A liberal supplemental oxygen approach is recommended for all severely injured trauma patients despite limited evidence. Liberal oxygen administration may cause oxidative stress. The aim of this study was to investigate the effect of an early restrictive oxygen strategy versus a liberal oxygen strategy in adult trauma patients on biomarkers of oxidative stress within 48 h of hospital admission.This was a single-centre, sub-study of an international, randomised controlled trial TRAUMOX2. In TRAUMOX2, patients were randomised shortly after trauma to a restrictive oxygen strategy (arterial oxygen saturation target of 94 %) or a liberal oxygen strategy (12-15 L of oxygen per minute or fraction of inspired oxygen of 0.6-1.0) for 8 h. Blood samplings were performed at four time points within 48 h after randomisation: upon arrival at the trauma centre, and at eight, 24, and 48 h post-randomisation. The primary outcome was the plasma level of malondialdehyde (MDA) 24 h post-randomisation. Secondary outcomes were numerous, and included the level of MDA at other time points, superoxide dismutase (SOD) at all time points, 30-day mortality, and major respiratory complications.The sub-study included 90 adult trauma patients. The median MDA levels at 24 h post-randomisation was 60.9 μM (95 % CI 49.5 to 73.4) in the restrictive oxygen group and 56.7 μM (95 % CI 46.9 to 68.2) in the liberal oxygen group, corresponding to a difference of -4.2 μM (95 % CI -19.8 to 10.5; P = 0.35). No significant differences were found in MDA or SOD at the other time points either. Neither did we find a significant difference in 30-day mortality or major respiratory complications.In this sub-study of the TRAUMOX2 trial, no significant differences were found in biomarkers of oxidative stress between a restrictive oxygen strategy and liberal oxygen strategy in adult trauma patients.