False-positive PCR results due to sporadic cross-reactivity with non-tuberculous mycobacteria has been suspected earlier also with other NAAT systems [8, 22, 23]. As the technical validation

of the hyplex® TBC kit had indeed shown some unspecific binding for single Mycobacterium species, it would be possible also for BMS345541 the M. intracellulare. The second false-positive specimen originated from a case without a known MTB infection. It cannot be ruled out completely that very low amounts of MTB nucleic acids originating from an early TB infection may have led to positive PCR results with hyplex® TBC. Among smear-negative, culture-positive specimens, 34 out of 62 were not detected by hyplex® TBC. This was, at least in part, due to the fact that the cut-off has been SU5402 concentration increased from OD 0.200 to OD 0.400 in order to reduce the false-positive rate to a minimum. It would certainly be worth trying, whether the sensitivity could be increased by applying higher volumes of sample. Our evaluation was performed

with a sample volume of 10 μl, but theoretically sample volumes up to 40 μl can be applied. However, too much DNA may considerably reduce the effectiveness of a PCR and, in return, would lead to a higher rate of inhibition. The optimal volume of specimen needs to be check details determined in further investigations. Seven percents of smear-positive, culture-positive samples also escaped the detection by hyplex® TBC. It is unlikely that this was caused solely by too low amounts of MTB DNA, since most of these specimens yielded clearly positive smear microscopy results (at least between 10 and 50 acid fast bacilli per 100 fields) and re-assessment by CTM PCR gave positive results with 14

of 15 specimens. The hyplex® TBC PCR is based on target sequences of a house keeping gene. It can be speculated that missing of some of these TB samples by hyplex® TBC was related to single nucleotide polymorphisms within this gene. This question should be studied and the results may certainly help to optimise the oligonucleotide probes used in the kit. Conclusions Hyplex® TBC is an accurate and reliable NAAT assay for the direct Farnesyltransferase detection of MTB in respiratory and non-respiratory specimens. Similar to other commercial NAATs, the hyplex® TBC assay is impacted by the compromise between specificity and sensitivity: specificity is maximised at the cost of sensitivity. Compared to other commercial NAAT systems, the hyplex® TBC assay shows excellent specificity estimates but slightly lower sensitivity, in particular for smear-negative TB specimens. Also, when the assay is used as rapid confirmation test for smear-positive specimens one should be aware of the fact that a small percentage of TB infections may be not detected.