Identification of the best MIRU-VNTR set technique for MDR-TB and XDR-TB surveillance in Lisbon, Portugal

Multidrug and extensively drug resistant tuberculosis (MDR/XDR-TB) remains a serious threat to TB control worldwide. Portugal registers a high number of MDR-TB and XDR-TB cases particularly in Lisbon, the capital city. The majority of MDR- and XDR-TB cases that circulates in Lisbon belongs either one od the two groups of genetically close strains, know as Lisboa family and Q1 cluster. These strains are responsible for more than 80% of MDR-TB cases; 50% of which are also XDR-TB (2008). Given the high prevalence and considering the drug resistant profile of such strains, we aimed to investigate the most discriminatory technique for MDR- and XDR-TB surveillance. For this purpose, 74 clinical Mycobacterium tuberculosis isolates collected in Lisbon Health Region were genotyped by 12, 15 and 24-loci Mycobacterial interspersed units – variable number of tandem repeats (MIRU-VNTR). Twenty-two isolates were susceptible and 54 were resistant to at least one drug (34 MDR-TB, 15 of which were XDR-TB). We verified that MIRU-VNTR analysis based on 12, 15 and 24 loci sets allowed the clustering of 22 (66.7%), 20 (60.6%) and 17 (51.5%) MDR-TB isolates; 13 (92.9%), 12 (85.7%) and 11 (78.6%) of these were also XDR-TB, respectively. It was also noticed that, as the number of analysed loci increased, the Lisboa3 and Q1 isolates that differ by 1 or 2 loci were gradually excluded from the clusters. In order to evaluate the clustering of MDR-TB and XDR-TB isolates, we analysed the discriminatory power of all MIRU-VNTR sets, applying the Simpson’s Diversity Index. As expected, the 24 loci set analysis presented the highest discriminatory index (D) in comparison with both 12 and 15 loci sets (D. MIRU 24= 0.971; D. MIRU15=0.964; D.MIRU12=0.931). The small difference observed between D.MIRU15 and D. MIRU24, leads to the conclusion that 15 loci MIRU-VNTR is a powerful discriminatory tool. Interestngly, although with a lower discriminatory power, the 12-loci MIRU-VNTR analysis was enough to cluster the XDR-TB isolates into 2 major clusters, Lisboa 3 and Q1 cluster, corresponding to 92.9% of XDR-TB cases in Lisbon Health Region. Despite having a lower discriminatory power, we conclude that, in such settings, 15 loci MIRU-VNTR has a suficient discriminatory power for MDR-TB and XDR-TB surveillance and that 12 loci MIRU-VNTR is a useful method to ascertain a common origin between drug resistant isolates that have undergone subsequent genetic diversification