Evaluation of Colorimetric Nitrate Reductase Assay in Liquid Medium for Detection of Resistance to First-line Antitubercular Drugs.

Background: On a global scale, India holds the distinction of having the greatest number of tuberculosis (TB) cases caused by Mycobacterium tuberculosis (MTB) complex. The study aimed at evaluating the sensitivity, specificity, accuracy, cost, rapidity, and feasibility of the performance of the colorimetric nitrate reductase-based antibiotic susceptibility (CONRAS) test against the indirect proportion method (IPM) on Lowenstein-Jensen media as the gold standard.

Methods: A comparative cross-sectional study was performed on 51 MTB isolates. Fresh subcultures were used for drug susceptibility testing by IPM on the Lowenstein-Jensen medium and the CONRAS method in liquid medium. Quality control for drug susceptibility testing was done using a known sensitive strain of MTB (H37Rv) and strains resistant to both isoniazid (INH) and rifampicin (RIF) - multidrug-resistant (MDR), mono-resistant to RIF, streptomycin (STM), and ethambutol (EMB). Statistical analysis was performed using MedCalc software (Version 20.027).

Results: CONRAS, carried out in microfuge tubes, was cost-efficient and easy to perform/interpret with most results being available in 10 days compared to 42 days in the case of IPM. The sensitivity, specificity, and accuracy of RIF and INH were 100%, 97.37%, and 98.04 and 93.33%, 97.59%, and 96.08%, respectively, which translates into an almost perfect agreement between the two methods as indicated by κ value of 0.905 and 0.949, respectively, for the two drugs. The performance of CONRAS was less satisfactory for STM and EMB when compared to IPM.

Conclusions: CONRAS may serve as a useful test for the detection of MDR-TB because of its accuracy, low cost, ease of performance/interpretation, and rapidity when compared to IPM on LJ medium. It does not involve the use of expensive reagents and equipment, as is the case with molecular methods like GeneXpert and line probe assay, making it a suitable option for the detection of MDR-TB in resource-poor settings.