Am J prospective trial. JAMA 2000;283(5):639 – 45.
Respir Crit Care Med 1996;153(5):1606 – 10. [170] Piersimoni C, Scarparo C, Piccoli P, et al. Perfor-
[157] Della-Latta P, Whittier S. Comprehensive evaluation mance assessment of two commercial amplification
of performance, laboratory application, and clinical assays for direct detection of Mycobacterium tuber-
usefulness of two direct amplification technologies culosis complex from respiratory and extrapulmo-
for the detection of Mycobacterium tuberculosis nary specimens. J Clin Microbiol 2002;40(11):
complex. Am J Clin Pathol 1998;110(3):301 – 10. 4138 – 42.
[158] Pfyffer GE, Kissling P, Wirth R, et al. Direct detection [171] Visca P, De Mori P, Festa A, et al. Evaluation of the
of Mycobacterium tuberculosis complex in respira- BDProbeTec strand displacement amplification assay
tory specimens by a target-amplified test system. in comparison with the AMTD II direct test for rapid
J Clin Microbiol 1994;32(4):918 – 23. diagnosis of tuberculosis. Clin Microbiol Infect 2004;
[159] Wang SX, Tay L. Evaluation of three nucleic acid 10(4):332 – 4.
amplification methods for direct detection of Myco- [172] Sperhacke RD, Mello FC, Zaha A, et al. Detection of
bacterium tuberculosis complex in respiratory speci- Mycobacterium tuberculosis by a polymerase chain
mens. J Clin Microbiol 1999;37(6):1932 – 4. reaction colorimetric dot-blot assay. Int J Tuberc
[160] Coll P, Garrigo M, Moreno C, et al. Routine use of Lung Dis 2004;8(3):312 – 7.
Gen-Probe Amplified Mycobacterium Tuberculosis [173] Lemaitre N, Armand S, Vachee A, et al. Comparison
Direct (MTD) test for detection of Mycobacterium of the real-time PCR method and the Gen-Probe
tuberculosis with smear-positive and smear-negative amplified Mycobacterium tuberculosis direct test for
specimens. Int J Tuberc Lung Dis 2003;7(9):886 – 91. detection of Mycobacterium tuberculosis in pulmo-
[161] Wobeser WL, Krajden M, Conly J, et al. Evaluation nary and nonpulmonary specimens. J Clin Microbiol
of Roche Amplicor PCR assay for Mycobacterium 2004;42(9):4307 – 9.
tuberculosis. J Clin Microbiol 1996;34(1):134 – 9. [174] Miller N, Cleary T, Kraus G, et al. Rapid and specific
[162] Carpentier E, Drouillard B, Dailloux M, et al. detection of Mycobacterium tuberculosis from acid-
Diagnosis of tuberculosis by Amplicor Mycobacte- fast bacillus smear-positive respiratory specimens and
rium tuberculosis test: a multicenter study. J Clin BacT/ALERT MP culture bottles by using fluoro-
Microbiol 1995;33(12):3106 – 10. genic probes and real-time PCR. J Clin Microbiol
[163] Dilworth JP, Goyal M, Young DB, et al. Comparison 2002;40(11):4143 – 7.
of polymerase chain reaction for IS6110 and Ampli- [175] Cleary TJ, Roudel G, Casillas O, et al. Rapid and
cor in the diagnosis of tuberculosis. Thorax 1996; specific detection of Mycobacterium tuberculosis by
51(3):320 – 2. using the Smart Cycler instrument and a specific
[164] Tevere VJ, Hewitt PL, Dare A, et al. Detection of fluorogenic probe. J Clin Microbiol 2003;41(10):
Mycobacterium tuberculosis by PCR amplification 4783 – 6.
with pan-Mycobacterium primers and hybridization [176] Centers for Disease Control and Prevention. Update.
to an M. tuberculosis-specific probe. J Clin Microbiol Nucleic acid amplification tests for tuberculosis [in
1996;34(4):918 – 23. process citation]. MMWR Morb Mortal Wkly Rep
[165] Lim T. Relationship between estimated pretest proba- 2000;49(26):593 – 4.
bility and accuracy of automated Mycobacterium [177] Sloutsky A, Han LL, Werner BG. Practical strategies
tuberculosis assay in smear-negative pulmonary for performance optimization of the enhanced Gen-
tuberculosis. Chest 2000;118:641 – 7. Probe amplified Mycobacterium tuberculosis direct
[166] Bergmann JS, Woods GL. Clinical evaluation of the test. J Clin Microbiol 2004;42(4):1547 – 51.
Roche AMPLICOR PCR Mycobacterium tuberculo- [178] Divinagracia RM, Harkin TJ, Bonk S, et al. Screening
sis test for detection of M. tuberculosis in respiratory by specialists to reduce unnecessary test ordering in
specimens. J Clin Microbiol 1996;34(5):1083 – 5. patients evaluated for tuberculosis [see comments].
[167] Laifer G, Widmer AF, Frei R, et al. Polymerase chain Chest 1998;114(3):681 – 4.
the diagnosis of tuberculosis 269
[179] Lim TK, Mukhopadhyay A, Gough A, et al. Role of diagnosis of tuberculosis in various biopsy and body
clinical judgment in the application of a nucleic acid fluid specimens by the AMPLICOR Mycobacterium
amplification test for the rapid diagnosis of pulmo- tuberculosis polymerase chain reaction test. Chest
nary tuberculosis. Chest 2003;124(3):902 – 8. 1998;113(5):1190 – 4.
[180] Van den Wijngaert S, Dediste A, VanLaetham Y, [194] Cloud JL, Shutt C, Aldous W, et al. Evaluation of a
et al. Critical use of nucleic acid amplification tech- modified Gen-Probe amplified direct test for detec-
niques to test for Mycobacterium tuberculosis in re- tion of Mycobacterium tuberculosis complex organ-
spiratory tract samples. J Clin Microbiol 2004;42(2): isms in cerebrospinal fluid. J Clin Microbiol 2004;
837 – 8. 42(11):5341 – 4.
[181] Roos BR, van Cleeff MR, Githui WA, et al. Cost- [195] Pai M, Flores LL, Pai N, et al. Diagnostic accuracy
effectiveness of the polymerase chain reaction versus of nucleic acid amplification tests for tuberculous
smear examination for the diagnosis of tuberculosis in meningitis: a systematic review and meta-analysis.
Kenya: a theoretical model. Int J Tuberc Lung Dis Lancet Infect Dis 2003;3(10):633 – 43.
1998;2(3):235 – 41. [196] Pai M, Flores LL, Hubbard A, et al. Nucleic acid
[182] Rajalahti I, Ruokonen EL, Kotomaki T, et al. amplification tests in the diagnosis of tuberculous
Economic evaluation of the use of PCR assay in pleuritis: a systematic review and meta-analysis.
diagnosing pulmonary TB in a low-incidence area. BMC Infect Dis 2004;4(1):6.
Eur Respir J 2004;23(3):446 – 51. [197] Thwaites GE, Caws M, Chau TT, et al. Comparison
[183] Dowdy DW, Maters A, Parrish N, et al. Cost- of conventional bacteriology with nucleic acid ampli-
effectiveness analysis of the Gen-Probe amplified fication (amplified mycobacterium direct test) for
Mycobacterium tuberculosis direct test as used diagnosis of tuberculous meningitis before and after
routinely on smear-positive respiratory specimens. inception of antituberculosis chemotherapy. J Clin
J Clin Microbiol 2003;41(3):948 – 53. Microbiol 2004;42(3):996 – 1002.
[184] Yuen KY, Chan KS, Chan CM, et al. Use of PCR in [198] Mazzarelli G, Rindi L, Piccoli P, et al. Evaluation of
routine diagnosis of treated and untreated pulmonary the BDProbeTec ET system for direct detection of
tuberculosis. J Clin Pathol 1993;46(4):318 – 22. Mycobacterium tuberculosis in pulmonary and
[185] Walker DA, Taylor IK, Mitchell DM, et al. Com- extrapulmonary samples: a multicenter study. J Clin
parison of polymerase chain reaction amplification of Microbiol 2003;41(4):1779 – 82.
two mycobacterial DNA sequences, IS6110 and the [199] Johansen IS, Lundgren B, Tabak F, et al. Improved
65 kDa antigen gene, in the diagnosis of tuberculosis. sensitivity of nucleic acid amplification for rapid
Thorax 1992;47(9):690 – 4. diagnosis of tuberculous meningitis. J Clin Microbiol
[186] Schluger NW, Kinney D, Harkin TJ, et al. Clinical 2004;42(7):3036 – 40.
utility of the polymerase chain reaction in the [200] Goto M, Noguchi Y, Koyama H, et al. Diagnostic
diagnosis of infections due to Mycobacterium tuber- value of adenosine deaminase in tuberculous pleural
culosis. Chest 1994;105(4):1116 – 21. effusion: a meta-analysis. Ann Clin Biochem 2003;
[187] Hellyer TJ, DesJardin LE, Teixeira L, et al. Detection 40(Pt 4):374 – 81.
of viable Mycobacterium tuberculosis by reverse [201] Greco S, Girardi E, Masciangelo R, et al. Adenosine
transcriptase-strand displacement amplification of deaminase and interferon gamma measurements for
mRNA. J Clin Microbiol 1999;37(3):518 – 23. the diagnosis of tuberculous pleurisy: a meta-analysis.
[188] Conde MB, Loivos AC, Rezende VM, et al. Yield of Int J Tuberc Lung Dis 2003;7(8):777 – 86.
sputum induction in the diagnosis of pleural tuber- [202] Nagesh BS, Sehgal S, Jindal SK, et al. Evaluation of
culosis. Am J Respir Crit Care Med 2003;167(5): polymerase chain reaction for detection of Mycobac-
723 – 5. terium tuberculosis in pleural fluid. Chest 2001;
[189] Menzies D. Sputum induction: simpler, cheaper, and 119(6):1737 – 41.
safer – but is it better? Am J Respir Crit Care Med [203] Villegas MV, Labrada LA, Saravia NG. Evaluation of
2003;167(5):676 – 7. polymerase chain reaction, adenosine deaminase, and
[190] Willcox PA, Potgieter PD, Bateman ED, et al. Rapid interferon-gamma in pleural fluid for the differential
diagnosis of sputum negative miliary tuberculosis diagnosis of pleural tuberculosis. Chest 2000;118(5):
using the flexible fibreoptic bronchoscope. Thorax 1355 – 64.
1986;41(9):681 – 4. [204] Lima DM, Colares JK, da Fonseca BA. Combined
[191] Pant K, Chawla R, Mann PS, et al. Fiberbronchos- use of the polymerase chain reaction and detection
copy in smear-negative miliary tuberculosis. Chest of adenosine deaminase activity on pleural fluid im-
1989;95(5):1151 – 2. proves the rate of diagnosis of pleural tuberculosis.
[192] Lang AM, Feris-Iglesias J, Pena C, et al. Clinical Chest 2003;124(3):909 – 14.
evaluation of the Gen-Probe amplified direct test for [205] Trajman A, Kaisermann MC, Kritski AL, et al.
detection of Mycobacterium tuberculosis complex Diagnosing pleural tuberculosis. Chest 2004;125(6):
organisms in cerebrospinal fluid. J Clin Microbiol 2366 [author reply: 2366 – 7].
1998;36(8):2191 – 4. [206] Corral I, Quereda C, Navas E, et al. Adenosine
[193] Shah S, Miller A, Mastellone A, et al. Rapid deaminase activity in cerebrospinal fluid of HIV-
270 brodie & schluger
infected patients: limited value for diagnosis of son of three molecular assays for rapid detection of
tuberculous meningitis. Eur J Clin Microbiol Infect rifampin resistance in Mycobacterium tuberculosis.
Dis 2004;23(6):471 – 6. J Clin Microbiol 1998;36(7):1969 – 73.
[207] Burgess LJ, Reuter H, Carstens ME, et al. The use of [219] El-Hajj HH, Marras SA, Tyagi S, et al. Detection of
adenosine deaminase and interferon-gamma as diag- rifampin resistance in Mycobacterium tuberculosis in
nostic tools for tuberculous pericarditis. Chest a single tube with molecular beacons. J Clin Micro-
2002;122(3):900 – 5. biol 2001;39(11):4131 – 7.
[208] Munk ME, Arend SM, Brock I, et al. Use of [220] Piatek AS, Telenti A, Murray MR, et al. Genotypic
ESAT-6 and CFP-10 antigens for diagnosis of ex- analysis of Mycobacterium tuberculosis in two
trapulmonary tuberculosis. J Infect Dis 2001;183(1): distinct populations using molecular beacons: impli-
175 – 6. cations for rapid susceptibility testing. Antimicrob
[209] Fan XY, Hu ZY, Xu FH, et al. Rapid detection of Agents Chemother 2000;44(1):103 – 10.
rpoB gene mutations in rifampin-resistant Mycobac- [221] Butt T, Ahmad RN, Kazmi SY, et al. Rapid diagnosis
terium tuberculosis isolates in shanghai by using the of pulmonary tuberculosis by mycobacteriophage
amplification refractory mutation system. J Clin assay. Int J Tuberc Lung Dis 2004;8(7):899 – 902.
Microbiol 2003;41(3):993 – 7. [222] Albay A, Kisa O, Baylan O, et al. The evaluation of
[210] Albert H, Trollip AP, Mole RJ, et al. Rapid FASTPlaqueTB test for the rapid diagnosis of tu-
indication of multidrug-resistant tuberculosis from berculosis. Diagn Microbiol Infect Dis 2003;46(3):
liquid cultures using FASTPlaqueTB-RIF, a manual 211 – 5.
phage-based test. Int J Tuberc Lung Dis 2002;6(6): [223] Albert H, Heydenrych A, Brookes R, et al. Perfor-
523 – 8. mance of a rapid phage-based test, FASTPlaqueTB, to
[211] Lin SY, Probert W, Lo M, et al. Rapid detection of diagnose pulmonary tuberculosis from sputum speci-
isoniazid and rifampin resistance mutations in Myco- mens in South Africa. Int J Tuberc Lung Dis 2002;
bacterium tuberculosis complex from cultures or 6(6):529 – 37.
smear-positive sputa by use of molecular beacons. [224] Albert H, Heydenrych A, Mole R, et al. Evaluation
J Clin Microbiol 2004;42(9):4204 – 8. of FASTPlaqueTB-RIF, a rapid, manual test for the
[212] Mokrousov I, Otten T, Vyshnevskiy B, et al. Allele- determination of rifampicin resistance from Myco-
specific rpoB PCR assays for detection of rifampin- bacterium tuberculosis cultures. Int J Tuberc Lung
resistant Mycobacterium tuberculosis in sputum Dis 2001;5(10):906 – 11.
smears. Antimicrob Agents Chemother 2003;47(7): [225] Albert H, Trollip A, Seaman T, et al. Simple, phage-
2231 – 5. based (FASTPplaque) technology to determine ri-
[213] Cooksey RC, Morlock GP, Glickman S, et al. fampicin resistance of Mycobacterium tuberculosis
Evaluation of a line probe assay kit for characteri- directly from sputum. Int J Tuberc Lung Dis 2004;
zation of rpoB mutations in rifampin-resistant Myco- 8(9):1114 – 9.
bacterium tuberculosis isolates from New York City. [226] Jacobs Jr WR, Barletta RG, Udani R, et al. Rapid
J Clin Microbiol 1997;35(5):1281 – 3. assessment of drug susceptibilities of Mycobacte-
[214] Hirano K, Abe C, Takahashi M. Mutations in the rium tuberculosis by means of luciferase reporter
rpoB gene of rifampin-resistant Mycobacterium phages [see comments]. Science 1993;260(5109):
tuberculosis strains isolated mostly in Asian countries 819 – 22.
and their rapid detection by line probe assay. J Clin [227] Riska PF, Su Y, Bardarov S, et al. Rapid film-based
Microbiol 1999;37(8):2663 – 6. determination of antibiotic susceptibilities of Myco-
[215] Marttila HJ, Soini H, Vyshnevskaya E, et al. Line bacterium tuberculosis strains by using a luciferase
probe assay in the rapid detection of rifampin- reporter phage and the Bronx Box. J Clin Microbiol
resistant Mycobacterium tuberculosis directly from 1999;37(4):1144 – 9.
clinical specimens. Scand J Infect Dis 1999;31(3): [228] Hazbon MH, Guarin N, Ferro BE, et al. Photographic
269 – 73. and luminometric detection of luciferase reporter
[216] Padilla E, Gonzalez V, Manterola JM, et al. Com- phages for drug susceptibility testing of clinical
parative evaluation of the new version of the INNO- Mycobacterium tuberculosis isolates. J Clin Micro-
LiPA Mycobacteria and genotype Mycobacterium biol 2003;41(10):4865 – 9.
assays for identification of Mycobacterium species [229] Carriere C, Riska PF, Zimhony O, et al. Conditionally
from MB/BacT liquid cultures artificially inoculated replicating luciferase reporter phages: improved
with Mycobacterial strains. J Clin Microbiol 2004; sensitivity for rapid detection and assessment of drug
42(7):3083 – 8. susceptibility of Mycobacterium tuberculosis. J Clin
[217] Johansen IS, Lundgren B, Sosnovskaja A, et al. Microbiol 1997;35(12):3232 – 9.
Direct detection of multidrug-resistant Mycobacte- [230] Riska PF, Jacobs Jr WR, Bloom BR, et al. Specific
rium tuberculosis in clinical specimens in low- and identification of Mycobacterium tuberculosis with the
high-incidence countries by line probe assay. J Clin luciferase reporter mycobacteriophage: use of p-nitro-
Microbiol 2003;41(9):4454 – 6. alpha-acetylamino-beta-hydroxy propiophenone.
[218] Watterson SA, Wilson SM, Yates MD, et al. Compari- J Clin Microbiol 1997;35(12):3225 – 31.
the diagnosis of tuberculosis 271
[231] Banaiee N, Bobadilla-Del-Valle M, Bardarov Jr S, [232] Banaiee N, Bobadilla-del-Valle M, Riska PF, et al.
et al. Luciferase reporter mycobacteriophages for Rapid identification and susceptibility testing of
detection, identification, and antibiotic susceptibility Mycobacterium tuberculosis from MGIT cultures
testing of Mycobacterium tuberculosis in Mexico. with luciferase reporter mycobacteriophages. J Med
J Clin Microbiol 2001;39(11):3883 – 8. Microbiol 2003;52(Pt 7):557 – 61.
Clin Chest Med 26 (2005) 197 – 205
In 1997, the Director General of the World Health multidrug-resistant tuberculosis (MDR-TB) and the
Organization (WHO) called the directly observed HIV epidemic.
treatment, short-course (DOTS) strategy ‘‘the most
important health breakthrough of the decade in terms
of the number of lives it will save’’ [1]. DOTS is
nothing new; Karel Styblo developed the essential Political and administrative commitment
principles of DOTS in the 1980s [1a]. The five
principles of the WHO-recommended DOTS strategy By any measure, the global burden of tuberculosis
[2] are: is staggering. There are nearly 9 million new cases
and 2 million deaths from tuberculosis worldwide
1. Political and administrative commitment. every year [3,4]. More than 100 million people have
2. Case detection, primarily by microscopic ex- died of tuberculosis since the tubercle bacillus was
amination of sputum of patients presenting to discovered by Koch in 1882 [5]. Cases are likely to
health facilities. increase, particularly in parts of the world with a
3. Standardized short-course chemotherapy given poorly controlled HIV epidemic. Tuberculosis dis-
under direct observation. proportionately affects young adults, impoverishes
4. Adequate supply of good-quality drugs. families, and undermines economic development [6].
5. Systematic monitoring and accountability for Despite the heavy burden of disease, tuberculosis-
every patient diagnosed. control efforts are often inadequately funded and
supported. Shortages of drugs and equipment are
This article reviews the principles, scientific basis, common, and key posts within tuberculosis-control
and experience with implementation of DOTS and programs suffer from low prestige or high turnover
discusses the relevance of DOTS in the context of and are often vacant. Physicians, in their role as
community leaders, are responsible for promoting the
existence of effective tuberculosis-control programs
and for supporting and coordinating their own efforts
with these programs. Although tuberculosis control is
This article is adapted from: Frieden TR. Directly
a core government function, it cannot be accom-
observed treatment, short course (DOTS): the strategy that
ensures cure of tuberculosis. In: Sharma SK, Mohan A, edi-
plished by any one individual or sector. It requires
tors. Tuberculosis. 2nd edition. New Delhi: Jaypee Brothers communication and collaboration among local and
Medical Publishers; 2005 [in press]; with permission. national health authorities, the primary health care
* Corresponding author. system, hospitals, medical schools, private physi-
E-mail address: tfrieden@health.nyc.gov (T.R. Frieden). cians, nongovernmental organizations, and others.
0272-5231/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccm.2005.02.001 chestmed.theclinics.com
198 frieden & munsiff
Diagnosis by sputum microscopy of patients numerous controlled clinical trials, [20 – 24] and is
attending health facilities effective for extrapulmonary tuberculosis [25,26].
Standard short-course regimens can cure more than
Two important concepts are combined in this 95% of cases of new, drug-susceptible tuberculosis
aspect of the DOTS strategy: first, diagnosis should [27]. Most recommended treatment regimens have
be based primarily on microscopy rather than on two phases: an initial intensive phase with four drugs
chest radiograph, clinical examination, or culture; and lasting 2 months, followed by a 4-month continua-
second, case finding should be done primarily among tion phase with at least two drugs. Corticosteroids
patients presenting at health facilities and not by have been shown to be useful in patients with peri-
active case finding in the community. cardial and meningeal tuberculosis and are recom-
Sputum microscopy is a highly specific test and mended in these situations [14,28]. Attempts to
should be the primary tool for diagnosis. In contrast, reduce the length of treatment to less than 6 months
the unreliability of chest radiograph is well docu- have failed, indicating that as of 2005 no practical
mented; 30% or more of patients classified as having regimen shorter than 6 months using currently avail-
active tuberculosis on the basis of chest radiograph, able medications is effective for patients with smear-
even by experts, are found not to have tuberculosis positive tuberculosis [29,30].
[7 – 10]. The acid-fast bacillus (AFB) smear also Patients previously treated for tuberculosis may
correlates with severity of disease, infectiousness, require a more intensive treatment regimen. In areas
and mortality [11,12] and is a low-cost, appropriate where drug-susceptibility testing is not available, the
technology that can be done reliably even in remote WHO re-treatment regimen should be used [14]. In
areas [13]. Decisions based on the standard WHO- areas where drug susceptibility can be determined,
recommended diagnosis algorithm [14] are often more the susceptibility results of the isolate should guide
rapid than those based on culture [15]. Technical and the treatment regimen [28].
logistic requirements make culture unsuitable as a It is now well documented that intermittent treat-
primary diagnostic tool in developing countries. ment given two or three times a week is, in general,
Serologic and amplification tests on sputum samples as effective as daily therapy. This finding should not
are currently more expensive and less informative be surprising, because Mycobacterium tuberculosis
than the AFB smear and are not of proven utility in doubles in 18 to 24 hours, compared with 12 to
global tuberculosis control; amplification techniques 20 minutes for most bacteria [31]. In fact, in one
can be useful in areas with low tuberculosis animal model, intermittent treatment was more effec-
prevalence. Future developments in this area may tive than daily treatment [32]. Supervised intermittent
further facilitate diagnosis, particularly of smear- short-course therapy for tuberculosis is shown to be
negative and extrapulmonary tuberculosis cases [16]. highly effective and extremely well tolerated in
The second component of this aspect of the DOTS patients whether or not they are HIV infected [24].
strategy is the approach to case finding. Active Intermittent treatment should be given only in a
tuberculosis case finding in the community should program of direct treatment observation [33].
not be undertaken; it usually results in low cure rates Once- or twice-weekly dosing with rifamycins in
and is of limited or no value in tuberculosis control HIV-infected tuberculosis patients with CD4 cell
[17]. Most patients with active tuberculosis seek care, counts less than 100/mm3 has been associated with
especially if care is provided free of charge, and the the development of acquired rifamycin resistance
few who do not seek care are less likely to complete [34]. The Centers for Disease Control (CDC) and
treatment [11,18]. In countries or regions with American Thoracic Society (ATS) now recommend
generalized HIV epidemics, however, programs for that patients with HIV-associated tuberculosis not be
intensified tuberculosis case finding should be im- treated with once-weekly regimens and that those
plemented in all HIV counseling and testing set- with CD4 cell counts of less than 100/mm3 not be
tings [19]. treated with any highly intermittent (ie, once- or
twice-weekly) regimens. The CDC and ATS now
recommend that such patients receive daily therapy
Standardized short-course chemotherapy given in during the intensive phase and daily or thrice-weekly
a program of directly observed treatment therapy during the continuation phase, regardless of
whether they are also receiving antiretroviral drugs.
This principle also has two aspects, both of which The international relevance of this recommendation
are crucial. The efficacy of short-course, intermittent is unclear and should be further examined; thrice-
treatment has been conclusively demonstrated in weekly regimens in the intensive phase are recom-
the dots strategy for controlling tuberculosis 199
mended by the WHO and the International Union the value of treatment success to patients and to
Against TB and Lung Disease (IUATLD) and are of their communities. It also implies recognition of
proven efficacy for tuberculosis treatment regardless the responsibility of the program and of the com-
of HIV status. If rifabutin is used, its dose may need munity to ensure successful treatment through
to be adjusted depending on the antiretroviral medi- showing respect for the patient and by providing
cations used [28,35]. treatment at convenient times and in appropriate fa-
Drugs can be effective only if they are taken [36]. cilities [55].
Virtually all clinical studies of short-course chemo- It may be impossible to arrange directly observed
therapy were conducted using directly observed ther- therapy for some patients, but directly observed
apy [37], and it can be argued that unobserved use of therapy should be possible in more than 90% of
rifampicin-containing regimens is experimental and cases in a well-functioning program. When directly
potentially dangerous. In particular, the initial phase observed treatment is impossible, the patient may
of treatment regimens that include rifampicin should need to be given self-administered treatment, in which
always be directly observed to ensure adherence and a family member may be able to assist the patient.
prevent emergence of resistance to rifampicin [37]. Assistance by a family member is unreliable, how-
Direct observation is the standard of care for tuber- ever [39,56]. There are no examples of successful
culosis in most countries [38]. About one third of large-scale DOTS programs in which immediate
patients do not take medications regularly as pre- family members have been used as primary providers
scribed, and it is not possible to predict accurately of directly observed therapy. Organizing DOTS,
which patients will not adhere to treatment [33,36, including directly observed therapy, may not be
39 – 42]. Nonadherence is not related to adverse ef- simple, but it is the only method that can ensure a
fects, dosage, or prior receipt of supervised treatment high cure rate on a program basis.
[40,43] and is as high with placebo as with active
drugs. Surprise home visits revealed a much greater
degree of nonadherence than did pill counts or urine Adequate supply of good-quality drugs
tests, despite ongoing efforts to obtain and maintain
the cooperation of patients and their families during Because long-term treatment with a combination
the full course of chemotherapy [44]. of drugs is required [57], it is important that sufficient
Directly observed therapy is the most difficult and supplies of all necessary antituberculosis drugs are
most controversial aspect of the DOTS strategy. available so that patients can complete the prescribed
(DOTS is the comprehensive five-point tuberculosis- treatment. In addition, it is important that drugs
control strategy; directly observed therapy is one are of good quality, with adequate bioavailability. Of
essential component of that strategy.) Observation particular concern are the bioavailability of rifampicin
must be done by a person who is accessible and in combination tablets, particularly when combined
acceptable to the patient and who is accountable to with pyrazinamide, and the stability of ethambutol,
the health system [14]. In some countries, directly which may be compromised by poor-quality pack-
observed therapy is given in hospital for the first aging or excessive humidity during storage.
2 months [45]. In other countries, health staff Combination tablets of proven bioavailability
[46 – 48], community volunteers [49,50], members have the theoretical advantage of preventing mono-
of nongovernmental organizations [51], religious therapy. Fixed-dose combination (FDC) tablets incor-
leaders [52], and combinations of health staff and porate two or more medications into the same tablet
a broad cross-section of community workers or vol- and prevent providers and patients from using a
unteers [53] have given directly observed therapy. single antituberculosis drug. This precaution should
Each community has particular leaders, and the reduce the likelihood of development of drug
challenge in implementing this aspect of the DOTS resistance and the possibility of physician prescrip-
strategy is to identify and enlist the support of tion error or patient medication error. FDCs can
these leaders. Even in the unstable environment of also simplify treatment for patients and logistics for
a refugee camp, directly observed therapy has been program managers. FDCs of low bioavailability
shown to be feasible [54]. could result in treatment failure and drug resistance,
Directly observed therapy involves more than however, and FDCs generally have a shorter shelf life
merely watching patients as they take medications. and increase the cost of antituberculosis drugs [14].
Rather, direct observation succeeds by building a The benefits of FDCs on a program basis are difficult
human bond between the patient and the health care to document, and, of course, FDCs still cannot ensure
worker or community volunteer and acknowledging that the drugs are taken.
200 frieden & munsiff
Where resources permit (eg, targets for case detection from tuberculosis exposure to disease was 3 months,
and treatment success are met and resources are in and in some outbreaks more than one third of
place to continue DOTS implementation), treatment exposed patients developed tuberculosis [85].
of MDR-TB can save lives and prevent further spread The HIV epidemic is a stress test for tuberculosis-
of disease. control programs and can relentlessly reveal program
There were an estimated 273,000 cases of MDR- weaknesses. The increase in tuberculosis incidence in
TB worldwide (3.1% of all tuberculosis cases) in Africa is strongly associated with the prevalence of
2000. Most MDR-TB is estimated to be concentrated HIV infection [86]. Tuberculosis case rates increased
in 10 countries [76]. These data may underestimate approximately twice as fast in countries with high
the number of MDR-TB cases worldwide, however, HIV infection rates, but the increase was lower in
because MDR-TB has been found in most regions of countries with good-quality tuberculosis-control ser-
the world that have been surveyed [77]. Systematic vices. Improving the quality of national tuberculosis
surveys of several areas of the world with high programs can mitigate HIV-associated tuberculosis
tuberculosis incidence, such as most parts of sub- [87]. Where prevalence of HIV infection is high,
Saharan Africa and most of Asia, have not yet been tuberculosis treatment alone cannot reverse the rise in
done. Resources need to be directed at areas with tuberculosis incidence. At present, the most effective
the highest burden of this disease. There is limited way to address HIV-associated tuberculosis is
experience treating MDR-TB in resource-limited through a sound DOTS program coupled with com-
settings, but successful programs can be imple- prehensive, effective HIV prevention and care that
mented [78]. International collaboration and re- incorporates active case finding in settings where
sources probably will be required to implement HIV counseling and testing are offered [19,88 – 92].
such programs [79]. In settings with high incidences of HIV and
Although MDR-TB can be successfully treated in tuberculosis, diagnosing active tuberculosis in HIV-
a variety of settings, including community-based infected persons is challenging. Sputum smear-
treatment by trained community health workers negative and extrapulmonary tuberculosis are more
[78], treatment should be given only under a program common in HIV-infected persons [90,93]. In such
of directly observed therapy in conjunction with settings, consideration should be given to using ra-
clinicians who are expert in the management of diographs and, if feasible, mycobacterial cultures to
MDR-TB. All attempts should be made to obtain assist the clinician in diagnosing tuberculosis if spu-
susceptibility results so that regimens can be tailored tum smears are negative.
for individual patients. Most studies of MDR-TB Recommended treatment regimens are similar for
treatment have used individualized regimens [80]. all tuberculosis patients, whether HIV-infected or
Patients must be treated with a regimen of at least uninfected [27], and patients with HIV respond well
three to five antituberculosis medications to which to standard antituberculosis treatment and do not need
the strain is known or likely to be susceptible, in- additional drugs [90,94,95]. Death during antituber-
cluding an injectable agent and a fluoroquinolone culosis treatment is more common in HIV-infected
[14,81]. Although longer use of injectable medica- individuals but is primarily from non – tuberculosis-
tions is associated with significant adverse effects, associated causes [33]. Patients with HIV infection
longer-term injectable therapy increases culture con- and tuberculosis survive longer if given short-course
version and survival rates in persons with MDR-TB chemotherapy with rifampicin-containing regimens
[82]. Intermittent regimens should not be used. The than if given non – rifamycin-containing regimens
drugs for treating MDR-TB are much more expensive [96] and longer still if short-course chemotherapy is
and have markedly more adverse effects than stan- given in a program of directly observed therapy
dard drugs [83]. The duration of treatment is 18 to [97,98]. In addition, antituberculosis regimens con-
24 months; patients need to be monitored carefully, taining rifampicin can be administered safely with
because the risk of default can be high. several effective, highly active antiretroviral drugs,
making it easier to deliver simultaneous treatment for
tuberculosis and HIV [35,90,99]. Rifabutin, a more
HIV expensive alternative to rifampicin, can be used with
most antiretroviral regimens with some dose adjust-
Infection with HIV is the most potent known risk ments [35].
factor for progression to active tuberculosis among Tanzania and Malawi have had DOTS programs
adults [84]. In outbreaks of tuberculosis in wards for for more than 10 years. Despite high rates of HIV
AIDS patients in the United States, the median time infection (which is present in more than 60% of tu-
202 frieden & munsiff
berculosis patients in some sub-Saharan African [9] Koppaka R, Bock N. How reliable is chest radiog-
countries, including Malawi, Zimbabwe, Ethiopia, raphy? In: Frieden T, editor. Toman’s tuberculosis: case
and Botswana [3]), there has been no evidence of detection, treatment, and monitoring—questions and
answers. 2nd edition. Geneva (Switzerland)7 World
an increased rate of relapse among HIV-infected
Health Organization; 2004. p. 51 – 60.
patients, and rates of drug resistance remain low
[10] Harries A. What are the relative merits of chest
[66,67]. Avoiding the creation of drug-resistant tu- radiography and sputum examination (smear micros-
berculosis is the best way to prevent the spread of copy and culture) in case detection among new outpa-
such strains. tients with prolonged chest symptoms? In: Frieden T,
editor. Toman’s tuberculosis: case detection, treatment,
and monitoring—questions and answers. 2nd edition.
Summary Geneva (Switzerland)7 World Health Organization;
2004. p. 61 – 5.
[11] Styblo K. Epidemiology of tuberculosis. The Hague
The technology to control tuberculosis has been (The Netherlands)7 Royal Netherlands Tuberculosis
known for decades. The DOTS strategy can double Association; 1991.
cure rates. Successful global implementation of [12] Rouillon A, Perdrizet S, Parrot R. Transmission of
DOTS could save millions of lives over the next tubercle bacilli: the effects of chemotherapy. Tubercle
10 years. Success will require effective governmental 1976;57(4):275 – 99.
programs along with active communication, collabo- [13] Shimao T. Tuberculosis case-finding. WHO/TB/
ration, and participation on the part of the health 82.131. Geneva (Switzerland)7 World Health Organi-
care system and governmental and nongovernmen- zation; 1982.
tal sectors. [14] Treatment of tuberculosis: guidelines for national pro-
grammes. 3rd edition. WHO/CDS/TB/2003.313. Geneva
(Switzerland)7 World Health Organization; 2003.
[15] Urbanczik R. Present position of microscopy and of
References culture in diagnostic mycobacteriology. Zentralbl
Bakteriol Mikrobiol Hyg [A] 1985;260(1):81 – 7.
[1] WHO calls for immediate use of new tuberculosis [16] American Thoracic Society Workshop. Rapid diagnos-
breakthrough [press release WHO/24, 1997]. Geneva tic tests for tuberculosis: what is the appropriate use?
(Switzerland)7 World Health Organization; March Am J Respir Crit Care Med 1997;155(5):1804 – 14.
21, 1997. [17] Rieder H. What is the role of case detection by peri-
[1a] Murray CJ, DeJonghe E, Chum HJ, et al. Cost effec- odic mass radiographic examination in tuberculosis
tiveness of chemotherapy for pulmonary tuberculosis control? In: Frieden T, editor. Toman’s tuberculosis:
in three sub-Saharan African countries. Lancet 1991; case detection, treatment, and monitoring—questions
338(8778):1305 – 8. and answers. 2nd edition. Geneva (Switzerland)7 World
[2] An expanded DOTS framework for effective tu- Health Organization; 2004. p. 72 – 9.
berculosis control. WHO/CDS/TB/2002.297. Geneva [18] Banerji D, Anderson S. A sociological study of
(Switzerland)7 World Health Organization Global Tu- awareness of symptoms among persons with pulmo-
berculosis Programme; 2002. nary tuberculosis. Bull World Health Organ 1963;
[3] Global tuberculosis control: surveillance, planning, 26:665 – 83.
financing. WHO Report 2005. WHO/HTM/TB/ [19] Interim policy on collaborative TB/HIV activities.
2005.49. Geneva (Switzerland)7 World Health Orga- WHO/HTM/TB/2004. Geneva (Switzerland)7 World
nization; 2005. Health Organization; 2004.
[4] The world health report 2004: changing history. Ge- [20] Mitchison DA, Nunn AJ. Influence of initial drug
neva (Switzerland)7 World Health Organization; 2004. resistance on the response to short-course chemo-
[5] TB deaths reach historic levels [press release WHO/22, therapy of pulmonary tuberculosis. Am Rev Respir Dis
1996]. Geneva (Switzerland)7 World Health Organiza- 1986;133(3):423 – 30.
tion; March 21, 1996. [21] Hong Kong Chest Service/British Medical Research
[6] Groups at risk: WHO report on the tuberculosis Council. Five-year follow-up of a controlled trial of
epidemic 1996. WHO/TB/96.198. Geneva (Switzer- five 6-month regimens of chemotherapy for pulmo-
land)7 World Health Organization; 1996. nary tuberculosis. Am Rev Respir Dis 1987;136(6):
[7] Newell RR, Chamberlain WE, Rigler L. Descriptive 1339 – 42.
classification of pulmonary shadows: a revelation of [22] Singapore Tuberculosis Service/British Medical
unreliability in the roentgenographic diagnosis of Research Council. Five-year follow-up of a clinical
tuberculosis. Am Rev Tuberc 1954;69(4):566 – 84. trial of three 6-month regimens of chemotherapy given
[8] Garland LH. Studies on the accuracy of diagnostic intermittently in the continuation phase in the treat-
procedures. Am J Roentgenol Radium Ther Nucl Med ment of pulmonary tuberculosis. Am Rev Respir Dis
1959;82(1):25 – 38. 1988;137(5):1147 – 50.
the dots strategy for controlling tuberculosis 203
[23] Cohn DL, Catlin BJ, Peterson KL, et al. A 62-dose, [36] Grzybowski S. Drugs are not enough: failure of short-
6-month therapy for pulmonary and extrapulmonary course chemotherapy in a district in India. Tuber Lung
tuberculosis: a twice-weekly, directly observed, and Dis 1993;74(3):145 – 6.
cost-effective regimen. Ann Intern Med 1990; [37] Mitchison DA. Basic mechanisms of chemotherapy.
112(6):407 – 15. Chest 1979;76(6 Suppl):771 – 81.
[24] Chaisson RE, Clermont HC, Holt EA, et al. Six-month [38] Iseman MD, Cohn DL, Sbarbaro JA. Directly observed
supervised intermittent tuberculosis therapy in Haitian treatment of tuberculosis: we can’t afford not to try it.
patients with and without HIV infection. Am J Respir N Engl J Med 1993;328(8):576 – 8.
Crit Care Med 1996;154(4 Pt 1):1034 – 8. [39] Fox W. Compliance of patients and physicians:
[25] Balasubramanian R, Ramachandran R. Management of experience and lessons from tuberculosis I. BMJ
non-pulmonary forms of tuberculosis: review of TRC 1993;287(6384):33 – 5.
studies over two decades. Indian J Pediatr 2000; [40] Sbarbaro JA. The patient-physician relationship: com-
67(2 Suppl):34 – 40. pliance revisited. Ann Allergy 1990;64(4):325 – 31.
[26] Yuen AP, Wong SH, Tam CM, et al. Prospective [41] Haynes RB, McKibbon KA, Kanani R. Systematic
randomized study of thrice weekly six-month and review of randomized trials of interventions to assist
nine-month chemotherapy for cervical tuberculous patients to follow prescriptions for medications [pub-
lymphadenopathy. Otolaryngol Head Neck Surg 1997; lished erratum in Lancet 1997;349(9059):1180]. Lan-
116(2):189 – 92. cet 1996;348(9024):383 – 6.
[27] Frieden TR, Sterling TR, Munsiff SS, et al. Tuber- [42] Pablos-Mendez A, Knirsch CA, Barr RG, et al. Non-
culosis. Lancet 2003;362(9387):887 – 99. adherence in tuberculosis treatment: predictors and
[28] American Thoracic Society/Centers for Disease Con- consequences in New York City. Am J Med 1997;
trol and Prevention/Infectious Diseases Society of 102(2):164 – 70.
America. Treatment of tuberculosis. Am J Respir Crit [43] Fox W. The problem of self-administration of drugs,
Care Med 2003;167(4):603 – 62. with particular reference to pulmonary tuberculosis.
[29] Balasubramanian R, Sivasubramanian S, Vijayan VK, Tubercle 1958;39(5):269 – 74.
et al. Five-year results of a 3-month and two 5-month [44] Fox W. Self-administration of medicaments. A review
regimens for the treatment of sputum-positive pulmo- of published work and a study of the problems. Bull Int
nary tuberculosis in South India. Tubercle 1990; Union Tuberc 1962;32:307 – 31.
71(4):253 – 8. [45] Murray CJ, DeJonghe E, Chum HJ, et al. Cost-
[30] Singapore Tuberculosis Service/British Medical Re- effectiveness of chemotherapy for pulmonary tuber-
search Council. Clinical trial of six-month and four- culosis in three sub-Saharan African countries. Lancet
month regimens of chemotherapy in the treatment of 1991;338(8778):1305 – 8.
pulmonary tuberculosis. Am Rev Respir Dis 1979; [46] Heldal E, Cruz JR, Arnadottir T, et al. Successful
119(4):579 – 85. management of a national tuberculosis programme
[31] North RJ, Izzo AA. Mycobacterial virulence. Virulent under conditions of war. Int J Tuberc Lung Dis 1997;
strains of Mycobacteria tuberculosis have faster in 1(1):16 – 24.
vivo doubling times and are better equipped to resist [47] Zhang LX, Kan GQ. Tuberculosis control programme
growth-inhibiting functions of macrophages in the in Beijing. Tuber Lung Dis 1992;73(3):162 – 6.
presence and absence of specific immunity. J Exp [48] Fujiwara PI, Larkin C, Frieden TR. Directly observed
Med 1993;177(6):1723 – 33. therapy in New York City: history, implementation,
[32] Mitchison DA, Dickinson JM. Laboratory aspects of results, and challenges. Clin Chest Med 1997;18(1):
intermittent drug therapy. Postgrad Med J 1971; 135 – 48.
47(553):737 – 41. [49] Wilkinson D. High-compliance tuberculosis treatment
[33] Datta M, Radhmani MP, Selvaraj R, et al. Critical programme in a rural community. Lancet 1994;
assessment of smear-positive pulmonary tuberculosis 343(8898):647 – 8.
patients after chemotherapy under the district tuber- [50] Wilkinson D, Davies GR. Coping with Africa’s
culosis programme. Tuber Lung Dis 1993;74(3): increasing tuberculosis burden: are community super-
180 – 6. visors an essential component of the DOT strategy?
[34] Warning to readers. Acquired rifamycin resistance in Trop Med Int Health 1997;2(7):700 – 4.
persons with advanced HIV disease being treated for [51] Chowdhury AM, Chowdhury S, Islam MN, et al. Con-
active tuberculosis with intermittent rifamycin-based trol of tuberculosis by community health workers in
regimens. MMWR Morb Mortal Wkly Rep 2002; Bangladesh [published erratum in Lancet 1997;
51(10):214 – 5. 350(9084):1108]. Lancet 1997;350(9072):169 – 72.
[35] Updated guidelines for the use of rifamycins for the [52] Manalo F, Tan F, Sbarbaro JA, et al. Community-based
treatment of tuberculosis among HIV-infected patients short-course treatment of pulmonary tuberculosis in a
taking protease inhibitors or nonnucleoside reverse developing nation. Initial report of an eight-month,
transcriptase inhibitors. January 20, 2004. Available at: largely intermittent regimen in a population with a high
http://www.cdc.gov/nchstp/tb/tb_hiv_drugs/toc.htm. prevalence of drug resistance. Am Rev Respir Dis
Accessed January 28, 2005. 1990;142(6 Pt 1):1301 – 5.
204 frieden & munsiff
[53] Khatri GR, Frieden TR. Controlling tuberculosis in [71] Ledru S, Cauchoix B, Yameogo M, et al. Impact of
India. N Engl J Med 2002;347(18):1420 – 5. short-course therapy on tuberculosis drug resistance in
[54] Miles SH, Maat RB. A successful supervised out- South-West Burkina Faso. Tuber Lung Dis 1996;
patient short-course tuberculosis treatment program in 77(5):429 – 36.
an open refugee camp on the Thai-Cambodian border. [72] Zhang LX, Kan GQ, Tu DH, et al. Trend of initial drug
Am Rev Respir Dis 1984;130(5):827 – 30. resistance of tubercle bacilli isolated from new patients
[55] Frieden TR, Driver CR. Tuberculosis control: past with pulmonary tuberculosis and its correlation with
10 years and future progress. Tuberculosis (Edinb) the tuberculosis programme in Beijing. Tuber Lung
2003;83(1 – 3):82 – 5. Dis 1995;76(2):100 – 3.
[56] Frieden T, Sbarbaro JA. The slippery slope to sloppy [73] Weis SE, Slocum PC, Blais FX, et al. The effect of
DOTS. Int J Tuberc Lung Dis 2002;6(5):371 – 2. directly observed therapy on the rates of drug
[57] Fox W, Ellard GA, Mitchison DA. Studies on the treat- resistance and relapse in tuberculosis. N Engl J Med
ment of tuberculosis undertaken by the British Medi- 1994;330(17):1179 – 84.
cal Research Council tuberculosis units, 1946 – 1986, [74] Weyer K, Kleeberg HH. Primary and acquired drug
with relevant subsequent publications. Int J Tuberc resistance in adult black patients with tuberculosis in
Lung Dis 1999;3(10 Suppl 2):231 – 79. South Africa: results of a continuous national drug
[58] Chaulk CP, Moore-Rice K, Rizzo R, et al. Eleven years resistance surveillance programme involvement. Tuber
of community-based directly observed therapy for Lung Dis 1992;73(2):106 – 12.
tuberculosis. JAMA 1995;274(12):945 – 51. [75] Sterling TR, Lehmann HP, Frieden TR. Impact of
[59] Frieden TR, Sterling T, Pablos-Mendez A, et al. The DOTS compared with DOTS-plus on multidrug
emergence of drug resistant tuberculosis in New York resistant tuberculosis and tuberculosis deaths: decision
City [published erratum in N Engl J Med 1993;329(2): analysis. BMJ 2003;326(7389):574 – 9.
148]. N Engl J Med 1993;328(8):521 – 6. [76] Dye C, Espinal MA, Watt CJ, et al. Worldwide
[60] Frieden TR, Fujiwara PI, Washko RM, et al. Tuber- incidence of multidrug-resistant tuberculosis. J Infect
culosis in New York City—turning the tide. N Engl J Dis 2002;185(8):1197 – 202.
Med 1995;333(4):229 – 33. [77] Anti-tuberculosis drug resistance in the world: report
[61] Gonzalez E, Armas L, Alonso A. Tuberculosis in the no. 3, prevalence and trends. WHO/HTM/TB/2004.
Republic of Cuba: its possible elimination. Tuber Lung 343. Geneva (Switzerland)7 World Health Organiza-
Dis 1994;75(3):188 – 94. tion; 2004.
[62] Suárez P, Watt CJ, Alarcon E, et al. The dynamics of [78] Mitnick C, Bayona J, Palacios E, et al. Community-
tuberculosis in response to 10 years of intensive based therapy for multidrug-resistant tuberculosis in
control effort in Peru. J Infect Dis 2001;184(4):473 – 8. Lima, Peru. N Engl J Med 2003;348(2):119 – 28.
[63] China Tuberculosis Control Collaboration. The effect [79] Guidelines for establishing DOTS-PLUS pilot projects
of tuberculosis control in China. Lancet 2004; for the management of multidrug-resistant tuberculosis
364(9432):417 – 22. (MDRTB). WHO/CDS/TB/2000.279. Geneva
[64] Sawert H, Kongsin S, Payanandana V, et al. Costs and (Switzerland)7 World Health Organization; 2000.
benefits of improving tuberculosis control: the case of [80] Mukherjee JS, Rich ML, Socci AR, et al. Programmes
Thailand. Soc Sci Med 1997;44(12):1805 – 16. and principles in treatment of multidrug-resistant
[65] Crofton J, Chaulet P, Maher D. Guidelines for the tuberculosis. Lancet 2004;363(9407):474 – 81.
management of drug-resistant tuberculosis. WHO/TB/ [81] Thorpe LE, Frieden TR, Laserson KF, et al. Season-
96.210. Geneva (Switzerland)7 World Health Organi- ality of tuberculosis in India: is it real and what does it
zation; 1996. tell us? Lancet 2004;364(9445):1613 – 4.
[66] Chum HJ, O’Brien RJ, Chonde TM, et al. An epi- [82] Frieden TR, Sherman LF, Maw KL, et al. A multi-
demiological study of tuberculosis and HIV infection institutional outbreak of highly drug-resistant tuber-
in Tanzania, 1991 – 1993. AIDS 1996;10(3):299 – 309. culosis: epidemiology and clinical outcomes. JAMA
[67] Glynn JR, Jenkins PA, Fine PE, et al. Patterns of 1996;276(15):1229 – 35.
initial and acquired antituberculosis drug resistance [83] Nathanson E, Gupta R, Huamani P, et al. Adverse
in Karonga District, Malawi. Lancet 1995;345(8954): events in the treatment of multidrug-resistant tuber-
907 – 10. culosis: results from the DOTS plus initiative. Int J
[68] Crofton J. The contribution of chemotherapy to the Tuberc Lung Dis 2004;8(11):1382 – 4.
control of tuberculosis as a community disease. Thorax [84] Selwyn PA, Hartel D, Lewis VA, et al. A prospective
1971;20(1):49 – 54. study of the risk of tuberculosis among intravenous
[69] Khalil A, Sathianathan S. Impact of anti-tuberculosis drug users with human immunodeficiency virus
legislation in Libya on the prevalence of primary and infection. N Engl J Med 1989;320(9):545 – 50.
acquired resistance to the three main drugs at a major [85] Dooley SW, Villarino ME, Lawrence M, et al.
tuberculosis centre. Tubercle 1978;59(1):1 – 12. Nosocomial transmission of tuberculosis in a hospi-
[70] Kim SJ, Bai GH, Hong YP. Drug-resistant tuber- tal unit for HIV-infected patients. JAMA 1992;
culosis in Korea, 1994. Int J Tuber Lung Dis 1997; 267(19):2632 – 4.
1(4):302 – 8. [86] Corbett EL, Steketee RW, ter Kuile FO, et al.
the dots strategy for controlling tuberculosis 205
HIV-1/AIDS and the control of other infectious dis- tation and management. Pharmacol Ther 1997;73(1):
eases in Africa [published erratum in Lancet 2002; 1 – 50.
360(9340):1178]. Lancet 2002;359(9324):2177 – 87. [94] Small PM, Schecter GF, Goodman PC, et al. Treatment
[87] Cantwell MF, Binkin NJ. Tuberculosis in sub-Saharan of tuberculosis in patients with advanced human
Africa: a regional assessment of the impact of the immunodeficiency virus infection. N Engl J Med
human immunodeficiency virus and national tuber- 1991;324(5):289 – 94.
culosis control program quality. Tuber Lung Dis 1996; [95] Perriens JH, St Louis ME, Mukadi YB, et al.
77(3):220 – 6. Pulmonary tuberculosis in HIV-infected patients in
[88] Guidelines for implementing collaborative TB and Zaire: a controlled trial of treatment for either 6 or
HIV programme activities. WHO/CDS/TB/2003. 12 months. N Engl J Med 1995;332(12):779 – 84.
319. Geneva (Switzerland)7 World Health Organiza- [96] Okwera A, Whalen C, Byekwaso F, et al. Randomized
tion; 2003. trial of thiacetazone and rifampicin-containing regi-
[89] Guidelines for HIV surveillance among tuberculosis mens for pulmonary tuberculosis in HIV-infected
patients. 2nd edition. WHO/HTM/TB/2004.339. Ge- Ugandans. Lancet 1994;344(8933):1323 – 8.
neva (Switzerland)7 World Health Organization; 2004. [97] Alwood K, Keruly J, Moore-Rice K, et al. Effective-
[90] TB/HIV. A clinical manual. 2nd edition. WHO/HTM/ ness of supervised, intermittent therapy for tuberculosis
TB/2004.329. Geneva (Switzerland)7 World Health Or- in HIV-infected patients. AIDS 1994;8(8):1103 – 8.
ganization; 2004. [98] Alpert PL, Munsiff SS, Gourevitch MN, et al. A
[91] Jones JL, Hanson DL, Dworkin MS, et al. HIV- prospective study of tuberculosis and human immuno-
associated tuberculosis in the era of highly active deficiency virus infection: clinical manifestations and
antiretroviral therapy. Int J Tuberc Lung Dis 2000; factors associated with survival. Clin Infect Dis
4(11):1026 – 31. 1997;24(4):661 – 8.
[92] Badri M, Wilson D, Wood R. Effect of highly active [99] Department of Health and Human Services. Guide-
antiretroviral therapy on incidence of tuberculosis in lines for the use of antiretroviral agents in HIV-1
South Africa: a cohort study. Lancet 2002;359(9323): infected adults and adolescents. October 29, 2004.
2059 – 64. Available at: http://aidsinfo.nih.gov/guidelines/adult/
[93] Harries AD. Tuberculosis in Africa: clinical presen- AA_102904. pdf. Accessed January 28, 2005.
Clin Chest Med 26 (2005) 341 – 347
Nearly 2 billion people, one third of the world’s of isoniazid, rifampin, pyrazinamide, and ethambutol.
population, are infected with tuberculosis (TB). Eight Patients find it difficult to adhere to this complicated
million of these people develop active TB every year, regimen, leading to an increase in drug-resistant
leading to 2 million deaths annually [1]. Although strains [3]. Second-line drugs for multidrug-resistant
an effective vaccine against active disease would be tuberculosis (MDR-TB) are expensive and more toxic
key in ultimately eradicating TB as a public health than the standard treatment.
problem, there is, as yet, no satisfactory vaccine for It has been more than 30 years since a novel TB
the disease. The commonly used bacille Calmette- drug has been introduced into clinical practice. The
Guérin vaccine is reasonably effective when admin- estimated costs of discovery and development of a
istered to infants but is much less effective against new TB drug (including the costs of failure) are
the pulmonary form of TB in adults. Despite a re- between $115 million and $240 million [4]. Adopting
cent increase in vaccine research, ongoing efforts a portfolio approach, the Global Alliance for TB
to develop a new, more effective vaccine are not Drug Development (TB Alliance) expects to pursue
expected to yield results for at least a decade, and an multiple candidate drugs. The global TB market
effective vaccine would not substantially diminish the could reach $700 million by 2010 but is concentrated
TB epidemic for several decades after that. Both in developing countries, and the pharmaceutical in-
effective drugs and an effective vaccine will probably dustry, guided by the perception that the TB market
be needed for many decades to come, until the pool remains too small and diverse to guarantee return on
of current TB patients and latently infected individu- investment, has not driven a comprehensive devel-
als is eliminated. Drugs that improve or shorten opment program for new TB drugs.
therapy, in contrast, would affect treatment of current
patients almost immediately upon their adoption.
The World Health Organization’s directly ob- The Global Alliance for Tuberculosis Drug
served treatment, short course (DOTS) strategy is an Development
effective TB control strategy but reaches only one
third of the people who need it [2]. Moreover, it is In response to these challenges, and with initial
expensive, labor-intensive, and usually involves a 6- to support from the Rockefeller Foundation and other
12-month, four-drug combination regimen consisting international organizations, the TB Alliance [5] was
established in 2000 with a mission to accelerate the
discovery and development of new drugs to fight TB
The views expressed by the authors do not necessarily and to ensure their affordability and accessibility. The
reflect the views of their institutions. initiative was born at a pivotal meeting in Cape
* Corresponding author. Town, South Africa [6], and soon gained substantial
E-mail address: cgardner@rockfound.org (C.A. Gardner). support from the Bill and Melinda Gates Foundation.
0272-5231/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccm.2005.02.008 chestmed.theclinics.com
342 gardner et al
As one of a new breed of nonprofit product- fined and measurable milestones, and clear go/no-go
development public – private partnerships (PD PPPs), decision points. So far, the TB Alliance has as-
the TB Alliance draws on strengths of the public and sembled a portfolio of projects that are in various
private sector to achieve its objectives—new and phases—lead identification, lead optimization, and
better TB treatments that preclinical development (Fig. 1).
Since 2000, the TB Alliance has catalyzed the
Shorten or significantly simplify the treatment expansion of the TB drug pipeline. In 2005, up to
of active tuberculosis seven compounds are expected to be in or to enter
Provide effective treatment of MDR-TB clinical trials, many of them with novel modes of
Improve the treatment of latent tuberculosis action. For comparison, in 2000 only two TB candi-
date drugs were in development, and these products
The TB Alliance operates like a nonprofit com- represented only slight modifications to existing
pany based on an innovative partnership designed to molecules. The sequencing of the genome of
share risks and incentives. The range of organizations Mycobacterium tuberculosis is expected to yield
that partner with the TB Alliance includes academic useful results for drug discovery, although not before
institutions, government research laboratories and substantial investment by the global research com-
public health institutions, nongovernmental organi- munity. Examples of compound families reviewed
zations, the pharmaceutical industry, and contract by the TB Alliance for drug candidacy over the last
research organizations worldwide. Its headquarters several years include
are in New York, but it also maintains offices in Cape
Town, South Africa, and Brussels, Belgium. Nitroimidazoles, with potentially sterilizing ac-
tivity and effectiveness against MDR-TB
Fluoroquinolones and quinolizine derivatives
that may hold the key to significantly shorter
The strategy of the Global Alliance for treatment regimens
Tuberculosis Drug Development Longer-acting rifamycins for more widely
spaced intermittent treatment that have a poten-
The TB Alliance has five strategic objectives: tial for avoiding cross-resistance to other rifa-
mycins and antiretroviral drug interactions
1. Identify and access promising compounds Macrolides, an excellent antibiotic class with
2. Oversee the preclinical development of candi- significant potential to yield a tuberculosis drug
date drugs because of its excellent pharmacologic features
3. Spearhead clinical trials and drive regula- and its promising antibacterial activity against
tory approval M. tuberculosis
4. Ensure affordability, adoption, and access Oxazolidinones, broad-spectrum antimicrobials,
(AAA strategy) which may have substantial antimycobacte-
5. Mobilize expertise and resources for tuber- rial activity;
culosis drug development Pleuromutilins, a novel class of antibiotics of a
natural product origin, which have tuberculosis
These objectives and selected accomplishments efficacy in early tests, do not have cross-
for each objective are outlined below [5]. resistance with other antibiotics, and seem to
produce resistance very slowly
Objective 1: identify and access promising Pyrroles, a previously largely unexplored class
compounds of compounds with antimycobacterial activity in
vitro and in preclinical animal models, with a
Through proactive business development and novel mechanism of action.
periodic requests for proposals from public, private,
and academic institutions, the TB Alliance selects, Selected lead compounds in the Global Alliance for
assembles, and manages a portfolio of promising Tuberculosis Drug Development portfolio
candidate drugs. The TB Alliance collaborates with A lead compound in the portfolio, the nitro-
or outsources (and provides funding for) the devel- imidazopyran PA-824, has successfully passed estab-
opment of these candidates to public and private lished milestones and is expected to be in clinical
partners but retains overall responsibility for these trials by the first half of 2005. PA-824 was the first
projects, with dedicated project management, prede- compound acquired by the TB Alliance, in June
global alliance for tuberculosis drug development 343
Nitroimidazole Analogs
Nitroimidazole PA-824 Moxifloxacin
(TB Alliance, Novartis Institute for Tropical Diseases,
(TB Alliance, Chiron) (TB Alliance, Bayer Pharmaceuticals)
NIAID)
Compounds, Analogs and Derivatives
Fig. 1. Global tuberculosis drug pipeline, March 2005. (Courtesy of the Global Alliance for TB Drug Development, New York,
NY; with permission.)
2002, through an exclusive license agreement with of rifampicin. Developed by Bayer AG (Leverkusen,
Chiron Corporation (Emeryville, California). The Germany), the drug currently has been approved by
compound has bactericidal activity similar to isonia- the Food and Drug Administration (FDA) for treat-
zid and sterilizing activity that rivals that of rifam- ment of skin and upper respiratory tract infections
picin. During the discovery stage, PA-824 and its and pneumonia. In vivo experiments using a murine
analogues demonstrated activity against both drug- model at the Center for Tuberculosis Research at
sensitive and MDR strains of TB. Preclinical develop- Johns Hopkins University and funded by the TB
ment results so far have demonstrated the feasibility Alliance have affirmed moxifloxacin’s early promise
of increasing synthesis for animal and clinical trials. for shortening therapy. Support from the TB Alliance
In preliminary toxicology testing, PA-824 did not helps ensure that this murine model, one of the TB
demonstrate teratogenic or toxic effects on normal Alliance’s platform technologies, will continue to be
metabolic or hormonal systems. Additional animal available for the development of other TB drugs. The
studies to assess the safety and efficacy of PA-824 Johns Hopkins team substituted moxifloxacin in
have also yielded encouraging results. As a result, various combinations to replace or enhance aspects
PA-824 will be entering phase I clinical trials in the of existing treatment. Substitution of moxifloxacin
first half of 2005. The TB Alliance is also working to for isoniazid significantly increases efficacy, and
optimize the synthesis of PA-824 to reduce production studies have suggested that inclusion of moxifloxacin
costs. To ensure further development of this promising in a combination regimen can shorten the time of TB
class of drug candidates, the TB Alliance is pursuing treatment [7].
research into analogues of PA-824. To this end, the The next step is to confirm these results in clinical
Alliance has also launched a partnership with Novartis trials. To that end, the TB Alliance is working with
Institute of Tropical Diseases in Singapore to explore Bayer and several partners, including the Centers for
new avenues within the nitroimidazole class, working Disease Control and Prevention (CDC), Johns Hop-
closely with the National Institutes of Allergy and kins University, and University College, London
Infectious Diseases (NIAID), among others. (with funding from the European and Developing
Moxifloxacin is a quinolone that has shown high Countries Clinical Trials Program), to formulate and
levels of activity against M. tuberculosis in vitro and support a global clinical development plan for
has generated significant interest from the TB moxifloxacin in the treatment of TB. This under-
community for its potential to become the first major taking would build on an existing Cooperative
advance in TB therapy since the 1965 introduction Research and Development Agreement between
344 gardner et al
Bayer AG and the CDC that the TB Alliance a diverse portfolio and working with countries to
facilitated in 2002. As part of this plan, the CDC build capacity for drug development. For instance,
TB Trials Consortium is conducting phase II clinical the TB Alliance is working with the International
trials in Spain, North America, and Africa to evalu- Union Against Tuberculosis and Lung Disease to
ate the safety and efficacy of a regimen in which build clinical trials capabilities at international loca-
moxifloxacin is substituted for ethambutol in the tions in Africa, South America, and Asia. The TB
treatment of patients who have newly diagnosed TB. Alliance is exploring multiple approaches to stream-
lining clinical development, including simultaneous
Objective 2: oversee the preclinical development of rather than sequential testing of new compounds and
drug candidates validation of novel surrogate markers.
The TB Alliance has established an effective Objective 4: ensure affordability, adoption, and
process and enabling technologies to develop can- access
didate drugs. It has developed a comprehensive
management plan, created a diverse network of In developing countries, the TB Alliance is
outsourced, global expertise, and established go/no- working to ensure the AAA strategy for the products
go decision gates to advance candidates along the it develops. Affordability of the final product depends
development pipeline. on both technical and business factors. The costs of
A research plan and development timeline is production and development of potential compounds
established for each compound or project selected are evaluated, and all agreements are structured to
for the portfolio. The development plan requires a limit royalties in countries where TB is endemic and
series of tests, each designed to provide a go/no-go to include licensing provisions and manufacturing
decision point for continued development. At pre- rights to keep prices low. The TB Alliance is also
clinical stages, a comprehensive program of safety working with the World Health Organization and
and toxicology pharmacologic investigations is per- national TB control programs to ensure early adop-
formed. Because TB drugs must be designed to tion of a new product in existing programs and thera-
optimize combination therapy, the cornerstone of peutic regimens. Access ensures that medicines reach
resistance control, these safety considerations must all patients, particularly the poorest of the poor. As
include drug – drug interactions with companion anti- part of the AAA strategy, the TB Alliance is ex-
TB medications. Furthermore, because an increas- ploring innovative intellectual-property strategies to
ing percentage of TB patients are also infected with balance access and incentives. The TB Alliance
HIV, it is critical to determine the compatibility with hopes to retain the ability to deliver new anti-TB
antiretroviral treatments and in particular to ensure drugs equitably to those areas most in need while
the lack of induction or metabolism by cytochrome helping to maintain incentives for the pharmaceutical
P450 enzymes. industry to develop new TB medicines. The costs of
production and development of potential compounds
Objective 3: spearhead clinical trials and drive are evaluated, and all agreements are structured to
regulatory approval limit royalties in endemic countries and to include
licensing provisions and manufacturing rights to keep
The TB Alliance is working to ensure that there prices low.
is adequate capacity for clinical trials to test the com- The TB Alliance is also working with the World
pounds that successfully pass preclinical milestones. Health Organization, the Stop TB Partnership, and
These facilities must meet the highest regulatory national TB control programs to ensure early adop-
standards, including those of the FDA and European tion of a new product in existing programs and
Medicines Agency. The aim is to reduce unnecessary therapeutic regimens. Unlike PD PPPs developing
regulatory hurdles to ensure that new drugs are made novel prevention tools lacking any existing distribu-
available to patients as soon as possible. Therefore, tion mechanisms, the TB Alliance will be able to rely
the TB Alliance is also working to harmonize regu- on a decade of investments in DOTS programs and
latory approval processes to expedite the inclusion of their drug procurement and distribution mechanisms
new products in standardized treatment. such as the Global Drug Facility. Although further
As in any drug discovery and development improvements in access to DOTS are imperative,
program, general delays or setbacks are expected in these mechanisms are in place, and the TB Alliance
the process of TB drug development. The TB can rely on them for the distribution of the new drugs
Alliance hopes to reduce these risks by maintaining in combination therapy.
global alliance for tuberculosis drug development 345
Objective 5: mobilize expertise and resources for portfolio and could offer their laboratories’ preclinical
tuberculosis drug development capacity to develop the portfolio. For instance, in April
2003 the TB Alliance signed a 2-year agreement with
Following initial funding by the Rockefeller and the Korea Research Institute of Chemical Technology,
The Bill and Melinda Gates Foundations, the TB where scientists will synthesize more than 400 com-
Alliance has secured further public funding from the pounds, including novel quinolones, pyridones, and
United States Agency for International Development quinolizines. This project aims to yield up to three
and the Dutch Ministry of Development Coopera- lead candidates for the TB Alliance portfolio and is
tion and in-kind support from the NIAID. The TB the TB Alliance’s first research and development
Alliance has leveraged this funding to position itself partnership in Asia as well as the first in a country that
as the catalyst in developing new TB drugs. The TB has a significant TB burden. Patient enrollment plays
Alliance has mobilized worldwide expertise and a central role in the clinical development stage.
resources by establishing solid relationships with Because of TB’s global impact, most patients reside
pharmaceutical and biotechnology companies, proac- in developing countries, where clinical trial infra-
tively seeking new sources of support, and actively structure is limited but where reasonable cost struc-
participating in global forums to promote its mission. tures help fulfill the ultimate goal of affordability.
For example, the TB Alliance is the lead agency of Furthermore, DOTS infrastructure, expanded during
the Stop TB Partnership Working Group on TB Drug the last 10 years, offers an excellent starting point for
Development, which coordinates a forum to facilitate clinical trial capacity building.
research and development collaborations worldwide The TB Alliance is also working closely with
for new TB drugs. While building its own portfolio other PD PPPs with similar goals, such as Medicines
through the partnerships described previously, the TB for Malaria Venture, the Malaria Vaccine Initiative,
Alliance has helped catalyze the global TB drug the International AIDS Vaccine Initiative, the Interna-
pipeline by engaging all relevant industry entities and tional Partnership for Microbicides, and others, to
by enhancing technologies to support drug develop- coordinate overall activities and to strengthen the
ment in general. The TB Alliance has taken both incentive for investment in new tools to solve global
advisory and advocacy roles to support corporate health problems. The TB Alliance has welcomed the
efforts for TB, such as those currently underway creation of a new research effort, the Foundation for
at Novartis, Anacor, AstraZeneca, Lupin, Glaxo- Innovative New Diagnostics (FIND) to develop bet-
SmithKline, and other companies. For example, ter diagnostic tests for infectious diseases, with an
AstraZeneca has committed $25 million over 5 years initial emphasis on TB [10]. The TB Alliance and
at its newly established research facility in Bangalore, FIND plan to work together to encourage the
India [8], and Novartis has established a $122 million development of new tools for tuberculosis.
tropical disease research institute in Singapore to
focus on tuberculosis and dengue fever [9]. More-
over, the TB Alliance Scientific Advisory Committee Global product development public – private
includes drug discovery and TB experts from partnerships
GlaxoSmithKline, Pfizer, and Johnson & Johnson.
Additionally, to enhance the scope and depth of Globally based nonprofit PD PPPs emerged in the
the global pipeline, the TB Alliance has invested in 1990s as a way to
platform technologies that expedite, support, and
lower hurdles in TB drug development. One es- 1. Link public-sector goals with private-sector
sential platform technology is the mouse model for expertise as a means of accelerating drug and
preclinical testing, which has been instrumental in vaccine development for neglected diseases
affirming preclinical efficacy of lead compounds. 2. Raise awareness of global health inequities and
Others include the establishment of a worldwide attract substantial new funding to the field
inventory of preclinical and clinical capacity/exper- 3. Promote culture change, incorporating methods
tise and clinical trial site standardization in selected and models from the private sector into public-
countries to ensure the necessary clinical infra- sector practice and encouraging more private
structure as portfolio compounds advance to clinical entities to enter the field of neglected dis-
trials. In this context, the involvement of endemic eases [11]
countries is central to the mission of the TB
Alliance. On the drug development front, these In many ways, PD PPDs can be seen as not-
countries may have compounds to expand the for-profit companies.
346 gardner et al
Building on its history, mission, and comparative Together, these organizations provide a true, new,
advantages, The Rockefeller Foundation provided social and coherent field of initiatives designed to bridge
venture capital toward the creation of the International the gap between basic research and product develop-
AIDS Vaccine Initiative, Medicines for Malaria Ven- ment, aligning public health goals with private-sector
ture, the TB Alliance, International Partnership expertise to develop essential new products to pre-
for Microbicides, and Pediatric Dengue Vaccine Initia- vent, control, and treat diseases of the poor. The field
tive [12]. At the same time, other organizations is still young. Sufficient time has not yet passed to
established a Malaria Vaccine Initiative, Aeras Global determine whether PD PPPs will achieve their
TB Vaccine Foundation, Foundation for Innovative ultimate goals. A recent analysis of interim indicators
New Diagnostics, Drugs for Neglected Diseases Ini- and organizational best-practice benchmarks gives
tiative, Institute for One World Health, and others. Most room for hope, even enthusiasm [14]. Impressive
of these PD PPPs, including the TB Alliance, owe their new public-sector support has arisen, and product-
continued financial well being to donors—principally development pipelines have expanded significantly
the Bill and Melinda Gates Foundation—along with both within and outside of PD PPP portfolios. To
other philanthropies, government development agencies, build on this progress successfully, still more re-
and multilateral organizations. sources will be needed to carry the work forward
The Rockefeller Foundation based its disease- to the next stages including regulatory approvals,
product priorities (eg, in the creation of the TB clinical trials, manufacturing, and distribution.
Alliance) on expert assessment of a combination of
high social demand and maturity of the science. The
Foundation pushed for pharmacoeconomic analy- Future directions for the Global Alliance for
ses, business plans, and scientific blueprints for the Tuberculosis Drug Development
organizations with which it was involved, and these
approaches have now become the norm. Among the TB Alliance’s achievements to date
Like private for-profit companies, PD PPPs use are the rapid scaling-up of their project portfolio,
business practices in staffing and in managing a securing of endorsements at the highest national and
portfolio of candidate products. Most have explicit international levels, and enlisting the assistance of
policies affecting portfolio turnover, including a guil- leading global pharmaceutical and biotechnology
lotine strategy based on milestones in the project’s companies. Most significantly, the TB Alliance is
business plan and criteria for acquiring new product developing a growing, robust pipeline, a quantum
candidates. Go/no-go decisions are based on advice leap from the situation 5 years ago when TB drug
from a scientific advisory board made up of experts development was at a standstill.
in the field. Thus, for some donors, PD PPPs repre- The ideal scenario in the next decade is one in
sent a way to delegate the tough decisions to profes- which the TB Alliance successfully creates and
sional experts. manages a portfolio of therapeutic alternatives and
Each PD PPP seeks to deliver products that will develops a novel anti-TB drug regimen that is as
be cheaper and easier to supply than existing in- effective or more effective than the current treatment
terventions [13]. The quintessential example would and that shortens the duration of treatment to less than
be a vaccine, just as the polio vaccine replaced the 2 months. The reduction in the duration of treat-
iron lung half a century ago. A shorter-course TB ment would represent a significant reduction in the
treatment regimen to replace the existing 6- to cost and complexity of DOTS. A drug that simplifies
12-month regimen represents a similar breakthrough the regimen and shortens treatment by 3 months
goal. Such products will make access to better health would also have a significant, positive impact
more attainable. In addition, PD PPPs focus directly (although not as dramatic as the previous scenario).
on ensuring that their products will be affordable In either case, the benefits of a new, faster-acting TB
and accessible to the poor. Criteria for acquisition of regimen would go beyond the direct impact on TB
new candidate products include low manufacturing incidence and prevalence and the indirect effect on
cost (often with manufacture in developing countries), the global economy. The success of the TB Alliance
ease of delivery in tropical conditions, and advanta- could encourage donors, mobilizing more funding
geous intellectual-property arrangements that entice for additional projects. Demonstration of the success
private-sector participation and help ensure fulfill- of the PD PPP model could spur support for other
ment of the AAA strategy in endemic countries. It is PD PPPs. Other institutions might apply the TB
important that a series of strategic partnerships be Alliance’s platform technologies and lessons learned
established to deliver products to the neediest. during the process of TB drug development to
global alliance for tuberculosis drug development 347
their own objectives. Finally, a success from one of [5] Global Alliance for TB Drug Development. Overview
the PD PPPs might energize the entire field of of the TB Alliance program. Available at: http//www.
global health. tballiance.org. Accessed March 4, 2005.
[6] Pablos-Méndez A for the Working Alliance for TB
Drug Development. The declaration of Cape Town. Int
J Tuberc Lung Dis 2000;4(6):489 – 90.
[7] Nuermberger EL, Yoshimatsu T, Tyagi S, et al.
Acknowledgments Moxifloxacin-containing regimen greatly reduces time
to culture conversion in murine tuberculosis. Am J
The authors thank Ann Ginsberg, Gwynne Ooster- Respir Crit Care Med 2004;169(3):421 – 6.
baan, Melvin Spigelman, and Joelle Tanguy at the [8] Astrazeneca. Astrazeneca opens multi-million dollar
TB Alliance for their valuable comments. Indian research facility to find new treatments for
tuberculosis [press release, June 2, 2003]. Available at:
http://www.astrazeneca.com/pressrelease/497.aspx.
Accessed March 4, 2005.
References [9] Novartis. Novartis Institute for Tropical Diseases opens
in Singapore’s state-of-the-art Biopolis research fa-
[1] World Health Organization Fact sheet no 104, revised cility. Novartis special report. Available at: http://www.
March 2004. Geneva (Switzerland)7 World Health novartis.com/special/nitd_opening_pr.shtml. Accessed
Organization; 2004. March 4, 2005.
[2] Médecins sans Frontières. Campaign for access to [10] Foundation for Innovative New Diagnostics (FIND).
essential medicines Running out of breath? Available Available at: http://www.finddiagnostics.org.
at: http://www.doctorswithoutborders.org/publications/ [11] Internal strategic planning document. New York7 The
reports/2004/tbreport_2004.pdf. Accessed March Rockefeller Foundation; 1998.
4, 2005. [12] Evans TG. Health-related global public goods: initia-
[3] World Health Organization. WHO/IUATLD Global tives of the Rockefeller Foundation 2002. Available at:
Project on Anti-Tuberculosis Drug Resistance Surveil- http://www.undp.org/ods/monterrey-papers/evans.pdf.
lance 1999 – 2002 Anti-tuberculosis drug resistance Accessed March 4, 2005.
in the world (third global report). Available at: http:// [13] Wheeler C, Berkley S. Initial lessons from public-
www.tballiance.org/MDR_report/WHO_MDRTB_ private partnerships in drug and vaccine development.
report.pdf. Accessed March 4, 2005. Bull World Health Organ 2001;79(no. 8):728 – 34.
[4] Global Alliance for TB Drug Development. The [14] The Rockefeller Foundation. Partnering to develop
economics of TB drug development, October 2001. new products for diseases of poverty—one donor’s per-
New York: Global Alliance for TB Drug Develop- spective 2004. Available at: http://www.rockfound.org.
ment; 2001. Accessed March 8, 2005.
Clin Chest Med 26 (2005) 207 – 216
With tuberculosis (TB) having plagued mankind Characteristics of the Mycobacterium tuberculosis
for centuries, there can be no doubt that Mycobacte- complex
rium tuberculosis, the causative agent of human TB,
has been successful in adapting for human infection. The MTC consists of bacteria that genetically
M. tuberculosis belongs to the Mycobacterium tuber- share identical 16S rRNA sequence and greater than
culosis complex (MTC), itself comprised of bacte- 99.9% nucleotide identity. M. tuberculosis, M. af-
rial agents responsible for TB or TB-like disease. ricanum, M. microti, and M. bovis have been re-
Members of the MTC are known to infect mam- garded as the four traditional species of the MTC,
malian hosts, and the extent and consequence of although the extent of MTC speciation is not yet
this infection is gaining greater recognition in part resolved. In this article, MTC organisms are referred
because of the availability of diagnostic tools to to as members, and the nomenclature provided in the
classify specific isolates appropriately. This article most recent literature is used.
introduces the tools and terminology used for this Members characteristically differ in their host
classification and illustrates their utility by discussing range, epidemiology, clinical presentation in humans,
work from independent laboratories that have es- and laboratory phenotype, although little is known
tablished a genome-based phylogeny for the MTC about these differences or why these differences have
[1 – 5]. Next, it considers the use of these markers evolved. The human form (M. tuberculosis sensu
to distinguish atypical isolates not conforming to stricto) and the bovine form (M. bovis) have been
attributes of traditional MTC members [6,7]. Finally, nominally distinct for more than a century; other mem-
it discusses the current genomic evidence regarding bers have been identified more recently (Table 1).
the origin and evolution of M. tuberculosis in the The members classically were described by their
context of its relevance for TB control in humans and biochemical properties or by targeting their specific
other mammalian hosts. genetic regions. Genomic insights now show a new
approach to MTC speciation outside the scope of
these more traditional tools [8].
0272-5231/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccm.2005.02.004 chestmed.theclinics.com
208 mostowy & behr
Table 1
Myobacterium tuberculosis complex members
Virulence
MTC member Natural host Mouse Guinea pig Rabbit Unique attribute
M.canettii Human? + + Most ancestral recognized MTC member,
anecdotal isolation
M. tuberculosis Human + + Predominant cause of human TB
a
M. africanum subtype II Human + + Reclassified as atypical M. tuberculosis
b
M. africanum subtype I (a) Human? + + Rarely isolated
c
M. africanum subtype I (b) Human? + + Phenotypically heterogeneous
M. pinnipedii Pinnipeds + + + Closely related to M. microti
M. microti Vole Attenuated, used as live vaccine in humans
Dassie bacillus Dassie More attenuated than M. microti
M. caprae Goat + + + Only described in Europe?
M. bovis Cow + + + Dynamic pathogen with wildlife reservoirs?
M. bovis BCG None Family of laboratory adapted strains of
M. bovis used as live vaccine
209
210 mostowy & behr
evolutionary selection. Both types of mutations have event polymorphisms (UEPs). These UEPs, which re-
been applied toward differentiation and diagnostics of present one-time events in the evolution of the organ-
MTC members [24 – 26]. In a landmark study, a first ism, can serve as robust markers of clonal organisms,
sequence analysis of MTC isolates revealed that useful for determining phylogenetic classification.
allelic polymorphism is impressively rare, occurring A valuable use of these genomic deletions pertains
on the order of 1 in 10,000 base pairs (bp), suggesting to their application in defining specific MTC mem-
that the complex could be dated to about 15,000 to bers and accurately assessing their prevalence in
20,000 years of age [24]. clinical specimens [8,26]. Unlike biochemical testing,
Genomic comparison of multiple sequenced MTC which for individual results had imperfect sensitivity
strains has made possible the identification of SNP and specificity, the use of genomic events in these
markers for studies of evolution, pathogenesis, and studies provided unambiguous classification, thereby
epidemiology in clinical M. tuberculosis [27] and simplifying the process considerably. To explore the
M. bovis [20], supporting a clonal evolution of the basis for the previously observed biochemical attri-
MTC without detectable lateral gene exchange. The butes used for MTC speciation, an association was
ratio of SNP types within a genome can act as a mo- sought between deleted sequences and phenotypic
lecular clock [28] in which the high ratio of non- results for isolates assigned as M. africanum. Results
synonymous to synonymous mutations across coding indicate that convergent biochemical profiles can be
sequences within MTC genomes suggests a recent independently obtained in different MTC members.
divergence of M. bovis and M. tuberculosis [20]. For instance, organisms presenting the distinct
deletion profile of M. africanum and M. bovis can
Large-sequence polymorphisms manifest the same biochemically based profile [7].
These results confirm the limitations of biochemically
Unlike other mycobacterial species in which derived speciation and, by extension, challenge the
horizontal gene transfer has been demonstrated [22], taxonomic divisions currently in place for classifying
this mode of generating genomic diversity has not members of the MTC.
been observed for the obligately intracellular MTC. Beyond diagnostics, different studies have all sup-
Genomic comparisons for the MTC reveal a promi- ported the potential value of most MTC genomic de-
nent role of genomic deletions relative to the letions (with the exception of mycobacteriophage
sequenced strains of M. tuberculosis. For example, DNA) as evolutionary markers. In separate genomic
the complete genome sequence of M. bovis 2122 studies of M. bovis BCG vaccine strains, it has been
contains 66,037 bp less than M. tuberculosis H37Rv, documented that BCG-specific deletions superimpose
and no genomic region exclusive to M. bovis but perfectly on the historical record [3,29]. In studies of
consistently absent from M. tuberculosis has been genomic deletions within clinical isolates of M. tu-
detected [20]. berculosis [31,32], mycobacterial clones shared the
To uncover deletions in nonsequenced strains same genomic deletions, again suggesting that
efficiently, one can hybridize whole genomic DNA deletions can be used to reconstruct phylogenetic
of a MTC member against a spotted array [29] or an trees. Finally, a recent analysis of 100 M. tuberculosis
Affymetrix GeneChip (Santa Clara, California) [30] clones from San Francisco has again confirmed that
representative of the entire M. tuberculosis H37Rv these deletions are UEPs [5], and therefore genomic
genome [17]. Regions of the prototype strain that deletions can effectively brand a particular clone [33].
seem to be absent from the test strain are then A practical use of this approach will be to provide
confirmed by performing polymerase chain reaction a secure genomic definition for prominent strains,
(PCR) with primers approximating the deleted region, such as the Beijing [34] and Manila [35] strains of
to amplify across the deletion. This amplicon is then M. tuberculosis, and to assess their prevalence
sequenced to define the deletion point precisely. through space and time.
Isolates are said to share a genomic deletion when
sequencing shows the deletion occurs in different
isolates at exactly the same cut point [2]. Because Genomic deletions and the origin and evolution of
independently arisen chromosomal rearrangements the Mycobacterium tuberculosis complex
sometimes involve the same strategically located
elements, only upon exact description of the specific The long-recognized presence of a human TB
genomic event (ie, genomic location within a refer- bacillus and a closely related bovine form has given
ence strain) can one determine with confidence rise to speculation that TB originally came to hu-
whether genomic deletions behave as unidirectional mans as a zoonotic infection from cattle [36]. In
mycobacterium tuberculosis complex evolution 211
retrospect, this view was probably influenced by considered a piscine/amphibian mycobacterium [44]
the types of M. tuberculosis isolates available for and M. avium an avian mycobacterium [45]. The
study, biased toward hosts (namely cattle and hu- MTC is a relatively broad-ranging mammalian
mans) for which a diagnosis of TB would lead to mycobacterium. Organisms of the four most ancestral
microbiologic investigation. To explore the evolu- lineages (M. canettii, M. tuberculosis, and both
tionary relationship of members of the MTC, the genotypes of M. africanum subtype I) have been
presence or absence of deletions was tested within cultured predominantly from humans. Because iso-
complex isolates derived from different hosts and lation of M. canettii has been extremely rare [37,38],
from isolates in various geographic locales [1,2]. an unrecognized nonhuman reservoir might exist,
Analysis revealed a stepwise accumulation of ge- with humans representing an accidental or circum-
nomic deletions among isolates interrogated, but stantial host. The next three MTC lineages (M. mi-
the distribution of genomic deletions argued against croti, M. pinnipedii, and the dassie bacillus) affect
present-day M. bovis as the evolutionary precursor undomesticated mammals irrespective of their geo-
of M. tuberculosis, making it improbable that human graphic location. M. microti, first identified in
TB originated with the domestication of cattle. In- Europe, infects the field vole [19], the dassie bacillus
stead, a number of MTC organisms, both long estab- infects the dassie and the surikat from Africa [6,43],
lished and more recently described, present genomic and M. pinnipedii globally infects a variety of seals
profiles that seem to be intermediate between the and sea lions from Oceania to South America [4,42].
ancestor of modern M. tuberculosis and that of Finally, more derivative forms of the MTC are seen in
present-day M. bovis. goats (M. caprae) and subsequently cattle (classic
The availability of improved laboratory tools has M. bovis), suggesting that the organism was intro-
facilitated the description of a number of novel duced into livestock in the order of their domestica-
variants of the MTC, including M. canettii [37,38], tion. More recently, spillover of M. bovis from farms
M. caprae [39,40], M. pinnipedii [41,42], and the has been seen in the case of badgers in the United
dassie bacillus [2,43] (Table 1). Kingdom [46] and the brushtail opossum in New
Before these tools were available, MTC members Zealand [46a]. Far from suggesting that human TB
had presented a well-established host range, presum- originated with livestock, the genomic record sug-
ably biased by expectations: M. tuberculosis (and gests that, directly or indirectly, humans were respon-
sometimes M. africanum) is classically isolated from sible for bringing MTC to the farm, with secondary
humans, M. microti from voles, and M. bovis from a foci of spread now observed in animals associated
broad range of hosts including (but not limited to) with this setting.
cows. More careful study, however, has revealed a
wider range of host animals. A practical issue arising
from these studies involves the generally held belief Geographic, chronologic, and ecologic origins of
that M. bovis infects an extensive range of animal the Mycobacterium tuberculosis complex
species, including the badger, opossum, elk, cougar,
and buffalo. Until recently, M. caprae and M. pin- An absolute chronology of the TB epidemic is
nipedii were considered to be forms of M. bovis difficult to discern by genomic deletions, because
[40,42]. Although M. bovis might be versatile enough they do not evolve on a predictable time scale. The
to accommodate such a dynamic host range, the genetic record can potentially point to the geographic
inclusion of such organisms probably overestimates origins, however, because the ancestral form M. ca-
the true host range of M. bovis. Detailed genomic nettii [1,4,25] has been isolated only in persons liv-
analysis of isolates from unusual hosts is underway, ing in Africa [37,38]. If the origins of human TB
with the expectation that results will continue to are situated in the same the continent as the origins of
challenge accepted notions of MTC speciation and man, it is conceivable that the organism spread with
taxonomy [2]. humans during the paleomigration, explaining the
Just as genomic deletions have proven unique to presence of MTC DNA in 5000-year-old samples
isolates of M. tuberculosis affecting only human from Egypt [47] and pre-Columbian mummies from
hosts [5,30], deletions unique to these other MTC Ecuador [48]. Because more derivative organisms are
members permit resolution of their phylogenetic found in hosts domesticated 10,000 to 12,000 years
situation (Fig. 1) [1,2]. A first observation from this ago, these clues suggest that the organism accessed
distribution of organisms is that the MTC affects a humans before that era and subsequently spread to
number of undomesticated and domesticated mam- other hosts, either from man directly or through an
mals, both terrestrial and aquatic. M. marinum can be unrecognized vector.
212 mostowy & behr
Fig. 1. Deletion-based phylogeny of the MTC based on deleted regions demonstrated through genomic analysis. The vertical axis
presents the stepwise accumulation of unidirectional evolutionary polymorphisms (RDs and N-RD) previously characterized
among members of the MTC [1,2]. Clustered along each horizontal axis are organisms for which one or more genomic deletions
specific to this evolutionary branch have been revealed in supporting citations [4,6,7,19,32,58] and unpublished observations.
N-RD, new deletions. (Serge Mostowy, Marcel Behr, MD, unpublished data, 2005.)
Using deletions to assign directionality to the reservoirs such as plant or insects deserve consid-
MTC phylogeny, one can employ sequence-based eration [50]. With the ability to test rapidly for
analysis to estimate the chronology of this scenario genomic deletions by PCR, one can test putative
and refine the previous nucleotide-based analysis that wildlife reservoirs for variants of the MTC to find
suggested a 20,000-year divergence between M. tu- the natural host of relatively ancestral forms such as
berculosis and M. bovis [24]. Another approach to M. canettii.
date these events uses testing for genomic regions
directly on paleo-DNA samples [49] (Mostowy et al,
unpublished data). Because these samples can be
carbon dated independently, it is possible to provide What is being deleted from the Mycobacterium
genomic signatures for samples of human or non- tuberculosis complex?
human provenance and to derive minimal estimates
for the ages of genomic events portrayed in Fig. 1. When compared with other bacterial species,
Turning to the ecologic origins of the MTC, a members of the MTC present relatively little genomic
livestock source seems to be unlikely, because human diversity. Estimates of large-sequence polymorphism
forms diverged before the modern caprine and bovine diversity among MTC members [32], in agreement
forms. Although it is attractive to consider another with similar conclusions drawn from estimates of
mammalian host as the ancestral niche, observations SNPs [25], have been consistently described as low
for other mycobacteria suggest that nonmammalian in comparison with other microbes. Nonetheless,
mycobacterium tuberculosis complex evolution 213
genomic flexibility does seem to exist within the M. microti and the dassie bacillus also were shown to
MTC for specific host adaptation, and a similar have deletions in the RD1 region; notably, both have
potential is beginning to reveal itself among other been characterized as having low virulence in animal
mycobacterial complexes. Although the amount of models [19,43]. More detailed analysis of the RD1
diversity revealed within the Mycobacterium avium region revealed it contains genes encoding a novel
complex is 10-fold more than that of the MTC [51], secretion system of two important secreted antigens
host-specific genomic contents are being observed (CFP-10, ESAT-6) [53,54]. Presumably a metabol-
there as well (M. Semret and M. Behr, unpublished ically expensive process, the loss of this region in
data). Taken together, these data highlight a com- BCG was probably advantageous with no selective
parative genomics approach to understanding an pressure in favor of synthesizing and secreting
evolutionary potential of mycobacterium pathogene- antigenic proteins in vitro. Although the independent
sis, in which genomic content can suggest DNA loss of CFP-10 and ESAT-6 in M. microti and the
features for host-specific adaptation. dassie bacillus explains their attenuated phenotype,
the selective pressures for their deletion in vivo are
Genomic deletions and virulence more speculative.
Given the documented impact of the RD1 region
With host-specific MTC extending beyond a on virulence, the observation of its deletion in both
domesticated setting, how TB spreads from host to the vole and dassie hosts is provocative. Nothing
host is difficult to ascertain. From what is known evident points to why the genetically distant vole
about humans and cows with TB, it is reasonable to (a rodent) and dassie (closely related to the elephant)
expect that transmission would occur through aero- would share some unique immunologic susceptibil-
sols from a diseased animal to a contact animal. If ity. A more likely explanation might involve social
so, a requisite of host adaptation is a certain degree conditions and transmissibility [55], given that voles
of virulence in that host. Although greater virulence and dassies congregate in high-density underground
might facilitate transmission, too much virulence communities, unlike other MTC hosts that predomi-
could be detrimental if host mortality is excessive nantly live aboveground in open-air conditions. Such
or if the organism causes an invasive form of TB that congregate living settings would be extremely favor-
is generally nontransmissible (such as TB meningi- able for TB transmission, and an organism of lesser
tis). Thus, optimal transmissibility requires some virulence might be successful in such burrowing
degree of virulence (ie, pulmonary pathology) but a hosts so long as host populations remain sufficiently
sufficiently contained disease process to generate the abundant [56]. Conversely, conditions for transmis-
agents required for spread (ie, aerosols). sion aboveground between goats and seals are less
Support for this notion comes from studying the ideal and would probably require an organism of
content of the genomic regions that have been deleted relatively high virulence to optimize transmissibility.
in different MTC members. A general observation is
that although each deletion noted in Fig. 1 is unique Summary of catalogued deletions
to the bp, both genomic regions and the predicted
function of implicated genes are nonrandom. Several From Fig. 1, deletions represented along the
regions of difference (RD) seem to be prone to vertical line of the phylogeny preceded spread of
genomic deletion, with different specific deletions the bacillus into new hosts; those along the horizontal
having occurred near the same locus [6,7]. Most axes arose during coevolution of the organism with
prominent among these is RD1, a series of nine genes new hosts. Evidence supporting this scenario is that
implicated in the attenuation of M. bovis BCG strains, organisms lacking RD7, RD8, RD9, and RD10 have
that has suffered three distinct genomic deletions. been recovered from the entire MTC host range,
Although this confluence of deletions might point whereas the precise deletions seen along the horizon-
to genetic instability at this locus, a study of 100 cir- tal lineages are observed in only restricted, one-host
culating M. tuberculosis clones documenting 176 de- settings. To derive a scenario for the loss of genomic
letions failed to detect a single deletion in this region, regions in vivo, genes lost on the vertical axis and
arguing against an inherently elevated mutation rate those lost along horizontal lineages can be directly
[32]. The absence of RD1 was first observed for compared, pointing to nonrandom distinction be-
BCG vaccines [51a]; subsequently, targeted disrup- tween the functional classification of these two sets
tion of RD1 from M. tuberculosis was shown to of deleted genes. Although such studies generate
decrease bacterial replication and educe pulmonary hypotheses regarding the evolution of MTC members
pathology in a mouse model [52,53]. More recently, in different hosts, these studies are naturally biased
214 mostowy & behr
[16] Mazars E, Lesjean S, Banuls AL, et al. High- culosis evolution and pathogenesis. J Bacteriol 2003;
resolution minisatellite-based typing as a portable 185:3392 – 9.
approach to global analysis of Mycobacterium tuber- [28] Kimura M. The neutral theory of molecular evolution.
culosis molecular epidemiology. Proc Natl Acad Sci Cambridge (UK)7 Cambridge University Press; 1983.
U S A 2001;98:1901 – 6. [29] Behr MA, Wilson MA, Gill WP, et al. Comparative
[17] Cole ST, Brosch R, Parkhill J, et al. Deciphering genomics of BCG vaccines by whole-genome DNA
the biology of Mycobacterium tuberculosis from microarray. Science 1999;284:1520 – 3.
the complete genome sequence. Nature 1998;393: [30] Salamon H, Kato-Maeda M, Small PM, et al.
537 – 44. Detection of deleted genomic DNA using a semi-
[18] Fleischmann RD, Alland D, Eisen JA, et al. Whole- automated computational analysis of GeneChip data.
genome comparison of Mycobacterium tuberculosis Genome Res 2000;10:2044 – 54.
clinical and laboratory strains. J Bacteriol 2002;184: [31] Kato-Maeda M, Rhee JT, Gingeras TR, et al.
5479 – 90. Comparing genomes within the species Mycobacte-
[18a] The Institute for Genomic Research. (Preliminary rium tuberculosis. Genome Res 2001;11:547 – 54.
sequence data.) Available at: http://www.tigr.org. [32] Tsolaki AG, Hirsh AE, DeRiemer K, et al. Functional
Accessed March 1, 2005. and evolutionary genomics of Mycobacterium
[19] Brodin P, Eiglmeier K, Marmiesse M, et al. Bacterial tuberculosis: insights from genomic deletions in
artificial chromosome-based comparative genomic 100 strains. Proc Natl Acad Sci U S A 2004;101:
analysis identifies Mycobacterium microti as a natu- 4865 – 70.
ral ESAT-6 deletion mutant. Infect Immun 2002;70: [33] Nguyen D, Brassard P, Menzies D, et al. Genomic
5568 – 78. characterization of an endemic Mycobacterium
[20] Garnier T, Eiglmeier K, Camus JC, et al. The com- tuberculosis strain: evolutionary and epidemiologic
plete genome sequence of Mycobacterium bovis. implications. J Clin Microbiol 2004;42:2573 – 80.
Proc Natl Acad Sci U S A 2003;100:7877 – 82. [34] Bifani PJ, Mathema B, Kurepina NE, et al. Global
[20a] The Sanger Institute. Available at: ftp://ftp.sanger.ac. dissemination of the Mycobacterium tuberculosis
uk/pub/pathogens/mb/. Accessed March 1, 2005. W-Beijing family strains. Trends Microbiol 2002;10:
[21] Cole ST, Eiglmeier K, Parkhill J, et al. Massive 45 – 52.
gene decay in the leprosy bacillus. Nature 2001;409: [35] Douglas JT, Qian L, Montoya JC, et al. Characteri-
1007 – 11. zation of the Manila family of Mycobacterium
[22] Stinear TP, Mve-Obiang A, Small PL, et al. Giant tuberculosis. J Clin Microbiol 2003;41:2723 – 6.
plasmid-encoded polyketide synthases produce the [36] Stead WW, Eisenach KD, Cave MD, et al. When did
macrolide toxin of Mycobacterium ulcerans. Proc Mycobacterium tuberculosis infection first occur in
Natl Acad Sci U S A 2004;101:1345 – 9. the New World? An important question with public
[22a] GenBank. Accession no NC_002944. Accessed health implications. Am J Respir Crit Care Med 1995;
March 1, 2005. 151:1267 – 8.
[22b] The Sanger Institute. Available at: ftp://ftp.sanger.ac. [37] Pfyffer GE, Auckenthaler R, van Embden JD, et al.
uk/pub/pathogens/mm/. Accessed March 1, 2005. Mycobacterium canettii, the smooth variant of M.
[23] Brosch R, Pym AS, Gordon SV, et al. The evolution tuberculosis, isolated from a Swiss patient exposed
of mycobacterial pathogenicity: clues from compara- in Africa. Emerg Infect Dis 1998;4:631 – 4.
tive genomics. Trends Microbiol 2001;9:452 – 8. [38] Van Soolingen D, Hoogenboezem T, de Haas PE, et
[24] Sreevatsan S, Pan X, Stockbauer KE, et al. Restricted al. A novel pathogenic taxon of the Mycobacterium
structural gene polymorphism in the Mycobacterium tuberculosis complex, Canetti: characterization of an
tuberculosis complex indicates evolutionarily recent exceptional isolate from Africa. Int J Syst Bacteriol
global dissemination. Proc Natl Acad Sci U S A 1997;47:1236 – 45.
1997;94:9869 – 74. [39] Aranaz A, Liebana E, Gomez-Mampaso E, et al.
[25] Gutacker MM, Smoot JC, Migliaccio CA, et al. Mycobacterium tuberculosis subsp. caprae subsp.
Genome-wide analysis of synonymous single nucleo- nov.: a taxonomic study of a new member of the
tide polymorphisms in Mycobacterium tuberculosis Mycobacterium tuberculosis complex isolated from
complex organisms: resolution of genetic rela- goats in Spain. Int J Syst Bacteriol 1999;49(Pt 3):
tionships among closely related microbial strains. 1263 – 73.
Genetics 2002;162:1533 – 43. [40] Aranaz A, Cousins D, Mateos A, et al. Elevation of
[26] Huard RC, Oliveira Lazzarini LC, Butler WR, et al. Mycobacterium tuberculosis subsp. caprae Aranaz
PCR-based method to differentiate the subspecies et al. 1999 to species rank as Mycobacterium caprae
of the Mycobacterium tuberculosis complex on the comb. nov., sp. nov. Int J Syst Evol Microbiol 2003;
basis of genomic deletions. J Clin Microbiol 2003;41: 53:1785 – 9.
1637 – 50. [41] Cousins DV, Williams SN, Reuter R, et al. Tuber-
[27] Alland D, Whittam TS, Murray MB, et al. Modeling culosis in wild seals and characterisation of the seal
bacterial evolution with comparative-genome-based bacillus. Aust Vet J 1993;70:92 – 7.
marker systems: application to Mycobacterium tuber- [42] Cousins DV, Bastida R, Cataldi A, et al. Tuberculosis
216 mostowy & behr
in seals caused by a novel member of the Myco- bacille Calmette-Guerin attenuation. J Infect Dis
bacterium tuberculosis complex: Mycobacterium 2003;187:117 – 23.
pinnipedii sp. nov. Int J Syst Evol Microbiol 2003; [53] Hsu T, Hingley-Wilson SM, Chen B, et al. The pri-
53:1305 – 14. mary mechanism of attenuation of bacillus Calmette-
[43] Cousins DV, Peet RL, Gaynor WT, et al. Tuberculosis Guerin is a loss of secreted lytic function required
in imported hyrax (Procavia capensis) caused by an for invasion of lung interstitial tissue. Proc Natl Acad
unusual variant belonging to the Mycobacterium tu- Sci U S A 2003;100:12420 – 5.
berculosis complex. Vet Microbiol 1994;42:135 – 45. [54] Pym AS, Brodin P, Majlessi L, et al. Recombinant
[44] Decostere A, Hermans K, Haesebrouck F. Piscine BCG exporting ESAT-6 confers enhanced protection
mycobacteriosis: a literature review covering the against tuberculosis. Nat Med 2003;9:533 – 9.
agent and the disease it causes in fish and humans. [55] Boots M, Hudson PJ, Sasaki A. Large shifts in
Vet Microbiol 2004;99:159 – 66. pathogen virulence relate to host population structure.
[45] Thoen CO, Karlson AG, Himes EM. Mycobacterial Science 2004;303:842 – 4.
infections in animals. Rev Infect Dis 1981;3:960 – 72. [56] Davis S, Begon M, De Bruyn L, et al. Predictive
[46] Donnelly CA, Woodroffe R, Cox DR, et al. Impact of thresholds for plague in Kazakhstan. Science 2004;
localized badger culling on tuberculosis incidence in 304:736 – 8.
British cattle. Nature 2003;426:834 – 7. [57] Monack DM, Mueller A, Falkow S. persistent
[46a] Morris RS, Pfeiffer DU. Directions and issues in bacterial infections: the interface of the pathogen
bovine tuberculosis epidemiology and control in New and the host immune system. Nat Rev Microbiol
Zealand. N Z Vet J 1995;43(7):256 – 65. 2004;2:747 – 65.
[47] Zink AR, Sola C, Reischl U, et al. Characterization [58] Behr MA. BCG—different strains, different vaccines?
of Mycobacterium tuberculosis complex DNAs from Lancet Infect Dis 2002;2:86 – 92.
Egyptian mummies by spoligotyping. J Clin Micro- [59] Griffin JFT, Chinn DN, Rodgers CR. Diagnostic
biol 2003;41:359 – 67. strategies and outcomes on three New Zealand deer
[48] Konomi N, Lebwohl E, Mowbray K, et al. Detection farms with severe outbreaks of bovine tuberculosis.
of mycobacterial DNA in Andean mummies. J Clin Tuberculosis 2004;84:293 – 302.
Microbiol 2002;40:4738 – 40. [60] Lantos A, Niemann S, Mezosi L, et al. Pulmonary
[49] Donoghue HD, Spigelman M, Greenblatt CL, et al. tuberculosis due to Mycobacterium bovis subsp.
Tuberculosis: from prehistory to Robert Koch, as re- caprae in captive Siberian tiger. Emerg Infect Dis
vealed by ancient DNA. Lancet Infect Dis 2004;4: 2003;9:1462 – 4.
584 – 92. [61] Thompson PJ, Cousins DV, Gow BL, et al. Seals, seal
[50] Marsollier L, Stinear T, Aubry J, et al. Aquatic plants trainers, and mycobacterial infection. Am Rev Respir
stimulate the growth of and biofilm formation by Dis 1993;147:164 – 7.
Mycobacterium ulcerans in axenic culture and har- [62] Wagner JC, Buchanan G, Bokkenheuser V, et al.
bor these bacteria in the environment. Appl Environ An acid fast bacillus isolated from the lungs of a cape
Microbiol 2004;70:1097 – 103. hyrax. Procavia capensis (Pallus). Nature 1958;181:
[51] Semret M, Zhai G, Mostowy S, et al. Extensive ge- 284 – 5.
nomic polymorphism within Mycobacterium avium. [63] Dannenberg Jr AM. Pathogenesis of pulmonary
J Bacteriol 2004;186:6332 – 4. Mycobacterium bovis infection: basic principles
[51a] Mahairas GG, Sabo PJ, Hickey MJ, et al. Molecular established by the rabbit model. Tuberculosis (Edinb)
analysis of genetic differences between Mycobacte- 2001;81:87 – 96.
rium bovis BCG and virulent M. bovis. J Bacteriol [64] Springer B, Stockman L, Teschner K, et al. Two-
1996;178:1274 – 82. laboratory collaborative study on identification of
[52] Lewis KN, Liao R, Guinn KM, et al. Deletion of mycobacteria: molecular versus phenotypic methods.
RD1 from Mycobacterium tuberculosis mimics J Clin Microbiol 1996;34:296 – 303.
Clin Chest Med 26 (2005) 273 – 282
In the past 5 years, the Tuberculosis Trials Con- emerged from the first studies of drug therapy of
sortium (TBTC) of the Centers for Disease Control tuberculosis initiated in the late 1940s. These studies
and Prevention has completed several large studies evaluated monotherapy with streptomycin and sub-
that have improved the understanding of pharmaco- sequently para-aminosalicylic acid (PAS) [7 – 9].
therapy of tuberculosis. Insights gained from these They demonstrated that drug resistance developed
studies have resulted in major changes in drug frequently in persons treated with monotherapy.
therapy of tuberculosis in HIV-infected and non- During 3 months of monotherapy with streptomycin,
infected individuals [1 – 5]. These advances require 92% of persons who remained culture-positive
that tuberculosis drug therapy now be individualized. developed streptomycin resistance [3]. Resistance
Recommended treatment regimens are based on a also developed commonly during monotherapy with
patient’s risk profile that is determined by a combi- PAS and was found in approximately one third of
nation of hematologic, microbiologic, clinical, and patients during 4 months of treatment [9]. It was also
radiographic findings [6]. These studies have resulted observed that resistance was much less common in
in substantial changes in the treatment guidelines. persons treated with the combination of streptomycin
Although they are more complicated than the pre- and PAS, and that many more patients treated with
vious guidelines, they allow treatment to be refined the two-drug regimen became bacteriologically nega-
so that it can be extended in patients at high risk for tive with 4 months of therapy [9]. Ten percent or less
treatment failure and allow shorter, more convenient of persons treated simultaneously with streptomycin
treatment regimens in patients who can be identified and PAS developed streptomycin resistance [7,8]. It
as being at very low risk for failure [2]. This article also was observed that development of resistance
reviews the basic principles of drug treatment of was associated with a worse prognosis and with more
tuberculosis, individual pharmacologic agents, cur- severe disease [3]. From these early observations
rent treatment recommendations, and several special came the principle that tuberculosis treatment must
situations that clinicians are likely to encounter in include simultaneous treatment with at least two ef-
medical practice. fective drugs.
The microbiologic basis for these early observa-
tions was not identified until the early 1960s and
Axioms of chemotherapy of tuberculosis remains as important today to understand the design
of current treatment regimens [10]. Persons with cavi-
Effective tuberculosis drug therapy requires not tary disease are estimated to have bacterial popu-
one but at least two effective drugs. This axiom lations of approximately 108 organisms in each cavity
[10,11]. During division, Mycobacterium tubercu-
losis bacilli mutate from drug-susceptible to drug-
* Corresponding author. resistant status spontaneously, randomly, and at a
E-mail address: sweis@hsc.unt.edu (S.E. Weis). predictable rate [12]. The proportion of naturally oc-
0272-5231/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccm.2005.02.011 chestmed.theclinics.com
274 sahbazian & weis
curring organisms that are resistant to antituberculosis with drug-resistant tuberculosis is testimony that
drugs is variable, approximately 10 5 for ethambutol, these principles are not being implemented success-
10 6 for isoniazid and streptomycin, and 10 8 for fully [14].
rifampin [8]. The probability of a single organism The last 50 years of tuberculosis drug treatment
mutating simultaneously and becoming resistant to can be summarized succinctly. First, it was demon-
two drugs is the product of individual probabilities strated that proper chemotherapy and the cooperation
of mutation. It can therefore be estimated that the of the patient are the most important factors influenc-
likelihood of an organism having mutations simulta- ing response to treatment [15]. Second, it has been
neously for isoniazid and rifampin is approximately proven that the social factors such as those corrected
[(1 10 6) (1 10 8)] or (1 10 14), and the during sanatoria treatment of tuberculosis (which pro-
bacillary burden in human tuberculosis is several vided bed rest, airy accommodations, a well-balanced
orders less than this [13]. This mutation rate is the diet, good nursing care, and psychologic balance)
basis for the observation that successful drug ther- have had no effect on outcome in persons prescribed
apy requires that at least two drugs be given drug therapy and cooperating with treatment [15].
concurrently to prevent selection of drug-resistant Third, a few new antituberculosis drugs have been
organisms. If a single drug is used for treatment, developed. For the most part, however, progress has
selection of the resistant organisms occurs, and the been made in learning to use available drugs more
patient rapidly becomes resistant to that drug. This effectively, with treatment regimens becoming re-
mutation rate is also the basis for designing regimens fined to the current treatments that are shorter, have
with an intensive initial phase that uses more medi- fewer side effects, and are more convenient [6].
cations and a less intense continuation phase that uses
fewer medications.
Corollaries of the treatment axiom that tubercu-
losis treatment must include simultaneous treatment Pharmacology and toxicity of antimycobacterial
with at least two effective drugs are important for agents
designing effective tuberculosis regimens and tu-
berculosis control programs. Because of the possi- The current drugs approved by the Food and Drug
bility of resistance, a single drug is never added to Administration (FDA) for the treatment of tuber-
a failing drug-treatment regimen. Optimal design of culosis include isoniazid, pyrazinamide, rifampin,
re-treatment regimens should include at least two rifapentine, ethambutol, cycloserine, ethionamide, ca-
medications to which the patient is naı̈ve, and preomycin, PAS, and streptomycin. Drugs that com-
clinicians designing initial treatment regimens must monly are recommended by expert panels for use in
consider prevailing tuberculosis-susceptibility pat- the treatment of tuberculosis but are not FDA ap-
terns in the community where the infection probably proved include rifabutin, the aminoglycosides includ-
was acquired. It is equally important to successful ing amikacin, kanamycin, and the fluoroquinolones
treatment that the patient actually take the two proba- including ciprofloxacin, moxifloxacin, and levofloxacin.
bly effective drugs. The only way to ensure that a Of the approved drugs, isoniazid, rifampin, ethambu-
patient actually takes drug therapy as prescribed is tol, and pyrazinamide are considered first-line anti-
direct observation of therapy. If three separate drugs tuberculosis drugs. Rifapentine and rifabutin can also
are prescribed for a patient with tuberculosis, the be considered first-line drugs under special condi-
patient may, for many reasons, take a single drug at tions discussed later. The others are categorized as
a time. Short-term single-drug therapy in a person second-line drugs, which are used when the first-line
with high bacillary burden can lead to emergence drugs are unsuitable because of drug intolerance
of drug resistance [7 – 9]. If a patient happens to or infection with drug-resistant tuberculosis. Addi-
be initially resistant to one drug and takes a combi- tionally clarithromycin, amoxicillin/clavulanate, and
nation of two drugs, including the one to which he linezolid have been used in the treatment of patients
or she is resistant, drug resistance to the second with drug-resistant tuberculosis.
drug will emerge. Similarly, if the patient is resistant Drug-level monitoring is not routinely an impor-
to two drugs and takes these two drugs and a single tant aspect of treatment in a patient with active
effective drug, resistance to the third will emerge. tuberculosis. Therapeutic drug monitoring is most
Therefore, poor adherence, inadequate prescribing, useful when there is a direct relationship between
or both may result in the development of multidrug serum concentrations and therapeutic response and
resistance. Although these axioms may seem self- when serum concentrations serve as a surrogate for
evident, the growing number of persons worldwide drug concentrations at the site of action. Therapeutic
treatment of active tuberculosis 275
drug monitoring is also important when there is a Peripheral neuropathy also is associated occasion-
narrow range of concentrations that are effective ally with use of isoniazid. Neuropathy occurs more
and safe and when toxicity or lack of effectiveness commonly among persons who have other risks
puts the patient at great risk [16,17]. Examples of for neuropathy. Persons at increased risk of periph-
situations in which therapeutic drug monitoring is eral neuropathy include those who are nutritionally
useful for safety include persons treated with amino- deficient, alcoholics, diabetics, pregnant women,
glycosides and persons treated with ethambutol or breastfeeding mothers, and patients with renal dis-
cycloserine with renal impairment. ease. Vitamin B6 (pyridoxine) supplements usually
are given with isoniazid to prevent development of
peripheral neuropathy [6].
Isoniazid Hypersensitivity reactions including arthralgias,
irritability, seizures, and lupuslike syndrome have
Isoniazid is used for the treatment of both latent also been reported in patients receiving isoniazid.
and active tuberculosis and works primarily by Although as many as 20% of patients treated with
inhibiting cell wall synthesis. It is usually adminis- isoniazid develop a positive antinuclear antibody test,
tered orally but has been given successfully intra- systemic lupus rarely occurs [26].
muscularly or intravenously [6]. Isoniazid is cleared Isoniazid has clinically important reactions with
predominantly through the liver by acetylation. A other concomitantly used medications. Isoniazid
patient’s acetylation status and the associated differ- can affect the levels of certain antiseizure medica-
ences in plasma isoniazid concentrations are not as- tions, such as phenytoin and carbamazepine. Levels
sociated with isoniazid-induced liver injury [18]. of these medications must be monitored during iso-
Additionally, no association was found between niazid therapy [6].
plasma isoniazid concentrations and isoniazid-induced
liver injury [19]. Isoniazid is distributed throughout
the body with peak concentrations occurring within Rifamycins
1 to 2 hours after the administration of an oral dose
[20]. The usual dose for isoniazid is 3 to 5 mg/kg The rifamycins, which include rifampin, rifabutin,
body weight/day in adults with a maximum dose and rifapentine, work by interfering with RNA
of 300 mg/day [6]. synthesis, even in bacilli with minimal metabolic
Isoniazid generally is well tolerated. Hepatic side activity [27]. The rifamycins are variable inducers of
effects are perhaps the best known of the untoward the cytochrome P450 system. Rifampin, rifabutin,
effects associated with isoniazid use. Less well and rifapentine are each first-line drugs for the treat-
known is the asymptomatic elevation of liver amino- ment of tuberculosis in different circumstances.
transferases of up to five times the upper limits of Rifampin generally is given orally, but formulations
normal, which occurs in approximately 20% of are available for parenteral therapy. The usual dose
patients receiving isoniazid. This asymptomatic mild for rifampin in adults is 10 mg/kg to a maximum of
elevation of liver aminotransferases is not progres- 600 mg daily. It is distributed well throughout the
sive, is not an indication of progressive liver toxicity, body and reaches effective concentrations in all tis-
and when asymptomatic does not require discon- sues [6]. Rifampin is a necessary component of all
tinuation of isoniazid treatment [6]. Isoniazid-induced short-course regimens [6].
hepatitis does occur, but recent studies indicate it is Rifampin is generally a well-tolerated drug. The
less common than previously thought. Isoniazid- most common side effect of rifampin use is an
induced hepatitis is estimated to occur in 0.15% of orange discoloration of the urine, tears, and other
those starting and in 0.15% of those completing body fluids. The change in the color of the urine or
treatment for latent tuberculosis infection [21]. The other body fluids can be disconcerting to persons
rate of isoniazid-induced hepatitis is higher when treated with rifampin if they are not warned. This
isoniazid is combined with rifampin [22]. It is also discoloration has been associated with discoloration
more common in older persons, heavy alcohol of soft contact lenses and clothing. This staining
consumers, and persons with underlying liver disease must be rare, however, because the author and col-
[23]. Based on a large survey, the risk of isoniazid- leagues have treated many contact lens wearers with
induced fatal hepatitis is much lower than previously rifampin and never have had a complaint of dis-
thought—0.001%—when patients are monitored rou- coloration of contacts lens, even though they rou-
tinely for liver toxicity [24,25]. The risk increases tinely warn patients of this potential side effect.
slightly in patients over the age of 35 years. Rifampin can also cause pruritus [28]. Gastroin-
276 sahbazian & weis
testinal upset, including diarrhea, nausea, and ab- concurrent medications must be checked for inter-
dominal pains, can occur but rarely require drug actions with rifampin.
discontinuation and are usually self limiting [29].
Transient elevation of serum bilirubin may be ob- Ethambutol
served during rifampin administration. Hepatitis is
more common when rifampin is administered with Ethambutol is used in combination with isoniazid
isoniazid [30]. and rifampin in the initial treatment of active tuber-
A more serious side effect of rifampin use is an culosis and has been proven effective in primary
influenzalike syndrome. Symptoms often mistaken treatment of pulmonary tuberculosis [35]. Ethambutol
by patients and physicians for influenza, including is used with isoniazid and rifampin to prevent
fevers, chills, faintness, headaches, myalgia, and ar- selecting resistant organisms when resistance to
thralgia, occur alone or in combination. This hyper- one of the primary drugs is present. Like isoniazid,
sensitivity syndrome seems to be immune mediated ethambutol inhibits cell wall synthesis. It is available
and develops primarily when rifampin is given only in the oral form. Because it is secreted through
intermittently or in larger doses than are currently the kidneys, it can accumulate in patients with renal
recommended. It most commonly develops after 3 to insufficiency. The usual dose for ethambutol is 15 to
6 months but can occur at any time during treatment 20 mg/kg/day or 50 mg/kg two times per week.
[31]. Among persons receiving once-weekly rifampin Ethambutol generally is very well tolerated, but,
as part of the antituberculosis regimen, 35% to 57% rarely, it can cause retrobulbar neuritis. This syn-
of persons who received 1200 to 1800 mg rifampin drome first manifests as decreased red-green color
developed a flulike syndrome; the rates were 22% to discrimination and visual acuity. Although it can
31% among those receiving 900 mg/week and 10% result in irreversible vision loss, recognition of the
among persons taking 600 mg/week [32]. In contrast, symptoms and prompt discontinuation of the drug
for persons who received twice-weekly rifampin, a usually results in return of normal vision. Reducing
flulike syndrome was reported in 8% of those the dose of ethambutol to 15 mg/kg/day can minimize
receiving 900 mg/week and in 4% of those receiving the risk [36].
600 mg/week. Symptoms usually appear 1 to 2 hours
after administration of the drug and last up to 8 hours Fluoroquinolones
[31,32]. The hypersensitivity syndrome can be ac-
companied by other manifestations that may be Levofloxacin, moxifloxacin, and gatifloxacin
severe and, rarely, life threatening. The incidence all are active against mycobacterium tuberculosis
of individual adverse drug reactions included in [37,38]. Although they are not approved by the
the hypersensitivity syndrome is not well described FDA for the treatment of tuberculosis, they are
for persons treated for tuberculosis. A study of used frequently in treating drug-resistant tubercu-
20,667 patients treated for leprosy with rifampin, losis or when patients are intolerant of first-line
600 mg/day for 3 months, noted the following agents [39,40]. The adult dose for levofloxacin is
incidence rates: rash (0.07%), acute renal failure 500 to 1000 mg/day orally. Moxifloxacin is admin-
(0.1%), thrombocytopenia (0.01%), and hypotension istered at 400 mg/day. Central nervous system (CNS)
(0.01%) [33]. Rifabutin use is also associated with concentrations of fluoroquinolones have been found
rare immune-related reactions. These reactions tend to be around 16% to 20% of serum after admin-
to be hematologic, such as leukopenia and thrombo- istration of a standard dose of levofloxacin [41].
cytopenia [33,34]. Microbial resistance to fluoroquinolones is common
Rifampin can interact with a large number of in the community setting; therefore it is imperative
medications because it is a potent inducer of several that fluoroquinolones be used only when appropriate.
enzymes. Rifampin induction of hepatic enzymes can The most common side effects reported with the use
reduce serum concentrations of oral contraceptives, of this group of antimicrobials are gastrointestinal
resulting in pregnancy, and women relying on hor- symptoms such as nausea, anorexia, dyspepsia, ab-
monal methods of contraception need to use addi- dominal pain, followed by CNS disturbances (head-
tional means of contraception. Rifampin can increase ache, dizziness, drowsiness, abnormal vision) and
the metabolism of methadone and glucocorticoids, liver enzyme abnormalities [42]. Fluoroquinolones
resulting in narcotic withdrawal syndrome and ad- were not developed with the expectation that they
renal insufficiency or exacerbation of the illness would be used for months; however most experts in
being treated by glucocorticoid. The interactions of the field of TB have reported a good safety profile
rifampin with other drugs are so extensive that all and tolerability with long-term use. Controlled stud-
treatment of active tuberculosis 277
sahbazian
3 INH 3/wk for 24 doses (8 wk) 3a INH/RIF 3/wk for 54 doses (18 wk) 78 (26 wk) B (I) B (II)
RIF
PZA
EMB
&
4 INH 7 d/wk for 56 doses (8 wk) or 5 d/wk 4a INH/RIF 7 d/wk for 217 doses (31 wk) or 273 – 195 (39 wk) C (I) C (II)
weis
RIF for 40 doses (8 wk)e 5 d/wk for 156 doses (31 wk)c
EMB 4b INH/RIF 2/wk for 62 doses (31 wk) 118 – 102 (39 wk) C (I) C (II)
Abbreviations: EMB, Ethambutol; HIV , HIV-negative; HIV+, HIV-positive; INH, isoniazid; PZA, pyrazinamide; RIF, rifampin; RPT, rifapentine.
a
Definitions of evidence ratings: A, preferred; B, acceptable alternative; C, offer when A and B cannot be given; E, should never be given.
b
Definitions of evidence ratings: I, randomized clinical trial; II, data from clinical trials that were not randomized or were conducted in other populations; III, expert opinion.
c
When directly observed therapy is used, drugs may be given 5 d/wk and the necessary number of doses adjuated accordingly. Although there are no studies that compare five
with seven daily doses, extensive experience indicates this would be an effective practice.
d
Patients with cavitation on initial chest radiograph and positive cultures at completion of 2 months of therapy should receive a 7-month (31 wk; either 217 doses [7/wk] or 62 doses
[2/wk]) continuation phase.
e
Five d/wk administration is always given by DOT. Rating for 5 d/wk regimens is A III.
f
Not recommended for HIV-infected patients with CD4+ cell counts <100 cells/ml.
g
Options 1c and 2b should be used only in HIV-negative patients who have negative sputum smears at the time of completion of 2 months of therapy and who do not have cavitation
on initial chest radiograph (see text). For patients started on this regimen and found to have a positive culture from 2-month specimen, treatment should be extended an extra 3 months.
From American Thoracic Society, Centers for Disease Control and Prevention, and Infectious Diseases Society of America. Treatment of tuberculosis. MMWR Recomm Rep 2003;
52(RR-11):3.
treatment of active tuberculosis 279
niazid and rifampin. It can also be given, more con- therapy group. There was no demonstrable improve-
veniently for patients and staff, once weekly using ment in the broader prespecified combined end points
isoniazid and rifapentine in patients with tuberculo- of death or severe disability after 9 months [52].
sis without cavitation on the chest radiograph. This There have been many reports of an increased risk
group is estimated to represent 40% of persons with of tuberculosis in patients receiving tumor necrosis
tuberculosis in the United States [3]. Persons who factor- alpha (TNF-a) antagonists [53,54]. These
have cavitation on the initial or follow-up chest agents, which include infliximab, etanercept, and
radiograph or who are culture positive at the end adalimumab, are used for the treatment of an ex-
of initial phase of therapy (usually completed after panding group of diseases and work by blocking
2 months) have an unacceptably high risk of treat- TNF-a; an inflammatory cytokine. TNF-a is ex-
ment failure [3,6]. For these patients, the continuation pressed by activating macrophages, T cells, and other
phase should be extended for an additional 3 months immune cells and is an important part of the host
[6]. It is critically important to have sputum cultures response against M. tuberculosis and other intra-
at the time of completion of the initial phase of treat- cellular organisms. Current expert opinion on this
ment to identify patients at increased risk of relapse. emerging problem in tuberculosis treatment is that
The treatment of tuberculosis in persons with HIV the TNF-a antagonist should be discontinued if
is discussed elsewhere in this issue. tuberculosis develops during TNF-a antagonist ther-
Treatment of tuberculosis may be delayed for apy. The optimal time for resuming TNF-a antagonist
many reasons. The current treatment guidelines de- therapy is undetermined. It is recommended that
fined completion of adequate therapy by the number TNF-a antagonist therapy be withheld at least until
of doses ingested as well as by the duration of treatment with the tuberculosis regimen has been
treatment administration [6]. The minimum goal for started, and the patient’s condition has improved [54].
adequate therapy is delivery of the full number of Tuberculosis occurring in pregnancy is a danger to
doses in no more than 150% of the expected delivery the pregnant woman and her child, and treatment
duration [6]. should not be delayed because of the pregnancy. In-
fants born to women with untreated tuberculosis may
Special situations be of lower birth weight than those born to women
without tuberculosis and can acquire congenital
Central nervous system tuberculosis is one of the tuberculosis [55 – 57]. Of the first-line medications,
most devastating presentations of human tuberculo- pyrazinamide is not recommended for general use in
sis. Disability and death occur despite antitubercu- pregnant women in the United States because of
losis therapy [49]. The best antimicrobial agents for insufficient data to determine safety. Aminoglyco-
the treatment of central nervous system tuberculosis sides should not be used to treat tuberculosis in preg-
have not been validated by well-designed, random- nancy, because they are associated with birth defects
ized, clinical trials. Isoniazid and pyrazinamide pene- [6]. There is little information about the safety of
trate the meninges in all stages of inflammation. second-line antituberculosis drugs during pregnancy.
Rifampin, ethambutol, and aminoglycosides pene- The recommended initial treatment regimen in preg-
trate the blood – brain barrier in the presence of men- nancy should consist of isoniazid, rifampin, and
ingeal inflammation but poorly in its absence. The ethambutol [6]. If the organism is confirmed to be
use of glucocorticoids in an attempt to reduce mor- susceptible to isoniazid and rifampin, the ethambutol
tality and morbidity has been controversial [50]. may be discontinued and isoniazid and rifampin con-
Recently, a large trial of dexamethasone adjunct tinued for a minimum of 9 months [6]. It is recom-
therapy for persons 14 years of age and older with mended that pregnant women receiving isoniazid also
tuberculous meningitis has clarified the role of glu- be given pyridoxine (25 mg/day) [6]. Breastfeeding
cocorticoids [51]. Dexamethasone treatment was should not be discouraged for women being treated
started as soon as possible after starting antituber- with first-line agents, because the small concentra-
culosis treatment. Patients were stratified by Glasgow tions of these drugs in breast milk do not produce
Coma Scale and given intravenous dexamethasone toxic effects in the nursing infant [58].
for 4 weeks for severe disease and for 2 weeks for Renal insufficiency increases the risk for devel-
mild disease. Subsequently, all patients were given oping tuberculosis, and treatment of the two con-
tapering doses of dexamethasone orally for an addi- ditions concurrently is a complex and common
tional 4 weeks. Dexamethasone adjunctive treatment situation. Isoniazid and rifampin are metabolized in
improved survival. Adverse and severe adverse events the liver, and dosages need not be changed in persons
were reduced significantly in the dexamethasone- with chronic renal failure [59 – 61]. Metabolites of
280 sahbazian & weis
pyrazinamide are excreted renally and can accumu- mens in patients who can be identified as being at
late in patients with renal insufficiency [61]. Approxi- very low risk for failure [2].
mately 80% of ethambutol is cleared by the kidneys,
so ethambutol may accumulate in patients with renal
insufficiency [59,61]. Reducing the dosage may avoid Acknowledgments
toxicity, but the peak serum concentrations achieved
may be too low to be effective. Therefore increasing The authors acknowledge Thaddeus Miller’s work
the dosing interval is recommended [60]. For patients in editing this article.
undergoing hemodialysis, administering all drugs for
tuberculosis after dialysis is a way to facilitate di-
rectly observed treatment and simultaneously to References
avoid removal of drugs such as pyrazinamide [6].
To avoid toxicity, it is important to monitor serum [1] Vernon A, Burman W, Benator D, et al. Relapse with
drug concentrations in persons with renal failure who rifamycin mono-resistant tuberculosis in HIV-infected
are taking aminoglycosides, cycloserine, or etham- patients treated with supervised once-weekly rifapen-
butol [60]. Data are unavailable for the effect of tine and isoniazid. Lancet 1999;353:1843 – 7.
peritoneal dialysis on the clearance of antitubercu- [2] Centers for Disease Control and Prevention. Acquired
losis drugs. rifamycin resistance in persons with advanced HIV
disease being treated for active tuberculosis with
The challenges facing patients with tuberculosis
intermittent rifamycin-based regimens. MMWR Morb
and underlying liver disease are great. Clinicians Mortal Wkly Rep 2002;51:214 – 5.
must choose antituberculosis agents that, with a few [3] Benator D, Bhattacharya M, Bozeman L, et al.
exceptions, are metabolized by the liver and can po- Tuberculosis Trials Consortium. Rifapentine and iso-
tentially cause additional liver damage [62 – 64]. This niazid once a week versus rifampicin and isoniazid
damage can be life threatening for a person with twice a week for treatment of drug-susceptible pul-
marginal hepatic function [62 – 65]. Hepatic dysfunc- monary tuberculosis in HIV-negative patients: a ran-
tion can also alter absorption and distribution of domised clinical trial. Lancet 2002;360:528 – 34.
drugs that are metabolized or excreted by the liver [4] Weiner M, Burman W, Vernon A, et al. Tuberculosis
[65]. In the setting of severe liver disease, it is Trials Consortium. Low isoniazid concentrations and
outcome of tuberculosis treatment with once-weekly
reasonable to include fewer hepatotoxic medications
isoniazid and rifapentine. Am J Respir Crit Care Med
and to extend the period of treatment [6,65]. This 2003;167:1341 – 7.
change can be accomplished using a single hepato- [5] Weiner M, Bock N, Peloquin CA. Tuberculosis Trials
toxic drug, generally rifampin, in combination with Consortium. Pharmacokinetics of rifapentine at 600,
ethambutol, a quinolone, and an aminoglycoside. 900, and 1,200 mg during once-weekly tuberculosis
Isoniazid can be substituted for rifampin, if rifampin therapy. Am J Respir Crit Care Med 2004;169:1191 – 7.
cannot be given [6]. For these complicated patients, [6] American Thoracic Society, Centers for Disease Con-
expert opinion should be obtained [6]. trol and Prevention, and Infectious Diseases Society of
America. Treatment of tuberculosis. MMWR Recomm
Rep 2003;52(RR-11):1 – 77.
[7] Medical Research Council. Treatment of pulmonary
Summary
tuberculosis with streptomycin and para-amino-
salicylic Acid. BMJ 1950;2:1073 – 85.
Insights gained from studies done by the TBTC [8] McDermott W, Mushenheim C, Hadley SJ, et al.
have resulted in major changes in the recommenda- Streptomycin in the treatment of tuberculosis in
tions for drug therapy of tuberculosis in HIV-infected humans. Ann Intern Med 1947;27:769 – 822.
and noninfected individuals [1 – 5]. Although the [9] Tempel CW, Hughes FJ, Mardis RE, et al. Combined
goals for the treatment of tuberculosis remain the intermittent regimes employing streptomycin and para-
same, these advances require that tuberculosis drug aminosalicycylic acid in the treatment of pulmonary
therapy now be more individualized. Treatment regi- tuberculosis. Am Rev Tuberc 1951;63:295 – 311.
mens are based on a patient’s risk profile based on a [10] Canetti G, Grosset J. Teneur des souches sauvages de
mycobacterium tuberculosis en variants resistants a
combination of hematologic, microbiologic, clinical,
l’isoniazide et en variants resistants a la streptomycine
and radiographic findings [6]. Although they are sur milieu de Lowenstein-Jensen. Ann Inst Pasteur
more complicated than the previous guidelines, they (Paris) 1961;101:28.
allow treatment to be refined so that it can be ex- [11] Canetti G, Rist N, Grosset J. Primary drug resistance
tended in patients at high risk for treatment failure in tuberculosis. Am Rev Resp Dis 1964;90:792 – 9.
and allow shorter, more convenient treatment regi- [12] David HL. Probability distribution of drug-resistant
treatment of active tuberculosis 281
[53] Centers for Disease Control and Prevention. Tuber- and ethambutol. Am J Respir Crit Care Med 1999;159:
culosis associated with blocking agents against tumor 1580 – 4.
necrosis factor-alpha – California, 2002 – 2003. MMWR [60] Peloquin CA. Antituberculosis drugs: pharmacoki-
Morb Mortal Wkly Rep 2004;53(30):683 – 6. netics. In: Heifets L, editor. Drug susceptibility in the
[54] Bieber J, Kavanaugh A. Consideration of the risk and chemotherapy of mycobacterial infections. Boca Raton
treatment of tuberculosis in patients who have rheu- (FL)7 CRC Press; 1991. p. 59 – 88.
matoid arthritis and receive biologic treatments. [61] Strauss I, Erhardt F. Ethambutol absorption, excretion
Rheum Dis Clin North Am 2004;30(2):257 – 70. and dosage in patients with renal tuberculosis. Chemo-
[55] Snider DE, Layde PM, Johnson MW, et al. Treatment therapy 1970;15:148 – 57.
of tuberculosis during pregnancy. Am Rev Respir Dis [62] Ziersky M, Bek E. Side effects of drug regimens used
1980;122:65 – 79. in short course chemotherapy for pulmonary tuber-
[56] Medchill MT. Gillum. Diagnosis and management of culosis. Tubercle 1980;61:41 – 9.
tuberculosis during pregnancy. Obstset Gynecol Surv [63] Girling DJ. Adverse effects of antituberculosis drugs.
1989;44:81 – 9. Drugs 1982;23:56 – 74.
[57] Davidson PT. Managing tuberculosis during preg- [64] Yee D, Valiquette C, Pelletier M, et al. Incidence of
nancy. Lancet 1995;346:199 – 200. serious side effects from first-line antituberculosis
[58] Snider DE, Powell KE. Should women taking anti- drugs among patients treated for active tuberculosis.
tuberculosis drugs breast-feed? Arch Intern Med 1984; Am J Respir Crit Care Med 2003;167:1472 – 7.
144:589 – 90. [65] Mclean AJ, Morgan DJ. Clinical pharmacokinetics in
[59] Malone RS, Fish DN, Spiegel DM, et al. The effect patients with liver disease. Clin Pharmacokinet 1991;
of hemodialysis on isoniazid, rifampin, pyrazinamide, 21:42 – 69.
Clin Chest Med 26 (2005) 313 – 326
It is estimated that one third of the global popu- Diagnosis of latent Mycobacterium tuberculosis
lation, or 2 billion people, are infected with Myco- infection
bacterium tuberculosis [1]. Among infected persons,
approximately 10% progress to the clinically im- As discussed elsewhere in this issue, the diagno-
portant (and infectious) stage of active tuberculosis sis of latent M. tuberculosis infection relies primarily
over their lifetime [2,3]. The risk is higher in persons on the tuberculin skin test, an intradermal test that
with concomitant HIV infection (>20%), evidence of utilizes purified protein derivative (PPD). Because
old healed tuberculosis on chest radiograph (>20%), the mycobacterial proteins in PPD are not specific for
or recent M. tuberculosis infection (10% to 20%) [4]. M. tuberculosis, persons infected with other myco-
In most infected persons, the host immune response bacteria (eg, environmental mycobacteria such as
contains the replication of M. tuberculosis and pre- M. avium intracellulare and M. bovis, the organism
vents the development of disease [5]. Among infected in the tuberculosis vaccine bacille Calmette-Guerin)
persons who develop active disease, progression may result in a false-positive test. Conversely, the
occurs either shortly after initial infection (progres- tuberculin skin test has low sensitivity, particularly
sive primary disease) or subsequent to the initial in immunocompromised persons. Because of these
infection, when there is a breakdown in the host im- limitations, the definition of a positive tuberculin
mune response. A person is at greatest risk of pro- skin test varies according to the person’s risk of
gressing to active disease during the first 2 years tuberculosis infection and risk of progressing to
after infection with M. tuberculosis [2]. In addition active disease if infected (Box 1) [6]. The different
to HIV infection and recent M. tuberculosis infec- criteria for a positive test increase its sensitivity in
tion, other risk factors for progression to active tu- high-risk persons and specificity in low-risk persons.
berculosis include silicosis, diabetes mellitus, chronic
renal failure, malnutrition, weight loss, leukemia, lym-
phoma, cancer of the head, neck, and lung, gastrec-
tomy, and jejunoileal bypass surgery [6]. Indications for treatment of Mycobacterium
tuberculosis infection
0272-5231/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccm.2005.02.003 chestmed.theclinics.com
314 dooley & sterling
315
316 dooley & sterling
status, so tolerability data are also presented accord- HIV-seronegative persons, and to ensure uniformity
ing to HIV serostatus. Recommended doses of of recommendations, the ATS/CDC/IDSA guide-
specific drugs have been published previously [6]. lines recommend 9 months of isoniazid for HIV-
seropositive persons [6]. Although HIV-infected
persons are at increased risk of having a negative
Isoniazid tuberculin skin test, particularly with advanced im-
munosuppression, isoniazid is not substantially more
Effectiveness effective than placebo in preventing tuberculosis
in such persons, and therefore is not recommended
HIV-seronegative persons (Table 3) [11].
Isoniazid is the best-studied regimen for the
treatment of latent M. tuberculosis infection. More Toxicity
than 20 randomized, controlled trials have been con-
ducted, which together enrolled more than 100,000 HIV-seronegative persons
persons. Most of these studies were conducted in the In the initial studies of isoniazid, drug discon-
1950s and 1960s and therefore enrolled only HIV- tinuation rates were low and did not differ from rates
seronegative persons. These trials are summarized among persons receiving placebo [17]. Subsequent
in Table 1. studies, however, have noted higher rates of drug
Most of the studies assessed the effectiveness discontinuation, as summarized in Tables 4 and 5.
of 12 months of isoniazid versus placebo. The trial Few of these studies have been placebo-controlled.
conducted by the International Union Against Tuber- Isoniazid can cause elevated hepatic transaminases
culosis (IUAT) assessed 3 versus 6 versus 12 months [18], but these liver function abnormalities are often
of therapy and found that 6 months of isoniazid transient and are not representative of clinically
was less effective (65%) than 12 months (75%) [9]. significant hepatitis. In studies in which serum trans-
George Comstock [10] subsequently performed an aminases were monitored regularly regardless of
analysis of previously conducted clinical trials and symptoms, 10% to 22% of participants had at least
found that 6 months of isoniazid provided insuffi- one elevated transaminase level during the course
cient protection; 9 to 10 months of isoniazid seemed of therapy [19 – 25]. Rates of clinically significant
to provide optimal protection. Based on these find- hepatitis are lower. In a surveillance study of the US
ings, the American Thoracic Society (ATS), Centers Public Health Service, 236 of 13,838 persons (1.7%)
for Disease Control and Prevention (CDC), and In- who received isoniazid developed hepatitis. When
fectious Diseases Society of America (IDSA) guide- considering only those persons in whom the hepatitis
lines recommend 9 months of isoniazid [6]. A was probably or possibly related to isoniazid, the rate
9-month regimen of isoniazid has never been com- was 174 in 13,838 (1.3%) [26]. Hepatitis risk
pared with a 6- or 12-month course of isoniazid in a increased with age and concomitant alcohol con-
clinical trial, however. sumption. In another study, a 7-year survey from one
public health clinic, 11 of 11,141 patients (0.10%)
HIV-seropositive persons who started isoniazid developed hepatotoxicity [27].
Isoniazid effectiveness has also been studied in Isoniazid-associated hepatotoxicity can be fatal,
HIV-seropositive persons, although not as extensively and the risk of death increases with age. It is
as in HIV-seronegative persons. The randomized, estimated that the hepatotoxicity-associated case-
controlled trials of treatment of latent M. tuberculo- fatality rate per 10,000 persons initiating isoniazid
sis infection in HIV-infected persons are summarized treatment is 0 for ages 20 to 34 years, 2 for ages 35 to
in Table 2, and the randomized, placebo-controlled 49 years, and 4 for ages 50 to 64 years [24,26,28,29].
trials of isoniazid are summarized in Table 3. Among Although never tested in a trial, rates of hepato-
tuberculin skin-test – positive persons, isoniazid is toxicity may be lower when there is regular monitor-
clearly more effective than placebo in preventing tu- ing of signs and symptoms of hepatitis [27,28]. In
berculosis; this finding has been confirmed in a meta- the United States it is currently recommended that
analysis [11]. Isoniazid has been associated with patients receiving isoniazid undergo monthly clinical
improved survival in some studies [12 – 14], but not assessments for adverse effects. They should also be
in all [11,15,16]. Although there has not been a evaluated whenever symptoms develop. Patients
direct comparison of 6 versus 9 versus 12 months should be educated regarding the signs and symptoms
of isoniazid, 6 months seems to be less effective of hepatotoxicity and instructed to discontinue the
than 12 months. Based on the rationale used for medicine and seek clinical evaluation if symptoms
Table 2
Randomized, controlled trials of treatment of latent Mycobacterium tuberculosis infection among HIV-seropositive persons
Mean
First author/date [reference] Trial type Population Regimens N follow-up Compliance/follow-up
INH versus placebo trials
Pape, 1993 [12] Randomized Haiti INH 300 mg/d for 12 months 118 33 months No loss to follow-up
Double-blind New HIV
Placebo-controlled diagnosis
317
318 dooley & sterling
Table 3
Results of randomized, controlled trials of isoniazid for the treatment of latent Mycobacterium tuberculosis infection among
HIV-seropositive persons
Rx n/N Control n/N 95% Confidence
First author/date [reference] (%) INH (%) placebo RR interval Reduction (%)
PPD positive
Pape, 1993 [12] 2/38 (5.2) 6/25 (24.0) 0.22 0.05, 1.00 78
Hawken, 1997 [76] 5/67 (7.5) 8/69 (11.6) 0.64 0.22, 1.87 36
Whalen, 1997 [15] 7/536 (1.3) 21/464 (4.5) 0.29 0.12, 0.67 71
PPD+ cohort
Mwinga, 1998 [41] 6/101 (5.9) 11/60 (18.3) 0.32 0.13, 0.83 68
PPD negative
Gordin, 1997 [77] 4/260 (1.5) 6/257 (2.3) 0.66 0.19, 2.31 34
Fitzgerald, 2001 [78] 6/126 (4.8) 4/111 (3.6) 1.32 0.38, 4.56 32
Hawken, 1997 [76] 11/235 (4.7) 8/224 (3.6) 1.31 0.54, 3.20 31
Pape, 1993 [12] 2/20 (10.0) 5/35 (14.3) 0.70 0.15, 3.28 30
Whalen, 1997 [15] 9/395 (2.3) 10/323 (3.1) 0.74 0.30, 1.79 26
anergic cohort
Mwinga, 1998 [41] 27/351 (7.7) 17/166 (10.2) 0.75 0.42, 1.34 25
Results are presented according to the tuberculin skin test status of study patients.
Abbreviations: INH, isoniazid; PPD, purified protein derivative; RR, relative risk.
Table 5
Toxicity of isoniazid for treatment of latent Mycobacterium tuberculosis infection in HIV-seropositive patients
Toxicity requiring discontinuation of therapy 95% Confidence
First author/date [reference] INH n/N (%) Placebo n/N (%) RR interval
Pape, 1993 [12] 0/58(0) 0/60 (0) 0
Gordin, 1997 [77] 24/260 (9.2) 24/257 (9.3) 0.99 0.58, 1.69
Hawken, 1997 [76] 11/342 (3.2) 5/342 (1.5) 2.2 0.77, 6.26
Whalen, 1997 [15] PPD+ 3/536 (0.6) 1/464 (0.2) 2.60 0.27, 24.88
Anergic 0/395 0/323 0
Mwinga, 1998 [41] 26/703 (3.7) 3/350 (0.9) 4.31 1.32, 14.16
Toxicity is defined as adverse events resulting in discontinuation of therapy.
Abbreviations: INH, isoniazid; PPD, purified protein derivative; RR, relative risk.
Table 6
Randomized, controlled trials of the effectiveness of pyrazinamide plus rifampin for the treatment of latent Mycobacterium
tuberculosis infection in HIV-seropositive persons
Control n/N 95% Confidence
First author/date [reference] Rx n/N (%) RIF/PZA (%) INH RR interval Reduction (%)
PPD positive
Gordin, 2000 [36] 28/791 (3.5) 29/792 (3.7) 0.97 0.58, 1.61 3
Halsey, 1998 [40] 19/380 (5.0) 14/370 (3.8) 1.32 0.67, 2.56 32
Mwinga, 1998 [41] 2/49 (4.1) 4/52 (7.7) 0.53 0.10, 2.77 47
PPD negative
Mwinga, 1998 [41] 13/173 (7.5) 14/178 (7.9) 0.96 0.46, 1.96 4
Results are presented according to the tuberculin skin test status of study patients.
Abbreviations: INH, isoniazid; PPD, purified protein derivative; PZA, pyrazinamide; RIF, rifampin; RR, relative risk.
320 dooley & sterling
Table 7
Toxicity of rifampin plus pyrazinamide for treatment of latent Mycobacterium tuberculosis infection in HIV-seronegative patients
Toxicity requiring discontinuation Hepatotoxicity requiring discontinuation
of therapy of therapy
First author/date [reference] RIF/PZA n/N (%) INH n/N (%) RR RIF/PZA n/N (%) INH n/N (%) RR
a
Bock, 2001 [58] 13/168 (7.7) No INH arm 1/168 (0.6) No INH arm
a
Chaisson, 2002 [81] 12/589 (2.0) No INH arm 10/589 (1.7) No INH arm
Jasmer, 2002 [60] 28/307 (9.1) 8/282 (2.8) 3.21 12/207 (5.8) 2/204 (1.0) 5.91
a
Lee, 2002 [82] 26/148 (17.6) No INH arm 11/148(7.4) No INH arm
a
McNeill, 2003 [83] 14/110 (12.7) 5/114 (4.4) 2.90
a
Stout, 2003 [84] 8/114 (7.0) No INH arm 6/114 (5.3) No INH arm
Van Hest, 2004 [85] 14/166 (8.4) 17/528 (3.2) 2.62 14/166 (8.4) 18/528 (3.4) 2.47
Toxicity is defined as adverse events resulting in discontinuation of therapy.
Abbreviations: INH, isoniazid; PZA, pyrazinamide; RIF, rifampin; RR, relative risk.
a
In these studies, the patient population was predominantly HIV-seronegative, with a small minority of HIV-
seropositive patients.
with this regimen so that the risk of toxicity could HIV-seropositive persons
be determined [39]. Data were collected for persons The regimen was very well tolerated in all of the
initiating therapy between January 2000 and June clinical trials conducted among HIV-seropositive
2002 and reported to CDC by June 6, 2003. Of the persons (Table 8) [36,40,41], even upon repeat analy-
7737 persons reported to have started rifampin plus sis specifically addressing hepatotoxicity [42]. Be-
pyrazinamide treatment during the survey period, 204 cause of the relatively small sample sizes of the
persons developed aspartate aminotransferase con- studies, and an event rate of severe hepatotoxicity of
centrations more than five times the upper limit of approximately 1 per 1000, it is possible that such
normal (2.6/100 treatment initiations), and an addi- serious events were not detected in these studies. The
tional 146 patients discontinued therapy because of CDC therefore recommends that rifampin plus
symptoms of hepatitis (1.9/100 treatment initiations). pyrazinamide should generally not be offered, regard-
There were 48 cases of severe hepatotoxicity (de- less of HIV serostatus [39].
fined as resulting in hospitalization or death); 11 pa-
tients died. The estimated rate of death caused by
hepatotoxicity was 0.9 per 1000 treatment initiations
[39]. The estimates were limited because data were Isoniazid plus rifampin for 3 months
obtained retrospectively, and the risk associated with
isoniazid was not determined concurrently. Nonethe- Effectiveness
less, the risk of severe hepatotoxicity and death
seemed to be approximately 10 times greater than There are few studies of isoniazid plus rifampin
the risk associated with isoniazid, and it was there- for the treatment of latent M. tuberculosis infection.
fore recommended that rifampin plus pyrazinamide In the only study among HIV-seronegative adults,
should generally not be offered for treatment of the efficacy among adherent patients of 3 months of
latent tuberculosis infection [39]. The toxicity data isoniazid plus rifampin was 41% [43]. Among HIV-
from clinical trials are summarized in Table 7. seropositive adults, the regimen was 59% effective in
Table 8
Toxicity of rifampin plus pyrazinamide for treatment of latent Mycobacterium tuberculosis infection in HIV-seropositive patients
Toxicity leading to discontinuation of therapy 95% Confidence
First author/date [reference] RIF/PZA n/N (%) INH n/N (%) RR interval
Halsey, 1998 [40] 0/380 (0) 0/370 (0)
Mwinga, 1998 [41] 14/351 (4.0) 12/352 (3.4) 1.17 0.55, 2.50
Gordin, 2000 [36,42] 75/791 (9.5) 48/792 (6.1) 1.56 1.09, 2.22
Narita, 2003 [86] 5/135 (3.7) 0/25 (0)
Toxicity is defined as adverse events resulting in discontinuation of therapy.
Abbreviations: INH, isoniazid; PZA, pyrazinamide; RIF, rifampin; RR, relative risk.
treatment of latent tuberculosis infection 321
The only treatment regimen that has been exten- Difficulties and challenges of treatment of latent
sively studied in children is isoniazid. Isoniazid may Mycobacterium tuberculosis infection
be more effective in children than in adults, with a
reported effectiveness of 70% to 90% [49,50]. Be- Low rates of treatment initiation
cause the effectiveness of short-course regimens
has not been studied in children, such regimens are Not all persons who are eligible for treatment
not recommended by the ATS/CDC/IDSA [6]. of latent tuberculosis infection initiate therapy. In a
study of close contacts of smear-positive pulmonary
Contacts of persons with drug-resistant tuberculosis tuberculosis cases, 95 HIV-infected persons were
eligible for treatment of latent infection; of these,
If persons are infected with a strain of M. tu- only 30 (32%) initiated therapy [56]. In another study
berculosis that is resistant to rifampin but suscep- of close contacts of tuberculosis cases, of the
tible to isoniazid, isoniazid is effective for treatment. 630 persons with newly documented positive tuber-
For persons exposed to a case of tuberculosis resis- culin skin tests (all of whom were eligible for ther-
tant to isoniazid, treatment with rifampin is recom- apy), treatment was recommended in 447 (71%),
mended, as discussed previously [6]. The optimal and was started in only 398 (63%) [57]. Efforts must
treatment is unknown for persons with evidence of be made to improve the use of treatment of latent
latent M. tuberculosis infection who have been ex- infection, particularly in those at highest risk of pro-
posed to multidrug-resistant tuberculosis (MDR-TB), gression to active disease, such as HIV-infected
defined as resistance to at least isoniazid plus ri- persons and close contacts.
fampin. No clinical studies have assessed the effec-
tiveness of specific regimens, and such studies will Low treatment-completion rates
probably never be conducted because of the diffi-
culty in enrolling a sufficient sample size. Recom- Among persons who start therapy, treatment-
mended regimens include a fluoroquinolone plus completion rates are low. In the second study of
pyrazinamide or ethambutol plus pyrazinamide [51]. close contacts mentioned previously, 203 of 398 per-
The optimal duration of these regimens is unknown, sons starting therapy (51%) completed it; 203 of
although 6 to 12 months is recommended. They 630 of those eligible for therapy (32%) completed it
often are poorly tolerated [52]. [57]. In a study of isoniazid treatment of latent in-
fection in an inner-city population, 84 of 409 persons
Tumor necrosis factor-alpha antagonists eligible for therapy (21%) completed it [58]. Low
treatment-completion rates can result in continued
Drugs that block the inflammatory cytokine tu- transmission of M. tuberculosis and additional cases
mor necrosis factor-alpha (TNF-a), such as inflixi- [59]. The low completion rates may result from a lack
treatment of latent tuberculosis infection 323
of understanding on the part of the patient of the tine for 3 months. A study assessing the tolerability
importance of treatment, the absence of symptoms of this regimen has recently completed enrollment in
related to tuberculosis, the toxicity of the regimen, Brazil, and data analysis is underway. A compari-
and the prolonged duration of therapy. Even if son of isoniazid plus rifapentine versus the standard
toxicity is relatively mild and infrequent, patients regimen of daily isoniazid is underway in South
may be less likely to tolerate adverse effects because Africa and in a multinational study being conducted
they do not have symptomatic tuberculosis disease. by the Tuberculosis Trials Consortium of the CDC
Shorter treatment duration may improve completion (US Public Health Service Study 26). If effective
rates [36,40], but this result has not been noted and well-tolerated, isoniazid plus rifapentine would
uniformly [60]. Education of the patient regarding the provide a relatively simple short-course regimen for
importance of such therapy is extremely important. the treatment of latent tuberculosis infection, which
Direct observation of treatment of latent infection can could improve treatment completion rates.
improve adherence [61,62] but is often not feasible Data from the mouse model of tuberculosis, which
for tuberculosis treatment programs because of its has correlated closely with human tuberculosis
high cost. [63,64], have demonstrated that moxifloxacin, a
newer fluoroquinolone, has excellent activity against
Monitoring for toxicity M. tuberculosis. This drug may allow shorter and
more effective treatment of active tuberculosis [65,66]
As discussed previously, the potentially severe and may be effective for the treatment of latent tuber-
toxicity associated with the treatment of latent culosis infection. Additional studies are warranted.
M. tuberculosis infection necessitates monitoring, The Tuberculosis Epidemiologic Studies Consor-
particularly in persons at increased risk for toxic- tium of the CDC has launched a large study of the
ity. Routine monitoring for symptoms of toxicity is treatment of latent tuberculosis infection in the United
recommended. Monitoring of laboratory tests (eg, he- States (Task Order 13), including an assessment of
patic transaminases) to prevent toxicity, or at least the regimens used and the factors associated with
identify toxicity in its early stages, is a reasonable acceptance of, adherence to, and tolerability of cur-
approach in persons at high risk for toxicity, although rent treatment regimens. A better understanding of
the utility of such a strategy (and the optimal these factors will allow the development of in-
monitoring strategy) has never been assessed in a terventions to improve the treatment of latent
clinical trial. M. tuberculosis infection in the United States and
throughout the world.
Logistical difficulties in implementing treatment of
latent tuberculosis infection
References
Treatment of latent M. tuberculosis infection has
been a cornerstone of efforts to control tuberculosis [1] Dye C, Scheele S, Dolin P, et al. Global burden of
in the United States for the past several decades. This tuberculosis. Estimated incidence, prevalence, and
approach has been possible because of the relatively mortality by country. JAMA 1999;282:677 – 86.
[2] Sutherland I. Recent studies in the epidemiology of
low number of tuberculosis cases and available pub-
tuberculosis, based on the risk of being infected with
lic health resources in the United States. In areas of tubercle bacilli. Adv Tuberc Res 1976;19:1 – 63.
the world where tuberculosis case rates are substan- [3] Comstock GW, Livesay VT, Woolpert SF. The prog-
tially higher and resources are more limited, treat- nosis of a positive tuberculin reaction in childhood
ment of latent M. tuberculosis infection is limited; and adolescence. Am J Epidemiol 1974;99:131 – 8.
the focus in these countries is to identify and treat [4] Horsburgh Jr CR. Priorities for the treatment of latent
cases of active tuberculosis. tuberculosis infection in the United States. N Engl J
Med 2004;350(20):2060 – 7.
[5] Schluger NW, Rom WN. The host immune response to
Prospects for improvements in the treatment of tuberculosis. Am J Respir Crit Care Med 1998;157:
679 – 91.
latent Mycobacterium tuberculosis infection
[6] American Thoracic Society. Centers for Disease Con-
trol and Prevention. Targeted tuberculin testing and
There is clearly a need for safe and effective short- treatment of latent tuberculosis infection. Am J Respir
course treatment of latent M. tuberculosis infection. Crit Care Med 2000;161(4):S221 – 47.
Studies are underway to assess the tolerability and [7] Institute of Medicine. In: Geiter L, editor. Ending
effectiveness of once-weekly isoniazid plus rifapen- neglect: the elimination of tuberculosis in the United
324 dooley & sterling
States. Washington (DC)7 National Academy of Sci- [24] Byrd RB, Horn BR, Solomon DA, et al. Toxic effects
ences; 2000. of isoniazid in tuberculosis chemoprophylaxis. Role
[8] Manabe YC, Bishai WR. Latent Mycobacterium tuber- of biochemical monitoring in 1,000 patients. JAMA
culosis—persistence, patience, and winning by wait- 1979;241:1239 – 41.
ing. Nat Med 2000;6(12):1327 – 9. [25] Stuart RL, Wilson J, Grayson ML. Isoniazid toxicity in
[9] International Union Against Tuberculosis. Efficacy of health care workers. Clin Infect Dis 1999;28:895 – 7.
various durations of isoniazid preventive therapy for [26] Kopanoff DE, Snider Jr DE, Caras GJ. Isoniazid-
tuberculosis: five years of follow-up in the IUAT trial. related hepatitis: a U.S. Public Health Service Coop-
Bull World Health Organ 1982;60:555 – 64. erative Surveillance Study. Am Rev Respir Dis 1978;
[10] Comstock GW. How much isoniazid is needed for 117:991 – 1001.
prevention of tuberculosis among immunocompetent [27] Nolan CM, Goldberg SV, Buskin SE. Hepatoxicity
adults? Int J Tuberc Lung Dis 1999;3(10):847 – 50. associated with isoniazid preventive therapy: a 7-year
[11] Bucher HC, Griffith LE, Guyatt GH, et al. Isoniazid survey from a public health tuberculosis clinic. JAMA
prophylaxis for tuberculosis in HIV infection: a meta- 1999;281:1014 – 8.
analysis of randomized controlled trials. AIDS 1999; [28] Dash LA, Comstock GW, Flynn JPG. Isoniazid
13(4):501 – 7. preventive therapy. Am Rev Respir Dis 1980;121:
[12] Pape JW, Jean SS, Ho JL, et al. Effect of isonia- 1039 – 44.
zid prophylaxis on incidence of active tuberculosis [29] Snider Jr DE, Caras GJ. Isoniazid-associated hepatitis
and progression of HIV infection. Lancet 1993;342: deaths: a review of available information. Am Rev
268 – 72. Respir Dis 1992;145:494 – 7.
[13] Moreno S, Miralles P, Diaz MD, et al. Isoniazid pre- [30] M. Tuberculosis Chemotherapy Centre. The prevention
ventive therapy in human immunodeficiency virus- and treatment of isoniazid toxicity in the therapy of
infected persons. Long-term effect on development of pulmonary tuberculosis. Bull World Health Organ
tuberculosis and survival. Arch Intern Med 1997; 1963;29:457 – 81.
157(15):1729 – 34. [31] Snider Jr DE. Pyridoxine supplementation during
[14] de Pinho AM, Santoro-Lopes G, Harrison LH, et al. isoniazid therapy. Tubercle 1980;61(4):191 – 6.
Chemoprophylaxis for tuberculosis and survival of [32] Ungo JR, Jones D, Ashkin D, et al. Antituberculosis
HIV-infected patients in Brazil. AIDS 2001;15(16): drug-induced hepatotoxicity. The role of hepatitis C
2129 – 35. virus and the human immunodeficiency virus. Am J
[15] Whalen CC, Johnson JL, Okwera A, et al. A trial of Respir Crit Care Med 1998;157:1871 – 6.
three regimens to prevent tuberculosis in Ugandan [33] Thomas DL, Vlahov D, Solomon L, et al. Correlates of
adults infected with the human immunodeficiency vi- hepatitis C virus infections among injection drug users.
rus. N Engl J Med 1997;337:801 – 8. Medicine 1995;74:212 – 20.
[16] Quigley MA, Mwinga A, Hosp M, et al. Long-term [34] Sadaphal P, Astemborski J, Graham NM, et al.
effect of preventive therapy for tuberculosis in a co- Isoniazid preventive therapy, hepatitis C virus infec-
hort of HIV-infected Zambian adults. AIDS 2001; tion, and hepatotoxicity among injection drug users
15(2):215 – 22. infected with Mycobacterium tuberculosis. Clin Infect
[17] Ferebee SH, Mount FW. Tuberculosis morbidity in a Dis 2001;33(10):1687 – 91.
controlled trial of the prophylactic use of isoniazid [35] Fernandez-Villar A, Sopena B, Vazquez R, et al.
among household contacts. Am Rev Respir Dis 1962; Isoniazid hepatotoxicity among drug users: the role
85:490 – 521. of hepatitis C. Clin Infect Dis 2003;36(3):293 – 8.
[18] Scharer L, Smith JP. Serum transaminase elevations [36] Gordin FM, Chaisson RE, Matts JP, et al. Rifampin
and other hepatic abnormalities in patients receiving and pyrazinamide vs. isoniazid for prevention of tu-
isoniazid. Ann Intern Med 1969;71:1113 – 20. berculosis in HIV-infected persons. An international
[19] Bailey WC, Weill H, DeRouen TA, et al. The effect randomized trial. JAMA 2000;283:1445 – 50.
of isoniazid on transaminase levels. Ann Intern Med [37] Centers for Disease Control and Prevention. Fatal and
1974;81:200 – 2. severe hepatitis associated with rifampin and pyrazin-
[20] Mitchell JR, Long MW, Thorgeirsson UP, et al. amide for the treatment of latent tuberculosis infec-
Acetylation rates and monthly liver function tests tion—New York and Georgia, 2000. MMWR Morb
during one year of isoniazid preventive therapy. Chest Mortal Wkly Rep 2001;50:289 – 91.
1975;68:181 – 90. [38] Centers for Disease Control and Prevention. Update:
[21] Litt IF, Cohen MI, McNamara H. Isoniazid hepatitis fatal and severe liver injuries associated with rifampin
in adolescents. J Pediatr 1976;89:133 – 5. and pyrazinamide for latent tuberculosis infection, and
[22] Byrd RB, Nelson R, Elliott RC. Isoniazid toxicity. A revisions in American Thoracic Society/CDC recom-
prospective study in secondary chemoprophylaxis. mendations—United States, 2001. MMWR Morb
JAMA 1972;220:1471 – 3. Mortal Wkly Rep 2001;50:733 – 5.
[23] Dickinson DS, Bailey WC, Hirschowitz BI, et al. Risk [39] Update. Adverse event data and revised American
factors for isoniazid (INH)-induced liver dysfunction. Thoracic Society/CDC recommendations against the
J Clin Gastroenterol 1981;3:271 – 9. use of rifampin and pyrazinamide for treatment of
treatment of latent tuberculosis infection 325
latent tuberculosis infection—United States, 2003. [55] Centers for Disease Control. Tuberculosis associated
MMWR Morb Mortal Wkly Rep 2003;52:735 – 9. with blocking agents against tumor necrosis factor-
[40] Halsey NA, Coberly J, Desormeaux J, et al. Random- alpha—California, 2002-2003. MMWR Morb Mortal
ised trial of isoniazid versus rifampicin and pyrazin- Wkly Rep 2004;53:683 – 6.
amide for prevention of tuberculosis in HIV-1 [56] Centers for Disease Control and Prevention. Missed
infection. Lancet 1998;351:786 – 92. opportunities for prevention of tuberculosis among
[41] Mwinga A, Hosp M, Godfrey-Faussett P, et al. Twice persons with HIV infection—selected locations, United
weekly tuberculosis preventive therapy in HIV infec- States, 1996-1997. MMWR Morb Mortal Wkly Rep
tion in Zambia. AIDS 1998;12(18):2447 – 57. 2000;49:685 – 7.
[42] Gordin FM, Cohn DL, Matts JP, et al. Hepatotoxicity [57] Reichler MR, Reves R, Bur S, et al. Treatment of la-
of rifampin and pyrazinamide in the treatment of latent tent tuberculosis infection in contacts of new tuber-
tuberculosis infection in HIV-infected persons: is it culosis cases in the United States. South Med J 2002;
different than in HIV-uninfected persons? Clin Infect 95(4):414 – 20.
Dis 2004;39(4):561 – 5. [58] Bock NN, Rogers T, Tapia JR, et al. Acceptability of
[43] Hong Kong Chest Service/British Medical Research short-course rifampin and pyrazinamide treatment of
Council. A double-blind placebo controlled clinical latent tuberculosis infection among jail inmates. Chest
trial of three antituberculosis chemoprophylaxis regi- 2001;119(3):833 – 7.
mens in patients with silicosis in Hong Kong. Am Rev [59] Centers for Disease Control and Prevention. Trans-
Respir Dis 1992;145:36 – 41. mission of Mycobacterium tuberculosis associated
[44] Ormerod LP. Rifampicin and isoniazid prophylactic with failed completion of treatment for latent tuber-
chemotherapy for tuberculosis. Arch Dis Child 1998; culosis infection—Chickasaw County, Mississippi,
78(2):169 – 71. June 1999-March 2002. MMWR Morb Mortal Wkly
[45] Steele MA, Burk RF, DesPrez RM. Toxic hepatitis Rep 2003;52:222 – 4.
with isoniazid and rifampin. A meta-analysis. Chest [60] Jasmer RM, Saukkonen JJ, Blumberg HM, et al. Short-
1991;99(2):465 – 71. course rifampin and pyrazinamide compared with
[46] Polesky A, Farber HW, Gottlieb DJ, et al. Rifampin isoniazid for latent tuberculosis infection: a multicenter
preventive therapy for tuberculosis in Boston’s home- clinical trial. Ann Intern Med 2002;137(8):640 – 7.
less. Am J Respir Crit Care Med 1996;154:1473 – 7. [61] Heal G, Elwood RK, FitzGerald JM. Acceptance and
[47] Villarino ME, Ridzon R, Weismuller PC, et al. Ri- safety of directly observed versus self-administered
fampin preventive therapy for tuberculosis infection. isoniazid preventive therapy in aboriginal peoples
Experience with 157 adolescents. Am J Respir Crit in British Columbia. Int J Tuberc Lung Dis 1998;2:
Care Med 1997;155:1735 – 8. 979 – 83.
[48] Menzies D, Dion MJ, Rabinovitch B, et al. Treatment [62] Chaisson RE, Link-Barnes G, Hackman J, et al. A
completion and costs of a randomized trial of rifampin randomized, controlled trial of interventions to im-
for 4 months versus isoniazid for 9 months. Am J prove adherence to isoniazid therapy to prevent tuber-
Respir Crit Care Med 2004;170(4):445 – 9. culosis in injection drug users. Am J Med 2001;110:
[49] Comstock GW, Hammes LM, Pio A. Isoniazid 610 – 5.
prophylaxis in Alaskan boarding schools: a com- [63] Grosset J, Lounis N, Truffot-Pernot C, et al. Once-
parison of two doses. Am Rev Respir Dis 1969;100: weekly rifapentine-containing regimens for treatment
773 – 9. of tuberculosis in mice. Am J Respir Crit Care Med
[50] Mount FW, Ferebee SH. Preventive effects of isonia- 1998;157(5 Pt 1):1436 – 40.
zid in the treatment of primary tuberculosis in children. [64] The Tuberculosis Trials Consortium. Rifapentine and
N Engl J Med 1961;265:713 – 21. isoniazid once a week versus rifampicin and isoniazid
[51] Centers for Disease Control and Prevention. Manage- twice a week for treatment of drug-susceptible pul-
ment of persons exposed to multidrug-resistant tu- monary tuberculosis in HIV-negative patients: a ran-
berculosis. MMWR Morb Mortal Wkly Rep 1992;41: domized clinical trial. Lancet 2002;360:528 – 34.
61 – 71. [65] Nuermberger EL, Yoshimatsu T, Tyagi S, et al.
[52] Ridzon R, Meador J, Maxwell R, et al. Asymptomatic Moxifloxacin-containing regimen greatly reduces time
hepatitis in persons who received alternative preven- to culture conversion in murine tuberculosis. Am J
tive therapy with pyrazinamide and ofloxacin. Clin Respir Crit Care Med 2004;169(3):421 – 6.
Infect Dis 1997;24(6):1264 – 5. [66] Nuermberger EL, Yoshimatsu T, Tyagi S, et al.
[53] Keane J, Gershon S, Wise RP, et al. Tuberculosis Moxifloxacin-containing regimens of reduced duration
associated with infliximab, a tumor necrosis factor produce a stable cure in murine tuberculosis. Am J
alpha-neutralizing agent. N Engl J Med 2001;345(15): Respir Crit Care Med 2004;170:1131 – 4.
1098 – 104. [67] Mount FW, Ferebee SH. The effect of isoniazid
[54] Gardam MA, Keystone EC, Menzies R, et al. Anti- prophylaxis on tuberculosis morbidity among house-
tumour necrosis factor agents and tuberculosis risk: hold contacts of previously known cases of tuber-
mechanisms of action and clinical management. Lancet culosis. Am Rev Respir Dis 1962;85:821 – 7.
Infect Dis 2003;3(3):148 – 55. [68] Ferebee SH, Mount FW, Murray FJ, et al. A con-
326 dooley & sterling
trolled trial of isoniazid prophylaxis in mental institu- negative HIV-infected adults living in a country with
tions. Am Rev Respir Dis 1963;88:161 – 75. endemic tuberculosis: results of a randomized trial.
[69] Comstock GW, Ferebee SH, Hammes LM. A con- J Acquir Immune Defic Syndr 2001;28(3):305 – 7.
trolled trial of community-wide isoniazid prophylaxis [79] Bailey WC, Taylor SL, Dascomb HE, et al. Disturbed
in Alaska. Am Rev Respir Dis 1967;95:935 – 43. hepatic function during isoniazid prophylaxis. Am Rev
[70] Chiba Y, Takahara T, Kondo K, et al. Chemoprophy- Respir Dis 1973;107:523 – 9.
laxis of tuberculosis for adults in Japan. Bull Int Union [80] Byrd RB, Horn BR, Griggs GA, et al. Isoniazid
Tuberc 1963;69:91 – 3. chemoprophylaxis. Association with detection and in-
[71] Nyboe J, Farah AR, Christensen OW. Report on cidence of liver toxicity. Arch Intern Med 1977;137:
tuberculosis chemotherapy pilot project (Tunisia 9). 1130 – 3.
WHO/TB/Tech Information/10. April 22, 1963. [81] Chaisson RE, Armstrong J, Stafford J, et al. Safety
[72] Egsmose T, Ang’awa JOW, Poti SJ. The use of and tolerability of intermittent rifampin/pyrazinamide
isoniazid among household contacts of open cases for the treatment of latent tuberculosis infection in
of pulmonary tuberculosis. Bull World Health Organ prisoners. JAMA 2002;288(2):165 – 6.
1965;33:419 – 33. [82] Lee AM, Mennone JZ, Jones RC, et al. Risk factors for
[73] Del Castillo H, Bautista LD, Jacinto CP, et al. hepatotoxicity associated with rifampin and pyrazin-
Chemoprophylaxis in the Philippines. A controlled amide for the treatment of latent tuberculosis infection:
pilot study among household contacts of tuberculosis experience from three public health tuberculosis
cases. Bulletin of the Quezon Institute 1965;7:277 – 90. clinics. Int J Tuberc Lung Dis 2002;6(11):995 – 1000.
[74] Horwitz O, Payne PG, Wilbek E. Epidemiological [83] McNeill L, Allen M, Estrada C, et al. Pyrazinamide
basis of tuberculosis eradication. 4. The isoniazid and rifampin vs isoniazid for the treatment of latent
trial in Greenland. Bull World Health Organ 1966;35: tuberculosis: improved completion rates but more
509 – 29. hepatotoxicity. Chest 2003;123(1):102 – 6.
[75] Veening GJJ. Long term isoniazid prophylaxis. Con- [84] Stout JE, Engemann JJ, Cheng AC, et al. Safety of
trolled trial on INH prophylaxis after recent tuberculin 2 months of rifampin and pyrazinamide for treatment
conversion in young adults. Bull Int Union Tuberc of latent tuberculosis. Am J Respir Crit Care Med
1968;41:169 – 71. 2003;167(6):824 – 7.
[76] Hawken MP, Meme HK, Elliott LC, et al. Isoniazid [85] van Hest R, Baars H, Kik S, et al. Hepatotoxicity of
preventive therapy for tuberculosis in HIV-1-infected rifampin-pyrazinamide and isoniazid preventive ther-
adults: results of a randomized controlled trial. AIDS apy and tuberculosis treatment. Clin Infect Dis 2004;
1997;11(7):875 – 82. 39(4):488 – 96.
[77] Gordin FM, Matts JP, Miller C, et al. A controlled [86] Narita M, Kellman M, Franchini DL, et al. Short-
trial of isoniazid in persons with anergy and human course rifamycin and pyrazinamide treatment for latent
immunodeficiency virus infection who are at high risk tuberculosis infection in patients with HIV infection:
for tuberculosis. N Engl J Med 1997;337:315 – 20. the 2-year experience of a comprehensive community-
[78] Fitzgerald DW, Severe P, Joseph P, et al. No effect of based program in Broward County, Florida. Chest
isoniazid prophylaxis for purified protein derivative- 2002;122(4):1292 – 8.
Clin Chest Med 26 (2005) 295 – 312
Tuberculosis in Children
Kristina Feja, MD, MPH, Lisa Saiman, MD, MPH*
Department of Pediatrics, College of Physicians and Surgeons, Columbia University, 622 West 168th Street,
PH4West, New York, NY 10032, USA
Tuberculosis (TB) accounts for significant mor- examines aspects of TB epidemiology, pathogenesis,
bidity and mortality in children and adolescents clinical presentations, diagnosis, treatment, and pub-
worldwide, with the majority of cases of latent TB in- lic health issues unique to children.
fection (LTBI) and disease occurring in developing
countries. Pediatric TB, or childhood TB, defined
by the World Health Organization (WHO) and Cen-
ters for Disease Control and Prevention (CDC) as TB Epidemiology of childhood tuberculosis
in children less than 15 years of age, presents health
care providers with unique challenges. In contrast to The knowledge of the global epidemiology of TB
adults and older adolescents, the clinical manifesta- in children is somewhat limited. In 1990, approxi-
tions of TB disease in children are usually related to mately 7,500,000 TB cases occurred worldwide, of
primary tuberculosis. Diagnostic and therapeutic chal- which 650,000 occurred in children [1]. In 2002, the
lenges arise because children have less specific signs WHO estimated that 8,800,000 TB cases occurred
and symptoms of disease, have fewer positive myco- worldwide, based on data from 209 countries with an
bacterial cultures, are at increased risk for progres- estimated mean case rate of 145 per 100,000 popu-
sion of disease once infected, and are at an increased lation (range, 2 – 1067 cases) [2]. Although estimates,
risk of disseminated disease. these numbers can be particularly useful when assess-
Pediatric TB is considered a sentinel event ing the effectiveness of TB control programs.
reflecting recent transmission from an undiagnosed Currently, the WHO reports only acid-fast bacillus
infectious source case in the community. Performing (AFB) smear – positive cases stratified by age. There-
contact investigations in adults with infectious TB to fore, age-specific estimates of all cases (smear-
detect other cases of TB or LTBI is an important negative and smear-positive) are unavailable. Data
means of diagnosing or preventing TB in children. from the WHO reflecting 2002 regional notification
Public health efforts vary according to country and rates of smear-positive cases among children ranged
available resources. Developed countries focus on from 0 per 100,000 persons in Europe to 6 per
identification and treatment of LTBI as well as active 100,000 in Africa and 43 and 45 per 100,000 in
disease. Children who have LTBI represent the res- South Africa and Zambia, respectively [2]. Reporting
ervoir for future disease. In resource-poor countries, only AFB smear – positive cases potentially has a
identification and treatment of TB disease, using the huge impact on accurate estimates of pediatric cases,
strategy of directly observed therapy (DOT), if avail- because young children have lower organism burdens
able, are the primary public health efforts. This article than adults and adolescents, and they are less likely
to be AFB smear – positive. Approximately 95% of
children less than 12 years of age are AFB smear –
negative [3]. Other challenges in accurately estimating
* Corresponding author. the burden of TB in children include the difficulties
E-mail address: LS5@columbia.edu (L. Saiman). establishing a definitive diagnosis, the increased
0272-5231/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccm.2005.02.010 chestmed.theclinics.com
296 feja & saiman
prevalence of extrapulmonary disease, the lack of a experienced an increase in TB case rates in the late
standard clinical case definition, and the generally 1980s and early 1990s with case rates of pediatric TB
lower public health priority given to childhood TB increasing from 2.4 cases per 100,000 persons (1261
compared with adult TB in many countries [3]. cases) in 1985 to 3.1 cases per 100,000 (1708 cases)
Nelson and Wells [3] reviewed recent epidemio- in 1992 [5]. The cause of this resurgence was
logic studies and surveillance data describing trends multifactorial. Contributing factors included the
in the global burden of TB in children. Childhood TB HIV epidemic, the dismantling of national TB control
accounted for only 2% to 7% of TB cases in programs, the increased immigration of persons from
developed countries, compared with 15% to 40% of countries with a high prevalence of TB, and, perhaps,
cases in developing countries. This difference may be improved reporting [6]. Happily, since 1993, there
explained partially by an older population structure in has been a steady decline of TB in the United States.
developed countries. Case rates in children seem to Nelson et al [7] recently reviewed the epidemiology
be increasing in Africa, countries of the former Soviet of 11,480 cases of childhood TB reported in the
Union, and some countries in Europe, the Middle United States between 1993 and 2001 and reported
East, and South America such as Sweden, England, that the number of childhood TB cases declined by
Wales, Greenland, Austria, Denmark, Israel, and 44%, from 1663 cases in 1993 to 931 cases in 2001,
Brazil. Possible reasons for this trend include with a respective 48% decline in incidence rates from
increasing immigration of children or families born 2.9 to 1.5 cases per 100,000 persons (Fig. 1). This
in countries with high case rates of TB, worsening decline, however, is less pronounced than that seen in
economic conditions, rising HIV infection rates, and a adults (Fig. 1). Seventy percent of all cases of
weakening public health infrastructure [3,4]. Con- pediatric TB occurred in eight states: California,
versely, Turkey and Peru have experienced a decline Texas, New York, Illinois, Georgia, Florida, New Jer-
in case rates among children. In Peru, the decline in sey, and Pennsylvania, and most cases presented in
case rates has been attributed to the institution of an urban areas with populations greater than 2,500,000.
aggressive DOT short-course program. Nelson and In the United States, a disproportionate burden of
Wells [3,4], however, cited several limitations of cases is seen among younger children, racial and
these epidemiologic data which included possible ethnic minorities, and those who are foreign-born
reporting bias because countries with smaller TB (Figs. 2, 3). In 2001, case rates per 100,000 persons
burdens may be less likely to report cases and some were 2.8, 1.0, and 0.9 in children less than 5 years of
studies included only hospitalized patients. Compar- age, 5 to 9 years of age, and 10 to 14 years of age,
ing rates across countries may not be feasible because respectively. From 1993 to 2001, 74% of pediatric
of variations in case definition, case finding, and TB cases occurred in Hispanic and non-Hispanic
contact investigation protocols. blacks and approximately 30% occurred in foreign-
The epidemiology of TB in the United States has born children [7]. Approximately two thirds of
undergone substantial changes during the past 20 foreign-born children with pediatric TB were born
years. After decades of decline, the United States in the following countries: Mexico (39.8%), the Phil-
20
15
10
0
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
<15 15-24 25-44 45-64 65+
Age Group (years)
Fig. 1. Tuberculosis case rates (per 100,000) by age group in the United States from 1993 to 2003. (From Centers for Disease
Control and Prevention. 2003 Surveillance slide sets. Available at: http://www.cdc.gov/nchstp/tb/pubs/slidesets/surv/surv2003/
default.htm. Accessed March 14, 2005.)
tuberculosis in children 297
800
Pediatric TB Cases
700
600
500
400
300
200
100
0
90
91
92
94
96
97
99
01
93
95
98
00
02
19
19
19
19
19
19
19
20
19
19
19
20
20
Year
White, non-Hispanic Black, non-Hispanic
Hispanic American Indian/Alaskan Native
Asian/Pacific Islander
Fig. 2. Pediatric tuberculosis cases by race and ethnicity from 1990 to 2002. (From Centers for Disease Control and Prevention.)
ippines (8.9%), Vietnam (5.7%), Somalia (4.4%), family member with LTBI, contact with a high-risk
Russia and the former countries of the Soviet Union adult (eg, an adult who is infected with HIV/AIDS,
(3.5%), and Haiti (3.3%). Approximately 60% of homeless, incarcerated, or a user of illicit drug) and
foreign-born cases were reported from three states: age greater than 11 years of age [8 – 11].
California (38.5%; n = 1070), New York (11.4%; Approximately one third of the world’s population
n = 317), and Texas (8.9%; n = 247) [7]. is infected with Mycobacterium tuberculosis. Most
infected persons do not progress to disease. In the
absence of treatment for LTBI, 5% to 10% of
Risk factors for latent tuberculosis infection and immunologically normal adults develop TB disease
progression to tuberculosis disease during their lifetimes, and half of the risk occurs in
the first 2 to 3 years after infection [12]. Infected
Several recent studies have assessed the risk children have a comparatively higher risk of pro-
factors for LTBI among children in the United States. gression to active disease: 43% of infants less than
Risk factors varied somewhat from study to study but 1 year of age, 24% of children 1 to 5 years old, and
generally included close contact with a TB case, birth 15% of those 11 to 15 years old develop TB disease if
in a country with high prevalence of TB, travel to or a not treated for LTBI [6]. Factors that increase the risk
household visitor from a high-prevalence country, a of progression from infection to disease usually affect
1400
Pediatric TB Cases
1200
1000
800
600
400
200
0
90
91
92
93
94
95
96
97
98
99
00
01
02
19
19
19
19
19
19
19
19
19
19
20
20
20
Year
Foreign-born U.S.-born
Fig. 3. Pediatric tuberculosis cases by birth country from 1990 to 2002. (From Centers for Disease Control and Prevention.)
298 feja & saiman
the immune system and include immunosuppressive or disseminate further. Progression within the lung,
therapy, HIV coinfection, malnutrition, medical con- pulmonary primary disease, is the most common
ditions (eg, renal and liver failure, diabetes mellitus, manifestation of primary TB and presents as enlarge-
or cancer), TB infection within the past 2 years, age ment of the affected regional lymph nodes. The
of 4 years or younger, or intercurrent viral infections primary focus may be smaller and may not be
such as measles [6,12,13]. visualized radiographically. Lymphadenopathy may
lead to further spread of disease by partial or com-
plete bronchial obstruction, by erosion of the bron-
Pathogenesis chial wall resulting in endobronchitis, or by external
compression of the bronchus resulting in segmental
The pathogenesis of pediatric TB is largely similar lesions caused by localized hyperinflation and atel-
to that described for adults and older adolescents, but ectasis of the adjacent lung tissue.
subtle differences in pathogenesis can lead to differ- Local progression in the lung may also occur from
ences in clinical presentations. As in adults, more the primary focus, particularly if the majority of
than 98% of infections in children occur when bacilli remain within this component of the primary
M. tuberculosis bacilli enter the lungs through complex. In young children, the primary focus
aerosolized droplets expelled when an infectious generally continues to grow even after the develop-
adult coughs, sneezes, or sings [6]. Other less ment of cellular immunity and may caseate centrally,
common portals of entry include the gastrointestinal liquefy, and empty into the bronchi resulting in
tract, the skin, mucous membranes, and conjunctiva. further spread [13]. This phenomenon is known as
When inhalation is the route of infection, the progressive pulmonary primary tuberculosis.
organisms multiply in alveolar macrophages, thus Progression of disease beyond the lungs occurs
creating a small area of inflammatory reaction known more frequently in younger children [13]. Younger
as the primary focus [14]. Early in infection, tubercle children are more likely to have involvement of distal
bacilli may also spread to the regional lymph nodes lymph nodes such as the mediastinal nodes, which
or disseminate through the bloodstream. The pri- may also compress or invade bronchi, blood vessels,
mary focus, regional lymph nodes, and adjoining or lymphatics. Occasionally disease progression oc-
lymphatics are collectively known as the primary curs by local spread to the pleural space, mediasti-
tuberculous complex. The pathology of the primary num, esophagus, or pericardium. Progression to these
complex is the same, regardless of the site of sites most often occurs through the bloodstream,
infection. Cellular immunity develops, usually 2 to however. Mycobacteria disseminated by the blood-
12 weeks after infection, and manifests as delayed stream can cause extrapulmonary disease (eg, super-
hypersensitivity to tuberculin, that is, a positive ficial lymphadenopathy generally of the cervical
tuberculin skin test (TST). At this time, the primary nodes, miliary TB, meningitis, or osteoarticular
focus becomes encapsulated, and perifocal inflamma- TB). Meningitis develops when caseating lesions on
tion increases and manifests as a granuloma, which the cerebral cortex invade the meninges and dissemi-
may be visible on chest radiograph. Radiographic nate into the subarachnoid space. Tuberculomas, a
evidence of perifocal infiltration is most common in less frequent manifestation of central nervous system
younger children [13]. In most cases, cellular immu- disease, form when caseous foci within the brain
nity controls the spread of infection. The pathologic enlarge and become encapsulated. Miliary TB occurs
events associated with this process include caseous when large numbers of bacilli disseminate through
necrosis, fibrosis, and healing of the primary complex the bloodstream and cause simultaneous disease in
components with or without calcification. The extent two or more organs [13].
of calcification depends on the relative degree of The pathogenesis of different forms of childhood
necrosis and caseation [13] and occurs sooner after TB may be viewed as a continuum in which host
infection in infants. Although this sequence of events factors play an important role in controlling the extent
occurs in LTBI [14], the resolution of infection may of disease. Immunologic control of TB infection is a
not be complete, and viable M. tuberculosis may function of macrophages and dendritic cells, man-
persist for many years, potentially reactivating and ifested by a Th1-type T-cell immunity characterized
causing TB disease months to years later. by strong CD8-positive cell response with production
Primary tuberculosis, defined as disease progres- of interferon gamma by CD4-positive cells. Children
sion of any part of the primary complex, is most have a relative deficiency of macrophage and
common in younger children. Disease from a primary dendritic cell function, however, and, in contrast with
pulmonary source may extend to the adjacent lung adults, tend to develop Th2-type T-cell responses to
tuberculosis in children 299
Fig. 6. Segmental atelectasis of the right upper lobe with Fig. 8. CT scan demonstrating bilateral upper lobe cavities
hyperinflation of right middle and lower lobe caused by ball- and infiltrates in a 14-year-old girl.
valve effect in an 8-month-old girl demonstrating partial
obstruction of the bronchus intermedius by hilar adenopathy
(anteroposterior view). Progressive pulmonary primary TB presents with
weight loss or failure to gain weight, anorexia,
fatigue, low-grade fevers, and intermittent cough.
patient populations, isolated hilar adenopathy has Rales may persist for days to weeks after the onset of
increased in frequency during the past 2 decades, initial symptoms [13]. Advanced disease may result
possibly because of more efficient contact investiga- in the development of a cavity or endobronchial
tions of infectious TB cases and subsequent detection spread [13].
of disease in children at an earlier stage of illness Chronic pulmonary TB or adult-type reactivation
[22]. The most common parenchymal findings in- disease is seen most commonly in adolescents whose
clude segmental hyperinflation and atelectasis caused primary infection occurred after 7 years of age [13].
by bronchial obstruction (Fig. 6), alveolar consoli- Like adults, such patients present with fever, weight
dation, interstitial densities, and, occasionally, bron- loss, a productive cough, hemoptysis, and night
chiectasis (Fig. 7) [21]. sweats. Radiographic findings include cavitary le-
sions, typically in the upper lobes (Fig. 8). An atlas of
radiographic images of intrathoracic TB disease in
children is available at www.iuatld.org/pdf/iuatld_
atlas.pdf. Recent United States data indicated that,
when compared with younger children, children 10 to
14 years of age had significantly higher rates of
cavitary disease (11.8% versus 3.5%; P< 0.001),
AFB smear – positive sputum (10.3% versus 1.7%;
P< 0.001), and culture-positive sputum (21.3% ver-
sus 4.2 – 5.0%; P< 0.001) [7].
Extrapulmonary tuberculosis
Tuberculosis in newborns
culous lesions in the infant, exclusion of postnatal cations in the lung parenchyma, nonenlarged
transmission by thorough contact investigation of regional lymph node, or both
close contacts including health care workers, and one The lack of the previously described clinical signs
of the following: or symptoms of disease
Lesions in the first week of life In recent years, a paradigm shift to targeted
Primary complex in the liver or caseating he- tuberculin skin testing has taken place in the United
patic granulomas States. This concept suggests that children should be
TB infection of the maternal genital tract or placenta screened for known risk factors for TB, elicited on a
risk factor questionnaire. Only children with one or
Congenital TB most often presents in the second more risk factors should be skin tested, thereby
to third week of life and may mimic other congenital increasing the sensitivity and specificity of the TST
infections such as syphilis, cytomegalovirus, or neo- (Box 2).
natal sepsis. Hepatosplenomegaly, respiratory distress, Current recommendations for screening, diag-
fever, lymphadenopathy, and abdominal distention nosis, and treatment of LTBI in children are found
are the most common signs and symptoms [37]. Most in a consensus document prepared by the Pediatric
infants have abnormal chest radiographs, usually Tuberculosis Collaborative Group entitled Targeted
showing a miliary pattern, hilar and mediastinal Tuberculin Skin Testing and Treatment of Latent
lymphadenopathy, or parenchymal infiltrates [33, Tuberculosis Infection in Children and Adoles-
37,38], and, less commonly, multiple rim-enhancing cent [19].
pulmonary nodules with central hypodense areas
[38]. In many cases, infants require mechanical Case definitions for tuberculosis disease
ventilation. Delayed diagnosis and onset of treatment
contribute to the approximately 40% mortality [37]. Although AFB smears and culture remain the
Perinatal TB is acquired from postnatal trans- cornerstones of diagnosing TB among adults, chil-
mission from the mother, adult caregiver, health care dren are often diagnosed clinically based on a
worker, or other infectious source and presents later positive TST, clinical and radiographic findings
than congenital TB. Clinical manifestations may be suggestive of TB, or a history of contact with an
similar to those noted in congenital TB. Of 38 South adult with TB [21,41]. Both the WHO and the CDC
African infants (including 7 cases of congenital TB) employ laboratory-confirmed or clinician-diagnosed
who were less than 3 months of age when diagnosed case definitions. In the United States, the laboratory
with TB, 87% had cough, 82% had tachypnea, 66% case definition consists of (1) isolation of M. tuber-
had hepatomegaly, and 53% had splenomegaly [39]. culosis complex (ie, M. tuberculosis, M. bovis, and
Other common findings are pneumonia, fever, and M. africanum) from a clinical specimen; (2) demon-
lymphadenopathy [40]. stration of M. tuberculosis from a clinical specimen
using nucleic acid amplification (NAA) test; or
(3) demonstration of AFB from a clinical specimen
when culture is unavailable. In the absence of
Diagnosis of pediatric latent tuberculosis infection laboratory confirmation, the clinical case definition
and tuberculosis is fulfilled by all of the following criteria [42]:
Children in close contact with known 1. Was your child born outside the United
or suspected contagious cases of States? If yes, this question would be
tuberculosis disease followed by where was your child born?
Children suspected to have tuberculosis If the child was born in Africa, Asia, Latin
disease because of America, or Eastern Europe, a tuberculin
Findings on chest radiograph skin test should be placed.
2. Has your child traveled outside the United
consistent with active or
States? If yes, this question would be
previously active tuberculosis followed by where did the child travel,
Clinical evidence of
with whom did the child stay, and how
tuberculosis disease long did the child travel. If the child
Children receiving immunosuppressive stayed with friends or family members in
therapy or with immunosuppressive Africa, Asia, Latin America, or Eastern Eu-
conditions, including HIV infection rope, for 1 week or longer, cumulatively,
a tuberculin skin test should be placed.
Induration 10 mm 3. Has your child been exposed to anyone
with TB disease? If yes, this question
should be followed by questions to deter-
Children at increased risk of
mine if the person had TB disease or LTBI,
disseminated disease:
when the exposure occurred, and the
Those younger than 4 years of age
nature of the contact. If the child’s
Those with other medical conditions, exposure to someone with suspected or
including Hodgkin’s disease, lym- known TB disease is confirmed, a tuber-
phoma, diabetes mellitus, chronic renal culin skin test should be placed, and the
failure, or malnutrition local health department should be notified
per local reporting guidelines.
Children with increased exposure to tu- 4. Does your child have a close contact with
berculosis disease: a person who has a positive TB skin test?
Those born, or whose parents were If yes, see question 3 follow-up
questions.
born, in high-prevalence regions of
the world
Those frequently exposed to adults Risk Assessment Questionnaires can in-
clude the following based on local epidemiology
who are HIV-infected, homeless,
and priorities:
users of illicit drugs, residents of
nursing homes, incarcerated or
institutionalized, or migrant farm 1. Does your child spend time with anyone
workers who has been in jail (or prison), a shelter,
Those who travel to high-prevalence who uses illegal drugs, or who has HIV?
2. Has your child drunk raw milk or eaten
regions of the world unpasteurized cheese?
3. Does your child have a household member
Induration 15 mm who was born outside the United States?
4. Does your child have a household member
Children 4 years of age or older without who has traveled outside the United
any risk factors States?
Table 1
Factors associated with false-negative or false-positive tuberculin skin test reactions
False-negative reactions False-positive reactions
Infections
Viral illnesses (HIV, measles, varicella) Exposure to nontuberculous mycobacteria
Bacterial (typhoid fever, brucellosis, typhus, leprosy) (eg, M. marinum, M. kansasii)
Early TB infection (< 12 wk)
TB disease (meningitis, miliary, pleural)
Fungal (Blastomycosis)
Live-virus vaccines
Measles Bacille Calmette-Guerin vaccine
Polio
Smallpox
Concomitant medical conditions
Metabolic abnormalities (chronic renal failure) Transfusion with whole blood from donors with
Malignancies (Hodgkin’s disease, lymphoma, leukemia) known positive tuberculin skin test
Sarcoidosis
Poor nutrition
Drugs and technical factors
Corticosteroids, chemotherapy Inexperienced or biased reader
Newborns and children <2 y
Material: poor quality, inadequate dose (1 tuberculin unit),
improper storage (exposure to heat/light), expired
Administration: not injected intradermally, too long in syringe
Reading: inexperienced or biased reader, recording error,
read too early/late
Interpretative
Decreasing mm of induration Increasing mm of induration
From Pediatric Tuberculosis Collaborative Group. Targeted tuberculin skin testing and treatment of latent tuberculosis infection
in children and adolescents. Pediatrics 2004;114(4):1186; with permission.
306 feja & saiman
include the lack of standardization of antigens and Standardization of gastric aspiration techniques is
poor reproducibility of the delayed-type hypersensi- critical and has been shown to increase the proportion
tivity reaction [46], lack of association of anergy with of positive cultures [50]. Hospitalization may not be
a high risk of progression to TB disease, and no necessary to obtain gastric aspirate specimens from
demonstrable benefit of treatment with empiric iso- children with suspected TB (providing hospitalization
niazid in anergic HIV-infected individuals to prevent is otherwise not clinically indicated). There was no
TB disease [47]. significant difference in the yield of outpatient
(37%) and inpatient (48%) gastric aspirates among
Effect of bacille Calmette-Guerin vaccination on 80 children [52]. In the United States, the relative
tuberculin skin test results yield of gastric aspirates is not reportable, but recent
epidemiologic studies indicate that only 2.4% and
Bacille Calmette-Guerin (BCG) vaccination is 8.5% of all pediatric cases had smear-positive and
used in most countries with a high incidence of TB culture-positive gastric aspirates, respectively [7].
to prevent the severe forms of TB in young children. To circumvent the low yield of positive cultures
According to the WHO, 161 member states have from expectorated sputum and gastric aspirates, Zar
BCG vaccination on their immunization schedule, et al [53] successfully performed sputum induction in
and global BCG vaccination coverage in 2002 among 142 of 149 children (median age, 9 months; 70%
infants less than 1 year of age was 81% [48]. Thus, it HIV-infected) in South Africa. After a 2- to 3-hour
is important to understand the potential effects of fast, pretreatment with inhaled salbutamol was
BCG vaccine on TST results in previously immu- followed by nebulized 5% sterile saline and chest
nized children who immigrate to developed countries physiotherapy. Sputum was then obtained by suction-
and undergo targeted tuberculin skin testing. Factors ing (90% of children) or expectoration (10% of
such as age at BCG immunization (less effect if children). Induced sputum cultures grew M. tuber-
vaccinated at birth), time since immunization (less culosis more frequently (11%; 15 of 142 children)
effect if immunization occurred years before TST), than gastric aspirates (6%; 9 of 142 children).
exposure to nontuberculous mycobacteria (increased Bronchoscopy does not aid significantly in the
exposure may cause false positive TSTs), prevalence diagnosis of pediatric TB; only 4% to 12% of
of LTBI (higher prevalence increases the positive bronchoalveolar lavage (BAL) cultures were positive
predictive value of reactive TSTs), and type of among children with suspected or presumed pulmo-
vaccine used (larger number of viable bacilli causes nary TB [54 – 56]. Furthermore, in the two studies
larger effect) may influence TST results [19]. examining a total of 70 children, the overall yield of
Children immunized with BCG at birth who were BAL specimens was lower than that of gastric
studied 12 years later and found to have a TST aspirates (10% – 12% versus 32% – 50%), and 83%
induration of 10 mm or more had a five- to 48-fold to 100% of children with positive BAL specimens
increased risk of developing TB disease within also had positive gastric aspirates [55,56]. Bronchos-
4 years [49]. Thus, positive TST reactions in BCG- copy in children may be useful to accurately diagnose
vaccinated children from countries with high case endobronchial disease, however [57]. Cultures from
rates of TB are likely be caused by LTBI, not im- children with extrapulmonary TB grow M. tuber-
munization. It is recommended that a history of BCG culosis from the affected site in about 50% of cases
immunization be ignored when planting, reading, and [20]. A summary of diagnostic procedures and
interpreting a TST [20]. findings is found in Table 2. More recent diagnos-
tic techniques include polymerase chain reaction,
Smear for acid-fast bacilli and mycobacterial culture NAA, serology tests, and immunoassays based on
detection of cellular responses to specific M. tuber-
Smear and culture techniques in children are culosis antigens.
similar to those in adults, but the types of specimens
and their yield differs. Younger children with
pulmonary TB rarely produce sputum and have a Treatment of latent tuberculosis infection and
relatively low organism burden. Thus, first morning tuberculosis disease
gastric aspirates have traditionally been the best
clinical specimens to obtain in young children with Children with LTBI and TB are treated with the
suspected pulmonary TB. These aspirates may be drugs used for adults, although regimens, side-effect
positive for M. tuberculosis in 30% to 50% of all profiles, and availability of antituberculous drugs in
cases [21,50] and in as many as 75% of infants [51]. pediatric doses may differ. Further, recommendations
tuberculosis in children 307
Table 2
Diagnostic methods for obtaining specimen to assess tubercular disease in children
Site Evaluation Findings
Pulmonary Induced sputum May be AFB smear – negative but culture positive
Gastric aspirates Role of rapid amplification tests is limited, not FDA
Bronchoscopy approved for AFB-smear negative specimens
Central nervous system Lumbar puncture Lymphocytosis, low glucose, high protein
Cerebrospinal fluid Smear usually negative
Culture may be positive
Rapid amplification tests are not FDA approved
Pleural, pericardial, peritoneal Fluid samplea Lymphocytosis, low glucose, high protein
Smear usually negative
Culture may be positive
Adenosine deaminase may be elevated
Adenitis Biopsy Caseous granulomas with AFB on pathology
Culture may be positive
Osteoarticular Joint aspirationa Smear and culture of sediment or biopsy
Bone or synovial biopsy Culture may be positive
Renal Urine sample Smears usually negative
Multiple samples
Abbreviations: AFB, acid-fast bacilli; FDA, Food and Drug Administration.
a
Greater volume and centrifugation increase yield.
for pediatric regimens and doses from national and gression to disease and disseminated disease. All
international organizations (eg, the British Thoracic infants and children with LTBI and no history
Society, American Thoracic Society, WHO) may of previous treatment for TB should be treated
differ. Nevertheless, the backbone of all recom- with isoniazid for 9 months. When the source
mended regimens for LTBI is isoniazid unless the case has isoniazid-resistant, rifampin-susceptible
child has been exposed to an isoniazid-resistant M. tuberculosis, rifampin is recommended for at least
source case. Although the recommended regimens 6 months [17].
for TB vary somewhat, they generally include two to In young children, especially those less than
four of the following first-line drugs: isoniazid, 4 years of age, with close exposure to an infectious
rifampin, pyrazinamide, and ethambutol. Generally, source case, window prophylaxis is recommended
antituberculosis drugs are well tolerated by children in efforts to initiate preventive therapy early and
and are better tolerated than by adults. potentially prevent progression to active disease [17].
The success of pediatric treatment regimens If the contact investigation demonstrates that a young
depends on using correct doses (Table 3) and com- child’s initial TST shows no induration, and chest
pleting a full course of therapy. Adherence to radiography is normal, treatment with isoniazid (or
treatment is facilitated by the use of child-friendly rifampin, if the source case is known to have an
(eg, good tasting, small volumes) drug formulations isoniazid-resistant, rifampin-susceptible organism) is
and DOT. Treatment should be started promptly recommended to prevent progression to active dis-
in young children when infection or disease is diag- ease. A follow-up TST is performed 12 weeks after
nosed, because they are at increased risk for pro- the last exposure to the infectious source case; if the
induration is less than 5 mm, preventive therapy is
discontinued. If the TST induration is 5 mm or greater
(see Box 1), a full course for LTBI is completed.
Table 3
Choice of treatment regimens in children with TB
Pediatric doses for first-line antituberculous drugs
disease must consider the identification of a source
Dose mg/kg (maximum dose) case and mycobacterial susceptibilities, extent of
First-line drugs Daily Intermittent disease, empiric versus targeted therapy, and local
Isoniazid 10 – 15 (300 mg) 20 – 30 (900 mg) resistance patterns. The lower bacillary load typical
Rifampin 10 – 20 (600 mg) 10 – 20 (600 mg) of primary pulmonary tuberculosis is associated with
Pyrazinamide 15 – 40 (2 g) 50 (2 g) less risk of drug resistance [58]. Therefore, a 6-month
Ethambutol 15 – 25 (1 g) 50 (2.5 g) regimen of isoniazid, rifampin, and pyrazinamide for
308 feja & saiman
the first 2 months (daily for at least 2 weeks) followed ing, malaise, fatigue, abdominal discomfort, or fever)
by 4 months of isoniazid and rifampin (daily or 2 – of liver toxicity develop during treatment [19]. Fur-
3 times/week) is recommended for children with thermore, parents and children should be educated
drug-susceptible intrathoracic tuberculosis [17]. Chil- to be alert for signs and symptoms of hepatitis while
dren with hilar adenopathy and no other evidence of the patient is receiving isoniazid.
disease may be treated with 6 months of isoniazid and The use of isoniazid may lead to peripheral neu-
rifampin if drug resistance is not a consideration [17]. ropathy caused by pyridoxine (vitamin B6) deficiency
If reactivation-type cavitary disease is diagnosed, an from increased excretion. This toxicity manifests as a
adult-type regimen is used that includes ethambutol tingling sensation of the fingers and toes. This phe-
in addition to isoniazid, rifampin, and pyrazinamide nomenon is rare in children, although factors such as
during the initial 2 months, until drug susceptibilities diabetes, uremia, a diet deficient in milk and meat,
are determined [58]. Ethambutol also should be nutritional deficiencies, symptomatic HIV infection,
added during the first 2 months if isoniazid resistance alcohol consumption, pregnancy, and breastfeeding
is documented or suspected, particularly in regions may increase the risk of peripheral neuropathy. Chil-
with high rates of isoniazid resistance [17]. dren and adolescents with these risk factors should
Most extrapulmonary disease in children can be receive pyridoxine [17].
treated like pulmonary TB with 6 months of therapy. Additional toxicities of isoniazid include a hy-
TB meningitis, miliary, and severe disseminated persensitivity reaction such as maculopapular or
disease require four drugs (isoniazid, rifampin, morbilliform skin rashes, which may necessitate
pyrazinamide, and ethambutol) used daily for discontinuation of isoniazid. Isoniazid inhibits the
2 months followed by daily or twice weekly isoniazid P450 cytochrome isozymes and may therefore in-
and rifampin for a total of 9 to 12 months [17,58]. crease serum concentrations of concomitant drugs
Some experts continue all four drugs to complete a such as anticonvulsants [58].
full course if susceptibilities are unknown. Rifampin may cause adverse effects such as cuta-
neous reactions, gastrointestinal reactions, flulike
Toxicities of first-line tuberculosis drugs syndrome with intermittent therapy, hepatotoxicity
(particularly when paired with isoniazid), and, rarely,
Isoniazid is generally very well tolerated by a severe immunologic reaction. In a study of
children, but, rarely, hepatic, neurologic, and gastro- 157 adolescents receiving 6 months of daily rifampin
intestinal toxicities may occur. Hepatotoxicity may for treatment of LTBI caused by isoniazid-resistant
manifest as asymptomatic transient elevation of M. tuberculosis, 26% reported anorexia, nausea,
transaminases (most common) to fulminant hepatitis fatigue, or rash, but only 1.3% (2/157) required
and liver failure (very rare). Risk factors for hepatitis, permanent drug discontinuation [64]. Patients should
such as preexisting liver disease, older age, mal- be made aware of the probable orange discoloration
nutrition, excessive alcohol consumption, and the use of body fluids (eg, urine, sweat, tears) as well as the
of concomitant hepatotoxic drugs, are less frequently need for additional birth control methods in adoles-
noted in children [19,58]. Palusci et al [59] performed cents taking oral contraceptives. Rifampin is a potent
a pooled analysis and calculated that 8% of cytochrome P450 enzyme inducer and can decrease
965 children developed transient abnormalities in the serum concentration of many drugs, including
liver function tests, but only 0.4% discontinued iso- oral contraceptives, HIV protease inhibitors, non-
niazid. Fewer than 1% of children receiving isoniazid nucleoside reverse transcriptase inhibitors, cortico-
may develop adverse effects that include rash, steroids, and phenytoin [58].
nausea, vomiting, and diarrhea [60,61], but severe Pyrazinamide generally is well tolerated in chil-
hepatotoxicity has been reported [59,62,63]. Thus, dren and rarely causes hepatotoxicity in doses of
before initiating isoniazid therapy, a history should 30 mg/kg/day or less, although toxicity can be
be obtained for risk factors, side effects with previ- observed during treatment with multidrug regimens.
ous isoniazid, and signs or symptoms of liver dis- Toxicities from pyrazinamide (20 – 25 mg/kg/day)
ease [19]. were evaluated among 114 children (aged 6 months
Baseline liver function tests and further testing to 15 years) given 2 months of isoniazid, rifampin,
during therapy are recommended only in children and pyrazinamide followed by 4 months of isoniazid
with risk factors for hepatitis. Liver function tests and rifampin for pulmonary TB. Adverse effects
should be obtained in children without risk factors if during the pyrazinamide treatment phase included
signs (scleral icterus, jaundice, brown urine, clay- fever (2.6%), gastrointestinal disturbances (4.4%),
colored stools) or symptoms (anorexia, nausea, vomit- and transient asymptomatic elevation of uric acid
tuberculosis in children 309
(9.8%). No patients had treatment interrupted [65]. In and cognitive dysfunction [69]. A prospective, ran-
another study, liver enzymes were monitored in domized study of 29 children with greater than
43 children receiving isoniazid, rifampin, and pyra- 50% bronchial obstruction caused by TB found a
zinamide [66]. Only 5 (12%) of the 43 developed more rapid improvement in radiographic and bron-
transaminase elevations more than twice the upper choscopy findings in children treated with concomi-
limit of normal for age. One child developed jaundice tant prednisolone than in those only given anti-TB
requiring temporary discontinuation of medications. medications [71]. Although no randomized trials in
Most laboratory abnormalities occurred in the first children are available, corticosteroids may be used to
2 weeks of treatment and resolved by 14 weeks of increase reabsorption of pleural and pericardial fluids
treatment [66]. Despite the rarity of symptomatic and to alleviate alveolocapillary block in severe
hepatitis, pyrazinamide is not recommended for miliary disease [17].
pediatric patients with underlying liver disease or Optimal treatment regimens for TB in HIV-
isoniazid-induced hepatotoxicity [17]. infected children have not yet been established.
Adverse reactions associated with ethambutol Prospective cohort studies have noted higher treat-
include retrobulbar neuritis, gastrointestinal distur- ment failures among HIV-infected children receiving
bances, and hypersensitivity. Retrobulbar neuritis is standard short-course therapy when compared with
usually a reversible, dose-dependent, and renal non – HIV-infected children (29%, 6/21, versus 3%,
function – dependent phenomenon that manifests as 5/156; P = 0.0004) [72]. Mortality rates also were
decreased visual acuity or decreased red-green color found to be higher in HIV-infected children with
discrimination. Monitoring for these signs and pulmonary TB (41%, 22/58 HIV-positive, versus 7%,
symptoms is recommended monthly in older children 26/459 HIV-negative; P < 0.001) or TB meningitis
and adults [17,58]. Past guidelines have advised (30%, 3/10 HIV-positive, versus 0/30 HIV-negative,
against using or have advised caution when using P = 0.01) [73,74]. The American Academy of Pedia-
ethambutol in younger children who cannot verbalize trics recommends at least 9 months of therapy with
symptoms of optic neuritis, although two reviews did isoniazid, rifampin, pyrazinamide, plus ethambutol
not detect visual toxicity in young children [67,68]. (or an aminoglycoside) for at least 2 months pending
Current recommendations of the American Academy drug-susceptibility results [17]. An expert in pediatric
of Pediatrics suggest consideration of the risks and TB and HIV should be consulted. Rifabutin can be
benefits involved with the use of ethambutol in substituted for rifampin to avoid drug interactions in
younger children [17]. children receiving protease inhibitors as part of the
The use of second-line drugs such as cycloserine, HIV treatment.
ethionamide, streptomycin, amikacin, kanamycin, It is paramount to ensure successful completion of
capreomycin, p-aminosalicylic acid, and fluoroquino- therapy. Pediatric formulations are not available for
lones should always be reserved for the treat- many antituberculosis drugs, and often pills or cap-
ment of multidrug-resistant TB. Consultation with sules must be crushed and mixed with food, fre-
a pediatric TB specialist is required, because the quently resulting in bitter-tasting mixtures. Despite
associated adverse effects may be greater than these barriers, between 1993 and 1999, 90% of
with first-line drugs, because clinical experience children with TB in the United States completed
and pharmacokinetic data are more limited in therapy [7]. A combination of DOT, parent/patient
children than in adults, and because inappropriate education, pediatric formulations, enablers (strategies
management may lead to life-threatening conse- to overcome logistic barriers such as funds for
quences [58]. transportation or extended clinic hours), and incen-
Corticosteroids as adjuvant therapy are indicated tives (strategies to enhance motivation such as
when the heightened inflammatory response may snacks, food coupons, or movie tickets) can increase
cause further tissue damage. High-dose steroids completion rates [19]. DOT is recommended for use
should always be used in conjunction with effective in all children with TB disease and multidrug-
antituberculosis therapy and should be tapered slowly resistant LTBI and in children with drug-susceptible
over weeks to avoid a rebound reaction. Generally, LTBI when resources permit [17,19,58].
1 to 2 mg/kg/day of prednisone (maximum, 60 mg/day)
or its equivalent tapered over 6 to 8 weeks is used
[17]. Prospective, randomized trials have shown that Public health aspects of pediatric tuberculosis
concomitant corticosteroids decreased mortality rates
in children with TB meningitis [69,70] and reduced As with adults, the public health sector plays an
neurologic complications, neurologic sequelae [70], important role in control of TB in children through a
310 feja & saiman
hierarchy of activities that range from identification are the future source of disease in adults. Therefore,
and treatment of patients with TB disease to close collaboration between public health services
conducting public health investigations of infectious and pediatric care providers is needed to ensure suc-
source cases and identifying and treating patients cessful completion of treatment. Future goals should
with LTBI [19]. Following recognition of an infec- continue to focus on developing more accurate
tious adult or adolescent case, a contact investigation epidemiologic data for childhood TB, supporting
is performed to identify exposed persons, including TB control programs to decrease morbidity and mor-
children and adolescents, and evaluate them for LTBI tality in children with TB and LTBI, and developing
or TB disease. Source-case investigations are con- additional pediatric formulations of antituberculo-
ducted to identify an infectious adult from whom a sis drugs.
child with active TB may have contracted M. tuber-
culosis. During this process, the close contacts of the
child (including other children) are evaluated for References
infection and disease. Last, associate investigations
evaluate household contacts of children with LTBI to [1] Dolin PJ, Raviglione MC, Kochi A. Global tuber-
find a potential source of infection and to identify culosis incidence and mortality during 1990 – 2000.
others with LTBI. Bull World Health Organ 1994;72:213 – 20.
As previously mentioned, younger children with [2] World Health Organization. Global tuberculosis control:
surveillance, planning, financing. WHO report 2004.
pulmonary TB are rarely infectious. Therefore, air-
Geneva (Switzerland): World Health Organization.
borne isolation is not generally required in a health [3] Nelson LJ, Wells CD. Global epidemiology of child-
care setting. Unique situations have been described, hood tuberculosis. Int J Tuberc Lung Dis 2004;8(5):
however, in which transmission may have occurred 636 – 47.
from a young child with unrecognized congenital TB [4] Walls T, Shingadia D. Global epidemiology of pae-
or cavitary TB. Such transmission is associated par- diatric tuberculosis. J Infect 2004;48(1):13 – 22.
ticularly with suctioning or respiratory care inter- [5] Ussery XT, Valway SE, McKenna M, et al. Epidemi-
ventions [75,76]. More commonly, accompanying ology of tuberculosis among children in the United
parents, relatives, or health care workers with un- States: 1985 to 1994. Pediatr Infect Dis J 1996;15(8):
diagnosed pulmonary TB are the source of nosoco- 697 – 704.
[6] Starke JR, Jacobs RF, Jereb J. Resurgence of tuber-
mial transmission in pediatric health care settings
culosis in children. J Pediatr 1992;120(6):839 – 55.
[77 – 81]. Infection control measures, therefore, [7] Nelson LJ, Schneider E, Wells CD, et al. Epidemiology
should focus on adults accompanying the child. of childhood tuberculosis in the United States, 1993 –
Promptly evaluating of the household contacts of a 2001: the need for continued vigilance. Pediatrics
child with suspected TB by chest radiograph before 2004;114(2):333 – 41.
giving visitation rights in the hospital may avoid [8] Lobato MN, Hopewell PC. Mycobacterium tuber-
health care – associated transmission of TB. culosis infection after travel to or contact with visitors
from countries with a high prevalence of tuberculosis.
Am J Respir Crit Care Med 1998;158(6):1871 – 5.
Summary [9] Saiman L, San Gabriel P, Schulte J, et al. Risk
factors for latent tuberculosis infection among chil-
dren in New York City. Pediatrics 2001;107(5):
The epidemiology of pediatric TB continues to be 999 – 1003.
shaped by risk factors such as age, race, immigration, [10] Besser RE, Pakiz B, Schulte JM, et al. Risk factors
poverty, overcrowding, and HIV/AIDS. Once in- for positive Mantoux tuberculin skin tests in children
fected, young children have an increased risk of TB in San Diego, California: evidence for boosting and
disease and progression to extrapulmonary disease possible foodborne transmission. Pediatrics 2001;
because of immunologic host factors. The patho- 108(2):305 – 10.
genesis of disease differs from that in adults, because [11] Froehlich H, Ackerson LM, Morozumi PA. Targeted
primary disease and its complications are more testing of children for tuberculosis: validation of a risk
common in children. This prevalence of primary assessment questionnaire. Pediatrics 2001;107(4):e54.
[12] Centers for Disease Control and Prevention. Core
disease in turn leads to differences in clinical and
curriculum on tuberculosis. 4th edition. Atlanta (GA)7
radiographic manifestations in pediatric TB. Diffi- Centers for Disease Control and Prevention; 2000.
culties in diagnosing children stem from the low yield [13] Lincoln EM, Sewell EM. Tuberculosis in children.
of mycobacteriology cultures and the subsequent New York7 McGraw-Hill; 1963.
reliance on clinical case definitions. Inadequately [14] American Thoracic Society/Centers for Disease Con-
treated TB infection and TB disease in children today trol and Prevention. Diagnostic standards and classi-
tuberculosis in children 311
fication of tuberculosis in adults and children. Am J [33] Ormerod P. Tuberculosis in pregnancy and the puer-
Respir Crit Care Med 2000;161(4 Pt 1):1376 – 95. perium. Thorax 2001;56:494 – 9.
[15] Lewinsohn DA, Gennaro ML, Scholvinck L, et al. [34] Figueroa-Damian R, Arrendondo-Garcia JL. Preg-
Tuberculosis immunology in children: diagnostic and nancy and tuberculosis: influence of treatment on peri-
therapeutic challenges and opportunities. Int J Tuberc natal outcome. Am J Perinatol 1998;15:303 – 6.
Lung Dis 2004;8(5):658 – 74. [35] Jana N, Vasishta K, Jindal SK, et al. Perinatal outcome
[16] American Thoracic Society. Targeted tuberculin testing in pregnancies complicated by pulmonary tuberculosis.
and treatment of latent tuberculosis infection. Am J Int J Gynecol Obstet 1994;44:119 – 24.
Respir Crit Care Med 2000;161(4 Pt 2):S221 – 47. [36] Jana N, Vasishta K, Shaha SC, et al. Obstetrical
[17] American Academy of Pediatrics. Tuberculosis. In: outcomes among women with extrapulmonary tuber-
Pickering LK, editor. Red book: 2003 report of the culosis. N Engl J Med 1999;341:645 – 9.
committee on infectious diseases. 25th edition. Elk [37] Cantwell MF, Shehab ZM, Costello AM, et al. Congeni-
Grove Village (IL)7 Pickering LK; 2003. p. 642 – 60. tal tuberculosis. New Engl J Med 1994;330:1051 – 4.
[18] Smuts NA, Beyers N, Gie RP, et al. Value of the lateral [38] Chen A, Shih SL. Congenital tuberculosis in two
chest radiograph in tuberculosis in children. Pediatr infants. Am J Radiol 2004;182:253 – 6.
Radiol 1994;24(7):478 – 80. [39] Schaaf HS, Gie RP, Beyers N, et al. Tuberculosis in
[19] Pediatric Tuberculosis Collaborative Group. Targeted infants less than 3 months of age. Arch Dis Child
tuberculin skin testing and treatment of latent tuber- 1993;69(3):371 – 4.
culosis infection in children and adolescents. Pediatrics [40] Adhikari M, Pillay T, Pillay DG. Tuberculosis in the
2004;114(4):1175 – 201. newborn: an emerging disease. Pediatr Infect Dis J
[20] Starke JR, Taylor-Watts KT. Tuberculosis in the pe- 1997;16(12):1108 – 12.
diatric population of Houston, Texas. Pediatrics 1989; [41] Starke JR. Diagnosis of tuberculosis in children.
84:28 – 35. Pediatr Infect Dis J 2000;19(11):1095 – 6.
[21] Khan EA, Starke JR. Diagnosis of tuberculosis in [42] The Centers for Disease Control and Prevention.
children: increased need for better methods. Emerg Reported tuberculosis in the United States, 2002.
Infect Dis 1995;1(4):115 – 23. Appendix B. Atlanta (GA)7 United States Department
[22] Sanchez-Albisua I, Baquero-Artigao F, Del Castillo F, of Health and Human Services, CDC; 2003.
et al. Twenty years of pulmonary tuberculosis in [43] San Gabriel P, Saiman L, Kaye K, et al. Completeness
children: what has changed? Pediatr Infect Dis J of pediatric TB reporting in New York City. Public
2002;21:49 – 53. Health Rep 2003;118(2):144 – 53.
[23] Maltezou HC, Spyridis P, Kafetzis DA. Extra-pulmo- [44] Steiner P, Rao M, Victoria MS, et al. Persistently
nary tuberculosis in children. Arch Dis Child 2000; negative tuberculin reactions: their presence among
83:342 – 6. children with culture positive for Mycobacterium
[24] Merino JM, Carpintero I, Alvarez T, et al. Tuberculous tuberculosis (tuberculin-negative tuberculosis). Am J
pleural effusion in children. Chest 1999;115:26 – 30. Dis Child 1980;134(8):747 – 50.
[25] Wallgren A. The time table of tuberculosis. Tubercle [45] The Centers for Disease Control and Prevention.
1948;29:245 – 56. Anergy skin testing and tuberculosis preventive
[26] van den Bos F, Terken M, Ypma L, et al. Tuberculous therapy for HIV-infected persons: revised recommen-
meningitis and miliary tuberculosis in young children. dations [published correction appears in MMWR Morb
Trop Med Int Health 2004;9:309 – 13. Mortal Wkly Rep. 1997;46:880]. MMWR Recomm
[27] Waecker NJ, Connor JD. Central nervous system Rep 1997;46(RR-15):1 – 10.
tuberculosis in children: a review of 30 cases. Pediatr [46] Chin DP, Osmond D, Page-Shafer K, et al. Reliability
Infect Dis J 1990;9:539 – 43. of anergy skin testing in persons with HIV infection.
[28] Delage G, Dusseault M. Tuberculous meningitis in The Pulmonary Complications of HIV Infection Study
children: a retrospective study of 79 patients, with Group. Am J Respir Crit Care Med 1996;153(6 Pt 1):
analysis of prognostic factors. Can Med Assoc J 1979; 1982 – 4.
120:305 – 9. [47] Gordin FM, Matts JP, Miller C, et al. A controlled trial
[29] Hussey G, Chisholm T, Kibel M. Miliary tuberculosis of isoniazid in persons with anergy and human
in children: a review of 94 cases. Pediatr Infect Dis J immunodeficiency virus infection who are at high risk
1991;10:832 – 6. for tuberculosis. Terry Beirn Community Programs for
[30] Munoz FM, Starke JR. Tuberculosis in children. Clinical Research on AIDS. N Engl J Med 1997;
In: Reichman LB, Harshfield ES, editors. Tuberculosis: 337(5):315 – 20.
a comprehensive international approach. 2nd edition. [48] World Health Organization. WHO vaccine-preventable
New York7 Marcel Dekker; 2000. p. 553 – 98. diseases: monitoring system 2003 global summary. Ge-
[31] Vallejo JG, Ong LT, Starke JR. Tuberculous osteo- neva (Switzerland)7 World Health Organization; 2003.
myelitis of the long bones in children. Pediatr Infect [49] Chee CB, Soh CH, Boudville IC, et al. Interpretation
Dis J 1995;14:542 – 6. of the tuberculin skin test in Mycobacterium bovis
[32] Teo HE, Peh WC. Skeletal tuberculosis in children. BCG- vaccinated Singaporean schoolchildren. Am J
Pediatr Radiol 2004;34(11):853 – 60. Respir Crit Care Med 2001;164(6):958 – 61.
312 feja & saiman
[50] Pomputius WF, Rost J, Dennehy PH, et al. Stan- [66] Corrigan D, Paton J. Hepatic enzyme abnormalities
dardization of gastric aspirate technique improves in children on triple therapy for tuberculosis. Pediatr
diagnosis of tuberculosis in children. Pediatr Infect Pulmonol 1999;27(1):37 – 42.
Dis J 1997;16(2):222 – 6. [67] Graham SM, Daley HM, Banerjee A, et al. Ethambu-
[51] Vallejo JG, Ong LT, Starke JR. Clinical features, tol in tuberculosis: time to reconsider? Arch Dis Child
diagnosis, and treatment of tuberculosis in infants. 1998;79(3):274 – 8.
Pediatrics 1994;94:1 – 7. [68] Trebucq A. Should ethambutol be recommended for
[52] Lobato MN, Loeffler AM, Furst K, et al. Detection of routine treatment of tuberculosis in children? A re-
Mycobacterium tuberculosis in gastric aspirates col- view of the literature. Int J Tuberc Lung Dis 1997;
lected from children: hospitalization is not necessary. 1(1):12 – 5.
Pediatrics 1998;102(4):e40. [69] Schoeman JF, Van Zyl LE, Laubscher JA, et al. Effect
[53] Zar HJ, Tannenbaum E, Apolles P, et al. Sputum of corticosteroids on intracranial pressure, computed
induction for the diagnosis of pulmonary tuberculosis tomographic findings, and clinical outcome in young
in infants and young children in an urban setting in children with tuberculous meningitis. Pediatrics 1997;
South Africa. Arch Dis Child 2000;82:305 – 8. 99(2):226 – 31.
[54] Bibi H, Mosheyev D, Shoseyov D, et al. Should [70] Girgis NI, Farid Z, Kilpatrick ME, et al. Dexametha-
bronchoscopy be performed in the evaluation of sus- sone adjunctive treatment for tuberculous meningitis.
pected pediatric pulmonary tuberculosis? Chest 2002; Pediatr Infect Dis J 1991;10(3):179 – 83.
122:1604 – 8. [71] Toppet M, Malfroot A, Derde MP, et al. Cortico-
[55] Abadco DL, Steiner P. Gastric lavage is better than steroids in primary tuberculosis with bronchial obstruc-
bronchoalveolar lavage for isolation of Mycobacterium tion. Arch Dis Child 1990;65(11):1222 – 6.
tuberculosis in childhood pulmonary tuberculosis. [72] Espinal MA, Reingold AL, Perez G, et al. Human
Pediatr Infect Dis J 1992;11(9):735 – 8. immunodeficiency virus infection in children with
[56] Somu N, Swaminathan S, Paramasivan CN, et al. tuberculosis in Santo Domingo, Dominican Republic:
Value of bronchoalveolar lavage and gastric lavage in prevalence, clinical findings, and response to anti-
the diagnosis of pulmonary tuberculosis in children. tuberculosis treatment. J Acquir Immune Defic Syndr
Tuber Lung Dis 1995;76(4):295 – 9. Hum Retrovirol 1996;13(2):155 – 9.
[57] Chan S, Abadco DL, Steiner P. Role of flexible [73] Palme IB, Gudetta B, Bruchfeld J, et al. Impact of
fiberoptic bronchoscopy in the diagnosis of childhood human immunodeficiency virus 1 infection on clinical
endobronchial tuberculosis. Pediatr Infect Dis J 1994; presentation, treatment outcome and survival in a
13(6):506 – 9. cohort of Ethiopian children with tuberculosis. Pediatr
[58] Centers for Disease Control and Prevention. Treatment Infect Dis J 2002;21(11):1053 – 61.
of tuberculosis, American Thoracic Socity, CDC, and [74] Topley JM, Bamber S, Coovadia HM, et al. Tuber-
Infectious Diseases Society of America. MMWR Morb culous meningitis and co-infection with HIV. Ann Trop
Mortal Wkly 2003;52(No. RR-110):1 – 77. Paediatr 1998;18(4):261 – 6.
[59] Palusci VJ, O’Hare D, Lawrence RM. Hepatotoxicity [75] Cardona M, Bek MD, Mills K, et al. Transmission of
and transaminase measurement during isoniazid che- tuberculosis from a seven-year-old child in a Sydney
moprophylaxis in children. Pediatr Infect Dis J 1995; school. J Paediatr Child Health 1999;35(4):375 – 8.
14(2):144 – 8. [76] Curtis AB, Ridzon R, Vogel R, et al. Extensive
[60] Mount F, Ferrebee S. Preventive effects of isoniazid transmission of Mycobacterium tuberculosis from a
in the treatment of primary tuberculosis in children. child. N Engl J Med 1999;341(20):1491 – 5.
N Engl J Med 1961;265:713 – 21. [77] Lee EH, Graham PL, O’Keefe M, et al. Nosocomial
[61] Hsu KH. Isoniazid in the prevention and treatment of transmission of Mycobacteria tuberculosis in a chil-
tuberculosis. A 20-year study of the effectiveness in dren’s hospital. Int J Tuberc Lung Dis, in press.
children. JAMA 1974;229(5):528 – 33. [78] Munoz FM, Ong LT, Seavy D, et al. Tuberculosis
[62] Vanderhoof JA, Ament ME. Fatal hepatic necrosis due among adult visitors of children with suspected tuber-
to isoniazid chemoprophylaxis in a 15-year-old girl. culosis and employees at a children’s hospital. Infect
J Pediatr 1976;88(5):867 – 8. Control Hosp Epidemiol 2002;23(10):568 – 72.
[63] Moulding TS, Redeker AG, Kanel GC. Twenty [79] Weinstein JW, Barrett CR, Baltimore RS, et al.
isoniazid-associated deaths in one state. Am Rev Nosocomial transmission of tuberculosis from a hos-
Respir Dis 1989;140(3):700 – 5. pital visitor on a pediatrics ward. Pediatr Infect Dis J
[64] Villarino ME, Ridzon R, Weismuller PC, et al. 1995;14(3):232 – 4.
Rifampin preventive therapy for tuberculosis infection: [80] George RH, Gully PR, Gill ON, et al. An outbreak of
experience with 157 adolescents. Am J Respir Crit tuberculosis in a children’s hospital. J Hosp Infect
Care Med 1997;155(5):1735 – 8. 1986;8(2):129 – 42.
[65] Sanchez-Albisua I, Vidal ML, Joya-Verde G, et al. [81] Centers for Disease Control and Prevention. Tuber-
Tolerance of pyrazinamide in short course chemo- culosis transmission in a renal dialysis center – Nevada,
therapy for pulmonary tuberculosis in children. Pediatr 2003. MMWR Morb Mortal Wkly Rep 2004;53(37):
Infect Dis J 1997;16(8):760 – 3. 873 – 5.
Clin Chest Med 26 (2005) ix – x
Preface
Tuberculosis
Neil W. Schluger, MD
Guest Editor
In 1952, Selman Waksman received the Nobel culosis rates in high-burden countries is not due to
Prize for his discovery of streptomycin. In his Nobel lack of efficacy of drugs in those countries, but rather
address, Waksman said: the confluence of medical and social factors that fuel
the ongoing tuberculosis epidemic: the coepidemic
In the treatment of tuberculosis, the more controlled of HIV; poverty; lack of a functioning public health
dosage of streptomycin and the supplementary use of
infrastructure; the economics of tuberculosis drug de-
PAS tended to overcome some of the limitations of
the antibiotic, notably its toxicity and the develop-
velopment; bureaucratic and doctrinaire approaches
ment of bacterial resistance. The recent introduction to tuberculosis control; lack of funding to support
of isonicotinic acid hydrazide suggests the possibil- basic research aimed at development of new drugs,
ity that its combined use with streptomycin will tend diagnostics, and vaccines; and apathy.
further to control the disease and overcome undesir- The articles in this issue of the Clinics in Chest
able complications. The conquest of the bGreat Medicine address many of these issues. They have
White Plague,Q undreamt of less than 10 years ago, been written by leaders in the basic science, medical,
is now virtually in sight. [1] and public health communities who have broad per-
spective and expertise in their fields, and who are
That same year, Rene and Jean Dubos, in their still
engaged in efforts locally and globally to control the
important and prescient book The White Plague, wrote:
worldwide tuberculosis epidemic.
However useful in specialized cases, vaccination, I would like to dedicate this issue to all patients
antimicrobial drug therapy, or other therapeutic who have tuberculosis, in hopes that the information
measures cannot possibly solve the social problem and knowledge offered within will improve their lives
of tuberculosis. . .. It is only through gross errors in in a direct and immediate way.
social organization, and mismanagement of individ-
ual life, that tuberculosis could reach the catastrophic
Neil W. Schluger, MD
levels that prevailed in Europe and North America
Division of Pulmonary, Allergy, and
during the nineteenth century, and that still prevail in
Asia and much of Latin America today. [2] Critical Care Medicine
Columbia University College of
Unfortunately, the Dubos’ assessment proved more Physicians and Surgeons
accurate than Waksman’s. Despite great advances in PH-8, Room 101
immunology, microbiology, and drug development, 630 West 168th Street
tuberculosis remains among the great public health New York, NY 10032, USA
challenges of our time. The failure to reduce tuber- E-mail address: ns311@columbia.edu
0272-5231/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccm.2005.03.001 chestmed.theclinics.com
x preface