Diagnostic yield of polymerase chain reaction on induced sputum for pulmonary tuberculosis: A single-center retrospective cross-sectional study.

BACKGROUND

Data on tuberculosis-polymerase chain reaction (TB-PCR) diagnostic yield in induced sputum (IS) samples is limited. This study was conducted to evaluate the diagnostic yield of TB-PCR in IS samples from patients with pulmonary TB and to identify factors that are associated with positive TB-PCR results.

METHODS

This retrospective cross-sectional study was conducted at the Faculty of Medicine Ramathibodi Hospital. Patients who underwent IS collection for the diagnosis of pulmonary TB were included. Sputum specimens were obtained for acid-fast bacilli (AFB) smear, TB-PCR (Anyplex Seegene MTB/NTM real-time detection assay or Xpert MTB/RIF assay), and TB culture. Multivariate logistic regression analysis was performed to identify factors associated with IS TB-PCR positivity. The McNemar test was used to compare the diagnostic yield of each test.

RESULTS

A total of 124 IS specimens of patients with pulmonary TB were evaluated. There were 65 (52.4%) men, with a mean age of 55.3 ± 19.5 years. The diagnostic yield of IS TB-PCR for the diagnosis of pulmonary TB was 31.5% (95% confidence interval [CI]: 23.2-39.7). The diagnostic yields were 34.4% (95% CI: 22.0-46.0) for Xpert MTB/RIF and 28.6% (95% CI: 17.8-40.2) for Anyplex MTB/NTM, with no significant difference between the two assays ( = 0.49). TB-PCR had a higher diagnostic yield than AFB smear (31.5% vs. 6.5%, < 0.01). Logistic regression analysis showed that moderately advanced (adjusted odds ratio [aOR] = 3.73, 95% CI: 1.24-11.21, = 0.019) and far advanced (aOR = 3.95, 95% CI: 1.05-14.82, = 0.042) radiographic extent of disease were associated with positive IS TB-PCR.

CONCLUSIONS

Induced sputum TB-PCR is an effective initial method for patients with suspected pulmonary TB who are unable to produce reliable sputum, especially those with moderately advanced or far advanced radiographic extent of disease.