Abstract
INTRODUCTION: Existing tests for the diagnosis of pleural tuberculosis (TB) have major limitations in terms of accuracy, time to diagnosis and drug resistance testing. A test which can diagnose pleural TB and detect resistance, like Xpert MTB/Rif, would be optimal for rapid diagnosis and treatment. METHODS: A prospective observational study was done in a tertiary care hospital in Eastern Nepal. Fifty-one patients with clinic-radiologic suspicion of pleural TB were included. The results of pleural fluid Xpert MTB/Rif were compared with two Composite Reference Standards. Composite Reference Standard-1 consisted of positive pleural fluid smear, positive culture, positive histology of pleural biopsy, and positive sputum results. Composite Reference Standard-2 included those with Composite Reference Standard-1 and those with high ADA values (>40 U/l) with response to anti-tubercular treatment at 8 weeks of follow-up. RESULTS: Thirty-six patients were diagnosed as Pleural TB. Nine fulfilled Composite Reference Standard-1. Pleural fluid Xpert MTB/Rif was positive in five cases with Composite Reference Standard-1 and nine cases with Composite Reference Standard-2. The sensitivity, specificity, positive predictive value and negative predictive value with reference to Composite Reference Standard-1 were 55.56%, 88.10%, 50%, and 90.24%, respectively. Using Composite Reference Standard-2 as reference, sensitivity, specificity, positive predictive value and negative predictive value were 25%, 93.33%, 90%, and 34.15%, respectively. Two cases were diagnosed Xpert Rif resistant on pleural fluid. CONCLUSION: Due to low sensitivity, the Xpert MTB/Rif test cannot be recommended as initial test of diagnosis in a high prevalence setting. At the same time its clinical utility lies in testing of patients suspected to have drug-resistant pleural tuberculosis.