Adriana Hristea, Dan Otelea, Anca Macri, Cristian Baicus, Olga Moldovan, Mihaela Tinischi, Simona Paraschiv, Ruxandra Moroti, Victoria Arama, Adrian Streinu-Cercel

Objectives The objectives of our study were 1) to define risk factors for antimicrobial-resistant Mycobacterium tuberculosis (MTB) 2) to evaluate the use of real-time PCR-based technique for the prediction of phenotypic resistance of MTB. Methods We performed a prospective study in which we considered all 120 patients with smear-positive tuberculosis who were admitted to two clinical wards of the Marius Nasta Institute of Pneumology within a period of one year. Only data from 103 patients who underwent bacteriological investigations, including sensitivity tests, and who had filled out a questionnaire was analyzed. Fourteen patients with culture-confirmed tuberculosis resistant to any first line drug were compared to patients with fully susceptible disease. Sixty-seven MTB strains (26 drug-resistant and 41 drug-susceptible) were tested using a method recommended for the Light Cycler platform. For rifampin resistance, two regions of rpoB gene were targeted, while for identification of isoniazid resistance, we searched for mutations in katG and inhA genes. The susceptibility testing was performed using the absolute concentration method. Results We noted that while MTB resistance tends to be lower in new cases (3.7%), in patients with retreatment this resistance is much higher (44%), especially multidrug resistance (MDR), which accounts for 36% of cases of retreatment vs. 1.2% of new cases. Drug resistance was significantly associated only with previous antituberculous treatment (OR 19.56, CI 95% (3.96-94.41, p<0.005).The sensitivity and specificity of the rapid detection of mutations for isoniazid were 96% (95% CI: 88-100), respectively 95% (95% CI: 89-100), while for the rifampin resistance, sensitivity and specificity were 92% (95% CI: 81-100%), and 74% (95% CI: 61-87%) respectively. In our isolates the main resistance mechanism to isoniazid was related to changes in the katG gene that encodes catalase. Using this method, the best accuracy for genotyping compared to phenotypic resistance testing was obtained in what concerned detecting isoniazid resistance mutations. We found that for rifampin resistance the correlation between the predicted and observed phenotype was less than satisfactory. Conclusion MTB resistance was significantly associated only with previous antituberculous treatment. Real-time PCR assays represent a valuable diagnostic tool but they are not yet completely satisfactory for the detection of drug-resistant MTB.

Keywords: Mycobacterium tuberculosis,risk factors,resistance,Real-Time PCR