This work aims to investigate the diagnostic accuracy of a nucleic acid amplification test (FluoroType MTB®) in pleural fluid (PF) and sputum to diagnose tuberculous pleural effusion (TPE). We also analyzed the increase in diagnostic accuracy of a second FluoroType MTB® test on a second thoracentesis sample when the first was negative.

We conducted a prospective single-center study that included 207 patients with pleural effusion (31 tuberculous and 176 due to other causes). Of the 31 cases of TPE, 21 (68%) were confirmed histologically or microbiologically; the other cases were considered probable.

The operational characteristics of FluoroType MTB® in PF for identifying tuberculosis were a sensitivity of 13%, a specificity of 99%, a positive likelihood ratio of 11, and a negative likelihood ratio of 0.9. The diagnostic efficacy data for sputum samples were 21%, 91%, 2.4, and 0.9, respectively. PF and sputum cultures in solid and liquid media had greater sensitivity (36% and 31%, respectively). A second FluoroType MTB® test in PF was negative for 24 patients who had TPE and for whom the first FluoroType MTB® test was also negative. Only two (6.5%) patients with TPE had a confirmed diagnosis based exclusively on the positive results of the FluoroType MTB® in PF.

Due to its low sensitivity, the FluoroType MTB® test in PF has a limited role in diagnosing tuberculous pleurisy.

This work aims to investigate the diagnostic accuracy of a nucleic acid amplification test (FluoroType MTB®) in pleural fluid (PF) and sputum to diagnose tuberculous pleural effusion (TPE). We also analyzed the increase in diagnostic accuracy of a second FluoroType MTB® test on a second thoracentesis sample when the first was negative.

We conducted a prospective single-center study that included 207 patients with pleural effusion (31 tuberculous and 176 due to other causes). Of the 31 cases of TPE, 21 (68%) were confirmed histologically or microbiologically; the other cases were considered probable.

The operational characteristics of FluoroType MTB® in PF for identifying tuberculosis were a sensitivity of 13%, a specificity of 99%, a positive likelihood ratio of 11, and a negative likelihood ratio of 0.9. The diagnostic efficacy data for sputum samples were 21%, 91%, 2.4, and 0.9, respectively. PF and sputum cultures in solid and liquid media had greater sensitivity (36% and 31%, respectively). A second FluoroType MTB® test in PF was negative for 24 patients who had TPE and for whom the first FluoroType MTB® test was also negative. Only two (6.5%) patients with TPE had a confirmed diagnosis based exclusively on the positive results of the FluoroType MTB® in PF.

Due to its low sensitivity, the FluoroType MTB® test in PF has a limited role in diagnosing tuberculous pleurisy.