Authors’ response

Sir,

We thank Kalhoro et al¹ for providing comments on our work on the validation of the PathoDetect™ MTB RIF & INH assay² and for acknowledging the uniqueness of the same.

The point regarding the exclusion of extra-pulmonary tuberculosis (EPTB) in our study is important and well received. EPTB constitutes a significant portion of the global TB burden (15-20%) and accounts for 20 to 24 per cent of all TB cases across age groups in India³. EPTB cases are often more complex to diagnose due to their pauci-bacillary nature and requires multifaceted approach. Molecular tests with their better sensitivity have the potential to improve the diagnosis of EPTB. Although the PathoDetect™ assay can reportedly detect TB in EPTB samples, this was not evaluated in our study. In addition, as highlighted by Kalhoro et al¹, the Abbott RealTime MTB RIF/INH assay showed 98 to 100 per cent accuracy for PTB and EPTB, however, the number of EPTB samples were less⁴. Due to variable sensitivity to specificity of available diagnostic kits for Mycobacterium tuberculosis (MTB), and across sample types, reported earlier^5,6, we do agree that there is a need for improved diagnostics across sample types for EPTB in future diagnostic accuracy studies.

Furthermore, we also acknowledge the importance of molecular testing in people living with HIV (PLHIV), where TB diagnosis is challenging due to paucibacillary nature. Regarding the comment related to use of two sputum samples to increase diagnostic yield, it is true that using two samples instead of one increase TB detection by around 10 per cent in most studies⁷, this primarily applies to acid fast Bacilli (AFB) smear microscopy, which is relatively less sensitive. However, for molecular tests such as Xpert MTB/RIF, there is no significant difference in TB detection between early morning and spot sputum samples in routine programmatic settings^8,9. It has also been recommended ruling out the TB disease with a single negative Xpert MTB/RIF test¹⁰. In this study, we attempted to collect early morning samples; however, if the patient could not return, one/two spot sample half an hour apart, wherever possible, was accepted, as both appear equivalent when using molecular test for TB diagnosis.

In addition, we used a cross-sectional study design, which is widely regarded as appropriate for evaluating the diagnostic accuracy of a new index test against established reference standards¹¹. Kalhoro et al¹‘s contention that long-term study would have provided information about how well the test works during treatment or disease progression, will not be applicable to molecular tests targeting DNA that detect both live and dead bacilli. Poor specificity of Xpert MTB/RIF for monitoring TB treatment response was also reported earlier¹². Sputum smear microscopy and culture are still considered the mainstay for monitoring the treatment response in TB patients.