Comparative diagnostic yield of raw sputum and decontaminated sputum pellets for tuberculosis detection among pulmonary TB suspects.

Early and reliable diagnosis of PTB is crucial for timely treatment and disease control. The Xpert MTB/RIF assay can be performed on either raw sputum or NALC-NaOH decontaminated concentrated pellets. However, direct comparisons of these approaches using culture as a reference standard are limited. In this cross-sectional study, 349 adults with presumptive pulmonary TB in Pune, India, provided paired sputum samples. One sample was tested directly by Xpert, while the other was decontaminated, concentrated, and tested by Xpert along with MGIT and LJ culture. MGIT culture was positive in 38.7% of participants. Xpert detected MTB in 40.7% of raw sputum and 37.2% of pellets with comparable sensitivity (88.9% vs 89.6%). However, decontaminated pellets showed higher specificity (95.8% vs 89.7%) and PPV (93.1% vs 84.5%) with improved concordance across Xpert, MGIT, LJ, and AFB smear microscopy. Semi-quantitative Xpert grades from pellets correlated more strongly with MGIT time-to-positivity (R2 = 0.44) than raw sputum (R2 = 0.22), reflecting better estimation of viable bacillary load. This indicates that pellet processing preserves sensitivity while enhancing diagnostic precision, reducing false positives, and improving culture concordance. These findings support the use of decontaminated pellets for TB detection in high-burden settings, particularly for paucibacillary, retreatment cases or HIV co-infected patients, where reliable detection is critical for clinical management.