JOURNAL CLUB · GO DEEPER
60% or 30% oxygen to start resuscitation of babies under 29 weeks (TORPIDO 30/60)
Oei JL, Kirby A, Travadi J, et al. JAMA. 2026;335(6):523-530. doi:10.1001/jama.2025.23327. An unblinded, multinational randomised trial.
NO DIFFERENCE
UNDER 29 WK
No difference in the primary outcome
Death or brain injury by 36 weeks was 46.9% (60%) vs 47.8% (30%). RR 0.98 (95% CI 0.89-1.09).
No difference in death or brain injury alone
Death 15.4% vs 15.8%; brain injury 42.4% vs 43.2%. No difference in severe IVH or PVL.
More ventriculomegaly with 60%
6.4% vs 2.4%. A secondary finding of uncertain significance, but not reassuring for higher starting oxygen.
Many went to 100% anyway
FiO2 was escalated to 1.0 in 38% (60% group) vs 41% (30% group). 94% received the allocated starting FiO2.
The verdict
For babies under 29 weeks, starting at 60% rather than 30% did not reduce death or brain injury. Consistent across gestation, steroid exposure, sex and consent type.
WHY THIS MATTERS
- The best starting oxygen for extremely preterm babies has been uncertain, and practice varies widely
- Low starting oxygen often leaves babies below saturation targets; high oxygen risks oxidative injury
- Earlier trials were too small to answer death or brain injury
- Brain injury-free survival is the outcome that matters most to families
- Babies under 29 weeks are occasionally born at Gosford before NETS arrives
STRENGTHS
- Large (1469 babies) and recruited to its planned sample size
- Individually randomised, with a hard, patient-important primary outcome
- Multinational, including middle-income settings, so broadly generalisable
- Imaging reported by masked radiologists
- High adherence (94%) to the allocated starting FiO2
LIMITATIONS
- Unblinded clinicians, who may have titrated oxygen differently
- Powered for a large effect (24% to 32% brain injury-free survival); a smaller benefit could be missed
- Did not test 40% or 100% as a starting point
- Could not examine interactions with cord clamping, PEEP strategy or device
- Brain injury defined on ultrasound and includes any-grade IVH; no neurodevelopmental follow-up yet
Practice implications
Extremely preterm births are rare in a special care nursery, but they happen, often before NETS arrives. TORPIDO 30/60 shows that starting at 60% rather than 30% oxygen does not improve survival without brain injury, and may increase ventriculomegaly. Starting low (around 30%) and titrating to target saturations remains appropriate - follow your local and ANZCOR guidance. What matters most: get a pulse oximeter on early, watch the heart rate, titrate promptly to targets and ventilate effectively. About 4 in 10 babies needed 100% oxygen whichever group they were in.
What starting oxygen does our unit use for a 26-week baby, and where is that written down?
About 40% of babies were escalated to 100% oxygen. What does that tell us about the starting point versus titration?
How should we interpret the higher rate of ventriculomegaly with 60%, given a neutral primary outcome?
Taken with AIROPLANE, what delivery-room oxygen approach would you teach a new RMO?
JAMA 2026ATTENDING A BIRTH
The Hub pageAIR vs OXYGEN
Resair 2 & landmark trialsJOURNAL CLUB
Article of the week & archive
Take-home message: In babies born at 23-28 weeks, starting resuscitation at 60% rather than 30% oxygen did not change death or brain injury by 36 weeks (about 47% in both groups), and ventriculomegaly was more common with 60%. Start low, titrate promptly to target saturations, and ventilate effectively.
For educational purposes only. Journal club appraisal - figures paraphrased from the published trial; read the full article for complete data. Always align practice to your local guidelines.