Objective: To evaluate the feasibility of melting curve assay technologies in quality assessment of drug susceptibility test (DST) on Mycobacterium tuberculosis.
Methods: During May 2009 to September 2012, a total of 2 204 MTB isolates were collected in Shanghai Songjiang District and 5 other county level TB institutes. DST results of these isolates were verified by the melting curve assay. Isolates with inconsistent phenotype and genotype results were retested and sequenced for any drug resistance mutations.
Results: Evaluation at Songjiang District showed high consistency of the melting curve assay with rifampicin and isoniazid DST results (kappa = 0.97 and 0.99, respectively). Consistency of the two methods were 97.6% (41/42), 100% (757/757) and 99.9% (788/799) in rifampicin resistant, susceptible, and total cases, respectively. As for isoniazid resistance detection, the consistency were 100% (72/72), 99.9% (726/727), and 99.9% (798/799) in resistant, susceptible, and total cases, respectively. Blinded re-examination of isolates with inconsistent genotype and phenotype resistance confirmed the good reproducibility of melting curve assay, which yielded the same results as the first test. Quality assessment at institutes A,B, C,D and E, however, showed consistency of only 87.5% (21/24), 80.8% (21/26), 87.5% (14/16), 82.9% (29/35), 60.9% (14/23) in rifampicinresistant cases, and 88.2% (30/34), 65.6% (21/32), 85.0% (17/20), 68.3% (41/60), 62.5% (15/24) in isoniazid resistant cases. As the reliability of DST results significantly varied among different areas, our finding indicated that there were misdiagnosed drug resistant cases in some field sites.
Conclusion: The melting curve assay for MTB DST is fast, easy to use, reproducible, and able to identify mistakes in daily work. Therefore, it is a suitable tool for the quality assessment of conventional DST results.