Comparison between the BACTEC MGIT 960 system and the agar proportion method for susceptibility testing of multidrug resistant tuberculosis strains in a high burden setting of South Africa

BACKGROUND The increasing problem of multi-drug-resistant (MDR) tuberculosis (TB) [ie resistant to at least isoniazid (INH) and rifampicin (RIF)] is becoming a global problem. Successful treatment outcome for MDR-TB depends on reliable and accurate drug susceptibility testing of first-line and second-line anti-TB drugs. METHOD Consecutive M. tuberculosis isolates identified as MDR-TB during August 2007 to January 2008 using the BACTEC MGIT 960 systems and the agar proportion method were included in this study. Susceptibility testing of MDR-TB isolates against ethambutol (EMB) and streptomycin (STR) as well as two second-line anti-TB drugs, kanamycin (KAN) and ofloxacin (OFX) was performed using the BACTEC MGIT 960 systems at a routine diagnostic laboratory. The results were compared to those obtained by the agar proportion method. RESULT The agreement between the BACTEC MGIT 960 system and the agar proportion method was 44% for EMB, 61% for STR and 89% for both KAN and OFX. The sensitivity and specificity of the BACTEC MGIT 960 system using the agar proportion method as a gold standard was 92% and 37% for EMB, 95% and 37% for STR, 27% and 97% for KAN and 84% and 90% for OFX, respectively. CONCLUSIONS The BACTEC MGIT 960 system showed acceptable sensitivity for EMB, STR, and OFX; however, the BACTEC MGIT 960 system was less specific for EMB and STR and demonstrated a low sensitivity for KAN. The lower agreement found between the two methods suggests the unreliability of the BACTEC MGIT 960 system for the drugs tested. The reasons for the lower agreement between the two methods need to be investigated and further studies are needed in this setting to confirm the study finding.