Whole-genome sequencing to establish relapse or re-infection with Mycobacterium tuberculosis: a retrospective observational study

Abstract

Background: Recurrence of tuberculosis after treatment makes management difficult and is a key factor for determining treatment efficacy. Two processes can cause recurrence: relapse of the primary infection or re-infection with an exogenous strain. Although re-infection can and does occur, its importance to tuberculosis epidemiology and its biological basis is still debated. We used whole-genome sequencing-which is more accurate than conventional typing used to date-to assess the frequency of recurrence and to gain insight into the biological basis of re-infection.

Methods: We assessed patients from the REMoxTB trial-a randomised controlled trial of tuberculosis treatment that enrolled previously untreated participants with Mycobacterium tuberculosis infection from Malaysia, South Africa, and Thailand. We did whole-genome sequencing and mycobacterial interspersed repetitive unit-variable number of tandem repeat (MIRU-VNTR) typing of pairs of isolates taken by sputum sampling: one from before treatment and another from either the end of failed treatment at 17 weeks or later or from a recurrent infection. We compared the number and location of SNPs between isolates collected at baseline and recurrence.

Findings: We assessed 47 pairs of isolates. Whole-genome sequencing identified 33 cases with little genetic distance (0-6 SNPs) between strains, deemed relapses, and three cases for which the genetic distance ranged from 1306 to 1419 SNPs, deemed re-infections. Six cases of relapse and six cases of mixed infection were classified differently by whole-genome sequencing and MIRU-VNTR. We detected five single positive isolates (positive culture followed by at least two negative cultures) without clinical evidence of disease.

Interpretation: Whole-genome sequencing enables the differentiation of relapse and re-infection cases with greater resolution than do genotyping methods used at present, such as MIRU-VNTR, and provides insights into the biology of recurrence. The additional clarity provided by whole-genome sequencing might have a role in defining endpoints for clinical trials.

Funding: Wellcome Trust, European Union, Medical Research Council, Global Alliance for TB Drug Development, European and Developing Country Clinical Trials Partnership.

Figure 1

Distribution of the case outcomes for study patients based on sequencing quality data, sequence comparison, and clinical evaluation

Figure 2

Pair-wise distances between pairs of isolates from the same patient For each patient pair, the calculated pair-wise distance is based on the number of high quality base differences between the samples.

Figure 3

MIRU-VNTR loci differing between pairs of isolates from the same patient (A) Number of loci differing between pairs of isolates from the same patient with relapse, re-infection, mixed infection, and single isolated positives identified by whole genome sequencing. (B) Comparison of differences detected by whole genome sequence data and MIRU-VNTR; red=relapse, blue=re-infection. MIRU-VNTR=mycobacterial interspersed repetitive unit-variable number of tandem repeat.