Tests based on the detection of mycobacterial lipoarabinomannan (LAM) antigen in urine have emerged as potential point-of-care tests for tuberculosis (TB). We aimed to assimilate the current evidence regarding the diagnostic performance of LAM assays and to ascertain their clinical indication in settings with high and low prevalence of HIV-TB co-infection. Owing to suboptimal sensitivity, the urinary LAM assays are unsuitable as general screening tests for TB. However, unlike traditional diagnostic methods, they demonstrate improved sensitivity in HIV-TB co-infection which further increases with low CD4 counts. Accordingly, these assays are indicated as rule-in tests for TB in patients with advanced HIV-induced immunosuppression, and facilitate the early initiation of antituberculous treatment in them. They also offer incremental sensitivity and specificity when used as adjunct tests to smear microscopy and chest radiography in HIV-TB co-infection. They obviate the biohazards associated with sputum samples and provide an alternative diagnostic tool in sputum-scarce patients. Notwithstanding these advantages, the specificity of these assays is variable, which is mostly attributable to misclassification bias and cross-reactivity with non-tuberculous mycobacteria or other commensal flora. Furthermore, the inability to detect low titres of antigen in HIV-uninfected patients makes these assays unsuitable for use in settings with a low HIV prevalence. Future research targeted towards inclusion of specific monoclonal antibodies and more sensitive immunoassay platforms might help to improve the diagnostic performance of these assays and extend their applicability to the general population of patients with TB.
Keywords: Diagnosis; Lipoarabinomannan antigen; Pulmonary tuberculosis.