Viewpoint: Scientific dogmas, paradoxes and mysteries of latent Mycobacterium tuberculosis infection


Corresponding Author Alimuddin Zumla, Department of Infection, University College London Medical School, Windeyer Institute of Medical Sciences, 46 Cleveland Street, London W1T 4JF, UK. Tel./Fax: +44 (0)207 6799311; E-mail:


Worldwide, there are nearly 10 million new cases of active TB and 1.8 million associated deaths every year. WHO estimates that one-third of the world’s population is infected with Mycobacterium tuberculosis (Mtb), forming a huge latent Mtb global reservoir. This renders the prospect of ever eliminating Mtb from the human race almost impossible. Several controversial issues regarding host-pathogen interactions and existing prevention and eradication strategies for latent Mtb infections need to be critically re-examined. In this viewpoint, widely held assumptions on Mtb latency and isoniazid monotherapy and chemoprophylaxis are challenged. We highlight the need for future research to resolve these issues and to develop evidence-based strategies for better understanding of equilibrium and escape of Mtb in the human body, eventually leading to global recommendations for elimination of the latent Mtb state through informed policy and practice. Until such strategies and policies are realized, WHO and TB experts will have to settle for global TB control rather than eradication.


Dans le monde, il ya près de 10 millions de nouveaux cas de tuberculose (TB) active et 1,8 millions de personnes en meurent chaque année, ce qui en fait l’une des plus importantes causes de décès. L’OMS estime qu’un tiers de la population mondiale est infecté par Mycobacterium tuberculosis (Mtb), formant ainsi un énorme réservoir mondial de Mtb. Ceci rend la perspective de l’élimination de Mtb de la race humaine presque impossible. Plusieurs questions controversées concernant les interactions hôte/pathogène et les stratégies de prévention et d’éradication des infections latentes de Mtb doivent être réexaminées de façon critique. Dans ce point de vue, les hypothèses largement répandues sur la latence de Mtb et la monothérapie/chimioprophylaxie à l’isoniazide sont contestées. Nous soulignons le besoin de futures recherches nécessaires pour résoudre ces problèmes et pour élaborer des stratégies fondées sur des preuves pour une meilleure compréhension de l’équilibre et de l’évasion de Mtb dans le corps humain, afin d’aboutir éventuellement à des recommandations globales pour l’élimination de l’état latent de Mtbà travers des politiques et des pratiques éclairées. Jusqu’à ce que ces stratégies et politiques soient réalisées, l’OMS et les experts de la TB devront se contenter du contrôle mondial de la TB plutôt que de son éradication.


A nivel mundial hay casi 10 millones de nuevos casos de TB activa, y 1.8 millones de personas mueren de esta enfermedad cada año, siendo una de las más importantes causas de muerte a nivel global. La OMS calcula que una tercera parte de la población mundial está infectada con Mtb, formando así un inmenso reservorio latente de Mtb. Esto hace que la posibilidad de eliminar algún día al Mtb de la raza humana sea casi imposible. Hay varios puntos controversiales sobre las interacciones huésped-patógeno y las estrategias existentes de prevención y erradicación de las infecciones latentes de Mtb deben ser reexaminadas con una visión crítica. En este artículo, se discuten asunciones ampliamente aceptadas sobre la latencia del Mtb y se plantean desafíos para la monoterapia con isoniazida y la quimioprofilaxis. Subrayamos la necesidad de realizar investigaciones futuras para resolver estos temas y desarrollar estrategias basadas en la evidencia para alcanzar un mayor entendimiento del equilibrio y escape del Mtb en el cuerpo humano, llevando eventualmente a recomendaciones globales para la eliminación del estado latente de Mtb mediante políticas y prácticas informadas. Hasta que se encuentren estas estrategias y políticas, los expertos en TB y de la OMS deberán conformarse con el control de la TB, más que con su erradicación.


Tuberculosis (TB), caused by Mycobacterium tuberculosis (Mtb), is a global health catastrophe. There are nearly 10 million new cases of active TB and 1.8 million people die from it every year (WHO 2010) making it one of the most important causes of death worldwide from an infectious disease. Global tuberculosis control ultimately depends on identification of the two billion people who, according to WHO estimates, are infected with Mycobacterium tuberculosis (Mtb) (Sudre et al. 1992; Corbett et al. 2003). Current strategies for identification and treatment for latent TB are inadequate (Sterling 2008; Parris et al. 1998; Lobue & Menzies 2010). The latest Europe TBNET consensus statement on latent Mtb infection re-emphasises the lack of accurate scientific knowledge on mycobacterial mechanisms of immune evasion, replication, equilibrium and escape (Mack et al. 2009). Furthermore, several controversial issues regarding host-pathogen interactions and existing prevention and eradication strategies for latent Mtb infections need to be critically re-examined, particularly the current scientific basis, dogmas, paradoxes and mysteries of latent Mtb infection.

Definitions: infection, active disease and latency

The probability of development of active clinical tuberculous disease after being infected with Mtb inhaled from infected aerosols from an infectious patient with active TB is very small (Behar et al. 2010). No more than 10% of those infected with Mtb develop symptoms and signs of active disease over a lifetime; the actual figure depends on geographical location, numbers of organisms and strain type, genetic background, immunosuppression including co-infection with HIV, social, environmental and other risk factors (Sterling 2008; Lin & Ottenhoff 2008). The risk of developing active tuberculous disease after initial infection is thought to be highest within the first year (estimated at 1%) and second year (estimated at 0.3%); thereafter the risk diminishes considerably. Thus the majority of immunocompetent individuals (>90% of those infected) either eliminate Mtb or contain it in a ‘latent state’ in which an equilibrium is established between host and pathogen.

‘Latent TB’ is a clinical condition that occurs after an individual is infected with Mtb, the infection is established, and the elicited host immune response holds the Mtb bacilli in a quiescent state, thereby preventing active replication and tissue damage. Thus Mtb bacilli are present in host tissue and yet there are no clinical symptoms or signs of active tuberculous disease. Latency has been proven to exist since Mtb bacilli have been cultured from tissues obtained from healthy individuals who died from traffic accidents and had no pathological evidence of active disease (Opie & Aronson 1927). This state of Mtb latency is assumed to be lifelong (Ehlers 2009). While this dogma persists, long term latency may be related to Mtb strain differences and high human density-adjusted strains. Importantly, re-activation of latent Mtb bacilli can occur at any time in the infected individual’s lifetime. The actual number of people developing active tuberculous disease later in life varies considerably and depends on the waning of effective host immune responses, occurrence of chronic diseases such as diabetes, alcoholic liver disease, malnutrition, immunosuppression due to HIV co-infection and steroids or other immunosuppressive drugs. When active disease occurs in later life, it becomes difficult to ascertain whether it is due to reactivation of latent Mtb bacilli or a new infection with a different Mtb strain.

Mtb latency and size of the global problem

According to WHO estimates, 2 billion people globally are latently infected with Mtb (Sudre et al. 1992). How this figure was calculated and why it has become a dogma in the literature over the past two decades requires explanation since this huge latent Mtb human reservoir would render the prospect of eliminating Mtb from the human race seemingly impossible. Latent TB is diagnosed by the absence of clinical disease (symptoms and signs) and positive intradermal skin test reactivity with purified protein derivative (PPD) of Mtb. The tuberculin skin test (TST) and/or the more recent positive interferon gamma release assays (IGRAs) lack accuracy for the purpose. An accurate study to survey worldwide prevalence of latent TB would be impossible due to logistical reasons. Thus epidemiological models based on sets of complex formulae, using epidemiological indicators of case notification rates, predicted incidence of disease, mortality from TB, annual risk of infection (from tuberculin skin test surveys), rate of change of the annual risk of infection in the past, and the age-specific prevalences of infection, are used by WHO to estimate latent TB prevalence rates and of dual Mtb/HIV co-infection rates for each country, region and continent. In 1990 WHO estimated that one third (1.7 billion people) of the world’s population was infected with Mtb and that 3 million people were co-infected with Mtb and HIV (2.375 million in Africa) (Sudre et al. 1992). Since then the TB and HIV epidemics have progressed, the figures have not reduced significantly (WHO 2010) and the emergence of drug resistant Mtb strains throughout the world adds an another complex dimension to TB control (Migliori et al. 2010).

Mycobacterial escape and equilibrium

Several clinical, immunological and pharmacological issues relating to Mtb latency need to be unravelled before effective progress can be made towards eradicating the Mtb reservoir. Mtb has over thousands of years developed an elaborate survival mechanism in humans, allowing it to remain in a dormant state in the human body after initial infection (Behar et al. 2010). In most immunocompetent individuals, the immune system engages a lifelong battle against latent Mtb infection, either eradicating it or limiting its replication and containing its spread (Parris et al. 1998; Lobue and Menzies 2010; Behar et al. 2010; Lin & Ottenhoff 2008; Opie & Aronson 1927; Ehlers 2009; Henderson et al. 2010). Among the unresolved mysteries of latent Mtb are the nature and anatomical situation of persisting tubercle bacilli (Grange 1992). Various morphological variants of the bacilli such as minute Much’s granules and cell wall-free forms have been postulated but none has been clearly proven (Stanford 1987). Also, while it is often stated that persisting bacilli are sequestered in dense fibrous scars of healed primary lesions, studies based on in situ PCR suggest that they are more widely distributed throughout apparently normal lung tissue (Hernández-Pando et al. 2000). In addition, such DNA is widely distributed in adipose tissue of individuals with latent TB (Neyrolles et al. 2006). Although most cases of post-primary TB occur in the upper regions of the lung, there is no reason to assume that latent bacilli are restricted to this anatomical situation (Balasubramanian et al. 1994).

One generally accepted postulate of latent TB assumes that Mtb bacilli are in a true dormant state, undergoing no replication. Such a state could be a response to tissue factors such as anoxia, but although the DosR regulon enables Mtb to adapt to anaerobic conditions in vitro, and to survive for long periods under such conditions, its relevance in human TB remains uncertain (Honaker et al. 2009). Recent data cast doubt on the assumption of an ‘inactive’ latent state, as there is constant metabolic activity within the Mtb bacilli. More recently, there is increasing evidence to support the hypothesis that molecular chaperones are secreted as intercellular signalling molecules, which in turn can control metabolic activity at, and composition of, the cell wall (Henderson et al. 2010). Resuscitation-promoting factors have also been identified and the role of these in Mtb latency and Mtb re-activation is being investigated (Biketov et al. 2007).

A further bacterial regulatory mechanism dependent on tissue factors favouring latency involves Wag31, a protein homologous to DivIVA proteins present in all Gram positive bacteria and which, by an effect on peptidoglycan synthesis, determines cell morphology and division. In mycobacteria, depletion of Wag31 under conditions conducive to dormancy causes a weakening of the cell wall, initially at one pole but with the eventual formation of circular forms (Kang et al. 2008). This could well be the initiating phase in formation of the postulated cell wall-free, non acid-fast persisting bacilli.

These proposed forms of latency are certainly not mutually exclusive and may coexist, though there may be considerable differences between the nature of bacilli persisting after initial infection with limited tissue damage, and those persisting after therapy for post-primary TB, when there is extensive tissue damage and dense fibrous scarring.

Mycobacterial eradication

TB eradication will require the identification and preventive treatment of high risk individuals who are latently infected with Mtb. However, accurate and direct identification of latent Mtb infection is not possible due to technical deficiencies. Available diagnostic tests used to identify individuals latently infected with Mtb, the in vivo tuberculin skin test (TST) and the ex vivo interferon-gamma release assays (IGRAs), are designed to identify an adaptive immune response against, but not necessarily latent infection with, Mtb. The proportion of individuals who truly remain infected with Mtb after tuberculin skin test or IGRA conversion is unknown. Emerging data show that IGRAs are not useful for distinguishing active disease from latent Mtb infection in high TB and HIV endemic areas. It is also uncertain how long adaptive immune responses towards Mtb antigens persist in the absence of live mycobacteria. Clinical management and public health care policies for preventive chemotherapy against Mtb need improving and will require more sensitive and specific tests which can distinguish between active disease and latent TB (Mack et al. 2009; Marais et al. 2009).

From the viewpoint of epidemiology and disease control, the most important form of Mtb latency is that occurring between primary infection and the emergence of post-primary, usually infectious, pulmonary TB. The classical rationale of preventing emergence of active disease from a latent state involves administering a limited duration course of a single or combination of anti-Mtb drugs. WHO currently recommends isoniazid monotherapy for treatment of latent TB and for secondary prophylaxis after treatment for active TB in those at risk (Sudre et al. 1992; Lobue & Menzies 2010). Isoniazid preventive therapy is recommended for risk groups who are HIV-infected, receiving immunosuppressive therapy, or in close contact with sputum smear positive active TB patients, such as health care workers, prison staff, household contacts and immigrants to TB non-endemic areas who are latently infected.

In high endemic HIV/AIDS and TB areas, particularly sub-Saharan Africa, a large proportion of HIV-infected patients are co-infected with Mtb. WHO recommends isoniazid preventive therapy (IPT) for preventing re-activation of latent Mtb infections progressing to active TB (WHO 1998). However, despite clear guidelines, this policy has not been implemented widely in most TB and HIV/AIDS endemic areas. There is also widespread concern that such preventive therapy may lead to the emergence of initial resistance to isoniazid, the most potent and important first line bactericidal drug crucial to the rapid reduction of the organism load in the early phase of chemotherapy. Furthermore, the issue of the role of isoniazid prophylaxis in patients infected with MDR or XDR strains of Mtb remains unclear.

The paradox of isoniazid monotherapy dogma

Isoniazid is a potent first-line bactericidal drug which can rapidly reduce the Mtb organism load at start of chemotherapy (Mitchison 1985). Despite several decades of research, numerous theories and expert discussion groups, the pharmacological and physiological basis of the use of INH monotherapy for IPT has eluded scientists and TB experts (Vilchèze & Jacobs 2007). For INH to be effective, it requires active division and replication of Mtb bacilli. There is strong evidence from clinical trials that isoniazid monotherapy is highly effective in preventing latent Mtb infection progressing to active TB (WHO 1998). This paradox, of a highly effective impact of INH on dormant, non-replicative metabolic state Mtb (that should render it physiologically resistant to killing by INH), has not yet been explained and continues to baffle scientists.

Revisiting alternative modes of action of isoniazid

Isoniazid is a pro-drug that requires oxidative activation by the mycobacterial catalase-peroxidase enzyme KatG and the most frequent resistance-determining mutations occur in the gene encoding this enzyme (Timmins et al. 2004). The usual target for INH is the enzyme enoyl-acyl carrier protein reductase which is involved in the synthesis of mycolic acid, a major constituent of the mycobacterial cell wall (Argyrou et al. 2006). Some INH resistance-determining mutations occur in the inhA gene that encodes this enzyme, or in its promoter region. If, however, the latent Mtb bacillus is in a cell wall-free form, and therefore lacking mycolic acid as a result of inhibition of Wag31 or of other mechanisms, INH would not be expected to have a bactericidal effect (Kang et al. 2008). This indicates that INH has a quite different mode of action on such latent Mtb bacilli. INH, to be effective in TB latency, requires effective T cell responses. This is based on the observation that INH only offers protection to TST-positive individuals in TB and HIV endemic areas, despite the fact that even TST-negative people are likely to have latent TB. It appears possible that an interaction between isoniazid and the human T cell response is required before isoniazid can be effective in latency. The isoniazid-T cell interaction may be mediated through endogenous vaccination by released mycobacterial antigens.

Multi-drug resistant TB (MDR-TB), defined as resistance of Mtb to INH and rifampicin, is an important problem globally (Migliori et al. 2010). Currently there are no methods by which patients with latent TB due to INH resistant strains of Mtb can be identified. This poses another dilemma for national TB programs which are following WHO guidelines using IPT. A fundamental and crucial question arises as to whether INH is effective against latency caused by infection with MDR strains of Mtb, which by definition are resistant to INH. This may be answered through current longitudinal cohort studies of patients in highly MDR-TB endemic areas who are being given INH by national TB programmes (Fraser et al. 2006).

Other drugs for treatment of latent TB

While unravelling the mysteries of latency and mechanisms of action of INH continues, other clinical approaches in eradicating the huge Mtb human reservoir include further studies for better selection of combination of drug candidates to eradicate latent Mtb. Two large-scale trials are ongoing (Lobue & Menzies 2010); one compares efficacy and effectiveness of 9 months of INH with 4 months rifampicin; the other compares 9 months of daily self-administered INH with 3 months of once-weekly INH combined with rifapentine. The results of these two trials may shape future recommendations for preventive therapy. However as with INH, these will not shed light on the fundamental scientific basis underlying the mechanism of action of these bactericidal drugs on latent Mtb bacilli.


The challenges posed by Mtb infection, its natural history, its interactions with the immune system and its mechanisms for evasion require more serious investments for basic science research (Sudre et al. 1992; Hernández-Pando et al. 2000). The task of unravelling the mysteries of survival mechanisms evolved by Mtb to overcome the human immune response, and its successful use of the human population to survive and thrive over centuries, is a daunting one. The current scientific basis underlying Mtb dormancy, its dogmas and paradoxes, and the persisting mysteries of Mtb and its interaction with the immune system leave a huge knowledge gap. From the current perspective, it appears that the human race will continue to be outwitted by Mtb, and WHO and TB experts will have to settle for trying to achieve global TB control rather than global eradication of Mtb.


Alimuddin Zumla, Justin O’Grady, Matthew Bates, Peter Mwaba, Keertan Dheda, Markus Maeurer, Michael Hoelscher receive grant support from EuropeAID (ADAT), EU-FW7, UK-MRC, SIDA, and EDCTP. Alimuddin Zumla also acknowledges support from NIHR-UCLH CBRC.