www.thelancet.com/infection Published online December 19, 2019 https://doi.org/10.1016/S1473-3099(19)30575-4 1 Articles Lancet Infect Dis 2019 Published Online December 19, 2019 https://doi.org/10.1016/ S1473-3099(19)30575-4 See Online/Comment https://doi.org/10.1016/ S1473-3099(19)30627-9 Respiratory Epidemiology and Clinical Research Unit, Montreal Chest Institute, McGill International TB Centre, McGill University Health Centre Research Institute (Prof D Menzies MD, J R Campbell PhD, Prof A Trajman MD, C Valiquette CNA, A Benedetti PhD, F Fregonese PhD, J C Johnston MD) and Department of Epidemiology and Biostatistics (Prof D Menzies, J R Campbell, A Benedetti), McGill University, Montréal, QC, Canada; Department of Biomedical Sciences, Division of Pharmacology and Therapy, Faculty of Medicine, Universitas Padjadjaran, Bandung, Indonesia (Prof R Ruslami MD); Federal University of Rio de Janeiro, Rio de Janeiro, Brazil (Prof A Trajman); Faculty of Medicine, University of British Columbia, Vancouver, BC, Canada (V J Cook MD, J C Johnston); Provincial TB Services, British Columbia Centre for Disease Control, Vancouver, BC, Canada (V J Cook, J C Johnston); National Hospitalier Universitaire de Pneumo- Phtisiologie, Cotonou, Benin (M Adjobimey MD); and University of Alberta, Edmonton, AB, Canada (L Eisenbeis MSc) Correspondence to: Prof Dick Menzies, McGill University, Montréal, QC H4A 3S5, Canada dick.menzies@mcgill.ca Introduction In the absence of new strategies, tuberculosis elimination will require identification and treatment of a substantial proportion of the 1·7 billion people with latent tuberculosis infection. 1 A key part of WHO’s End TB Strategy, latent tuberculosis infection treatment has gained momentum with the goal of initiating treatment in 30 million patients by 2022. 2 Current diagnostics are quite sensitive for latent tuberculosis infection but have poor predictive value in identifying the few individuals who will progress to tuberculosis disease. 3 Therefore, its treatment must be safe above all else. Yet, safety has been the Achilles heel of latent tuber- culosis infection treatment. The risk of adverse events, particularly hepatotoxicity, with most current first-line tuberculosis drugs, are orders of magnitude higher than other drugs routinely used in primary care. 4 Daily isoniazid, the most widely used treatment, was reported to cause fatal hepatotoxicity shortly after its widespread introduction for latent tuberculosis infection in 1971. 5,6 In addition, a 2-month regimen of daily rifampicin and pyrazinamide was recommended in 2000, but this recommendation was withdrawn in 2001 7 after more widespread use led to cases of fatal hepatotoxicity. Despite these concerns, 9 months of daily isoniazid remains the most widely used regimen for latent tuberculosis infection treatment in many high-resource settings. Adverse events in adults with latent tuberculosis infection receiving daily rifampicin or isoniazid: post-hoc safety analysis of two randomised controlled trials Jonathon R Campbell, Anete Trajman, Victoria J Cook, James C Johnston, Menonli Adjobimey, Rovina Ruslami, Lisa Eisenbeis, Federica Fregonese, Chantal Valiquette, Andrea Benedetti, Dick Menzies Summary Background An important problem limiting treatment of latent tuberculosis infection is the occurrence of adverse events with isoniazid. We combined populations from phase 2 and phase 3 open-label, randomised controlled trials, to establish risk factors for adverse events during latent tuberculosis infection treatment. Methods We did a post-hoc safety analysis based on data from two open-label, randomised controlled trials done in health-care facilities in Australia, Benin, Brazil, Canada, Ghana, Guinea, Indonesia, Saudi Arabia, and South Korea. Participants were consenting adults (aged ≥18 years) with a positive latent tuberculosis infection diagnostic test, indication for treatment, and without contraindications to rifampicin or isoniazid. Patients were centrally randomly assigned 1:1 to 4 months of daily 10 mg/kg rifampicin or 9 months of daily 5 mg/kg isoniazid. The primary outcome evaluated was adverse events (including grade 1–2 rash and all events of grade 3–5) resulting in permanent discontinuation of study medication and judged possibly or probably related to study drug by a masked, independent, three-member adjudication panel (trial registration: NCT00170209; NCT00931736). Findings Participants were recruited from April 27, 2004, up until Jan 31, 2007 (phase 2), and Oct 1, 2009, up until Dec 31, 2014 (phase 3). The safety populations for each group comprised 3205 individuals receiving isoniazid and 3280 receiving rifampicin. Among those receiving isoniazid, 86 (2·7%) of 3205 had grade 1–2 rash or any grade 3–5 adverse events, more than the 50 (1·5%) of 3280 who had these events with rifampicin (risk difference –1·2%, 95% CI –1·9 to –0·5). Age was associated with adverse events in adults receiving isoniazid. Compared with individuals aged 18–34 years, the adjusted odds ratio (OR) for adverse events was 1·8 (95% CI 1·1–3·0) for individuals aged 35–64 years and 3·0 (1·2–6·8) for individuals aged 65–90 years. With rifampicin, adverse events were associated with inconsistent medication adherence (adjusted OR 2·0, 1·1–3·6) and concomitant medication use (2·8, 1·5–5·2), but not age, with an adjusted OR of 1·1 (0·6–2·1) for individuals aged 35–64 years and 1·7 (0·5–4·7) for individuals aged 65–90 years. One treatment-related death occurred in the isoniazid group. Interpretation In patients without a contraindication, rifampicin is likely to be the safest latent tuberculosis infection treatment option. With more widespread use of rifampicin, rare, but serious adverse events might be seen. However, within these randomised trials, rifampicin was safer than isoniazid and adverse events were not associated with older age. Therefore, rifampicin should become a primary treatment option for latent tuberculosis infection based on its safety. Funding Canadian Institutes of Health Research. Copyright © 2019 Elsevier Ltd. All rights reserved.