RESUMO
OBJECTIVE: To determine the accuracy of drug-susceptibility testing (DST) for isoniazid, rifampicin, ethambutol and streptomycin in a provisional network of 22 regional laboratories in Italy. METHODS: Methods, definitions and reference Mycobacterium tuberculosis strains were derived from the WHO/IUATLD Global Project on Anti-tuberculosis Drug Resistance Surveillance. The laboratories were selected based on technical skills required by the project, the number of DST performed annually and geographic localisation. The results (sensitive/resistant strain) were compared with the gold standard (global project results). Sensitivity (ability to detect true resistance), specificity (ability to detect true susceptibility), positive predictive values for resistance and susceptibility, efficiency and reproducibility were calculated in two rounds. RESULTS: Eighteen of 22 laboratories completed the first round of proficiency testing for the four drugs. Sensitivity was 76.6%, specificity 97.2%, predictive value of a resistant test 89.8% and of a susceptible test 86.8%, efficiency 87.8% and reproducibility 92.8%. A second round was performed by all those laboratories that did not achieve > or = 90% agreement with the results of the Global Project. Overall, after the second round, all the parameters except specificity improved, exceeding 90%. CONCLUSIONS: A network of 15 regional laboratories that fulfil the quality criteria for determining the susceptibility of M. tuberculosis to the four primary antituberculosis drugs was established in Italy.