Background: Delays in identifying multidrug-resistant tuberculosis (MDR-TB) contribute to higher TB morbidity and mortality, and ongoing transmission. The line-probe assay (LiPA) is a rapid, commercially available polymerase chain reaction based assay that detects most mutations in the rpoB gene for rifampicin (RMP) resistance. We validated and compared this assay with conventional drug susceptibility testing (DST).
Methods: We re-cultured a random sample of stored isolates known to be either RMP-resistant or RMP-susceptible according to DST (proportion method). We performed a blinded comparison between LiPA and conventional DST. Genetic sequencing of the rpoB gene was performed on RMP-resistant isolates and discordant results.
Results: We tested 79 RMP-resistant and 64 RMP-susceptible strains. Concordance of LiPA with DST was 94%. For detecting RMP resistance, LiPA sensitivity was 90% and specificity was 100%. Molecular analysis of possible false-negative isolates by LiPA revealed an absence of mutations in the rpoB gene. RMP resistance was a good proxy for MDR-TB, as 66 (93%) of 71 RMP-resistant isolates were also isoniazid-resistant.
Conclusion: The LiPA provided rapid results that were highly predictive of RMP resistance and MDR-TB. False-negatives occurred, but only among isolates with mutations in regions not assessed by LiPA. Performance and cost-effectiveness should be evaluated in patients during routine program conditions.