Non-commercial phenotypic assays for the detection of Mycobacterium tuberculosis drug resistance: a systematic review.

Several rapid non-commercial culture-based methods and assays for drug susceptibility testing (DST) of Mycobacterium tuberculosis have emerged over the last decades. The aim of the current review was to summarise evidence on the performance of microscopic observation of drug susceptibility (MODS), thin-layer agar (TLA) and colorimetric redox-indicator (CRI) assays for detection of resistance to first- and second-line anti-tuberculosis (TB) drugs. Forty-three publications satisfying selection criteria were selected for data extraction. MODS and CRI assays demonstrated pooled sensitivity and specificity of > 93% for the detection of resistance to rifampicin and isoniazid and confirmed their utility for an accurate detection of multidrug-resistant TB (MDR-TB) in various settings. Sensitivity and specificity values for indirect DST for ethambutol (EMB) using CRI assays were 94.0% and 82.0%, respectively, suggesting that CRIs could be used to rule out resistance to EMB. Performance for other drugs varied more substantially across the reports. There was no sufficient evidence on the performance of the TLA assay for making any conclusion on its utility for DST. Our data suggests that non-commercial assays could be used for a rapid and accurate DST in settings where the use of commercial World Health Organization-endorsed assays could be limited due to a variety of reasons including limited resources, laboratory facilities or trained personnel. While inexpensive and easy-to-perform MODS and TLA assays can be used in low-income settings, using CRI assays for determination of minimal inhibitory concentrations may be implemented in middle- and high-income countries with high MDR-TB burden to guide clinical management of TB patients.