Xpert MTB/RIF Ultra-resistant and MTBDRplus-susceptible rifampicin results in people with tuberculosis: utility of FluoroType MTBDR and deep sequencing.

Abstract

Xpert MTB/RIF Ultra (Ultra)-detected rifampicin-resistant tuberculosis (TB) is often programmatically confirmed using MTBDRplus. There are limited data on discordant results, including when re-tested using newer methods, like FluoroType MTBDR (FT-MTBDR) and targeted deep sequencing. MTBDRplus rifampicin-susceptible isolates from people with Ultra rifampicin-resistant sputum were identified from a South African programmatic laboratory. FT-MTBDR and single molecule-overlapping reads (SMOR; rpoB, inhA, katG) on isolate DNA were done (SMOR was used as a reference standard). Between 1 April 2021 and 30 September 2022, 8% (109/1347) of Ultra rifampicin-resistant specimens were MTBDRplus-susceptible. Of 89% (97/109) isolates with a sequenceable rpoB, SMOR resolved most in favor of Ultra (79% [77/97]). Sputum with lower mycobacterial load was associated with Ultra false-positive resistance (46% [11/24] of "very low" Ultra had false resistance vs 12% [9/73; P = 0.0004] of ≥"low"), as were Ultra heteroresistance calls (all wild-type probes, ≥1 mutant probe) (62% [23/37 vs 25% 15/60] for Ultra without heteroresistance calls; P = 0.0003). Of the 91% (88/97) of isolates successfully tested by FT-MTBDR, 55% (48/88) were FT-MTBDR rifampicin-resistant and 45% (40/88) susceptible, translating to 69% (47/68) sensitivity and 95% (19/20) specificity. In the 91% (99/109) of isolates with inhA and katG sequenced, 62% (61/99) were SMOR isoniazid-susceptible. When Ultra and MTBDRplus rifampicin results are discordant, Ultra is more likely to be correct, and FT-MTBDR agrees more with Ultra than MTBDRplus; however, lower load and the Ultra heteroresistance probe pattern were risk factors for Ultra false rifampicin-resistant results. Most people with Ultra-MTBDRplus discordant resistance results were isoniazid-susceptible. These data have implications for drug-resistant TB diagnosis.