Characterization of Mycobacterium tuberculosis isolates with discordant rifampicin susceptibility test results
Master Thesis
2018
Permanent link to this Item
Authors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher
University of Cape Town
Faculty
License
Series
Abstract
Background: The Xpert MTB/RIF assay was adopted as the initial diagnostic test for patients with presumptive tuberculosis (TB) by the South African National TB Control programme in December 2010. Rifampicin (RIF) resistance detected by the Xpert MTB/RIF (Xpert) is confirmed by a line probe assay (LPA) (GenoType MTBDRplus) and/or phenotypic (culture-based) drug susceptibility testing (DST) by MGIT (Mycobacterial Growth Indicator Tube) on the culture isolate from a 2nd specimen. Although both the Xpert and LPA target the rifampicin resistance determining region (RRDR) of the rpoB gene, discordant RIF results (Xpert RIF resistant (RIFR ), LPA RIF susceptible (RIFS )) have been reported. In addition, in cases where genotypic tests detect an rpoB mutation, inferring RIF resistance, routine phenotypic DST may report a RIF susceptible result. This is usually due to disputed rpoB mutations.
Aim: The aims of this study are to determine 1) whether the discordance between Xpert and LPA is due to false RIFR by Xpert or false RIFS by LPA and to elucidate the causes of false results and 2) the frequency and types of rpoB mutations expected to test susceptible on routine phenotypic DST and their corresponding RIF MIC (minimum inhibitory concentration).
Methods: Consecutive isolates with discordant Xpert RIFR and LPA RIFS results were selected during routine review. For the Xpert, parameters including bacterial DNA load and cycle threshold (Ct) of the probes were evaluated. In addition, isolates with a pattern of any absent rpoB WT band and absent MUT band on the LPA strip (“miscellaneous rpoB mutations”) were selected for MIC testing using the MGIT 960 system and EpiCenter TB eXiST software. Sanger sequencing of the rpoB gene from codon 462 to 591 was performed on all selected isolates.
Results and discussion: Discordant Xpert/LPA results: From the total of 1542 patients with RIFR results by Xpert, 106 (6.9%) had a discordant LPA RIFS result. Sequencing results were available for 101 isolates of which 78 (77.2%) had no rpoB mutation detected and these were categorized as false RIFR by Xpert. Mutations were detected by sequencing in the remaining 23 (22.8%); these were categorized as false RIFS by LPA. Probe delay occurred in 56/76 (73.7%) cases compared with 104/1436 (7.2%) controls (p 4 and there is a Very Low bacterial load has a positive predictive value (PPV) of 64.2 % of being false and where the Ct max is between 4.1 and 4.9 with Very Low bacterial load, the PPV of a false result increases to 85.7%. For the false RIFS results by LPA, the majority 11/23 (47.8%) were due to technical errors. In 6/23 (26.1%) it was due to mixed infection and in 2/23 (8.7%) there was laboratory mix up. In the remaining 4/23 (17.4%) the cause could not be determined and mixed infection or a laboratory mix up could not be excluded.
Discordant genotypic/phenotypic results: RIF resistance was detected in 1502 patients by LPA, of which 169 (11.3%) had a miscellaneous mutation. In addition, a further 21 isolates were selected from “Part 1” of the study, where sequencing confirmed that the rpoB mutation was not one of the high level / high confidence rpoB mutations. A total of 178 isolates had both MIC and rpoB sequencing results. In our study 140/178 (78.7%) isolates with miscellaneous rpoB mutations (n=158) or previously described disputed rpoB mutations (n=20) had MIC values ranging from ≤0.0625 µg/ml to 1.0 µg/ml. An MIC >1.0 µg/m was determined for 38/178 (21.3%) that would have tested RIFR by MGIT DST.
Conclusion: Arising from this study is a laboratory based guideline that is now used within NHLS TB laboratories detailing steps on how to detect possible false RIFR results by Xpert MTB/RIF and on how to troubleshoot discordant Xpert RIFR and LPA RIFS results. A database has been created from the results obtained in this study that lists specific rpoB mutations and their corresponding MIC value and has the potential to assist clinicians in individualizing the patient TB treatment regimen.
Description
Keywords
Reference:
Ghebrekristos, Y. 2018. Characterization of Mycobacterium tuberculosis isolates with discordant rifampicin susceptibility test results. University of Cape Town.