Diagnostic accuracy of Lipoarabinomannan detection by lateral flow assay in pleural tuberculosis

Abstract

Abstract Background Lipoarabinomannan (LAM) antigen serves as an attractive biomarker to diagnose Tuberculosis (TB). Given the limitations of current diagnostic modalities for Pleural TB, current study evaluated LAM’s potential to serve as a point-of-care test to diagnose pleural TB. Methods A cross sectional, diagnostic accuracy study was conducted during February to November 2021 in a tertiary care hospital in India. LAM antigen detection was performed on pleural fluid as well as early morning urine specimen of suspected pleural TB patients by "Alere/ Abott Determine TB LAM" lateral flow assay (LAM-LFA). The results were compared to microbiological reference standards/MRS (Mycobacterial culture or NAAT) and Composite reference standards/CRS (MRS plus Clinico-radiological diagnosis). Results A total of 170 subjects were included in the analysis, including 26 with Definite TB, 22 with Probable TB, and 122 with No TB. Compared to MRS and CRS, the sensitivity (61.54% & 45.83%) and positive predictive value (PPV) (57.14 & 78.57%) of Pleural LAM-LFA testing were found to be suboptimal, whereas the specificity (91.67% & 95.08%) and negative predictive value (NPV) (92.96% & 81.69%) of the assay were found to be good. Urinary LAM-LFA performed even worse than pleural LAM-LFA, except for its higher specificity against MRS and CRS (97.2% and 98.3%, respectively). Specificity and PPV of pleural LAM detection increased to 100% when analysed in a subgroup of patients with elevated ADA levels (receiver operating curve analysis-derived cut off value > 40 IU/ml). Conclusion Detection of LAM antigen by LFA directly from pleural fluid performed fairly accurate in diagnosing pleural TB and was highly predictive of the disease in a selected cohort of patients making it a valuable POCT.