P075

Do we use the right breakpoints for delamanid susceptibility testing ?

N Lequerré(1) C Poignon(1,2) E Gyde(2) J Robert(1,2) N Veziris(1,2,3) A Saffarian(2) A Aubry(1,2)

1:UPMC - Paris 6; 2:Hôpital Pitié-Salpetriere; 3:Hôpital Saint-Antoine

As delamanid (DLM) is increasingly used in the treatment of drug-resistant tuberculosis (MDR/XDR-TB), accurate DLM drug susceptibility testing (DST) is essential. DST methods recommended by WHO are the proportion method in 7H11 (PM) and the BACTEC™ Mycobacterial Growth Indicator Tube 960™ (MGIT).

We aimed to evaluate the concordance between these methods for DLM and to compare those results to whole genome sequencing data (WGS).

We performed DST following WHO recommendations in MGIT (CC=0.06 mg/L) and in PM (WHO CC=0.016 mg/L + additional CC=0.06 mg/L) against 126 MDR/XDR-TB clinical isolates. Additionally, DLM genotypic resistance markers were identified by WGS (fbiA-D, fgd1, ddn, Rv2983).

The agreement between MGIT and PM was moderate using CCᵂᴴᴼ and almost perfect using CC⁰·⁰⁶ (k=0.51 and 0.90, respectively). Among discrepant isolates (n=10 with CCᵂᴴᴼ, n=1 with CC⁰·⁰⁶), no mutation known to be involved in DLM-R was identified by WGS, but mutations of unknown impact were found in 5/10 using CCᵂᴴᴼ and in the strain using CC⁰·⁰⁶.

Among strains categorized as DLM-R by both methods (n=6 with CCᵂᴴᴼ, n=5 with CC⁰·⁰⁶), WGS identified mutations known as involved in DLM-R in 3/6 using CCWHO and 3/5 using CC⁰·⁰⁶.

Among strains categorized as DLM-S by both methods (n=110 with CCᵂᴴᴼ, n=120 with CC⁰·⁰⁶), WGS identified mutations of unknown impact in 54.2% (n=60) using CCᵂᴴᴼ and 54.2% (n=65) using CC⁰·⁰⁶.

The discrepancies results suggest that increasing PM's CC to 0.06 mg/L allows better concordance with MGIT DST and improving our knowledge regarding genotypic markers for DLM-R is crucial.