Quality assurance programme for drug susceptibility testing of Mycobacterium tuberculosis in the WHO/IUATLD Supranational Reference Laboratory Network: five rounds of proficiency testing, 1994-1998.

SETTING: Quality assurance for the WHO/IUATLD Global Tuberculosis Drug Resistance Surveillance Programme.
OBJECTIVE: To implement an ongoing proficiency-testing programme for drug susceptibility testing (DST) of Mycobacterium tuberculosis within the WHO/IUATLD Supranational Reference Laboratories Network (SRLN).
DESIGN: Five culture panels, each consisting of 10 duplicate drug-susceptible and drug-resistant clinical isolates (100 strains) of M. tuberculosis were tested for resistance to streptomycin (SM), isoniazid (INH), rifampicin (RMP) and ethambutol (EMB). DST procedures included the proportion, absolute concentration and resistance ratio methods, as well as the radiometric BACTEC 460 method.
RESULTS: The efficiency, sensitivity and specificity of M. tuberculosis DST as well as the intra-laboratory reproducibility showed that the laboratories tested susceptibility to RMP and to INH very reliably, with values ranging from 97% to 99%. The testing of SM and EMB was less dependable, with values ranging from 90% to 95%. The sensitivity of testing of EMB increased from 60% in Round 1 to 98% in Round 5, without a concomitant decrease in specificity.
CONCLUSIONS: This study has shown that regular proficiency testing can significantly improve the quality of DST, even in the most sophisticated TB laboratories. Mean DST efficiency levels of 92% for both SM and EMB and 97% and 99% for INH and RMP, respectively, are proposed as reasonable performance goals for the SRL network. Efficiency, consistently lower than these values, would require remedial action. Efficiency levels lower than mean -1 standard error, i.e., 80% for SM and EMB, 89% for INH and 95% for RMP, should always be considered as sub-standard performance for DST.