The deadly gap in diagnosing children with tuberculosis


Cathy Hewison, MSF Tuberculosis Working Group Leader
Diana Gomez, MSF Laboratory Advisor
Stijn Deborggraeve, Diagnostics Advisor, MSF Access Campaign

Note: This blog was first published in March 2022 and exact figures may have changed but the content remains current.

This week, the World Health Organization (WHO) has released updated guidance for treatment providers on the management of TB in children and adolescents.

These new guidelines are urgently needed. Today it is estimated that more than 60% of all children with TB worldwide are never even diagnosed with the disease, and 96% of children who die from TB are never put on treatment, 80% of them younger than 5 years old.

Drawing on Médecins Sans Frontières (MSF) extensive experience treating children with TB, our authors unpack some of the most important challenges in diagnosing paediatric TB, and how the new guidelines could improve the situation to save many more young lives.

MSF is providing testing and treatment for people with TB currently in 38 countries, among them many children © Sabir Sabirov

MSF works with people who are particularly vulnerable to TB. This includes children living in places that are riven by conflict or who are displaced or weakened by malnutrition.

With TB, as with all diseases that impact children, it is vitally important to diagnose disease quickly and accurately in order to get them on to lifesaving treatment as soon as possible.

Currently this is not the reality we in MSF face in many of our projects, nor is it the experience of other treatment providers in the places where populations are particularly vulnerable to TB. Diagnosing TB in children remains a major challenge to our medical staff.

In our clinic, every day we see children with symptoms that could be TB. We are an experienced team and have the laboratory TB tests available but for most children who come to us, we just cannot confirm the diagnosis of TB.
- Dr Lazro Fidelle, Malakal, South Sudan

What are the difficulties in confirming the diagnosis in a child?

There are several reasons it is hard to confirm the diagnosis of TB in a child. First, the clinical presentation of a child with TB is often different than for adults, with more non-specific symptoms such as fatigue, fever or slow weight gain, especially in malnourished children. This leads to misdiagnosis and children are put on treatment for other diseases, instead of TB, thereby missing out on the treatment they actually need.

Children also become sick with much lower levels of TB bacteria compared to adults, so the diagnostic tests available today — most designed and evaluated to detect TB in adults — often cannot detect the low levels of TB bacteria, and the child does not test positive for TB.

Another important difference is that children, unlike most adults, often have TB infection outside the lungs, for instance in lymph nodes or bones where it is difficult to collect samples to test.

This all adds up to the devastating situation that, with the current tests, the majority of children with TB are never diagnosed, and only a minority have a positive test that confirms the disease.

New algorithms to improve clinical diagnosis of TB in children

Most of the time we go ahead and treat the child without any positive lab test results. However, this decision to treat must always balance the need for timely and lifesaving treatment with the need to avoid unnecessary treatment of children who may not have TB but another respiratory infection. So we welcome the updated WHO guidelines which now provide clear evidence-based clinical algorithms to diagnose TB, even when test results are not available.
- Dr Lazro Fidelle, Malakal, South Sudan

The most important recommendation in this week’s new WHO guidelines is how to diagnose children in places where laboratory tests remain currently unavailable or when test don’t give a reliable result. This could have a real and positive impact on the number of children diagnosed with TB and how quickly they are put on treatment, leading to a reduction of unnecessary deaths.

These include new treatment-decision algorithms* to diagnose TB on clinical criteria alone, giving doctors the confidence to start TB treatment in children without confirmation from laboratory tests. This could be when treatment is urgent but the testing time too long, or in places where there are no laboratory tests available, but also when challenges persist such as getting the right sample from the child or when we know we can’t always rely on the accuracy of the test result.

WHO now also recommends stool samples to test for TB in children

Our biggest challenge is collecting a sputum sample from the child to test for TB. While sputum is the standard specimen that we use to detect TB, small children are often not able to produce sufficient sputum to test.
- Dr Lazro Fidelle, Malakal, South Sudan

As well as the challenges of existing tests to detect the low levels of TB bacteria in children, the sputum samples which we typically test for TB are very difficult to get from children who are often unable to cough up enough sputum. We can induce the production of more sputum by various methods, including irritating the lungs with nebulised saline, but this is so difficult for the child that in reality we rarely do it.

This video features a painful procedure used to extract sputum from children with suspected TB, and illustrates the need for better diagnostic tools for children with TB. The procedure is very rarely used today, and only in the most difficult cases.

Radiographical tools such as chest X-ray can be another useful tool as part of a treatment-decision algorithm for TB in children. However, X-rays are often available only in hospital-level care, and thus not available in most places where children need to get diagnosed for TB.

These difficulties in detecting TB are why WHO’s new recommendation, that children who cannot produce sputum can also be tested for TB using stool, is a real advance for us and other TB treatment providers. It’s simply much more feasible to collect stool samples from children than sputum for testing. There is however the drawback that the current TB tests for stool samples have low sensitivity, meaning they don’t pick up on the TB bacteria when children carry low levels of bacteria, so we will likely still miss diagnosing TB in many children.

Limitations of current WHO-recommended TB test for children

The test that WHO recommends for diagnosing TB in children is the GeneXpert Ultra test, which runs on the GeneXpert machine from the US-based diagnostics company Cepheid. This device can test TB in many sample types, including sputum and stool. There are however limitations with this test.

First the machine needed to run the test is not often available outside urban settings and the places where most children need to be diagnosed. In addition, the machine needs a continuous power supply and an air-conditioned room, which again are rarely available in the remote settings where most children need to get tested. The machine is also fragile, so in the places where we have the system already set up, it breaks down frequently and getting it repaired by Cepheid can take more than three months and is expensive. This means long periods where we cannot test children — or adults — on this device.

Cepheid had moved ahead and created a newer GeneXpert technology— a battery-powered, robust and portable GeneXpert system called Omni — to counter some of the limitations of its original model. But last year, Cepheid unexpectedly announced it will not bring the product to the market. The company has given no adequate reason for this decision despite receiving large amounts of public funding for its development, and this point-of-care machine could make a huge difference to our diagnostic capacity in the places where we work.

The high price of the GeneXpert machine and tests are also a limiting factor in the ability of countries to scale up testing. Cepheid charges US$9.98 per TB test in low- and middle-income countries, which is more than double the price it costs the company to produce the test, based on MSF’s analysis.

MSF works to try to improve the diagnosis of TB in children and adults. In Malakal, South Sudan, we are undertaking operational research on a new technology, POCUS (point-of-care ultrasound), to increase and improve diagnosis of TB in children.

What more needs to happen to get more children diagnosed?

We make do with what we have, but we are still really in desperate need of better TB tests for children that can be used in even the most remote settings and that can test TB at high sensitivity in samples that are easy to collect in children, such as mouth swabs or finger-prick blood.
- Dr Lazro Fidelle, Malakal, South Sudan

Today, as regards the existing tests, there is a worrying dependency on Cepheid as the sole supplier of WHO-recommended TB tests to diagnose children. Cepheid’s monopoly market position means countries have little room to negotiate prices and better service and maintenance contracts with the company, and that, as we have seen, has resulted in high prices, and low standards of service and maintenance provision.

What we need to see is funding made available for other manufacturers and developers of TB diagnostics to accelerate the development of alternative and competing tests. This is the proven way to secure sustainable supply, lower prices and adequate service.

Looking ahead, we urgently need better laboratory tests made available to all children with TB and that work regardless of where they may live. This will mean more research to develop more sensitive tests that can accurately detect TB in children and widen the range of sample types we can use in tests, and it means a test that works without the need for electricity or other laboratory infrastructure, that is robust to extremes of temperature and humidity, and does not require a high level of maintenance.

What does MSF recommend for the future?

For us to see the improvements set out in the new guidelines made a reality on the ground, it is essential that governments and national TB programmes adopt them as soon as possible into their own national TB policies, and at the same time provide sufficient training and support to staff to allow their practical implementation at all levels of care.

As well, we need to see governments, funders and manufacturers step up to the challenge of developing better TB diagnostic tests for children. We need to ensure a more diverse supply with several manufacturers so that these tests are available at affordable prices and with adequate service support through market competition.

Only then can we hope to see real change and a reduction in the unacceptably high number of children who currently go undiagnosed and die each year unnecessarily from this curable yet killer disease.

Diana Gomez, Stijn Deborggraeve and Cathy Hewison

* Algorithm: In clinical medicine, an algorithm is a step-by-step decision pathway for management of a health care problem.

Subscribe to our Medium blog here and get more updates and insights in to our work to open up access to medicines for all



MSF Access Campaign — Medicines Are Not a Luxury

This blog is a place to reflect on our experiences working for access to medicines. For the official MSF Access Campaign website please visit