Melting the Eis: non-detection of kanamycin resistance markers by routine diagnostic tests and identification of new eis-promoter variants

Until 2017, eis-promoter mutations and phenotypic kanamycin (KAN) resistance were not routinely tested for in South Africa. This could have led to undetected resistance, leading to partially empiric treatment regimens with decreased effectiveness. In this study, we analyzed two different datasets to investigate the presence, type, and detection of eis-promoter mutations in clinical Mycobacterium tuberculosis isolates collected in the Western Cape Province of South Africa. Using MTBDRsl, Sanger sequencing, whole genome sequencing and phenotypic drug susceptibility testing, the prevalence of eis-promoter mutations missed by routine diagnostic tools were determined and associations between genotype and phenotype were assessed. Across both datasets, 17/410 tested isolates were wrongly classified as eis-promoter wildtype by MTBDRsl. Six of these isolates harbored eis-promoter mutations known to confer low-level KAN resistance. In addition, three previously undescribed eis-promoter mutations not conferring KAN resistance were detected. The combination of whole genome sequencing and repeated pDST furthermore revealed a mixed infection with an underlying high-level KAN resistant strain that would have been missed or misinterpreted using a singular method only.Although the proportion of missed low-level KAN resistance was low in this setting, it nevertheless represents a potential threat for treatment failure and increasing drug resistance due to partially ineffective treatment regimens. Unambiguous classification of bacteria as susceptible or resistant remains crucial for diagnosis, treatment, and control of tuberculosis, especially with the increasing thread of extensively drug resistant strains where every remaining effective drug could make the difference between life and death.