Rifampicin-resistant TB: discordance between Xpert MTB/RIF and MTBDR results.

South Africa´s diagnostic algorithm for TB diagnosis from 2011 to 2017 employed the Xpert MTB/RIF assay as the initial screening test for TB diagnosis and rifampicin (RIF) susceptibility, followed by submission of a specimen for GenoType MTBDR. This study aimed to determine the concordance between the two assays in terms of RIF susceptibility and explore reasons for discordance. This was a retrospective laboratory-based study that included all MTBDR results of tests performed at the Braamfontein Mycobacteriology Referral Laboratory between 1 September 2014 and 31 August 2015. The patient´s Xpert RIF result was linked with the MTBDR result. The overall concordance between RIF susceptibility results was 96.4%. There were 68 discordant RIF results. The most common reasons for discordance identified were possible false Xpert RIF-resistant results (22%), mixed infection/heteroresistance (16%), transcription errors (7%) and erroneous manual interpretation of the MTBDR strip (7%). Xpert RIF resistance detected using delayed hybridisation was associated with discordance. The overall concordance between the MTBDR and Xpert RIF results were very good. Management of discordance should include repeat specimens for Xpert and MTBDR and B sequencing. All variables should then be considered before treatment regimens are altered.