top of page


Excellent specificity and moderate to robust sensitivity of Xpert MTB/XDR for TB drug-susceptibility testing compared against Deeplex Myc-TB in nine African countries

F Massou(1,2) C Vuchas(3) J C Semuto Ngabonziza(4,15) S C Agbla(5) O El Tayeb(6) M K Kaswa(7,8,9) G Abebe(10) L Camara(11) B Diarra(12) C Merle(13) B C de Jong(14) D Affolabi(1) L Rigouts(2,14)

1:Laboratoire de Référence de Mycobactéries, Cotonou, Benin; 2:Biomedical Sciences, Antwerp University, Antwerpen, Belgium; 3:The Bamenda Center for Health Promotion and Research, Bamenda, Cameroon; 4:Rwanda Biomedical Center, Kigali, Rwanda; 5:London School of Hygien and Tropical Medicine, London,The UK; 6:Damien Foundation, Ibadan, Nigeria; 7:Institut National de Recherche Biomedicale, Kinshasa, DRC; 8:DRC National TB Program; 9:University of Kinshasa, School of Medicine, DRC; 10:Jimma University, Jimma, Ethiopia; 11:Service de Pneumophtisiologie, Conakry, Guinea; 12:Université des Sciences, des Techniques et des technologies de Bamako, Bamako, Mali; 13:UNICEF/UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases, Geneva, Switzerland; 14:Mycobacteriology Unit, Institute of Tropical Medicine (ITM), Antwerpen, Belgium; 15:University of Rwanda, Kigali, Rwanda

The GeneXpert MTB/XDR® cartridge (Cepheid, USA) is a near point-of-care tool for rapid drug-susceptibility testing directly on clinical samples for isoniazid (INH), fluoroquinolones (FQ), ethionamide (ETH), and amikacin (AMK).

We evaluated Xpert-XDR under field conditions against Deeplex Myc-TB® (Genoscreen, France), a targeted next generation sequencing test, using samples from the DIAMA project (equal number of rifampicin-resistant and -susceptible patients collected between 2017 and 2021 in nine African countries). Both tests were conducted on the same sputum sample.

A total of 1138 samples were included in this evaluation. The sensitivity (Se) and specificity (Sp) with 95% confidence intervals for the Xpert-XDR cartridge were:

  • INH: Se = 93% [91-95], Sp = 96% [95-97]

  • FQ: Se = 65% [48-79], Sp = 100% [99-100]

  • ETH: Se = 54% [48-60], Sp = 99% [99-100]

  • AMK: Sp = 100% [99-100]

Part of mutations missed by Xpert-XDR were attributable to heteroresistance: 11/45 for INH and 10/12 for FQ. For ETH, 130 mutants were not detected: 94% had mutations in regions not covered by the test, and the remaining 6% all shared the same mutation (S94A) in the inhA promoter. Sensitivity for AMK-resistance detection could not be assessed within our sample.

The Xpert-XDR demonstrated a high specificity across all tested drugs and robust sensitivity for INH, highlighting its utility in rapid detection of second-line TB drug resistance, despite the limited ability in detecting heteroresistance and mutations outside targeted regions. Extended evaluations in regions representing other resistance profiles is warranted.

bottom of page