Background: The diagnosis of pleural tuberculosis continues to be a challenge due to the low sensitivity of traditional diagnostic methods. Histopathological examination of pleural tissue is the most accurate method, with a sensitivity of up to 80%. Determination of adenosine deaminase levels is a recently introduced method, although its usefulness in the diagnosis of pleural tuberculosis in Brazil has yet to better elucidated. Objective: To verify the sensitivity and specificity of an experimental method of measuring adenosine deaminase activity in pleural fluid in a series of patients with pleural effusion patients evaluated between August 1998 and November 2002 in Rio de Janeiro (RJ). Results: Out of 137 cases, 111 pleural fluid samples were available. Of those, 83 were from pleural tuberculosis patients. Among the 67 pleural tuberculosis patients tested, 10 (14.9%) presented human immunodeficiency virus. The adenosine deaminase cutoff value of 35U/L was determined by a receiver operator characteristic curve. The sensitivity, specificity and likelihood ratios (positive and negative) were 92.8%, 93.3%, 25.8 and 13.9, respectively. Mean adenosine deaminase in the pleural tuberculosis group was 84.7 ± 43.1 U/L, versus 15.9 ± 11.1 U/L in the group with other diseases. There was no significant difference in adenosine deaminase activity between patients with and without human immunodeficiency virus co-infection. Conclusions: Adenosine deaminase measurement in pleural fluid is a sensitive and specific method for the diagnosis of pleural tuberculosis and its use can preclude the need for pleural biopsy in the initial workup of pleural effusion patients. An adenosine deaminase cutoff value of 35U/L is recommended.

Keywords: Pleural fluid. Adenosine deaminase. Diagnosis. Tuberculosis. HIV