Surveillance of pyrazinamide susceptibility among multidrug-resistant Mycobacterium tuberculosis isolates from Siriraj Hospital, Thailand

Background: Susceptibility testing of pyrazinamide (PZA) against Mycobacterium tuberculosis is difficult to perform because the acidity of culture medium that is required for drug activity also inhibits the growth of bacteria. In Thailand, very limited information has been generated on PZA resistance, particularly among multidrug-resistant tuberculosis (MDR-TB) isolated from Thailand. Only two studies on PZA susceptibility among Thai M. tuberculosis strains have been reported; one used a pyrazinamidase assay, and the other used the BACTEC 460 TB for PZA susceptibility testing. In this study, we determined the percentage of strains possessing pyrazinamide resistance among pan-susceptible M. tuberculosis and MDR-TB isolates by using the pyrazinamidase assay, BACTEC MGIT 960 PZA method and pncA sequencing, and assessed the correlation in the data generated using these methods. The type and frequency of mutations in pncA were also determined. Results: Overall, 150 M. tuberculosis isolates, consisting of 50 susceptible and 100 MDR-TB isolates, were tested for PZA susceptibility by BACTEC MGIT 960 PZA, the pyrazinamidase assay and pncA sequencing. The study indicated PZA resistance in 6% and 49% of susceptible and MDR-TB isolates, respectively. In comparison to the BACTEC MGIT 960 PZA, the PZase assay showed 65.4% sensitivity and 100% specificity, whereas pncA sequencing showed 75% sensitivity and 89.8% specificity. Twenty-four mutation types were found in this study, with the most frequent mutation (16%) being His71Asp. Of these mutations, eight have not been previously described. The Ile31Ser and Ile31Thr mutations were found both in PZA susceptible and resistant isolates, suggesting that mutation of this codon might not play a role on PZA resistance. Conclusions: Our findings suggest that phenotypic susceptibility testing is still essential for the detection of PZA resistance, especially for MDR-TB isolates. Some mutations were not associated with resistance and could lead to misinterpretation of the genotypic methods. This information could be helpful for clinicians in managing tuberculosis patients and frequencies, and the types of pncA mutations should offer baseline information on PZA resistance.