Monaz: Hello, I'm Monaz Mehta, editor in the Cochrane Editorial and Methods department. The Cochrane Library contains more than 140 Cochrane Reviews of the accuracy of diagnostic tests, covering a very wide range of conditions. One of these reviews, on a test used to detect tuberculosis in the lungs, was updated for the third time in February 2021 and Karen Steingart from the Liverpool School of Tropical Medicine in the UK speaks with first author, Jerry Zifodya from Tulane University in New Orleans, USA about the latest findings in this podcast. This review is also included in the Cochrane Library Special Collection for World Tuberculosis Day on 'Diagnosing tuberculosis'.

Karen: Hello Jerry. First, perhaps you could say a few words about how big a problem tuberculosis is?

Jerry: Thanks Karen. Tuberculosis, or TB, remains in the top 10 causes of death worldwide; and people with TB are often poor and disadvantaged, have limited access to health care, and often face stigma and discrimination. According to the World Health Organization, in 2019, 10 million people developed TB of whom 1.2 million HIV‐negative people and nearly 210,000 HIV‐positive people died. This situation could worsen further owing to challenges in healthcare systems arising from the COVID-19 pandemic. The good news is that when TB is detected early and effectively treated, it is largely curable.

Karen: Your review looks at some of the ways to achieve that early diagnosis. What in particular did you study?

Jerry: Most people with TB have the disease in their lungs, called 'pulmonary TB' and we focused on this. Its diagnosis is often difficult and it's also important to find out if the person's TB is resistant to rifampicin, one of the most important drugs used to treat TB. 
With that in mind, we looked at two World Health Organization-recommended tests that simultaneously detect Mycobacterium tuberculosis complex and rifampicin resistance. These are called Xpert Ultra and Xpert MTB/RIF, and they are automated and rapid, providing results within two hours. In this update, we directly compared the accuracy of the two tests.

Karen: As well as this narrowing of the focus of the review, were there other reasons to perform this update?

Jerry: In the previous version of the review, in 2019, nearly all the included studies assessed the diagnostic accuracy of just one of the two tests, Xpert MTB/RIF. Overall, we found it to be sensitive and specific for detecting pulmonary TB and rifampicin resistance, but it was less sensitive in people with smear-negative sputum and people with HIV. There is now more research on the newer test version, Xpert Ultra, and we wanted to bring this evidence into this update so that it could be used in the development of updated World Health Organization guidelines about these tests which were published in 2020.  

Karen: And has the new evidence changed things?

Jerry: Yes, absolutely, we are now able to include nine studies involving 3,500 participants that directly compared Xpert Ultra to Xpert MTB/RIF for diagnosing pulmonary TB, and five studies, with about 900 participants, for rifampicin resistance.
Regarding detection of TB, the updated review shows that, if you imagine a population of 1,000 people of whom 100 have pulmonary TB, Xpert Ultra will miss 9 cases but Xpert MTB/RIF will miss 15 cases, showing that Xpert Ultra has higher sensitivity. On the other hand, its specificity is lower. In the same population of 1,000 people, 40 would be wrongly diagnosed as having pulmonary TB with Xpert Ultra, compared to only 14 with Xpert MTB/RIF.

Karen: Thinking about the subgroups of patients you mentioned earlier, how did the tests perform in those difficult to diagnose patients?

Jerry: In people with smear-negative culture-positive sputum and people living with HIV, Xpert Ultra was more likely than Xpert MTB/RIF to detect people with TB. Both tests performed well and were unlikely to make a false-positive diagnosis in those without TB.  However, for people with a history of TB, compared with Xpert MTB/RIF, Xpert Ultra was indeed more likely to make a false-positive diagnosis in those without TB.

Karen: Turning to your other question, was there any difference in the detection of rifampicin resistance?

Jerry: Xpert Ultra and Xpert MTB/RIF were both very good at detecting rifampicin resistance when it was present and ruling it out if someone didn't have it. However, we found that Xpert Ultra did have a higher number of rifampicin resistance indeterminate results compared to Xpert MTB/RIF, with a pooled proportion of 8% rather than 1%.

Karen: Thanks Jerry. To finish, what would you say are the main implications of these updated results and where can people find the review?

Jerry: Thanks Karen. In summary, both tests are accurate for diagnosing rifampicin resistance but, for diagnosing pulmonary TB, Xpert Ultra has higher sensitivity and lower specificity than Xpert MTB/RIF, especially in people with smear-negative sputum and people living with HIV. The sensitivity and specificity trade-off would be expected to vary by setting, but the takeaway message is that both Xpert Ultra and Xpert MTB/RIF provide accurate results and can allow for rapid initiation of treatment for rifampicin-resistant and multidrug-resistant TB. Finally, getting hold of the review is straightforward, it's available online and people just need to go to Cochrane Library dot com and type 'ultra for pulmonary tuberculosis in adults' to find it.