Multicentre quality control study for detection of Mycobacterium tuberculosis in clinical samples by nucleic amplification methods.

The aim of this study was to evaluate the laboratory performance of nucleic acid amplification tests (NATs) for detection of the Mycobacterium tuberculosis complex. A proficiency panel consisting of eight sputum specimens and four specimens diluted in phosphate-buffered saline (PBS) was sent to 82 laboratories in 23 countries by the Quality Control for Molecular Diagnostics (QCMD) TB programme. The performance of different NATs was analysed in combination with a questionnaire on the applied methods. Seventy-eight participants (95.2%) contributed a total of 85 evaluable data sets. The percentage of correct results on the eight sputum samples was 86.3% (586/679). Of the sputum specimens considered as 'smear-negatives' (650 CFU/250 micro L), only 61.2% (104/170) were reported positive. The percentage of correct results for the three scored PBS samples was 75.7% (193/255). The total number of false-positive results was 11 (4.3%); these were reported for seven (8.2%) of the 85 data sets. In 32 (37.6%) data sets an 'in-house' NAT method was used, and in 53 (62.4%) sets a commercial assay was tested. The percentage of data sets achieving correct results on all sputum samples was 35.3% and 37.8%, respectively. For the PBS samples this was 45.8% and 41.5%. Overall, the results of this study demonstrated that the performance of NATs for the detection of M. tuberculosis has improved since previous studies. The percentage of false-positives has decreased considerably. However, a large number of procedures still lack sufficient sensitivity for application to smear-negative samples.