Diagnosis of tuberculosis and detection of drug-resistance

Publication Year 28 Oct 2024

Download this article in PDF format

Background:
The World Health Organization (WHO) End TB Strategy calls for early diagnosis of tuberculosis (TB) and universal access to drug-susceptibility testing (DST) 1 but the diagnostic gaps remain significant in many countries worldwide. In 2022, 10.6 million people were estimated to have fallen ill with TB worldwide, approximately 7.5 million people were newly diagnosed with TB and notified, and, of them, only 4.0 million were bacteriologically confirmed with TB.2 Among those who were bacteriologically confirmed with TB, 73% received DST for rifampicin, the most powerful first-line anti-TB drug.

The detection of TB remains particularly challenging in people with HIV and in children. The use of concurrent testing offers an important new approach to closing the diagnostic gaps in these populations. The objectives of the current update are the following:

  • Evaluate concurrent3 use of WHO-recommended rapid diagnostic tests for diagnosis of pulmonary TB in children and TB in people with HIV;
  • Consolidate individual product-specific recommendations into class-based recommendations and update the recommendations for technologies falling into the low-complexity nucleic acid amplification test (NAAT) classes.


Method:
In 2023, WHO commissioned a series of systematic reviews of published and unpublished data on the low complexity automated NAATs (LC-aNAATs) and low complexity manual NAATs (LC-mNAATs) and on concurrent testing approaches for TB diagnosis in children and people with HIV. The systematic reviews included data on diagnostic accuracy, impact on patient important outcomes, economic and implementation considerations, and qualitative evidence on feasibility, acceptability, equity, end-user values and preferences. WHO convened a Guideline Development Group (GDG) on 6-10 May 2024 to discuss the findings of the systematic reviews and make recommendations on the use of these technology classes.

The evidence was assessed and synthesized following the GRADE method. Evidence summaries in standard GRADE format were prepared in GRADEpro. The outcomes in the tables were those relevant to the PICO questions.

Standard methods were used to describe and analyze the aggregated and individual data. Estimates of effect were expressed as risk ratios, odds ratios, or hazard ratios with their 95% confidence limits. The absolute risk was also calculated where possible. During the discussion, the GDG members formulated successive drafts of the recommendations based on their assessment of the evidence. The GRADEpro “evidence to decision” template guided this process. Several factors determined the direction and strength of the recommendations (e.g., strong or conditional), including the certainty in the estimates of effect (“quality of the evidence”), values and preferences, how substantial the anticipated desirable and undesirable effects were, certainty on the balance of the benefits and harms, resource implications, health equity, acceptability, and feasibility.

The GDG made recommendations about which populations required a particular diagnostic approach, outlining considerations for implementation when possible. All GDG decisions were reached by discussion and consensus on the recommendations, including their strength and, where appropriate, the conditions to be attached to them.

This rapid communication aims to inform national TB programmes, clinicians and other stakeholders about the key findings and considerations on the use of concurrent testing approaches and low complexity NAATs for the detection of TB and drug-resistant TB. The concurrent use of tests for the diagnosis of TB in people with HIV and children is expected to result in a significant shift in the way these vulnerable populations are to be managed in health programs, and rapid preparation for this change is critical. Furthermore, implementation of these recommendations will result in an earlier and more timely diagnosis of TB in these important populations.