Commercial Nucleic-Acid Amplification Tests for Diagnosis of Pulmonary Tuberculosis in Respiratory Specimens: Meta-Analysis and Meta-Regression
2008

Meta-Analysis of Commercial Tests for Diagnosing Pulmonary Tuberculosis

Sample size: 125 publication 10 minutes Evidence: moderate

Author Information

Author(s): Daphne I. Flores, Laura L. Riley, Lee W. Riley, Madhukar Pai

Primary Institution: University of California, Berkeley

Hypothesis

What factors are associated with higher accuracy in commercial nucleic-acid amplification tests for diagnosing pulmonary tuberculosis?

Conclusion

Commercial nucleic-acid amplification tests for diagnosing pulmonary tuberculosis show high variability in sensitivity and specificity, indicating they cannot replace conventional tests.

Supporting Evidence

  • The pooled sensitivity was 0.85, indicating that the tests missed some cases of tuberculosis.
  • The pooled specificity was 0.97, showing that the tests were good at identifying those without the disease.
  • Significant heterogeneity was found in the results, suggesting variability in test performance across studies.
  • Subgroup analyses indicated that the type of specimen used affected test accuracy.
  • Studies using respiratory specimens had higher diagnostic odds ratios compared to those using only sputum.

Takeaway

This study looked at tests that help doctors find out if someone has tuberculosis. It found that while these tests can be fast, they don't always work well, especially for some patients.

Methodology

The study performed a meta-analysis of 125 studies evaluating the accuracy of commercial nucleic-acid amplification tests for tuberculosis diagnosis.

Potential Biases

Many studies did not report blinding status or recruitment methods, which could introduce bias.

Limitations

The variability in test results and the lack of a perfect reference standard limit the clinical applicability of the findings.

Participant Demographics

The studies included a mix of patients with both smear-positive and smear-negative tuberculosis.

Statistical Information

P-Value

p<0.001

Confidence Interval

95% CI: 26.02, 41.22

Statistical Significance

p<0.001

Digital Object Identifier (DOI)

10.1371/journal.pone.0001536

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication