Abstract
Rapid, accurate and reliable assays are required for timely detection of drug-resistant tuberculosis and early initiation of second-line TB treatment as well as to minimize transmission of resistant strains. This study assessed diagnostic performance characteristics of two rapid molecular assays, line probe assay (LPA) and GeneXpert (MTB/RIF), in the detection rifampicin monoresistance using the phenotypic proportion method on Lowenstein−Jensen media as the gold standard. This study involved a total of 357 isolates, 74 rifampicin-resistant and 283 rifampicin-susceptible, collected at the Central Tuberculosis Reference Laboratory (CTRL) in Dar es Salaam, Tanzania, between 2016 and 2019. Sensitivity, specificity and positive and negative predictive values were used to assess the performance characteristics of the two assays while kappa coefficient was used to determine agreement of test results. The receiver operating curve (ROC) was used to determine the discriminatory ability of the test in distinguishing resistant and susceptible TB isolates. Our results showed that GeneXpert had sensitivity, specificity and positive and negative predictive values of 93.2, 82.7, 58.5 and 97.9%, respectively; the corresponding performance for LPA was 86.5, 97.5, 90.1 and 96.5%, respectively. Compared with conventional phenotypic DST results, GeneXpert had a moderate agreement (kappa 0.621, p < 0.001), while LPA had high agreement (0.853, p < 0.001). LPA showed an accuracy of 95.2% compared to GeneXpert’s 84.9%. ROC curve depicted the ability of the tests to distinguish rifampicin-sensitive and rifampicin-resistant strains to be 87.9% for GeneXpert and 92.0% for LPA. Our results indicate the superiority of LPA over GeneXpert regarding detection of rifampicin monoresistance. However, logistic challenges such as longer turnaround time and need for skilled laboratory personnel may limit its use.