Within-Host Heterogeneity ofMycobacterium tuberculosisInfection Is Associated With Poor Early Treatment Response: A Prospective Cohort Study

Abstract

The clinical management of tuberculosis is a major challenge in southern Africa. The prevalence of within-host genetically heterogeneous Mycobacterium tuberculosis infection and its effect on treatment response are not well understood. We enrolled 500 patients with tuberculosis in KwaZulu-Natal and followed them through 2 months of treatment. Using mycobacterial interspersed repetitive units-variable number of tandem repeats genotyping to identify mycobacterial heterogeneity, we report the prevalence and evaluate the association of heterogeneity with treatment response. Upon initiation of treatment, 21.1% of participants harbored a heterogeneous M. tuberculosis infection; such heterogeneity was independently associated with a nearly 2-fold higher odds of persistent culture positivity after 2 months of treatment (adjusted odds ratio, 1.90; 95% confidence interval, 1.03-3.50).

Evaluation of the impact of polyclonal infection and heteroresistance on treatment of tuberculosis patients

“…In contrast, Cohen et al . showed that the negative effect of polyclonal infections on the response to early treatment appeared to be independent of a relationship with drug resistance in TB patients in KwaZulu-Natal, South Africa15. This difference may be related to the follow-up period; which ended at six and two months in our and Cohen’s study, respectively.…”

M. tuberculosis microvariation is common and is associated with transmission: analysis of three years prospective universal sequencing in England

“…Examining the bases mapped to each site using per-base binomial testing (see Methods) shows sample-C from patient-1 contains multiple mixed positions (M-sites) at the sites differing between A and B. This kind of within-patient heterogeneity, which is reported to be associated with impaired treatment response, 10, 13, 14 is not evident using standard methods of consensus base calling, using which C appears identical to (that is, has a zero pairwise SNV distance from) sample A. Consequently, we decided to develop automated methods detecting mixed TB infection using our routinely sequenced data.…”

“…Secondly, although we observed microvariation to be more commonly detected in lineages 1-3 than in the most common UK lineage-4, the biological basis for this needs further study. Thirdly, although mixed infection has been associated with delayed response to therapy in high incidence settings, 10 we were unable to determine whether this occurs in the UK as the UK’s national TB surveillance scheme does not collect suitable data on treatment response; it records only rates of treatment completion within 12 months of diagnosis, or within 24 months for cases with rifampicin or multidrug resistance.…”

“…15, 16 Co-circulating strains are commonly closely related, 15 and co-circulation is ongoing in hyperendemic settings, in which mixed isolates are reported in 10-30% of cases. 8-10 In either the chronic infection or superinfection scenario, mixed infection might be associated with raised transmission risk, although data to support this is not available at present.…”

“…Such infections are currently of clinical interest since response to treatment is known to be delayed in individuals with mixed infection. 10, 13, 14 Mixed infection may reflect either the development of intra-host diversity during chronic infection 6 , or co-infection by more than one strain circulating in the community in which the affected individual resides. 15, 16 Co-circulating strains are commonly closely related, 15 and co-circulation is ongoing in hyperendemic settings, in which mixed isolates are reported in 10-30% of cases.…”