Reply to “Mycobacterial Interspersed Repetitive-Unit Locus PCR Amplification and Beijing Strains of Mycobacterium tuberculosis”†

We thank Luo et al. for their comments on our previous paper describing the usefulness of mycobacterial interspersed repetitive-unit (MIRU) PCR amplification in rapid diagnosis of Beijing lineage strain infections among pediatric tuberculosis patients (4). We agree that the negative results of Mtub2 amplification in Beijing strains might be explained by the deletion of RD105 in Beijing strains, given the location of the primers used. However, we do not think that this invalidates our previous conclusion that a negative Mtub2 amplification result is predictive of the presence of Beijing strains. The advantage of using this assay covering both the Mtub2 loci and RD105 is its ability to concurrently differentiate Beijing from non-Beijing strains and among non-Beijing lineage strains on the basis of a simple test. We agree that the RD105 deletion remains a reliable marker of Beijing strains. That is why we proposed combining the Mtub02 PCR assay with the deletiontargeted multiplex PCR (DTM-PCR) assay to achieve higher sensitivity and specificity instead of using any particular MIRU locus alone (4). We appreciate the efforts of Luo et al. to review the literature and share their findings and unpublished data. We are not surprised by their finding that the sensitivity of a specific MIRU-based assay to detect Beijing strains differed among different studies that had used isolates from different geographic regions and samples of different types (2, 3, 5, 6), given that it is well known that the coadaptation of Mycobacterium tuberculosis and the human host population in different environments has resulted in distinct lineages and sublineages of M. tuberculosis clinical strains (1) and that the Beijing family contains a diverse population of strains (5). Despite these variations, the data from the literature review by Luo et al. show high levels of sensitivity for three of the four reviewed MIRU loci in detection of Beijing strains in countries such as China and Russia, where Beijing strain infections are prevalent. It was with great interest that we read the statement of Luo et al. that the high specificities resulting from the use of MIRU loci to predict Beijing strains observed in our study were due to the high prevalence of the European-American lineage in our study region that is genetically distant from Beijing family and shows distinct MIRU profiles. While we recognize the potential impact of convergent evolution at some MIRU loci on their ability to differentiate the Beijing lineage from other lineages in some settings, we do not think it is appropriate to judge the broad utility of these loci based on extrapolation of sensitivity and specificity from published data without the consideration of potential bias resulting from small samples and the nature of the study samples (3, 4). Frequency distribution of different genetic lineages of M. tuberculosis can vary not only by geographic region but also by type of study sample (Table 1). Therefore, comparisons of the population structures of M. tuberculosis among different geographic regions as well as evaluations of the usefulness of different genotyping methods based on published data should be performed with caution.