Development and evaluation of a model to predict hypersensitivity reaction when using 3HP for TB infection.

Authors

S Ghosh, F L Garden, K L Boi, T D Duong, L T Hien, T P Luong, T N Nguyen, T T Nguyen, N N Tran, T A Nguyen, G J Fox, G B Marks

Published in

The international journal of tuberculosis and lung disease : the official journal of the International Union against Tuberculosis and Lung Disease. Volume 29. Issue 12. Pages 541-547. Nov 28, 2025.

Abstract

A 12-dose weekly regimen of isoniazid and rifapentine (3HP) is effective treatment for TB infection (TBI), but some patients experience hypersensitivity reactions (HRs). We investigated whether preceding mild symptoms predict subsequent HRs, helping to improve safety by allowing early cessation of 3HP.We conducted a case-control study with 10,092 participants who commenced 3HP as TBI treatment in Ca Mau, Vietnam. Cases were those who developed HRs, defined as specified symptoms and/or signs. Controls were randomly selected from those who completed treatment without HRs. Participants completed weekly symptom questionnaires, and we developed a predictive model for subsequent HRs using this data and baseline characteristics, which was validated in a separate cohort of 1,581 participants.Of those initiating 3HP, 776 (7.7%) developed HRs. Cases were older and more likely female than controls. Symptoms in the 2 weeks prior to HRs were more common in cases (69.7%) than controls (26.8%). Key predictive symptoms were fever, 'feeling sick', nausea, rash, aches, and dizziness. The multivariate predictive model showed good discrimination (c-statistic 0.78). At a cut-off point giving 80% sensitivity, the specificity was 61.3%. In the validation cohort, the sensitivity was 85.1% and specificity 57.8%.This predictive model, based on baseline characteristics and reported symptoms, may allow early cessation of 3HP and avert subsequent HRs..

PMID:
42396739
Bibliographic data and abstract were imported from PubMed on 03 Jul 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.