Resolution of blood RNA signatures fails to discriminate sputum culture status after eight weeks of tuberculosis treatment

Abstract

SummaryBackgroundThere is concerted effort to reduce the burden of 6 months antimicrobial treatment for tuberculosis (TB). Early treatment cessation at 8 weeks is effective for most but incurs increased risk of disease relapse. We tested the hypothesis that blood RNA signatures of TB disease or C-reactive protein (CRP) measurements discriminate microbiological cure after 8 weeks of treatment, as a pre-requisite for a biomarker to stratify risk of relapse.MethodsWe identified blood RNA signatures of TB disease or cure by systematic review. We evaluated CRP measurements and blood RNA signatures that could be reproduced in genome-wide transcriptomic data from a previously reported longitudinal dataset in pulmonary TB, spanning samples collected pre-treatment, at 2 and 8 weeks of treatment, and after 2 years of follow up. In our primary analysis, we tested discrimination of sputum culture positivity at 8 weeks by contemporary blood RNA and CRP measurements using area under the receiver operating characteristic curve (AUROC) analysis. In secondary analyses, we tested the relationship between biomarker measurements and time to culture positivity as a surrogate for bacterial load in sputum culture positive cases at 8 weeks, and discrimination of sputum culture status at 8 weeks by biomarker measurements at any other time point.FindingsWe evaluated 12 blood RNA signatures. Blood RNA signature scores normalised over time from TB treatment initiation. 11/44 cases with available blood RNA, CRP and sputum culture results, were sputum culture positive at 8 weeks of treatment. None of the 12 blood RNA signature scores tested achieved statistically significant discrimination between sputum culture-positive vs. negative patients at this time point, with AUROC point estimates of 0.48-0.61. CRP achieved the best AUROC of 0.69 (95% confidence interval 0.52-0.87). None of the contemporary biomarker measurements correlated with bacterial load, and no measurements pre-treatment or at 2 weeks discriminated sputum culture status at 8 weeks.InterpretationThe current repertoire of blood RNA signatures of TB and CRP will not provide host response surrogates of microbiological cure to support cessation of TB treatment at 8 weeks. Decoupling of blood transcriptional host-response from the presence of viable bacteria is indicative of subpopulations ofMycobacterium tuberculosisable to colonise the respiratory tract without triggering a detectable immune response.Research in contextEvidence before this studyWe performed a systematic review, using comprehensive terms for “tuberculosis”, “transcriptional” and “biomarker” with no language or date restrictions in Medline on October 4, 2023. Many studies have described normalisation of blood RNA signatures during the course of tuberculosis treatment. Five studies have evaluated blood RNA signatures as a test of microbiological cure after completion of 6 months of treatment. However, there is growing interest in their application as a test of cure to support shortened treatment regimens. The performance of one blood RNA signature has been reported to provide modest discrimination of contemporary sputum culture status at 8 weeks of treatment among HIV co-infected patients with recurrent tuberculosis. We found no reports of whether these findings are generalisable to other blood RNA signatures or to HIV negative patients with their first episode of tuberculosis, who are most likely to be candidates for shortened treatment regimens.Added value of this studyTo our knowledge, we provide the first evaluation and comparison of multiple blood RNA signatures of tuberculosis for discrimination of microbiological cure after 8 weeks of tuberculosis therapy among HIV negative patients. 12 previously validated blood RNA signatures of tuberculosis identified by systematic review underwent head-to-head evaluation, alongside blood C-reactive protein measurement as an alternative biomarker of disease, to determine whether they discriminated contemporary sputum culture status after 8 weeks of treatment among 44 HIV negative patients with smear-positive drug-sensitive tuberculosis enrolled to a previously reported randomised controlled trial of adjunctive vitamin D therapy. None of the blood RNA signatures showed statistically significant discrimination of contemporary sputum culture status after 8 weeks tuberculosis treatment, or quantitative relationships with sputum bacterial load among sputum culture positive cases. Importantly, most sputum culture positive cases at this time point, showed contemporary blood RNA signature scores within the normal range.Implications of all the available evidenceAssuming that microbiological cure is a pre-requisite for early cessation of antimicrobial treatment for tuberculosis, the current repertoire of blood RNA signatures of tuberculosis does not provide host response surrogates of microbiological cure to support introduction of 8-week treatment regimens. Therefore, there is a need for further discovery and validation of new biomarkers to support risk stratification for truncated therapy both for research and clinical practice applications. The lack of association between blood transcriptomic signatures and sputum culture status after 8 weeks of treatment suggests the existence of microbial sub-populations that do not trigger a host response. Whether this reflects a state of latency or bacterial persistence in immune privileged compartments requires further investigation.