Diagnostic associations between pleural and pulmonary tuberculosis

Abstract

To evaluate diagnostic differences and associations between pleural tuberculosis (PLTB), and pulmonary tuberculosis alone (PTB), as well as, pleural and pulmonary co-infection (PLPTB) and PTB. Data from a 6-year, population-based study in Houston, Texas, were analyzed by logistic regression modeling. Study personnel interviewed 2226 tuberculosis cases for TB risk assessment and reviewed medical records. Mycobacterium tuberculosis isolates from available culture-positive patients underwent genotyping. Among the categories analyzed [PTB, n=2070; PLTB, n=95; and PLPTB, n=61], mortality within 180 days of TB presentation was highest for PLPTB patients (HR=1.8, P=0.045), followed by PTB (HR=1.0, referent), and PLTB (HR=0.97, P=0.9), after adjustment for age and HIV. Multivariate analysis identified history of liver disease, MTB culture negativity, having symptoms for <60 days, chest pain, and age>64 years as independent risk factors for PLTB. Symptoms for <60 days and HIV seropositivity predicted PLPTB. Patients with PLTB were significantly less likely to be genotypically clustered than patients presenting with PTB and patients diagnosed with PLPTB were significantly more likely to be identified has part of a cluster than pulmonary TB patients. PLTB presented as a more acute infection with additional symptoms and fewer symptomatic days, while PLPTB presented as a more severe disease with higher mortality compared with PTB. Given the different manifestations seen in this study, the biological mechanisms by which the pleural space is involved for PLTB and PLPTB may be substantially different.