P15

CFZ is a group B medicine recommended by WHO to treat rifampicin-/multidrug-resistant tuberculosis (RR/MDR-TB). False-resistance was reported when drug-susceptibility testing (DST) was performed with CFZ stock solutions subsequently diluted in water. In the context of clinical trials we observed increased MIC values among reference susceptible isolates and baseline clinical isolates. Root cause analysis showed that in addition to the solvent, the polypropylene tubes yielded higher MICs, yet polystyrene (PS) tubes restored the expected MICs. We aimed to validate the MIC99 and MIC100 determination in Middlebrook 7H10 and 7H11 in PS tubes using DMSO throughout.

Analysis of accuracy and agreement were performed. A set of 23 reference CFZ-susceptible isolates (including H37Rv) and 11 Rv0678 mutants were analyzed. With a tentative breakpoint of 1 µg/ml the accuracy of MIC₉₉ and MIC₁₀₀ met the acceptance criterium (≥95%) in 7H10 (96.6% and 100% respectively) but not in 7H11 (91.2% and 94.1% respectively). Using  0.5 µg/ml as breakpoint in 7H11, the accuracies were 97.1% and 100% respectively. The agreement to classify isolates as susceptible or resistant on 7H10 or 7H11 was 93.1% using 1 µg/ml as breakpoint for 7H11, with two mutants being CFZ-S on 7H11, and 100% using 0.5 µg/ml for 7H11. 

Using PS tubes and DMSO in subsequent dilutions the MIC in 7H10 and 7H11 showed acceptable values for H37Rv, all CFZ-susceptible isolates and majority of mutants. A breakpoint of 0.5 µg/ml for 7H11 suggests improved accuracy and agreement. Determination of epidemiological cut-off is warranted to allow confident CFZ resistance calling.