Threads
HomeArticlesThere’s a new way to test kids for TB — but the...

There’s a new way to test kids for TB — but the NHLS cyberattack has delayed its roll-out

  • The National Health Laboratory Service’s plans to pilot a new way for finding tuberculosis (TB) in kids have been put on hold after a cyberattack shut down computer systems at the state labs. 
  • South Africa isn’t doing well at finding children with TB, and just 60% of those with the disease are on treatment.
  • That’s because finding TB in young children is tricky in the first place, as they usually have so little of the germs in their lungs that sputum tests often come back negative.
  • In 2021, the World Health Organisation recommended using stool samples to test for TB, and now South Africa’s new childhood TB guidelines will include this method.

In today’s newsletter, health reporter Zano Kunene explains why testing children’s stools can get more of them on TB medication. Sign up now.

Plans to kick off a study this month to test a new way for finding tuberculosis (TB) in kids — by looking at their poop — have been put on hold after a cyberattack in late June shut down the National Health Laboratory Service’s (NHLS) computer systems, but Farzana Ismail, a clinical microbiologist with the lab network, says it “remains a priority”.

Because of the IT problems at the country’s public pathology lab service, test results now have to be processed and reported manually, slowing down feedback to doctors and patients — and upsetting plans like those for the TB testing study.

The NHLS initially said its systems would be up an running again by 15 July, but issued a press release on 16 July saying only some systems are back up, “with more being available as soon as appropriate security measures are completed”.

The TB test pilot, in which stool samples were to be tested for signs of the TB germ, would have taken place over the next two months at six labs in Gauteng, the Eastern Cape and the Western Cape, she told Bhekisisa at the 8th TB Conference in Durban in early June. The new method will be rolled out at state labs nationwide, the NHLS confirmed.  

Finding TB in children is difficult, because young kids (those younger than five) often have only a small amount of the bacteria in their lungs. This means the bug might not show up in a sputum sample (a mix of saliva and mucus from the airways), and so the test result could come back negative even though the child is actually sick with TB. 

Around 27 000 children younger than 15 in South Africa are thought to have TB — almost 10% of the country’s total cases — and about 40% of them went untreated in 2022. TB is the third biggest killer in this age group even though it is a preventable and curable disease.

Roughly six out of every 10 childhood TB cases are among kids under the age of five, the group in whom it’s particularly difficult to pick up the TB bug.  

The difficulty in getting good sputum samples from kids means other, often more uncomfortable, and sometimes painful, ways have to be used to get a specimen that will be good enough to give a reliable result. 

For example, a health worker can take a sample from someone’s stomach (because people often swallow phlegm, which can contain TB germs if they’re infected). But for this, a tube has to be fed through the child’s nose down into their stomach, and they can’t eat for three or more hours before the test. 

South Africa’s guidelines on how to test and treat children for TB were last revised in 2013 and, says Norbert Ndjeka, chief director for TB at the national health department, they are “impatiently waiting for the updated paediatric TB guidelines [to be published]” to help them better diagnose children, because “currently we’re not doing particularly well in this area”. 

Stool is cool 

“To improve specimen collection [in children], we’re trying to get to child-friendly methods, like [taking] a [mouth] swab or stool [sample], so that it’s not so invasive,” says Karen du Preez, a senior clinical researcher at the Desmond Tutu TB Centre. “The reality is that because kids typically don’t have a lot of bugs, the diagnostic accuracy in the current tests is expected to be low.” 

Using stool samples to test for TB will be included in the new guidelines, which are being written jointly by the health department and the TB Think Tank, a group of experts who advise the government on policies for curbing TB infections, and follow after the World Health Organisation (WHO) published an update of their standards document in 2022. 

The WHO has recommended testing faeces for TB since 2021 — and in countries like Vietnam, Zambia and Ukraine the method has been widely used. Moreover, because the same rapid-result equipment can be used as for sputum tests, namely the GeneXpert machine, any lab that’s already set up for this will be able to run tests on stool samples too

With a GeneXpert test, a sample (such as sputum or stool) can be analysed for TB within two hours, and in South Africa, most labs where TB gets tested are equipped with such machines.

Findings from a systematic review of studies in Africa and Asia, where stool samples were used to test for TB, found that the bug was correctly picked up in, on average, about 67% of close to 1 700 cases. This is in line with the WHO’s ideal sensitivity rate of 66% for rapid non-sputum-based TB tests in children. (Africa and Asia make up more than two-thirds of all TB cases globally.)  

Correctly identifying kids who did not have TB was almost perfect when using the stool sample test, the review showed. (The stool test’s specificity was 99%, meaning that only one out of every 100 test results was a false positive.) Having a TB test with a high specificity means that fewer people will incorrectly be put on antibiotic treatment when they don’t need it, which can last between six and nine months and so add to the chances of drug-resistant germs developing.

In the same way, using a TB test that yields accurate results fast and so identifies kids who do have the germ, means that they can get onto treatment early, rather than having to wait for a bacterial culture test to be done, which can take two to six weeks.  

“The longer we wait [to make a diagnosis], the sicker someone can get,” says Du Preez.

Getting results from a culture test takes long because a small dish with a gel-like food source in which bacteria can grow has to stand at a suitable temperature for a few weeks to allow enough time for a clump of bacteria to grow.  

Moreover, given that taking only one type of sample from children isn’t guaranteed to identify the bug in someone who is indeed sick, the WHO says using stool together with another specimen type may improve the chance of getting an accurate diagnosis.

What about other options? 

Researchers are also looking at adding other ways of sample collection to the TB test choices. For example, taking a swab from someone’s mouth and or tongue may be promising. 

An oral swab is taken by scraping something that looks like a long earbud against the lining inside someone’s cheek or at the back of their tongue. It’s more comfortable for the patient than having to cough up sputum, easier to collect and safer for health workers too, because seeing that the patient doesn’t have to cough hard, there’s little chance of breathing in the TB germ from tiny droplets of spit expelled into the air when someone coughs.

While the WHO doesn’t have guidelines on oral swab tests yet, it was the topic of many presentations at the International Union Against TB and Lung Disease (The Union) conference on lung health in November. Moreover, earlier research has found a similar accuracy level (of around 83%) when comparing sputum and oral swab samples. 

However, as children often have a smaller amount of germs in their lungs than adults, it’s not sure how well oral swabs will work in finding the bug in children, says Du Preez.

Indeed, findings of a systematic review of studies in which mouth samples were used for TB tests, found that correct results showed up in only 5–42% of children.

Says Ismail about the research on oral swabs for finding TB in kids: “It’s shown sensitivity [how accurate a test is in identifying a disease] of around 30%, which means that three in 10 tests will be confirmed and the other seven will need further testing.”

[Updated] 16 July 2024, 2:10pm: The text of this article was updated to include the contents of an NHLS press release, issued on the morning of 16 July, after this story had been published.

Zano Kunene is a health journalist at Bhekisisa.

RELATED ARTICLES
MORE FROM AUTHOR