Objective: To evaluate the performance of MIRUs (mycobacterial interspersed repetitive units) genotyping alone, IS6110-RFLP (IS6110 restriction fragment length polymorphism) genotyping alone and their combination applied in the molecular epidemiological study of Mycobacterium tuberculosis (MTB) isolates circulating in rural China.
Methods: A cross-sectional study was designed to collect MTB isolates from the TB patients registered in local TB dispensaries of Deqing county and Guanyun county from 2004 to 2005. The proportion method was used to determine drug susceptibility of MTB isolates to the first line anti-TB drugs (isoniazid, rifampin, ethambutol and streptomycin). The Beijing family of MTB was identified by Spoligotyping. All isolates were genotyped by MIRUs alone, IS6110-RFLP alone and their combination.
Results: Of the 351 studied MTB isolates, 243 (69.2%) had the genotypes that belonged to the Beijing family, and 223 (63.5%) were resistant to at least 1 anti-TB drug, including 53 (15.1%) resistant to isoniazid and rifampin simultaneously or multidrug resistant (MDR). The heterogeneity of 12 MIRUs loci differed from 0.76 in MIRU26 to 0.003 in MIRU2. MIRUs alone identified 235 genotypes (HGI = 0.9317), including 46 "clusters" containing 162 isolates and 189 "unique" pattern/isolates. Thirty-eight isolates comprised the largest MIRUs defined cluster and presented MIRUs type 2233 2517 3533. MIRUs based clusters were further identified by IS6110-RFLP into 28 subgroups containing 80 isolates (HGI = 0.9989). In comparison, IS6110-RFLP determined 267 genotypes from all isolates, including 46 "clusters" containing 130 isolates (HGI = 0.9684) and 221 "unique" pattern/isolates. All the 46 IS6110-RFLP defined clusters could also be further identified by MIRUs into 31 subgroups (HGI = 0.9992). The performance of MIRUs prior to IS6110-RFLP combination was comparable to that of IS6110-RFLP prior to MIRUs combination, especially in Beijing genotype strain (HGI: 0.9930 vs 0.9933) and MDR-TB strains (HGI: 0.9965 vs 0.9963).
Conclusion: For feasibility, cost and discriminatory power, MIRUs prior to IS6110-RFLP combination is more suitable for the massive epidemiological investigation of MTB in rural China.