Comparative Evaluation of Sputum GeneXpert, Smear Microscopy, and Culture in the Diagnosis of Pulmonary Tuberculosis

Abstract

Background: Early and accurate diagnosis of pulmonary tuberculosis (PTB) is critical for timely treatment and reduction of transmission. Conventional smear microscopy, though widely available, has limited sensitivity, whereas culture is time-consuming. The GeneXpert MTB/RIF assay, a rapid molecular diagnostic tool, has emerged as a promising alternative for simultaneous detection of Mycobacterium tuberculosis and rifampicin resistance.

Aim: To compare the diagnostic efficacy of sputum GeneXpert, smear microscopy, and culture in detecting pulmonary tuberculosis among suspected patients.

Methods: A cross-sectional analytical study was conducted among 160 patients with clinical suspicion of PTB at a tertiary care hospital. Two sputum samples per patient were processed for Ziehl-Neelsen smear microscopy, GeneXpert MTB/RIF assay, and Löwenstein-Jensen (LJ) culture. Culture served as the reference standard. Diagnostic accuracy indices-sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV)-were computed with 95% confidence intervals. Concordance and rifampicin resistance detection were also analyzed.

Results: Of 160 samples, 94 (58.8%) were culture-positive. GeneXpert showed sensitivity 93.6% and specificity 90.9%, compared to smear microscopy (71.3% and 93.9%, respectively). The overall diagnostic agreement with culture was 92.5% for GeneXpert (κ = 0.85) and 80.6% for smear microscopy. Among culture-positive cases, rifampicin resistance was detected in 11.6% by culture-DST and 12.6% by GeneXpert, with near-perfect concordance (sensitivity 90.9%, specificity 98.8%, p = 1.000).

Conclusion: GeneXpert MTB/RIF demonstrated superior sensitivity and diagnostic concordance with culture while rapidly identifying rifampicin resistance. It offers a robust, rapid, and accurate diagnostic option for PTB, supporting its role as a frontline test in TB diagnostic algorithms.