[EVALUATION STANDARD OF EXTERNAL QUALITY ASSESSMENT PROGRAMME FOR DRUG SUSCEPTIBILITY TESTING OF MYCOBACTERIUM TUBERCULOSIS IN JAPANESE LABORATORIES: PROFICIENCY TESTING IN 2004-2010].

Abstract

Objective: To analyze the results of the external quality assessments (EQA) for anti-tuberculosis drug susceptibility testing (DST) and to set-up its rational passing criterion.

Method: Each participating laboratory in EQA performed DST, and the sensitivity, specificity, agreement (efficiency) and kappa coefficient were calculated from the results. We analysed the data of seven EQA results for DST from 2004 to 2010.

Results: A total of 20, 20, 10, 5, 10, 10, and 10 strains of M. tuberculosis with known susceptibility were sent to each participating laboratory in 2004, 2005, 2006, 2007, 2008, 2009, and 2010, respectively. The total of participating laboratories was 564. Each laboratory was asked to perform DST with its routine methods and reported 25,100 test results in these seven years. The laboratories showed relatively high specificity than sensitivity, and an improving sensitivity through the years. Sixteen laboratories participated the EQA continuously, and the sensitivity and specificity to isoniazid (INH), rifampicin (RFP), streptomycin (SM) and ethambutol (EB) were 0.999 (95% CI 0.992-1.000) and 0.998 (95% CI 0.991-1.000), 0.985 (95% CI 0.973-0.992) and 0.997 (95% CI 0.989-0.999), 0.932 (95% CI 0.912-0.948) and 0.977 (95% CI 0.962-0.986), and 0.965 (95% CI 0.947-0.977) and 0.978 (95% CI 0.966-0.986), respectively.

Discussion: The analyses revealed that the accuracy of DST for INH and RFP was highly maintained throughout the years. However, SM showed a high unevenness of performance quality and required situational considerations for evaluation. In conclusion, the EQA for DST would require a minimum number of 10 strains for each assessment, and INH and RFP should show over 95% of sensitivity and specificity with over 90% of efficiency to SM and EB as passing remark.