Back

A whole-blood transcriptional signature associated with obstructive post-tuberculosis lung disease

Newby, R. E.; Kijaro, L.; Dill-McFarland, K. A.; Njagi, L. N.; Murithi, W. B.; Mwongera, Z.; Agata, M.; Byers, J.; Peterson, G. J.; Fennelly, K. P.; Nduba, V.; Horne, D. J.; Zifodya, J. S.; Hawn, T. R.

2026-03-12 infectious diseases
10.64898/2026.03.11.26348051 medRxiv
Show abstract

BackgroundPost-tuberculosis (TB) lung disease (PTLD) affects approximately 50% of persons with pulmonary TB. We recently discovered whole blood transcriptional signatures associated with PTLD. We examined whether a minimum gene signature predicts PTLD as a clinically useful biomarker. MethodsWe prospectively enrolled 301 treatment naive adults with newly diagnosed pulmonary TB (PTB) (cohort A). We collected whole blood at 0 and 6-month visits, isolated RNA, and measured a modified MTB Host Response (HR) signature (mHR) based on expression of DUSP3, GBP5, and TMBIM6. We recorded spirometry at 6 (n=216) and 12 months (n=210) after treatment initiation and examined the association of the mHR score with PTLD and Mtb aerosolization. We recruited household contacts of cohort A to compare mHR score with non-PTB participants (cohort B). FindingsmHR was associated with TB (p=4.15e-66) when compared to HHCs, treatment response (p=1.07e-53), and characteristics including CD4 count (p=0.003), bacillary load (p=3.02e-05), lung cavities (p=1.59e-04), and lung quadrants involved (p=3.87e-06). The mHR score was not associated with Mtb aerosolization. In total, 105 (50%) participants had PTLD at 12 months including 61 with restriction, 26 with obstruction, and 18 with mixed obstruction and restriction. Baseline mHR was associated with obstructive PTLD at both 6 (p=0.003) and 12 months (p=0.012) in bivariate and multivariate analyses. The mHR score was not associated with restrictive lung disease. InterpretationBaseline mHR was associated with obstructive PTLD at 6 and 12 months and may have applications in targeting treatment and prognostication.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.