Diagnostic approach — MSF Tuberculosis — Médecins Sans Frontières 2026

MSFClinical guideline (chapter)Current edition

From Tuberculosis › Chapter 4: Diagnosis of active tuberculosis in children · Médecins Sans Frontières, Current online edition, 2026

Clinical guidance on diagnosing active tuberculosis in children, particularly those under 5 years, for use in resource-limited settings.

In one sentenceDiagnose TB in children using exposure history, clinical assessment (symptoms >2 weeks), and Xpert MTB/RIF Ultra testing on respiratory/stool/extrapulmonary specimens; treat high-risk children without awaiting bacteriological confirmation.

Issuer: Médecins Sans Frontières
Edition: Current online edition, 2026
Audience: Clinicians working in resource-limited settings
Evidence: Expert consensus based on field experience and WHO recommendations
Reviewed: 21 Oct 2026 · GMJ

Read the original on MSF Whole guideline (PDF)

Summary by GMJ Guidelines Desk•Reviewed by Prof. Giorgi Pkhakadze, MD, MPH, PhD•Record updated 21 Oct 2026•Next check Apr 2027

What this guideline says

GMJ summary

This MSF guideline section addresses the diagnostic approach to active tuberculosis in children, a population in which diagnosis is particularly challenging. Children with TB typically present with non-specific symptoms, and bacteriological confirmation is achieved in only 20–30% of cases. The guideline emphasises the importance of systematically seeking TB in high-risk groups: children under 2 years, those with HIV infection or severe acute malnutrition (SAM), those in contact with a person with TB, and those not responding to antibacterial or nutritional treatment.

Diagnosis is often based on a combination of history of exposure, clinical assessment, and investigations such as radiology when available. The guideline strongly recommends rapid molecular tests (RMTs) on respiratory, stool or extrapulmonary specimens as the initial diagnostic test, with preference for Xpert MTB/RIF Ultra over Xpert MTB/RIF due to higher sensitivity. For children at high risk of death from TB, treatment should be initiated as soon as TB is considered likely, even without bacteriological confirmation.

The clinical assessment includes screening for symptoms such as cough or fever for more than 2 weeks, night sweats, weight loss or poor weight gain, and enlarged lymph nodes. If diagnosis is uncertain at the first consultation, reassessment within 1–2 weeks is recommended. WHO diagnostic algorithms are available for pulmonary TB (PTB) but not yet for extrapulmonary TB (EPTB). The guideline explicitly states that a trial of TB treatment is not recommended as a diagnostic method; once a decision is made to treat, a full course should be given.

Key points

For clinicians — doses and thresholds as stated in the guideline

Actively look for TB in children under 2 years of age, with HIV infection or SAM, in contact with a person with TB, or not responding to antibacterial/nutritional treatment
TB is bacteriologically confirmed in only 20–30% of children; diagnosis often relies on exposure history, clinical assessment and investigations
Use rapid molecular tests (Xpert MTB/RIF Ultra preferred) on respiratory, stool or extrapulmonary specimens as the initial diagnostic test
For children at high risk of death from TB, initiate treatment as soon as TB diagnosis is considered likely
Key symptoms include cough or fever for more than 2 weeks, night sweats, weight loss or poor weight gain, and enlarged lymph nodes
If diagnosis is uncertain, reassess the child within 1–2 weeks; look for persistent symptoms despite antibiotic or nutritional treatment
Perform lateral flow urine lipoarabinomannan assay (LF-LAM) in children with HIV infection who have TB symptoms, are hospitalised with advanced HIV disease, or have low CD4 count
Do not use a trial of TB treatment as a diagnostic method; once treatment is started, complete the full course

From the guideline

Verbatim

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

— Médecins Sans Frontières, Diagnostic approach — MSF Tuberculosis, Section 4.2 (2026)

Editor’s perspective

Why this matters for Georgia and the region

Prof. Giorgi Pkhakadze, MD, MPH, PhD

Editor-in-Chief, Georgian Medical Journal · Chair, Public Health Institute of Georgia · WHO consultant, health and migration

As a WHO consultant who has worked extensively on health systems strengthening in the Caucasus region, I recognise that childhood TB diagnosis remains a critical gap in our national TB programmes across Georgia, Armenia and Azerbaijan. This MSF guidance is particularly valuable because it acknowledges the reality we face: most children with TB cannot provide sputum and will never have bacteriological confirmation, yet we must not miss these cases. I would stress to Georgian clinicians the recommendation to use stool specimens for Xpert testing—this is feasible even in district hospitals and avoids the need for gastric aspiration, which is rarely available outside Tbilisi. The most significant implementation challenge will be ensuring that Xpert MTB/RIF Ultra cartridges (not the older Xpert MTB/RIF) are consistently available at regional level and that primary care physicians are empowered to initiate treatment based on clinical and radiological findings in high-risk children, without waiting for specialist confirmation that may delay care by weeks.

ORCID 0000-0001-7609-4515 · Perspective written for the GMJ Guidelines Library, 21 Oct 2026

Applicability

Georgia and the South Caucasus

This MSF guidance is highly relevant for Georgia and neighbouring countries in the South Caucasus and Eastern Europe, where childhood TB diagnosis remains challenging due to limited paediatric TB expertise and variable access to rapid molecular diagnostics. Georgia has made significant progress in TB control through its National TB Programme, but childhood TB is frequently underdiagnosed. National programmes should assess their capacity to perform Xpert MTB/RIF Ultra testing on multiple specimen types (including stool specimens, which avoid invasive procedures), ensure availability of lateral flow urine LAM tests for children with HIV, and provide chest X-ray access particularly in rural areas. The emphasis on clinical diagnosis in the absence of bacteriological confirmation is particularly applicable to primary care settings in the region, where clinicians must balance the risk of underdiagnosis against overtreatment. Countries with higher HIV prevalence among children (such as some areas with mobile populations) should prioritise the LF-LAM testing recommendations.

Questions this guideline answers

Quick answers

Why is TB so difficult to diagnose in children compared to adults?

Children, particularly those under 5 years, have low bacillary load, making bacteriological tests often negative. They also present with non-specific symptoms and cannot produce sputum spontaneously. TB is bacteriologically confirmed in only 20–30% of children, compared to much higher rates in adults.

Can I use stool specimens instead of sputum for TB diagnosis in children?

Yes, stool specimens (which contain swallowed sputum) are an alternative to respiratory specimens for diagnosing pulmonary TB in children and can be tested with rapid molecular tests. Respiratory specimens are more likely to give a positive result, but stool specimens avoid invasive collection procedures like gastric aspiration.

Should I start TB treatment in a child with suspected TB but negative tests?

Yes, if the child is at high risk of death from TB (e.g., under 2 years, HIV-infected, severely malnourished, or with severe symptoms), treatment should be initiated as soon as TB is considered likely based on clinical assessment and exposure history, even without bacteriological confirmation. For children not at high risk, reassess within 1–2 weeks before making the treatment decision.

Is it acceptable to give a trial of TB treatment to see if the child improves?

No, a trial of treatment with TB drugs is explicitly 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.

Related on GMJ

Interconnected resources

Dictionary of Public Health

Conditions A–Z

Tuberculosis

GMJ News coverage

Quality & accreditation

Same guideline

Diagnosis of Active Tuberculosis in Children: Introduction — MSF TuberculosisChapter 4: Diagnosis of active tuberculosis in children
Other investigations — MSF TuberculosisChapter 3: Diagnosis and follow-up investigations
Latent tuberculosis infection — MSF TuberculosisChapter 3: Diagnosis and follow-up investigations
Active tuberculosis — MSF TuberculosisChapter 3: Diagnosis and follow-up investigations
Summary of clinical presentations of tuberculosis — MSF TuberculosisChapter 2: Clinical presentation
Clinical presentation of tuberculosis in HIV-infected persons — MSF TuberculosisChapter 2: Clinical presentation

Diagnostic approach — MSF Tuberculosis

What this guideline says

Key points

From the guideline