FeedbackThe use of an mWRD as the initial test to diagnose TB greatly increases the sensitivity of the diagnostic process compared with the use of sputum smear microscopy (57). However, certain subpopulations will not fully benefit from the high sensitivity of the molecular tests. For example, people living with HIV, especially those who are seriously ill, are known to have difficulties producing sputum; similarly, children, especially those aged below 5 years, often cannot easily produce sputum. In addition, if a person living with HIV or a young child can produce sputum, it will frequently have low or variable amounts of mycobacteria. These issues mean that it can be difficult to confirm TB disease bacteriologically; inclusion of alternate specimen types that can be collected in non-invasive ways could overcome some of these challenges, improve bacteriological confirmation of disease and speed time to diagnosis.
To assess whether testing of multiple specimen types with different tests – referred to by WHO as “concurrent testing” – enhances diagnostic accuracy. Therefore, WHO commissioned a systematic review in 2024, in which a GDG evaluated evidence for three risk groups: adult and adolescent people living with HIV, children aged below 10 years without HIV or with unknown HIV status, and children living with HIV. The tests for which data were available included LC-aNAAT (used on respiratory and stool samples), and the urine LF-LAM for HIV-positive individuals. The GDG assessed the following combinations:
- adults and adolescents with HIV – LC-aNAAT on one respiratory sample and urine LF-LAM;
- children – LC-aNAAT on one respiratory and one stool sample; and
- children living with HIV – LC-aNAAT on one respiratory and one stool sample, and urine LF-LAM.
For all three combinations, the GDG concluded that concurrent testing improved testing accuracy compared with use of a single test, with moderate cost requirements and possible increases in testing workload. Based on the evidence, the GDG issued a strong recommendation to test concurrently with urine LF-LAM and LC-aNAAT on a respiratory sample for adults and adolescents living with HIV; and with LC-aNAAT on a respiratory and a stool sample for children without HIV or with unknown HIV status. However, for children living with HIV, the recommendation to test concurrently with urine LF-LAM and LC-aNAAT on a respiratory and a stool sample is conditional, owing to the low certainty of the evidence and the increased number of false positive results. The recommendation for use of LC-aNAATs on respiratory samples in the concurrent testing strategies applies to both Xpert MTB/RIF Ultra and Truenat MTB Plus and MTB-Rif Dx, whereas concurrent testing of stool samples is limited to the use of Xpert MTB/RIF Ultra, given insufficient evidence on the use of Truenat tests with this sample type at the time of evidence assessment (5).
The recent update of the consolidated guidelines includes these new recommendations on concurrent testing for children and for people (of all ages) living with HIV (see Table 1.1 in the guidelines (5)). The idea of concurrent testing is that all tests should be performed as soon as possible. People eligible for testing should be provided with information to help them understand this approach, including the need for more than one sample type and the need for any follow-up. Where possible, all sample types required for the concurrent testing strategy should be collected during the initial visit, and promptly tested or sent for analysis. Staffing and training needs should be assessed to ensure they are adequate for collection of quality samples, sample referral, and any changes or volume increase associated with onsite testing. Similarly, sites collecting urine should be equipped with private collection areas that have handwashing facilities, sterile collection containers, and clear procedures on the sanitary collection of midstream urine samples, to minimize contamination.
A positive result on either test is a positive result overall; it is sufficient to confirm TB diagnosis and should trigger a decision to treat, to reduce delays. Although conducting multiple tests for anyone who is eligible can increase the sensitivity of the diagnostic process, the approach must be balanced against reduced specificity and programmatic implications (e.g. higher test costs and access to multiple sample types and tests). It is important to monitor and prevent patient loss to follow-up while awaiting a second diagnostic test or drug susceptibility test, or test results.
LF-LAM is a POC test that can be performed onsite without delay and can provide results within 30 minutes, whereas results from LC-aNAATs require more time (≥90 minutes) for onsite testing (or longer when samples require referral for offsite testing). For example, concurrent testing in one setting involves an adult person living with HIV being tested onsite with LF-LAM while a sputum sample is transported to a site with LC-aNAAT. If the onsite LF-LAM yields a positive result, treatment can begin immediately and can later be adjusted based on the RIF result from the LC-aNAAT. If LF-LAM is unavailable, treatment decisions will rely on LC-aNAAT results alone; however, efforts should be made to ensure that LF-LAM testing is done.
If a test included in the concurrent testing strategy is not available at a site, this should not hinder the use of existing diagnostic tools. For example, if a site has LC-mNAAT (i.e. TB-LAMP) available onsite – a test not included in this recommendation – and also has LF-LAM, both tests should be conducted as soon as possible. A positive result from either test warrants treatment initiation. An additional sample should be referred for testing with LC-aNAAT, to determine the individual’s RIF resistance status; this should also be done in cases where both tests are negative.
For children who are HIV-negative or whose HIV status is unknown, one stool sample and one respiratory sample should be tested on LC-aNAAT according to the concurrent testing strategy. Stool can be collected at peripheral sites and, if there is also capacity to collect respiratory samples from children, the samples can be tested with LC-aNAAT concurrently. If testing capacity is not available onsite, samples should be transported to a site with LC-aNAAT capacity. For children living with HIV, a urine sample should be collected in addition to the respiratory and stool samples. Testing with LF-LAM should be conducted onsite if the test is available, regardless of whether LC-aNAAT is available, and a positive result should trigger a treatment decision while awaiting LC-aNAAT results.
In situations where collection of quality respiratory samples is not possible, it may be necessary to refer a child to a facility that has the capacity to perform the specialized procedures of collecting respiratory samples from children. If a stool sample is available and the LC-aNAAT result is negative, additional respiratory sampling should be undertaken (e.g. induced sputum, nasopharyngeal or gastric aspirate) for LC-aNAAT testing at a higher level facility, to maximize opportunities for bacteriological confirmation of disease and determine RIF resistance status. If a stool sample is available and the result is positive, and RIF resistance results are available, treatment should be initiated; in such cases, collection of a respiratory sample is not necessary.
Owing to the limited sensitivity of the available diagnostic assays in children, clinical diagnosis continues to play a critical role in the diagnosis of TB. The aim of the concurrent testing strategy in this population is to increase testing sensitivity, but diagnosis should not be delayed because of the unavailability of a sample type or test, or the availability of negative test results alone. Delaying diagnosis leads to higher rates of loss to follow-up and additional burdens on clients. Therefore, laboratory-based investigations in children should not happen in isolation but rather should be embedded as part of the clinical evaluation process. Clinical diagnoses are usually based on suggestive signs and symptoms and a history of TB contact, with or without chest radiography. The WHO operational handbook on tuberculosis: module 5: management of tuberculosis in children and adolescents contains example integrated treatment decision algorithms aimed at encouraging health workers at the primary care level to make an evidencebased decision to start TB treatment in children aged below 10 years (58).