Surgery trial finds uncertain mortality effects of oxygen level and antioxidants
Neither oxygen levels nor antioxidants produced a statistically clear difference in deaths, hospital admissions or heart attacks. A treatment interaction needs further investigation.
Based on the published abstract. The full paper may contain additional methods, results and limitations.
The 30-second takeaway
Neither treatment comparison showed statistically significant differences in mortality, hospital admissions or heart attacks. Mortality estimates were compatible with benefit or harm, so the results do not establish equivalence. An interaction between oxygen level and antioxidant assignment raised a separate question, but it does not establish that either treatment improves outcomes.
Mortality confidence intervals were wide. The abstract’s editorial comment calls the interaction hypothesis-generating, so it remains a question for further trials rather than proof of a reliable treatment difference.
What researchers found
Hazard ratio for death from any cause
Oxygen fraction 0.80 versus 0.30 over 1 year: estimated death hazard was 1.46 times as high, but the interval allows lower or higher hazard (p = 0.23).
Understanding the comparison
The trial tested oxygen levels and antioxidants together, using medical records to assess outcomes. A hazard ratio compares the rate of dying at a given time; it does not represent an absolute probability of death.
How to read the interaction
The abstract reports more deaths among patients given higher oxygen with placebo and a statistically significant interaction. That pattern does not establish that antioxidants protect against oxygen-related harm or settle the uncertain overall treatment comparisons.
The original publication
Effects of Hyperoxia and Antioxidants on Mortality, Hospital Admissions, and Myocardial Infarction After Noncardiac Surgery: 1-Year Follow-Up of a Randomized Controlled Trial.
Loft FC, Holse C, Aasvang EK et al.
Acta Anaesthesiol Scand · 2025
- PubMed ID
- 41055553
- Record checked
AI-assisted research and writing. This explains one selected publication; it is not a complete review of everything known. Our approach.
One more question, understood.
Keep track of the research you’ve explored.