top of page

OR20

Nationwide evaluation of Xpert MTB/RIF Ultra reveals challenges in accurate diagnosis of rifampicin-resistant tuberculosis

I Cuella-Martin(1,4) F Hakizayezu(2) H Niyompano(2) D Runyambo(2) J Keysers(1) W B De Rijk(1) W Mulders(1) Y M Habimana(2) P Migambi(2) B C de Jong(1) L Rigouts(1,4) J C S Ngabonziza(1,2,3)

1:Institute of Tropical Medicine Antwerp; 2:Rwanda Biomedical Center; 3:University of Rwanda; 4:University of Antwerp

Xpert MTB/RIF Ultra (Ultra) was expected to improve accuracy of rifampicin-resistant tuberculosis (RR-TB) detection. To assess the diagnostic performance of Ultra in real-world settings, all RR-TB results in Rwanda had Ultra repeated on a new sample, with rpoB target vs whole genome sequencing and phenotypic drug-susceptibility testing as reference standard, with any (hetero)resistant result overriding.


This nationwide observational study from December 2021 to February 2024, included 129 patients initially identified as RR-TB by Ultra. Upon Ultra retesting only 41 (32%) remained classified as RR-TB. Reference testing, available for 40 (98%) patients, confirmed all as true RR-TB. Conversely, of the remaining 88 patients (68%) not classified as RR-TB on retesting, reference methods were available for 61 (69%), identifying only seven (11%) as true RR-TB; the other 54 (89%) were RS. Notably, this high rate of initial false RR affected samples with very low bacillary loads (49/55, 89%). The underlying technical reason has yet to be clarified; melting temperature peaks of the MUT probes did not match with expected values for known mutations, with melt curve analysis ongoing by the manufacturer.


Our study highlights a high rate of false-RR, especially in samples with very low bacillary loads, challenging the expected specificity enhancement in the Ultra system. At least 54 patients with RS-TB (42% of initial RR-TB Ultra) received unnecessary RR-TB treatment. These findings emphasize the necessity to adjust diagnostic strategies to accurately identify true RR-TB cases among paucibacillary patients, and prevent unnecessary treatments.

bottom of page