Mycobacterial antigens in pleural fluid mononuclear cells to diagnose pleural tuberculosis in HIV co-infected patients

Abstract

BACKGROUND: Extra pulmonary manifestation of tuberculosis (TB) accounts for approximately one-half of TB cases in HIV-infected individuals with pleural TB as the second most common location. Even though mycobacteria are cleared, mycobacterial antigens may persist in infected tissues, causing sustained inflammation and chronicity of the disease. The aim of this study was to explore various mycobacterial antigens in pleural effusions, the impact of HIV infection and CD4+ T-cell depletion on the presence of antigens, and the diagnostic potential of antigens for improved and rapid diagnosis of pleural TB. METHODS: Pleural fluid specimens were collected from patients presenting with clinically suspected pleural TB, and processed routinely for culture, cytology, and adenosine deaminase activity analysis. HIV status and CD4+ T-cell counts were recorded. Pleural fluid mononuclear cells (PFMC) were isolated, and cell smears were stained with acid-fast staining and immunocytochemistry for various mycobacterial antigens. Real-time and nested-PCR were performed. Patients were categorized as pleural TB or non-TB cases using a composite reference standard. Performance of the mycobacterial antigens as diagnostic test was assessed. RESULTS: A total of 41 patients were enrolled, of which 32 were classified as pleural TB and 9 as non-TB. Thirteen patients had culture confirmed pleural TB, 26 (81%) were HIV-TB co-infected, and 64% had < 100 CD4+ T-cells/microL. Both secreted and cell-wall mycobacterial antigens were detected in PFMC. Lipoarabinomannan (LAM) was the most frequently detected antigen. There was no direct correlation between positive culture and antigens. Cases with low CD4+ T-cell counts had higher bacterial and antigen burden. By combining detection of secreted antigen or LAM, the sensitivity and specificity to diagnose pleural TB was 56 and 78%, respectively, as compared to 41 and 100% for culture, 53 and 89% for nested PCR, and 6 and 100% for real-time PCR. CONCLUSION: Mycobacterial antigens were detectable in PFMC from tuberculous pleural effusions, even in cases where viable mycobacteria or bacterial DNA were not always detected. Thus, a combination of secreted antigen and LAM detection by immunocytochemistry may be a complement to acid-fast staining and contribute to rapid and accurate diagnosis of pleural TB.