Diagnosis and treatment of latent infection with Mycobacterium tuberculosis


  • The Authors: Cynthia B.-E. Chee, MBBS, FRCP, is senior consultant respiratory physician at the Tuberculosis Control Unit and Department of Respiratory and Critical Care Medicine, Tan Tock Seng Hospital, Singapore. Her research interests are clinical and public health aspects of TB. Martina Sester, PhD, is a Professor of Transplant and Infection Immunology at Saarland University, Germany, and her research interests focus on the characterization of cellular immune responses for the control of clinically relevant pathogens. Wenhong Zhang, MD, PhD, is a Professor of Medicine at Fudan University, China. He is the director of the Department of Infectious Diseases. His research interests include diagnosis and treatment of infectious diseases, especially TB. Christoph Lange, Dr. med. Dipl.-Biol, FIDSA, is a Professor of Medicine at the University of Lübeck, Germany. At the Research Center Borstel, Germany, he is the principal attending physician of the Medical Clinic and the Head of the Division of Clinical Infectious Diseases. His research interests include topics of the prevention, diagnosis and treatment of TB.
  • Conflict of Interest Statement: MS and CL have received kits free of charge from Cellestis and Oxford Immunotec to study the performance of interferon-γ release assay responses in immune-compromised patients. CL has conducted investigator-initiated clinical trials, where QuantiFERON-TB Gold In-Tube and T-SPOT.TB test kits were provided by the manufacturers free of charge. MS has a patent application entitled ‘In vitro process for the quick determination of a patient's status relating to infection with Mycobacterium tuberculosis’ (international patent number WO2011113953/A1).

Correspondence: Christoph Lange, Clinical Infectious Diseases, Research Center Borstel, 23845 Borstel, Germany. Email: clange@fz-borstel.de


In clinical practice, latent infection with Mycobacterium tuberculosis is defined by the presence of an M. tuberculosis-specific immune response in the absence of active tuberculosis. Targeted testing of individuals from risk groups with the tuberculin skin test or an interferon-γ release assay is currently the best method to identify those with the highest risk for progression to tuberculosis. Positive predictive values of the immunodiagnostic tests are substantially influenced by the type of test, the age of the person who is tested, the prevalence of tuberculosis in the society and the risk group to which the person belongs. As a general rule, testing should only be offered when preventive chemotherapy will be accepted in the case of a positive test result. Preventive chemotherapy can effectively protect individuals at risk from the development of tuberculosis, although at least 3 months of combination therapy or up to 9 months of monotherapy are required, and overall acceptance rate is low. Improvements of the current generation of immunodiagnostic tests could substantially lower the number of individuals that need to be treated to prevent a case of tuberculosis. If shorter treatment regimens were equally effective than those currently recommended, acceptance of preventive chemotherapy could be much improved.


US Centre for Disease Control and Prevention


confidence interval


human immunodeficiency virus




interferon-γ release assay




latent infection with Mycobacterium tuberculosis


preventive therapy










tuberculin skin test


Based on information available on the frequency of positive tuberculin skin test (TST) responses, the World Health Organization estimates that one-third of the world's population is infected with Mycobacterium tuberculosis, the causative microorganism of tuberculosis (TB).[1] However, only a small minority of individuals with latent infection with M. tuberculosis (LTBI), by definition diagnosed on the basis of a positive TST or an interferon-γ release assay (IGRA) result, develops TB in the future.[2, 3] It is unclear how well immunological tests, such as the TST or IGRA, identify individuals that are truly infected with M. tuberculosis.[4] Nevertheless, these tests are currently the best available diagnostic methods for the routine risk evaluation of future TB development in clinical practice.

Identification of individuals with a positive TST or IGRA result from risk groups for the future development of TB and preventive chemotherapy are efficient measures for TB prevention.

This review describes the concept and immunopathology of LTBI, current methods for the diagnosis and treatment for the prevention of TB, and discusses current areas of controversy and possibilities to improve prevention of TB.

Immunopathology of LTBI

LTBI is defined by the presence of an M. tuberculosis-specific immune response in the absence of clinical and radiological disease. It is, however, unclear whether such individuals harbour viable M. tuberculosis or just maintained the immune response after eradication of M. tuberculosis. More recently, it has been proposed that there is a spectrum of latent infection ranging from those with obvious TB lesions containing live bacilli but without symptoms (‘near-active’ TB) to those who have eradicated the infection with virtually no chance of reactivation.[6-8]

During latent infection, the host immune system is able to contain the bacilli in a state of non-replicating persistence. The outcome of M. tuberculosis infection depends on the interaction between the host immune system and the bacteria. The bacilli enter the host lung and survive in alveolar macrophages by arresting phagosome maturation at an early stage and preventing its destruction and degradation within macrophages.[9] Interactions between T cells and infected macrophages thereafter initiate formation of a productive granuloma. Antigens derived from M. tuberculosis are processed by antigen-presenting cells. Subsequently, CD4 T cells are involved, which mature into predominantly two functionally different phenotypes, T helper 1 (Th1) and Th2 cells. The former principally secrete interleukin 2 and interferon (IFN)-γ, while the latter largely secrete interleukins 4, 5, 6 and 10. Containment of M. tuberculosis in granulomatous lesions is under the control of cell-mediated protective immunity that is largely mediated by T lymphocytes and their cytokines.[10]

Recent advances in global gene expression profiling provided valuable insights into host defence against M. tuberculosis infection. Apart from producing cytokines that activate macrophages and initiate granuloma formation, T cells also express direct microbicidal activities via a concerted action of perforins and granulysins, and antimicrobial peptides such as cathelicidin.[11, 12] The important role of tumour necrosis factor in controlling LTBI is demonstrated by the observation that tumour necrosis factor antagonists used in the treatment of rheumatoid arthritis can cause reactivation of active TB.[13] This study was the first showing in humans, rather than animal models, that a cytokine played a key role in the immunopathogenesis from LTBI to active disease. Recently, more clinical trials confirmed this new mechanism for the progression from latency to active TB.[14, 15]

Conventional CD4 and CD8 T cells recognize mycobacterial peptides in the context of gene products of the major histocompatibility class I or class II molecules.[16] In addition, unconventional T cells comprising γδ T cells and CD1-restricted T cells with specificity for non-proteinaceous antigens exist.[17] A study in HIV patients suggested that robust immune responses of Vgamma2Vdelta2+ and CD8+ T effector cells provide protective immunity to prevent the development from latent infection to active TB.[18] Moreover, increasing evidence is emerging that memory T cells participate in protective immunity against TB.[19] It has been suggested that memory T cells can persist in the absence of nominal antigen, which could play a role in a number of purified protein derivative-positive individuals that developed an immune response strong enough to eradicate the pathogen. The roles of regulatory T cells[20] and Th17 in LTBI and active TB are worth commenting. It is generally assumed that regulatory T cells reduce the risk of immunopathology by suppressing ongoing immune response after pathogen eradication. Reduced Th17 responses were associated with the clinical outcome of M. tuberculosis infection, and suppression of Th17 responses through downregulation of interleukin 6R expression may be an important mechanism in the development of active TB.[21, 22]

Methods for the Diagnosis of LTBI

As opposed to active TB, where the diagnosis relies on the detection of M. tuberculosis bacilli and/or clinical symptoms, an LTBI is indirectly diagnosed by the presence of a specific cellular immune response directed towards mycobacterial antigens in the absence of clinical disease.[4] As such, the presence of a specific immune response is a good proxy of a prior or actual encounter with mycobacterial antigens. However, when used in a clinical setting, a specific immune response does not provide any direct evidence for viable bacilli[4, 7] and hence does not represent a very specific indicator of an increased risk for progression towards active TB.

Assays for the diagnosis of LTBI comprise the in vivo TST or more recent developments of ex vivo methods to detect a specific cellular immune response based on the induction of cytokines after stimulation of lymphocytes with mycobacterial antigens.[4] Obviously, the specificity of detecting an immune response resulting from an encounter with M. tuberculosis critically depends on the use of antigens derived from bacilli of the M. tuberculosis complex. The sensitivity of an immune-based assay is determined by the overall percentage of responding cells and by the general immune status of the tested individual.

The TST and in vitro assays share common principles. The current TST has been in use for more than a century. It is an in vivo assay that elicits a delayed type hypersensitivity reaction towards tuberculin, which represents a crude mixture of proteins obtained from the sterile supernatant of liquid cultures of M. tuberculosis. The extent of cellular infiltrates correlates with a skin induration, which is quantified 48–72 h after intradermal tuberculin inoculation.[4] Although well suited for use in resource poor settings, a major drawback of the TST is its low specificity, as a tuberculin-specific skin reaction may not only originate from a previous encounter with M. tuberculosis but also from M. bovis bacillus Calmette–Guérin vaccination or infection with non-tuberculous mycobacteria. In addition, the TST is frequently falsely negative, especially in immunocompromised patients due to cutaneous anergy.[23] Finally, the test requires two visits, and the intradermal application of defined amounts of tuberculin and correct measurement of induration requires well-trained personnel to obtain standardized results.

In the recent years, in vitro tests have been developed, which overcome many of the operational and conceptual disadvantages of the TST. These tests may be applied ex vivo directly from whole-blood or isolated lymphocyte fractions. An increase in specificity is conferred by the use of antigens encoded in the region of difference 1 (RD-1) genomic region of M. tuberculosis that is deleted in the M. bovis bacillus Calmette–Guérin strain and most non-tuberculous mycobacteria species. These antigens termed early secretory antigenic target-6 and culture filtrate protein-10 may be used in vitro to stimulate T cell activation and cytokine induction in an antigen-specific manner.[24] Typical read-out systems for cytokine analysis require stimulation times as short as 6–20 h and include the quantitation of cytokine concentrations from supernatants using an enzyme-linked immunosorbent assay, the enumeration of cytokine-producing cells using an enzyme-linked immunosorbent spot assay or the quantitation of cytokine-producing cells using flow cytometric intracellular cytokine staining.[4] Although T cells secrete a variety of mainly Th1 cytokines such as interleukin 2, IFN-γ or tumour necrosis factor-α, the cytokine IFN-γ is most commonly used as read-out. Whole blood tests relying on IFN-γ analysis by enzyme-linked immunosorbent assay or enzyme-linked immunosorbent spot assay are commercially available as QuantiFERON-TB Gold In-Tube test (Cellestis, a Qiagen company, Hilden, Germany) and as T-SPOT.TB assay (Oxford Immunotec, Abingdon, UK), respectively, and are commonly referred to as IGRA. In addition to an increase in specificity, in vitro assays are faster as compared with skin test and may be performed together with additional stimulatory reactions such as negative or positive controls to internally control for antigen-non-specific reactivity or general immune function, respectively. Although an excess of T cell reactivity in the negative control or a limited reactivity in the positive control is formally scored as an indeterminate result, these outcomes are still valuable to estimate the overall response profile of a given individual in vivo, especially in the situation of low T cell counts or any other type of immunodeficiency.[23, 25]

When formally evaluating the performance of TST or IGRA, it is important to distinguish between their potential to detect a specific immune response and the clinical use of a given test result for assessing the risk of progression towards active TB. Meta-analyses on the performance of TST and IGRA to detect an immune response towards M. tuberculosis in low-risk controls have shown that the specificity of IGRA is very high, especially in bacillus Calmette–Guérin-vaccinated individuals.[26] Evaluations of sensitivities have used various clinical situations where the presence of a specific immune response is likely, such as patients with TB, a history of active disease or close contacts of index cases with active TB.[26-28] Those groups are suboptimal as a gold standard, as T cell responses may frequently be absent. Nevertheless, these studies have revealed that IGRA shows better correlation with surrogate measures for M. tuberculosis exposure and is of similar or even superior sensitivity as compared with TST. Studies on superiority of IGRA in very young children are more controversial in outcome, as the percentage of indeterminate results may be high.[4] Nevertheless, a large European multicentre study on the comparative analysis of IGRA and TST indicated a large percentage of IGRA-positive results among TST-negative children, which indicates that LTBI is underestimated based on the use of TST only.[29] Thus, IGRA seems superior to detect evidence for the presence of a specifically immune response towards M. tuberculosis, although this immune response does not necessarily correspond to a status of latent infection with viable bacilli. Instead, a specific immune response may be present in a wide spectrum of clinical states that include individuals with active TB, non-symptomatic subclinical infections and, finally, individuals where bacilli may have been completely eliminated.[7, 23]

The fact that TST- or IGRA-positive individuals may originate from a diverse spectrum of clinical states clearly indicates that TST or IGRA cannot by itself be used to diagnose TB or to assess the potential risk for the development of TB. According to a recent meta-analysis, neither TST nor IGRA should be used in the diagnosis of active TB, as the sensitivity of both assays is not sufficiently high.[30] In addition, analyses of TB suspects with final diagnoses other than TB showed that specificity for active TB was also low due to a considerable extent of non-symptomatic individuals with detectable immunity (i.e. individuals with LTBI).[30] Recently, studies have accumulated, which have analysed the negative and positive predictive values of TST or IGRA to assess potential risk for development of TB in close contacts of contagious patients with active TB or in patients at risk for progression due to immunosuppressive conditions (Table 1).[3, 26] Of note, both TST and IGRA have a very high negative predictive value in immunocompetent individuals indicating that individuals without evidence of specific immunity are unlikely to progress.[26] When assessing positive predictive values for progression, IGRA seems to be superior over TST in the setting of contact tracing in low-incidence countries. Two remarkable studies from single centres in Germany and the UK demonstrated a progression rate of 12.9%[32] and 12.5%,[33] respectively, in adult IGRA test-positive close contacts of acid-fast bacilli smear-positive index cases who did not receive preventive chemotherapy. The corresponding progression rate for TST-positive individuals was only 3.1%.[32] The progression rates of IGRA in both studies are substantially higher than in other settings that were potentially more heterogeneous in underlying risk conditions or burden of remote infection.[34-36] Nevertheless, it is plausible that IGRA is superior to the TST in predicting progression to TB due to the independence of the IGRA test result from previous vaccination with M. bovis bacillus Calmette–Guérin.[26] Positive predictive values of TST and IGRA do not appear to differ widely in high-prevalence countries presumably due to a higher specificity of TST for M. tuberculosis infection in this setting.[3]

Table 1. Progression rates towards active tuberculosis in non-treated individuals with positive and negative results in TST or IGRA depending on risk group for progression and origin of study
Risk factor


Origin of study

TestProgression rate if test is positiveProgression rate if test is negativeObservation period (median, months)Comments
  1.  Information is given for patients who did not receive or  did not receive or complete chemotherapy.
  2. HIV, human immunodeficiency virus; IGRA, interferon-γ release assay; QFT-G-IT, QuantiFERON Gold In-Tube; TNF, tumour necrosis factor; TST, tuberculin skin test.
Close contactsDiel et al.[32]QFT-G-IT12.9% (19/147)0% (0/756)43
GermanyTST (5 mm)3.1% (17/555)0.6% (2/348)
Haldar et al.[33]QFT-G-IT12.5% (14/112)1.0% (6/601)24No TST performed
Immigrant close contactsKik et al.[34]QFT-G-IT2.8% (5/178)2.0% (3/149)22
The NetherlandsT-SPOT.TB3.3% (6/181)1.7% (2/118)
TST (15 mm)2.7% (5/184)0.7% (1/138)
HIV infectionAichelburg et al.[43]QFT-G-IT8.1% (3/37)0% (0/738)19No TST performed
Renal transplantationKim et al.[31]T-SPOT.TB5.6% (4/71)0% (0/171)21Bias for TST-negative patients, as those did not receive chemotherapy
South Korea
SilicosisLeung et al.[35]T-SPOT.TB8.0% (12/151)1.1% (1/90)29
ChinaTST (5 mm)5.6% (9/161)5.0% (4/80)
TNF antagonist therapyJung et al.[36]T-SPOT.TB2.6% (3/114)0% (0/71)38Bias for TST-negative patients, as those did not receive chemotherapy
South KoreaTST (10 mm)8.0% (2/25)1.8% (3/163)

When comparatively analysing results of immune-based assays in patients with various types of immunodeficiencies, the percentage of IGRA-positive individuals is in general higher as compared with TST-positive patients, which is indicative of a higher sensitivity.[23] Among immunocompromised patients, the percentage of positive IGRA test results in patients with rheumatoid arthritis,[37, 38] psoriasis[39] or chronic renal failure[40] is higher than in patients with HIV infection,[41, 42] yet the progression rate to TB in patients from these groups is likely lower when compared with HIV-infected individuals.[43] Thus, despite considerable improvement in detection of M. tuberculosis-specific immune responses, the predominant proportion of individuals with positive tests does not develop active TB. This emphasizes the need for screening strategies where testing is targeted to individuals with risk factors for TB to finally increase specificity of risk assessment and to reduce the extent of overtreatment.

LTBI Treatment Regimens

Persons deemed to have LTBI and who are at increased risk for progression to active TB should be considered candidates for preventive therapy (PT). The groups with the highest risk are those recently infected (e.g. close contacts of infectious TB cases, recent TST converters), HIV-infected individuals and persons with fibrotic lesions on chest radiograph consistent with old, healed but previously untreated TB.[44] Other risk groups include organ transplant recipients on immunosuppressive treatment[45, 46] and persons with silicosis, end-stage renal failure or those on immunomodulatory therapy.[15, 47] Targeted testing and treatment of LTBI in close contacts and other risk groups are a key TB control and elimination strategy in many high-income, low-TB prevalence countries. In contrast, the top priority of TB programmes in high-TB prevalence countries, of which many are resource-constrained, should be the prompt detection and successful treatment of infectious TB cases (i.e. ‘turning off the tap’ of TB transmission). In such settings, international authorities recommend that children under 5 years of age and persons of any age with HIV infection who are close contacts of an infectious index patient and who, after careful evaluation, do not have active TB should be treated for presumed LTBI with isoniazid (INH).[48] The threshold for implementation of LTBI testing and treatment as a TB control strategy and the priority groups to be targeted should be determined by the health authorities of each country or region according to the local situation and resources available.

At the individual level, the decision for or against PT must consider the potential risk and consequences of developing active TB weighed against the risk of adverse drug reaction (particularly hepatotoxicity) and the likelihood of treatment adherence and completion. It is of utmost importance to exclude active TB by history, physical examination, chest radiograph and, if necessary, sputum examination before commencing PT. The patient should be educated regarding symptoms that may suggest hepatotoxicity (such as nausea, vomiting, abdominal discomfort, unexplained fatigue), and to stop the medication and return for evaluation should such symptoms occur. Patients should be reviewed monthly for close clinical monitoring of adverse drug effects and treatment adherence. Table 2 shows the various recommended PT regimens, the evidence for which is summarized later. The interested reader is also referred to two comprehensive reviews on this topic by Leung et al.,[44] and Lobue and Menzies.[50] Because these reviews were published in 2010, new evidence has emerged, which has influenced recommendations, and these are included in the present review.

Table 2. Recommended LTBI treatment regimens
INH monotherapy
RegimenDrug dosage, dosing scheduleRecommending authority and remarks
INH for 6 months Daily, self-administered WHO, ISTC[48, 96]
5 mg/kg up to maximum of 300 mg/dayChildren <5 years old and HIV-infected persons of any age who are household contacts of infectious TB cases
Twice weekly, DOT (ATS/CDC)WHO[49]
15 mg/kg, up to maximum of 900 mg/doseChildren, adults and adolescents living with HIV should receive at least 6 months of INH PT as part of a comprehensive package of HIV care (strong recommendation)
NICE, UK[66]
Equal alternative to 3RIF + INH
Recommended regimen for HIV-infected persons of any age
Alternative regimen to 9INH
Except for HIV-infected, children or those with fibrotic lesions on chest radiograph
INH for 9 monthsDaily, self-administeredATS/CDC
5 mg/kg up to maximum of 300 mg/dayPreferred PT regimen for all risk groups
Twice weekly, DOT
15 mg/kg, up to 900 mg/dose maximum
INH for at least 36 monthsWHO[86]
Adult and adolescents living with HIV in high-TB prevalence and transmission settings (conditional recommendation)
Alternative regimens to INH monotherapy
RegimenDrug dosage, dosing scheduleRecommending authority and remarks
  1. ART, antiretroviral therapy; ATS, American Thoracic Society; CDC, US Centres for Disease Control and Prevention; DOT, directly observed treatment; HIV, human immunodeficiency virus; INH, isoniazid; ISTC, International Standards for Tuberculosis Care; LTBI, latent infection with Mycobacterium tuberculosis; PT, preventive therapy; RIF, rifampicin; RPT, rifapentine; TB, tuberculosis; WHO, World Health Organization.
RIF for 4 monthsDaily, self-administeredATS/CDC
10 mg/kg, 600 mg/day maximumAlternative to 9INH
Contacts of INH-resistant, RIF-susceptible index cases
RIF for 6 monthsDaily, self-administeredUK
10 mg/kg, 600 mg/day maximumContacts, aged 35 or younger, of INH-resistant, RIF-susceptible TB cases
RIF + INH for 3 monthsDaily, self-administeredUK
RIF: 10 mg/kg, maximum 600 mg/dayEqual alternative to 6INH
INH: 5 mg/kg up to maximum of 300 mg/dayNot recommended for HIV-infected persons
RPT + INH for 3 months (to be administered only under DOT)Once weekly, DOTCDC[80]
INH: 15 mg/kg, 900 mg maximumEqual alternative to self-administered 9INH in persons ≥12 years old
RPT: 10.0–14.0 kg: 300 mgNot recommended for children <2 years, HIV-infected persons on ART, pregnant women or women expecting to become pregnant during treatment, those with LTBI with presumed INH or RIF resistance
14.1–25.0 kg: 450 mg
25.1–32.0 kg: 600 mg
32.1–49.9 kg: 750 mg
≥50 kg: 900 mg maximum

Persons with risk factors other than HIV infection for progression to TB disease

INH monotherapy

INH, synthesized in 1912, was found to have bactericidal activity against M. tuberculosis in the 1950s. The United States Public Health Service conducted a series of randomized, controlled trials in the 1950s and 1960s that demonstrated the efficacy of 12 months of daily INH (12INH) in reducing the rate of active TB among immunocompetent, TST-positive household contacts,[51, 52] residents of mental institutions[53] and native Alaskans in high-TB prevalence communities.[54] The protective effect of INH persisted for up to 19 years in the Alaskan study. In the 1970s, the International Union Against TB conducted a trial in Eastern Europe evaluating 12, 24 and 52 weeks of INH PT in 28 000 tuberculin-positive persons with previously untreated fibrotic lesions on chest radiograph. This showed a reduction in development of culture-positive TB of 21%, 65% and 75% among those randomized to 12, 24 and 52 weeks of INH, respectively. Among those who completed and complied with treatment, the reduction in TB rates was 30%, 69% and 93%, respectively.[55] Cost-effectiveness analysis of this trial data indicated that the 24-week regimen was more cost-effective than the 12- or 52-week regimen. This led most public health programmes to adopt 6 months of INH (6INH) as the standard duration for PT.[56] Subsequent review of data from the International Union Against TB and United States Public Health Service studies (including the Alaskan study) by Comstock in 1999 concluded that the optimal protection from INH appeared to be obtained by 9–10 months.[57] Based on this reanalysis, the American Thoracic Society and US Centres for Disease Control and Prevention (CDC) in 2000 recommended 9 months of INH (9INH) as the standard for LTBI treatment.[58] A meta-analysis of 11 trials involving 73 375 non-HIV-infected persons, however, showed no significant difference between 6- and 12-month courses of INH in reducing the risk of developing active TB compared with placebo.[59] Randomized, controlled trials on the efficacy of INH PT in persons with HIV infection and silicosis are discussed later. The use of INH PT for transplant recipients, those on immunomodulatory therapy, and those with medical conditions other than HIV infection and silicosis is extrapolated from evidence from the earlier studies.

INH hepatotoxicity

Following two fatalities in an outbreak of hepatitis among TB contacts receiving INH PT in 1970,[60] the United States Public Health Service undertook a survey of 21 health departments that found a 1% rate of hepatitis and eight deaths among 13 838 individuals who received INH. Hepatotoxicity increased with older age and daily alcohol consumption.[61] A meta-analysis of six studies involving 38 257 adult patients treated with INH monotherapy showed a 0.6% incidence of clinical hepatitis (range 0–2.9%).[62] Subsequent observational studies among 11 141 patients in US Public Health Clinics that utilized clinical monitoring reported INH hepatotoxicity rates of 0.1% with no fatalities in either study, providing reassurance as to the safety of INH PT when patients are carefully selected and clinically monitored for adverse effects.[63, 64]

Rifampicin monotherapy and rifampicin-containing regimens

Concern regarding INH hepatotoxicity and the prolonged duration of INH PT resulting in low rates of patient acceptance and adherence has prompted the search for shorter and better tolerated regimens. Rifampicin (RIF) is bactericidal, with good sterilizing activity against M. tuberculosis. A randomized, controlled trial in Chinese men with silicosis in Hong Kong compared 3 months of RIF + INH (3RIF + INH) with 3 months of RIF monotherapy (3RIF), 6INH and placebo.[65] The cumulative percentage of patients with active pulmonary TB over 5 years among those who received placebo was significantly higher than among any of the treatment groups with no significant difference between the three treatment regimens (placebo 27%, 3RIF + INH: 16%, 6INH 14% and 3RIF 10%). This study formed the basis for the recommendation of 3RIF + INH for non-HIV-infected persons by the UK health authorities[66] and that of 4 months of RIF (4RIF) as an alternative to 9INH by the American Thoracic Society and CDC.[58] Meta-analysis of pooled data from four studies involving 3586 patients showed that compared with 9INH, 4RIF was associated with significant reduction in risk of non-completion (relative risk 0.53; 95% confidence interval (CI) 0.44–0.63) and hepatotoxicity (relative risk 0.12; 95% CI 0.05–0.30). The authors also calculated that the 4RIF strategy was cost-effective, resulting in US$213 savings per patient treated.[67] Apart from the Hong Kong study, there are no other randomized, controlled trials comparing the efficacy of RIF monotherapy with standard INH PT regimens in reducing active TB. A multicentre trial comparing the effectiveness of 4RIF with 9INH is currently underway and is expected to be completed in 2016.[68] In addition to the Hong Kong study, 3–4 months of RIF + INH was evaluated against standard therapy (6–12INH) for efficacy in a Spanish non-HIV-infected cohort and three HIV-infected cohorts. A meta-analysis of these five studies showed that short-course RIF + INH was equivalent to standard INH therapy in terms of efficacy, proportion of severe side-effects and mortality.[69] An observational study comparing 4RIF + INH with 12INH in two sequential cohorts with radiographic evidence of previous TB showed that both regimens had similar rates of treatment completion and adverse effects, both increased life expectancy compared with no treatment and 4RIF + INH was cost-saving compared with 12INH.[70] A randomized, controlled study in Greece comparing 3–4 months of RIF + INH with 9INH in children found that patients randomized to 9INH were less compliant than those who received RIF + INH. Although none developed clinical disease during follow up, new radiographic findings suggestive of possible active disease were more common in those who received 9INH. Serious adverse effects were not detected.[71] Several randomized, controlled studies in the 1990s showed that 2–3 months of RIF + pyrazinamide (PZA) was as efficacious as 6–12INH in reducing active TB in the HIV-infected population.[72-75] This led to the recommendation of 2RIF + PZA in the 2000 American Thoracic Society/CDC statement. The resultant widespread use of this regimen in the US was unfortunately followed by reports of severe and fatal hepatotoxicity among non-HIV-infected persons, leading the American Thoracic Society and CDC in 2003 to issue a recommendation against its use.[76]

Three months of once-weekly, directly observed rifapentine and INH

Rifapentine (RPT) is a rifamycin with a long half-life and greater potency against M. tuberculosis than RIF. A regimen of once-weekly doses of directly observed RPT + INH for 12 weeks was compared with self-administered 2RIF + PZA in 399 TST-positive, adult, largely HIV-negative household contacts in Brazil. Follow up for more than 2 years showed a TB incidence rate ratio of 2.8 for RPT + INH versus RIF + PZA (95% CI 0.2–26.8), and hepatotoxicity was 10% in the RIF + PZA group versus 1% in the RPT + INH group.[77] This regimen was evaluated against three other regimens (twice-weekly, directly observed RIF + INH for 12 weeks, self-administered 6INH and INH continuously for up to 6 years) in 1148 HIV-infected, TST-positive South African adults who were not on antiretroviral therapy. No significant difference was found in the rate of active TB over 4 years (incidence rates of 1.4–2.0 per 100 person-years). Serious adverse reactions were more common in the continuous INH group (18.4 per 100 person-years) than in the other treatment groups (8.7–15.4 per 100 person-years).[78] A large-scale, multicentre study in the US, Canada, Brazil and Spain involving 7731 subjects compared once-weekly, directly observed RPT + INH for 12 weeks (‘combination therapy’) with self-administered 9INH. This predominantly HIV-negative study population comprised TST-positive persons of whom 71% were close contacts and 25% recent TST converters. Over a follow-up period of 33 months, the cumulative TB incidence rates were 0.19% and 0.43% among subjects in the combination therapy group and INH monotherapy group, respectively. There was less hepatotoxicity in the combination therapy group (0.4% vs 2.7%) but more adverse (mostly hypersensitivity) reactions leading to discontinuation of therapy (4.9% vs 3.7%).[79] Based on these three studies, the US CDC in December 2011 recommended once-weekly, directly observed RPT + INH for 12 weeks as an equivalent alternative to self-administered 9INH in otherwise healthy persons more than 12 years of age with risk factors for developing active TB.[80]

PT regimens for HIV-infected persons

HIV infection is the strongest known risk factor for the development of TB disease. A meta-analysis of seven randomized, controlled trials that compared INH with placebo in 4539 subjects showed a risk ratio of 0.40 (95% CI 0.24–0.65) and 0.84 (95% CI 0.54–1.30) for TST-positive and -negative persons, respectively.[81] Shorter, RIF-containing regimens (2RIF + PZA, 3RIF + INH) were compared with 6–12INH in several randomized, controlled studies. A systematic review of 12 studies that included 8578 participants showed that LTBI treatment (regardless of regimen, frequency or duration of treatment) reduced the risk of active TB, especially in those with positive TST (relative risk 0.38; 95% CI 0.35–0.57) versus TST-negative (relative risk 0.89; 95% CI 0.54–1.24). Compared with INH monotherapy, multidrug regimens were much more likely to require discontinuation due to side-effects.[82] A major concern is the durability of protective effect of PT in HIV-infected persons in high-TB prevalence settings. It was observed among HIV-positive individuals in sub-Saharan Africa that the protective effect of 6INH was lost within 6–18 months of its completion.[83, 84] A recent study in Botswana, a TB-endemic setting, that randomized 1995 HIV-positive subjects to 6INH or 36 months of INH (36INH) showed that 36INH was more effective than 6INH for preventing TB in those who were TST-positive. antiretroviral therapy was provided to those with CD4 cell count <200/μL. TB incidence was reduced by 50% in those who received 360 days of antiretroviral therapy versus those who did not receive antiretroviral therapy. Severe INH-associated hepatotoxicity rates in those who received 36INH were similar or below that with shorter courses of INH, with almost all cases occurring during the first 9 months of treatment.[85] In another TB-endemic setting, 12-week courses of intermittent RPT + INH or RIF + INH, and continuous INH (for up to 6 years) were not found to be superior to 6INH in TST-positive, HIV-infected South African adults (median CD4 cell count 484/μL) who were not on antiretroviral therapy.[78] Based on the Botswana study, the World Health Organization 2011 guidelines included a conditional recommendation that adults and adolescents living with HIV in high-TB prevalence and transmission settings should receive at least 36 months of INH PT.[86]

Regimens for contacts of drug-resistant TB

To date, there are no published randomized, controlled trials on PT for persons presumed to be latently infected with INH-resistant or INH- and RIF-resistant (i.e. multidrug-resistant (MDR)) strains. Two case series reported that contacts of INH-resistant cases who completed 6 months of RIF did not develop active TB; in one of these studies, contacts who received INH had the same rate of active TB as those who did not receive any PT.[87, 88]

A recent systematic review concluded that based on the available evidence, it is not possible to support or reject the use of PT for contacts of MDR-TB.[89] Currently recommended PT regimens for MDR-TB contacts are based on expert opinion. The US CDC recommends at least two drugs (combinations of PZA, ethambutol and/or fluoroquinolone), to which the source case's M. tuberculosis isolate is susceptible, for 6–12 months.[90] However, poor tolerance with high rates of hepatotoxicity and premature discontinuation has been reported in persons prescribed with these regimens (all of whom received PZA).[91-93] Fluoroquinolone monotherapy has been suggested as a safer alternative;[94] this, however, raises concern regarding the generation of resistance to this key second-line anti-TB drug. Some paediatric experts support the use of high-dose INH and a fluoroquinolone (ideally levofloxacin) for a minimum of 6 months in children younger than 5 years or who are HIV-infected.[95] In view of the lack of evidence for efficacy and poor tolerability of MDR PT regimens, a reasonable option would be close observation for at least 2 years for MDR-TB contacts who are otherwise healthy and do not have risk factors for rapid disease progression and dissemination. This approach is recommended by UK health authorities, World Health Organization and European CDC.[66, 96, 97]

Areas of Controversy and Future Directions

Immunodiagnosis of LTBI has the primary purpose to identify individuals with the highest risk among risk groups for the future development of TB. The key factors that allow an understanding of the impact of immunodiagnosis for TB prevention are predictive values of the diagnostic tests and the efficacy of preventive chemotherapy. Predictive values of the diagnostic tests are dependent upon the prevalence of the disease and on the pretest probability of the individual for the development of TB.

A person with a positive TST or IGRA result following recent exposure to a patient with infectious TB has a higher risk for the progression to active TB than a person with a positive test result not belonging to any risk group. As background frequencies of positive results in immunodiagnostic tests can be higher than 10% in low-TB incidence countries, it is important to target immunodiagnostic testing only to risk groups. In countries of low TB prevalence, the negative predictive values of the IGRA and the TST are close to 100%, indicating that individuals with a negative test result in either test have a very low likelihood of progression to active disease in the years following testing, even if belonging to a risk group.[26] When the risk is already very low, it cannot be substantially reduced by preventive chemotherapy.

Recent studies provide convincing information that IGRA is superior to the TST to identify individuals at risk for the progression to TB. It is apparent that without preventive chemotherapy, >85% of close adult TB contacts with a positive IGRA result do not develop TB within 2 years when the risk for TB is highest.[26, 32, 33] In a recent UK study, 35 adult contacts were required to be screened to identify one contact developing TB at 2 years.[33] Despite this success, better characterization of the predictive values of immunodiagnostic tests for the future development of TB in close contacts of patients with contagious TB and in other patients at risk is needed.

An ideal test for LTBI should have a substantially improved positive predictive value to justify the indication for preventive chemotherapy. Based on promising findings on immunophenotypical changes in specific immunity in patients with active TB,[98, 99] improvements in the diagnosis of patients at risk could be possible when cytokines other than IFN-γ or combinations of different read-outs are being considered.[100] Interleukin 2[98, 99, 101-103] and interferon-gamma induced protein 10 (IP-10)[104-106] are among the cytokines that are being evaluated to optimize immunodiagnostic assays to diagnose TB and to predict TB development in individuals from risk groups. As this requires simultaneous analysis of multiple parameters, flow cytometry is particularly well suited to define more complex immunophenotypical or functional signatures characteristic for individuals with LTBI to progress to active TB. Other less hypothesis-driven approaches such as transcriptomics, proteomics or metabolomics could further aid in the discovery of biomarkers to identify patients at risk for progression towards active TB, but technical requirements are not yet suitable for use in a clinical setting.[107-109] Using blood transcriptional profiling, an IFN-inducible, neutrophil-driven whole-blood transcript signature was identified for active TB, which was also found in a subset of 10–25% of individuals with LTBI, suggesting that this transcript signature could potentially identify patients with LTBI with higher microbiological burden or subclinical disease.[107] Likewise, an analysis of the transcription profiles of purified protein derivative-stimulated peripheral blood mononuclear cells identified a combination of IP-10, the cation transport adenosine triphosphatase 10A and Toll-like receptor 6 gene expression that discriminated active TB from LTBI with a sensitivity 71% and a specificity 89%.[108] Together, these findings provide further insight as to the immunological determinants promoting progression of LTBI to active TB and should be developed further towards establishing biomarkers suitable for point-of-care diagnosis of patients at risk.

Additional efforts should focus on more targeted screening of patients at risk, an improvement of acceptance rates for PT and finally patients on shorter treatment regimens. Even though the duration of preventive treatment regimens is becoming shorter,[79] the duration of preventive chemotherapy is far too long to be acceptable, for example, compared with the eradication for Helicobacter pylori to prevent gastritic ulcerative disease or gastric carcinoma.[110] Although the efficacy of preventive chemotherapy has been ascertained to be 90% for 9 months of daily INH monotherapy,[57, 59, 111, 112] 69% for 6 months of daily INH therapy[111] and 65% for 3 months of daily combination therapy with INH and RIF,[65, 69] acceptance of preventive chemotherapy is as low as 20% in some countries.[113] These areas need to be targeted to increase public health impact on TB prevention. Resources could be saved if tests were only offered to individuals from risk groups willing to accept preventive chemotherapy in case of a positive test result—intention to test is intention to treat.[114]


Preventive chemotherapy can be an effective, efficacious and efficient intervention to reduce the risk for the development of TB. Targeted testing for M. tuberculosis-specific immune responses should be performed in individuals from groups where positive test results are related to a substantial increase in the risk for the progression to TB, such as specifically in recent contacts in countries of low TB incidence and HIV-infected individuals, and more generally in cases where preventive chemotherapy has an impact on the risk reduction to progress to TB. Resources could be saved, if candidates for immunodiagnostic testing were only investigated if they agree to be treated in the event of a positive test result.

In patients with psoriasis, rheumatoid arthritis, diabetes mellitus and chronic renal failure, and transplant recipients, health-care workers and contacts of patients with MDR/extensively drug-resistant TB, the evidence for targeted immunodiagnostic testing and preventive chemotherapy is less clear. In children younger than 5 years of age who are household contacts of infectious index cases, prophylactic chemotherapy may be warranted without immunodiagnosis, as their overall risk for the progression to TB is very high.

Awareness for and acceptance of preventive chemotherapy could be raised if the positive predictive value of diagnostic tests could be improved and future drug regimens allowed for much shorter treatment periods than those currently recommended.