The new child-friendly fixed dose combinations (FDCs) were introduced at Port Moresby General Hospital, Papua New Guinea, in 2016 for the first-line treatment of children (aged 15 years) with tuberculosis (TB) who weighed 25 kg. Objective: To describe the characteristics and outcomes for children treated with the new FDCs, and to identify risk factors for unfavourable treatment outcomes. Design: This was a retrospective cohort study of all children treated for TB with the FDCs from August 2016 to August 2017. Results: Of 713 children included, 488 (68%) were diagnosed with pulmonary TB. Only 6 (0.8%) TB cases were bacteriologically confirmed and human immunodeficiency virus (HIV) status was known in 50%. Treatment outcomes were favourable in 425 (60%) children. Of 288 children with unfavourable outcomes, there were 242 (84%) with loss to follow-up (LTFU) and 25 (8.4%) were known to have died. Children who were severely underweight (weight-for-age Z score -3) on presentation were at greater risk of LTFU compared to children of normal weight on multivariable analysis (aRR 1.3, 95%CI 1.0-1.6, P  0.05). Conclusion:Alternative models of care to decrease LTFU during treatment are needed, including integration with nutritional support. Improving diagnosis through microbiological confirmation of TB and HIV are major challenges to be addressed. Public Health Action (PHA)The voice for operational research.Published by The Union (www.theunion.org), PHA provides a platform to fulfil its mission, 'Health solutions for the poor'. PHA publishes high-quality scientific research that provides new knowledge to improve the accessibility, equity, quality and efficiency of health systems and services.