Latent TB

The central challenge of treatment for LTBI is identifying those with the highest risk of progression. Since only approximately 5% of TST-positive TB contacts progress to active TB within 2–5 years of exposure,4 the number of TST-positive contacts requiring treatment to prevent one case of active TB (number needed to treat, NNT) is about 20. Given that the TB burden in low-incidence countries is largely driven by LTBI progressing to active disease, a more powerfully predictive test to stratify latently infected individuals by progression risk would improve targeting of preventive treatment thereby reducing the NNT.

Several longitudinal studies have confirmed that IGRAs predict progression from LTBI to active TB. However, their prognostic power compared with the TST has been a key area of uncertainty. One systematic review of 15 studies with 26 680 individuals concluded that the prognostic power of IGRAs and the TST were low and broadly equivalent,8 while another review of 28 studies including >30 000 individuals found IGRAs to be more predictive than the TST.9 Updated guidelines from the National Institute for Health and Clinical Excellence (NICE) on the use of IGRAs10 highlighted the paucity of high-quality studies of the prognostic power of IGRAs relevant to low-incidence settings on which to base clinical recommendations.

A new study published in this issue of Thorax makes an important contribution to this evidence base and may help to inform IGRA-based screening policies.11 A single-step QFT-GIT screening strategy in 1769 recent TB contacts was evaluated in routine practice in Leicester, UK. Among untreated contacts, QFT-GIT had 70% sensitivity and 86% specificity for identifying LTBI progressing to active disease in 2 years, and the NNT of QFT-positive cases to prevent one case of active TB was 7, substantially lower than previously reported for the TST. This figure needs to be interpreted with caution, however, as it is lower than that reported by almost all previous studies and meta-analyses and may reflect inherent bias in the study design because the investigators making the diagnoses were not blinded to the IGRA results. Nonetheless, the study shows promise for the use of IGRAs in contact investigations to improve targeting of preventive therapy. However, IGRA results were not predictive of progression to disease in older contacts (≥36 years), most likely due to a higher burden of infection that was remotely acquired (and therefore less likely to progress) in this mostly foreign-born population, analogous to the disappointing prognostic power of IGRAs in contacts in high-burden countries.

Available longitudinal outcome data and the higher specificity of IGRAs collectively suggest that the NNT for IGRA-positive contacts is lower than for TST-positive contacts, but most available evidence suggests it is still far from that required to effect a step change in TB control and prevention. Enhancing the prognostic power of IGRAs, thereby reducing the NNT, is therefore an urgent research priority (box 1).

Since recently infected contacts are at higher risk of progression to active TB, a test able to distinguish recently acquired infection from remotely acquired infection would substantially reduce the NNT. Neither IGRAs nor the TST are capable of this discrimination,12 ,13 and targeting recently infected persons will continue to depend on epidemiological risk factors until such a test is developed.

Some studies suggest that a larger size of baseline IGRA response may be predictive of progression,14 but there are currently insufficient data to use this clinically to risk-stratify, and more data correlating quantitative baseline responses with clinical outcomes are required (box 1). Therefore, results should currently only be interpreted in a binary fashion as positive or negative. Borderline results are uninterpretable and have no clinical significance, and should therefore be repeated. Indeterminate results, which usually result from a failed positive control and are more common in the setting of suppressed cellular immunity, are also uninterpretable and should also be repeated.

Where IGRAs are repeated, results may change from positive to negative or vice versa.15 ,16 Conversion from negative to positive has similar clinical implications to TST conversion, reflecting recent acquisition of infection. The clinical significance of reversion from positive to negative is uncertain, although it may reflect spontaneous clearance of infection and low/absent risk of progression. Only longitudinal clinical outcome studies will determine if this is the case (box 1).

Active TB

IGRAs measure a cellular immune response to Mtb-specific antigens. A positive result therefore indicates infection, but does not distinguish active from latent TB. Since Mtb infection is a prerequisite for TB disease, a negative IGRA result might provide a rapid ‘rule-out’ for active TB. However, this requires very high diagnostic sensitivity and, although IGRAs may have higher sensitivity than the TST, neither is sensitive enough to exclude a diagnosis of active TB.17 ,18

It is also proposed that a positive IGRA may provide a useful adjunct to ‘rule-in’ the diagnosis in paediatric patients with a clinical picture consistent with active TB, as the TST is already used.19 Furthermore, NICE guidelines suggest that TST/IGRA can be beneficial in diagnosing cases where histological appearances are not conclusive during the investigation of non-pulmonary TB.20 Clinical studies already indicate that next-generation IGRAs, incorporating additional Mtb antigens or measuring additional Mtb-induced cytokines, provide significantly improved diagnostic sensitivity (without compromising specificity) that could enable rapid rule-out of suspected active TB (box 1).21 ,17 Newer T cell-based diagnostic platforms utilise T cell functional signatures that reflect mycobacterial burden to differentiate active from latent TB and provide treatment monitoring tools.22–24