Role of Sequential Approach to Sample Collection in Microbiological Diagnosis of Smear Negative Pulmonary Tuberculosis.

Abstract

Background: The current diagnostic tools for pulmonary tuberculosis (PTB) fail in detecting up to 43% of cases, leading to inappropriate use of antibiotics and delayed diagnosis. Induced sputum and bronchoalveolar lavage (BAL) are more accurate samples, but their use is not clearly recommended. This study evaluates the real-world feasibility and effectiveness of using a sequential collection of respiratory samples to diagnose PTB.

Materials and methods: PTB cases with sputum-scarce or smear-negative status were included. Sputum induction (SI) and BAL samples were collected. We used a composite reference standard (CRS) method as the reference.

Results: Of 220 cases screened, 156 were enrolled. Eighty cases underwent successful SI (group A), and 76 cases failed induction (group B). In group A, 53.75% of cases were positive for Mycobacterium tuberculosis (MTB) by cartridge-based nucleic acid amplification test (CBNAAT). Of the 112 cases that underwent flexible optic bronchoscopy (FOB), 85 were detected positive for PTB, of which eight were detected rifampicin-resistant by CBNAAT in BAL samples. The remaining cases were found to have an alternative diagnosis. Using the sequential approach, CBNAAT had a higher cumulative sensitivity (97.72%), specificity (100%), and diagnostic accuracy (DA) (98.08%) with a p-value < 0.001, as compared to either smear examination or culture. 27.56% of cases were able to avoid bronchoscopy using the sequential technique.

Conclusion: The sequential technique of sample collection increased the bacteriological confirmation of PTB.