4.2 Diagnostic approach

Select language:
On this page

    Children with TB usually have non-specific symptoms. Clinicians should therefore look for TB, particularly in children:

    • Under 2 years of age, or
    • With HIV infection or SAM, or
    • In contact with a person with TB, or
    • Not responding to antibacterial and/or nutritional treatment.


    The diagnosis of TB in children, particularly those under 5 years, is often based on a combination of history of exposure to a person with TB, clinical assessment and investigations, such as radiology, when available.


    In children at high risk of death from TB, treatment should be initiated as soon a TB diagnosis is considered likely.

    In children not at high risk of death from TB, the diagnosis may not be made at the first consultation. A second consultation after one to two weeks is often necessary to reassess the clinical status.


    The diagnosis is often made without bacteriological confirmation as:

    • Children under 5 years have low bacillary load and bacteriological tests are often negative.  

    • Specimens for diagnosis of EPTB may be difficult to collect.

    TB is bacteriologically confirmed in only 20 to 30% of children [1] Citation 1. Seddon JA, Jenkins HE, Liu L, Cohen T, Black RE, Vos T, Becerra MC, Graham SM, Sismanidis C, Dodd PJ. Counting children with tuberculosis: why numbers matter. Int J Tuberc Lung Dis. 2015 Dec;19 Suppl 1(0 1):9-16.


    To facilitate the diagnosis of PTB and enable rapid treatment in children, WHO has developed diagnostic algorithms (Section 4.3).
    The diagnosis of EPTB uses the same diagnostic approach. However, no evidence-based algorithms are currently available.


    A trial of treatment with TB drugs is not recommended as a method to diagnose TB.

    Once a decision is made to treat TB in a child, a full course of treatment should be given.

    4.2.1 History of exposure to tuberculosis

    Children are at risk of TB if they are exposed to a person with TB.
    They are at higher risk of TB if:

    • The index case is a household or close contact.
    • The index case has PTB, sputum smear positive or cavities on chest x-ray.
    • The exposure to the index case has occurred in the past 12 months.


    Note: conversely, when TB is diagnosed in children, it is important to detect the index case and any other undiagnosed household member(s) or close contact(s).

    4.2.2 Clinical assessment

    Symptoms suggestive of tuberculosis

    Ask if the child has symptoms commonly associated with TB:

    • Cough for more than 2 weeks.
    • Fever for more than 2 weeks.
    • Night sweats that soak the bed or clothes.
    • Weight loss or poor/no weight gain.
    • Fatigue, reduced playfulness, loss of appetite.
    • Haemoptysis (rare in children).
    • Non-painful, enlarged cervical, submandibular, or axillary lymph nodes.
    • Rapid breathing.
    Physical examination and growth assessment

    Look for signs suggestive of TB:

    • Fever, tachypnoea, tachycardia.
    • Weight loss, growth curve flattening, underweight or malnourished according to weight for height and/or mid-upper arm circumference.
    • Abnormal pulmonary auscultation.
    • Signs of respiratory distress and SpO2 < 90-92%.
    • Lethargy, altered mental status (may indicate TB meningitis).
    • Signs of EPTB:
      • Highly suggestive, e.g.:
        • Angular deformity of the spine, loss of ability to walk.
        • Cervical lymph node with fistula formation.
      • Requiring further investigation, e.g.:
        • Sub-acute meningitis not responding to antibiotic treatment.
        • Ascites.
        • Lymph node without fistula formation.
        • Non-painful enlarged joint.


    HIV status should be assessed in all children with presumed or confirmed TB.

    Clinical review

    If diagnosis is not made at the first consultation, reassess the child (signs/symptoms suggestive of TB and growth) within one to two weeks maximum.


    The following are suggestive of TB:

    • Persistent or worsening pneumonia despite non-TB antibiotic treatment.
    • No weight gain or weight loss despite nutritional support or treatment.
    • Persistent fever after other causes have been ruled out or treated (e.g. malaria).
    • Persistent or worsening fatigue, reduced playfulness, loss of appetite.

    4.2.3 Baseline investigations

    When PTB or EPTB is suspected, perform bacteriological tests, lateral flow urine lipoarabinomannan assay if indicated, and radiography if available.

    Bacteriological tests

    Rapid molecular tests (RMTs) should be performed on respiratory, stool or extrapulmonary (EP) specimens as the initial diagnostic test.

    As the sensitivity of Xpert MTB/RIF Ultra is higher than that of Xpert MTB/RIF, preferably use MTB/RIF Ultra for the detection of TB and rifampicin-resistance (Chapter 3).


    Sputum specimens can be difficult to obtain in children. Explanation and encouragement are important. Chest clapping may help expectoration.

    If sputum cannot be obtained spontaneously, more invasive procedures, such as nasopharyngeal aspiration, sputum induction or gastric aspiration (Appendix 3), can be performed, but only if the specimen is collected for rapid molecular tests, culture or genome sequencing. These procedures should not be performed for smear microscopy.


    Stool specimens (which may contain swallowed sputum) are an alternative to respiratory specimens for the diagnosis of PTB in children. Respiratory specimens are more likely to give a positive result, but the use of stool specimens can avoid invasive collection procedures.


    For children at risk of DR-TB, i.e. contact with a person with DR-TB or coming from an area with high DR-TB prevalence:

    • Multiple specimens (respiratory, stool and EP) should be tested with RMTs. Multiple testing increases the likelihood of detecting TB and obtaining the resistance profile.
    • Every effort should be made to perform culture and phenotypic drug susceptibility tests (Chapter 3).


    For the diagnostic accuracy of Xpert MTB/RIF in specimens other than sputum, see Appendix 1.

    Lateral flow urine lipoarabinomannan assay (LF-LAM)

    LF-LAM should be performed in children with HIV infection:

    • With signs and symptoms of TB, or
    • Hospitalised with advanced HIV disease, or
    • Followed as outpatients with a low CD4 count.
    Chest x-ray (CXR)

    CXR is particularly useful when bacteriological tests are negative or not available. It is also useful to assess the severity of TB and to determine eligibility for the 4-month drug-susceptible TB regimen (Chapter 9).

    Children with PTB usually have abnormalities on CXR, but a normal CXR does not rule out TB.


    For young children unable to stand alone, perform anteroposterior and lateral CXRs if possible (lateral CXR can improve detection of enlarged hilar/mediastinal lymph nodes). 

    For other children, perform a standard posteroanterior CXR.


    CXR findings suggestive of TB in children include a Citation a. For more information, see Diagnostic CXR atlas for tuberculosis in children: A guide to chest X-ray interpretation. International Union Against Tuberculosis and Lung Disease. Second edition, 2022.
    : enlarged hilar/mediastinal lymph nodes, miliary pattern, and cavities. Although generally less specific, consolidation and pleural/pericardial effusion in a child not acutely ill is also suggestive of TB.


    See Chapter 3.

    Tuberculin skin test (TST)

    In children, a positive TST may be one element among many to establish the diagnosis of active TB. However, it has many limitations (Chapter 3 and Appendix 9).

    4.2.4 Follow-up investigations

    For children able to expectorate spontaneously, smear microscopy is used to monitor treatment progress (Chapter 3).

    For children unable to expectorate spontaneously, monitoring of treatment progress is clinical. Invasive procedures should not be performed to obtain respiratory specimens for smear microscopy.

    • 1.Seddon JA, Jenkins HE, Liu L, Cohen T, Black RE, Vos T, Becerra MC, Graham SM, Sismanidis C, Dodd PJ. Counting children with tuberculosis: why numbers matter. Int J Tuberc Lung Dis. 2015 Dec;19 Suppl 1(0 1):9-16.