Use of whole-genome sequencing to distinguish relapse from reinfection in a completed tuberculosis clinical trial

BackgroundWe previously described a tuberculosis chemotherapy trial in southern Africa using high-Dose rifapentine with moxifloxacin. Here we evaluate the utility of whole genome sequencing (WGS) as a way of inferring whether new infections in treated patients are relapses or reinfections, and compare it with MIRU-VNTR typing. This is the first completed trial in which WGS has been used in this way.MethodsDNA from 37 paired samples of Mycobacterium tuberculosis cultured from patients before and after treatment were typed using 24 locus MIRU-VNTR, and WGS. With WGS, the sequences were mapped against strain H37Rv, a phylogenetic tree was generated, and single nucleotide polymorphism (SNP) differences counted . FindingsWGS indicated that two pairs were problematic and were excluded. 32 of the remaining paired samples had a very low number of SNP differences (0-5; likely relapses). One pair proved to be a mixed infection that we also classified as a relapse. The remaining two pairs had very high SNP differences (>1000; likely reinfections). InterpretationWGS and MIRU-VNTR both differentiated relapses and reinfections, but we found WGS was technically easier, and provided significant extra information. In light of the low proportion of reinfections seen, we suggest that in standard chemotherapy trials, there is no cost benefit in typing the strains using either method. However for research interest, WGS was far more useful.