The ongoing challenge of latent

tuberculosis
H. Esmail1,2, C. E. Barry 3rd3, D. B. Young1,4 and R. J. Wilkinson1,2,4
1

Department of Medicine, Imperial College, London W2 1PG, UK

Clinical Infectious Diseases Research Initiative, Institute of Infectious Diseases and Molecular Medicine,
University of Cape Town, Observatory 7925, South Africa
3
Tuberculosis Research Section, NIAID, NIH, Bethesda, MD 20892, USA
4
MRC National Institute for Medical Research, London NW7 1AA, UK
2

rstb.royalsocietypublishing.org

Review
Cite this article: Esmail H, Barry CE 3rd,
Young DB, Wilkinson RJ. 2014 The ongoing
challenge of latent tuberculosis. Phil.
Trans. R. Soc. B 369: 20130437.
http://dx.doi.org/10.1098/rstb.2013.0437
One contribution of 12 to a Theme Issue After
2015: infectious diseases in a new era of
health and development.
Subject Areas:
health and disease and epidemiology,
immunology, microbiology

The global health community has set itself the task of eliminating tuberculosis
(TB) as a public health problem by 2050. Although progress has been made in
global TB control, the current decline in incidence of 2% yr21 is far from the
rate needed to achieve this. If we are to succeed in this endeavour, new strategies to reduce the reservoir of latently infected persons (from which new cases
arise) would be advantageous. However, ascertainment of the extent and risk
posed by this group is poor. The current diagnostics tests (tuberculin skin test
and interferon-gamma release assays) poorly predict who will develop active
disease and the therapeutic options available are not optimal for the scale of
the intervention that may be required. In this article, we outline a basis for
our current understanding of latent TB and highlight areas where innovation
leading to development of novel diagnostic tests, drug regimens and vaccines
may assist progress. We argue that the pool of individuals at high risk of progression may be significantly smaller than the 2.33 billion thought to be
immune sensitized by Mycobacterium tuberculosis and that identifying and
targeting this group will be an important strategy in the road to elimination.

1. Introduction
Keywords:
latent tuberculosis, Mycobacterium tuberculosis,
elimination, diagnosis, treatment,
natural history
Author for correspondence:
H. Esmail
e-mail: h.esmail@imperial.ac.uk

Mycobacterium tuberculosis (Mtb) is a pathogen that has coevolved with anatomically modern humans [13], co-migrating from Africa as our population
expanded to cover every area of the globe (see box 1). It has been estimated
that in 2006 there were more cases of tuberculosis (TB) than in any other year in
recent history [7], and yet the ambitious vision adopted by the World Health
Organization (WHO) and Stop TB partnership is to eliminate TB as a public
health problem by 2050 [8]. This has been defined as achieving an incidence
rate of less than 1 case per million of the global population; for comparison, the
2012 rate is 1220 cases per million [9]. Various regions of the world are in different
phases of the TB epidemic and will require different strategies to make progress
towards elimination. In the twentieth century, much of Western Europe, North
America and parts of East Asia saw dramatic reductions in TB incidence through
major social and economic progress and implementation of improved TB control
and treatment programmes, with reduction in TB cases of up to 8.8% yr21 being
achieved after the second world war [10]. By contrast, sub-Saharan Africa and
Eastern Europe/Central Asia in particular suffered a steep increase in incidence during the 1990s owing to the HIV epidemic and the social and
economic disruption following collapse of the Soviet Union, respectively.

(a) Targeting tuberculosis

Progress recently has certainly been made in global TB control aided by a number
of internationally agreed targets over the last two decades. In the 1990s, as part of
the WHO DOTS strategy, a commitment to identify 70% and cure 85% of TB cases
by 2005 was made, and this was largely achieved in many parts of the world [11].

& 2014 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution
License http://creativecommons.org/licenses/by/3.0/, which permits unrestricted use, provided the original
author and source are credited.

Subsequently, as part of the response to the United Nations Millennium Development Goal 6 to combat HIV/AIDS, malaria
and other major diseases, the Stop TB partnership set a target
of reducing the global mortality and prevalence of TB disease
by 50% compared with 1990 levels. By 2012, a 37% reduction
in global prevalence of TB had been achieved (although not
on track to achieve 50% reduction by 2015) and a 45% reduction
in mortality (on track to achieve 50% reduction by 2015) [12].
Latest estimates suggest that in 2012 there were 8.6 million
new cases of TB and 1.3 million deaths, with the global incidence of TB falling 2% yr21 over recent years [9]. In this
context, the 2050 elimination target seems particularly bold,
requiring a historically unprecedented 20% yr21 reduction in
global incidence [10]. A detailed strategy to make the 2050
vision a reality is currently being developed and will be
announced in 2014 with interim targets for 2025 and 2035
being proposed. In the initial phase, scale-up and widespread
implementation of current TB control measures coupled with
continued socioeconomic development particularly within the
BRICS (Brazil, Russia, India, China and South Africa) countries
along with continued antiretroviral therapy (ART) roll out in
sub-Saharan Africa could result in reductions in TB incidence
of 10% yr21. However, to bring global incidence down towards
current levels seen in North America and parts of Western
Europe, considered to be in the elimination phase (less than
100 cases/million yr21), by 2035 will require development of
novel technologies and approaches though research and innovation. Whereas until now TB control has focused on
detection and management of active disease, which will continue to be important, a renewed focus on understanding and
managing the important reservoir of infected humans with
latent infection will be critical to future progress. Modelling
suggests that mass treatment of latent TB would be one of
the most effective ways to reduce incidence of TB [10,13], but
with current treatment and diagnostics this would involve up
to one-third of the worlds population taking three to nine
months of anti-tuberculous therapy, which is neither desirable
or feasible. In this article, we will highlight our current understanding of latent TB and the gaps in our knowledge that need
to be filled to develop more predictive diagnostic tests, effective
short-course treatments and vaccines.

2. Diagnosing latent infection

Latent tuberculosis may be defined for convenience as that which
is unaccompanied by symptoms and physical signs, causes no

obvious disturbance and is not recognised by the physician.

There is no sharp distinction between latent and manifest tuberculosis, and in some instances latent tuberculosis is more
extensive than that which is recognisable. Ability to distinguish
between latent and manifest disease will vary with the means
available for diagnosis. (Opie & McPhedran 1926 [14, p. 347])

Opie & McPhedrans [14] characterization of latent TB remains

accurate and insightful 88 years after writing, and it is striking
that the means physicians have available to distinguish
active and latent TB remain essentially unchanged today.
Chest X-ray, tuberculin skin testing and sputum investigation
for evidence of Mtb remain the cornerstones of diagnosis. Interferon-gamma release assays developed over the last decade
represent an evolution in diagnosis of latent TB rather than a
conceptual shift as they also indicate immune sensitization.
Advanced imaging such as computed tomography (CT) and
combined positron emission tomography/computed tomography (PET/CT) used to identify minimal disease and
invasive sampling also now play a role in specialist centres in
high-income countries in selected cases. The diagnostic pathway nevertheless still ends in a binary classification with
those considered latently infected being potentially eligible for preventive therapy and those with active disease,
standardized treatment.

(a) The development of the tuberculin skin test

By the end of the nineteenth century, the notion that a latent
period of infection could occur in TB was widely accepted. Several autopsy case series carried out around the turn of the
century demonstrated that Mtb was frequently present in persons who died of causes other than TB by inoculating material
into rabbits or guinea pigs, where secondary infection was
observed [15,16]. Detecting the presence of infection in asymptomatic living people was more challenging. Kochs discovery
and development of tuberculin, a heat-killed culture filtrate of
TB that was proposed unsuccessfully as a cure in 1890, provided a useful diagnostic test, unmasking occult infection by
inducing systemic reaction following subcutaneous injection
[17]. Over a period of 60 years, this was refined into an intradermal skin test using a standardized purified protein derivative
of tuberculin (PPD) and measuring the induration formed
after 4872 h. The result, expressed as millimetres (mm) of
induration, is a continuous variable where the threshold for a
positive result can be varied to modify diagnostic sensitivity
and specificity. The dose of PPD was optimized to maximize
sensitivity in distinguishing healthy close contacts of TB from
healthy non-contacts, and the tuberculin skin test (TST) remains

Phil. Trans. R. Soc. B 369: 20130437

Current evidence suggests that Mtb was already established as an infection of ancient human populations prior to migration out
of Africa. In these small isolated hunter gatherer populations, sustained infection would be favoured by low-virulence pathogens capable of persisting within the human host by chronic or latent infection and transmitting to susceptible new birth cohorts
years or decades after initial infection. Higher virulence pathogens with shorter incubation would result in self-terminating epidemics owing to elimination of susceptible hosts [2,4]. It has been speculated that increases in human population density
associated with the Neolithic Revolution in farming and the Industrial Revolution in Europe may have favoured the emergence
of Mtb strains with greater virulence and shorter incubation periods [1,3,5]. According to this model, carriage as an asymptomatic commensal may have been the predominant mode of Mtb infection in ancient human populations and may have shaped
the natural immune response. The current predominant high mortality form of TB would then represent a relatively recent challenge to human health. This model is consistent with phylogenetic analysis of global Mtb and with epidemiological differences
between the spread of modern Beijing strains and that of ancient Mycobacterium africanum [6].

rstb.royalsocietypublishing.org

Box 1. Origins and evolution of latency in tuberculosis.

Defining the prevalence of infection at a global and a regional

level is critical to understanding the potential size of the
reservoir of infection and planning intervention strategies.
One of the most widely quoted statistics is that one-third of
the worlds population is infected by Mtb, emphasizing the

Phil. Trans. R. Soc. B 369: 20130437

3. Estimating the global burden of latent

tuberculosis

huge scale of the problem [26]. However, no test actually

demonstrates the presence of infection and it is useful to consider the data upon which the statement is made. Prevalence
of infection in a population is not directly estimated from population-wide tuberculin surveys but derived from the annual
rate of infection (ARI), which can be either directly determined
from focused tuberculin surveys (usually in school children) or
indirectly estimated from incidence of active TB (itself usually
estimated from case notification, disease prevalence or mortality data), using the equation ARI incidence/coefficient,
as risk of infection is determined by contact with infectious
cases [27]. In addition, the change in ARI over time should
be known in order to accurately determine prevalence of
infection throughout the population.
In 1999, WHO convened a consensus group comprising 86
experts and epidemiologists who evaluated the best available
data for all countries up to 1997 for a number of TB indicators
including prevalence of infection [26]. Recent good quality
tuberculin surveys were available for only 24 countries and
the rate of change in ARI was only accurately known in a minority. For the majority of countries for which good quality
tuberculin surveys were not available (or it was not possible
to confidently extrapolate from countries with good data),
ARI was derived from the incidence of smear positive disease
using the equation above with the coefficient of 50 (for countries
where HIV prevalence in TB cases was less than 5%) coming
from Styblos rule [28] (which states that a smear positive pulmonary TB incidence of 50/100 000 yr21 corresponds to ARI of
1%, potentially an overestimatevide infra). The authors estimated that 32% of the worlds population was infected with
Mtb but acknowledged the lack of good data and limitations
of the models to determine prevalence of infection and did
not provide an uncertainty estimate.
Tuberculin surveys, while widely performed and a potentially cost effective way to monitor changes in the burden of
infection, have several limitations. Aside from the inherent issue
of observer variability, interpreting results to determine the proportion that are immune sensitized by Mtb is challenging as the
specificity of the test varies between populations depending
upon exposure to environmental mycobacteria and BCG vaccination. Analysing the distribution of TST reactions using
mixture models and other techniques to identify bimodal patterns (as the reaction to TST is greater following immune
sensitization by Mtb than environmental mycobacteria or BCG)
can be performed but not all distributions lend themselves to
this form of analysis [29,30]. Systematic surveys using IGRA
have not often been performed and the need for venepuncture,
specialist laboratories and cost may limit widespread use. In
addition, the dynamics of IGRA conversion and reversion over
time and hence sensitivity to detect remote infection are less
well understood, although data from IGRA surveys could be
used with tuberculin surveys to refine estimates of latent TB
infection (LTBI) prevalence [31]. A recombinant ESAT-6/
CFP-10 skin test is currently in development with early clinical
studies showing superior specificity to TST and correlation
with whole blood Quantiferon results [32]; such a test may
ultimately prove particularly useful for surveys of this kind.
Modelling the risk of infection from the incidence of smear
positive active disease seems appropriate but Styblos rule
makes some key assumptions informed by six studies between
1921 and 1971: each smear positive incident case is infectious for
2 years (corresponding to two prevalent cases), and each prevalent case results in 10 new infections per year. Hence the rule, an

rstb.royalsocietypublishing.org

widely used globally today and is generally considered to

demonstrate the presence of infection [18].
But to what extent can this relationship be assumed to be
true in the obvious absence of autopsy studies to relate premorbid TST to post-mortem identification of viable bacilli? The
guinea pig natural infection model, initially developed by
Riley to investigated airborne transmission in which air from
side rooms or wards where TB-infected patients are resident
is vented over chambers housing guinea pigs, can provide
some evidence. In these studies, the distribution and magnitude of TST reactions found in guinea pigs was similar to
humans; at autopsy, evidence of infection was found in 0% of
guinea pigs with tuberculin reactions of 05 mm (negative
reaction), 92% of guinea pigs with TST of 14 mm or more or
evidence of necrosis, but only 25% with TST of 613 mm [19].
This finding, that not all naturally infected guinea pigs with
positive TST have evidence of infection at autopsy, was subsequently confirmed by others [20,21]. Equally, if positive TST
indicated presence of infection and negative TST absence of
infection, then treatment of latent infection might be expected
to cause a reversion of status. In the United States Public
Health Service Trials 13 176 household contacts who were
initially tuberculin reactors had TST repeated at 12 months;
6.5% of contacts who received placebo converted to negative
and 7.9% of isoniazid treated subjects converted to negative.
However, isoniazid reduced the 10-year incidence of TB by
59% in those that remained TST positive at 12 months and by
38% in those that converted at 12 months, indicating that isoniazids efficacy to prevent disease was not associated with a
capacity to induce TST reversion [22]. From this, we conclude
that TST provides evidence of immune sensitization by Mtb
and is a correlate of TB infection but usually remains positive
even if infection is treated.
Because TST can register low-level false positive results due
to sensitization by environmental bacteria and BCG, interferon
(IFN)-gamma release assays (IGRAs) were developed to
improve specificity of the diagnosis. While these tests provide
a reasonably good measure of TB exposure and their negative
predictive values are very high (IGRA 99.7%, TST 99.4%),
they are, like TST, poorly predictive of progression to active disease (positive predictive value: IGRA 2.7%, TST 1.5%) [23,24].
As a result, when these immunodiagnostic tests are used
as a guide for administration of preventive therapy, the
number needed to treat (NNT) to prevent one case of active
TB is high. In a systematic review of 11 studies involving
73 375 HIV-uninfected participants where presence of infection was mainly determined by TST and participants were
randomized to isoniazid or placebo, the pooled NNT was
100, ranging from 36 in recently infected household contacts
and up to 179 in those remotely infected [25]. After the initial
need to develop sensitive and then specific tests for latent TB,
we now need to develop tests that are better able to predict
who will develop active TB.

known

7 billion

unknown

? immune sensitized at
some stage but reverted

? high risk of active TB

8.6 million active TB/yr

Figure 1. Reservoir of TBwe currently have estimates for proportion of population that are immune sensitized (large circle) and number of cases of active TB annually
(small filled circle). As TST and IGRA reversion can occur, total number of exposed persons may be greater than this (larger dashed circle), in addition TST and IGRA are
only moderately sensitive for active TB. A much smaller pool of people may be at much higher risk of TB (bottom small dashed circle) and also a proportion of people
may receive considerable protection against reinfection (top small dashed circle). Identifying these additional populations may be very valuable. (Online version in colour.)
incidence of 50/100 000 yr21 results in an infection rate of 1000/
100 000 yr21. This may be less applicable in the contemporary
era as improved diagnosis and treatment may have reduced
the average duration of infectiousness significantly and factors
such as population density, success of TB control programmes,
the prevalence of HIV infection and the prevalence of drugresistant TB will also have influenced transmission [33]. A
recent analysis of data from East Asian countries between
1975 and 1994 by van Leth et al. [34] determined that the
number of infections per prevalent smear positive case was
2.65.9 yr21, so at least in some parts of the world Styblos
rule overestimates infection [33,34]. Smear positive TB is also
not homogeneous. Jones-Lopez et al. [35] have shown that in
only 45% of smear positive cases could Mtb be cultured from
cough aerosols generated through 10 min of strong coughing.
In addition, the variability in colony-forming units (cfu) generated was great (1378 cfu). They also showed that infection of
household contacts was significantly greater when the index
cases had a high cough aerosol cfu compared with low or no
cough aerosol cfu [35]. Confirming that cases of TB transmit variably, Escombe et al. [36] using the Riley guinea pig model of
airborne infection in an HIV/TB ward in Peru showed that
8.5% of admissions were responsible for 98% of infections in
guinea pigs. Further complicating the situation is the fact that
although previously Mtb was considered a highly invariant
pathogen, recent large-scale whole genome sequencing projects
have made it clear that the pathogen has continued to evolve.
As human population density has expanded exponentially and
living conditions have shifted from low-density agrarian conditions to high-density urban conditions, new genetic variants
of Mtb have emerged, displacing formerly resident strains [37]
(see box 1). More modern strains differ from their ancient progenitors, notably at a cellular level in the magnitude of the
innate immune response they elicit [38]. More variation in
these strains has been observed than previously expected, and
plausible links to enhanced transmissibility have been inferred

from cluster size in human populations exposed to multiple

strain clades [39]. Comparison of experimental infection outcome
in non-human primates with modern and ancient strains reveals
strikingly different outcomes of infection among currently prevalent clades [40]. The contribution of strain variability to
differences in transmissibility is likely to be geography and population-density specific, and employing any general rule to
establish prevalence of infection base on reported cases is likely
to be extremely inaccurate.
Accurate estimation of the proportion of the world that
is infected using the currently available tools is extremely difficult. Better data and a better understanding of how parameters
change regionally or in certain situations (such as drug resistance) may allow for the development of more sophisticated
models [41] that will allow more accurate assessment of the
prevalence of infection and estimates of useful subgroups
such as the prevalence of drug-resistant LTBI or the proportion
of LTBI related to a recent infection. However, ultimately what
we want to know is the proportion that is highly likely to
develop active disease (figure 1).

4. The natural history of tuberculosis

The natural history of TB is more complex than most bacterial
pathogens. The incubation period is prolonged and the outcome of infection variable depending upon both host and
pathogen. Much of our current understanding still arises
from piecing together historical studies and evidence from
animal models that often fail to replicate key aspects of
human disease. However, understanding this natural history
of infection is critical to accurate categorization of TB-infected
persons, identification of correlates of risk and protection,
and development of novel interventions.
Although TB can develop in virtually any part of the
body, disease involving the lungs (which occurs in 60 75%

Phil. Trans. R. Soc. B 369: 20130437

2.33 billion TST +ve

rstb.royalsocietypublishing.org

? degree of protection
from re-infection by Mtb

The lifetime age-weighted risk of TB following infection in settings with low exogenous reinfection is estimated to be 12%
[60]. Careful follow-up in placebo-controlled intervention
studies has demonstrated that disease is most likely to occur
in the first year following infection, with stepwise reduction
year on year over the following 510 years (figure 2), by
which time incidence approaches that of uninfected contacts
[22]. The different manifestations of TB occur at different intervals following infection, with pleural TB, TB meningitis and
miliary TB occurring after a shorter interval than pulmonary
or other extra-pulmonary sites.
Reactivation several decades after initial infection occurs
[61], but as observational studies with close follow-up
rarely continue beyond 10 years it is difficult to assess how
common reactivation is outside this timeframe. In addition,
conventional observational studies make it difficult to evaluate whether disease relates to the initial infection event or
subsequent reinfection. Borgdorff et al. [62] applied a molecular epidemiology approach (using restriction fragment length
polymorphism of IS6110 /2 polymorphic GC-rich
sequence) to 12 222 cases of TB over a 15-year period in the

1200
TST +ve placebo

incidence /100 000

1000

TST +ve isoniazid

TST ve

800
600
400
200
0
0

4
6
years since contact

10

Figure 2. Incidence of TB in household contacts of TB treated with isoniazid

or placebo over 10 years by TST status (TST 2ve less than 5 mm). Based on
data for 26 833 persons from Ferebee [22].
Netherlands and identified 1095 linked secondary cases
from 688 source cases. The median incubation period (time
between predicted date of infection and onset of symptoms
in the secondary case) was calculated to be 1.26 years, and
the serial interval (time between symptom onset in source
and secondary case) was found to be 1.44 years with 83%
of secondary cases occurring within 5 years of the source
case and more than 95% within 10 years [62].
Studies of immigrants from high to low burden countries
provide further insight. Risk of TB is especially high in the
first few years following migration, but migrants remain at
higher risk of TB for decades after entry [63]. It is difficult
to establish whether this relates to delayed reactivation or
(re)infection following recent transmission either from visiting country of origin or from the local community.
However, McCarthy [64] has shown that diagnosing TB in
migrants is very rare more than 15 years after arrival if they
have low risk of re-exposure. In this study, of 230 migrants
from high burden countries in Asia diagnosed with TB in
London in the 1980s (low burden setting), 10.4% had arrived
in the UK 1115 years previously and 5.6% more than 15
years previously; however, in those who had never returned
to Asia since migrating and had no known UK TB contact,
only 3.9% had arrived 11 15 years previously and 0.8%
more than 15 years previously [64].
Furthermore, the observation that the elderly in low TB
burden settings have a higher incidence of TB is often, possibly
incorrectly, interpreted as providing evidence of prolonged
latency and reactivation following immunosenescence. However, careful evaluation of birth cohorts shows that this
apparent increased risk is an artefact of falling transmission,
and younger adults are still invariably at greater risk of TB
than the elderly [65]. These data show that the common view
that reactivation TB disease often occurs decades after initial
infection may be overstated; the majority of cases occur
within 18 months of infection and disease resulting from reactivation more than 10 years after infection may be rare. The risk of
disease is also not constant over time; following a single

Phil. Trans. R. Soc. B 369: 20130437

(a) Lifetime risk of infection progressing to disease

incidence of TB in household contacts

by infection status and intervention

rstb.royalsocietypublishing.org

of cases) is necessary for transmission of infection, in particular pulmonary cavitation facilitates efficient Mtb replication
and transmission. There is some evidence that suggests Mtb
may specifically exploit the immune response through conservation of immunodominant epitopes [42], which could
promote the induction of immunopathology that leads to
lung cavitation. It is immunocompetent adults that contribute
most to disease transmission. These individuals who are most
effective at transmitting are often sputum smear positive and
cough spontaneously, thereby generating infectious particles.
Infection is initiated by droplet nuclei of less than 5 mm that
can remain suspended in the air for hours (if not disrupted by
turbulence) and be inhaled by contacts sharing the same
environment [43]. Around 3050% of close household contacts will develop evidence of immune sensitization as a
result of infection [44].
A single droplet nucleus ( probably containing 110 bacilli)
can initiate infection, with the site of implantation following
chance distribution strongly influenced by the particle size
across the lung lobes [45]. The early stages of infection are
characterized by a localized macrophage-rich alveolitis, lymphatic spread to regional mediastinal lymph nodes and a
low-grade bacillaemia allowing distant dissemination [46,47].
Approximately 210 weeks following initial infection, a cellmediated immune response develops, signified by tuberculin
conversion, facilitating the development of granulomas
which promote control of infection [48] (box 2). This initial
infection is often asymptomatic but may be associated with
fever, mild chest symptoms and increased inflammatory markers [52,53]. The primary infiltrate may be visible on chest
radiograph in 26% of older children and adults [54] (this
may be considerably higher in young children [55,56]). In a
small proportion (less than 15%), the visible primary infiltrate
may progress ( progressive primary TB), but in general the
lesion heals and often eventually calcifies. If progressive TB
disease subsequently develops within the lungs, it does so at
a distant site, most commonly arising apically or sub-apically;
the mechanism for this characteristic localization is poorly
understood and the source of some speculation [5759].

infection, the risk is 12% over a persons lifetime; if no disease

develops after 5 years the lifetime risk might only be 2% and
after 10 years 0.5%.
It is likely that at least in some people for whom there is a
prolonged time interval between infection and eventual
disease presentation, episodes of subclinical reactivation had
occurred much sooner. Evidence for this comes from
twentieth century mass chest radiograph (CXR) screening programmes, which identified asymptomatic persons with no
previous history of disease but with apical fibrotic scarring
felt to represent inactive or arrested TB. These individuals
were up to 15 times more likely to develop TB than those
having normal CXR, with the risk of developing TB steadily
falling over a 510 year period of observation [66]. In addition,
studies in Europe and America at a time of rapidly falling TB
incidence showed that up to 70% of persons developing TB
(with no history of TB and usually no clear contact history)
had evidence of fibrotic scarring on previous CXR [67,68].
This suggests that, in a proportion of people, the disease
may follow a cyclical waxing and waning course with earlier
reactivation initially arrested by the host delaying disease presentation. In addition, it is clear that the subclinical phase of
active disease prior to clinical presentation may be several
months, as evidenced by the demonstration of culture positivity in asymptomatic persons. In HIV-infected persons in high
burden settings, prevalent asymptomatic TB has been shown
to be present in up to 8.5% [69].

(b) Immunosuppression
A number of conditions are associated with increased risk of
progression of TB, with HIV infection and anti-tumour necrosis
factor (TNF) therapy being two well-documented examples.
The effect of anti-TNF therapy is particularly striking in the
macaque model of latent TB treatment, with anti-TNF resulting
in almost universal reactivation in animals that had initially no
signs or symptoms of active disease for at least six months from
the time of infection [70]. In humans treated with anti-TNF
therapy, especially with infliximab, the risk of TB is increased
initially up to 20-fold with 43% of TB cases occurring within
the first 90 days of administration of anti-TNF therapy, demonstrating how rapidly active disease can be precipitated [71].
However, reactivation is by no means universal. In an evaluation of the implementation of LTBI screening prior to

anti-TNF therapy in Spain, 56 patients with positive TST

(more than 5 mm) were identified who did not receive any isoniazid prophylaxis and in only one case did TB occur following
anti-TNF treatment [72]. The impact of HIV on latent infection
can be more challenging to evaluate as the majority of HIV/TB
studies are performed in high burden settings where reinfection
complicates understanding of the natural history of a single
infectious episode. In studies in low burden settings, widespread ART use can also be a confounder. In addition and in
contrast to anti-TNF therapy, HIV-associated immunosuppression is slowly progressive. In studies from the high burden
setting of the South African mines, the risk of TB infection has
been shown to double within the first year, with Mtb strains
significantly more likely to be unique within 2 years of HIV-seroconversion than later on in the disease, suggesting that TB may
more likely be precipitated by reactivation early in the course of
HIV and that recent reinfection with subsequent rapid progression occurs in more advanced immunosuppression [73].
Studies from low prevalence settings (Spain and Switzerland)
suggest that the rate of TB in HIV-infected persons who are
TST positive but untreated reduces over time, with most of
the excess cases of TB compared with TST negative occurring
within the first 2 years of follow-up; in total, only 1012% of
this very high-risk group developed TB over a follow-up
period of up to 5 years [74,75].

(c) Protection from reinfection

Although latently infected persons are at greater risk of developing disease through reactivation than are those uninfected
there is some evidence that at least a proportion are protected
against subsequent (re)infection. Andrews et al. [76] reviewed
18 studies in which 19 886 persons with or without latent TB,
as evidenced by tuberculin reactivity, were followed up for
active disease in the absence of intervention. The majority of
these studies were published before 1950, and largely involved
nursing and medical students entering clinical practice and
exposed to extremely high annual rates of infection (median
33.6%). The incidence of disease in those with LTBI at entry
was 5.1/1000 person-years and in those uninfected at entry
was 18.2/1000 person-years. Once adjustments were made for
reactivation and average timing of infection, risk reduction in
those with LTBI was found to be 79%. Historical healthcare settings clearly represent an extreme scenario, however other

Phil. Trans. R. Soc. B 369: 20130437

Mtb adapts to environmental triggers such as hypoxia, nutrient starvation and reduced pH encountered during infection by
altering metabolism and arresting replication [49]. Adaptation often includes transient transcriptional activation of a characteristic set of approximately 50 genes under the control of the DosR dormancy regulator, together with additional genes
appropriate to the specific environmental cue [50]. The products of these induced genes are currently being explored as
potential biomarkers. Resumption of replication following exposure to a more favourable environment is presumed to
involve analogous transcriptional and metabolic reprogramming, including cell wall changes mediated by a family of transglycosylase enzymes [51]. The ability of Mtb to persist in a reversible non-replicating state is a key virulence factor but the
direct equation of clinical latency with non-replicating mycobacteria and active disease with replicating mycobacteria is an
oversimplification. Although active disease is characterized by uncontrolled increases in bacillary numbers, imaging and
autopsy studies show that there are numerous micro-environments that exhibit varying degrees of progression and healing.
A partially overlapping heterogeneous spectrum is seen in latent stages of infection. The prolonged courses of treatment
required to prevent relapse following treatment of active TB are thought to be due to persistent populations of bacilli,
whereas the efficacy of isoniazid as preventive therapy in latent infection is thought to be due to its effect on replicating bacilli.
In short, although absolute numbers and proportion clearly differ it seems likely that both replicating and non-replicating
bacilli are present in both latent infection and active disease.

rstb.royalsocietypublishing.org

Box 2. The nature of Mtb in latent infection.

(a)

disease

rstb.royalsocietypublishing.org

1 progression

infection

unstable
Prc

Prd

Prc

Prc

Prc

stable

possible precipitating factors

HIV
anti-TNF therapy
malnutrition
Vit D deficiency
viral infection

possible predisposing factors

HIV
malnutrition
diabetes
alcoholism
pro/anti inflammatory imbalance

(b)

Prc

pathology

unstable LTBI

subclinical active phase

clinical TB

culture
imaging
time (months)

Figure 3. (a) Following infection, there may be a critical period where fate of infection is determined with predisposing factors (Prd) influencing this outcome. In a
small proportion, the primary infection may be progressive; in those that control primary infection, a proportion may eliminate TB or exert highly effective control
and be at very low risk of reactivation. In the third group, control may be unstable waxing and waning in response to a variety of precipitating factors (Prc) with
reactivation of TB most likely to occur in this high-risk group. (b) Precipitating factors (Prc) may lead to progression of disease. Prior to presentation these individuals
may pass through a subclinical phase of active infection which may last months; during this phase Mtb may be isolated by culture or pathology may be visible
through imaging prior to symptomatic presentation. (Online version in colour.)
approaches have also suggested a protective effect of infection.
Using national datasets over long periods of time to model the
dynamics of TB, Vynnycky & Fine [60] from data for England
and Wales 19001990 predicted a 1641% protection from an
initial infection against reinfection, and Sutherland et al. [77]
from data for the Netherlands predicted 6381% protection.
Taking an alternative strategy, Brooks-Pollock et al. [78] used
cross-sectional household data from Lima, Peru from 1996
2002 to propose 35% protection. In detailed studies of reinfection in the rabbit model, Lurie [79] demonstrated that control
of a re-infecting strain in previously infected rabbits was
mediated by tissue resident mononuclear cells, with efficiency
of control relating to the extent of the primary lesion from the
initial infection. The mechanism of protection in humans is
not known but clearly an improved understanding of this
could greatly facilitate vaccine development.

5. Integrating the spectrum of tuberculosis with

natural history
We and others have suggested that asymptomatic people
considered to have latent TB might be better considered
as part of a spectrum of infection states where at one end

infection may have been eliminated, while at the other end disease may be active but in a subclinical form, and between these
two extremes infection is controlled in a quiescent state [8084].
When carefully considering the natural history of infection, it
seems plausible that soon after initial infection and immune
sensitization there are three main possible outcomes influenced
by predisposing factors that determine the course of infection
during this critical phase and alter the proportions in each
group (figure 3). Some may initially develop primary progressive disease; this may be a very small proportion in adults but
would likely be more common in advanced immunosuppression and infants. A second group (a high-risk groupthe
main group from which reactivation disease arises) enter a
more unstable state with infection taking a waxingwaning
course during which periods of progression triggered by
precipitating factors may be followed by control (which may
lead to evidence of immunopathology) or the development of
clinical disease. Some precipitating factors may be more
potent than others; very potent precipitating factors (such as
anti-TNF and HIV) may have the effect of causing rapid progression over a short time interval. It is also in this group that
isoniazid preventive therapy may be most effective. A third
group may rapidly and effectively control infection and
eventually may even sterilize the organism and may be at

Phil. Trans. R. Soc. B 369: 20130437

control/ elimination

extremely low risk of progressing to active disease even in the

presence of precipitating factors.

Phil. Trans. R. Soc. B 369: 20130437

Some predisposing and precipitating factors may overlap (figure

3). Many predisposing factors from the host side are known and
can be considered as generally immunosuppressive (HIV, malnutrition, chronic kidney disease, type 2 diabetes mellitus, etc.)
but many more may be unknown, poorly characterized, have
more subtle effects and may or may not be genetically mediated.
In particular, rather than just immunosuppression alone it is
becoming more apparent that the extremes of the immune
response may lead to detrimental outcome in TB, with a more
balanced response being optimal (the so-called Goldilocks
effect); weak responses may lead to unopposed bacillary replication whereas aggressive responses may lead to tissue damage
and necrosis which may provide a more favourable environment
for the bacillus. One of the implications of this, when considering
biomarker discovery for novel diagnostics (see below), is that
there may be at least two distinct correlates of risk.
Leukotriene (LT) A4 hydrolase (LTA4H) mediates the balance of pro-inflammatory eicosanoid LTB4 and antiinflammatory lipoxin A. Zebrafish larvae in which LTA4H is
over- or under-expressed are made hyper-susceptible to
Mycobacterium marinum infection compared with wild-type
either by low levels of LTA4H resulting in increased lipoxin A
and impaired TNFa production, or excessive LTA4H resulting
in increased LTB4 and increased TNFa production [85].
Humans who are heterozygous for a single-nucleotide polymorphism of LTA4H promoter rs17525495 (C/T) appear to
have the best clinical outcome from TB meningitis; those who
are homozygotes for the T allele (T/T) have increased LTA4H
expression and inflammatory cerebrospinal fluid, but derive
significantly greater benefit from dexamethasone therapy compared with the C/C genotype [8587]. In humans, plasma
prostaglandin E2 and lipoxin A levels likewise correlate with
disease susceptibility in a similar bimodal fashion with both
insufficient and abundant responses tied to disease exacerbation
(Mayer-Barber & Sher 2014, unpublished results, personal
communication).
Another intriguing and poorly understood predisposing
factor is age. It is a consistent and striking feature of TB that
the age of infection affects the risk of subsequently developing
disease [56,88]. Infants and young children, especially those
less than 2 years, are at considerable risk of developing disease
following infection. Older children (510 years old) have consistently been shown to be at the lowest risk of TB following
infection especially, whereas peri-pubescent adolescents and
young adults are at much greater risk of developing cavitary
TB compared with children less than 10 years old [56,88]. It
has been suggested that this may relate to the immunoendocrine
effects mediated by the balance between dehydroepiandrosterone (DHEAa precursor of sex steroids) and glucocorticoids
[89]. DHEA levels start increasing from 7 years old, peak in
early adulthood and reduce in older adults [88]. One of
DHEAs many effects is as a glucocorticoid antagonist, and
the cortisol : DHEA ratio has important immunological consequences. Recently, DHEA has been shown to influence
dendritic cell function to promote Th1 responses by increasing
interleukin (IL)-12 and diminishing IL-10 production following
Mtb stimulation, with increased expression of MHCI, MHCII
and CD86 expression resulting in enhanced T cell proliferation

rstb.royalsocietypublishing.org

(a) Predisposing and precipitating factors

and IFNg production [90]. So it seems plausible that pro-inflammatory responses in healthy adolescents and young adults have
detrimental effects leading to cavitary disease, anti-inflammatory responses in infants lead to their inability to control
replication and there is a more optimal balanced response in
older children.
A further incompletely understood historical observation
recently being revisited is the effect of monocyte : lymphocyte
ratio (M : L) on risk of disease. It has been found in both
animal models and clinical observation that extremes of
both low and high M : L result in a greater risk of developing
TB, but what is still not clear is whether this is a predisposing
factor [91] or a marker of progressive disease as TB treatment
normalizes the M : L ratio [92].
While some precipitating factors may be well known
and potent resulting in rapid progression of the at risk group
(e.g. anti-TNF therapy), some may just contribute to a fluctuating
course triggering disease in a minority. To consider an example,
an interesting observation is the seasonality of TB with increased
case notification that can be 2025% higher in spring/summer
compared with autumn/winter [9396]. This is striking for an
infectious disease with relatively prolonged and variable incubation. In common with other respiratory pathogens, one
explanation would be behavioural, with winter crowding leading to greater transmission, but modelling evidence and
analysis of unique and clustered Mtb strains suggest that this
cannot fully explain seasonality of TB [97,98]. A seasonal precipitating factor such as vitamin D deficiency or viral respiratory
infection (e.g influenza) is an alternative explanation. Vitamin
D, synthesized within the skin requiring UV light, acts as an
immunomodulatory and anti-inflammatory agent primarily
exerting its effect on the macrophage, facilitating enhanced control of mycobacteria through pleiotropic mechanisms [92,99]. A
number of clinical observations provide some support for the
role of vitamin D deficiency in inducing reactivation. TB patients
are well documented to have significantly lower vitamin D levels
than healthy household controls [100,101], and the spring/
summer peak in TB notifications is preceded by a winter
trough in vitamin D levels in Cape Town [102]. There is some evidence to suggest that seasonality is more pronounced in foreignborn cases (who may be at greater risk of vitamin D deficiency
owing to skin tone) compared with native cases in Europe
[103,104]. An alternative seasonal precipitant could be viral infection. It has recently been shown that the type 2 interferon (IFNg)
response critical for mycobacterial control can be impaired by the
downstream effects of type 1 interferons (IFNa/b) [105]. It has
therefore been hypothesized that viral respiratory infections
inducing a type 1 interferon response could lead to reactivation
by impairment of type 2 interferon facilitated control of Mtb. In
the mouse model, mycobacterial growth is enhanced and survival decreased in mice previously exposed to influenza by a
mechanism dependent on type 1 interferon signalling [106]. In
addition, historical observations and modelling of the 1918 influenza pandemic suggest a negative impact of influenza on TB
[107]. It is also possible that aside from host factors, variability
of the bacillus may influence rate of progression and disease outcome, and it is worth noting that there is clear evidence for
diversifying selection in genes of the bacillus whose functional
roles are less than clear [108].
Having a fuller understanding of these predisposing and
precipitating factors and the magnitude of their effects might
allow us to consider the impact of novel intervention strategies: for instance, whether widespread or targeted vitamin

ce

ac

si

ea

r
nc

ng

NNT to prevent progression

Figure 4. Acceptability of treatment relates to the duration and tolerability of

treatment and the likelihood of benefit ( prevention of progression to active disease). Improvements in drug regimens and/or improvements in predictability of
diagnostic tests should lead to improved acceptability of treating LTBI.
D replacement and/or influenza vaccination may impact the
incidence of TB.

(a) Markers of exposure

6. Novel diagnostic and preventive treatment

strategies for LTBI
Current approaches to management of latent TB centre on
those at greatest epidemiological risk of progressing to TB,
namely close contacts of active TB, HIV-infected and other
persons about to undergo, or with established, immunosuppression. Following diagnosis of latent TB, preventive
treatment is most commonly six to nine months of isoniazid
or a three month course of rifampicin or rifapentine and isoniazid [82] (in HIV-uninfected persons). However, in many
low- and middle-income countries with the highest burdens
of TB, many of these high-risk groups do not receive preventive therapy, especially household contacts. In these settings,
operational research may enhance the better implementation
of clinical trials evidence and recommendations. However, to
make significant progress towards TB elimination will also
require providing preventive treatment to an even wider
group of latently infected persons. One of the critical barriers
to this is the acceptability of the current intervention to individuals at risk as well as healthcare professionals and policy
makers. This is to a large extent influenced by the prolonged
duration of treatment and the high NNT to prevent a case of
active disease; hence acceptability should improve if either
or both of these are improved (figure 4). An inexpensive
on-the-spot diagnostic that provided an accurate assessment
of likelihood of progression coupled with a single-dose
prophylactic would be game-changing for TB control and
enable mass testing and treatment programmes with a
realistic chance of achieving eradication.
More predictive diagnostic tests are a goal of TB research
agendas but how these tests might be implemented in practice
also requires careful evaluation. The perfect diagnostic for
latent TB would be a cheap, low resource, point-of-care test

As discussed, TST and IGRA are poorly predictive but also

are suboptimal measures of Mtb exposure. Both TST and
IGRA show evidence of spontaneous reversion over time
and hence some people previously exposed to and immune
sensitized by Mtb may have a negative test. A proportion
of these individuals may be identified using a two-step testing strategy where the initial negative TST boosts the
immune response so subsequent TST or IGRA reverts to positive [109]. In addition, the sensitivity of TST and IGRA for
active TB infection is only 70 90% [110112]; not withstanding arguments as to why immunodiagnostic tests might be
compromised in someone with active TB, a test that is not
able to identify the most heavily infected individuals leaves
room for improvement. A more sensitive test for Mtb
exposure would give a better understanding of TB transmission dynamics and more accurate estimation of the
annual rate of infection. Because IGRAs detect IFN-g production after short-term incubation (16 24 h), they identify
primarily ESAT6/CFP10-specific effector cells and it is possible that prolonged assays measuring alternative cytokines or
flow cytometric assays could identify central memory populations that may be a better marker of exposure and history of
immune sensitization than the currently available tests.

(b) Predictive markers

Predictive biomarkers could be developed using two broad
approaches. Careful follow-up of an at-risk population over
time during which some develop TB might allow identification of biomarkers that were correlates of risk, although
such natural history approaches would not only require
very large numbers but would also have considerable ethical
considerations unless carried out in groups where observation is standard of care, such as contacts of multi-drug
resistant (MDR)-TB, where guidelines do not recommend
preventive therapy (e.g. over 35-year-olds in some countries)

Phil. Trans. R. Soc. B 369: 20130437

duration/intolerability of drug regimen

ty

ili

ab
pt

rstb.royalsocietypublishing.org

current
LTBI tx

with very high positive and negative predictive value that

maintained sensitivity in immunocompromised persons,
notably HIV infection and children. However, it is worth
noting that a single test cannot be both highly predictive for
active TB and a sensitive marker of exposure as these test
characteristics are to a degree a mutually exclusive. Another
consideration is over what period of time should these tests
be predictive? It may be substantially easier to develop tests
(from both a development and validation perspective) that
are predictive over a short period of timee.g. risk over 12
months rather than predictive of life time riskas it may be
over this time frame that transition into a subclinical phase
prior to symptomatic presentation occurs with characteristic
changes in the host response and bacillary numbers and metabolic state. Such shorter term predictive tests may be of
particular use in high burden settings where reinfection is
common and in HIV-infected persons and other immunocompromised groups (such as in type 2 diabetes mellitus and
chronic renal failure) where regular contact with healthcare
many allow for regular screening. In addition, recent contacts
of TB could be followed up annually during the period of greatest risk. Optimal test characteristics might change as the
elimination phase progresses; understanding these requirements will be helped greatly by models showing the impact
of different types of tests in different settings.

Shorter drug regimens are highly desirable for treatment of

both latent and active TB, with the current prolonged treatment
duration required to ensure that recrudescence of persisting
organisms does not occur after cessation of therapy. In addition,
a prolonged treatment course is associated with poor treatment
adherence in routine settings. Drugs that contribute most to
treatment shortening of active or latent TB (rifampicin and pyrazinamide) have the most potent sterilizing ability (usually
determined by evaluating relapse rate in the murine model),
whereas isoniazid, although rapidly bactericidal, largely acts
on replicating bacilli and has poor sterilizing activity [115]. As
an alternative strategy to shortening treatment for LTBI, rather
than targeting persisting organisms, might be to provoke resuscitation of non-replicating bacilli and couple this with rapidly
bactericidal therapies such as isoniazid. Such approaches will
require a far more sophisticated understanding of the mechanisms of resuscitation and ability to define the metabolic states of
single organisms.
The optimal duration of isoniazid as preventive therapy is
nine months [116]; with addition of rifampicin or rifapentine
this reduces to three months and rifampicin combined with
pyrazinamide is effective after two months of administration
(although unacceptable toxicity prevents widespread use of
this regimen) [117]. There are several novel or re-purposed
drugs in later stages of the TB drug pipeline [118] that have
impressive sterilizing ability either alone or in drug combinations (usually in pyrazinamide-containing regimens).
Bedaqualine (recently FDA approved for active MDR-TB) and
sutezolid (an oxalozidinone in phase 2 studies) seem most
promising in this respect. Nitroimidazole derivatives (delamanid and PA-824) and moxifloxacin are also possibilities [119].
Whether reductions in treatment duration beyond two months
could be feasible is not certain, however it is possible that this
pipeline provides several options for preventive treatment in
drug-resistant TB contacts. A major consideration for novel regimens in addition to cost [120] is toxicity and side effects, which

(d) Post-exposure vaccines and immunomodulation

TB vaccines can be administered either pre-infection, designed
to prevent infection from occurring, or post-infection, designed
to prevent latent infection progressing. H56 is a multistage vaccine comprising the Ag85B, ESAT-6 and Rv2660 antigens, is
one of the first vaccines to be designed to be used post-infection, and is soon entering phase I/IIa studies in South Africa
in latently infected and uninfected adults [126].
However, recently the MVA85A vaccine, a novel TB vaccine (administered pre-infection) that was in the most
advanced stage of clinical development, yielded disappointing results that demonstrated safety but no significant
efficacy over placebo in preventing TB when administered
to BCG-vaccinated, HIV-negative infants with no evidence
of latent TB infection, in a high burden setting [127]. The
resulting debate has highlighted how limited our understanding of protective immunity in humans is and the
dangers of extrapolating findings in animal models to clinical
practice [128]. Our understanding of how to use a vaccine to
safely prevent reactivation of an established latent infection is
even more limited, as there are fewer appropriate animal
models and there is always concern about the induction of
Koch reactions, where rapid induction of vigorous immune
response in persons with an asymptomatic subclinical infection may result in immunopathology and symptomatic
deterioration. It is critical that we develop a better understanding of what the immunological correlates of protection
from disease are in humans to inform vaccine design. One
approach may be in the careful evaluation of the heterogeneous effects of vaccination either within a single trial or
between trials performed in different populations and the
correlation with clinical outcomes. Another approach would
be to try and better understand the protective effects of natural infection especially the immunological basis for this (see
above), or to characterize the immunological differences in
response to infection between older children and young

10

Phil. Trans. R. Soc. B 369: 20130437

(c) Drugs

are a major factor when treating otherwise asymptomatic persons (as demonstrated by pyrazinamide), especially while we
are unable to more precisely define who will derive greatest
benefit from preventive therapy.
A further challenge is how to best evaluate novel LTBI regimens. Currently, the only endpoint for clinical trials is the
absence of disease and hence in order to demonstrate clinical
efficacy studies they will need large numbers and prolonged
follow-up. A surrogate marker of clinical response (analogous
to 14-day early bactericidal activity or two-month culture conversion for active TB) may allow for more rapid evaluation of
different regimens to select which should go forward to
larger clinical studies. Peripheral blood biomarkers that signify
treatment success for LTBI would be very useful in this regard.
An alternative approach that is being developed similar to
oncology studies is the use of PET/CT imaging to evaluate
response of therapy. 18F-Fluorodeoxyglucose (FDG) is the
most widely used tracer, is a non-specific marker of metabolic
activity and is taken up avidly by activated neutrophils and
macrophages [121]. Sites of active TB even in the absence of
symptoms accumulate FDG avidly and a number of studies
have demonstrated that uptake is markedly reduced following
TB treatment [122125]. Tracers that are more specific for Mtb
would be a great advance and these are currently in early
stages of development.

rstb.royalsocietypublishing.org

or in placebo arms of intervention studies. Another approach

would be to identify biomarkers for people at the transition
of latent to active disease with subclinical and minimally
active pathology or evidence of immunopathology (e.g. fibrotic scaring) and then validate these markers prospectively.
The most useful predictive biomarkers will most likely be
mycobacterial products or markers of host response identified within blood or urine or through skin testing. Omics
approaches are useful as exploratory tools to identify key
components of diagnostic tests; transcriptomic approaches
have been particularly successful at differentiating the
extremes of active and latent TB [113,114], but these signatures may not be predictive of active TB if the signatures
relate to response to disease process itself. More predictive
tests would certainly be a huge advance allowing treatment
decisions to be based on a biological as well as epidemiological markers of risk, but it is sobering to note how slow
progress in the cancer field has been in developing markers
of early detection into diagnostic tests. Such biomarkers
will be challenging to develop, requiring a large amount of
support to move down the pipeline from concept through
to development, validation and implementation; but unlike
vaccine, drug and diagnostic tests for active TB, no such
pipeline exists for predictive tests for latent TB.

A coordinated strategy will be required to effectively tackle

the reservoir of latent infection. Improved data are needed
to more accurately estimate the scale of the problem and

References
1.

Gagneux S. 2012 Host pathogen coevolution in

human tuberculosis. Phil. Trans. R. Soc. B 367,
850859. (doi:10.1098/rstb.2011.0316)
2. Comas I et al. 2013 Out-of-Africa migration and
Neolithic coexpansion of Mycobacterium tuberculosis
with modern humans. Nat. Genet. 45, 1176 1182.
(doi:10.1038/ng.2744)
3. Wirth T et al. 2008 Origin, spread and demography
of the Mycobacterium tuberculosis complex. PLoS
Pathog. 4, e1000160. (doi:10.1371/journal.ppat.
1000160)
4. Blaser MJ, Kirschner D. 2007 The equilibria that
allow bacterial persistence in human hosts. Nature
449, 843849. (doi:10.1038/nature06198)
5. Hershberg R et al. 2008 High functional diversity in
Mycobacterium tuberculosis driven by genetic drift
and human demography. PLoS Biol. 6, e311.
(doi:10.1371/journal.pbio.0060311)
6. de Jong BC et al. 2008 Progression to active
tuberculosis, but not transmission, varies by
Mycobacterium tuberculosis lineage in The Gambia.
J. Infect. Dis. 198, 10371043. (doi:10.1086/
591504)
7. WHO. 2011 Global tuberculosis control 2011.
Geneva, Switzerland: World Health Organization.
8. WHO. 2006 The global plan to stop TB 20062015:
Stop TB partnership. Geneva, Switzerland: World
Health Organization.
9. WHO. 2013 Global tuberculosis report 2013. Geneva,
Switzerland: World Health Organization.
10. Dye C, Glaziou P, Floyd K, Raviglione M. 2013
Prospects for tuberculosis elimination. Annu. Rev.

11.

12.

13.

14.
15.

16.

17.

18.

19.

Public Health 34, 271 286. (doi:10.1146/annurevpublhealth-031912-114431)

Dye C, Hosseini M, Watt C. 2007 Did we reach the
2005 targets for tuberculosis control? Bull. World
Health Organ. 85, 364 369. (doi:10.2471/BLT.06.
037580)
WHO. 2013 Countdown to 2015: global tuberculosis
report 2013 supplement. Geneva, Switzerland:
World Health Organization.
Abu-Raddad LJ, Sabatelli L, Achterberg JT,
Sugimoto JD, Longini Jr IM, Dye C, Halloran ME.
2009 Epidemiological benefits of more-effective
tuberculosis vaccines, drugs, and diagnostics. Proc.
Natl Acad. Sci. USA 106, 13 98013 985. (doi:10.
1073/pnas.0901720106)
Opie EL, McPhedran FM. 1926 The contagion of
tuberculosis. Am. Rev. Tuberc. 14, 347 385.
Yik Wang C. 1916 An experimental study of latent
tuberculosis. Lancet 188, 417 419. (doi:10.1016/
S0140-6736(00)58936-3)
Newsholme A. 1908 The history of phthisis. In The
prevention of tuberculosis, 2nd edn, pp. 35 42.
London, UK: Methuen and Co.
Koch R. 1890 A further communication on a remedy
for tuberculosis. Br. Med. J. 2, 11931199. (doi:10.
1136/bmj.2.1560.1193)
Rieder HL. 1999 Epidemiologic basis of tuberculosis
control. Paris, France: International Union Against
Tuberculosis and Lung Disease.
Mills CC, OGrady F, Riley RL. 1960 Tuberculin
conversion in the naturally infected guinea pig.
Bull. Johns Hopkins Hosp. 106, 36 45.

20. Dharmadhikari AS et al. 2011 Natural infection of

guinea pigs exposed to patients with highly drugresistant tuberculosis. Tuberculosis 91, 329 338.
(doi:10.1016/j.tube.2011.03.002)
21. Escombe AR et al. 2007 The detection of airborne
transmission of tuberculosis from HIV-infected
patients, using an in vivo air sampling model. Clin.
Infect. Dis. 44, 13491357. (doi:10.1086/515397)
22. Ferebee SH. 1970 Controlled chemoprophylaxis trials
in tuberculosis: a general review. Bibl. Tuberc. 26,
28 106.
23. Diel R, Loddenkemper R, Nienhaus A. 2012
Predictive value of interferon-gamma release assays
and tuberculin skin testing for progression from
latent TB infection to disease state: a meta-analysis.
Chest 142, 63 75. (doi:10.1378/chest.11-3157)
24. Rangaka MX, Wilkinson KA, Glynn JR, Ling D,
Menzies D, Mwansa-Kambafwile J, Fielding K,
Wilkinson RJ, Pai M. 2011 Predictive value of
interferon-gamma release assays for incident active
tuberculosis: a systematic review and meta-analysis.
Lancet Infect. Dis. 12, 45 55. (doi:10.1016/S14733099(11)70210-9)
25. Smieja MJ, Marchetti CA, Cook DJ, Smaill FM. 2000
Isoniazid for preventing tuberculosis in non-HIV
infected persons. Cochrane Database Syst. Rev.
2000, CD001363. (doi:10.1002/14651858.
CD001363)
26. Dye C, Scheele S, Dolin P, Pathania V, Raviglione MC.
1999 Consensus statement. Global burden of
tuberculosis: estimated incidence, prevalence, and
mortality by country. WHO global surveillance and

11

Phil. Trans. R. Soc. B 369: 20130437

7. Concluding remarks

quantify the number of new infections occurring each year,

and a redoubling of effort will be required to reduce this as
far as possible by implementing currently recommended
interventions. However, in order for widespread treatment
of latent TB to be acceptable to the public, healthcare providers and policy makers, major advances on the currently
available diagnostic and interventional tools will be required.
Progress in identifying who is most likely to reactivate and
how this occurs will assist the development of more predictive diagnostic tests allowing interventions to be focused on
those that will benefit most. The development of drugs that
effectively target and rapidly sterilize the subset of persistent
bacilli should allow for significant reductions in the duration
of preventive treatment. Both of these should improve acceptability of more widespread treatment of latent infection. In
addition, greater understanding of who is protected from
reinfection and how this occurs would provide key pieces
of knowledge to facilitate progress with development of
effective vaccines and immunomodulatory agents that
could have a major impact.
The aim to eliminate TB by 2050 is a bold one and the
development of the post-2015 TB strategy and targets provides an opportunity to identify the critical gaps in our
knowledge and to focus the scientific community, policy
makers, advocates and funding agencies on achieving this
challenging goal.

rstb.royalsocietypublishing.org

adults, who have the best and worse outcomes following

infection, respectively (see above). In particular, if we are to
develop vaccines that prevent reactivation or successfully
eliminate latent infection, we need to better understand
the immunological mechanisms that precipitate reactivation
and control in those with unstable latent infection, which
may require refinement of existing animal models to more
accurately reflect the natural history of TB in humans.
An alternative and innovative approach would be to combine immunomodulation with anti-tuberculous treatment as a
method to shorten therapy. This immunomodulation could
take the form of either a vaccine or a drug. RUTI is a novel vaccine comprising heat inactivated, liposomed fragments of Mtb
grown under different conditions of stress designed to
be administered after one month of chemotherapy of LTBI
and facilitate immune clearance of persisting bacilli [129].
The vaccine is currently in phase 2 studies in HIV-infected
and -uninfected persons with LTBI (http://clinicaltrials.gov/
show/NCT01136161)

28.

29.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.
56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

critical review of literature from the pre-chemotherapy

era. Int. J. Tuberc. Lung Dis. 8, 392402.
Goodwin RA, Des Prez RM. 1983 Apical localization
of pulmonary tuberculosis, chronic pulmonary
histoplasmosis, and progressive massive fibrosis of
the lung. Chest 83, 801 805. (doi:10.1378/chest.
83.5.801)
Balasubramanian V, Wiegeshaus EH, Taylor BT,
Smith DW. 1994 Pathogenesis of tuberculosis:
pathway to apical localization. Tuberc. Lung Dis. 75,
168178. (doi:10.1016/0962-8479(94)90002-7)
Rich A. 1951 The factors responsible for the
characteristics of tuberculous lesions and symptoms.
In The pathogenesis of tuberculosis, 2nd edn (ed.
CC Thomas), pp. 704778. Baltimore, MD:
Springfield.
Vynnycky E, Fine PE. 1997 The natural history of
tuberculosis: the implications of age-dependent
risks of disease and the role of reinfection.
Epidemiol. Infect. 119, 183201. (doi:10.1017/
S0950268897007917)
Lillebaek T, Dirksen A, Baess I, Strunge B, Thomsen
VO, Andersen AB. 2002 Molecular evidence of
endogenous reactivation of Mycobacterium
tuberculosis after 33 years of latent infection.
J. Infect. Dis. 185, 401 404. (doi:10.1086/338342)
Borgdorff MW, Sebek M, Geskus RB, Kremer K,
Kalisvaart N, van Soolingen D. 2011 The incubation
period distribution of tuberculosis estimated with a
molecular epidemiological approach. Int. J. Epidemiol.
40, 964970. (doi:10.1093/ije/dyr058)
Pedrazzoli D, Anderson D, Lalor M, Davidson J,
Abubakar I, Thomas L. 2013 Tuberculosis in the UK:
annual report on tuberculosis surveillance in the UK.
London, UK: Tuberculosis Section, Centre for
Infectious Disease Surveillance and Control, Public
Health England.
McCarthy OR. 1984 Asian immigrant tuberculosis:
the effect of visiting Asia. Br. J. Dis. Chest 78,
248253. (doi:10.1016/0007-0971(84)90136-0)
Wiker HG, Mustafa T, Bjune GA, Harboe M. 2010
Evidence for waning of latency in a cohort study of
tuberculosis. BMC Infect. Dis. 10, 37. (doi:10.1186/
1471-2334-10-37)
Steinbruck P, Dankova D, Edwards LB, Doster B,
Livesay VT. 1972 Tuberculosis risk in persons with
fibrotic X-ray lesions. Bull. Int. Union Tuberc. 47,
135159.
Vaara J, Kokkola K. 1975 Inactive pulmonary lesions:
a potent risk factor of tuberculosis. Ann. Clin. Res. 7,
331333.
Stead WW. 1967 Pathogenesis of the sporadic case
of tuberculosis. N. Engl. J. Med. 277, 10081012.
(doi:10.1056/NEJM196711092771906)
Oni T et al. 2011 High prevalence of subclinical
tuberculosis in HIV-1-infected persons without
advanced immunodeficiency: implications for TB
screening. Thorax 66, 669673. (doi:10.1136/thx.
2011.160168)
Lin PL et al. 2010 Tumor necrosis factor
neutralization results in disseminated disease in
acute and latent Mycobacterium tuberculosis
infection with normal granuloma structure in a

12

Phil. Trans. R. Soc. B 369: 20130437

30.

41.

(Callithrix jacchus). Infect. Immun. 81, 29092919.

(doi:10.1128/IAI.00632-13)
Dowdy DW, Dye C, Cohen T. 2013 Data needs for
evidence-based decisions: a tuberculosis modelers
wish list. Int. J. Tuberc. Lung Dis. 17, 866877.
(doi:10.5588/ijtld.12.0573)
Comas I, Chakravartti J, Small PM, Galagan J,
Niemann S, Kremer K, Ernst JD, Gagneux S. 2010
Human T cell epitopes of Mycobacterium
tuberculosis are evolutionarily hyperconserved. Nat.
Genet. 42, 498 503. (doi:10.1038/ng.590)
Wells WF. 1955 Airbourne contagion and air
hygiene: an ecological study of droplet infections, 1st
edn. Boston, MA: Harvard University Press.
Fox GJ, Barry SE, Britton WJ, Marks GB. 2012
Contact investigation for tuberculosis: a systematic
review and meta-analysis. Eur. Respir. J. 41,
140 156. (doi:10.1183/09031936.00070812)
Medlar EM. 1948 The pathogenesis of minimal
pulmonary tuberculosis; a study of 1,225 necropsies
in cases of sudden and unexpected death. Am. Rev.
Tuberc. 58, 583611.
Canetti G. 1955 Histology of pulmonary
tuberculosis: a review. In The tubercle bacillus in the
pulmonary lesion of man; histobacteriology and its
bearing on the therapy of pulmonary tuberculosis,
1st edn, pp. 7 28. New York, NY: Springer.
Saunders BM, Orme IM, Basaraba RJ. 2008
Immunopathology of tuberculosis. In Handbook of
tuberculosis: immunology and cell biology (eds
SH Kaufmann, WJ Britton), pp. 245277.
Weinheim, Germany: Wiley-VCH.
Allen EA. 1995 Tuberculosis and other mycobacterial
infections of the lung. In Pathology of the lung, 2nd
edn (eds WM Thurlbeck, AM Churg), pp. 230265.
New York, NY: Thieme.
Chao MC, Rubin EJ. 2010 Letting sleeping dogs lie:
does dormancy play a role in tuberculosis? Annu.
Rev. Microbiol. 64, 293311. (doi:10.1146/annurev.
micro.112408.134043)
Rustad TR, Harrell MI, Liao R, Sherman DR. 2008
The enduring hypoxic response of Mycobacterium
tuberculosis. PLoS ONE 3, e1502. (doi:10.1371/
journal.pone.0001502)
Kana BD, Mizrahi V. 2010 Resuscitation-promoting
factors as lytic enzymes for bacterial growth and
signaling. FEMS Immunol. Med. Microbiol. 58,
39 50. (doi:10.1111/j.1574-695X.2009.00606.x)
Wallgren A. 1948 The time-table of tuberculosis.
Tubercle 29, 245 251. (doi:10.1016/S00413879(48)80033-4)
Poulsen A. 1950 Some clinical features of
tuberculosis. 1. Incubation period. Acta Tuberc.
Scand. 24, 311 346.
Myers JA. 1965 The natural history of tuberculosis in
the human body; forty-five years of observation.
JAMA 194, 10861092. (doi:10.1001/jama.1965.
03090230054013)
Poulsen A. 1957 Some clinical features of
tuberculosis. Acta Tuberc. Scand. 33, 37 92.
Marais BJ, Gie RP, Schaaf HS, Hesseling AC, Obihara CC,
Starke JJ, Enarson DA, Donald PR, Beyers N. 2004 The
natural history of childhood intra-thoracic tuberculosis: a

rstb.royalsocietypublishing.org

27.

monitoring project. JAMA 282, 677686. (doi:10.

1001/jama.282.7.677)
Dye C, Bassili A et al. 2008 Measuring tuberculosis
burden, trends, and the impact of control
programmes. Lancet Infect. Dis. 8, 233 243.
(doi:10.1016/S1473-3099(07)70291-8)
Styblo K. 1985 The relationship between the risk of
tuberculous infection and the risk of developing
infectious tuberculosis. Bull. Int. Union Tuberc. 60,
117119.
Rieder HL, Chadha VK, Nagelkerke NJ, van Leth F,
van der Werf MJ. 2011 Guidelines for conducting
tuberculin skin test surveys in high-prevalence
countries. Int. J. Tuberc. Lung Dis. 15(Suppl. 1),
S1 S25.
Rieder HL. 1995 Methodological issues in the
estimation of the tuberculosis problem from
tuberculin surveys. Tuberc. Lung Dis. 76, 114121.
(doi:10.1016/0962-8479(95)90552-9)
Pai M, Dendukuri N, Wang L, Joshi R, Kalantri S,
Rieder HL. 2008 Improving the estimation of
tuberculosis infection prevalence using T-cell-based
assay and mixture models. Int. J. Tuberc. Lung Dis.
12, 895902.
Aggerbeck H, Giemza R, Joshi P, Tingskov PN, Hoff
ST, Boyle J, Andersen P, Lewis DJ. 2013 Randomised
clinical trial investigating the specificity of a novel
skin test (C-Tb) for diagnosis of M. tuberculosis
infection. PLoS ONE 8, e64215. (doi:10.1371/journal.
pone.0064215)
Dye C. 2008 Breaking a law: tuberculosis disobeys
Styblos rule. Bull. World Health Organ. 86, 4.
(doi:10.2471/BLT.07.049510)
van Leth F, van der Werf MJ, Borgdorff MW. 2008
Prevalence of tuberculous infection and incidence of
tuberculosis: a re-assessment of the Styblo rule.
Bull. World Health Organ. 86, 20 26. (doi:10.2471/
BLT.06.037804)
Jones-Lopez EC et al. 2013 Cough aerosols of
Mycobacterium tuberculosis predict new infection: a
household contact study. Am. J. Respir. Crit. Care
Med. 187, 1007 1015. (doi:10.1164/rccm.2012081422OC)
Escombe AR et al. 2008 The infectiousness of
tuberculosis patients coinfected with HIV. PLoS Med.
5, e188. (doi:10.1371/journal.pmed.0050188)
Gagneux S. 2013 Genetic diversity in Mycobacterium
tuberculosis. Curr. Top. Microbiol. Immunol. 374,
1 25. (doi:10.1007/82_2013_329)
Portevin D, Gagneux S, Comas I, Young D. 2011
Human macrophage responses to clinical isolates
from the Mycobacterium tuberculosis complex
discriminate between ancient and modern lineages.
PLoS Pathog. 7, e1001307. (doi:10.1371/journal.
ppat.1001307)
Yang C, Luo T, Sun G, Qiao K, DeRiemer K, Mei J,
Gao Q. 2012 Mycobacterium tuberculosis Beijing
strains favor transmission but not drug resistance in
China. Clin. Infect. Dis. 55, 1179 1187. (doi:10.
1093/cid/cis670)
Via LE et al. 2013 Differential virulence and disease
progression following Mycobacterium tuberculosis
complex infection of the common marmoset

72.

74.

75.

76.

77.

78.

79.

80.

81.

82.

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.
94.

95.

96.

97. Soetens LC, Boshuizen HC, Korthals Altes H. 2013

Contribution of seasonality in transmission of
Mycobacterium tuberculosis to seasonality in
tuberculosis disease: a simulation study. Am. J.
Epidemiol. 178, 12811288. (doi:10.1093/aje/kwt114)
98. Top R, Boshuizen H, Dekkers A, Korthals Altes H.
2013 Similar seasonal peak in clustered and unique
extra-pulmonary tuberculosis notifications: winter
crowding hypothesis ruled out? Int. J. Tuberc. Lung
Dis. 17, 1466 1471. (doi:10.5588/ijtld.13.0226)
99. Martineau AR. 2012 Old wine in new bottles:
vitamin D in the treatment and prevention of
tuberculosis. Proc. Nutr. Soc. 71, 84 89. (doi:10.
1017/S0029665111003326)
100. Nnoaham KE, Clarke A. 2008 Low serum vitamin D
levels and tuberculosis: a systematic review and
meta-analysis. Int. J. Epidemiol. 37, 113 119.
(doi:10.1093/ije/dym247)
101. Wilkinson RJ, Llewelyn M, Toossi Z, Patel P, Pasvol
G, Lalvani A, Wright D, Latif M, Davidson RN. 2000
Influence of vitamin D deficiency and vitamin D
receptor polymorphisms on tuberculosis among
Gujarati Asians in west London: a case-control
study. Lancet 355, 618621. (doi:10.1016/S01406736(99)02301-6)
102. Martineau AR et al. 2011 Reciprocal seasonal
variation in vitamin D status and tuberculosis
notifications in Cape Town, South Africa. Proc. Natl
Acad. Sci. USA 108, 19 013 19 017. (doi:10.1073/
pnas.1111825108)
103. Korthals Altes H, Kremer K, Erkens C, van Soolingen D,
Wallinga J. 2012 Tuberculosis seasonality in the
Netherlands differs between natives and non-natives: a
role for vitamin D deficiency? Int. J. Tuberc. Lung Dis. 16,
639644. (doi:10.5588/ijtld.11.0680)
104. Douglas AS, Ali S, Bakhshi SS. 1998 Does vitamin D
deficiency account for ethnic differences in
tuberculosis seasonality in the UK? Ethn. Health 3,
247253. (doi:10.1080/13557858.1998.9961867)
105. Teles RM et al. 2013 Type I interferon suppresses
type II interferon-triggered human antimycobacterial responses. Science 339, 14481453.
(doi:10.1126/science.1233665)
106. Redford PS, Mayer-Barber KD, McNab FW,
Stavropoulos E, Wack A, Sher A, OGarra A. 2013
Influenza A virus impairs control of Mycobacterium
tuberculosis coinfection through a type i interferon
receptor-dependent pathway. J. Infect. Dis. 209,
270274. (doi:10.1093/infdis/jit424)
107. Oei W, Nishiura H. 2012 The relationship between
tuberculosis and influenza death during the
influenza (H1N1) pandemic from 1918 19.
Comput. Math. Methods Med. 2012, 124861.
(doi:10.1155/2012/124861)
108. Osorio NS, Rodrigues F, Gagneux S, Pedrosa J, PintoCarbo M, Castro AG, Young D, Comas I, Saraiva M.
2013 Evidence for diversifying selection in a set of
Mycobacterium tuberculosis genes in response to
antibiotic- and nonantibiotic-related pressure. Mol.
Biol. Evol. 30, 13261336. (doi:10.1093/molbev/
mst038)
109. van Zyl-Smit RN, Zwerling A, Dheda K, Pai M. 2009
Within-subject variability of interferon-g assay

13

Phil. Trans. R. Soc. B 369: 20130437

73.

83.

Drug Discov. Today 17, 514521. (doi:10.1016/j.

drudis.2011.12.013)
Robertson BD et al. 2012 Detection and treatment
of subclinical tuberculosis. Tuberculosis 92,
447 452. (doi:10.1016/j.tube.2012.06.004)
Lin PL, Flynn JL. 2010 Understanding latent
tuberculosis: a moving target. J. Immunol. 185,
15 22. (doi:10.4049/jimmunol.0903856)
Tobin DM et al. 2010 The lta4h locus modulates
susceptibility to mycobacterial infection in zebrafish
and humans. Cell 140, 717730. (doi:10.1016/j.
cell.2010.02.013)
Tobin DM et al. 2012 Host genotype-specific
therapies can optimize the inflammatory response
to mycobacterial infections. Cell 148, 434446.
(doi:10.1016/j.cell.2011.12.023)
Lalvani A, Behr MA, Sridhar S. 2012 Innate
immunity to TB: a druggable balancing act. Cell
148, 389 391. (doi:10.1016/j.cell.2012.01.026)
Donald PR, Marais BJ, Barry 3rd CE. 2010 Age and
the epidemiology and pathogenesis of tuberculosis.
Lancet 375, 18521854. (doi:10.1016/S01406736(10)60580-6)
Hernandez-Pando R, de la Luz Streber M, Orozco H,
Arriaga K, Pavon L, Marti O, Lightman SL, Rook GA.
1998 Emergent immunoregulatory properties of
combined glucocorticoid and anti-glucocorticoid
steroids in a model of tuberculosis. QJM 91,
755 766. (doi:10.1093/qjmed/91.11.755)
Angerami M, Suarez G, Pascutti MF, Salomon H,
Bottasso O, Quiroga MF. 2013 Modulation of the
phenotype and function of Mycobacterium
tuberculosis-stimulated dendritic cells by adrenal
steroids. Int. Immunol. 25, 405 411. (doi:10.1093/
intimm/dxt004)
Naranbhai V, Hill AV, Abdool Karim SS, Naidoo K,
Abdool Karim Q, Warimwe GM, McShane H, Fletcher
H. 2013 Ratio of monocytes to lymphocytes in
peripheral blood identifies adults at risk of incident
tuberculosis among HIV-infected adults initiating
antiretroviral therapy. J. Infect. Dis. 209, 500 509.
(doi:10.1093/infdis/jit494)
Coussens AK et al. 2012 Vitamin D accelerates
resolution of inflammatory responses during
tuberculosis treatment. Proc. Natl Acad. Sci.
USA 109, 15 449 15 454. (doi:10.1073/pnas.
1200072109)
Fares A. 2011 Seasonality of tuberculosis. J. Glob.
Infect. Dis. 3, 4655. (doi:10.4103/0974-777X.77296)
Koh GC, Hawthorne G, Turner AM, Kunst H, Dedicoat
M. 2013 Tuberculosis incidence correlates with
sunshine: an ecological 28-year time series study.
PLoS ONE 8, e57752. (doi:10.1371/journal.pone.
0057752)
Willis MD, Winston CA, Heilig CM, Cain KP, Walter
ND, Mac Kenzie WR. 2012 Seasonality of
tuberculosis in the United States, 19932008.
Clin. Infect. Dis. 54, 15531560. (doi:10.1093/
cid/cis235)
Maclachlan JH, Lavender CJ, Cowie BC. 2012 Effect
of latitude on seasonality of tuberculosis, Australia,
2002 2011. Emerg. Infect. Dis. 18, 18791881.
(doi:10.3201/eid1811.120456)

rstb.royalsocietypublishing.org

71.

cynomolgus macaque model. Arthritis Rheum. 62,

340350. (doi:10.1002/art.27271)
Wallis RS, Broder M, Wong J, Lee A, Hoq L. 2005
Reactivation of latent granulomatous infections
by infliximab. Clin. Infect. Dis. 41(Suppl. 3),
S194 S198. (doi:10.1086/429996)
Gomez-Reino JJ, Carmona L, Angel Descalzo M. 2007
Risk of tuberculosis in patients treated with tumor
necrosis factor antagonists due to incomplete
prevention of reactivation of latent infection. Arthritis
Rheum. 57, 756761. (doi:10.1002/art.22768)
Sonnenberg P, Glynn JR, Fielding K, Murray J,
Godfrey-Faussett P, Shearer S. 2005 How soon after
infection with HIV does the risk of tuberculosis start
to increase? A retrospective cohort study in South
African gold miners. J. Infect. Dis. 191, 150 158.
(doi:10.1086/426827)
Martinez-Pino I, Sambeat MA, Lacalle-Remigio JR,
Domingo P. 2013 Incidence of tuberculosis in HIVinfected patients in Spain: the impact of treatment
for LTBI. Int. J. Tuberc. Lung Dis. 17, 1545 1551.
(doi:10.5588/ijtld.13.0070)
Elzi L, Schlegel M, Weber R, Hirschel B, Cavassini M,
Schmid P, Bernasconi E, Rickenbach M, Furrer H.
2007 Reducing tuberculosis incidence by tuberculin
skin testing, preventive treatment, and antiretroviral
therapy in an area of low tuberculosis transmission.
Clin. Infect. Dis. 44, 94 102. (doi:10.1086/510080)
Andrews JR, Noubary F, Walensky RP, Cerda R,
Losina E, Horsburgh CR. 2012 Risk of progression to
active tuberculosis following reinfection with
Mycobacterium tuberculosis. Clin. Infect. Dis. 54,
784791. (doi:10.1093/cid/cir951)
Sutherland I, Svandova E, Radhakrishna S. 1982 The
development of clinical tuberculosis following
infection with tubercle bacilli. 1. A theoretical
model for the development of clinical tuberculosis
following infection, linking from data on the risk of
tuberculous infection and the incidence of clinical
tuberculosis in the Netherlands. Tubercle 63,
255268. (doi:10.1016/S0041-3879(82)80013-5)
Brooks-Pollock E, Becerra MC, Goldstein E, Cohen T,
Murray MB. 2011 Epidemiologic inference from the
distribution of tuberculosis cases in households in
Lima, Peru. J. Infect. Dis. 203, 1582 1589. (doi:10.
1093/infdis/jir162)
Lurie MB. 1933 A correlation between the
histological changes and the fate of living
tubercle bacilli in the organs of reinfected
rabbits. J. Exp. Med. 57, 181 201. (doi:10.1084/
jem.57.2.181)
Young DB, Gideon HP, Wilkinson RJ. 2009
Eliminating latent tuberculosis. Trends Microbiol. 17,
183188. (doi:10.1016/j.tim.2009.02.005)
Barry 3rd CE, Boshoff HI, Dartois V, Dick T,
Ehrt S, Flynn J, Schnappinger D, Wilkinson RJ,
Young D. 2009 The spectrum of latent tuberculosis:
rethinking the biology and intervention strategies.
Nat. Rev. Microbiol. 7, 845855. (doi:10.1038/
nrmicro2236)
Esmail H, Barry 3rd CE, Wilkinson RJ. 2012
Understanding latent tuberculosis: the key to
improved diagnostic and novel treatment strategies.

112.

113.

114.

115.

123.

124.

125.

126.

127.

128.

129.

Assist. Tomogr. 35, 387 393. (doi:10.1097/RCT.

0b013e318219f810)
Kim IJ, Lee JS, Kim SJ, Kim YK, Jeong YJ, Jun S,
Nam HY, Kim JS. 2008 Double-phase 18F-FDG PETCT for determination of pulmonary tuberculoma
activity. Eur. J. Nucl. Med. Mol. Imaging 35,
808814. (doi:10.1007/s00259-007-0585-0)
Goo JM, Im JG, Do KH, Yeo JS, Seo JB, Kim HY,
Chung JK. 2000 Pulmonary tuberculoma evaluated
by means of FDG PET: findings in 10 cases.
Radiology 216, 117 121. (doi:10.1148/radiology.
216.1.r00jl19117)
Sathekge M, Maes A, Kgomo M, Stoltz A, Van de
Wiele C. 2011 Use of 18F-FDG PET to predict
response to first-line tuberculostatics in HIVassociated tuberculosis. J. Nucl. Med. 52, 880 885.
(doi:10.2967/jnumed.110.083709)
Aagaard C et al. 2011 A multistage tuberculosis
vaccine that confers efficient protection before and
after exposure. Nat. Med. 17, 189194. (doi:10.
1038/nm.2285)
Tameris MD et al. 2013 Safety and efficacy of
MVA85A, a new tuberculosis vaccine, in infants
previously vaccinated with BCG: a randomised,
placebo-controlled phase 2b trial. Lancet
381, 1021 1028. (doi:10.1016/S0140-6736(13)
60177-4)
Bishai W, Sullivan Z, Bloom BR, Andersen P. 2013
Bettering BCG: a tough task for a TB vaccine? Nat.
Med. 19, 410 411. (doi:10.1038/nm.3153)
Vilaplana C, Montane E, Pinto S, Barriocanal AM,
Domenech G, Torres F, Cardona PJ, Costa J. 2010
Double-blind, randomized, placebo-controlled phase
I clinical trial of the therapeutical antituberculous
vaccine RUTI. Vaccine 28, 11061116. (doi:10.
1016/j.vaccine.2009.09.134)

14

Phil. Trans. R. Soc. B 369: 20130437

111.

116. Comstock GW. 1999 How much isoniazid is needed

for prevention of tuberculosis among
immunocompetent adults? Int. J. Tuberc. Lung Dis.
3, 847 850.
117. Sharma SK, Sharma A, Kadhiravan T, Tharyan P.
2013 Rifamycins (rifampicin, rifabutin and
rifapentine) compared to isoniazid for preventing
tuberculosis in HIV-negative people at risk of active
TB. Cochrane Database Syst. Rev. 7, CD007545.
(doi:10.1002/14651858.CD007545.pub2)
118. Grosset JH, Singer TG, Bishai WR. 2012 New drugs
for the treatment of tuberculosis: hope and reality.
Int. J. Tuberc. Lung Dis. 16, 10051014. (doi:10.
5588/ijtld.12.0277)
119. Williams K, Minkowski A, Amoabeng O, Peloquin
CA, Taylor D, Andries K, Wallis RS, Mdluli KE,
Nuermberger EL. 2012 Sterilizing activities of novel
combinations lacking first- and second-line drugs in
a murine model of tuberculosis. Antimicrob. Agents
Chemother. 56, 31143120. (doi:10.1128/AAC.
00384-12)
120. Shepardson D et al. 2013 Cost-effectiveness of a 12dose regimen for treating latent tuberculous
infection in the United States. Int. J. Tuberc. Lung
Dis. 17, 15311537. (doi:10.5588/ijtld.13.0423)
121. Kaim AH, Weber B, Kurrer MO, Gottschalk J, Von
Schulthess GK, Buck A. 2002 Autoradiographic
quantification of 18F-FDG uptake in experimental
soft-tissue abscesses in rats. Radiology 223, 446
451. (doi:10.1148/radiol.2232010914)
122. Treglia G, Taralli S, Calcagni ML, Maggi F, Giordano
A, Bonomo L. 2011 Is there a role for fluorine 18
fluorodeoxyglucose-positron emission tomography
and positron emission tomography/computed
tomography in evaluating patients with
mycobacteriosis? A systematic review. J. Comput.

rstb.royalsocietypublishing.org

110.

results for tuberculosis and boosting effect of

tuberculin skin testing: a systematic review.
PLoS ONE 4, e8517. (doi:10.1371/journal.
pone.0008517)
Metcalfe JZ, Everett CK, Steingart KR, Cattamanchi A,
Huang L, Hopewell PC, Pai M. 2011 Interferon-gamma
release assays for active pulmonary tuberculosis
diagnosis in adults in low- and middle-income
countries: systematic review and meta-analysis.
J. Infect. Dis. 204(Suppl. 4), S1120S1129. (doi:10.
1093/infdis/jir410)
Pai M, Zwerling A, Menzies D. 2008 Systematic
review: T-cell-based assays for the diagnosis of
latent tuberculosis infection: an update. Ann. Intern.
Med. 149, 177 184. (doi:10.7326/0003-4819-1493-200808050-00241)
Menzies D, Pai M, Comstock G. 2007 Meta-analysis:
new tests for the diagnosis of latent tuberculosis
infection: areas of uncertainty and
recommendations for research. Ann. Intern. Med.
146, 340354. (doi:10.7326/0003-4819-146-5200703060-00006)
Berry MP et al. 2010 An interferon-inducible
neutrophil-driven blood transcriptional signature in
human tuberculosis. Nature 466, 973 977. (doi:10.
1038/nature09247)
Kaforou M et al. 2013 Detection of tuberculosis in
HIV-infected and -uninfected African adults using
whole blood RNA expression signatures: a casecontrol study. PLoS Med. 10, e1001538. (doi:10.
1371/journal.pmed.1001538)
Fox W, Ellard GA, Mitchison DA. 1999 Studies on
the treatment of tuberculosis undertaken by the
British Medical Research Council tuberculosis units,
1946 1986, with relevant subsequent publications.
Int. J. Tuberc. Lung Dis. 3, S231S279.