TB-EFI, a novel 18-Plex microbead-based method for prediction of second-line drugs and ethambutol resistance in Mycobacterium tuberculosis complex.

Several diagnostic tests are being developed to detect drug resistance in tuberculosis. In line with previous developments detecting rifampicin and isoniazid resistance using microbead-based systems (spoligoriftyping and TB-SPRINT), we present here an assay called TB-EFI detecting mutations involved in resistance to ethambutol, fluoroquinolones and the three classical injectable drugs (kanamycin, amikacin and capreomycin) in Mycobacterium tuberculosis. The proposed test includes both wild-type and mutant probes for each targeted locus. Basic analysis can be performed manually. An upgraded interpretation is made available in Excel 2016®. Using a reference set of 61 DNA extracts, we show that TB-EFI provides perfect concordance with pyrosequencing. Concordance between genotypic resistance and phenotypic DST was relatively good (72 to 98% concordance), with lower efficiency for fluoroquinolones and ethambutol due to some untargeted mutations. When compared to phenotypical resistance, performances were similar to those obtained with Hain MTBDRsl assay, possibly thanks to the use of automatized processing of data although some mutations involved in fluoroquinolone resistance could not be included. When applied on three uncharacterized sets, phenotype could be predicted for 51% to 98% depending on the setting and the drug investigated, detecting one extensively drug-resistant isolate in each of a Pakistan and a Brazilian set of 91 samples, and 9 XDR among 43 multi-resistant Kazakhstan samples. By allowing high-throughput detection of second-line drugs resistance and of resistance to ethambutol that is often combined to second-line treatments, TB-EFI is a cost-effective assay for large-scale worldwide surveillance of resistant tuberculosis and XDR-TB control.