Treatment of Latent Tuberculosis Infection

ABSTRACT

There are approximately 56 million people who harbor Mycobacterium tuberculosis that may progress to active tuberculosis (TB) at some point in their lives. Modeling studies suggest that if only 8% of these individuals with latent TB infection (LTBI) were treated annually, overall global incidence would be 14-fold lower by 2050 compared to incidence in 2013, even in the absence of additional TB control measures. This highlights the importance of identifying and treating latently infected individuals, and that this intervention must be scaled up to achieve the goals of the Global End TB Strategy. The efficacy of LTBI treatment is well established, and the most commonly used regimen is 9 months of daily self-administered isoniazid. However, its use has been hindered by limited provider awareness of the benefits, concern about potential side effects such as hepatotoxicity, and low rates of treatment completion. There is increasing evidence that shorter rifamycin-based regimens are as effective, better tolerated, and more likely to be completed compared to isoniazid. Such regimens include four months of daily self-administered rifampin monotherapy, three months of once weekly directly observed isoniazid-rifapentine, and three months of daily self-administered isoniazid-rifampin. The success of LTBI treatment to prevent additional TB disease relies upon choosing an appropriate regimen individualized to the patient, monitoring for potential adverse clinical events, and utilizing strategies to promote adherence. Safer, more cost-effective, and more easily completed regimens are needed and should be combined with interventions to better identify, engage, and retain high-risk individuals across the cascade from diagnosis through treatment completion of LTBI.

The World Health Organization (WHO) has recently reported that the global tuberculosis (TB) epidemic is larger than previously projected, with an estimated 10.4 million new (incident) cases occurring in 2015 (1). It has thus become clear that the WHO End TB Strategy targets of a 90% reduction in TB incidence and a 95% reduction in TB deaths by 2035 can be achieved only by combining the effective detection and treatment of active TB with measures to prevent new infection with Mycobacterium tuberculosis and to eradicate existing latent TB infections (LTBI) (1–6). Recent estimates indicate that approximately 1.7 billion people, nearly one-quarter of the world’s population, are latently infected with Mycobacterium tuberculosis and are at risk of progression to active TB without treatment (7). Moreover, an estimated 11% of those are likely infected with an isoniazid-resistant strain. With ongoing transmission of M. tuberculosis and a high rate of reactivation from LTBI to active TB, a heightened global commitment to the identification and treatment of infected persons is thus critical for achievement of TB elimination (1–6, 8–10).

In resource-limited countries with high TB incidence rates, the spread of M. tuberculosis is controlled primarily through identification and treatment of infectious persons with TB disease; active contact tracing and screening of other high-risk persons for LTBI are rarely implemented due to resource constraints (3, 6). In such areas, immunization with bacillus Calmette-Guérin (BCG) is used to reduce the morbidity and mortality of TB among young children but is not effective for preventing primary M. tuberculosis infection or reactivation from latent infection to active disease later in life (11–13). In countries with lower TB disease incidence and higher levels of resources, identification and treatment of the large reservoir of persons with LTBI has been demonstrated an important and feasible component of TB control and elimination (4, 5, 9, 10, 14–17). Recognizing the potential benefit that a combined treatment and prevention approach could have towards TB elimination, the WHO has recently issued evidence-based guidelines to expand the management of LTBI in high-risk individuals in high- or middle-upper-income countries with a low TB incidence (<100 new TB cases per 100,000 per year) (6). This review focuses on the identification and treatment of persons at increased risk of LTBI to prevent future development of TB disease. The epidemiology of latent and active TB, BCG vaccination, diagnosis of LTBI, and treatment of children and human immunodeficiency virus (HIV)-infected persons with LTBI are covered in greater detail elsewhere.

LATENT TB INFECTION

LTBI is characterized by infection with M. tuberculosis without evidence of active TB disease, including the absence of clinical signs or symptoms and a normal chest radiograph. Although patients are classified as having either latent infection or active TB disease for the purpose of clinical management, the spectrum of M. tuberculosis infection is more accurately described as a dynamic continuum between exposure, infection, and disease rather than a binary process (13, 18, 19) (Fig. 1). After initial exposure, the mycobacteria may be fully eliminated by the host innate immune response, may rapidly progress to primary TB disease, or may be at least partially contained in a semidormant state of latency that may include periods of subclinical disease with undetected mycobacterial replication. While the majority of individuals maintain a persistent, life-long asymptomatic infection, an estimated 8 to 10% reactivate into active TB with overt clinical manifestations. Infection with M. tuberculosis can be detected only indirectly by measuring immune sensitivity to mycobacterial proteins with the tuberculin skin test (TST) or interferon gamma release assay (IGRA). However, these tests are not able to distinguish between persons whose immune systems may have fully eliminated the infection and those who remain infected and at risk of future TB disease (18–20).

FIGURE 1.

The spectrum of TB, from Mycobacterium tuberculosis infection to active (pulmonary) TB disease. Although TB disease can be viewed as a dynamic continuum from Mycobacterium tuberculosis infection to active infectious disease, patients are categorized as having either LTBI or active TB disease for simplicity in clinical and public health settings. Individuals can advance or reverse positions, depending on changes in host immunity and comorbidities. Exposure to M. tuberculosis can result in the elimination of the pathogen, either because of innate immune responses or because of acquired T cell immunity. Individuals who have eliminated the infection via innate immune responses or acquired immune response without T cell priming or memory (indicated by an asterisk) can have negative TST or IGRA results. Some individuals eliminate the pathogen but retain a strong memory T cell response and are positive on the TST or the IGRA. These individuals do not benefit from LTBI treatment. If the pathogen is not eliminated, bacteria persist in a quiescent or latent state that can be detected as positive TST or IGRA results; these tests elicit T cell responses against M. tuberculosis antigens. These patients would benefit from receiving one of the recommended LTBI preventive therapy regimens (mostly 6 to 9 months of isoniazid). Patients with subclinical TB might not report symptoms but are culture positive (but generally smear negative because of the low bacillary load). Patients with active TB disease experience symptoms such as cough, fever, and weight loss, and the diagnosis can usually be confirmed with sputum smear, culture and molecular tests. Patients with active TB disease might sometimes be negative on the TST or the IGRA because of anergy that is induced by the disease itself or immunosuppression caused by comorbid conditions, such as HIV infection or malnutrition. Individuals with subclinical or active TB disease should receive one of the recommended treatment regimens for active TB disease, which consist of an intensive phase with four drugs, followed by a longer continuation phase with two drugs. Reprinted from reference 13, with permission.

The risk of developing active TB disease is highest in the first 2 years after infection, declining thereafter for a lifetime risk of approximately 10% (8, 13, 21–23). The likelihood of progression to TB disease from primary infection or reactivation from LTBI varies based on characteristics of the infected person, including age and underlying medical conditions, with the highest risk found among young children (especially <4 years old), persons exposed in the preceding 1 to 2 years, and those with silicosis, HIV infection, or other immunosuppressive conditions (organ transplantation, treatment with tumor necrosis factor alpha [TNF-α] inhibitors, etc.) (8, 13, 24–29). However, this risk of progression to active TB can be significantly reduced by treating latent M. tuberculosis infection, with the greatest benefit being demonstrated among high-risk individuals (6, 8, 16, 21, 30–33).

TARGETED TESTING AND TREATMENT OF LTBI

The identification and treatment of persons with LTBI constitutes an essential component of TB elimination through two fundamental mechanisms (8, 10). The first is the individual clinical benefit conferred through the prevention of morbidity and mortality associated with active TB disease. The second benefit is gained at the population level through the prevention of spread of M. tuberculosis infection within the community and the associated reduction in health care spending. In order to accelerate the decline of TB in the United States, the WHO, the Institute of Medicine, the Centers for Disease Control and Prevention (CDC), the U.S. Public Health Task Force (USPHTF), and several professional societies, including the Infectious Diseases Society of America (IDSA), the American Thoracic Society (ATS), and the American Academy of Pediatrics (AAP), recommend a strategy of targeted tuberculin testing among high-risk groups who would benefit from treatment of LTBI (6, 8–10, 16, 17, 34). The 2000 ATS/CDC guidelines provided the first evidence-based recommendations for which risk groups should be tested, regimens for LTBI therapy, and strategies for monitoring and adherence during treatment (8). Screening of low-risk groups and testing for administrative purposes are discouraged, and treatment of high-risk persons diagnosed with LTBI is recommended regardless of age unless there are clinical contraindications. Recommendations that clinicians screen for LTBI only among high-risk populations and when treatment is feasible were reiterated by the CDC in 2005 and again in 2013 (10, 35), and a more comprehensive revision of the CDC’s targeted testing and treatment of LTBI recommendations that will incorporate the most recent evidence regarding the screening and management of high-risk persons is under development (T. Sterling, personal communication). Other developed countries, including the United Kingdom and Canada, follow a similar practice of screening high-risk persons for LTBI and providing treatment when indicated (36, 37). In 2015, the WHO issued guidelines for managing LTBI in all low-TB-burden countries to support the global strategy for TB elimination (3, 6).

Recently published epidemiologic data underscore the critical importance of the United States’ approach of concurrently prioritizing identification and treatment of persons with active TB while also promoting prevention through targeted testing and treatment of LTBI. First, recent National Health and Nutrition Examination Survey (NHANES) estimates indicate that the U.S. prevalence of LTBI as measured by both the TST and an IGRA has remained constant, at 4.4% in 2011 and 2012, compared to the prevalence measured by the TST alone in the 1999–2000 NHANES data (38). Second, the 2016 CDC surveillance report showed that the incidence of active TB disease in the United States has also leveled off, at approximately 3.0 cases per 100,000 persons during 2013 to 2015, following 2 decades of consistent annual decreases (39). To evaluate the potential scope and impact of treatment of LTBI at a population level, the Tuberculosis Epidemiologic Studies Consortium conducted a survey of clinics in the United States (19 sites) and Canada (2 sites) that initiated LTBI treatment for ≥10 patients in 2002 (40). Extrapolating study data to the entire U.S. population and using an estimated 20 to 60% treatment effectiveness and 5% lifetime risk of active TB without treatment, the study estimated that targeted screening and treatment of LTBI could prevent between 4,000 and 11,000 active TB cases in the United States and is thus an effective strategy to reduce the national burden of tuberculosis. In 2016, the United States Public Health Task Force also issued recommendations to conduct screening for LTBI among populations at increased risk (B recommendation, “offer or provide this service”) (17). The accompanying evidence report and systematic review support the screening and management of LTBI among high-risk groups by primary care providers, which should expand these activities and enhance progress towards TB elimination (16).

PRIORITIZATION OF GROUPS FOR TARGETED TESTING AND TREATMENT OF LTBI

Persons at high risk of TB and LTBI are categorized into two groups: those with increased risk of exposure to M. tuberculosis and those with medical conditions that increase the risk of progression to active TB once infected (8, 10, 35). Groups with increased risk of recent exposure have a high prevalence of LTBI and are thus likely high-yield targets for population-based screening programs as well as for individual screening for latent and active TB. Such groups include immigrants to the United States from high-prevalence countries, persons in recent contact with person with a case of infectious TB, persons who inject illicit drugs, and residents and employees of high-risk congregate settings where local epidemiology indicates a high rate of TB disease (e.g., correctional facilities, long-term care facilities, residential centers for patients with AIDS, and homeless shelters) (8, 10, 16, 17, 35). Although nosocomial transmission of TB has become much less common in developed countries due to the decreased incidence of TB and higher standards of infection control practiced in health care settings, health care workers may also have increased risk of potential exposure to M. tuberculosis, and guidelines for risk assessment and screening of this group have been published (41).

Persons with medical conditions associated with increased risk of progression from latent to active TB should also be evaluated for TB and LTBI and are a priority for treatment if found to be infected (8, 35). Medical conditions with the highest risk of reactivation of TB include HIV infection; diabetes; silicosis or exposure to silica dust; low body weight; chronic renal failure or hemodialysis; gastrectomy; jejunoileal bypass; cirrhosis of the liver; organ transplantation; anticancer chemotherapy; other immunosuppressive therapy (e.g., TNF-α antagonists); carcinoma of the head or neck; other neoplasms, such as lung cancer, lymphoma, and leukemia; and fibrotic changes on chest radiograph compatible with previous TB (8, 10, 35). Although indoor air pollution (42), excessive alcohol use (43), and smoking (44) have also been associated with increased risk of TB disease in some studies, the evidence for these factors is weak, and their high prevalence in the community makes them less useful markers for screening programs (13). TB control guidelines from both Canada and the United Kingdom include similar recommendations for targeted screening of high-risk groups, though there are a few minor differences specific to each country (36, 37). Clinical factors that have been recognized as associated with risk of progression from LTBI to active disease are listed in Table 1 (28).

TABLE 1.

Risk factors for the development of active TB among persons infected with Mycobacterium tuberculosis (28)^a

Risk factor	Estimated risk for TB relative to persons with no known risk factor
High risk (testing and treatment for LTBI recommended for all ages)
AIDS (not on anti-HIV therapy)	110–170
HIV (not on anti-HIV therapy)	50–110
Transplantation (related to immunosuppressive therapy)	20–74
Silicosis	30
Chronic renal failure requiring hemodialysis	10–25
Carcinoma of head and neck	16
Recent TB infection (<2 yrs)	15
Abnormal chest X ray—with upper lobe fibronodular disease typical of healed TB infection	6–19
TNF-α inhibitors	2–9
Moderate risk (testing and treatment for LTBI recommended if age < 65 yrs)
Treatment with glucocorticoids	5
Diabetes mellitus (all types)	2–4
Young age when infected (0–4 yrs)	2–5
Slightly increased risk (testing and treatment for LTBI recommended if age < 50 yrs)
Underweight (<90% ideal body weight; for most persons, this is a BMI of 20)	2–3
Cigarette smoker (1 pack/day)	2–3
Abnormal chest X ray—granuloma	2
Low risk (testing and treatment for LTBI recommended if age < 35 yrs)
Infected person, no known risk factor, normal chest X ray (“low-risk reactor”)	1
Very low risk (treatment of LTBI not usually recommended)
Person with positive two-step (“boosting”), no other known risk factor, and normal chest X ray	0.5

^a Modified from the work of Lobue and Menzies (140) and the CDC.

Both the WHO and the USPHTF recommendations provide further guidance regarding which high-risk groups should be considered a priority for LTBI testing and treatment (6, 16, 17). The WHO’s 2015 guidelines address countries classified as high or upper income and having a low TB burden (incidence of <100 per 100,000 per year) and recommend systematic screening and treatment of LTBI in people living with HIV, adult and child contacts of pulmonary TB cases, patients with silicosis, and persons who are initiating TNF-α inhibitors, preparing for organ or hematologic transplantation, or receiving dialysis. Systematic testing and treatment of LTBI are conditionally recommended for prisoners, health care workers, immigrants from high-TB-burden countries, homeless persons, and illicit drug users based on available resources and local TB epidemiology (6). In all countries, regardless of TB burden or national economic status, a high priority should be given to the identification and treatment of HIV-seropositive individuals infected with M. tuberculosis, since HIV infection is the most potent risk factor for the rapid progression of LTBI to TB disease and is associated with high TB incidence rates and a greater likelihood of disseminated and extrapulmonary disease (5, 6, 8, 10, 27, 45, 46). Risk factors in Table 1 are listed according to priority for treatment based on the relative risk for progression to active TB.

The 2016 USPHTF recommendations also support the testing of people born in, or who frequently travel to, countries where TB disease is common (e.g., Mexico, the Philippines, Vietnam, India, China, Haiti, and Guatemala) or other countries with high rates of TB (16, 17). Persons born in Canada, Australia, New Zealand, or Western and Northern European countries are not considered at high risk for TB infection, unless they have other TB risk factors. Testing is also recommended for persons who have lived in large group settings where TB is more common, such as homeless shelters or prisons and jails. Of note, the USPHTF did not review evidence pertaining to several populations with the highest risk and the greatest benefit for targeted screening and treatment, notably persons living with HIV, close contacts, patients with silicosis, and patients receiving immunosuppressive medications, because “screening in these populations may be considered standard care as part of disease management or indicated prior to the use of certain medications….” USPHTF recommendations also excluded persons with diabetes, citing currently insufficient evidence on screening for and treatment of LTBI in persons with this condition. Additional support of targeted testing and treatment of LTBI has been provided by a cost-effectiveness study indicating that progress towards TB elimination can be maximized by prioritizing LTBI screening of close contacts, persons infected with HIV, and persons born in high-TB-burden countries regardless of their time living in the United States (46).

A decision to test should be considered a decision to treat if LTBI is diagnosed (8, 35–37). For this reason, screening of low-risk persons is widely discouraged, as the risk-benefit ratio may not favor treatment among persons likely to have a false-negative TST or IGRA result. However, routine administrative screening for LTBI is necessary for low-risk persons at baseline prior to employment at a high-risk worksite such as a hospital, long-term care facility, or correctional facility to enable a distinction between existing and potential future infections (8, 41). In these situations, low-risk persons with evidence of LTBI not included in the preceding risk groups may also be considered for therapy to prevent potential reactivation of active TB in a high-risk setting. There may also be other situations in which a health care provider may be asked to test individuals who are not necessarily regarded as high risk, such as day care center workers, teachers, and U.S.-born students. A risk assessment should be conducted for all individuals to determine if testing for LTBI is indicated, and treatment should be provided to those who are likely to be infected and can safely complete a full course of preventive treatment (35). A sample risk assessment tool can be found at http://www.cdc.gov/tb/publications/ltbi/pdf/targetedltbi.pdf. For persons found to have low risk of TB (i.e., no TB risk factors and no symptoms of TB disease), it is preferable to provide documentation of the negative assessment and discourage further testing given the risk of false-positive test results and unnecessary LTBI therapy.

Local public health programs may also conduct targeted testing and recommend LTBI treatment among other groups defined as high risk based on the incidence of TB, the prevalence of LTBI, and the likelihood of a population-level benefit resulting from such an intervention. This may include medically underserved and low-income minority groups, such as U.S.-born Hispanics or African Americans who live or work in a community with a high proportion of persons who have a traditional TB risk factor (8, 10). As an example, six unrelated TB cases (i.e., not exposed to each other) occurred over a 2-year period at a poultry processing facility in rural Tennessee, where there was a high prevalence of LTBI and the majority of workers were either U.S.- or foreign-born Hispanics (unpublished data). The Tennessee Department of Health implemented a successful tuberculin testing and treatment program among employees of the plant, which resulted in screening of several thousand employees, diagnosis of three additional persons with active TB, and diagnosis and treatment of 844 persons with LTBI (79% of whom completed treatment) (C. Haley, personal communication).

TREATMENT OF LTBI

Prior to the initiation of treatment, all persons with evidence of LTBI should be evaluated for the presence of pulmonary and extrapulmonary TB disease. This includes a thorough review for TB symptoms, a clinical examination, and a chest radiograph (6, 8, 35, 47). If the chest radiograph is abnormal or pulmonary symptoms are present, a sputum sample for smear and culture of acid-fast bacilli should also be obtained, and further evaluation and management of suspected TB should be conducted according to current guidelines (20, 47). If TB has been definitively ruled out, treatment for LTBI should be provided to infected persons who have not already received an adequate course of therapy (6, 8, 35). Patients with LTBI should be evaluated for preexisting medical conditions that may increase the risk of adverse events during treatment, in particular viral hepatitis, pregnancy (including the postpartum period), regular use of alcohol and other medications with hepatotoxic potential, and previous reactions to anti-TB medications (Fig. 2). Testing for coinfection with HIV should be offered as part of the evaluation of LTBI treatment candidates, as HIV coinfection has significant implications for diagnosis, treatment, and clinical outcomes. Treatment of LTBI is recommended for the majority of individuals with evidence of LTBI, but the individual risk-benefit ratio of treatment and the patient’s willingness to complete a full course of therapy must be considered. Additional clinical and laboratory monitoring may be indicated if certain comorbid conditions exist, as is described later (Fig. 3). Prior to therapy, patients should also be evaluated for concomitant use of other medications that may cause drug-drug interactions with the LTBI regimen, educated about potential adverse effects of treatment, and counseled regarding the importance of adherence to the recommended course of therapy.

FIGURE 2.

LTBI pretreatment clinical evaluation and counseling. Dotted lines signify management according to physician’s discretion. INR, international normalized ratio; PTT, partial thromboplastin time. DILI, drug-induced liver injury. Reprinted with permission of the American Thoracic Society (87).

FIGURE 3.

Monitoring for hepatotoxicity during LTBI treatment. Dotted lines signify management according to physician’s discretion. ALT, alanine aminotransferase; AST, aspartate aminotransferase; BeAg, Hepatitis B e antigen; Bili, bilirubin; HAV, hepatitis A virus; HepBcAb, hepatitis B core antibody; HepBsAg, hepatitis B virus surface antigen; ULN, upper limit of normal. Reprinted with permission of the American Thoracic Society (87).

The concept of using a single anti-TB agent to prevent active TB originated during the 1950s, when Edith Lincoln noted that children hospitalized at Bellevue Hospital in New York City no longer experienced complications of their primary TB following treatment with isoniazid (22). At her suggestion, the U.S. Public Health Service organized a multiclinic controlled trial among 2,750 children with asymptomatic primary TB or a recent tuberculin conversion. Preventive therapy with isoniazid proved to be remarkably effective, producing a 94% reduction in the development of TB during a year of LTBI treatment and a 70% reduction over the following 9-year period. Several other placebo-controlled trials found that isoniazid was effective for treatment of infected contacts of TB patients and other persons at high risk (e.g., those with radiographic evidence of prior untreated TB, inmates of mental health institutions, and native Alaskans) (22, 30, 48–52). More than five decades after the United States first adopted the treatment of LTBI as a primary strategy to prevent TB disease (53), isoniazid monotherapy remains the most widely used LTBI regimen globally (5, 6, 8, 16). Table 2 includes a list of studies evaluating the efficacy of isoniazid for LTBI treatment. Despite the potential effectiveness of isoniazid for reducing the incidence of TB, an estimated 90% for adherent patients compared to placebo, concerns of toxicity have limited treatment initiation and the completion rates for 6 to 9 months of isoniazid are low, thus substantially reducing the actual benefit of this regimen (8, 16, 30, 54–57).

TABLE 2.

Placebo-controlled studies of isoniazid efficacy for treatment of LTBI^a

Study(ies)	Yr	Location(s)	Population	Duration of INHb (mo)	Reduction in TB rates
Ferebee (22), Mount and Ferebee (184)	1956–1957	United States, multiple sites	Household contacts	12	68% reduction in first 15 mo of follow-up; 60% reduction after 10 yrsc
Ferebee (22), Mount and Ferebee (184)	1957–1960	United States, multiple sites	Household contacts	12	76% reduction in first 15 mo; 60% reduction after 10 yrsc
Ferebee (22), Ferebee et al. (51)	1957–1960	United States, multiple sites	Residents of mental institutions	12	88% reduction in first 15 mo; 62% reduction after 10 yrs
Comstock et al. (48)	1957–1964	Alaska	Native Alaskans	12	59% reduction after 43–76 mo
International Union Against Tuberculosis Committee on Prophylaxis (30)	Started 1969	Eastern Europe	Person with fibrotic pulmonary lesions (inactive TB)	3, 6, 12	After 5 yrs of follow-up in all randomized: 21% reduction for 3 mo of INH 65% reduction for 6 mo of INH 75% reduction for 12 mo of INH
					After 5 yrs of follow-up in completer/compliers: 30% reduction for 3 mo of INH 69% reduction for 6 mo of INH 93% reduction for 12 mo of INH
Pape et al. (75)	1983–1989	Haiti	HIV-infected persons	12	71% reduction after 60 mo
Whalen et al. (77)	1993–1995	Uganda	HIV-infected persons	6	For TST-positive persons: 67% reduction after 15 mo
					In anergic persons: no reduction

^a Reprinted from Respirology with permission of the publisher (140).

^b INH, isoniazid.

^c Sixty percent reduction for 10-year follow-up was calculated from aggregate results of first two studies listed as reported in references 8 and 50.

Shorter rifampin-based regimens have been evaluated for their potential to overcome the low treatment completion rates and perceived risk of toxicity associated with 6 to 9 months of isoniazid (8, 32, 58, 59). Rifamycin antibiotics have greater potency against dormant and semidormant M. tuberculosis organisms that characterize latent infection (60, 61). Several studies of short-course rifampin-based LTBI therapy have demonstrated efficacy equal to or greater than that of the longer isoniazid regimen (32, 58, 62–65). In 2000, ATS/CDC guidelines (also endorsed by the IDSA and the AAP) provided evidence-based recommendations for the use of two short-course LTBI regimens: rifampin monotherapy for 4 months and a 2-month course of rifampin combined with pyrazinamide (8). Although earlier studies conducted among HIV-infected persons found that rifampin plus pyrazinamide for 2 months was both safe and effective, subsequent cases of severe and fatal hepatotoxicity (also referred to as drug-induced liver injury) were reported when this regimen was widely incorporated into clinical use in the general population (62, 63, 65–71). In 2003, the CDC and ATS updated the guidelines for LTBI treatment with recommendations against the use of rifampin plus pyrazinamide for both HIV-seropositive and HIV-seronegative persons (66). Because the safety, tolerability, and adherence rates associated with 4 months of rifampin have been deemed favorable to date, this regimen remains a recommended LTBI regimen(6, 16, 17, 36). Many providers now routinely use the 4-month rifampin regimen as an acceptable alternative LTBI treatment regimen, particularly for persons assessed as having a high risk for progression to active TB but who are unlikely to complete a longer (6- to 9-month) course of isoniazid treatment. In 2011, the CDC issued an additional recommendation supporting the use of a combination regimen of isoniazid plus rifapentine administered weekly for 12 weeks as directly observed therapy (DOT) based on evidence from three randomized controlled trials demonstrating good efficacy, tolerability, and treatment completion rates (59, 72, 73). A 3-month regimen of isoniazid plus rifampin has also been recommended for many years in Canadian and British TB control guidelines but has not been widely used in the United States since it was not included in the 2000 CDC/ATS guidelines (36, 37). Specific LTBI treatment regimens are described below and summarized in Table 3 according to their efficacies and potential adverse effects. Current guidelines for several different institutions, including the WHO, the CDC, the United Kingdom’s National Institute for Health and Care, and the Canadian TB Standards, are listed in Table 4.

TABLE 3.

Regimens for latent TB treatment, according to pooled efficacy, risk of hepatotoxicity, adverse events, and drug interactions

Drug	Regimen dosage	OR (95% CI) fora:			Adverse events
		Efficacy vs placebo	Efficacy vs 6 mo of isoniazid	Hepatotoxicity vs 6 mo of isoniazid
Isoniazid alone for 6 mo or 9 mo	Adults, 5 mg/kg; children, 10 mg/kg (maximum, 300 mg)	6-mo regimen, 0.61 (0.48–0.77); 9-mo regimen, 0.39 (0.19–0.83)	Not applicable for 6-mo regimen, and not available for 9-mo regimen	Not applicable for 6-mo regimen and not available for 9-mo regimen	Drug-induced liver injury, nausea, vomiting, abdominal pain, rash, peripheral neuropathy, dizziness, drowsiness, and seizure
Rifampin alone for 3 to 4 mo	Adults, 10 mg/kg; children, 10 mg/kg (maximum if <45 kg, 450 mg; maximum if >45 kg, 600 mg)	0.48 (0.26–0.87)	0.78 (0.41–1.46)	0.03 (0.00–0.48)	Influenza-like syndrome, rash, drug-induced liver injury, anorexia, nausea, abdominal pain, neutropenia, thrombocytopenia, and renal reactions (e.g., acute tubular necrosis and interstitial nephritis)
Isoniazid plus rifampin for 3 to 4 mo	Adults, 10 mg/kg; children, 10 mg/kg (maximum if <45 kg, 450 mg; maximum if >45 kg, 600 mg)	0.52 (0.33–0.84)	0.89 (0.65–1.23)	0.89 (0.52–1.55)	Influenza-like syndrome, rash, drug-induced liver injury, anorexia, nausea, abdominal pain, neutropenia, thrombocytopenia, and renal reactions (e.g., acute tubular necrosis and interstitial nephritis)
Weekly rifapentine plus isoniazid for 3 mo	Adults and children: rifapentine, 15–30 mg/kg (maximum, 900 mg)b; isoniazid, 15 mg/kg (maximum, 900 mg)	Not available	0.44 (0.18–1.07)c	0.16 (0.10–0.27)c	Hypersensitivity reactions, petechial rash, drug-induced liver injury, anorexia, nausea, abdominal pain, and hypotensive reactions

^a Data on efficacy and hepatotoxicity are from the work of Stagg et al. (33). Reprinted from New England Journal of Medicine, with permission of the publisher (5).

^b The following incremental adjustments are required for persons weighing less than 50 kg: 10.0 to 14.0 kg, 300 mg; 14.1 to 25.0 kg, 450 mg; 25.1 to 32.0 kg, 600 mg; and 32.1 to 49.9 kg, 750 mg.

^c The comparison is with 9 months of isoniazid.

TABLE 4.

Current guidelines for the treatment of latent tuberculosis infection^a

Institution (year)	Recommended treatment for LTBI that is presumed to be drug susceptible	Recommended treatment for LTBI that is presumed to be MDR
WHO (2014) (6)	6INH or 9INH or	Strict clinical observation for 2 yrs is preferred over provision of preventive therapy. Benefits of preventive therapy may outweigh harm for children <5 yrs of age. If preventive therapy is given, monitor for acquired drug resistance.
	3 mo of wkly RPT plus INH under DOT or
	3–4 mo of INH + RIF or
	3–4 mo of RIF
CDC (2000) (8)	9INH daily or twice wkly	Not stated
	6INH daily or twice wkly
	4RIF dailyb
	3 mo of wkly RPT + INH under DOT (if HIV positive, 9 mo is preferred)
UK NICEc guidelines (2016) (37)	Close contacts aged <65 yrs or HIV positive: either 6INH (with pyridoxine) or 3 mo of INH + RIF (with pyridoxine)	Not stated
	Close contacts aged <35 yrs for whom hepatotoxicity is a concern: 3 mo of INH + RIF (with pyridoxine)
	For people living with HIV and for transplant recipients: 6INH (with pyridoxine)
Canadian TB Standards (2013) (37)	9 mo of INH (first choice)	INH-R: treat contacts of patients with INH resistance with 4RIF. RIF-R: contacts with RIF resistance with 9INH.MDR-TB: 9LFX or 9MOX with close monitoring.
	Alternative regimens:6INH3–4 of mo INH + RIF3 mo of wkly INH/RPT under DOT; intermittent regimens only recommended when daily regimens cannot be used (6-9INH twice wkly; 3 mo of INH + RIF twice wkly; under DOT)d

^a This table has been adapted from the work of Fox et al. (185). Note that the European Union Standards for TB Care are not included as they have not been updated since 2012, and the WHO guidelines apply to these countries. The duration of treatment is indicated by the number of months followed by the drug name (e.g., 6INH is 6 months of isoniazid). Treatment is given daily under self-administered therapy unless otherwise stated. DOT, directly observed therapy; INH, isoniazid; RIF, rifampin; RPT, rifapentine; LFX, levofloxacin; MOX, moxifloxacin.

^b ATS guidelines previously recommended rifampin with pyrazinamide for 2 months; however, this is no longer recommended on account of high rates of hepatotoxicity after implementation (53).

^c The UK National Institute for Health and Care Excellence (NICE) guidelines recommend that preventive therapy be offered to 35- to 65-year-olds if hepatotoxicity is not a concern.

^d In pregnancy, deferral of preventive therapy until 3 months after delivery is recommended unless there is a very high risk of disease (e.g., HIV or recent infection). Isoniazid (with pyridoxine) and rifampin are considered safe in pregnancy (45).

Isoniazid

More than 20 randomized, placebo-controlled trials of LTBI treatment with isoniazid involving more than 100,000 subjects have been conducted (Table 2) (22, 52, 74). The combined average reduction in TB reported in these studies was 60% during the period of observation, being somewhat higher during the year of treatment. These results were based on the total study populations treated, regardless of how regularly medication was taken. Among these trials, the five reporting less than 50% effectiveness included one that used small doses of isoniazid, one in which compliance was poor, and one that included patients who had undergone previous isoniazid therapy, a group now known not to benefit from additional treatment (22). When analyses were limited to participants who took their medication for most of the treatment year, efficacy approximated 90% (30). Protection also appears to be long-lasting, being demonstrable nearly 20 years after initiation of treatment (21).

The optimal duration of isoniazid treatment was addressed in a large International Union Against Tuberculosis trial conducted in six Eastern European countries among persons with untreated inactive TB (30). Regimens of daily isoniazid for 3, 6, and 12 months were tested against daily placebo for the same durations. The results of 5 years’ observation of the total population showed that treatment for 12 months resulted in a 75% reduction in TB, compared with reductions of 65% in those treated for 6 months and 21% for those treated for only 3 months. When the analysis was restricted to those who took at least 80% of the prescribed regimen, efficacy increased to 93% for the 12-month group but improved only slightly for those treated for 6 and 3 months. In the U.S. Public Health Service trials among household contacts and Alaskan villagers, the optimal duration of treatment appeared to be 9 to 10 months (21, 22, 48, 74) (Fig. 4). In the contact trial, irregular treatment was still effective as long as 80% of the 12-month dose (i.e., 9 to 10 months) was taken within a reasonable time (48). Although the effectiveness of 9 months of isoniazid has never been directly compared with a 6- or 12-month regimen, the United States and Canada currently recommend isoniazid given for a treatment duration of 9 months as the preferred option for HIV-negative individuals with LTBI (8, 36). Isoniazid taken for only 6 months is considered acceptable, though less effective, if it allows a higher likelihood of treatment completion. A 6-month duration of isoniazid treatment is favored in British TB control guidelines (37), and the WHO recommendations support using either 6 or 9 months of isoniazid (6).

FIGURE 4.

TB case rates in the Bethel Isoniazid Studies population according to the number of months that isoniazid was taken in the combined programs. Dots represent observed values; dashed line, the calculated curve (y = a + b/x); and dotted lines, the calculated values based on the first four and the last five observations (y = a + bx). Reprinted with permission of the International Union Against Tuberculosis and Lung Disease. © The Union (21).

Most of the studies referenced above that evaluated the use of isoniazid monotherapy for LTBI treatment were conducted prior to the 1960s and therefore did not include HIV-seropositive persons. A number of randomized placebo-controlled trials have subsequently been conducted to determine if treatment of LTBI is effective in preventing TB disease in HIV-infected individuals (63, 65, 75–78), as well as several systemic reviews and meta-analyses of these trials (24, 79, 80). There is now strong evidence that isoniazid treatment of LTBI reduces the risk of active TB in HIV-positive individuals, especially in those with a positive TST (79, 80). Because LTBI treatment and antiretroviral therapy act independently to decrease the risk of TB disease, use of both interventions is recommended for those who have LTBI and are HIV infected (81–86). Additional details regarding the treatment of LTBI in HIV-infected persons are included elsewhere.

For HIV-seronegative and -seropositive adults, the recommended dose of daily isoniazid is 5 mg/kg of body weight, not to exceed 300 mg (Table 3). For children, the dose is 10 to 20 mg/kg, with a maximum of 300 mg. Isoniazid can also be administered twice weekly, at doses of 15 mg/kg in adults and 20 to 40 mg/kg in children, for a maximum of 900 mg in either group, but must be given as DOT. Overall, isoniazid is one of the least toxic of the anti-TB drugs; most of the reactions are mild and transient, including dose-related peripheral neurotoxicity, central nervous system effects (irritability, dysphoria, seizures, impaired concentration, etc.), hypersensitivity reactions, lupus-like syndrome, and mild gastrointestinal discomfort (Table 4) (8). Neuropathy is more common among persons who are already predisposed due to conditions such as HIV infection, diabetes, renal failure, poor nutrition, and alcoholism, as well as among women who are pregnant or breastfeeding. Supplementation with pyridoxine (25 mg/day) is recommended for these persons during treatment with isoniazid (8). Elevations in transaminase values (alanine aminotransferase [ALT] and aspartate aminotransferase [AST]) up to five times the upper limit of normal (ULN) occur in 10 to 20% of persons receiving isoniazid monotherapy, and levels usually return to normal even with continued administration of the drug (87).

The side effect of principle concern during treatment with isoniazid is hepatotoxicity. Although this was rarely reported in early studies using isoniazid to treat LTBI, the potential for this drug to cause both asymptomatic transaminase elevation and clinically significant hepatitis, including death, was initially recognized in the late 1960s and 1970s (87–90). Subsequent studies from the 1970s to 1990s reported much lower rates of isoniazid-related hospitalization and death, which has been attributed to careful patient selection, education, and active monitoring for adverse reactions during treatment (66, 87–89). More recent larger reviews have reported a rate of significant transaminase elevation, 0.1 to 0.56% (87, 91–93). In one of these studies, over 11,000 patients who started isoniazid preventive therapy between 1989 and 1995 in a Seattle, WA, public health TB clinic reported very low (0.1%) rates of hepatotoxic reactions (elevated transaminases more than five times the ULN), though the risk did increase with age (P = 0.02) (93). Of note, rates of transaminase elevation could have been underestimated, since routine monitoring of levels was not performed among asymptomatic patients, and rates of clinically significant hepatotoxicity were measured based on all persons initiating treatment rather than on those actually taking medication. A subsequent study conducted in San Diego, CA, reported a 0.3% rate of transaminase elevation among 3,788 LTBI patients treated with isoniazid (defined as three times the ULN for symptomatic patients and five times the ULN for asymptomatic persons) (92). Another observational study conducted in Memphis, TN, from 1996 to 2003 reported significant AST elevations among 19 of 3,377 LTBI patients taking isoniazid monotherapy, only 1 of whom was symptomatic (91).

Even in the context of guidelines for improved patient selection and clinical monitoring during therapy (Fig. 2 and 3), significant hepatotoxicity and deaths during isoniazid treatment have been reported, especially among persons who continued to take the drug after symptoms of hepatitis had appeared. In January 2004, the CDC began a national passive surveillance system to quantify the frequency of severe adverse events associated with isoniazid for LTBI treatment and to characterize the clinical features of affected patients (57). Between 2004 and 2008, 15 adults and 2 children (aged 11 and 14 years) who received isoniazid therapy experienced severe idiosyncratic drug-induced hepatotoxicity; 5 required liver transplants (including 1 child), and 5 adults died. These findings, though uncommon, underscore the potential for severe adverse events during isoniazid therapy and emphasize the importance of sustained clinical monitoring throughout LTBI treatment in accordance with ATS/CDC recommendations (Fig. 3) (8, 57, 87). Predictors of hepatotoxicity during isoniazid therapy include older age and preexisting liver disease, particularly from hepatitis C virus (HCV) infection, concomitant use of other hepatotoxic medications, prior isoniazid-related hepatotoxicity, and regular alcohol consumption (87). The risk of infected persons developing TB must be weighed against the potential risk of developing this adverse event during preventive therapy, and appropriate monitoring must be conducted according to current guidelines (6, 8, 34, 36, 37). For tuberculin reactors with no additional risk factors, the balance is most strongly in favor of preventive treatment among children and young adults (6, 94). For persons with additional TB risk factors, the benefit-to-risk ratio is increased at all ages.

In addition to the actual and perceived risk of toxicity associated with isoniazid, the effectiveness of this therapy for preventing active TB among infected persons is significantly limited by poor adherence (6, 54, 56, 95, 96). Reported treatment completion rates associated with 6 to 9 months of isoniazid are typically around 50% but can be much lower for specific high-risk groups, such as inner city residents, jail inmates, homeless persons, and injection drug users (97–103). Even in a research setting, completion of isoniazid therapy reached only 69% (58). Although adherence has been improved in some situations using the twice-weekly regimen given as DOT, the data regarding the effectiveness of intermittent isoniazid compared to daily treatment are weak (54). Thus, the likelihood of completing the longer duration of isoniazid should also be considered when selecting a treatment regimen for persons with LTBI, particularly for those in the risk groups listed above.

Rifampin Monotherapy

Rifampin-based regimens were first considered promising for shorter LTBI treatment because of the potent bactericidal activity of rifamycins against Mycobacterium tuberculosis and findings from animal model studies suggesting that rifampin alone or in combination could be at least as effective as isoniazid monotherapy (60, 99, 104, 105). Since recommendations for the use of rifampin plus pyrazinamide were withdrawn, daily self-administered rifampin monotherapy has become widely accepted as an alternative LTBI treatment regimen, especially for persons who are infected with an isoniazid-resistant strain of M. tuberculosis and those who are intolerant of or unlikely to complete a longer course of isoniazid (6, 8, 32, 33, 106).

In contrast to the extensive experience evaluating the efficacy of isoniazid for preventing progression to active TB, only one published randomized clinical trial has evaluated rifampin monotherapy for patients with LTBI (64) (Table 3). From 1981 to 1987, a cohort of older Chinese men with silicosis and LTBI was randomly assigned to receive either placebo, rifampin for 3 months, isoniazid for 6 months, or isoniazid plus rifampin for 3 months. While all treatment groups had a reduced cumulative incidence of active TB over the 5-year follow-up period compared to the placebo group, rifampin monotherapy was more effective at preventing active TB than the isoniazid-rifampin group and isoniazid monotherapy regimens (Fig. 5). The effectiveness of 3 months of rifampin compared to placebo was calculated at 50% among persons who completed the 5-year study and at 46% among all persons who initiated treatment. TB rates were relatively high in this cohort of patients with silicosis, a potent facilitator for the progression from LTBI to active TB, experts concluded that the use of 4 months of rifampin would be more prudent than 3 months (8). Several small nonrandomized clinical studies also suggest that the efficacy of 4 months of rifampin is at least equivalent, if not superior, to that of 6 months of isoniazid (107, 108). In one study of homeless persons who developed a TST conversion during an epidemic of isoniazid-resistant TB, no patients treated with rifampin monotherapy for an average of 6 months developed TB disease, compared to 8.6% of untreated persons (107). In a second observational study of 157 adolescents who developed skin test conversions after exposure to a source case with isoniazid-resistant M. tuberculosis, none developed active TB during the 2 years following completion of 6 months of rifampin therapy (108). A systematic review and network meta-analysis conducted by Stagg and colleagues to guide the 2014 WHO LTBI guidelines found through indirect comparison that 3 to 4 months of rifampin monotherapy ranked highly for efficacy compared to other LTBI regimens, including isoniazid monotherapy (33). Currently, a large-scale multisite international randomized trial is under way to compare the effectiveness of rifampin monotherapy for 4 months and isoniazid for 9 months in preventing active TB (https://clinicaltrials.gov/ct2/show/NCT00931736).

FIGURE 5.

Effectiveness of three regimens for treatment of LTBI in elderly Chinese men with silicosis. Based on 503 patients at 1 year, 474 at 2 years, 418 at 3 years, 367 at 4 years, and 304 at 5 years who received their regimen without known interruption. The x axis shows the months from start of the LTBI treatment regimen. The y axis shows the percentage of patients who developed TB disease. HR3, isoniazid and rifampin for 3 months; H6, isoniazid for 6 months; Pl, placebo; R3, rifampin for 3 months (140). Reprinted with permission of the American Thoracic Society (64).

Reported treatment completion rates for 4 months of self-administered rifampin have been consistently high, ranging from 60 to 91% (32, 109–114). Rifampin also appears to be well tolerated and has been associated with low rates of drug-induced hepatotoxicity (32, 64, 87, 107, 108, 110, 112, 114). Two open-label randomized controlled trials conducted in university-based TB clinics in Canada, Brazil, and Saudi Arabia directly compared the rates of both adverse events and treatment completion using 4 months of rifampin and 9 months of isoniazid (112, 113). In both studies, treatment of LTBI using the rifampin regimen resulted in better adherence and fewer serious adverse events than with isoniazid. The systematic review conducted by Stagg and colleagues also reported that 3 to 4 months of rifampin monotherapy had fewer hepatotoxic events than did 6- and 9-month isoniazid regimens in an indirect comparison (33). Superior treatment completion rates, good tolerability, and lower rates of hepatotoxicity have also been found in several observational studies comparing 4 months of rifampin to 9 months of isoniazid (107, 111, 114, 115).

Other common side effects associated with rifampin include mild cutaneous reactions and gastrointestinal side effects such as nausea, anorexia, and abdominal pain (Table 3) (8, 47). More severe hypersensitivity reactions and immunologic reactions, such as thrombocytopenia and hemolytic anemia, renal effects (including acute tubular necrosis and interstitial nephritis), and a flu-like syndrome, are rare, occurring more often when therapy is intermittent, and may be associated with higher doses (116–120). Additionally, rifamycins induce cytochrome P450 enzymes, particularly CYP3A4, which, in turn, increase the metabolism of a wide variety of drugs, including warfarin, prednisone, digitoxin, quinidine, ketoconazole, itraconazole, propranolol, clofibrate, sulfonylureas, phenytoin, HIV protease inhibitors, and HIV nonnucleoside reverse transcriptase inhibitors (47). Regular measurements of serum drug concentrations of these medications should therefore be conducted during rifampin therapy. Additionally, patients should be warned that rifamycins cause orange discoloration of bodily fluids (sputum, urine, sweat, and tears) and permanently stain soft contacts and clothing.

Rifampin for LTBI therapy is recommended at a dose of 10 mg/kg for adults, not to exceed 600 mg, for a duration of 4 months. For children, the dose is 10 to 20 mg/kg up to a maximum of 600 mg given daily, and the AAP has decreased the recommended duration of treatment in the United States from 6 months to 4 months (121). As with isoniazid, it is important to exclude active TB before initiating rifampin monotherapy, particularly in HIV-infected persons. There is concern that inadvertent treatment of active TB with rifampin monotherapy can result in the development of resistance to rifampin. However, spontaneous chromosomal mutations of M. tuberculosis leading to rifampin resistance are 2 to 3 orders of magnitude less frequent than for isoniazid resistance, and there is little evidence that induced resistance occurs when either isoniazid or rifampin has been used alone in preventive treatment (22, 28, 32, 64, 99, 122). In a systematic review and meta-analysis including six randomized controlled trials of rifamycin-containing regimens for LTBI treatment that reported drug resistance, no statistically significant increased risk of rifamycin resistance was detected after LTBI treatment with rifamycin-containing regimens compared to non-rifamycin-containing regimens or a placebo (122). Nonetheless, particular caution should be used with HIV-infected persons, in whom active TB is often difficult to exclude and who may have a larger bacterial burden and thus an increased risk of developing rifampin-resistant TB disease (74, 111, 123). In addition, drug interactions are a significant concern in this population, since rifampin interacts with many antiretrovirals as well as other antimicrobial medications used for other concurrent infections (47, 85, 86). For this reason, rifampin monotherapy is not often used for LTBI patients with HIV infection. More comprehensive information regarding the treatment of LTBI in HIV-infected persons can be found elsewhere.

Isoniazid plus Rifapentine

Since 2011, the CDC has recommended the use of a newer LTBI regimen based on rifapentine, a long-acting rifamycin derivative, combined with isoniazid once weekly for 12 weeks provided as DOT (59). Murine models of LTBI first demonstrated that weekly isoniazid-rifapentine was likely to be effective for preventing reactivation of TB (124–126). Subsequently, three randomized controlled trials have shown that this regimen is at least as effective as isoniazid monotherapy and has superior rates of treatment completion (58, 72, 73). One small study in Brazil found that the incidence rate of developing active TB after completing 12 weeks of directly observed isoniazid-rifapentine was substantially lower (1.46%) than published rates for untreated household contacts (7 to 9%) (72). Another trial conducted in South Africa randomized adults with HIV infection and a positive TST who were not taking antiretroviral therapy to either isoniazid-rifapentine weekly for 12 weeks by DOT, isoniazid-rifampin twice weekly for 12 weeks by DOT, isoniazid self-administered daily for up to 6 years (continuous isoniazid), or isoniazid self-adminstered daily for only 6 months (control group) (73). TB incidence rates were not significantly different among the treatment groups, but treatment completion was superior in both the isoniazid-rifapentine group and in the isoniazid-rifampin group. Additionally, no acquired rifamycin resistance was detected among the few patients who developed culture-positive TB.

In the PREVENT TB trial conducted by the Tuberculosis Trials Consortium in Brazil, Canada, Spain, and the United States, 8,053 patients with LTBI were randomized to either directly observed 12-dose, once-weekly isoniazid-rifapentine (900 mg/900 mg) or self-administered isoniazid for 9 months (58). Among 7,731 participants 2 years of age or older who could be monitored for 33 months, the isoniazid-rifapentine regimen was found to be noninferior to 9 months of isoniazid in preventing active TB among high-risk individuals with LTBI. Treatment completion rates were significantly higher with this regimen compared to isoniazid (82% versus 69%, respectively; P < 0.01), and the rate of treatment discontinuation due to hepatotoxicity was lower (0.3% versus 2.0%, respectively; P < 0.01). However, more patients taking isoniazid-rifapentine than isoniazid had to permanently stop treatment as a result of serious adverse effects (4.9% versus 3.7%; P < 0.01), inclulding a hypersensitivity reaction (2.9% versus 0.4%; P < 0.01).

Although not reported when intermittent isoniazid and rifapentine were used concurrently to treat active TB disease, this hypersensitivity or “flu-like syndrome” consisting of fever, chills, fatigue, malaise, headache, myalgia, and arthralgia has been reported for patients taking intermittent rifampin therapy at higher doses (118, 127–129). This reaction most often begins within 2 h after ingestion of a rifampin dose and lasts up to 8 h afterwards; in some cases (excluding anaphylaxis), rifampin has been safely rechallenged at lower daily doses. In contrast, hypersensitivity reactions due to other medications that are likely immune-mediated can occur as late as 2 months after therapy is initiated, are not typically related to dose, and may intensify with continuation or a rechallenge of treatment. A similar syndrome has also been reported for patients taking isoniazid, with symptoms including rash, loss of appetite, weight loss, myalgia, arthralgia, fatigue, malaise, headache, fever, red eyes, leukocytosis, and hypotension (127, 129–131).

To better characterize the severity and risk factors associated with clinically significant systemic drug reactions (SDR) that occurred during treatment with isoniazid-rifapentine (including hypersensitivity), the PREVENT TB study investigators evaluated all participants who reported reactions with systemic manifestations (127). Among 7,552 persons who took at least one dose of study drug, 153 reported a SDR, including 3.5% of those taking isoniazid-rifapentine and 0.4% of those taking isoniazid (P < 0.001). Of the 138 patients with an SDR during isoniazid-rifapentine therapy, 83 (63%) developed a flu-like syndrome and 23 (17%) developed a cutaneous reaction, whereas in the isoniazid arm there were 15 events, including 9 (60%) cutaneous and 2 (13%) flu-like. Among persons taking isoniazid-rifapentine, SDR occurred after a median of 3 doses (interquartile range [IQR], 2 to 5 doses), began a median of 4 h after ingestion of a dose (IQR, 1.0 to 8.0 h), and resolved within a median of 24 h (IQR, 12 to 28 h). Although the underlying mechanism of SDR during isoniazid-rifapentine therapy remains unclear, most were flu-like and had features differing from typical hypersensitivity or immune-mediated drug reactions. The risk of SDR was found to be increased among persons of white race, female sex, older age, and lower body mass index. Considering all participants, SDR were rare (0.3%) during treatment and were associated with white non-Hispanic race/ethnicity (adjusted odds ratio [aOR], 5.4; 95% confidence interval [CI], 1.8 to 16.3) and concomitant use of nonstudy medications (aOR, 5.9; 95% CI, 1.3 to 27.1). Among 73 subjects rechallenged with a study drug, 36 of 51 (71%) tolerated rifapentine, 3 of 20 (15%) tolerated isoniazid, and 2 tolerated both isoniazid and rifapentine, though none completed the study regimen. There were no deaths or permanent sequelae as a result of any treatment.

Other side effects that may occur during treatment with isoniazid-rifapentine are similar to those for rifampin and isoniazid when administered as monotherapy (see sections for isoniazid and for rifampin above). With regard to hepatotoxicity, the 12-week isoniazid-rifapentine regimen has been found to be safer than isoniazid in published trials (58, 72, 73, 132). In the PREVENT TB study, the risk of hepatotoxicity was 4-fold higher among persons receiving isoniazid than among those receiving isoniazid-rifapentine (1.9% and 0.4%, respectively; relative risk, 4.42; 95% CI, 2.52 to 7.75) (58, 132). Hepatotoxicity was associated with increasing age, female sex, white race, non-Hispanic ethnicity, decreased body mass index, elevated baseline AST, and taking isoniazid for 9 months. In a nested case-control trial to assess the role of HCV, coinfection with HCV was an independent risk factor for hepatotoxicity, but study findings suggest that isoniazid-rifapentine may be the preferred choice for treatment of LTBI among patients who are at high risk of hepatotoxicity. Recent studies have also provided evidence that the 12-week isoniazid-rifapentine regimen is also both safe and effective in adults with HIV infection and in children (133–137). This regimen is now recommended by CDC, WHO, and the Canadian Tuberculosis Standards and is being increasingly used in clinical practice (6, 36, 59).

Several additional research studies that should expand the body of evidence regarding the safest and most effective regimen for treating LTBI are under way. The HALT trial (Hepatitis and Latent Tuberculosis), a multisite, unblinded, randomized trial comparing isoniazid-rifapentine with isoniazid-rifampin (Rifinah, two 150mg/300mg tablets) to assess completion of and adverse reactions to LTBI treatment, is being conducted in the United Kingdom (ISRCTN04379941: http://www.isrctn.com/ISRCTN04379941). There is also an open-label, multicenter, randomized controlled clinical trial (iADHERE) evaluating patients diagnosed with LTBI who are recommended for treatment: three-arm DOT (standard control), self-administered therapy, and self-administered therapy with text messaging reminders. The primary objective is to evaluate adherence to a 3-month (12-dose) regimen of weekly isoniazid-rifapentine (3INH/RPT) given by DOT compared to self-administered therapy (https://clinicaltrials.gov/ct2/show/NCT01582711).

Isoniazid plus Rifampin

While not included in the 2000 ATS/CDC guidelines for treatment of LTBI in the United States, the 3-month regimen of isoniazid plus rifampin is recommended in the United Kingdom and Canada as well as in the 2014 WHO guidelines for the management of LTBI in low-TB-burden countries (6, 36, 37). A meta-analysis of five randomized trials comprising 1,926 adults from Hong Kong, Spain, and Uganda was conducted in 2005 to determine the equivalence of daily isoniazid-rifampin for 3 months and isoniazid monotherapy for 6 to 12 months with regard to the development of TB, severe adverse drug reactions, and death (138). During a follow-up period varying from 13 to 37 months, the rates of developing active TB were equivalent for the two regimens. A total of 41 patients (4.2%) who received isoniazid-rifampin developed TB, compared with 39 patients (4.1%) who received only isoniazid (pooled risk difference, 0%; 95% CI, −1% to 2%). Severe adverse events requiring drug discontinuation were reported with similar frequencies for the two regimens (4.9% for isoniazid-rifampin and 4.8% for isoniazid), and a subanalysis of high-quality trials suggested that the two regimens were equally safe. Several other studies and a systematic review have also reported that the combination of isoniazid-rifampin is well tolerated and has rates of treatment discontinuation due to adverse drug reactions similar to those of isoniazid monotherapy, even in patients with HIV infection (32, 33, 64, 73, 77, 139). Mortality rates for the two regimens were also equivalent in the studies that provided mortality data. In the Hong Kong study evaluating Chinese men with silicosis treated for LTBI, the efficacy of 3 months of isoniazid-rifampin in preventing active TB was 41% compared to placebo, somewhat less than the efficacy of rifampin monotherapy for 3 months (51%) (64). Moreover, the network meta-analysis conducted by Stagg and colleagues found that 3 to 4 months of isoniazid-rifampin was more efficacious than placebo (OR, 0.52 [CI, 0.34 to 0.79]) and ranked favorably compared to isoniazid for 6 or 9 months, though this finding was based on limited data (33). This meta-analysis also reported that the isoniazid-rifampin regimen also potentially had lower hepatotoxicity than isoniazid-only regimens, although good evidence for this was found only when it was compared with isoniazid given for 12 to 72 months. Of note, rifampin has been shown to potentiate the hepatotoxicity of other anti-TB medications, and another meta-analysis including patients with TB disease estimated the rate of symptomatic hepatitis for isoniazid-rifampin combination at 2.55% compared with 1.6% for those treated only with isoniazid (64, 87). Although the combination of isoniazid and rifampin may be at least equivalent to isoniazid for 6 to 9 months in terms of efficacy, safety, and completion rates, rifampin monotherapy is likely safer and less costly (6, 28, 33, 87). Moreover, there is strong evidence that the 12-week regimen of isoniazid-rifapentine is superior to isoniazid in terms of effectiveness, safety, and treatment completion rates (6, 28, 32, 33) It is important to note, however, that all of these regimens are considered acceptable for treatment of patients with LTBI, and several ongoing trials discussed above may further inform the process of choosing among these regimens in the near future.

If prescribing combined isoniazid-rifampin therapy, the daily doses for isoniazid and rifampin given together are the same as for each drug individually, and the same precautions regarding monitoring and drug-drug interactions apply. As with monotherapy, intermittent combination LTBI treatment should be given only as DOT.

Rifampin plus Pyrazinamide

Earlier studies evaluating the 2-month regimen of rifampin plus pyrazinamide among HIV-infected persons with LTBI demonstrated that this regimen had an effectiveness similar to that of isoniazid and was well tolerated, with a low occurrence of toxicity (8, 140). Based on these data, this regimen was recommended in the targeted tuberculin testing and treatment guidelines issued by the ATS and CDC in 2000 (8). When rifampin-pyrazinamide was subsequently used widely in the general population, an unacceptably high risk of severe and fatal hepatotoxicity was recognized and the recommendations for using this regimen for LTBI therapy were withdrawn (66). Toxicity severe enough to require treatment discontinuation ranged from 2.0 to 17.6% in HIV-seronegative persons and 0 to 9.5% in HIV-seropositive persons (66, 69, 74, 141). A survey of state and city TB programs in the United States conducted by the CDC determined that the rate of symptomatic hepatitis was 18.7 per 1,000 persons among 8,087 patients initiating therapy (69). In this survey, the risk of hepatitis-associated death among persons taking rifampin-pyrazinamide was 10-fold higher than historical isoniazid rates. Another 50 cases of severe liver injury, including 12 deaths, were subsequently reported among persons taking the 2-month rifampin-pyrazinamide regimen (142). Given that the risk of rifampin-pyrazinamide far outweighs the benefits in most situations and the availability of effective alternatives for LTBI therapy, this regimen should not be used (66, 69, 142). One caveat that should be mentioned is that patients who are initially suspected of having TB disease and who receive at least 2 months of treatment containing both rifampin and pyrazinamide (usually with isoniazid and possibly ethambutol) can be considered to have completed LTBI therapy if active TB disease is ruled out (47).

MONITORING DURING LTBI THERAPY

Hepatotoxicity, or drug-induced liver injury, is perhaps the most concerning adverse effect of LTBI therapy and is most often a result of metabolic idiosyncratic reactions (8, 57, 59, 66, 87, 143). Other side effects can also occur (see sections on individual regimens above), although the vast majority of patients do not experience any problems during therapy. Monitoring can mitigate the potential of adverse sequelae; however, a systematic review conducted to inform the WHO LTBI management guidelines failed to identify any study providing direct evidence on best practices for monitoring of treatment. In this review, seven national LTBI guidelines from low-incidence countries were found to include recommendations for monitoring (6, 144). All seven recommended that all patients be monitored clinically throughout LTBI therapy to detect the occurrence of focal and systemic drug reactions. Clinical monitoring involves an assessment of symptoms and physical manifestations of possible adverse drug effects and is indicated for all patients at baseline (Fig. 2) and on a monthly basis (Fig. 3) until treatment is completed (8, 35, 59, 87, 144). Only one month’s supply of medication should be provided at each visit to facilitate frequent clinical monitoring and reinforce medication adherence. Patients’ medication lists should be regularly reviewed for potential drug-drug interactions, and patients should be instructed to stop therapy and immediately contact their providers if problems occur, in particular, symptoms of hepatitis such as jaundice, loss of appetite, emesis, abdominal pain, fatigue, and/or muscle and joint aches. Although recommendations vary by country, in the United States, baseline laboratory testing is not recommended before any LTBI treatment regimen except for patients with HIV infection, pregnant women and those within 3 months of delivery, persons with chronic liver disease, and those who use alcohol regularly or are at increased risk of chronic liver disease (8, 59, 87). Laboratory monitoring of ALT, AST, and bilirubin is recommended in addition to monthly clinical assessment only for patients whose baseline testing is abnormal and for those at risk for hepatic disease (Fig. 3). If monitoring is performed, LTBI treatment should be held if either the ALT or AST exceed three times the ULN with symptoms present or five times the ULN in an asymptomatic individual (6, 8, 35, 59, 87). For liver enzyme elevations less than three times the ULN in symptomatic patients, close clinical and laboratory monitoring should be instituted if treatment is to be continued.

Ensuring treatment completion is a critical aspect of monitoring patients who are taking one of the recommended LTBI regimens (8, 35). Treatment completion is estimated by counting the doses taken by each patient in a predefined period (8, 145). For example, completion of 9 months of isoniazid is defined as taking a minimum of 270 daily doses or 76 twice-weekly doses within 12 months. If a 6-month regimen of isoniazid is prescribed, 180 daily doses or 52 twice-weekly doses should be completed in 9 months. For rifampin monotherapy, 120 daily doses taken within 6 months are required, and for isoniazid-rifapentine, 12 doses are required within 16 weeks. The extended duration for each regimen allows for minor interruptions in therapy, but if a person has a gap in treatment of more than several months, consideration should be given to restarting treatment after ensuring that active TB has not developed, as long as the patient is motivated and there is a plan to ensure completion of therapy.

ADHERENCE TO LTBI THERAPY

Cascade of care has been defined for patients with LTBI. Steps include the identification of high-risk persons who should be tested for LTBI, completion of testing (including receipt of TST or IGRA results), provider recommendations for LTBI treatment, patient acceptance of treatment, and completion of a full recommended LTBI regimen (146). Patient dropout at each of these steps results in missed opportunities to prevent future development of active TB disease. A systematic review conducted to explore the extent of and reasons for patient loss along the cascade found that two of the most critical points of loss were provider recommendation to initiate LTBI therapy and patient failure to complete treatment once started (146), which have also been described in other studies (55, 95, 96, 147). Fewer patients drop out at the steps of receiving testing results, being referred for evaluation of a positive test, and acceptance of LTBI therapy if it is recommended (146, 148), though rates of both initiation and treatment completion vary by risk group and treatment setting (95, 96). In a CDC-funded Tuberculosis Epidemiologic Studies Consortium study evaluation of clinics prescribing LTBI therapy in the United States and Canada, 83% of persons recommended LTBI treatment accepted it, and 47% of those completed the full course. The combined loss from acceptance and completion means that only a minority (39%) of persons with LTBI eligible for treatment achieved the full benefit (55). This demonstrates the need to implement strategies to improve retention across the cascade of preventive therapy.

Poor adherence has a substantial impact on outcomes such as the effectiveness for preventing reactivation of TB, complications of both the disease and treatment, and the emergence of potential drug resistance (54, 95, 114, 145). At a community level, low completion rates among patients treated for LTBI result in failure to prevent reactivation of TB and continued risk of M. tuberculosis transmission, with increased overall health care costs (9, 10, 54, 149). Published treatment completion rates for 6 to 9 months of isoniazid have been unacceptably low, not even reaching 70% in a controlled research setting (8, 54–56, 58). A large, multisite, prospective evaluation of close contacts to persons with culture-positive pulmonary TB found that despite the high risk of progression to active TB, only 53% of contacts recommended LTBI treatment completed a full course (150). Contacts were significantly less likely to complete therapy for LTBI if they were prescribed isoniazid. Of contacts taking isoniazid, 14 developed active TB (9 who completed <6 months, 2 who completed ≥6 months, and 3 with unknown duration of treatment). Reported completion rates in studies using shorter rifampin-based regimens, however, have consistently reported much higher completion rates (8, 58, 96, 115). Moreover, several studies have demonstrated that a shorter duration of treatment is a significant and independent predictor of adherence (55, 110, 114, 146, 151, 152). Among LTBI patients taking rifampin, the favorable tolerability and lower occurrence of severe side effects have also been shown to contribute to enhanced adherence (70, 110, 115, 153). Similarly, the use of DOT, a convenient once-weekly dosing schedule for only 12 weeks, and improved safety profile have likely facilitated high treatment completion rates for both adults and children taking the isoniazid-rifapentine regimen (58, 132–134, 152).

Despite the advantages of adherence, tolerability, and safety associated with shorter rifamycin-based treatment, isoniazid remains the most commonly used regimen to prevent TB disease worldwide (6, 40, 55, 151). Anecdotally, the lower uptake of rifampin monotherapy despite the long-standing endorsement of national agencies from the United States, Canada, and the United Kingdom has likely been due to the lack of established efficacy data using this regimen and concerns for the development of rifampin resistance if active TB is inadvertently missed during the pretreatment evaluation (6, 28, 99, 122). Of significance, despite this concern, no significant reaction has been found between either isoniazid or rifampin monotherapy for LTBI and acquired drug resistance (6, 28, 32). Conclusive data demonstrating the long-term efficacy of the shorter rifamycin-based regimens and most recent recommendations from the WHO and the USPHTF should faciliate the scale-up of rifamycin-based LTBI treatment in the community.

Predicting which individuals initiating LTBI therapy are likely to need assistance to complete their recommended course may be challenging because adherence is simultaneously affected by multiple factors. In addition to the characteristics of the regimen (e.g., duration, pill burden, and tolerability), concurrent use of other medications, individual patient characteristics, socioeconomic factors, the structure and nature of health care services offered, the quality of the patient-provider communication, and the nature of social support that patients receive all contribute to acceptance and completion of LTBI therapy (54, 96, 114, 145, 146, 152, 154). This may explain why studies examining various predictors of adherence to LTBI therapy have reported conflicting findings (Table 5) regarding the association of adherence with sociodemographic characteristics such as age, gender, education, or occupation (54, 56, 96, 152). Low LTBI treatment completion rates have been more consistently reported among certain high-risk groups, such as inner-city residents, substance abusers, released jail inmates, homeless persons, migrants, and the mentally ill (54, 95, 97, 98, 100–103, 145, 146, 152, 155, 156). Clinic-based factors such as hours of operation, convenient location, costs, and availability of culturally competent services have also been shown to affect adherence (56, 96, 99, 114, 147, 152). Additionally, many patients have a hard time understanding the importance of taking medication with potential side effects for prolonged duration in order to treat an asymptomatic condition that may never result in disease (54, 148, 154). Indeed, low perception of the risk of progression from latent to active TB has been demonstrated as a prominent predictor of failure to complete therapy (54, 147, 157). Other knowledge, attitudes, and beliefs have also been associated with patients’ willingness to initiate or complete LTBI treatment, in particular, a misconception among persons born in TB-endemic countries that BCG vaccine is protective against TB in adults. Because foreign-born individuals comprise the majority of patients treated in high-income low-burden countries (6, 10, 39, 148), addressing this and other cultural factors that may affect treatment among this population could promote the acceptance and completion of LTBI therapy (147, 148, 158).

TABLE 5.

Overview of determinants of LTBI treatment initiation, adherence, and completion in the general population diagnosed with LTBI^a

Determinant	Specification determinant (vs. reference group)	No. of articles
		Positive association		Inverse association
		P (reference)	R (reference)	P (reference)	R (reference)
Determinants of LTBI treatment initiation
Age	Older age (vs younger age)		1 (187)		2 (188, 189)
Gender	Men (vs women)		1 (189)		1 (187)
Subpopulation within general population with LTBI	Refugee/immigrants (vs born in country of study)	1 (190)	1 (189)
	Immigrants born in WHO category 3 or 5 country (vs category 1 country)b	1 (190)
	HCW (vs no HCW)				2 (55, 188)
	Case contact (vs no case contact)	1 (191)	2 (55, 188)
Education	Lower education level (vs NR)	1 (191)
Behavior	Alcohol use reported at baseline (vs no alcohol use reported)				1 (187)
Other	Continuity of primary care by consulting a regular physician (vs NR)	1 (191)
	Pregnant (vs not pregnant)				1 (169)
	Prior incarceration (vs NR)	1 (191)
	Fear of getting sick with TB without medicine (vs no fear of getting sick)	1 (191)
	Previous BCG vaccination (vs NR)				1 (188)
	Abnormal chest X-ray findings consistent with previous TB (vs NR)		1 (188)
	A nonemployment reason for screening (vs NR)	1 (191)
Determinants of LTBI treatment adherence
Age	Older age (vs younger age)			1 (192)
Ethnicity	Biculturalc (vs Hispanic or non-Hispanic)	1 (192)
Education	Higher grades in school (vs lower grades)	1 (192)
Behavior	Risk behaviors (vs NR)d			2 (192, 193)
Adverse events	Some somatic complaints (vs NR)			1 (193)
Determinants of LTBI treatment completion
Age	Older (vs younger)	3 (194, 195)e,f	4 (110, 151, 196, 197)g	3 (91, 190, 198)	6 (55, 114, 199–202)
Gender	Male (vs female)				2 (197, 199)
Ethnicity	Hispanic/Latino ethnicity (vs Asian ethnicity)			1 (198)
	White Hispanic (vs black, non-Hispanic)		1 (199, 201, 203)
	Country of birth (i.e., Haiti, Dominican Republic, China with Hong Kong, or Vietnam) (vs other countries)	Various results found between countries (158)
	Asian/Pacific Islander (vs white)		2 (196, 197)
	Region of origin (i.e., Latin America and Caribbean or Asia and others) (vs USA, Canada, Europe)		1 (114)
	Black race (vs NR)				1 (110)g
	Ethnicity (i.e., Asian, Non-Hispanic black, or Hispanic) (vs non-Hispanic white)		1 (151)
Subpopulation within source population	HCW (vs no HCW)				1 (55)
	Case contact (vs no case contact)		1 (151)		1 (110)h
	Currently homeless (vs not currently homeless)				2 (199, 204)
	PWID (vs no PWID)				2 (55, 203)
	Refugees/immigrants (vs born in country of study)	1 (205)	4 (110, 151, 199, 200)g		2 (204, 206)
	Indication for LTBI treatment immunosuppression (vs case contact)	1 (194)f
Health	History of hepatitis A, B, or C (vs no history of liver disease)	1 (91)
	Other medications reported at baseline (vs NR)				1 (110)h
	Use of concomitant medications by women (vs no use of concomitant medication)				1 (187)
Behavior	(Excess) alcohol use (vs no alcohol use)				4 (110, 187, 199, 204)h
	Smoking (vs nonsmoking)	1 (194)f
Treatment	Treatment without H (vs treatment with H)	1 (194)f	5 (111, 114, 151, 196, 207)
	9 mo of H (vs other regimens)				1 (55)
	Regimen choice offered (vs no regimen choice offered)		1 (202)
	Twice-wkly RZ (vs daily RZ)		1 (69)
	DOT (vs SAT)		3 (151, 197, 208)
Adverse events	Adverse events (vs no adverse events)				7 (114, 169, 187, 199, 201, 206, 209)
	Adverse events (i.e., grade 1 or 2 hepatotoxicity, grade 3 or 4 hepatotoxicity, or adverse events other than hepatotoxicity) (vs NR)	Conflicting results found between adverse events (68)
Other	Not having been incarcerated within 6 mo of diagnosis (vs NR)	1 (189)
	Referral reason (i.e., correctional/rehabilitation or postpartum women) (vs TST positive from screening)				1 (199)
	Risk group (i.e., contact, medical risk,i population riskj) (vs low riskk)		1 (152)
	Cause of screening/referral (i.e., asylum seekers or contacts) (vs anti-TNF-α candidates)				1 (209)
	Fear for venipuncture (vs NR)			1 (158)
	Low TB risk perception (vs NR)			1 (158)
	Plan to tell friends or family about LTBI diagnosis (vs NR)	1 (190)
	Home situation (i.e., child living with no or one natural parent) (vs living with both natural parents)			1 (204)
	Spanish language (vs non-Spanish language)		1 (210)
	Resident in a congregate setting (vs never or unknown)				1 (55)
	Missed appointment call or letter (vs no missed appointment call)				1 (210)
	No medical insurance (vs medical insurance)				1 (169)
	Clinic attendance before treatment (vs clinic nonattendance before treatment)		1 (201)
	Presumed nonrecent TB infection (vs presumed recent TB infection)				1 (202)
	Public health nurse referral (vs no public health nurse referral)				1 (210)

^a Reprinted from reference 97, under the International CC-BY 4.0 license. Abbreviations: P, ••; R, ••; HCW, health care worker; NR, none reported; PWID, ••; H, ••; RZ, ••; SAT, ••.

^b The WHO defined 5 categories of TB prevalence based on 1st (least prevalent) to 5th (most prevalent).

^c “Bicultural” is defined by questions separated into the domains Hispanic and non-Hispanic, considering language use, linguistic proficiency, and electronic media use. Individuals scoring high in both domains are considered bicultural.

^d Risk behaviors: ever having used alcohol, cigarettes, or marijuana, been expelled or suspended from school, or been in a physical fight.

^e Data analyzed for individuals that underwent three QFT-GIT.

^f Data analyzed for individuals who underwent at least one serial QFT-GIT.

^g Data analyzed for non-Hispanic subjects for one study.

^h Data analyzed for Hispanic subjects for one study.

ⁱ Persons with medical risk factors such as having a TST conversion within 2 years of a negative TST, HIV infection, untreated or partially treated prior TB, suspected TB with an abnormal chest radiograph, being younger than 5 years of age with a positive TST, or having a clinical condition associated with an increased risk of TB disease.

^j Persons with population risk factors, such as recent immigrants to the United States (5 years) from countries with high TB prevalence, homeless persons, and residents and employees of congregate settings such as prisons, jails, and health care facilities.

^k Persons with low risk for developing TB disease (no case contact, no medical risk, and no population risk factors).

Various measures have been used to enhance LTBI treatment adherence, but no single intervention has been shown to be consistently effective (35, 54, 56, 96) (Table 5). The use of DOT has been associated with improved adherence to LTBI therapy in many (but not all) settings, and it is recommended for all intermittent regimens. DOT is also recommended for certain populations at increased risk of nonadherence and subsequent development of active TB, such as children, recent contacts, and HIV-infected persons (54, 96, 151, 159–162). DOT is usually provided or coordinated by the public health department and is most feasible when given in medical clinics, schools, work sites, day care programs, or in homes, particularly when delivered at the same time and site as DOT for persons with TB disease (54, 99, 151, 160). Despite the benefits of DOT for ensuring ingestion of medication and facilitating close patient monitoring during treatment, this practice is more costly, requiring additional staff and financial resources, and is thus not widely available (161, 163, 164). Preliminary findings from the iADHERE study indicate that overall treatment completion was 87.2% [95%CI 83.1%-90.5%] by DOT, 74.0% [68.9%–78.6%] by self-administered therapy, and 76.4% [71.3%–80.8%] by enhanced SAT with weekly text reminders; SAT was only found to be non-inferior to therapy given by DOT in the United States (http://www.croiconference.org/sessions/adherence-once-weekly-self-administered-inh-and-rifapentine-latent-tb-iadhere).

Other strategies that have been successfully used to enhance adherence include intensive case management using a patient-centered approach, education and counseling, adherence coaching, peer support, contingency contracting, and provision of incentives (e.g., cash, gift cards, coupons, stickers or other rewards for children, etc.) and enablers (e.g., public transportation passes, food, evening clinic hours, fast-track scheduling, etc.) (32, 56, 96, 152). Several studies have shown that case managers attuned to ethnocentric beliefs, patient concerns, and other barriers have been able to improve the acceptance of and adherence to LTBI therapy (96, 145, 147, 165–167). In addition, pill taking at a consistent time during the day, follow-up phone calls or text messages to remind patients to continue medication, minimizing clinic wait times, reminders before appointments, tracking adherence to appointments, and rapid rescheduling of missed clinic visits, have been employed to encourage treatment acceptance and completion (54, 151, 168).

Although some adherence strategies such as DOT may be too time-consuming and expensive to apply broadly (54, 168), screening patients for risk of nonadherence at baseline and during therapy can enable the targeted use of such measures when indicated (54, 145, 151, 152, 168, 169). One study of LTBI patients treated at an academic medical center TB clinic in Boston, MA, found that patient views indicative of poor adherence were often evident at the first clinic visit (157), and several studies have demonstrated that poor adherence in the first month of LTBI treatment strongly predicts subsequent loss to follow-up and low likelihood of treatment completion (151, 157, 158, 168). In the PREVENT TB study comparing isoniazid-rifapentine to isoniazid alone, the majority of discontinuations occurred early, and participants who missed an early visit and then returned at least once were more likely to permanently discontinue LTBI therapy (152). Selective use of shorter LTBI treatment regimens may also enhance adherence among individuals with obvious risk factors such as homelessness, substance abuse, and mental illness.

COST-EFFECTIVENESS OF LTBI TREATMENT REGIMENS

Cost-effectiveness is an important consideration for public health programs when determining the most appropriate LTBI treatment regimen to use for targeted testing and LTBI treatment activities. Isoniazid has consistently been found to be a cost-saving strategy compared to no treatment, particularly in younger populations and in persons at increased risk of progression from latent infections to active TB disease (28, 94). A meta-analysis of published studies found that the 4-month rifampin regimen was more cost-effective than 9 months of isoniazid, with a calculated cost savings of $213 per patient treated ($90/patient without physician fees) (115). Although the medication cost of isoniazid is much lower than for rifampin, the total cost of isoniazid treatment is increased by the additional monitoring and laboratory costs incurred with a longer treatment period and higher risk of adverse events (112, 115, 170). In another study using a computerized Markov model to estimate and compare total costs of treatment, quality-adjusted life years gained, and cases of active TB prevented for four LTBI treatment regimens, rifampin was found to be less costly and more effective than both self-administered and directly observed isoniazid regimens over a wide range of estimates for adherence and efficacy (163). Isoniazid plus rifapentine given as DOT for 3 months became cost-effective only for patients with the highest risk, such as HIV infection, with most of the higher costs of this regimen attributed to the use of rifapentine as well as DOT (170). However, since the cost of rifapentine has decreased, a more recent analysis found that 3 months of isoniazid-rifapentine is likely more cost-effective than 9 months of isoniazid (171, 172).

TREATMENT OF LTBI FOR PERSONS EXPOSED TO DRUG-RESISTANT MYCOBACTERIUM TUBERCULOSIS

When there is strong evidence that a person has become infected with a strain of M. tuberculosis resistant only to isoniazid, treatment with rifampin for 4 months is recommended (8, 36, 106). It is particularly important to exclude active TB disease prior to initiating rifampin monotherapy in patients with existing isoniazid resistance to avoid additive resistance to multidrug-resistant TB (MDR-TB) (i.e., having resistance to both isoniazid and rifampin) (6, 28). However, when considering treatment of persons exposed to infectious persons with MDR-TB, there is a paucity of data to inform this decision (6, 28, 173). There are currently no published randomized controlled trials and only a few observational studies to assess the most effective regimen for treating these contacts, though several are under way (174). Management of persons exposed to MDR-TB must be based largely on clinical and epidemiologic judgment, taking into consideration the infectiousness of the case and closeness, intensity, and duration of exposure and, where known, the drug susceptibility results for the index patient (36, 106, 173). Contacts to patients with MDR-TB should be quickly identified and carefully evaluated for active TB disease; if TB disease is suspected, patients should be managed according to current standards. If active TB is ruled out, high-priority contacts should be evaluated for LTBI as previously described, including an assessment of whether each individual has had previous diagnosis or treatment of LTBI before their exposure to MDR-TB.

Regarding treatment of contacts to drug-resistant TB, a prospective cohort study among South African children who were household contacts to an MDR-TB source case demonstrated that LTBI treatment individualized according to the drug susceptibility pattern of the source case was effective compared to no treatment (175). Following this approach, the current Drug-Resistant Tuberculosis: A Survival Guide for Clinicians from the Francis J. Curry National Tuberculosis Center (San Francisco, CA) supported by the CDC provides guidance for suggestions for tailoring treatment according to resistance pattern (Table 6) (106). The use of fluoroquinolones alone or in combination with ethambutol, ethionamide, cycloserine, or p-aminosalicylic acid is recommended if the source case isolate is susceptible to these drugs. These second-line drugs are more expensive and have an increased side effect profile; thus, the risks and benefits must be carefully considered prior to initiating therapy with any of these drugs. Where feasible, it is also best to use drugs which are more likely effective against nonreplicating or slowly replicating organisms that constitute a latent infection. Although CDC/ATS recommendations from the 1990s encouraged the use of multidrug regimens including pyrazinamide, these regimens were poorly tolerated and associated with unacceptable toxicity, leading to premature discontinuation (106, 176–178). Another concern is that resistance to quinolones is increasing globally, and in many settings isolates from the index case are not tested to determine susceptibility to these drugs (1, 179).

TABLE 6.

Specific treatment options for drug-resistant TB dependent on susceptibility of source case isolate^a

Resistance pattern	LTBI treatment optionsb
INH and RIF	Fluoroquinolonec monotherapy or
	Fluoroquinolone + EMB
INH, RIF, EMB	Fluoroquinolone monotherapy or
	Fluoroquinolone and ETA
INH, RIF, PZA	Fluoroquinolone monotherapy or
	Fluoroquinolone and EMB
INH, RIF, PZA, EMB, ± injectable	Fluoroquinolone monotherapy or
	Fluoroquinolone and ETA
INH, RIF, PZA, EMB, injectable, ETA	Fluoroquinolone monotherapy or
	Fluoroquinolone and CS
INH, RIF, PZA, EMB, and fluoroquinolone	No treatment with close clinical monitoring
	(In selected cases, CS + PAS or
	PAS + ETA, or
	ETA + CS)

^a Reprinted with permission from the Drug-Resistant Tuberculosis: A Survival Guide for Clinicians. EMB, ethambutol; ETA, ethionamide; CS, cycloserine; INH, isoniazid; PAS, para-aminosalicylate; PZA, pyrazinamide; RIF, rifampin.

^b Recommendations are not evidence based; there have been no clinical trials for the use of these regimens in contacts of patients with MDR-TB. Recommendations are based on the expert opinion of the CDC and the Francis J. Curry National Tuberculosis Center (106).

^c Fluoroquinolones that can be used include moxifloxacin and levofloxacin.

Several randomized controlled trials to evaluate specific regimens for preventing future TB disease among household MDR-TB contacts are under way. The TB CHAMP study in South Africa is comparing the outcomes of levofloxacin versus placebo among infants and children less than 5 years with and without HIV coinfection (http://www.isrctn.com/ISRCTN92634082). A second trial, V-QUIN, is being conducted among both adult and child contacts to MDR-TB cases who are randomized to receive either levofloxacin or placebo (https://anzctr.org.au/Trial/Registration/TrialReview.aspx?id=369817). A third trial is PHOENix, which is being conducted at different AIDS Clinical Trials Group sites to compare the use of delamanid to isoniazid among adult and child contacts (http://www.impaactnetwork.org/studies/IMPAACT2003B.asp). Results of these trials are not expected until 2020 (173). Additionally, bedaquiline has been shown to have sterilizing activity that may be as good as the rifampin monotherapy and rifampin-isoniazid regimens (but not isoniazid-rifapentine) in paucibacillary murine models of LTBI with drug-susceptible strains (28, 180, 181). However, concerns regarding potential side effects, such as a prolonged QT interval, have decreased interest in testing bedaquiline in healthy humans until there is more evidence to support its safety (28, 182).

The most effective duration of LTBI treatment for contacts exposed to MDR-TB is unknown, but a 6- to 12-month regimen at standard dosages is typically used (160). Given the increased potential for adverse events during treatment with second-line TB drugs, treated patients should be closely monitored during therapy for both adherence and signs of toxicity. Medication fact sheets are available for every anti-TB medication in chapter 5 of the Drug-Resistant Tuberculosis: A Survival Guide for Clinicians (http://www.currytbcenter.ucsf.edu/products/drug-resistant-tuberculosis-survival-guide-clinicians-3rd-edition/chapter-5-medication-fact) (106). It is also strongly recommended that an expert in the management of drug-resistant TB be consulted regarding appropriate treatment and monitoring of MDR-TB contacts. Finally, these contacts should be monitored closely for at least 2 years following completion of therapy.

Given the challenges and often uncertainties of treating contacts to MDR-TB with currently available regimens, an acceptable alternative is to provide no LTBI treatment and instead monitor the patient closely for at least 2 years (6, 106). If the person does develop active TB during the initial 2 years when progression to active TB is most likely (5% of the total 10% lifetime risk in most immunocompetent adults), presumably they will be diagnosed and treated early in the course of their TB disease, which should improve long-term outcomes. This approach may be most appropriate when the risk-benefit ratio favors close contact surveillance rather than using a potentially toxic treatment of uncertain efficacy (106, 173).

TREATMENT OF LTBI IN PERSONS WITH A NEGATIVE TST OR IGRA

In developed countries such as the United States and Canada, treatment of persons without evidence of LTBI (i.e., positive TST or IGRA result) is recommended only in a few specific circumstances (36, 37, 160). Because children aged <5 years are more susceptible to infection following exposure to M. tuberculosis and are more vulnerable to invasive, fatal forms of TB disease, treatment for presumptive M. tuberculosis infection (i.e., window prophylaxis) is recommended after TB disease has been excluded if the interval since the last exposure is >8 weeks, even if an initial skin test or IGRA result is negative. If a second LTBI test result is negative 8 to 10 weeks postexposure, treatment should be discontinued. However, if the second result is positive, the full course of treatment for latent M. tuberculosis infection should be completed (8, 160).

Similarly, HIV-infected persons, those taking immunosuppressive therapy for organ transplantation, and persons taking therapies such as TNF-α antagonists are at increased risk of rapidly developing active TB disease once infected. Therefore, after known exposure to M. tuberculosis, such persons should complete a full course of LTBI treatment as soon as active TB is ruled out, regardless of the results of the follow-up TST or IGRA even 8 or more weeks after exposure. The risks for TB are less clear for patients who chronically take corticosteroids, though the 2000 CDC/ATS guidelines consider 15 mg/day of prednisone (or its equivalent) administered for 1 month a risk factor for developing TB, largely because this dose can suppress tuberculin reactivity and may limit the accuracy of the IGRA (8, 183). A case-control study based on a large general practice population in the United Kingdom found that not only was the risk for developing TB significantly increased among patients using the equivalent of >15 mg of prednisone per day versus nonusers (aOR, 7.7; 95% CI, 2.8 to 21.4), but also the risk was signficantly higher among persons using <15 mg per day (aOR, 2.8; 95% CI, 1.0 to 7.9). Thus, for individuals with negative TST or IGRA who are taking corticosteroids and have a significant exposure to infectious TB, treatment for LTBI should be considered on a case-by-case basis (183). When determining whether to initiate window period or prophylactic treatment of a potential contact, the contribution of certain factors must also be considered, including the extent of disease in the index patient, the duration that the source and the contact are together and their proximity, and local air circulation (160). The public health department can provide consultation regarding the likelihood of clinically significant exposure and whether LTBI treatment is indicated. Of note, isoniazid has not been demonstrated to be effective in preventing TB in skin test-negative HIV-infected persons without known contact to active TB in low-TB-incidence areas and is thus not recommended (8, 74, 80).

CONCLUSIONS

An estimated 56 million people from among the 1.7 billion harboring infection with M. tuberculosis may develop active TB over their lifetimes (7). However, modeling studies have suggested that if 8% of individuals with LTBI were treated each year, the global incidence in 2050 would be 14-fold lower than the incidence in 2013, even without other TB control interventions (2). Thus, measures to identify and treat latently infected persons must be scaled up to achieve the goals of the current Global End TB Strategy. Although clearly efficacious in preventing TB disease, the effectiveness of LTBI treatment has been limited by factors such as concern for hepatotoxicity, insufficient recommendations to start therapy, and poor adherence. Clinicians should recognize that there is no one-size-fits-all approach; thus, treatment should be individualized. Shorter rifamycin-based regimens are as effective, better tolerated, and more likely to be completed than 6 to 9 months of isoniazid, including 4 months daily of rifampin monotherapy, 3 months of once-weekly isoniazid-rifapentine, and 3 to 4 months of daily isoniazid-rifampin. Choosing the most appropriate regimen, clinical monitoring for potential adverse events, and utilization of adherence-promoting strategies are critical elements for the success of LTBI treatment to prevent additional TB disease. Regimens that are even safer, more cost-effective, and more easily completed must be developed and combined with interventions to identify, engage, and retain high-risk individuals across the cascade from LTBI diagnosis through completion of treatment.