JOURNAL CLUB · ARTICLE OF THE WEEK
Whole-Body Hypothermia for Neonatal Encephalopathy in Preterm Infants 33-35 Weeks' Gestation
Faix RG, Laptook AR, Shankaran S, et al. JAMA Pediatrics. 2025;179(4):396-406. A randomised clinical trial.
PRACTICE-CHANGING
TRIAL
No reduction in death or disability
Primary outcome 35% (cooled) vs 29% (normothermia); adjusted RR 1.11 (95% CrI 0.74-2.00).
Signal towards harm
Bayesian analysis: 74% probability of increased death or disability with cooling.
Mortality numerically higher
Death 20% (cooled) vs 12% (normothermia); adjusted RR 1.38 (0.79-2.85); 87% probability of increased death.
Safety signals with cooling
More severe hyponatraemia (9% vs 0%), hyperglycaemia, and more ventilator days.
💡WHY THIS MATTERS
- Cooling is standard care for HIE at ≥36 weeks, but evidence below 36 weeks is lacking
- Off-label cooling of preterm infants has been increasing despite minimal evidence
- Preterm infants may be more vulnerable to cooling harms (intracranial haemorrhage, NEC, coagulopathy, shock)
- This is the first RCT to test cooling at 33-35 weeks
- The findings do not support routine cooling in this group
✅STRENGTHS
- First randomised trial in 33-35 week infants
- Rigorous multicentre NICHD NRN methodology, modelled on prior cooling trials
- Long-term follow-up at 18-22 months with blinded, standardised assessment
- Bayesian design suited to a small, hard-to-recruit population
- Control arm used targeted normothermia, avoiding harmful hyperthermia
⚠️LIMITATIONS
- Small sample (168) with wide credibility intervals - underpowered for a definitive answer
- Unblinded intervention
- Slow recruitment over >5 years; most infants were 34 weeks, few at 33 weeks
- Published correspondence has raised data and analysis concerns - results are debated
- Caution generalising to all preterm HIE
Practice implications
NSW ACI guidance limits therapeutic hypothermia to infants ≥35 weeks' gestation (≥1800g, started within 6 hours of birth). This trial found no benefit - and a possible harm signal - from cooling 33-35 week infants, consistent with current NSW practice of not cooling below that threshold (most of the trial population was 33-34 weeks). It supports confining cooling of 33-34 week infants to a clinical trial, and adds a note of caution about the borderline 35-week group that sits at the edge of eligibility.
Does this change your practice for cooling infants <36 weeks?
How do you interpret a Bayesian "probability of harm" compared with a traditional p-value and confidence interval?
Should 35-week infants be grouped with term infants or with the 33-34 week group?
Is 168 infants enough to change practice, or do we need more data?
JAMA Pediatrics 🧾TRIAL REGISTRY
NCT01793129 💬CORRESPONDENCE
Editorial debate 🩺NSW ACI GUIDELINE
HIE & cooling
Take-home message: In late-preterm infants (33-35 weeks) with HIE, whole-body hypothermia did not improve survival or neurodevelopment at 18-22 months, and Bayesian analysis favoured harm (74% probability of increased death or disability; 87% probability of increased death). The trial does not support routine cooling below 36 weeks.
For educational purposes only. Journal club appraisal - figures paraphrased from the published trial; read the full article for complete data. Always align management to current ANZCOR/NRP guidelines and your local SCN/NICU or NETS protocols.