Design: school-age follow-up of a multicentre, double-blind, placebo-controlled randomised trial (hPOD)
Population: 1,067 children aged 6-7 years, born at 35 weeks or more and 2.2 kg or more with at least one risk factor for transitional neonatal hypoglycaemia, across 9 New Zealand hospitals
Primary outcome: neurocognitive impairment at 6-7 years, defined as more than 1 SD below the normative mean on 1 or more of 7 NIH Toolbox items
Effect: 59% vs 57% (adjusted risk difference 3%; 95% CI -3% to 9%; P = .36); exploratory: emotional-behavioural difficulty 24% vs 18% (aRD 7%; 1-12) and low psychosocial function 17% vs 12% (aRD 6%; 1-10)

The hPOD trial gave a single 0.2 g/kg buccal dose of dextrose gel or placebo at one hour of age to newborns at 35 weeks or more with a risk factor for transitional hypoglycaemia — prematurity, maternal diabetes, small or large for gestational age, or a birth weight outside 2.5 to 4.5 kg. This follow-up assessed 1,067 of those children at 6 to 7 years, mostly at school.

Neurocognitive impairment was 59% in the dextrose gel group and 57% in the placebo group, an adjusted risk difference of 3% with an interval spanning zero. The intervention did not deliver the long-term benefit it was designed to test.

Two exploratory outcomes moved the other way. Emotional-behavioural difficulty was reported in 24% of the dextrose gel group against 18% of placebo, and low psychosocial function in 17% against 12%, both with intervals excluding zero. The authors are appropriately cautious about these, but they are not the sort of finding to set aside.

The conclusion the investigators draw is the one to act on: current evidence does not support routine prophylactic dextrose gel to prevent transitional neonatal hypoglycaemia. This is distinct from treatment of established hypoglycaemia with dextrose gel, which is a different question with its own evidence and is not challenged here.

Stop using prophylactic dextrose gel routinely in at-risk but normoglycaemic newborns
Continue treating established neonatal hypoglycaemia as before — this trial did not test that
Keep risk-factor-based glucose monitoring; the alternative to prophylaxis is surveillance, not nothing
Where a unit protocol mandates prophylactic gel, bring this follow-up to the next guideline review
Support feeding and thermal care, which remain the interventions with the least to argue against them

Don't overread it

The emotional-behavioural and psychosocial findings were exploratory outcomes among eleven; they raise a concern rather than establish harm.

The statistics, in plain English

The primary outcome, an adjusted risk difference of 3% with a 95% confidence interval of -3% to 9% and P = .36, is a clear null. The two adverse signals come from eleven exploratory outcomes, and with eleven comparisons one or two crossing significance by chance is expected — that is why they are labelled exploratory. Their intervals are also close to zero at the lower bound (1% for both), meaning the effect could be very small. Set against that, both point in the same direction and concern related domains, which is harder to explain by chance than two unrelated findings. Note also that 57-59% of children in both arms met the impairment definition, which reflects how the threshold was set rather than a catastrophic outcome in this cohort.