Comparison of quality of drug susceptibility testing of Mycobacterium tuberculosis by absolute concentration method and BACTEC 460

Abstract number: 1733_1139

Baranov A., Presnova S., Nizovtseva N., Trekin I., Gvozdovskaja L., Yendaurova L., Mariandyshev A., Bjune G., Heldal E., Mannsaker T.

The study has been carried out in 4 administrative territories of the North-western federal region of Russia. Namely, in Archangel oblast, Murmansk oblast, Republic of Karelia and Republic of Komi.

Objectives: The aim of our study was to find rates of agreement of DST between absolute concentration method, routinely performed at the Russian laboratories, and the reference method – BACTEC 460, performed at the National Reference Laboratory for Mycobacteria of the Norwegian Institute of Public Health (NIPH).

Laboratory at the Archangel Regional Anti-tuberculosis dispensary has been involved in DST quality assurance programmes on different levels for many years. The main component of these programmes was the collaboration with the reference laboratory at NIPH. Data on this collaboration from previous years is not presented in this manuscript.

Methods: In total, 88 strains of M. tuberculosis were analysed and results for these strains were compared for both methods. In addition to the computing of the overall percent agreement for each drug, kappa test has been used. The following arbitrary divisions for interpreting the results of kappa test have been used: under 0.20 is negligible; from 0.21 to 0.40 is minimal; from 0.41 to 0.60 is fair; from 0.61 to 0.80 is good; and over 0.81 is excellent.

Results: The highest rates of agreement were observed for isoniazid and rifampicin, 95.5% (CI; 88.8–98.8) and 97.7% (CI; 92.0–99.7), respectively. They also had the highest k-values 0.90 (CI; 0.80–1.0) and 0.96 (CI; 0.89–1.0), which correspond to excellent agreement beyond chance between investigating methods. Streptomycin and kanamycin had agreements in 93.2% (CI; 85.8–97.5) and 90.9% (CI; 82.9–96.0) in all tested strains, respectively, with k-values 0.85 (CI; 0.73–0.97) and 0.78 (CI; 0.64–0.92). These k-values mean from good to excellent agreements. The lowest rate of agreement was found for ethambutol, in 85.2% (CI; 76.1–91.9) of cases. This is consistent with relatively low k-value for this drug 0.70 (CI; 0.56–0.85), which means from fair to good agreement between investigating methods.

Conclusion: These findings suggest that, currently, the results of absolute concentration method for isoniazid and rifampicin can serve as a reliable diagnostic tool. This is very important for diagnosis and treatment of MDR tuberculosis.

Session Details

Date: 31/03/2007
Time: 00:00-00:00
Session name: European Society of Clinical Microbiology and Infectious Diseases
Location: ICC, Munich, Germany
Presentation type:
Back to top