Multidrug-resistant (MDR) and extensively drug-resistant tuberculosis (XDR-TB) is a serious
problem in South Africa. Standardised and accurate drug susceptibility testing for first-line
and second-line anti-TB drugs are critical elements in battling the MDR- and XDR-TB
epidemic. Furthermore, understanding of the population structure and transmission patterns
of circulating drug-resistant Mycobacterium tuberculosis (M. tuberculosis) strains is of utmost
importance. Previous studies have shown that different genotypes of M. tuberculosis strains
predominate in different geographical regions and strain to strain variations may have
important consequences when it comes to transmissibility, diagnostics and management.
Despite the high prevalence of MDR- and XDR-TB in South Africa, data with regard to the
population structure of drug-resistant strains is lacking in many regions. Therefore, the aim of
the study was to evaluate phenotypic and genotypic methods for susceptibility testing of firstline
and second-line anti-TB drugs and molecular characterisation of drug-resistant strains
from high-burden TB areas in South Africa.
Consecutive MDR-TB isolates from four provinces were included in this study. The results
of routine drug susceptibility testing of ethambutol, streptomycin, kanamycin and ofloxacin
using the BACTEC MGIT 960 system in MDR-TB isolates was compared to the agar
proportion method. The GenoType® MTBDRsl assay was evaluated for the detection of
resistance against ethambutol, capreomycin, kanamycin and ofloxacin. Finally, the study investigated the population structure and transmission patterns of drug-resistant
M. tuberculosis isolates using spoligotyping and MIRU-VNTR typing. Additionally, the
possible association of genotypes with drug-resistance patterns and demographic information
The sensitivity and specificity of the BACTEC MGIT 960 system using the agar proportion
method as a gold standard was 18.8% and 96.5% for ethambutol, 95.2% and 37.4% for
streptomycin, 54.6% and 91.8% for kanamycin and 100% and 89% for ofloxacin,
The GenoType® MTBDRsl assay gave comparable results for ofloxacin; however, a low
performance was found for the detection of resistance to capreomycin, kanamycin and
ethambutol. In order to resolve the discrepancies between the two methods, DNA sequencing
was done for the target genes (gyrA, rrs, embB) from the discrepant results as well as two
additional genes (gyrB, tlyA) that were not included in the assay. The DNA sequencing
identified mutations in the gyrA, gyrB, rrs and tlyA genes that were not detected by the
GenoType® MTBDRsl assay.
The prevalence of XDR-TB and pre-XDR TB was 7.1% and 9.5%, respectively. A highdiversity
of M. tuberculosis strains were found in this study, with the Beijing and EAI1_SOM
families being predominant. No association was found between genotypes and specific drugresistance
or demographic information.
The sensitivity of the BACTEC MGIT 960 system for ofloxacin and streptomycin was
excellent; however, the sensitivity was low for ethambutol and kanamycin. The GenoType®
MTBDRsl assay was promising for the detection of ofloxacin; however, the sensitivity of the
assay needs to be improved for capreomycin, kanamycin and ethambutol. The high-level of
diversity and the geographical distribution of the drug-resistant M. tuberculosis isolates in this
study suggested that the transmission of drug-resistant TB in these study settings is not caused
by clonal spread of a specific M. tuberculosis strain.