[Efficacy of PCR-microwell plate hybridization method (Amplicor Mycobacterium) for detection of M. tuberculosis, M. avium and/or M. intracellulare in clinical specimens].

Recently, a new kit to detect and identify mycobacteria in clinical specimens was developed by Japan Roche Co. Limited. The new method is based on amplification of DNA of mycobacteria in clinical specimens by PCR and hybridization of amplified DNA by microwell plate hybridization method, which is the "Amplicor Mycobacteria, Roche, (AMP-M)". Cooperative study was organized with 15 tuberculosis hospitals and institutions throughout Japan, and 349 clinical specimens from newly admitted tuberculosis patients and/or suspects were collected during July and August, 1993. All the specimens were examined by smear microscopy (Ziehl-Neelsen's staining), culture on Ogawa egg media, culture on variant 7H9 liquid media and by AMP-M. Excluding 25 specimens which had failed to identify the species of mycobacteria because of contamination, disability to multiply on the transplanted solid media and so on, the results of the examinations in 324 specimens consisting of 167 specimens from previously untreated cases and those of 157 specimens from previously treated cases were analysed. Main results obtained were as follows; 1. Of 70 smear positive specimens from previously untreated cases, culture positive on Ogawa media and 7H9 media, and by AMP-M positive were 59 (84.3%), 61 (87.1%) and 66 (94.3%), respectively. Of 97 smear negative specimens, culture positive were 20 (20.6%), 22 (22.7%) and 27 (27.8%), respectively. The AMP-M showed the highest positive rate in both groups. 2. The sensitivity and the specificity of AMP-M in previously untreated cases were calculated by assuming that positive on Ogawa and/or variant 7H9 media is "positive". The sensitivity was 95.8% (68/71) and the specificity was 94.8% (91/96) for M. tuberculosis in previously untreated cases. The sensitivity and the specificity for M. avium and M. intracellulare were all 100%, although the numbers observed were small. 3. So-called false positive of the AMP-M were observed in 5 cases out of 96 culture negatives on both Ogawa and variant 7H9 media. However, all 5 cases were positive by repeated AMP-M, 3 become culture positive later, and another 2 showed clinical findings consistent with tuberculosis. Hence, the authors considered that the false positive rate of the AMP-M method is to be very low in previously untreated cases. 4. Of 86 smear positive cases with history of previous chemotherapy, the positive culture on Ogawa media, variant 7H9 media and that by AMP-M method were 64 (74.4%), 77 (89.5%) and 85 (98.8%), respectively. In the smear negative cases, culture positive was 10 out of 71 (14.1%), 13 (18.3%) and 24 (33.8%), respectively.