RESUMO
BACKGROUND: Mycobacterial interspersed repetitive unit-variable number tandem repeat (MIRU-VNTR) analysis is a recently developed method which could be suitable as a 'real-time' genotyping tool for Mycobacterium tuberculosis. METHODS: One hundred and thirty-four M. tuberculosis isolates were analysed using the reference method, IS6110-restriction fragment length polymorphism (RFLP), and by MIRU, alone and together with spoligotyping. RESULTS: MIRU reduced the genotyping turnaround time by 21 days. The discriminatory power (HGDI) for MIRU and RFLP was 0.978 and 0.989, respectively. RFLP clustered 41.8% of the isolates (17 clusters; 2-9 representatives), whereas MIRU increased the number and size of the clusters (57.5% of the isolates in 20 clusters; 2-14 representatives). With respect to the RFLP clusters, MIRU data showed full correlation in only 7/ 17 (41%) clusters and low correlation in 8/17 (47%) clusters. When MIRU and spoligotyping were considered together, the analysis fitted better with RFLP data: 1) 42.5% of the isolates were grouped in 20 clusters of 2-6 representatives, and 2) the number of clusters with full correlation with RFLP data increased to 11/17 and those with low correlation decreased to 2/17. CONCLUSION: MIRU-VNTR analysis showed low correlation with RFLP. The addition of spoligotyping to MIRU analysis fitted much better with RFLP analysis, although full correlation was still not achieved.