Colorimetric redox-indicator methods for the rapid detection of multidrug resistance in Mycobacterium tuberculosis: a systematic review and meta-analysis

Abstract

With the spread of multidrug-resistant tuberculosis (MDR-TB) there is increasing demand for new accurate and cost-effective tools for rapid drug susceptibility testing (DST), particularly for developing countries. The reference standard method used today for DST is very slow and cumbersome. Colorimetric assays using redox indicators have been proposed to be used in low-resource countries as rapid alternative culture methods for the detection of resistance especially to rifampicin and isoniazid. These methods appear as promising new tools but their accuracy has not been systematically evaluated. We did a meta-analysis to evaluate the accuracy of the colorimetric assays for the detection of rifampicin and isoniazid-resistant tuberculosis among clinical isolates. We searched Medline, PubMed (NCBI), Global health-CAB, EJS-E (EbscoHost), ISI Web, Web of Science and IFCC databases and contacted authors if additional information was needed. Eighteen studies met our inclusion criteria for rifampicin resistance detection and 16 for isoniazid. We used a summary receiver operating characteristic (SROC) curve to perform meta-analysis and summarize diagnostic accuracy. For both drugs, all studies had a sensitivity and specificity that ranged between 89% and 100%. There is evidence that colorimetric methods are highly sensitive and specific for the rapid detection of MDR-TB. These new tools could offer affordable technologies for TB laboratories especially in places where resources are limited and where the prevalence of MDR-TB is important and make TB control efforts more effective. Additional studies are needed in high MDR prevalence countries and cost-effectiveness analysis to have more evidence on the utility of these methods. Future developments to detect resistance directly from smear-positive sputum specimens should be taken into consideration to speed up the process.