Background: Diagnosis of tuberculous pleuritis is difficult because of its nonspecific clinical presentation and insufficient efficiency of traditional diagnostic methods. We investigated the use of adenosine deaminase (ADA) activity in tuberculous pleuritis diagnosis.
Methods: We optimized a kinetic assay and retrospectively analysed 210 patients with exudative pleural effusions. Using the ROC curve, we determined the optimal cutoff for TB pleurisy.
Results: One hundred forty-seven exudative samples were nontuberculous (non-TB) and 63 were tuberculous (TB). There was statistically significant difference (p<0.0001) between the means of pleural fluid ADA levels among the TB and non-TB populations. The disease prevalence of TB pleurisy in the studied population was 30%. The cutoff value for diagnosing TB effusions was >55.8 U/L, with a sensitivity of 87.3% (95% CI: 76.5-94.3%) and specificity of 91.8% (95% CI: 86.2-95.7%). The positive predictive value (PPV) was 82.1% and the negative predictive value (NPV) was 94.4%. A pleural fluid ADA value <16.81 IU/L suggests that a tuberculous effusion is highly unlikely (100% sensitive with 100% NPV and 0% negative likelihood ratio for a pleural fluid ADA level>/=16.81 U/L). In addition, the area under the ROC curve was 0.933 (S.E.=0.0230, 95% CI: 0.890-0.963).
Conclusion: Pleural fluid total ADA assay is a sensitive and specific test suitable for rapid diagnosis of TB pleurisy.