A new point-of-care test to diagnose tuberculosis

In 2017, tuberculosis (TB) caused an estimated 1.6 million deaths, including 300,000 deaths among people living with HIV (PLHIV), and had surpassed HIV/AIDS to become the leading infectious cause of mortality worldwide.¹ This was due in part to not diagnosing or not reporting approximately 36% of TB cases (~3.5 million) each year.² Current diagnostic tools in routine clinical use, including Cepheid’s GeneXpert MTB/RIF assay, rely on sputum-based testing, which has consistently demonstrated suboptimal diagnostic sensitivity, especially in immunocompromised PLHIV either unable to produce sputum during hospitalization or at increased risk of extrapulmonary disease. Since research and development of new TB diagnostics has been lagging behind knowledge of the spectrum for TB pathogenesis, which includes incipient and subclinical TB,³ the World Health Organization (WHO) has prioritized a biomarker-based non-sputum test that could be used at the clinical point of care (POC) to rapidly diagnosis of all forms of TB (including extrapulmonary) and among all persons (including children).⁴

In this issue of The Lancet Infectious Diseases, Dr. Broger and colleagues present exciting data to evaluate a new urine-based POC test for detecting urine lipoarabinomannan (LAM).⁵ The first commercial LAM assay, the Determine™ TB LAM assay (“AlereLAM”; Alere was recently acquired by Abbott, USA), correlates with clinical disease severity and risk of mortality,⁶ and can improve outcomes for select hospitalized adult PLHIV in randomized trials,^7,8 but has only moderate diagnostic sensitivity.⁹ The research team then compared the new SILVAMP TB LAM assay (“FujiLAM”; FujiFilm, Japan) against the AlereLAM assay by testing stored, thawed urine samples from three independent cohorts of hospitalized adult PLHIV in South Africa. Qualitative results were compared to a microbiological reference standard (MRS), and a clinical reference standard (CRS) that included an empiric diagnosis of TB. Among 968 participants, the prevalence of pulmonary TB and CD4 counts were consistent with high-risk immunocompromised PLHIV who may be recommended for LAM testing,⁹ but not widely representative of PLHIV at risk for active TB. When compared to the MRS, FujiLAM had a diagnostic sensitivity of 70.4% and specificity of 90.8%, while the AlereLAM had a diagnostic sensitivity of 42.3% and specificity of 95.0%. The differences were statistically significant for diagnostic sensitivity, but not for specificity. Based on these results, the authors concluded that the FujiLAM assay had improved diagnostic sensitivity, without compromising specificity, as compared to the AlereLAM assay.

Appropriate validation of point-of-care tests intended for use in resource-limited settings can be complicated, and separating comparisons between diagnostic sensitivity and specificity may not accurately represent a test’s real clinical value.¹⁰ Although it would be desirable to have a POC test with higher diagnostic sensitivity, in this clinical setting and with these performance metrics (lower sensitivity, higher specificity), the two urine LAM assays would primarily be used as a diagnostic “rule in” test. An appropriate comparison may be a positive likelihood ratio (LR+), which accounts for both sensitivity and specificity. The LR+ is used in clinical medicine to determine if a diagnostic test result changes the pre-test probability that a disease exists (i.e. active TB). Using reported diagnostic accuracy results against either the MRS or the CRS, both of which the authors pointed out may be imperfect reference standards, the calculated LR+ values were similar. Regardless, the new urine LAM assay will require further characterization and validation in prospective studies using appropriate clinical, laboratory, and biomarker reference standards, with careful collection of participant outcomes and latent class modeling to adjudicate discordant results.

The global health community now has two non-sputum biomarker assays that may be useful in clinical POC settings to diagnose TB in PLHIV in TB-endemic countries. Compared to the AlereLAM assay, the new FujiLAM assay includes novel monoclonal antibodies and enhanced detection technology to enable higher diagnostic sensitivity.¹¹ However, the usability of FujiLAM may be less desirable due to more operator steps and a longer time to result. Similar to AlereLAM, the FujiLAM assay may need further optimization in order to be useful as a diagnostic test among the larger population of HIV-negative persons. Further, validation and implementation studies in both adults and children will be necessary to broaden current recommendations for urine LAM testing, in order to reliably diagnose TB, rapidly initiate appropriate therapy, and reduce TB mortality worldwide.

The development of a second simple, rapid, POC test is a major step forward for advancing TB diagnostics and could lead to lives saved from early detection and treatment. However, as has been observed with the AlereLAM assay, WHO endorsement⁹ and inclusion on the Essential Diagnostics List¹² may not necessarily lead to rapid uptake.¹³ For lives to be saved by the use of these POC tests, implementation and modeling studies will be needed to provide more guidance for national TB programs.