Transcriptional profiles predict treatment outcome in patients with tuberculosis and diabetes at diagnosis and at two weeks after initiation of anti-tuberculosis treatment

Doorn, C.L.R. van; Eckold, C.; Ronacher, K.; Ruslami, R.; Veen, S. van; Lee, J.S.; Kumar, V.; Kerry-Barnard, S.; Malherbe, S.T.; Kleynhans, L.; Stanley, K.; Hill, P.C.; Joosten, S.A.; Crevel, R. van; Wijmenga, C.; Critchley, J.A.; Walzl, G.; Alisjahbana, B.; Haks, M.C.; Dockrell, H.M.; Ottenhoff, T.H.M.; Vianello, E.; Cliff, J.M.; TANDEM Consortium

doi:doi:10.1016/j.ebiom.2022.104173

Background Globally, the tuberculosis (TB) treatment success rate is approximately 85%, with treatment failure, relapse and death occurring in a significant proportion of pulmonary TB patients. Treatment success is lower among people with diabetes mellitus (DM). Predicting treatment outcome early after diagnosis, especially in TB-DM patients, would allow early treatment adaptation for individuals and may improve global TB control.Methods Samples were collected in a longitudinal cohort study of adult TB patients from South Africa (n = 94) and Indonesia (n = 81), who had concomitant DM (n = 59), intermediate hyperglycaemia (n = 79) or normal glycaemia/no DM (n = 37). Treatment outcome was monitored, and patients were categorized as having a good (cured) or poor (failed, recurrence, died) outcome during treatment and 12 months follow-up. Whole blood transcriptional profiles before, during and at the end of TB treatment were characterized using unbiased RNA-Seq and targeted gene dcRT...Show moreBackground Globally, the tuberculosis (TB) treatment success rate is approximately 85%, with treatment failure, relapse and death occurring in a significant proportion of pulmonary TB patients. Treatment success is lower among people with diabetes mellitus (DM). Predicting treatment outcome early after diagnosis, especially in TB-DM patients, would allow early treatment adaptation for individuals and may improve global TB control.Methods Samples were collected in a longitudinal cohort study of adult TB patients from South Africa (n = 94) and Indonesia (n = 81), who had concomitant DM (n = 59), intermediate hyperglycaemia (n = 79) or normal glycaemia/no DM (n = 37). Treatment outcome was monitored, and patients were categorized as having a good (cured) or poor (failed, recurrence, died) outcome during treatment and 12 months follow-up. Whole blood transcriptional profiles before, during and at the end of TB treatment were characterized using unbiased RNA-Seq and targeted gene dcRT-MLPA.Findings We report differences in whole blood transcriptome profiles, which were observed before initiation of treatment and throughout treatment, between patients with a good versus poor TB treatment outcome. An eight-gene and a 22-gene blood transcriptional signature distinguished patients with a good TB treatment outcome from patients with a poor TB treatment outcome at diagnosis (AUC = 0.815) or two weeks (AUC = 0.834) after initiation of TB treatment, respectively. High accuracy was obtained by cross-validating this signature in an external cohort (AUC = 0.749).Interpretation These findings suggest that transcriptional profiles can be used as a prognostic biomarker for treatment failure and success, even in patients with concomitant DM. Copyright (C) 2022 The Authors. Published by Elsevier B.V.Show less

Leiden University Scholarly Publications

View statistics

Documents

In Collections

Transcriptional profiles predict treatment outcome in patients with tuberculosis and diabetes at diagnosis and at two weeks after initiation of anti-tuberculosis treatment