Tuberculosis disease in children

INTRODUCTION — Issues related to tuberculosis (TB) disease in children will be reviewed here. Issues related to diagnosis and treatment of latent TB infection in children are discussed in detail separately. (See "Tuberculosis infection (latent tuberculosis) in children".)

TERMINOLOGY — TB terminology is inconsistent in the literature [1]. Relevant terms are defined in the table (table 1).

EPIDEMIOLOGY

Worldwide — Estimating the global burden of TB disease in children is challenging due to the lack of a standard case definition, the difficulty in establishing a definitive diagnosis, the frequency of extrapulmonary disease in young children, and the relatively low public health priority given to TB in children relative to adults [2].

The World Health Organization (WHO) publishes global TB data including new and relapse cases by age. In its 2019 report, the WHO estimates that of the estimated 10 million incident cases of TB in 2018, approximately 1.1 million (11 percent) occurred among children <15 years; similar numbers of males and females were affected [3]. In 2019, the WHO estimated that there were 205,000 deaths due to TB in children <15 years (32,000 occurring in children with human immunodeficiency virus [HIV] infection) [3]. These deaths represent 13.8 percent of all TB deaths (which is higher than the estimated proportion of cases in children), suggesting a higher mortality rate in this age group [3].

Children under age five represent an important demographic group for understanding TB epidemiology, since TB frequently progresses rapidly from primary or latent infection to disease, and severe disease manifestations, such as miliary TB and meningitis, are more common in this age group. Therefore, these children serve as sentinel cases, indicating recent and/or ongoing transmission in the community.

Most children are infected by household or other close contacts with TB disease, particularly parents or other caretakers. Even in circumstances when adult index cases are sputum smear negative, transmission to children has been documented in 30 to 40 percent of households [4].

It has been estimated that, of nearly one million children who developed TB disease in 2014, 58,000 had isoniazid-monoresistant TB, 25,000 had multidrug-resistant (MDR)-TB, and 12,000 had extensively drug-resistant TB [5]. Additional effort is needed to improve detection of drug-resistant TB among children.

United States — Risk factors for pediatric TB in the United States include being born outside of the United States, having a parent who is non-United States-born, or having lived outside the United States for more than two months [6]. In the United States, TB among children is relatively rare.

Between 2007 and 2017, 9276 cases of TB were reported among children and adolescents <18 years [7]; 76 percent of cases among children and adolescents <15 years were United States born, while 42 percent of cases among adolescents ages 15 to 17 years were United States born. Overall, the incidence of TB was 12.9 times higher among children and adolescents born outside the United States. More than half of patients <15 years were 1 to 4 years of age [7]. About two-thirds of children <15 years with TB would have been recommended for TB testing under targeted testing guidelines [7]. Cases were discovered in a variety of ways: 37 percent of cases were prompted by the evaluation of symptoms. Cases were discovered via contact tracing among 41 percent of United States-born cases but only 10 percent of non-United States-born cases.

Among children and adolescents with TB in the United States between 2007 and 2017 with documented race/ethnicity data, 42 percent were of Hispanic origin, 27 percent were Black, 22 percent were of Asian origin or Pacific Islander, 7 percent were White, and 2 percent were American Indian or Alaska Natives [7]. Between 1993 and 2015, HIV status was known for 28 percent of the pediatric patients reported (n = 21,222); of these, 3.1 percent were HIV infected [8]. Drug susceptibility testing data from 2015 reveal that the isolates from 21 pediatric TB cases (15 percent of those with culture-confirmed TB) had detectable resistance to one or more drugs; none in that year had MDR-TB [6].

CLINICAL MANIFESTATIONS

Pulmonary tuberculosis — Pulmonary disease and associated intrathoracic adenopathy are the most frequent presentations of TB in children [9,10]. Common symptoms of pulmonary TB in children include [11]:

●Chronic, unremitting cough that is not improving and has been present for more than three weeks

●Fever of more than 38°C for at least two weeks, other common causes having been excluded

●Weight loss or failure to thrive (based on child's growth chart)

However, these symptoms are fairly nonspecific. In one study comparing symptoms of children with culture-proven TB with children with other lung diseases, there was no difference between the two groups with respect to weight loss, chronic cough, and duration of symptoms [12]. The only factors differentiating the groups were history of contact with an infectious TB case and a positive tuberculin skin test (TST). In a study of more than 1000 infants without HIV infection in South Africa, cough >2 weeks' duration (present in 17 percent) was the only diagnostic symptom associated with severe pulmonary TB disease; this symptom was twice as common in severe TB compared with nonsevere TB [13].

Physical exam findings may suggest the presence of a lower respiratory infection, but there are no specific clinical signs or findings to confirm that pulmonary TB is the cause. Children ages 5 to 10 may present with clinically silent (but radiographically apparent) disease, particularly in the setting of contact tracing [9]. In contrast, infants are more likely to present with signs and symptoms of lung disease. Common radiographic findings are discussed below. (See 'Chest radiography' below.)

Extrapulmonary tuberculosis — The clinical presentation of extrapulmonary TB depends on the site of disease. The most common forms of extrapulmonary disease in children are TB of the superficial lymph nodes and of the central nervous system [14]. Neonates have the highest risk of progression to TB disease with miliary and meningeal involvement [14]. Some forms of TB and their common physical signs are as follows [15]:

●Tuberculous meningitis – meningitis not responding to antibiotic treatment, with a subacute onset, communicating hydrocephalus, stroke, and/or elevated intracranial pressure (see "Central nervous system tuberculosis: An overview")

●Pleural TB – Pleural effusion (see "Tuberculous pleural effusion")

●Pericardial TB – Pericardial effusion (see "Tuberculous pericarditis")

●Abdominal TB – Distended abdomen with ascites, abdominal pain, jaundice, or unexplained chronic diarrhea (see "Abdominal tuberculosis")

●TB adenitis – Painless, fixed, enlarged lymph nodes, especially in the cervical region, with or without fistula formation (see "Tuberculous lymphadenitis")

●TB of the joint – Nontender joint effusion (see "Bone and joint tuberculosis")

●Vertebral TB – Back pain, gibbus deformity, especially of recent onset (rarely seen) (see "Bone and joint tuberculosis")

●Skin – Warty lesion(s), papulonecrotic lesions, lupus vulgaris; erythema nodosum may be a sign of tuberculin hypersensitivity

●Renal – Sterile pyuria, hematuria (see "Urogenital tuberculosis")

●Eye – Iritis, optic neuritis, phlyctenular conjunctivitis (see "Tuberculosis and the eye")

In the context of exposure to TB, presence of these signs should prompt further investigation of extrapulmonary TB.

Perinatal infection — Perinatal TB can be a life-threatening infection; the mortality in the setting of congenital and neonatal TB is about 50 percent [16-18]:

●Congenital TB is rare and most often is associated with tuberculous endometritis or disseminated TB in the mother. It can be acquired hematogenously via the placenta and umbilical vein or by fetal aspiration (or ingestion) of infected amniotic fluid [16,18].

Clinical manifestations of congenital TB include respiratory distress, fever, hepatomegaly, splenomegaly, poor feeding, lethargy, irritability, and low birthweight [17,19]. Clinical evaluation of the infant in the setting of suspected congenital TB should include TST and interferon-gamma release assay (IGRA), HIV testing, two-view chest radiograph, lumbar puncture, cultures (blood and respiratory specimens), and evaluation of the placenta with histologic examination (including acid-fast bacilli [AFB] staining and culture). TST and IGRA are usually negative among infants due to immaturity of the Th1 response; however, a positive result for either test supports a diagnosis of TB infection or disease. A positive TST may also be due to recent Bacille Calmette-Guérin (BCG) immunization; IGRA tests are not positive after BCG.

●Neonatal TB develops following postnatal exposure to the aerosolized respiratory secretions of a contagious person, often the infant’s mother. This is more common than congenital TB, and diagnosis of neonatal TB can lead to identification of previously unrecognized diagnosis of TB in the mother or infectious contact [20].

In the setting of congenital or neonatal TB, the mother or other symptomatic person should be evaluated as outlined in detail separately. (See "Diagnosis of pulmonary tuberculosis in adults".)

Adolescent infection — Adolescents with TB can present with features common in children or adults. In one review including 145 cases of adolescent TB, the following features were noted [21]:

●Most adolescents presented with clinical symptoms.

●Rates of extrathoracic TB were high, including six immunocompetent adolescents with TB meningitis.

●Most cases were AFB sputum smear negative.

●Only half of patients with intrathoracic TB had positive cultures.

●Antituberculous medications were generally well tolerated.

DIAGNOSIS

Overview of approach — TB in children is often diagnosed clinically. Because pulmonary TB in children typically presents with paucibacillary, noncavitary pulmonary disease, bacteriologic confirmation is achievable in less than 50 percent of children and 75 percent of infants; in such cases, pulmonary TB is diagnosed by other clinical criteria [22].

It is suggested that mycobacterial culture of respiratory specimens be performed for children suspected of having pulmonary TB [23]. However, obtaining sputum samples from young children is challenging due to lack of sufficient tussive force to produce adequate sputum samples by expectoration alone [24]. For these reasons, gastric aspiration is the principal means of obtaining material for culture from young children; induced sputum may also be collected if feasible. In addition, most experts recommend that children <12 months who are suspected of having pulmonary or extrapulmonary TB undergo lumbar puncture, regardless of whether neurological symptoms are present [22].

For diagnosis of extrapulmonary TB, specimens for culture should be collected from any site where infection is suspected. Each specimen should be cultured regardless of acid-fast bacilli (AFB) smear results [22]. The most common extrapulmonary specimens include whole blood, bone marrow, tissue specimens (such as lymph node or bone), cerebrospinal fluid, urine, and pleural fluid. Diagnostic yield is variable but the organism is detected in less than 30 percent of specimens. In pleural TB, adenosine deaminase (ADA) levels over 40 units/L in the pleural fluid are observed in the majority of patients [9]. (See "Tuberculous pleural effusion".)

A diagnosis of TB (pulmonary or extrapulmonary) in a child is often based on the presence of the classic triad: (1) recent close contact with an infectious case, (2) a positive tuberculin skin test (TST) or interferon-gamma release assay (IGRA), and (3) suggestive symptoms and findings on chest radiograph or physical examination [15].

The approach outlined by the World Health Organization (WHO) for evaluation of a child suspected of having TB includes [11]:

●Careful history (including history of TB contact and symptoms consistent with TB)

●Clinical examination (including growth assessment)

●TST and/or IGRA (both tests, if available, to increase sensitivity)

●Bacteriological confirmation whenever possible

●Investigations relevant for suspected pulmonary and extrapulmonary TB

●HIV testing (eg, in high HIV-prevalence areas)

All data, including thorough history, physical exam, and diagnostic testing, must be considered carefully. A history of recent close contact with a contagious case of TB (often marked by a positive sputum smear) is a critical factor in making the diagnosis of TB in children, especially for those under the age of five years, because it raises the positive predictive value of all of tests and clinical findings [25]. However, the ill adult contact may have not yet been diagnosed, so asking about ill contacts and facilitating evaluation for ill adults can also expedite diagnosis for children.

Diagnosis of TB should prompt HIV testing. (See "Screening and diagnostic testing for HIV infection".)

Clinical diagnosis — In many cases of TB in children, laboratory confirmation is never established (particularly among children under five years of age). In such cases, the diagnosis relies either primarily or exclusively on clinical criteria. Even with improvements in TB diagnostics, clinical assessment remains a common and critical feature of pediatric TB diagnosis [26,27].

A number of algorithms and scoring systems (with weighted point value assigned to clinical features and risk factors) have been developed for use in resource-limited settings [26-30]. Thus far, two algorithms have demonstrated high sensitivity for diagnosis of TB in children [29,30]. One algorithm applied to 438 children <14 years with HIV infection in Burkina Faso, Cambodia, Cameroon, and Vietnam had a sensitivity and specificity of 89 and 61 percent, respectively [29]. Another algorithm utilized clinical evidence initially and, if inconclusive, reflexed to chest radiography and Xpert MTB/RIF; among 478 children <13 years without HIV infection (including 176 children <2 years and 197 children with low weight for age), sensitivity and specificity were 91 and 52 percent, respectively [30]. While these findings are promising, further validation is needed before broad implementation; the efficacy is likely to be dependent upon the population and setting in which they are used [27].

A clinical diagnosis of TB is strengthened in the setting of clinical and radiographic response to empiric treatment. Treatment is often guided by the culture and drug susceptibility results from the index case (eg, the adult TB contact), again emphasizing the importance of contact tracing for optimal management of TB in children.

Screening tests

Tuberculin skin test — A positive TST may be present in both latent TB infection (LTBI) and in TB disease. Thus, although a positive TST may help support a diagnosis of disease, this finding alone is not diagnostic of disease; it must be considered together with other diagnostic criteria. The TST is helpful for diagnosis of TB in children only in circumstances when it is positive. Criteria for positive TST are outlined in the table (table 2) [15]. A positive TST may be falsely positive due to prior vaccination with Bacille Calmette-Guérin (BCG), infection with nontuberculous mycobacteria, and improper administration or interpretation (table 3).

A negative TST does NOT rule out TB disease, since false-negative results can occur in a variety of circumstances (eg, incorrect administration or interpretation of the TST, age less than six months, immunosuppression by HIV, other disease or medication, certain viral illnesses or recent live-virus immunization, and overwhelming TB disease) [15,31]. (See "Use of interferon-gamma release assays for diagnosis of tuberculosis infection (tuberculosis screening) in adults".)

Because the TST cannot distinguish between TB disease, latent Mycobacterium tuberculosis infection, and infection due to nontuberculous mycobacteria, the result must be interpreted in the context of the clinical features and history of TB exposure [32]. Overall, up to 40 percent of immunocompetent children with culture-confirmed TB disease may have a negative TST [22,33]. TST positivity rates vary by form of disease; in pulmonary and extrapulmonary TB, the TST is typically positive (90 and 80 percent, respectively), while in miliary TB and TB meningitis, the TST is usually positive in only 50 percent of cases [34-36].

Interferon-gamma release assays — IGRAs are in vitro blood tests of cell-mediated immune response to relatively TB-specific antigens (such antigens are absent in the BCG vaccine and most nontuberculous mycobacteria). These assays have sensitivity similar to the TST but greater specificity for diagnosis of LTBI and are most useful for evaluation of LTBI in BCG-vaccinated individuals [37-43]. As with the TST, IGRAs cannot distinguish LTBI from TB disease. Like the TST, IGRAs can be falsely negative in children with advanced TB disease because the disease itself is immunosuppressive.

We are in agreement with most experts who favor the use of IGRAs in children >2 years of age, especially in the setting of BCG vaccination [38-43]. A number of studies and meta-analyses have demonstrated that IGRAs perform well in children ≥4 years, and there is growing evidence for their use in children ≥2 years [44-51]. In one study including more than 3500 children under age 15 (including 219 under age 2) at risk for LTBI or progression to TB disease, the specificity and negative predictive value of IGRAs were both high (>90 percent and 100 percent, respectively) [51]. Among 533 children (including 54 under age 2) who had positive TST but negative IGRA results and did not receive LTBI treatment, none developed TB disease [51]. Both TST and IGRA were poor for predicting progression to TB disease [51].

A positive IGRA result should be considered indicative of infection with M. tuberculosis or Mycobacterium bovis (as LTBI or TB disease). However, the IGRA antigens are not found in BCG. A negative IGRA result cannot conclusively exclude a diagnosis of LTBI (or TB disease) and should be interpreted in the context of other clinical data. Indeterminate IGRA results may occur more often in children <2 years [50]; an indeterminate result should not be used for clinical decision-making.

In immunocompromised children, IGRAs should be interpreted with caution. IGRAs include positive control assays for nonimmune (mitogen)-mediated lymphocyte responsiveness; if insufficient control reactivity is documented, IGRA test results are reported as indeterminate or invalid and are considered invalid. (See "Use of interferon-gamma release assays for diagnosis of tuberculosis infection (tuberculosis screening) in adults", section on 'Uninterpretable results'.)

Issues related to dosing, administration, and false-positive and false-negative IGRA results in children are similar as for adults; these are discussed in detail separately. (See "Use of interferon-gamma release assays for diagnosis of tuberculosis infection (tuberculosis screening) in adults".)

Use of both TST and IGRA may increase sensitivity for diagnostic evaluation of children with suspected TB. In one study including 69 children age 5 to 18 years with TB who underwent both IGRA and TST, the sensitivity of IGRA was greater than TST (95 versus 83 percent); among children <5 years, the sensitivity was similar [52].

Additional issues related to use of IGRAs are discussed further separately. (See "Use of interferon-gamma release assays for diagnosis of tuberculosis infection (tuberculosis screening) in adults".)

Imaging

Chest radiography — Frontal and lateral chest radiography is a critical tool for diagnosis of intrathoracic TB in children (image 1A-K) [53,54]. The most common chest radiograph finding in a child with TB disease is a primary complex, which consists of opacification with hilar, mediastinal, or subcarinal lymphadenopathy, in the absence of notable parenchymal involvement [11]. When adenopathy advances, obstruction of the neighboring airway may result in consolidation or a segmental lesion, leading to collapse of a lung segment or lobe in the setting of infiltrate and atelectasis.

In a study of 326 traced contacts under five years of age, 9 percent of children diagnosed with intrathoracic TB were asymptomatic and had radiographic findings only of the primary complex [55]. A miliary pattern of opacification is highly suspicious for TB, as is opacification that does not improve or resolve following a course of antibiotics [11].

Adolescents with TB generally present with typical adult disease findings of upper lobe infiltrates, pleural effusions, and cavitation on chest radiograph [11]. However, they also can present with intrathoracic adenopathy and radiographic shadows similar to those seen in younger children. (See "Diagnosis of pulmonary tuberculosis in adults".)

Computed tomography scan — Computed tomography (CT) scan of the chest may be used to further delineate the anatomy for cases in which radiographic findings are equivocal. Endobronchial involvement, necrosis of lymph nodes, bronchiectasis, and cavitation may be more readily visualized on CT scans than chest radiographs [56]. However, there is no role for routine use of CT scans in the evaluation of an asymptomatic child with a normal chest radiograph as treatment regimens are based on chest radiography findings [9]. One review of CT scans use among adolescents with intrathoracic TB found that prompt ordering of sputum specimens for microscopy, molecular testing and culture would avoid unnecessary CT scans [57].

In the setting of tuberculous meningitis, CT scan or magnetic resonance imaging of the head is useful. Hydrocephalus and basilar meningeal enhancement are observed in 80 and 90 percent of cases, respectively, and evidence of vasculitis or stroke is not uncommon; chest radiography may be normal [9].

Laboratory studies — The likelihood of achieving bacteriological confirmation depends on the extent of disease and the type of specimen. The initial approach for diagnosis of TB in children consists of sputum examination: expectorated (for older children or adolescents), swallowed and collected as gastric contents (young children), or induced (all ages). Gastric aspiration is the primary method of obtaining material for AFB smear and culture from young children.

Sputum specimens should be sent for examination by smear microscopy and mycobacterial culture. Nucleic acid amplification (NAA) testing can be used for rapid diagnosis of an organism belonging to the M. tuberculosis complex (24 to 48 hours) in patients for whom the suspicion for TB is moderate to high [58]. (See "Diagnosis of pulmonary tuberculosis in adults".)

Acid-fast bacilli smear and culture — At least three respiratory specimens (including gastric aspirates) should be obtained for AFB smear and mycobacterial culture for children with suspected pulmonary TB [59].

●Sputum — Obtaining expectorated sputum from children for detection of AFB is difficult and its examination is of low yield (15 percent or less for microscopic examination and 30 percent or less for culture) [60,61]. However, most adolescents can produce expectorated sputum spontaneously.

Sputum induction has higher yield than expectorated sputum in children, and the use of sputum induction for obtaining TB diagnostic specimens in children is increasing. Sputum induction is performed via administration of aerosolized heated saline combined with salbuterol (or similar drug to minimize wheezing), followed by suctioning to capture the expectorated sputum. A key feature is suctioning the induced sputum before the young child swallows it. In a study of 250 children (median age 13 months), sputum induction was found to be a safe and effective procedure in children as young as one month of age [60]. In two studies, outpatient sputum induction yielded culture results comparable to or better than inpatient gastric aspiration [33,60], but other studies have demonstrated lower yields. Minimal adverse effects associated with the procedure included coughing, epistaxis, vomiting, and wheezing. Children with underlying reactive airways disease should receive pretreatment with a bronchodilator to prevent bronchospasm during or following the procedure [60].

●Gastric aspirate - Early morning gastric contents collected from a fasting child contain sputum swallowed during the night. Gastric aspiration specimens may be obtained in the inpatient or outpatient setting [62,63]. Ideally, three early morning samples collected on different days before the child eats or ambulates optimize specimen yield [64].

Gastric aspiration remains the most common method for obtaining respiratory samples from children (in facilities where this procedure may be performed). In general, cultures of gastric aspirate specimens are positive for M. tuberculosis in only 30 to 40 percent of cases [65]. Smears are even less reliable, with positive results in fewer than 10 percent of cases [65]; in addition, false-positive smear results caused by the presence of nontuberculous mycobacteria can occur [22].

●Nasopharyngeal aspirate - In settings where neither gastric aspiration nor sputum induction is readily available, a similar yield has been observed with less invasive laboratory testing including nasopharyngeal aspiration [66]. In one study including 294 Kenyan children <5 years of age, testing (via Xpert MTB/RIF and culture) of two nasopharyngeal aspirate samples, one nasopharyngeal aspirate sample and one stool sample, or one nasopharyngeal aspirate sample and one urine sample was associated with sensitivities of 74, 71, and 69 percent (compared with culture), respectively [67,68].

●Bronchoscopy - Bronchoscopy may be helpful for evaluation of patients with severe disease and nondiagnostic test results with other specimens. In a South African study including 146 children age 3 months to 13 years, diagnostic confirmation via bronchoalveolar (BAL) lavage was observed in 25 percent of case; diagnostic confirmation via BAL as well as other respiratory samples and was observed in 44 percent of cases [69].

●Other specimens - Other body fluid and/or tissue samples may be necessary in some circumstances, depending on suspicion for extrapulmonary TB. The approach to these diagnostic tools is outlined separately. (See "Clinical manifestations, diagnosis, and treatment of miliary tuberculosis" and "Tuberculous lymphadenitis" and "Diagnosis of pulmonary tuberculosis in adults".)

Xpert MTB/RIF and other rapid tests

●Xpert MTB/RIF and Xpert MTB/RIF Ultra  

•General principles − The Xpert MTB/RIF assay is an automated nucleic acid amplification (NAA) test that can simultaneously detect M. tuberculosis complex organisms and rifampin resistance. This test is substantially more sensitive than smear microscopy in children [70,71]. The Xpert MTB/RIF is the only NAA test approved by the US Food and Drug Administration (FDA) for use in the United States; it is approved only for sputum (induced or spontaneous), in individuals who have been on antituberculous therapy for less than three days.

Xpert MTB/RIF Ultra is a newer version of the test which was developed to improve sensitivity by using an updated cartridge and improved software. Xpert MTB/RIF Ultra is not available in the United States.

The WHO supports use of Xpert MTB/RIF Ultra with sputum and nasopharyngeal aspirate specimens for diagnosis of TB in children <10 years [27]. In addition, in 2021, the WHO endorsed use of Xpert MTB/ RIF Ultra with gastric aspirate or stool as an initial diagnostic test for TB and detection of rifampicin resistance in children age <10 years [27].

•Use with respiratory tract specimens − In a randomized trial including 452 children in South Africa with suspected pulmonary TB, 6 percent had a positive sputum smear, 16 percent had a positive sputum culture, and 13 percent had a positive sputum Xpert MTB/RIF result [70]. The initial Xpert MTB/RIF test detected 100 percent of culture-positive cases that were smear positive but only 33 percent of those that were smear negative; a second Xpert MTB/RIF test improved the detection of smear-negative cases to 61 percent. Overall, with induced sputum specimens, the sensitivity and specificity were 59 and 99 percent, respectively, for one Xpert MTB/RIF test and 76 and 99 percent for two Xpert MTB/RIF tests. Test performance was unaffected by patient HIV status. Results for Xpert MTB/RIF were available within a median of one day (versus 12 days for culture).

In one Chinese study including 93 children with pulmonary TB and 128 children with other respiratory tract infections, Xpert MTB/RIF Ultra on bronchoalveolar lavage samples detected M. tuberculosis in 58 percent of TB cases with negative culture or AFB stain results; the specificity was 98 percent [71]. Among 164 children for whom both Ultra and Xpert were performed, the sensitivity was 80 and 67 percent, respectively, and Ultra identified an additional six children who had negative Xpert results.

•Low sensitivity with negative sputum smear − While the Xpert MTB/RIF test appears to be highly specific, its sensitivity for sputum smear negative TB in children remains low. Since culture was used as the gold standard in both studies described above, the sensitivity of Xpert MTB/RIF is expected to be even lower in sputum culture-negative, clinically confirmed cases. Therefore, it cannot replace current methods used to suspect and diagnose TB in infants and children. Most children in the study presented with symptomatic pulmonary TB and extensive disease. The Xpert MTB/RIF test is meant to be a rapid diagnostic test that may take the place of sputum microscopy but not mycobacterial culture [72]. A negative Xpert MTB/RIF test should be interpreted in the context of the child's clinical and radiographic findings. Sputum culture remains a more sensitive test and is required to detect the full drug susceptibility profile of the infecting organism. Further study of the assay is needed in areas with high and low prevalence of TB. (See "Diagnosis of pulmonary tuberculosis in adults".)

•Use with alternate specimens − Use of the Xpert MTB/RIF test on alternate specimens may be beneficial, especially in settings where induced sputum and mycobacterial culture are not feasible. While Xpert MTB/RIF is not FDA approved in the United States for specimens other than sputum, some laboratories may offer validated, “off-label” testing for other sample types (eg, bronchial fluid).

The WHO endorsement of Xpert MTB/ RIF Ultra with gastric aspirate or stool is based on a systematic review and meta-analysis of six studies from nine countries (including four high-TB and five high-TB-HIV burden countries), evaluated the sensitivity and specificity of Xpert MTB/RIF Ultra among 659 participants for gastric aspirate samples and 1278 participants for stool samples [27]. Compared with culture, the sensitivity and specificity for gastric aspirate samples were 64 and 95 percent, respectively; the sensitivity and specificity for stool samples were 53 and 98 percent, respectively [27].

In a study of 454 Bangladeshi children with presumptive pulmonary TB who underwent stool-based Xpert and Xpert Ultra testing, sensitivity and specificity for Xpert were 38 and 100 percent, respectively; sensitivity and specificity for Xpert Ultra were 59 and 88 percent, respectively (using sputum culture, Xpert, and/or Xpert Ultra as the reference) [73].

In a systematic review and meta-analysis including 12 studies and 2177 children, Xpert MTB/RIF used with stool specimens and compared with bacteriologically confirmed TB with respiratory specimens, the pooled sensitivity of Xpert MTB/RIF with stool specimens was 50 percent (with a high degree of heterogeneity between studies); the pooled specificity was 99 percent [74].

•Use in setting of extrapulmonary TB − Data on the use of Xpert MTB/RIF in children with extrapulmonary tuberculosis are promising. In one study including 23 children in South Africa with musculoskeletal tuberculosis (confirmed by histology), the sensitivity and specificity of Xpert MTB/RIF were 74 and 100 percent, respectively; the sensitivity and specificity of culture were 61 and 100 percent, respectively [75]. In the systematic review of Xpert MTB/RIF in children, Xpert MTB/RIF pooled sensitivity and specificity verified by culture were both 90 percent [76]. In another study including 55 children in South Africa with tuberculous meningitis, combining results from culture, GenoType MTBDRplus, and Xpert MTB/RIF yielded sensitivity and specificity of 56 and 98 percent, respectively [77].

●Molecular line probe assays − Molecular line probe assays are rapid tests that can be used to detect the presence of M. tuberculosis as well as genetic mutations that confer rifampin resistance alone or in combination with isoniazid resistance. These assays have high sensitivity (90 to 97 percent) and specificity (99 percent) compared with drug susceptibility testing [78].

Urine antigen detection — Urine-based detection of mannosylated mycobacterial cell wall glycolipid lipoarabinomannan (urine LAM assay) is an assay for diagnosis of TB.

For regions of the world with high incidence of HIV and TB, we are in agreement with the WHO, which favors use of urine LAM testing (in addition to routine diagnostic tests) for patients with HIV infection who have signs and symptoms of pulmonary and/or extrapulmonary tuberculosis and CD4 ≤100 cells/microL, and for any patient with HIV infection who is seriously ill.

The high specificity of LAM testing suggests this can be used as a “rule-in” test among those at highest risk. In one cohort including 204 children in South Africa, the specificity was 92 percent overall and 96 percent in those >2 years; the sensitivity was 42 percent overall, but 60 percent among the 40 children with HIV infection and 62 percent among the 44 children who were malnourished [79,80]. (See "Clinical manifestations, diagnosis, and treatment of miliary tuberculosis", section on 'Urine antigen test in HIV infection'.)

The urine LAM assay may also have prognostic utility. In one study including 137 children with HIV infection in Malawi initiating antiretroviral therapy, the six-month mortality rate was 3.7-fold higher among those with a positive urine LAM assay [81].

Detection of drug resistance — Mycobacterial culture with second-line drug susceptibility testing (DST) should be performed whenever possible [82,83]. Concerted effort should be made to obtain multiple high-quality specimens from the most accessible site(s) of disease [83].

Rapid molecular tests are useful for providing some information regarding susceptibility in the absence of culture data. These include Xpert MTB/RIF (provides information regarding susceptibility to rifampin) and MTBDRsl (a line-probe assay, not approved in the United States, that provides information regarding susceptibility to fluoroquinolones and injectable antituberculous agents). Patients diagnosed with rifampin-resistant TB (RR-TB) via Xpert MTB/RIF are typically treated as multidrug-resistant (MDR)-TB whether or not isoniazid resistance is ever confirmed, hence the new designation of MDR/RR-TB.

Whole genome sequencing with molecular determination of drug resistance is available for patients in the United States from the United States Centers for Disease Control and Prevention (CDC) laboratory for Xpert-positive patient sputum sediment or for culture growth from patients at risk for drug resistance. This service generally provides results to the submitter within five working days from receipt of the specimen and may be initiated through state laboratories [84].

In 2016, the World Health Organization first recommended use of MTBDRsl for identifying patients with MDR-TB or RR-TB who are candidates for treatment with a shortened treatment regimen [85]. The assay may be used as an initial diagnostic test; however, phenotypic culture-based drug-susceptibility testing is required to detect resistance to other drugs and to monitor for emergence of additional resistance during treatment. Issues related to diagnosis of drug resistance are discussed further separately. (See "Diagnosis of pulmonary tuberculosis in adults".)

Pending definitive diagnostic information, in some circumstances it may be reasonable to presume presence of drug-resistant TB based on clinical criteria including signs and symptoms, radiographic findings, history of contact with a presumed or confirmed source case with drug-resistant TB, and failure to respond to first-line TB drugs [83].

Investigational diagnostic tools — Because of the difficulty in achieving microbiologic confirmation of clinically suspected TB in children, interest has grown in alternate methods of laboratory diagnosis. One candidate method is microarray analysis of blood samples to identify a pattern of ribonucleic acid (RNA) expression that is associated with active TB infection. One study identified an RNA expression risk score that distinguished with high sensitivity and specificity culture-confirmed TB from latent TB and diseases other than TB among children in sub-Saharan Africa. However, the risk score did not perform as well among children with clinically diagnosed, culture-negative TB [86]. Moreover, in order to be a practical tool in resource-limited settings, where its use would be most relevant, the technology would require substantial modification to reduce cost and complexity.

TREATMENT OF INTRATHORACIC TB

General treatment principles

●Empiric treatment – In many cases of TB in children, laboratory confirmation is never established (particularly among children under five years of age). In such cases, a presumptive diagnosis may be made based on clinical and radiographic response to empiric treatment.

●Drug susceptibility testing – Drug susceptibility testing should be performed on initial isolates from each site of disease. If sputum cultures are negative, the drug susceptibility test results for isolates from a presumed source case (if known or available) should guide decisions about treatment with respect to susceptibility.

●Nitrosamine impurities – In August 2020, the US Food and Drug Administration (FDA) announced detection of nitrosamine impurities in samples of rifampin and rifapentine [87]. Some such compounds have been implicated as possible carcinogens in long-term animal studies, with toxicity largely related to cumulative exposure. For treatment of TB disease, we favor continued use of rifampin if acceptable to the patient, as the risk of not taking rifampin likely outweighs any potential risk from nitrosamine impurities. This is concordant with FDA and CDC recommendations.  

Drug-susceptible TB

Available regimens — There are three available treatment regimens; these include the traditional (≥6 months) regimen and two shortened regimens:

●Traditional six-month regimen – Isoniazid, rifampin, pyrazinamide, and ethambutol for eight weeks followed by isoniazid and rifampin for 16 weeks, for most forms of TB (excluding TB meningitis and osteoarticular TB) (table 4 and table 5).

●Shortened four-month regimen (for nonsevere, smear-negative disease) – Isoniazid, rifampin, and pyrazinamide (with or without ethambutol, according to local guidelines) for eight weeks, followed by isoniazid and rifampin for eight weeks (table 4 and table 5).

●Rifapentine-moxifloxacin-based four-month regimen – Rifapentine, isoniazid, pyrazinamide, and moxifloxacin for eight weeks, followed by rifapentine, isoniazid, and moxifloxacin for nine weeks (table 6). This regimen is for drug-susceptible pulmonary TB, in the absence of extrapulmonary involvement.

Details and indications for each of the regimens are discussed below.

Traditional regimen (≥6 months) — The traditional regimen for treatment of pediatric TB is outlined in the tables (table 4 and table 5) [22,88]. Issues related to administration logistics and monitoring for this regimen are discussed separately. (See "Treatment of drug-susceptible pulmonary tuberculosis in nonpregnant adults without HIV infection", section on 'Traditional regimen (≥6 months)'.)

Shortened (four-month) regimen for nonsevere, smear-negative disease — The shortened (four-month) regimen is outlined in the tables (table 4 and table 5); it was endorsed by the WHO in 2022 for treatment of children between 3 months and 16 years of age with nonsevere, smear-negative, presumed drug-susceptible disease (defined as pulmonary TB confined to one lobe); exclusions include cavitary disease, miliary TB, complex pleural effusion, clinically significant airway obstruction, and peripheral lymph node involvement [88]. In settings where rapid molecular testing with Xpert MTB/RIF is available, a result of negative, very low, low, or trace positive may be used as a proxy for smear-negative disease and to confirm rifampin susceptibility. (See 'Regimen selection' below.)

The regimen was evaluated in the SHINE trial (Shorter Treatment for Minimal Tuberculosis in Children) which included 1204 children <16 years (median age 3.5 years) with nonsevere, smear-negative, presumed drug-susceptible TB (including 127 patients with HIV infection) in South Africa, Uganda, Zambia, and India [89,90]. Children were randomly assigned to treatment with a four-month regimen (as summarized above) or a traditional six-month regimen (which included a continuation phase of 16 weeks rather than 8 weeks).

The four-month regimen was found to be noninferior to the six-month regimen; the primary outcome (defined as treatment failure [extension, change, or restart of treatment or TB recurrence], loss to follow-up during treatment, or death by 72 weeks) was observed among 3 percent of patients in each group (adjusted difference -0.4 percentage points, 95% CI -2.2 to 1.5). Adverse events were comparable between the groups (occurring in approximately 8 percent of participants in each group); adherence and retention rates were high (94 and 95 percent, respectively).

Rifapentine-moxifloxacin-based four-month regimen — This regimen is outlined in the table (table 6) [88,91,92]. It was endorsed by the WHO in June 2021, and the CDC issued guidance for use in February 2022 [88,92]. The regimen may be used for nonpregnant patients (age ≥12 years, body weight ≥40 kg) with drug-susceptible pulmonary TB, in the absence of extrapulmonary involvement. (See 'Regimen selection' below.)

Important considerations for use of this regimen include ensuring fluoroquinolone susceptibility testing and monitoring for toxicity (eg, QT prolongation) in the setting of prolonged moxifloxacin administration. Issues related to administration logistics and monitoring for this regimen are discussed separately. (See "Treatment of drug-susceptible pulmonary tuberculosis in nonpregnant adults without HIV infection", section on 'Rifapentine-moxifloxacin-based four-month regimen'.)

The study evaluating the rifapentine-moxifloxacin 4-month regimen included 63 adolescents ≥12 years of age; a 4-month regimen consisting of rifapentine, isoniazid, pyrazinamide, and moxifloxacin was found to be noninferior to a traditional 6-month regimen consisting of rifampin, isoniazid, pyrazinamide, and ethambutol [91]. The data from this trial are discussed further separately. (See "Treatment of drug-susceptible pulmonary tuberculosis in nonpregnant adults without HIV infection", section on 'Rifapentine-moxifloxacin-based four-month regimen'.)

Regimen selection — The approach to regimen selection should be guided by patient age and clinical factors (algorithm 1):

●Age <3 months − For children age <3 months, we treat with the traditional regimen (table 4 and table 5). (See 'Traditional regimen (≥6 months)' above.)

●Age 3 months to 11 years − For children age 3 months to 11 years with nonsevere, smear-negative, presumed drug-susceptible disease, we treat with the shortened (four-month) regimen (table 4 and table 5). (See 'Shortened (four-month) regimen for nonsevere, smear-negative disease' above.)

For children in this age group who do not meet criteria for the shortened regimen, we treat with the traditional regimen.

●Age 12 to 16 years − For children age 12 to 16 years with nonsevere, smear-negative, presumed drug-susceptible disease, we treat with the shortened (four-month) regimen. (See 'Shortened (four-month) regimen for nonsevere, smear-negative disease' above.)

For children in this age group who do not meet criteria for the shortened regimen, options include the traditional regimen or the rifapentine-moxifloxacin-based regimen (table 6).

While some favor rifapentine-moxifloxacin-based regimen given the shorter duration, others favor the traditional regimen pending further outcome data with the rifapentine-moxifloxacin-based regimen. Additional factors warranting consideration in selection of the rifapentine-moxifloxacin-based regimen (including screening for risk of QT prolongation and other contraindications) are discussed separately. (See "Treatment of drug-susceptible pulmonary tuberculosis in nonpregnant adults without HIV infection", section on 'Rifapentine-moxifloxacin-based four-month regimen'.)

●Age 17 to 18 years − For children age 17 to 18 years, options include the traditional regimen or the rifapentine-moxifloxacin-based regimen (table 6).

While some favor rifapentine-moxifloxacin-based regimen given the shorter duration, others favor the traditional regimen pending further outcome data with the rifapentine-moxifloxacin-based regimen. Additional factors warranting consideration in selection of the rifapentine-moxifloxacin-based regimen (including screening for risk of QT prolongation and other contraindications) are discussed separately. (See "Treatment of drug-susceptible pulmonary tuberculosis in nonpregnant adults without HIV infection", section on 'Rifapentine-moxifloxacin-based four-month regimen'.)

Drug-resistant TB — In general, the approach for treatment of drug-resistant TB in children is similar to the approach for adults [93]. Children may be treated with a conventional regimen or, if they meet appropriate criteria, a shortened 9- to 12-month regimen [94,95]. However, the approach to selection of antituberculous agents for children may differ from that of adults in some circumstances; this is discussed further below, and expert consultation is advised. Pediatric dosing of second-line antituberculous drugs is summarized in the table (table 7). Issues related to treatment of drug-resistant TB are discussed in detail separately. (See "Treatment of drug-resistant pulmonary tuberculosis in adults", section on 'General principles'.)

Whenever possible, it is desirable to design a regimen that avoids use of an injectable agent. This is particularly true for very young children and those with mild TB disease (ie, those with good nutrition, mild forms of extrapulmonary disease, and the absence of cavitation on chest radiography and HIV infection) [94]. Use of second-line antituberculous agents in children is complicated by the lack of pediatric formulations for most of these drugs, which can lead to under- or over-dosing.

Bedaquiline and delamanid are two relatively new oral agents for treatment of drug-resistant TB.

According to WHO guidance issued in 2021, bedaquiline and delamanid may be used to treat children and adolescents of all ages with multidrug-resistant (MDR)/rifampin-resistant TB [27]. Bedaquiline may be used in the shorter, all-oral bedaquiline-containing regimen, and either bedaquiline or delamanid may be used as part of a longer treatment regimen [27]. This guidance facilitates the feasibility of designing all-oral treatment regimens for children and adolescents of all ages.

Development of a pediatric bedaquiline formulation is underway; in the meantime, the available tablet formulation may be mixed with soft food or dissolved in water for pediatric administration [96]. In one study including 27 children without HIV infection ages 10 to 17 who received bedaquiline for MDR-TB, good treatment responses were observed and there was no drug discontinuation due to adverse reactions [97].

Coadministration of bedaquiline with efavirenz or lopinavir/ritonavir should be avoided. Coadministration of bedaquiline with efavirenz is associated with reduced bedaquiline levels; coadministration of bedaquiline with lopinavir/ritonavir is associated with increased bedaquiline levels [98].

There is interim guidance for use of delamanid in children ages 6 to 17 years [99]. Dosing of delamanid may be difficult in children between 3 and 5 years of age because the only available formulation is a 50 mg (adult) tablet that cannot be split accurately (which may affect bioavailability); in addition, the contents are bitter and unpalatable [94]. Limited clinical outcome data and pharmacokinetic studies suggest higher doses of delamanid are necessary for effective treatment of children <3 years [27].

If an injectable agent must be used, kanamycin and capreomycin are preferred; amikacin or streptomycin should be used in children only when other options are not possible, when DST confirms susceptibility, and monitoring for both ototoxicity and nephrotoxicity can be performed [94]. Regular audiometry is critical for monitoring children receiving these drugs, given potential association between early hearing loss and diminished language acquisition [94].

Drug toxicity is common; in one meta-analysis of children treated for MDR-TB, it was reported in 39 percent of cases [100]. Similarly, in the series from Peru, adverse effects occurred in 42 percent of cases, although no events required suspension of therapy for >5 days [101]. Children on treatment for drug-resistant TB should be monitored at least monthly for adherence, response to treatment (eg, sputum analysis for those with pulmonary TB), and potential adverse events.

Overall, high rates of treatment success have been achieved among children with MDR-TB [102-104]. In a systematic review and meta-analysis including data from 875 patients in 18 countries, the treatment success rate was 78 percent [104]. Use of bedaquiline, delamanid, or linezolid can improve outcomes in children with MDR-TB; in one study including 119 children (of whom 18 received one or more of these drugs), favorable outcome (successful treatment completion or documented microbial cure) was observed more frequently among those who received these agents than among those who received standard second-line drugs (94 versus 80 percent) [105].

Among those with bacteriologically confirmed TB disease and HIV infection who did not receive any ART during their MDR-TB therapy, the treatment success rate was 56 percent; among those who did receive the treatment success rate was 82 percent [104]. Most children receive individualized treatment which may contribute to the generally good outcomes observed.

Similar outcomes have been observed among children with extensively drug-resistant tuberculosis (XDR-TB); however, data are limited. In a systematic review that included 38 children with XDR-TB from 11 countries, 81 percent had favorable treatment outcomes, while 11 percent died, and 3 percent failed treatment [106].

HIV coinfection — In children with HIV infection who are not on antiretroviral therapy (ART), ART should be initiated as soon as possible and within eight weeks of starting antituberculous therapy or within two weeks if the CD4 count is <50 cells/mm³ [107,108]. The optimal time to start ART in children with central nervous system TB is unclear; studies are few and results are mixed with some showing better outcomes when ART is delayed (likely due to fewer instances of Immune Reconstitution Inflammatory Syndrome [IRIS], see below) while others show worse outcomes with delayed ART [109-111]. Selection of an optimal ART regimen should be made in consultation with a pediatric HIV specialist. (See "Immune reconstitution inflammatory syndrome" and "Central nervous system tuberculosis: An overview".)

Paradoxical reactions consistent with IRIS can occur following the initiation of antituberculous therapy and/or ART [112,113]. Unexplained deterioration among immunocompetent children receiving appropriate therapy for pulmonary and/or extrapulmonary TB consistent with TB-IRIS has also been described [114,115]. In one study of 110 children, clinical or radiographic deterioration was observed in 14 percent of cases after initiating therapy (range 10 to 181 days; mean 80 days) [114]. The most common complication was enlarging intrathoracic lymphadenopathy, often causing airway compromise. Deterioration was more likely among children with weight-for-age ≤25th percentile and multiple sites of disease. All children achieved clinical or radiographic cure; corticosteroids were administered in 60 percent of cases. In another study of 115 immunocompetent children, 12 developed paradoxical worsening within 15 to 75 days (median 39 days) of starting TB therapy; children with paradoxical reactions tended to be younger (median age at diagnosis of 26 months versus 66 months) and had never received BCG vaccination [115]. The most common manifestation was worsening of preexisting pulmonary lesions, observed in 75 percent, while 25 percent had new disease present in new anatomic locations.

PREVENTION — Measures for prevention of TB include routine infant Bacille Calmette-Guérin (BCG) immunization, infection control interventions, and prompt identification and treatment of latent TB infection (LTBI). Suspicion of TB disease in a child should be reported to the health department so that contact tracing, which may lead to identification of the source case, can be started right away. Children, particularly those that are prepubertal, are rarely contagious. Therefore, whenever facility-based transmission is suspected, it is important to evaluate not only the health care workers but also the child’s caregivers. In a prospective evaluation of 105 adults accompanying children hospitalized for suspected TB, 15 percent had previously undetected TB [116]. In some hospitals, chest radiographs of caregivers of children with suspected TB are routinely obtained to detect previously undiagnosed TB, thereby ensuring prompt detection and treatment of the caregiver and halting potential further transmission in the hospital environment. (See "Tuberculosis infection (latent tuberculosis) in children" and "Tuberculosis transmission and control in health care settings", section on 'Contact investigation'.)

Issues related to treatment of LTBI following contact with a source case are discussed separately. (See "Tuberculosis infection (latent tuberculosis) in children".)

In countries where TB is endemic, routine childhood BCG immunization is an important preventive measure. Issues related to use of BCG are discussed separately. (See "Vaccines for prevention of tuberculosis".)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Diagnosis and treatment of tuberculosis".)

SUMMARY AND RECOMMENDATIONS

●Estimating the global burden of tuberculosis (TB) disease in children is challenging due to the lack of a standard case definition, the difficulty in establishing a definitive diagnosis, the frequency of extrapulmonary disease in young children, and the relatively low public health priority given to TB in children relative to adults. As a result, there is likely significant underreporting of childhood TB from high-prevalence countries. (See 'Epidemiology' above.)

●Children under the age of five years represent an important demographic group for understanding TB epidemiology; in this group, TB frequently progresses rapidly from primary infection or latent infection to TB disease. Therefore, these children serve as sentinel cases, indicating recent and/or ongoing transmission in the community. (See 'Epidemiology' above.)

●Common symptoms of pulmonary TB in children include cough (chronic, without improvement for more than three weeks), fever (more than 38ºC for more than two weeks), and weight loss or failure to thrive. Physical exam findings may suggest the presence of a lower respiratory infection, but there are no specific findings to confirm that pulmonary TB is the cause. (See 'Pulmonary tuberculosis' above.)

●The clinical presentation of extrapulmonary TB depends on the site of disease. The most common forms of extrapulmonary disease in children are TB of the superficial lymph nodes and of the central nervous system. Infants have the highest risk of progression to TB disease with dissemination (miliary TB) and meningeal involvement. (See 'Extrapulmonary tuberculosis' above.)

●Forms of perinatal TB include congenital and neonatal disease. Congenital TB is very rare and most often is associated with maternal tuberculous endometritis or miliary TB. Neonatal TB is more common and develops following exposure of an infant to aerosolized respiratory secretions of a contagious person, often the infant’s mother. (See 'Perinatal infection' above.)

●TB in children is often diagnosed clinically; in many cases, laboratory confirmation is never established (particularly among children under five years of age). Diagnosis is often based on the presence of the classic triad: (1) recent close contact with an infectious case, (2) a positive tuberculin skin test (TST) or interferon-gamma release assay (IGRA), and (3) suggestive symptoms and findings on chest radiograph or physical examination. (See 'Diagnosis' above.)

●In children, the TST or IGRA may be used as a tool to support diagnosis of TB disease or latent TB infection (LTBI; although, in adults, the TST or IGRA may be used only for diagnosis of LTBI, not TB disease). The TST or IGRA is helpful for diagnosis of TB in children only in circumstances when it is positive (table 2). (See 'Tuberculin skin test' above.)

●The most common chest radiograph finding in a young child with TB disease is a primary complex, which consists of opacification with hilar or subcarinal lymphadenopathy, in the absence of notable parenchymal involvement. In older children and adolescents, chest radiography findings may appear similar to those observed among adults. (See 'Imaging' above.)

●Gastric aspiration is the primary method of obtaining material for diagnostic testing (including acid-fast bacilli smear, rapid tests, and culture) from young children, since these patients lack sufficient tussive force to produce adequate sputum samples by expectoration alone. Alternative approaches include sputum induction or expectoration (for older children). For diagnosis of extrapulmonary TB, specimens for culture should be collected from any site where infection is suspected. Diagnosis of TB should also prompt HIV testing. (See 'Laboratory studies' above.)

●The approach to regimen selection should be guided by patient age and clinical factors (algorithm 1). (See 'Regimen selection' above.)

•For children age 3 months to 16 years with nonsevere, smear-negative, presumed drug-susceptible disease, we suggest treatment with the shortened (four-month) regimen (table 4 and table 5) (Grade 2B). (See 'Shortened (four-month) regimen for nonsevere, smear-negative disease' above.)

•For children age <3 months, and for children age 3 months to 11 years who do not meet criteria for the shortened regimen, we suggest treatment with the traditional regimen (table 4 and table 5) (Grade 2C). (See 'Traditional regimen (≥6 months)' above.)

•For children age 12 to 16 years who do not meet criteria for the shortened regimen, and for children age 17 to 18 years, options include the traditional regimen or the rifapentine-moxifloxacin-based regimen (table 6). While some favor rifapentine-moxifloxacin-based regimen given the shorter duration, others favor the traditional regimen pending further outcome data with the rifapentine-moxifloxacin-based regimen.

Additional factors warranting consideration in selection of the rifapentine-moxifloxacin-based regimen (including screening for risk of QT prolongation and other contraindications) are discussed separately. (See "Treatment of drug-susceptible pulmonary tuberculosis in nonpregnant adults without HIV infection", section on 'Rifapentine-moxifloxacin-based four-month regimen'.)